prerelease: 0.93: investigate smp_ihk_os_panic_notifier

Change-Id: I997b41f80038603261de2e8232b6b8ca200cd8cd

CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required(VERSION 2.6)
+cmake_minimum_required(VERSION 3.11)
 
 if (NOT CMAKE_BUILD_TYPE)
 	set (CMAKE_BUILD_TYPE "Debug" CACHE STRING "Build type: Debug Release..." FORCE)
@@ -7,10 +7,10 @@ endif (NOT CMAKE_BUILD_TYPE)
 enable_language(C ASM)
 
 project(mckernel C ASM)
-set(MCKERNEL_VERSION "1.7.0")
+set(MCKERNEL_VERSION "1.7.1")
 
-# See "Fedora Packaging Guidlines -- Versioning"
-set(MCKERNEL_RELEASE "0.3")
+# See "Fedora Packaging Guidelines -- Versioning"
+set(MCKERNEL_RELEASE "0.93")
 
 set(CMAKE_MODULE_PATH ${CMAKE_SOURCE_DIR}/cmake/modules)
 # for rpmbuild
@@ -50,6 +50,64 @@ if (ENABLE_WERROR)
 	add_compile_options("-Werror")
 endif(ENABLE_WERROR)
 
+execute_process(COMMAND bash -c "ls -ld /proc/tofu/ 2>/dev/null | wc -l"
+  OUTPUT_VARIABLE PROC_TOFU OUTPUT_STRIP_TRAILING_WHITESPACE)
+
+if(PROC_TOFU STREQUAL "1")
+	option(ENABLE_TOFU "Built-in tofu driver support" ON)
+else()
+	option(ENABLE_TOFU "Built-in tofu driver support" OFF)
+endif()
+
+if(ENABLE_TOFU)
+	add_definitions(-DENABLE_TOFU)
+	set(KBUILD_C_FLAGS "${KBUILD_C_FLAGS} -DENABLE_TOFU")
+endif()
+
+# when compiling on a compute-node
+execute_process(COMMAND bash -c "grep $(hostname) /etc/opt/FJSVfefs/config/fefs_node1.csv 2>/dev/null | cut -d, -f2 | grep -o CN"
+  OUTPUT_VARIABLE FUGAKU_NODE_TYPE OUTPUT_STRIP_TRAILING_WHITESPACE)
+if(FUGAKU_NODE_TYPE STREQUAL "CN")
+	option(ENABLE_FUGAKU_HACKS "Fugaku hacks" ON)
+else()
+	option(ENABLE_FUGAKU_HACKS "Fugaku hacks" OFF)
+endif()
+
+if(ENABLE_FUGAKU_HACKS)
+	add_definitions(-DENABLE_FUGAKU_HACKS)
+	set(KBUILD_C_FLAGS "${KBUILD_C_FLAGS} -DENABLE_FUGAKU_HACKS")
+endif()
+
+# krm that mandates reserved memory amount >= available at boot time?
+execute_process(COMMAND bash -c "rpm -qi FJSVpxkrm-plugin-mckernel | awk '$1 == \"Version\" && $2 == \":\" { print $3 }'"
+  OUTPUT_VARIABLE KRM_VERSION OUTPUT_STRIP_TRAILING_WHITESPACE)
+message("KRM_VERSION: ${KRM_VERSION}")
+
+if(NOT "${KRM_VERSION}" STREQUAL "" AND "${KRM_VERSION}" VERSION_LESS_EQUAL 4.0.1)
+	option(ENABLE_KRM_WORKAROUND "krm workaround" ON)
+else()
+	option(ENABLE_KRM_WORKAROUND "krm workaround" OFF)
+endif()
+
+if(ENABLE_KRM_WORKAROUND)
+	add_definitions(-DENABLE_KRM_WORKAROUND)
+	set(KBUILD_C_FLAGS "${KBUILD_C_FLAGS} -DENABLE_KRM_WORKAROUND")
+endif()
+
+# SIGSTOP instead of SIGSEGV, additional IHK Linux kmsg
+option(ENABLE_FUGAKU_DEBUG "Fugaku debug instrumentation" OFF)
+
+if(ENABLE_FUGAKU_DEBUG)
+	add_definitions(-DENABLE_FUGAKU_DEBUG)
+	set(KBUILD_C_FLAGS "${KBUILD_C_FLAGS} -DENABLE_FUGAKU_DEBUG")
+endif()
+
+option(PROFILE_ENABLE "System call profile" ON)
+if(PROFILE_ENABLE)
+	add_definitions(-DPROFILE_ENABLE)
+	set(KBUILD_C_FLAGS "${KBUILD_C_FLAGS} -DPROFILE_ENABLE")
+endif()
+
 option(ENABLE_LINUX_WORK_IRQ_FOR_IKC "Use Linux work IRQ for IKC IPI" ON)
 if (ENABLE_LINUX_WORK_IRQ_FOR_IKC)
 	set(KBUILD_C_FLAGS "${KBUILD_C_FLAGS} -DIHK_IKC_USE_LINUX_WORK_IRQ")
@@ -136,17 +194,19 @@ if (NOT LIBIBERTY)
 	message(FATAL_ERROR "error: couldn't find libiberty")
 endif()
 
-# libdwarf-devel provides /usr/lib64/libdwarf.so
 find_library(LIBDWARF dwarf)
 
-# elfutils-devel provides /usr/include/dwarf.h
-find_library(LIBEBL ebl)
-
-if ((NOT LIBDWARF) OR (NOT LIBEBL))
+if (NOT LIBDWARF)
+	if (CMAKE_CROSSCOMPILING)
+		message(FATAL_ERROR "Could not find libdwarf.so, install libdwarf-devel to ${CMAKE_FIND_ROOT_PATH}")
+	endif()
 	message("WARNING: libdwarf will be compiled locally")
-	set(LIBDWARF LIBDWARF-NOTFOUND)
-	set(LIBEBL LIBEBL-NOTFOUND)
 	enable_language(CXX)
+else()
+	# Note that libdwarf-devel provides /usr/include/libdwarf/dwarf.h
+	# but elfutils-devel provides /usr/include/dwarf.h
+	# while mcinspect.c performs "#include <dwarf.h>"
+	find_path(DWARF_H dwarf.h PATH_SUFFIXES libdwarf)
 endif()
 
 if (ENABLE_QLMPI)
@@ -250,6 +310,11 @@ message("KBUILD_C_FLAGS: ${KBUILD_C_FLAGS}")
 message("MAP_KERNEL_START: ${MAP_KERNEL_START}")
 message("ENABLE_MEMDUMP: ${ENABLE_MEMDUMP}")
 message("ENABLE_PERF: ${ENABLE_PERF}")
+message("ENABLE_TOFU: ${ENABLE_TOFU}")
+message("ENABLE_FUGAKU_HACKS: ${ENABLE_FUGAKU_HACKS}")
+message("ENABLE_FUGAKU_DEBUG: ${ENABLE_FUGAKU_DEBUG}")
+message("ENABLE_KRM_WORKAROUND: ${ENABLE_KRM_WORKAROUND}")
+message("PROFILE_ENABLE: ${PROFILE_ENABLE}")
 message("ENABLE_RUSAGE: ${ENABLE_RUSAGE}")
 message("ENABLE_QLMPI: ${ENABLE_QLMPI}")
 message("ENABLE_UTI: ${ENABLE_UTI}")

NEWS.md

@@ -1,537 +0,0 @@
-=============================================
-What's new in version 1.7.0rc4 (Apr 15, 2020)
-=============================================
-
-----------------------
-McKernel major updates
-----------------------
-1. arm64: Contiguous PTE support
-2. arm64: Scalable Vector Extension (SVE) support
-3. arm64: PMU overflow interrupt support
-4. xpmem: Support large page attachment
-5. arm64 port: Direct access to Mckernel memory from Linux
-6. arm64 port: utility thread offloading, which spawns thread onto
-   Linux CPU
-7. eclair: support for live debug
-8. Crash utility extension
-9. Replace mcoverlayfs with a soft userspace overlay
-10. Build system is switched to cmake
-11. Core dump includes thread information
-
-------------------------
-McKernel major bug fixes
-------------------------
-1. shmobj: Fix rusage counting for large page
-2. mcctrl control: task start_time changed to u64 nsec
-3. mcctrl: add handling for one more level of page tables
-4. Add kernel argument to turn on/off time sharing
-5. flatten_string/process env: realign env and clear trailing bits
-6. madvise: Add MADV_HUGEPAGE support
-8. mcctrl: remove in-kernel calls to syscalls
-9. arch_cpu_read_write_register: error return fix.
-10. set_cputime(): interrupt enable/disable fix.
-11. set_mempolicy(): Add mode check.
-12. mbind(): Fix memory_range_lock deadlock.
-13. ihk_ikc_recv: Record channel to packet for release
-14. Add set_cputime() kernel to kernel case and mode enum.
-15. execve: Call preempt_enable() before error-exit
-16. memory/x86_64: fix linux safe_kernel_map
-17. do_kill(): fix pids table when nr of threads is larger than num_processors
-18. shmget: Use transparent huge pages when page size isn't specified
-19. prctl: Add support for PR_SET_THP_DISABLE and PR_GET_THP_DISABLE
-20. monitor_init: fix undetected hang on highest numbered core
-21. init_process_stack: change premapped stack size based on arch
-22. x86 syscalls: add a bunch of XXat() delegated syscalls
-23. do_pageout: fix direct kernel-user access
-24. stack: add hwcap auxval
-25. perf counters: add arch-specific perf counters
-26. Added check of nohost to terminate_host().
-27. kmalloc: Fix address order in free list
-28. sysfs: use nr_cpu_ids for cpumasks (fixes libnuma parsing error on ARM)
-29. monitor_init: Use ihk_mc_cpu_info()
-30. Fix ThunderX2 write-combined PTE flag insanity
-31. ARM: eliminate zero page mapping (i.e, init_low_area())
-32. eliminate futex_cmpxchg_enabled check (not used and dereffed a NULL pointer)
-33. page_table: Fix return value of lookup_pte when ptl4 is blank
-34. sysfs: add missing symlinks for cpu/node
-35. Make Linux handler run when mmap to procfs.
-36. Separate mmap area from program loading (relocation) area
-37. move rusage into kernel ELF image (avoid dynamic alloc before NUMA init)
-38. arm: turn off cpu on panic
-39. page fault handler: protect thread accesses
-40. Register PPD and release_handler at the same time.
-41. fix to missing exclusive processing between terminate() and
-    finalize_process().
-42. perfctr_stop: add flags to no 'disable_intens'
-43. fileobj, shmobj: free pages in object destructor (as opposed to page_unmap())
-44. clear_range_l1, clear_range_middle: Fix handling contiguous PTE
-45. do_mmap: don't pre-populate the whole file when asked for smaller segment
-46. invalidate_one_page: Support shmobj and contiguous PTE
-47. ubsan: fix undefined shifts
-48. x86: disable zero mapping and add a boot pt for ap trampoline
-49. rusage: Don't count PF_PATCH change
-50. Fixed time processing.
-51. copy_user_pte: vmap area not owned by McKernel
-52. gencore: Zero-clear ELF header and memory range table
-53. rpm: ignore CMakeCache.txt in dist and relax BuildRequires on cross build
-54. gencore: Allocate ELF header to heap instead of stack
-55. nanosleep: add cpu_pause() in spinwait loop
-56. init_process: add missing initializations to proc struct
-57. rus_vm_fault: always use a packet on the stack
-58. process stack: use PAGE_SIZE in aux vector
-59. copy_user_pte: base memobj copy on range & VR_PRIVATE
-60. arm64: ptrace: Fix overwriting 1st argument with return value
-61. page fault: use cow for private device mappings
-62. reproductible builds: remove most install paths in c code
-63. page fault: clear writable bit for non-dirtying access to shared ranges
-64. mcreboot/mcstop+release: support for regular user execution
-65. irqbalance_mck: replace extra service with service drop-in
-66. do_mmap: give addr argument a chance even if not MAP_FIXED
-67. x86: fix xchg() and cmpxchg() macros
-68. IHK: support for using Linux work IRQ as IKC interrupt (optional)
-69. MCS: fix ARM64 issue by using smp_XXX() functions (i.e., barrier()s)
-70. procfs: add number of threads to stat and status
-71. memory_range_lock: Fix deadlock in procfs/sysfs handler
-72. flush instruction cache at context switch time if necessary
-73. arm64: Fix PMU related functions
-74. page_fault_process_memory_range: Disable COW for VM region with zeroobj
-75. extend_process_region: Fall back to demand paging when not contiguous
-76. munmap: fix deadlock with remote pagefault on vm range lock
-77. procfs: if memory_range_lock fails, process later
-78. migrate-cpu: Prevent migration target from calling schedule() twice
-79. sched_request_migrate(): fix race condition between migration req and IRQs
-80. get_one_cpu_topology: Renumber core_id (physical core id)
-81. bb7e140 procfs cpuinfo: use sequence number as processor
-82. set_host_vma(): do NOT read protect Linux VMA
-
-===========================================
-What's new in V1.6.0 (Nov 11, 2018)
-===========================================
-
------------------------------------------------
-McKernel new features, improvements and changes
------------------------------------------------
-1. McKernel and Linux share one unified kernel virtual address space.
-   That is, McKernel sections resides in Linux sections spared for
-   modules.  In this way, Linux can access the McKernel kernel memory
-   area.
-2. hugetlbfs support
-3. IHK is now included as a git submodule
-4. Debug messages are turned on/off in per souce file basis at run-time.
-5. It's prohibited for McKernel to access physical memory ranges which
-   Linux didn't give to McKernel.
-6. UTI (capability to spawn a thread on Linux CPU) improvement:
-   * System calls issued from the thread are hooked by modifying
-     binary in memory.
-
----------------------------
-McKernel bug fixes (digest)
----------------------------
-#<num> below corresponds to the redmine issue number
-(https://postpeta.pccluster.org/redmine/).
-
-1. #926: shmget: Hide object with IPC_RMID from shmget
-2. #1028: init_process: Inherit parent cpu_set
-3. #995: Fix shebang recorded in argv[0]
-4. #1024: Fix VMAP virtual address leak
-5. #1109: init_process_stack: Support "ulimit -s unlimited"
-6. x86 mem init: do not map identity mapping
-7. mcexec_wait_syscall: requeue potential request on interrupted wait
-8. mcctrl_ikc_send_wait: fix interrupt with do_frees == NULL
-9. pager_req_read: handle short read
-10. kprintf: only call eventfd() if it is safe to interrupt
-11. process_procfs_request: Add Pid to /proc/<PID>/status
-12. terminate: fix oversubscribe hang when waiting for other threads on same CPU to die
-13. mcexec: Do not close fd returned to mckernel side
-14. #976: execve: Clear sigaltstack and fp_regs
-15. #1002: perf_event: Specify counter by bit_mask on start/stop
-16. #1027: schedule: Don't reschedule immediately when wake up on migrate
-17. #mcctrl: lookup unexported symbols at runtime
-18. __sched_wakeup_thread: Notify interrupt_exit() of re-schedule
-19. futex_wait_queue_me: Spin-sleep when timeout and idle_halt is specified
-20. #1167: ihk_os_getperfevent,setperfevent: Timeout IKC sent by mcctrl
-21. devobj: fix object size (POSTK_DEBUG_TEMP_FIX_36)
-22. mcctrl: remove rus page cache
-23. #1021: procfs: Support multiple reads of e.g. /proc/*/maps
-24. #1006: wait: Delay wake-up parent within switch context
-25. #1164: mem: Check if phys-mem is within the range of McKernel memory
-26. #1039: page_fault_process_memory_range: Remove ihk_mc_map_virtual for CoW of device map
-27. partitioned execution: pass process rank to LWK
-28. process/vm: implement access_ok()
-29. spinlock: rewrite spinlock to use Linux ticket head/tail format
-30. #986: Fix deadlock involving mmap_sem and memory_range_lock
-31. Prevent one CPU from getting chosen by concurrent forks
-32. #1009: check_signal: system call restart is done only once
-33. #1176: syscall: the signal received during system call processing is not processed.
-34. #1036 syscall_time: Handle by McKernel
-35. #1165 do_syscall: Delegate system calls to the mcexec with the same pid
-36. #1194 execve: Fix calling ptrace_report_signal after preemption is disabled
-37. #1005 coredump: Exclude special areas
-38. #1018 procfs: Fix pread/pwrite to procfs fail when specified size is bigger than 4MB
-39. #1180 sched_setaffinity: Check migration after decrementing in_interrupt
-40. #771, #1179, #1143 ptrace supports threads
-41. #1189 procfs/do_fork: wait until procfs entries are registered
-42. #1114 procfs: add '/proc/pid/stat' to mckernel side and fix its comm
-43. #1116 mcctrl procfs: check entry was returned before using it
-44. #1167 ihk_os_getperfevent,setperfevent: Return -ETIME when IKC timeouts
-45. mcexec/execve: fix shebangs handling
-46. procfs: handle 'comm' on mckernel side
-47. ihk_os_setperfevent: Return number of registered events
-48. mcexec: fix terminating zero after readlink()
-
-===========================================
-What's new in V1.5.1 (July 9, 2018)
-===========================================
-
------------------------------------------------
-McKernel new features, improvements and changes
------------------------------------------------
-1. Watchdog timer to detect hang of McKernel
-   mcexec prints out the following line to its stderr when a hang of
-   McKernel is detected.
-
-       mcexec detected hang of McKernel
-
-   The watchdog timer is enabled by passing -i <timeout_in_sec> option
-   to mcreboot.sh. <timeout_in_sec> specifies the interval of checking
-   if McKernel is alive.
-   Example: mcreboot.sh -i 600: Detect the hang with 10 minutes interval
-
-   The detailed step of the hang detection is as follows.
-   (1) mcexec acquires eventfd for notification from IHK and perform
-       epoll() on it.
-   (2) A daemon called ihkmond monitors the state of McKernel periodically
-       with the interval specified by the -i option. It judges that
-       McKernel is hanging and notifies mcexec by the eventfd if its
-       state hasn't changed since the last check.
-
-2. Documentation
-   man page: Installed directory is changed to <install_dir>/share/man
-
----------------------------
-McKernel bug fixes (digest)
----------------------------
-1.  #1146: pager_req_map(): do not take mmap_sem if not needed
-2.  #1135: prepare_process_ranges_args_envs(): fix saving cmdline
-3.  #1144: fileobj/devobj: record path name
-4.  #1145: fileobj: use MCS locks for per-file page hash
-5.  #1076: mcctrl: refactor prepare_image into new generic ikc send&wait
-6.  #1072: execve: fix execve with oversubscribing
-7.  #1132: execve: use thread variable instead of cpu_local_var(current)
-8.  #1117: mprotect: do not set page table writable for cow pages
-9.  #1143: syscall wait4: add _WALL (POSTK_DEBUG_ARCH_DEP_44)
-10. #1064: rusage: Fix initialization of rusage->num_processors
-11. #1133: pager_req_unmap: Put per-process data at exit
-12. #731: do_fork: Propagate error code returned by mcexec
-13. #1149: execve: Reinitialize vm_regions's map area on execve
-14. #1065: procfs: Show file names in /proc/<PID>/maps
-15. #1112: mremap: Fix type of size arguments (from ssize_t to size_t)
-16. #1121: sched_getaffinity: Check arguments in the same order as in Linux
-17. #1137: mmap, mremap: Check arguments in the same order as in Linux
-18. #1122: fix return value of sched_getaffinity
-19. #732: fix: /proc/<PID>/maps outputs a unnecessary NULL character
-
-===================================
-What's new in V1.5.0 (Apr 5, 2018)
-===================================
-
---------------------------------------
-McKernel new features and improvements
---------------------------------------
-1. Aid for Linux version migration: Detect /proc, /sys format change
-   between two kernel verions
-2. Swap out
-   * Only swap-out anonymous pages for now
-3. Improve support of /proc/maps
-4. mcstat: Linux tool to show resource usage
-
----------------------------
-McKernel bug fixes (digest)
----------------------------
-1. #727: execve: Fix memory leak when receiving SIGKILL
-2. #829: perf_event_open: Support PERF_TYPE_HARDWARE and PERF_TYPE_HW_CACHE
-3. #906: mcexec: Check return code of fork()
-4. #1038: mcexec: Timeout when incorrect value is given to -n option
-5. #943 #945 #946 #960 $961: mcexec: Support strace
-6. #1029: struct thread is not released with stress-test involving signal
-          and futex
-7. #863 #870: Respond immediately to terminating signal when
-              offloading system call
-8. #1119: translate_rva_to_rpa(): use 2MB blocks in 1GB pages on x86
-11. #898: Shutdown OS only after no in-flight IKC exist
-12. #882: release_handler: Destroy objects as the process which opened it
-13. #882: mcexec: Make child process exit if the parent is killed during
-          fork()
-14. #925: XPMEM: Don't destroy per-process object of the parent
-15. #885: ptrace: Support the case where a process attaches its child
-16. #1031: sigaction: Support SA_RESETHAND
-17. #923: rus_vm_fault: Return error when a thread not performing
-          system call offloading causes remote page fault
-18. #1032 #1033 #1034: getrusage: Fix ru_maxrss, RUSAGE_CHILDREN,
-                       ru_stime related bugs
-19. #1120: getrusage: Fix deadlock on thread->times_update
-20. #1123: Fix deadlock related to wait_queue_head_list_node
-21. #1124: Fix deadlock of calling terminate() from terminate()
-22. #1125: Fix deadlock related to thread status
-    * Related functions are: hold_thread(), do_kill() and terminate()
-23. #1126: uti: Fix uti thread on the McKernel side blocks others in do_syscall()
-24. #1066: procfs: Show Linux /proc/self/cgroup
-25. #1127: prepare_process_ranges_args_envs(): fix generating saved_cmdline to
-           avoid PF in strlen()
-26. #1128: ihk_mc_map/unmap_virtual(): do proper TLB invalidation
-27. #1043: terminate(): fix update_lock and threads_lock order to avoid deadlock
-28. #1129: mcreboot.sh: Save /proc/irq/*/smp_affinity to /tmp/mcreboot
-29. #1130: mcexec: drop READ_IMPLIES_EXEC from personality
-
---------------------
-McKernel workarounds
---------------------
-1. Forbid CPU oversubscription
-   * It can be turned on by mcreboot.sh -O option
-
-
-===================================
-What's new in V1.4.0 (Oct 30, 2017)
-===================================
-
------------------------------------------------------------
-Feature: Abstracted event type support in perf_event_open()
------------------------------------------------------------
-PERF_TYPE_HARDWARE and PERF_TYPE_CACHE types are supported.
-
-----------------------------------
-Clean-up: Direct user-space access
-----------------------------------
-Code lines using direct user-space access (e.g. passing user-space
-pointer to memcpy()) becomes more portable across processor
-architectures. The modification follows the following rules.
-
-1. Move the code section as it is to the architecture dependent
-   directory if it is a part of the critical-path.
-2. Otherwise, rewrite the code section by using the portable methods.
-   The methods include copy_from_user(), copy_to_user(),
-   pte_get_phys() and phys_to_virt().
-
---------------------------------
-Test: MPI and OpenMP micro-bench
---------------------------------
-The performance figures of MPI and OpenMP primitives are compared with
-those of Linux by using Intel MPI Benchmarks and EPCC OpenMP Micro
-Benchmark.
-
-
-===================================
-What's new in V1.3.0 (Sep 30, 2017)
-===================================
-
---------------------
-Feature: Kernel dump
---------------------
-1. A dump level of "only kernel memory" is added.
-
-The following two levels are available now:
-   0: Dump all
-  24: Dump only kernel memory
-
-The dump level can be set by -d option in ihkosctl or the argument
-for ihk_os_makedumpfile(), as shown in the following examples:
-
-   Command:		ihkosctl 0 dump -d 24
-   Function call:	ihk_os_makedumpfile(0, NULL, 24, 0);
-
-2. Dump file is created when Linux panics.
-
-The dump level can be set by dump_level kernel argument, as shown in the
-following example:
-
-   ihkosctl 0 kargs "hidos dump_level=24"
-
-The IHK dump function is registered to panic_notifier_list when creating
-/dev/mcdX and called when Linux panics.
-
------------------------------
-Feature: Quick Process Launch
------------------------------
-
-MPI process launch time and some of the initialization time can be
-reduced in application consisting of multiple MPI programs which are
-launched in turn in the job script.
-
-The following two steps should be performed to use this feature:
-1. Replace mpiexec with ql_mpiexec_start and add some lines for
-   ql_mpiexec_finalize in the job script
-2. Modify the app so that it can repeat calculations and wait for the
-   instructions from ql_mpiexec_{start,finalize} at the end of the
-   loop
-
-The first step is explained using an example. Assume the original job
-script looks like this:
-
-/* Execute ensamble simulation and then data assimilation, and repeat this
-   ten times */
-for i in {1..10}; do
-
-     /* Each ensamble simulation execution uses 100 nodes, launch ten of them
-        in parallel */
-    for j in {1..10}; do
-         mpiexec -n 100 -machinefile ./list1_$j p1.out a1 & pids[$i]=$!;
-    done
-
-    /* Wait until the ten ensamble simulation programs finish */
-    for j in {1..10}; do wait ${pids[$j]}; done
-
-    /* Launch one data assimilation program using 1000 nodes */
-    mpiexec -n 1000 -machinefile ./list2 p2.out a2
-done
-
-The job script should be modified like this:
-
-for i in {1..10}; do
-    for j in {1..10}; do
-        /*  Replace mpiexec with ql_mpiexec_start */
-        ql_mpiexec_start -n 100 -machinefile ./list1_$j p1.out a1 & pids[$j]=$!;
-    done
-
-    for j in {1..10}; do wait ${pids[$j]}; done
-
-    ql_mpiexec_start -n 1000 -machinefile ./list2 p2.out a2
-done
-
-/* p1.out and p2.out don't exit but are waiting for the next calculation.
-   So tell them to exit */
-for j in {1..10}; do
-   ql_mpiexec_finalize -machinefile ./list1_$i p1.out a1;
-done
-ql_mpiexec_finalize -machinefile ./list2 p2.out a2;
-
-
-The second step is explained using a pseudo-code.
-
-    MPI_Init();
-    Prepare data exchange with preceding / following MPI programs
-loop:
-    foreach Fortran module
-        Initialize data using command-line argments, parameter files,
-	environment variables
-    Input data from preceding MPI programs / Read snap-shot
-    Perform main calculation
-    Output data to following MPI programs / Write snap-shot
-    /* ql_client() waits for command of ql_mpiexec_{start,finish} */
-    if (ql_client() == QL_CONTINUE) { goto loop; }
-    MPI_Finalize();
-
-qlmpilib.h should be included in the code and libql{mpi,fort}.so
-should be linked to the executable file.
-
-
-========================
-Restrictions on McKernel
-========================
-
- 1. Pseudo devices such as /dev/mem and /dev/zero are not mmap()ed
-    correctly even if the mmap() returns a success. An access of their
-    mapping receives the SIGSEGV signal.
-
- 2. clone() supports only the following flags. All the other flags
-    cause clone() to return error or are simply ignored.
-
-    * CLONE_CHILD_CLEARTID
-    * CLONE_CHILD_SETTID
-    * CLONE_PARENT_SETTID
-    * CLONE_SETTLS
-    * CLONE_SIGHAND
-    * CLONE_VM
-
- 3. PAPI has the following restriction.
-
-    * Number of counters a user can use at the same time is up to the
-      number of the physical counters in the processor.
-
- 4. msync writes back only the modified pages mapped by the calling process.
-
- 5. The following syscalls always return the ENOSYS error.
-
-    * migrate_pages()
-    * move_pages()
-    * set_robust_list()
-
- 6. The following syscalls always return the EOPNOTSUPP error.
-
-    * arch_prctl(ARCH_SET_GS)
-    * signalfd()
-
- 7. signalfd4() returns a fd, but signal is not notified through the
-    fd.
-
- 8. set_rlimit sets the limit values but they are not enforced.
-
- 9. Address randomization is not supported.
-
-10. brk() extends the heap more than requestd when -h
-    (--extend-heap-by=)<step> option of mcexec is used with the value
-    larger than 4 KiB.  syscall_pwrite02 of LTP would fail for this
-    reason. This is because the test expects that the end of the heap
-    is set to the same address as the argument of sbrk() and expects a
-    segmentation violation occurs when it tries to access the memory
-    area right next to the boundary. However, the optimization sets
-    the end to a value larger than the requested. Therefore, the
-    expected segmentation violation doesn't occur.
-
-11. setpriority()/getpriority() won't work. They might set/get the
-    priority of a random mcexec thread. This is because there's no
-    fixed correspondence between a McKernel thread which issues the
-    system call and a mcexec thread which handles the offload request.
-
-12. mbind() can set the policy but it is not used when allocating
-    physical pages.
-
-13. MPOL_F_RELATIVE_NODES and MPOL_INTERLEAVE flags for
-    set_mempolicy()/mbind() are not supported.
-
-14. The MPOL_BIND policy for set_mempolicy()/mbind() works as the same
-    as the MPOL_PREFERRED policy. That is, the physical page allocator
-    doesn't give up the allocation when the specified nodes are
-    running out of pages but continues to search pages in the other
-    nodes.
-
-15. Kernel dump on Linux panic requires Linux kernel CentOS-7.4 and
-    later. In addition, crash_kexec_post_notifiers kernel argument
-    must be given to Linux kernel.
-
-16. setfsuid()/setfsgid() cannot change the id of the calling thread.
-    Instead, it changes that of the mcexec worker thread which takes
-    the system-call offload request.
-
-17. mmap (hugeTLBfs): The physical pages corresponding to a map are
-    released when no McKernel process exist. The next map gets fresh
-    physical pages.
-
-18. Sticky bit on executable file has no effect.
-
-19. Linux (RHEL-7 for x86_64) could hang when offlining CPUs in the
-    process of booting McKernel due to the Linux bug, found in
-    Linux-3.10 and fixed in the later version.  One way to circumvent
-    this is to always assign the same CPU set to McKernel.
-
-20. madvise:
-    * MADV_HWPOISON and MADV_SOFT_OFFLINE always returns -EPERM.
-    * MADV_MERGEABLE and MADV_UNMERGEABLE always returns -EINVAL.
-    * MADV_HUGEPAGE and MADV_NOHUGEPAGE on file map returns -EINVAL
-      (It succeeds on RHEL-8 for aarch64).
-
-21. brk() and mmap() doesn't report out-of-memory through its return
-    value. Instead, page-fault reports the error.
-
-22. Anonymous mmap pre-maps requested number of pages when contiguous
-    pages are available. Demand paging is used when not available.
-
-23. Mixing page sizes in anonymous shared mapping is not allowed. mmap
-    creates vm_range with one page size. And munmap or mremap that
-    needs the reduced page size changes the sizes of all the pages of
-    the vm_range.
-
-24. ihk_os_getperfevent() could time-out when invoked from Fujitsu TCS
-    (job-scheduler).
-
-25. The behaviors of madvise and mbind are changed to do nothing and
-    report success as a workaround for Fugaku.

README.md

@@ -1,285 +0,0 @@
-![McKernel Logo](https://www.sys.r-ccs.riken.jp/members_files/bgerofi/mckernel-logo.png)
--------------------------
-
-IHK/McKernel is a light-weight multi-kernel operating system designed for high-end supercomputing. It runs Linux and McKernel, a light-weight kernel (LWK), side-by-side inside compute nodes and aims at the following:
-
-- Provide scalable and consistent execution of large-scale parallel scientific applications, but at the same time maintain the ability to rapidly adapt to new hardware features and emerging programming models
-- Provide efficient memory and device management so that resource contention and data movement are minimized at the system level
-- Eliminate OS noise by isolating OS services in Linux and provide jitter free execution on the LWK
-- Support the full POSIX/Linux APIs by selectively offloading (slow-path) system calls to Linux
-
-## Contents
-
-- [Background](#background-and-motivation)
-- [Architectural Overview](#architectural-overview)
-- [Installation](#installation)
-- [The Team](#the-team)
-
-## Background and Motivation
-
-With the growing complexity of high-end supercomputers, the current system software stack faces significant challenges as we move forward to exascale and beyond. The necessity to deal with extreme degree of parallelism, heterogeneous architectures, multiple levels of memory hierarchy, power constraints, etc., advocates operating systems that can rapidly adapt to new hardware requirements, and that can support novel programming paradigms and runtime systems. On the other hand, a new class of more dynamic and complex applications are also on the horizon, with an increasing demand for application constructs such as in-situ analysis, workflows, elaborate monitoring and performance tools. This complexity relies not only on the rich features of POSIX, but also on the Linux APIs (such as the */proc*, */sys* filesystems, etc.) in particular.
-
-
-##### Two Traditional HPC OS Approaches
-
-Traditionally, light-weight operating systems specialized for HPC followed two approaches to tackle scalable execution of large-scale applications. In the full weight kernel (FWK) approach, a full Linux environment is taken as the basis, and features that inhibit attaining HPC scalability are removed, i.e., making it light-weight. The pure light-weight kernel (LWK) approach, on the other hand, starts from scratch and effort is undertaken to add sufficient functionality so that it provides a familiar API, typically something close to that of a general purpose OS, while at the same time it retains the desired scalability and reliability attributes. Neither of these approaches yields a fully Linux compatible environment.
-
-
-##### The Multi-kernel Approach
-
-A hybrid approach recognized recently by the system software community is to run Linux simultaneously with a lightweight kernel on compute nodes and multiple research projects are now pursuing this direction. The basic idea is that simulations run on an HPC tailored lightweight kernel, ensuring the necessary isolation for noiseless execution of parallel applications, but Linux is leveraged so that the full POSIX API is supported. Additionally, the small code base of the LWK can also facilitate rapid prototyping for new, exotic hardware features. Nevertheless, the questions of how to share node resources between the two types of kernels, where do device drivers execute, how exactly do the two kernels interact with each other and to what extent are they integrated, remain subjects of ongoing debate.
-
-
-
-## Architectural Overview
-
-At the heart of the stack is a low-level software infrastructure called Interface for Heterogeneous Kernels (IHK). IHK is a general framework that provides capabilities for partitioning resources in a many-core environment (e.g.,CPU cores and physical memory) and it enables management of lightweight kernels. IHK can allocate and release host resources dynamically and no reboot of the host machine is required when altering configuration. IHK also provides a low-level inter-kernel messaging infrastructure, called the Inter-Kernel Communication (IKC) layer. An architectural overview of the main system components is shown below.
-
-
-![arch](https://www.sys.r-ccs.riken.jp/members_files/bgerofi/mckernel.png)
-
-
-McKernel is a lightweight kernel written from scratch. It is designed for HPC and is booted from IHK. McKernel retains a binary compatible ABI with Linux, however, it implements only a small set of performance sensitive system calls and the rest are offloaded to Linux. Specifically, McKernel has its own memory management, it supports processes and multi-threading with a simple round-robin cooperative (tick-less) scheduler, and it implements signaling. It also allows inter-process memory mappings and it provides interfaces to hardware performance counters.
-
-### Functionality
-
-An overview of some of the principal functionalities of the IHK/McKernel stack is provided below.
-
-#### System Call Offloading
-
-System call forwarding in McKernel is implemented as follows. When an offloaded system call occurs, McKernel marshals the system call number along with its arguments and sends a message to Linux via a dedicated IKC channel. The corresponding proxy process running on Linux is by default waiting for system call requests through an ioctl() call into IHK’s system call delegator kernel module. The delegator kernel module’s IKC interrupt handler wakes up the proxy process, which returns to userspace and simply invokes the requested system call. Once it obtains the return value, it instructs the delegator module to send the result back to McKernel, which subsequently passes the value to user-space.
-
-#### Unified Address Space
-
-The unified address space model in IHK/McKernel ensures that offloaded system calls can seamlessly resolve arguments even in case of pointers. This mechanism is depicted below and is implemented as follows.
-
-
-![unified_ap](https://www.sys.r-ccs.riken.jp/members_files/bgerofi/img/unified_address_space_en.png)
-
-
-First, the proxy process is compiled as a position independent binary, which enables us to map the code and data segments specific to the proxy process to an address range which is explicitly excluded from McKernel’s user space. The grey box on the right side of the figure demonstrates the excluded region. Second, the entire valid virtual address range of McKernel’s application user-space is covered by a special mapping in the proxy process for which we use a pseudo file mapping in Linux. This mapping is indicated by the blue box on the left side of the figure.
-
-
-## Installation
-
-For a smooth experience, we recommend the following combination of OS distributions and platforms:
-
-- CentOS 7.3+ running on Intel Xeon / Xeon Phi
-
-
-##### 1. Change SELinux settings
-Log in as the root and disable SELinux:
-
-~~~~
-vim /etc/selinux/config
-~~~~
-
-Change the file to SELINUX=disabled
-
-##### 2. Reboot the host machine
-~~~~
-sudo reboot
-~~~~
-
-##### 3. Prepare packages, kernel symbol table file
-You will need the following packages installed:
-
-~~~~
-sudo yum install cmake kernel-devel binutils-devel systemd-devel numactl-devel gcc make nasm git
-~~~~
-
-Grant read permission to the System.map file of your kernel version:
-
-~~~~
-sudo chmod a+r /boot/System.map-`uname -r`
-~~~~
-
-##### 4. Obtain sources and compile the kernel
-
-Clone the source code:
-
-~~~~
-mkdir -p ~/src/ihk+mckernel/
-cd ~/src/ihk+mckernel/
-git clone --recursive -b development https://github.com/RIKEN-SysSoft/mckernel.git
-~~~~
-
-(Optional) Checkout to the specific branch or version:
-
-~~~~
-cd mckernel
-git checkout <pathspec>
-git submodule update
-~~~~
-
-Foe example, if you want to try the development branch, use "development" as the pathspec. If you want to try the prerelease version 1.7.0-0.2, use "1.7.0-0.2".
-
-###### 4.1 Install with cmake
-
-Configure and compile:
-
-~~~~
-mkdir -p build && cd build
-cmake -DCMAKE_INSTALL_PREFIX=${HOME}/ihk+mckernel $HOME/src/ihk+mckernel/mckernel
-make -j install
-~~~~
-
-The IHK kernel modules and McKernel kernel image should be installed under the **ihk+mckernel** folder in your home directory.
-
-###### 4.2 Install with rpm
-
-Configure, compile and build rpm:
-
-~~~~
-mkdir -p build && cd build
-cmake $HOME/src/ihk+mckernel/mckernel
-make dist
-cp mckernel-<version>.tar.gz <rpmbuild>/SOURCES
-rpm -ba scripts/mckernel.spec
-sudo rpm -ivh <rpmbuild>/RPMS/<arch>/mckernel-<version>-<release>_<linux_kernel_ver>_<dist>.<arch>.rpm
-~~~~
-
-The IHK kernel modules and McKernel kernel image are installed under the system directory.
-
-##### 5. Boot McKernel
-
-A boot script called mcreboot.sh is provided under sbin in the install folder. To boot on logical CPU 1 with 512MB of memory, use the following invocation:
-
-~~~~
-export TOP=${HOME}/ihk+mckernel/
-cd ${TOP}
-sudo ./sbin/mcreboot.sh -c 1 -m 512m
-~~~~
-
-You should see something similar like this if you display the McKernel's kernel message log:
-
-
-~~~~
-./sbin/ihkosctl 0 kmsg
-
-IHK/McKernel started.
-[ -1]: no_execute_available: 1
-[ -1]: map_fixed: phys: 0xfee00000 => 0xffff860000009000 (1 pages)
-[ -1]: setup_x86 done.
-[ -1]: ns_per_tsc: 385
-[ -1]: KCommand Line: hidos    dump_level=24
-[ -1]: Physical memory: 0x1ad3000 - 0x21000000, 525520896 bytes, 128301 pages available @ NUMA: 0
-[ -1]: NUMA: 0, Linux NUMA: 0, type: 1, available bytes: 525520896, pages: 128301
-[ -1]: NUMA 0 distances: 0 (10),
-[ -1]: map_fixed: phys: 0x28000 => 0xffff86000000a000 (2 pages)
-[ -1]: Trampoline area: 0x28000
-[ -1]: map_fixed: phys: 0x0 => 0xffff86000000c000 (1 pages)
-[ -1]: # of cpus : 1
-[ -1]: locals = ffff880001af6000
-[  0]: BSP: 0 (HW ID: 1 @ NUMA 0)
-[  0]: BSP: booted 0 AP CPUs
-[  0]: Master channel init acked.
-[  0]: vdso is enabled
-IHK/McKernel booted.
-~~~~
-
-
-##### 5. Run a simple program on McKernel
-
-The mcexec command line tool (which is also the Linux proxy process) can be used for executing applications on McKernel:
-
-~~~~
-./bin/mcexec hostname
-centos-vm
-~~~~
-
-
-##### 6. Shutdown McKernel
-
-Finally, to shutdown McKernel and release CPU/memory resources back to Linux use the following command:
-
-~~~~
-sudo ./sbin/mcstop+release.sh
-~~~~
-
-##### 7. Advanced: Enable Utility Thread offloading Interface (UTI)
-
-UTI enables a runtime such as MPI runtime to spawn utility threads such as MPI asynchronous progress threads to Linux cores.
-
-1. Install capstone
-
-Install EPEL capstone-devel:
-
-~~~~
-sudo yum install epel-release
-sudo yum install capstone-devel
-~~~~
-
-2. Install syscall_intercept
-
-~~~~
-git clone https://github.com/RIKEN-SysSoft/syscall_intercept.git
-cmake ../arch/aarch64 -DCMAKE_INSTALL_PREFIX=<syscall-intercept-install> -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=gcc -DTREAT_WARNINGS_AS_ERRORS=OFF
-~~~~
-
-3. Install UTI for McKernel
-
-Install:
-
-~~~~
-git clone https://github.com/RIKEN-SysSoft/uti.git
-mkdir build && cd build
-../uti/configure --prefix=<mckernel-install> --with-rm=mckernel
-make && make install
-~~~~
-
-4. Install McKernel
-
-~~~~
-CMAKE_PREFIX_PATH=<syscall-intercept-install> cmake -DCMAKE_INSTALL_PREFIX=${HOME}/ihk+mckernel -DENABLE_UTI=ON $HOME/src/ihk+mckernel/mckernel
-~~~~
-
-5. Run executable
-
-~~~~
-mcexec --enable-uti <command>
-~~~~
-
-6. Install UTI for Linux for performance comparison
-
-Install by make:
-
-~~~~
-git clone https://github.com/RIKEN-SysSoft/uti.git
-mkdir build && cd build
-../uti/configure --prefix=<uti-install> --with-rm=linux
-make && make install
-~~~~
-
-Install by rpm:
-
-~~~~
-git clone https://github.com/RIKEN-SysSoft/uti.git
-mkdir build && cd build
-../uti/configure --prefix=<uti-install> --with-rm=linux
-rm -f ~/rpmbuild/SOURCES/<version>.tar.gz
-rpmbuild -ba ./scripts/uti.spec
-rpm -Uvh uti-<version>-<release>-<arch>.rpm
-~~~~
-
-## The Team
-
-The McKernel project was started at The University of Tokyo and currently it is mainly developed at RIKEN.
-Some of our collaborators include:
-
-- Hitachi
-- Fujitsu
-- CEA (France)
-- NEC
-
-
-## License
-
-McKernel is GPL licensed, as found in the LICENSE file.
-
-## Contact
-
-Please give your feedback to us via one of the following mailing lists. Subscription via [www.pccluster.org](http://www.pccluster.org/mailman/listinfo/mckernel-users) is needed.
-
-* English: mckernel-users@pccluster.org
-* Japanese: mckernel-users-jp@pccluster.org

README.rst

@@ -0,0 +1,23 @@
+.. figure:: docs/mckernel-logo.png
+
+IHK/McKernel is a light-weight multi-kernel operating system designed
+for high-end supercomputing. It runs Linux and McKernel, a light-weight
+kernel (LWK), side-by-side inside compute nodes and aims at the
+following:
+
+-  Provide scalable and consistent execution of large-scale parallel
+   scientific applications, but at the same time maintain the ability to
+   rapidly adapt to new hardware features and emerging programming
+   models
+-  Provide efficient memory and device management so that resource
+   contention and data movement are minimized at the system level
+-  Eliminate OS noise by isolating OS services in Linux and provide
+   jitter free execution on the LWK
+-  Support the full POSIX/Linux APIs by selectively offloading
+   (slow-path) system calls to Linux
+
+Documentation
+=============
+
+Documentation is available
+`here <https://ihkmckernel.readthedocs.io>`__.

arch/arm64/kernel/cpu.c

@@ -143,6 +143,11 @@ void arch_save_panic_regs(void *irq_regs)
 
 	clv = get_arm64_this_cpu_local();
 
+	/* If kernel mode PF occurred, unroll the causing call stack */
+	if (cpu_local_var(kernel_mode_pf_regs)) {
+		regs = cpu_local_var(kernel_mode_pf_regs);
+	}
+
 	/* For user-space, use saved kernel context */
 	if (regs->pc < USER_END) {
 		memset(clv->arm64_cpu_local_thread.panic_regs,
@@ -725,6 +730,49 @@ static void show_context_stack(struct pt_regs *regs)
 	}
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+void __show_context_stack(struct thread *thread,
+        unsigned long pc, uintptr_t sp, int kprintf_locked)
+{
+    uintptr_t stack_top;
+    unsigned long irqflags = 0;
+
+    stack_top = ALIGN_UP(sp, (uintptr_t)KERNEL_STACK_SIZE);
+
+    if (!kprintf_locked)
+        irqflags = kprintf_lock();
+
+    __kprintf("TID: %d, call stack (most recent first):\n",
+        thread->tid);
+    __kprintf("PC: %016lx, SP: %016lx\n", pc, sp);
+    for (;;) {
+        extern char _head[], _end[];
+        uintptr_t *fp, *lr;
+        fp = (uintptr_t *)sp;
+        lr = (uintptr_t *)(sp + 8);
+
+        if ((*fp <= sp)) {
+            break;
+        }
+
+        if ((*fp > stack_top)) {
+            break;
+        }
+
+        if ((*lr < (unsigned long)_head) ||
+            (*lr > (unsigned long)_end)) {
+            break;
+        }
+
+        __kprintf("PC: %016lx, SP: %016lx, FP: %016lx\n", *lr - 4, sp, *fp);
+        sp = *fp;
+    }
+
+    if (!kprintf_locked)
+        kprintf_unlock(irqflags);
+}
+#endif
+
 void handle_IPI(unsigned int vector, struct pt_regs *regs)
 {
 	struct ihk_mc_interrupt_handler *h;
@@ -786,6 +834,19 @@ void cpu_safe_halt(void)
 	cpu_enable_interrupt();
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+/*@
+  @ assigns \nothing;
+  @ ensures \interrupt_disabled == 0;
+  @*/
+void cpu_halt_panic(void)
+{
+	extern void __cpu_do_idle(void);
+	cpu_enable_interrupt();
+	__cpu_do_idle();
+}
+#endif
+
 #if defined(CONFIG_HAS_NMI)
 #include <arm-gic-v3.h>
 
@@ -851,6 +912,21 @@ unsigned long cpu_enable_interrupt_save(void)
 	return flags;
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+int cpu_interrupt_disabled(void)
+{
+	unsigned long flags;
+	unsigned long masked = ICC_PMR_EL1_MASKED;
+
+	asm volatile(
+		"mrs_s  %0, " __stringify(ICC_PMR_EL1)
+		: "=&r" (flags)
+		:
+		: "memory");
+	return (flags == masked);
+}
+#endif
+
 #else /* defined(CONFIG_HAS_NMI) */
 
 /* @ref.impl arch/arm64/include/asm/irqflags.h::arch_local_irq_enable */
@@ -1279,7 +1355,7 @@ long ihk_mc_show_cpuinfo(char *buf, size_t buf_size, unsigned long read_off, int
 
 		/* generate strings */
 		loff += scnprintf(lbuf + loff, lbuf_size - loff,
-				  "processor\t: %d\n", cpuinfo->hwid);
+				  "processor\t: %d\n", i);
 		loff += scnprintf(lbuf + loff, lbuf_size - loff, "Features\t:");
 
 		for (j = 0; hwcap_str[j]; j++) {
@@ -1372,6 +1448,14 @@ void arch_print_stack(void)
 {
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+unsigned long arch_get_instruction_address(const void *reg)
+{
+	const struct pt_regs *regs = (struct pt_regs *)reg;
+	return regs->pc;
+}
+#endif
+
 void arch_show_interrupt_context(const void *reg)
 {
 	const struct pt_regs *regs = (struct pt_regs *)reg;

arch/arm64/kernel/fault.c

@@ -223,7 +223,12 @@ static int do_translation_fault(unsigned long addr,
 					  unsigned int esr,
 					  struct pt_regs *regs)
 {
+#ifdef ENABLE_TOFU
+	// XXX: Handle kernel space page faults for Tofu driver
+	//if (addr < USER_END)
+#else
 	if (addr < USER_END)
+#endif
 		return do_page_fault(addr, esr, regs);
 
 	do_bad_area(addr, esr, regs);

arch/arm64/kernel/include/arch-lock.h

@@ -9,6 +9,9 @@
 #include "affinity.h"
 #include <lwk/compiler.h>
 #include "config.h"
+#ifdef ENABLE_FUGAKU_HACKS
+#include <ihk/debug.h>
+#endif
 
 //#define DEBUG_SPINLOCK
 //#define DEBUG_MCS_RWLOCK
@@ -31,6 +34,10 @@ typedef struct {
 #endif /* __AARCH64EB__ */
 } __attribute__((aligned(4))) ihk_spinlock_t;
 
+#ifdef ENABLE_FUGAKU_HACKS
+extern ihk_spinlock_t *get_this_cpu_runq_lock(void);
+#endif
+
 extern void preempt_enable(void);
 extern void preempt_disable(void);
 
@@ -98,6 +105,18 @@ static int __ihk_mc_spinlock_trylock_noirq(ihk_spinlock_t *lock)
 	: "memory");
 
 	success = !tmp;
+
+#ifdef ENABLE_FUGAKU_HACKS
+#if 0
+	if (success) {
+		if (get_this_cpu_runq_lock() == lock &&
+				!cpu_interrupt_disabled()) {
+			kprintf("%s: WARNING: runq lock held without IRQs disabled?\n", __func__); \
+		}
+	}
+#endif
+#endif
+
 	if (!success) {
 		preempt_enable();
 	}
@@ -182,6 +201,14 @@ static void __ihk_mc_spinlock_lock_noirq(ihk_spinlock_t *lock)
 	: "=&r" (lockval), "=&r" (newval), "=&r" (tmp), "+Q" (*lock)
 	: "Q" (lock->owner), "I" (1 << TICKET_SHIFT)
 	: "memory");
+#ifdef ENABLE_FUGAKU_HACKS
+#if 0
+	if (get_this_cpu_runq_lock() == lock &&
+			!cpu_interrupt_disabled()) {
+		kprintf("%s: WARNING: runq lock held without IRQs disabled?\n", __func__); \
+	}
+#endif
+#endif
 }
 
 #ifdef DEBUG_SPINLOCK

arch/arm64/kernel/include/arch-memory.h

@@ -94,7 +94,11 @@ extern char _end[];
 # define LD_TASK_UNMAPPED_BASE	UL(0x0000080000000000)
 # define TASK_UNMAPPED_BASE	UL(0x0000100000000000)
 # define USER_END		UL(0x0000400000000000)
+#ifdef ENABLE_TOFU
+# define MAP_VMAP_START		UL(0xffff7bdfffff0000)
+#else
 # define MAP_VMAP_START		UL(0xffff780000000000)
+#endif
 # define MAP_VMAP_SIZE		UL(0x0000000100000000)
 # define MAP_FIXED_START	UL(0xffff7ffffbdd0000)
 # define MAP_ST_START		UL(0xffff800000000000)
@@ -142,6 +146,7 @@ extern char _end[];
 # define __PTL1_SHIFT  16
 # define PTL4_INDEX_MASK 0
 # define PTL3_INDEX_MASK ((UL(1) << 6) - 1)
+# define PTL3_INDEX_MASK_LINUX ((UL(1) << 10) - 1)
 # define PTL2_INDEX_MASK ((UL(1) << 13) - 1)
 # define PTL1_INDEX_MASK PTL2_INDEX_MASK
 # define __PTL4_CONT_SHIFT (__PTL4_SHIFT + 0)
@@ -829,7 +834,13 @@ static inline int pte_is_head(pte_t *ptep, pte_t *old, size_t cont_size)
 	return page_is_contiguous_head(ptep, cont_size);
 }
 
-struct page_table;
+typedef pte_t translation_table_t;
+struct page_table {
+	translation_table_t* tt;
+	translation_table_t* tt_pa;
+	int asid;
+};
+
 void arch_adjust_allocate_page_size(struct page_table *pt,
 				    uintptr_t fault_addr,
 				    pte_t *ptep,
@@ -849,7 +860,6 @@ void *map_fixed_area(unsigned long phys, unsigned long size, int uncachable);
 void set_address_space_id(struct page_table *pt, int asid);
 int get_address_space_id(const struct page_table *pt);
 
-typedef pte_t translation_table_t;
 void set_translation_table(struct page_table *pt, translation_table_t* tt);
 translation_table_t* get_translation_table(const struct page_table *pt);
 translation_table_t* get_translation_table_as_paddr(const struct page_table *pt);

arch/arm64/kernel/include/arch-string.h

@@ -10,4 +10,13 @@ extern void *__inline_memcpy(void *to, const void *from, size_t t);
 
 extern void *__inline_memset(void *s, unsigned long c, size_t count);
 
+#define ARCH_MEMCLEAR
+
+extern void __memclear(void *addr, unsigned long len, void *tmp);
+inline static void memclear(void *addr, unsigned long len)
+{
+	uint64_t q0q1[4];
+	__memclear(addr, len, (void *)&q0q1);
+}
+
 #endif	/* __HEADER_ARM64_COMMON_ARCH_TIMER_H */

arch/arm64/kernel/include/io.h

@@ -80,6 +80,10 @@ static inline uint64_t __raw_readq(const volatile void *addr)
 	return val;
 }
 
+/* IO barriers */
+#define __iormb()		rmb()
+#define __iowmb()		wmb()
+
 /*
  * Relaxed I/O memory access primitives. These follow the Device memory
  * ordering rules but do not guarantee any ordering relative to Normal memory
@@ -95,5 +99,20 @@ static inline uint64_t __raw_readq(const volatile void *addr)
 #define writel_relaxed(v,c)	((void)__raw_writel((uint32_t)(v),(c)))
 #define writeq_relaxed(v,c)	((void)__raw_writeq((uint64_t)(v),(c)))
 
+/*
+ * I/O memory access primitives. Reads are ordered relative to any
+ * following Normal memory access. Writes are ordered relative to any prior
+ * Normal memory access.
+ */
+#define readb(c)		({ uint8_t  __v = readb_relaxed(c); __iormb(); __v; })
+#define readw(c)		({ uint16_t __v = readw_relaxed(c); __iormb(); __v; })
+#define readl(c)		({ uint32_t __v = readl_relaxed(c); __iormb(); __v; })
+#define readq(c)		({ uint64_t __v = readq_relaxed(c); __iormb(); __v; })
+
+#define writeb(v,c)		({ __iowmb(); writeb_relaxed((v),(c)); })
+#define writew(v,c)		({ __iowmb(); writew_relaxed((v),(c)); })
+#define writel(v,c)		({ __iowmb(); writel_relaxed((v),(c)); })
+#define writeq(v,c)		({ __iowmb(); writeq_relaxed((v),(c)); })
+
 #endif /* __KERNEL__ */
 #endif /* __ASM_IO_H */

arch/arm64/kernel/include/rlimit.h

@@ -85,7 +85,11 @@ enum __rlimit_resource
   __RLIMIT_RTPRIO = 14,
 #define RLIMIT_RTPRIO __RLIMIT_RTPRIO
 
-  __RLIMIT_NLIMITS = 15,
+  /* timeout for RT tasks in us */
+  __RLIMIT_RTTIME = 15,
+#define RLIMIT_RTTIME __RLIMIT_RTTIME
+
+  __RLIMIT_NLIMITS = 16,
   __RLIM_NLIMITS = __RLIMIT_NLIMITS
 #define RLIMIT_NLIMITS __RLIMIT_NLIMITS
 #define RLIM_NLIMITS __RLIM_NLIMITS

arch/arm64/kernel/include/syscall_list.h

@@ -83,6 +83,7 @@ SYSCALL_HANDLED(175, geteuid)
 SYSCALL_HANDLED(176, getgid)
 SYSCALL_HANDLED(177, getegid)
 SYSCALL_HANDLED(178, gettid)
+SYSCALL_HANDLED(179, sysinfo)
 SYSCALL_DELEGATED(188, msgrcv)
 SYSCALL_DELEGATED(189, msgsnd)
 SYSCALL_DELEGATED(192, semtimedop)
@@ -117,12 +118,13 @@ SYSCALL_HANDLED(241, perf_event_open)
 SYSCALL_DELEGATED(241, perf_event_open)
 #endif // PERF_ENABLE
 SYSCALL_HANDLED(260, wait4)
+SYSCALL_HANDLED(261, prlimit64)
 SYSCALL_HANDLED(270, process_vm_readv)
 SYSCALL_HANDLED(271, process_vm_writev)
 SYSCALL_HANDLED(281, execveat)
 SYSCALL_HANDLED(700, get_cpu_id)
 #ifdef PROFILE_ENABLE
-SYSCALL_HANDLED(__NR_profile, profile)
+SYSCALL_HANDLED(PROFILE_EVENT_MAX, profile)
 #endif // PROFILE_ENABLE
 SYSCALL_HANDLED(730, util_migrate_inter_kernel)
 SYSCALL_HANDLED(731, util_indicate_clone)
@@ -143,3 +145,8 @@ SYSCALL_DELEGATED(1049, stat)
 SYSCALL_DELEGATED(1060, getpgrp)
 SYSCALL_HANDLED(1062, time)
 SYSCALL_DELEGATED(1069, epoll_wait)
+
+/* Do not edit the lines including this comment and
+ * EOF just after it because those are used as a
+ * robust marker for the autotest patch.
+ */

arch/arm64/kernel/include/thread_info.h

@@ -2,7 +2,7 @@
 #ifndef __HEADER_ARM64_COMMON_THREAD_INFO_H
 #define __HEADER_ARM64_COMMON_THREAD_INFO_H
 
-#define MIN_KERNEL_STACK_SHIFT	15
+#define MIN_KERNEL_STACK_SHIFT	18
 
 #include <arch-memory.h>
 

arch/arm64/kernel/irq-gic-v3.c

@@ -7,6 +7,9 @@
 #include <process.h>
 #include <syscall.h>
 #include <ihk/debug.h>
+#ifdef ENABLE_FUGAKU_HACKS
+#include <ihk/monitor.h>
+#endif
 #include <arch-timer.h>
 #include <cls.h>
 
@@ -313,14 +316,27 @@ void handle_interrupt_gicv3(struct pt_regs *regs)
 	struct cpu_local_var *v = get_this_cpu_local_var();
 	//unsigned long irqflags;
 	int do_check = 0;
+#ifdef ENABLE_FUGAKU_HACKS
+	struct ihk_os_cpu_monitor *monitor = cpu_local_var(monitor);
 
+	++v->in_interrupt;
+#endif
 	irqnr = gic_read_iar();
 	cpu_enable_nmi();
 	set_cputime(from_user ? CPUTIME_MODE_U2K : CPUTIME_MODE_K2K_IN);
 	while (irqnr != ICC_IAR1_EL1_SPURIOUS) {
 		if ((irqnr < 1020) || (irqnr >= 8192)) {
 			gic_write_eoir(irqnr);
+#ifndef ENABLE_FUGAKU_HACKS
 			handle_IPI(irqnr, regs);
+#else
+			/* Once paniced, only allow CPU stop and NMI IRQs */
+			if (monitor->status != IHK_OS_MONITOR_PANIC ||
+					irqnr == INTRID_CPU_STOP ||
+					irqnr == INTRID_MULTI_NMI) {
+				handle_IPI(irqnr, regs);
+			}
+#endif
 		}
 		irqnr = gic_read_iar();
 	}
@@ -335,7 +351,12 @@ void handle_interrupt_gicv3(struct pt_regs *regs)
 	}
 	//ihk_mc_spinlock_unlock(&v->runq_lock, irqflags);
 
+#ifndef ENABLE_FUGAKU_HACKS
 	if (do_check) {
+#else
+	--v->in_interrupt;
+	if (monitor->status != IHK_OS_MONITOR_PANIC && do_check) {
+#endif
 		check_signal(0, regs, 0);
 		schedule();
 	}

arch/arm64/kernel/memory.c

@@ -150,12 +150,6 @@ void flush_tlb_single(unsigned long addr)
 	arch_flush_tlb_single(asid, addr);
 }
 
-struct page_table {
-	translation_table_t* tt;
-	translation_table_t* tt_pa;
-	int asid;
-};
-
 extern struct page_table swapper_page_table;
 static struct page_table *init_pt = &swapper_page_table;
 static ihk_spinlock_t init_pt_lock;
@@ -223,6 +217,13 @@ static inline int ptl4_index(unsigned long addr)
 	int idx = (addr >> PTL4_SHIFT) & PTL4_INDEX_MASK;
 	return idx;
 }
+#ifdef ENABLE_TOFU
+static inline int ptl3_index_linux(unsigned long addr)
+{
+	int idx = (addr >> PTL3_SHIFT) & PTL3_INDEX_MASK_LINUX;
+	return idx;
+}
+#endif
 static inline int ptl3_index(unsigned long addr)
 {
 	int idx = (addr >> PTL3_SHIFT) & PTL3_INDEX_MASK;
@@ -281,6 +282,40 @@ static inline pte_t* ptl4_offset(const translation_table_t* ptl4, unsigned long
 	}
 	return ptep;
 }
+
+#ifdef ENABLE_TOFU
+static inline pte_t* ptl3_offset_linux(const pte_t* l4p, unsigned long addr)
+{
+	pte_t* ptep = NULL;
+	pte_t pte = 0;
+	unsigned long phys = 0;
+	translation_table_t* ptl3 = NULL;
+	int idx = 0;
+
+	switch (CONFIG_ARM64_PGTABLE_LEVELS)
+	{
+	case 4:
+		pte = ptl4_val(l4p);
+		phys = pte & PT_PHYSMASK;
+		ptl3 = phys_to_virt(phys);
+		idx = ptl3_index_linux(addr);
+		ptep = (pte_t*)ptl3 + idx;
+		break;
+	case 3:
+		ptl3 = (translation_table_t*)l4p;
+		idx = ptl3_index_linux(addr);
+		ptep = (pte_t*)ptl3 + idx;
+		break;
+	case 2:
+	case 1:
+		/* PTL3が無いときにはエントリではなくページテーブルのアドレスを引渡していく。*/
+		ptep = (pte_t*)l4p;
+		break;
+	}
+	return ptep;
+}
+#endif
+
 static inline pte_t* ptl3_offset(const pte_t* l4p, unsigned long addr)
 {
 	pte_t* ptep = NULL;
@@ -959,7 +994,14 @@ static void init_normal_area(struct page_table *pt)
 	int i;
 	
 	tt = get_translation_table(pt);
-	
+
+#ifdef ENABLE_TOFU
+	setup(tt,
+			arm64_st_phys_base,
+			arm64_st_phys_base  + (1UL << 40));
+	return;
+#endif
+
 	for (i = 0; i < ihk_mc_get_nr_memory_chunks(); i++) {
 		unsigned long map_start, map_end;
 		int numa_id;
@@ -1287,6 +1329,58 @@ out:
 	return ret;
 }
 
+#ifdef ENABLE_TOFU
+int ihk_mc_linux_pt_virt_to_phys_size(struct page_table *pt,
+                           const void *virt,
+						   unsigned long *phys,
+						   unsigned long *size)
+{
+	unsigned long v = (unsigned long)virt;
+	pte_t* ptep;
+	translation_table_t* tt;
+
+	unsigned long paddr;
+	unsigned long lsize;
+
+	tt = get_translation_table(pt);
+
+	ptep = ptl4_offset(tt, v);
+	if (!ptl4_present(ptep)) {
+		return -EFAULT;
+	}
+
+	ptep = ptl3_offset_linux(ptep, v);
+	if (!ptl3_present(ptep)) {
+		return -EFAULT;
+	}
+	if (ptl3_type_block(ptep)) {
+		paddr = ptl3_phys(ptep);
+		lsize = PTL3_SIZE;
+		goto out;
+	}
+
+	ptep = ptl2_offset(ptep, v);
+	if (!ptl2_present(ptep)) {
+		return -EFAULT;
+	}
+	if (ptl2_type_block(ptep)) {
+		paddr = ptl2_phys(ptep);
+		lsize = PTL2_SIZE;
+		goto out;
+	}
+
+	ptep = ptl1_offset(ptep, v);
+	if (!ptl1_present(ptep)) {
+		return -EFAULT;
+	}
+	paddr = ptl1_phys(ptep);
+	lsize = PTL1_SIZE;
+out:
+	*phys = paddr | (v & (lsize - 1));
+	if(size) *size = lsize;
+	return 0;
+}
+#endif
 
 int ihk_mc_pt_virt_to_phys_size(struct page_table *pt,
                            const void *virt,
@@ -1348,7 +1442,6 @@ int ihk_mc_pt_virt_to_phys(struct page_table *pt,
 	return ihk_mc_pt_virt_to_phys_size(pt, virt, phys, NULL);
 }
 
-
 int ihk_mc_pt_print_pte(struct page_table *pt, void *virt)
 {
 	const unsigned long v = (unsigned long)virt;
@@ -1360,6 +1453,15 @@ int ihk_mc_pt_print_pte(struct page_table *pt, void *virt)
 	}
 	tt = get_translation_table(pt);
 
+	__kprintf("%s: 0x%lx, CONFIG_ARM64_PGTABLE_LEVELS: %d, ptl4_index: %ld, ptl3_index: %ld, ptl2_index: %ld, ptl1_index: %ld\n", 
+		__func__,
+		v,
+		CONFIG_ARM64_PGTABLE_LEVELS,
+		ptl4_index(v),
+		ptl3_index(v),
+		ptl2_index(v),
+		ptl1_index(v));
+
 	ptep = ptl4_offset(tt, v);
 	__kprintf("l4 table: 0x%lX l4idx: %d\n", virt_to_phys(tt), ptl4_index(v));
 	if (!(ptl4_present(ptep))) {
@@ -2147,6 +2249,198 @@ static void unmap_free_stat(struct page *page, unsigned long phys,
 	}
 }
 
+/*
+ * Kernel space page table clearing functions.
+ */
+struct clear_kernel_range_args {
+	int free_physical;
+};
+
+static int clear_kernel_range_middle(void *args0, pte_t *ptep, uint64_t base,
+			      uint64_t start, uint64_t end, int level);
+
+static int clear_kernel_range_l1(void *args0, pte_t *ptep, uint64_t base,
+		uint64_t start, uint64_t end)
+{
+	const struct table {
+		unsigned long pgsize;
+		unsigned long cont_pgsize;
+	} tbl = {
+		.pgsize = PTL1_SIZE,
+		.cont_pgsize = PTL1_CONT_SIZE
+	};
+
+	struct clear_kernel_range_args *args = args0;
+	uint64_t phys = 0;
+	pte_t old;
+	size_t clear_size;
+
+	if (ptl1_null(ptep)) {
+		return -ENOENT;
+	}
+
+	old = xchg(ptep, PTE_NULL);
+	if (!pte_is_present(&old))
+		return 0;
+
+	arch_flush_tlb_single(0, base);
+	clear_size = pte_is_contiguous(&old) ?
+		tbl.cont_pgsize : tbl.pgsize;
+
+	dkprintf("%s: 0x%lx:%lu unmapped\n",
+		__func__, base, clear_size);
+
+	if (args->free_physical) {
+		phys = ptl1_phys(&old);
+		ihk_mc_free_pages(phys_to_virt(phys), clear_size >> PAGE_SHIFT);
+	}
+
+	return 0;
+}
+
+static int clear_kernel_range_l2(void *args0, pte_t *ptep, uint64_t base,
+		uint64_t start, uint64_t end)
+{
+	return clear_kernel_range_middle(args0, ptep, base, start, end, 2);
+}
+
+static int clear_kernel_range_l3(void *args0, pte_t *ptep, uint64_t base,
+		uint64_t start, uint64_t end)
+{
+	return clear_kernel_range_middle(args0, ptep, base, start, end, 3);
+}
+
+static int clear_kernel_range_l4(void *args0, pte_t *ptep, uint64_t base,
+		uint64_t start, uint64_t end)
+{
+	return clear_kernel_range_middle(args0, ptep, base, start, end, 4);
+}
+
+static int clear_kernel_range_middle(void *args0, pte_t *ptep, uint64_t base,
+			      uint64_t start, uint64_t end, int level)
+{
+	const struct table {
+		walk_pte_t* walk;
+		walk_pte_fn_t* callback;
+		unsigned long pgsize;
+		unsigned long cont_pgsize;
+	} table[] = {
+		{walk_pte_l1, clear_kernel_range_l1, PTL2_SIZE, PTL2_CONT_SIZE}, /*PTL2*/
+		{walk_pte_l2, clear_kernel_range_l2, PTL3_SIZE, PTL3_CONT_SIZE}, /*PTL3*/
+		{walk_pte_l3, clear_kernel_range_l3, PTL4_SIZE, PTL4_CONT_SIZE}, /*PTL4*/
+	};
+	const struct table tbl = table[level-2];
+
+	struct clear_kernel_range_args *args = args0;
+	uint64_t phys = 0;
+	translation_table_t *tt;
+	int error;
+	pte_t old;
+	size_t clear_size;
+
+	if (ptl_null(ptep, level)) {
+		return -ENOENT;
+	}
+
+	dkprintf("%s(level: %d): 0x%lx in 0x%lx-0x%lx\n",
+			__func__, level, base, start, end);
+
+	if (ptl_type_page(ptep, level)
+			&& ((base < start) || (end < (base + tbl.pgsize)))) {
+		error = -EINVAL;
+		ekprintf("clear_range_middle(%p,%p,%lx,%lx,%lx,%d):"
+			 "split page. %d\n",
+			 args0, ptep, base, start, end, level, error);
+		return error;
+	}
+
+	if (ptl_type_page(ptep, level)) {
+		old = xchg(ptep, PTE_NULL);
+
+		if (!ptl_present(&old, level)) {
+			return 0;
+		}
+
+		arch_flush_tlb_single(0, base);
+
+		clear_size = pte_is_contiguous(&old) ?
+			tbl.cont_pgsize : tbl.pgsize;
+
+		dkprintf("%s(level: %d): 0x%lx:%lu unmapped\n",
+				__func__, level, base, clear_size);
+
+		if (args->free_physical) {
+			phys = ptl_phys(&old, level);
+			ihk_mc_free_pages(phys_to_virt(phys), clear_size >> PAGE_SHIFT);
+		}
+
+		return 0;
+	}
+
+	tt = (translation_table_t*)phys_to_virt(ptl_phys(ptep, level));
+	error = tbl.walk(tt, base, start, end, tbl.callback, args0);
+	if (error && (error != -ENOENT)) {
+		return error;
+	}
+
+	if (args->free_physical) {
+		if ((start <= base) && ((base + tbl.pgsize) <= end)) {
+			ptl_clear(ptep, level);
+			arch_flush_tlb_single(0, base);
+			ihk_mc_free_pages(tt, 1);
+		}
+	}
+
+	return 0;
+}
+
+static int clear_kernel_range(uintptr_t start, uintptr_t end, int free_physical)
+{
+	const struct table {
+		walk_pte_t* walk;
+		walk_pte_fn_t* callback;
+	} tables[] = {
+		{walk_pte_l2, clear_kernel_range_l2}, /*second*/
+		{walk_pte_l3, clear_kernel_range_l3}, /*first*/
+		{walk_pte_l4, clear_kernel_range_l4}, /*zero*/
+	};
+	const struct table initial_lookup = tables[CONFIG_ARM64_PGTABLE_LEVELS - 2];
+
+	int error;
+	struct clear_kernel_range_args args;
+	translation_table_t* tt;
+	unsigned long irqflags;
+
+	dkprintf("%s: start: 0x%lx, end: 0x%lx, free phys: %d\n",
+		 __func__, start, end, free_physical);
+
+	if (start <= USER_END)
+		return -EINVAL;
+
+	args.free_physical = free_physical;
+
+	irqflags = ihk_mc_spinlock_lock(&init_pt_lock);
+	tt = get_translation_table(get_init_page_table());
+	error = initial_lookup.walk(tt, 0,
+			(start & ~(0xffff000000000000)),
+			(end & ~(0xffff000000000000)),
+			initial_lookup.callback, &args);
+	dkprintf("%s: start: 0x%lx, end: 0x%lx, free phys: %d, ret: %d\n",
+		 __func__, start, end, free_physical, error);
+
+	ihk_mc_spinlock_unlock(&init_pt_lock, irqflags);
+	return error;
+}
+
+int ihk_mc_clear_kernel_range(void *start, void *end)
+{
+#define	KEEP_PHYSICAL	0
+	return clear_kernel_range((uintptr_t)start, (uintptr_t)end, KEEP_PHYSICAL);
+}
+
+/*
+ * User space page table clearing functions.
+ */
 struct clear_range_args {
 	int free_physical;
 	struct memobj *memobj;
@@ -2344,6 +2638,14 @@ static int clear_range(struct page_table *pt, struct process_vm *vm,
 	if (memobj && ((memobj->flags & MF_PREMAP))) {
 		args.free_physical = 0;
 	}
+
+	if (vm->proc->straight_va &&
+			(void *)start == vm->proc->straight_va &&
+			(void *)end == (vm->proc->straight_va +
+				vm->proc->straight_len)) {
+		args.free_physical = 0;
+	}
+
 	args.memobj = memobj;
 	args.vm = vm;
 

arch/arm64/kernel/memset.S

@@ -218,3 +218,41 @@ ENTRY(__inline_memset)
 	ret
 ENDPIPROC(__inline_memset)
 ENDPROC(____inline_memset)
+
+
+/*
+ * Non-temporal vector memory clear
+ *
+ * Parameters:
+ *     x0 - buf (assumed to be aligned to page size)
+ *     x1 - n (assumed to be at least page size)
+ */
+ENTRY(__memclear)
+       stp     q0, q1, [x2] /* Preserve two 128 bit vector regs */
+       eor v0.16B, v0.16B, v0.16B
+       eor v1.16B, v1.16B, v1.16B
+1:
+       stnp q0, q1, [x0, #32 * 0]
+       stnp q0, q1, [x0, #32 * 1]
+       stnp q0, q1, [x0, #32 * 2]
+       stnp q0, q1, [x0, #32 * 3]
+       stnp q0, q1, [x0, #32 * 4]
+       stnp q0, q1, [x0, #32 * 5]
+       stnp q0, q1, [x0, #32 * 6]
+       stnp q0, q1, [x0, #32 * 7]
+       stnp q0, q1, [x0, #32 * 8]
+       stnp q0, q1, [x0, #32 * 9]
+       stnp q0, q1, [x0, #32 * 10]
+       stnp q0, q1, [x0, #32 * 11]
+       stnp q0, q1, [x0, #32 * 12]
+       stnp q0, q1, [x0, #32 * 13]
+       stnp q0, q1, [x0, #32 * 14]
+       stnp q0, q1, [x0, #32 * 15]
+       add x0, x0, #512
+       subs x1, x1, #512
+       cmp x1, #0
+       b.ne    1b
+
+       ldp     q0, q1, [x2] /* Restore vector regs */
+       ret
+ENDPROC(__memclear)

arch/arm64/kernel/syscall.c

@@ -142,17 +142,6 @@ int obtain_clone_cpuid(cpu_set_t *cpu_set, int use_last)
 	return min_cpu;
 }
 
-int
-arch_clear_host_user_space()
-{
-	struct thread *th = cpu_local_var(current);
-
-	/* XXX: might be unnecessary */
-	clear_host_pte(th->vm->region.user_start,
-			(th->vm->region.user_end - th->vm->region.user_start), 0);
-	return 0;
-}
-
 /* archtecture-depended syscall handlers */
 extern unsigned long do_fork(int clone_flags, unsigned long newsp,
 			     unsigned long parent_tidptr, unsigned long child_tidptr,
@@ -185,33 +174,6 @@ SYSCALL_DECLARE(clone)
 	return ret;
 }
 
-SYSCALL_DECLARE(rt_sigaction)
-{
-	int sig = ihk_mc_syscall_arg0(ctx);
-	const struct sigaction *act = (const struct sigaction *)ihk_mc_syscall_arg1(ctx);
-	struct sigaction *oact = (struct sigaction *)ihk_mc_syscall_arg2(ctx);
-	size_t sigsetsize = ihk_mc_syscall_arg3(ctx);
-	struct k_sigaction new_sa, old_sa;
-	int rc;
-
-	if (sigsetsize != sizeof(sigset_t))
-		return -EINVAL;
-
-	if(act)
-		if(copy_from_user(&new_sa.sa, act, sizeof new_sa.sa)){
-			goto fault;
-		}
-	rc = do_sigaction(sig, act? &new_sa: NULL, oact? &old_sa: NULL);
-	if(rc == 0 && oact)
-		if(copy_to_user(oact, &old_sa.sa, sizeof old_sa.sa)){
-			goto fault;
-		}
-
-	return rc;
-fault:
-	return -EFAULT;
-}
-
 SYSCALL_DECLARE(prctl)
 {
 	struct process *proc = cpu_local_var(current)->proc;
@@ -1109,6 +1071,9 @@ static int setup_rt_frame(int usig, unsigned long rc, int to_restart,
 
 	if (k->sa.sa_flags & SA_RESTORER){
 		regs->regs[30] = (unsigned long)k->sa.sa_restorer;
+#ifdef ENABLE_FUGAKU_HACKS
+		kprintf("%s: SA_RESTORER: 0x%lx\n", __func__, regs->regs[30]);
+#endif
 	} else {
 		regs->regs[30] = (unsigned long)VDSO_SYMBOL(thread->vm->vdso_addr, sigtramp);
 	}
@@ -1368,70 +1333,6 @@ out:
 	return restart;
 }
 
-static struct sig_pending *
-getsigpending(struct thread *thread, int delflag){
-	struct list_head *head;
-	mcs_rwlock_lock_t *lock;
-	struct mcs_rwlock_node_irqsave mcs_rw_node;
-	struct sig_pending *next;
-	struct sig_pending *pending;
-	__sigset_t w;
-
-	w = thread->sigmask.__val[0];
-
-	lock = &thread->sigcommon->lock;
-	head = &thread->sigcommon->sigpending;
-	for(;;) {
-		if (delflag) {
-			mcs_rwlock_writer_lock(lock, &mcs_rw_node);
-		}
-		else {
-			mcs_rwlock_reader_lock(lock, &mcs_rw_node);
-		}
-
-		list_for_each_entry_safe(pending, next, head, list){
-			if(!(pending->sigmask.__val[0] & w)){
-				if(delflag) 
-					list_del(&pending->list);
-
-				if (delflag) {
-					mcs_rwlock_writer_unlock(lock, &mcs_rw_node);
-				}
-				else {
-					mcs_rwlock_reader_unlock(lock, &mcs_rw_node);
-				}
-				return pending;
-			}
-		}
-
-		if (delflag) {
-			mcs_rwlock_writer_unlock(lock, &mcs_rw_node);
-		}
-		else {
-			mcs_rwlock_reader_unlock(lock, &mcs_rw_node);
-		}
-
-		if(lock == &thread->sigpendinglock)
-			return NULL;
-
-		lock = &thread->sigpendinglock;
-		head = &thread->sigpending;
-	}
-
-	return NULL;
-}
-
-struct sig_pending *
-hassigpending(struct thread *thread)
-{
-	if (list_empty(&thread->sigpending) &&
-	    list_empty(&thread->sigcommon->sigpending)) {
-		return NULL;
-	}
-
-	return getsigpending(thread, 0);
-}
-
 int
 interrupt_from_user(void *regs0)
 {
@@ -1455,191 +1356,6 @@ void save_syscall_return_value(int num, unsigned long rc)
 	}
 }
 
-void
-check_signal(unsigned long rc, void *regs0, int num)
-{
-	__check_signal(rc, regs0, num, 0);
-}
-
-void
-check_signal_irq_disabled(unsigned long rc, void *regs0, int num)
-{
-	__check_signal(rc, regs0, num, 1);
-}
-
-static void
-__check_signal(unsigned long rc, void *regs0, int num, int irq_disabled)
-{
-	ihk_mc_user_context_t *regs = regs0;
-	struct thread *thread;
-	struct sig_pending *pending;
-	int	irqstate;
-
-	if(clv == NULL)
-		return;
-	thread = cpu_local_var(current);
-
-	if(thread == NULL || thread->proc->pid == 0){
-		struct thread *t;
-
-		irqstate = cpu_disable_interrupt_save();
-		ihk_mc_spinlock_lock_noirq(&(cpu_local_var(runq_lock)));
-		list_for_each_entry(t, &(cpu_local_var(runq)), sched_list){
-			if(t->proc->pid <= 0)
-				continue;
-			if(t->status == PS_INTERRUPTIBLE &&
-			   hassigpending(t)){
-				t->status = PS_RUNNING;
-				break;
-			}
-		}
-		ihk_mc_spinlock_unlock_noirq(&(cpu_local_var(runq_lock)));
-		cpu_restore_interrupt(irqstate);
-		goto out;
-	}
-
-	if(regs != NULL && !interrupt_from_user(regs)) {
-		goto out;
-	}
-
-	if (list_empty(&thread->sigpending) &&
-	    list_empty(&thread->sigcommon->sigpending)) {
-		goto out;
-	}
-
-	for(;;){
-		/* When this function called from check_signal_irq_disabled,
-		 * return with interrupt invalid.
-		 * This is to eliminate signal loss.
-		 */
-		if (irq_disabled == 1) {
-			irqstate = cpu_disable_interrupt_save();
-		}
-		pending = getsigpending(thread, 1);
-		if(!pending) {
-			dkprintf("check_signal,queue is empty\n");
-			goto out;
-		}
-		if (irq_disabled == 1) {
-			cpu_restore_interrupt(irqstate);
-		}
-		if (do_signal(rc, regs, thread, pending, num)) {
-			num = -1;
-		}
-	}
-
-out:
-	return;
-}
-
-static int
-check_sig_pending_thread(struct thread *thread)
-{
-	int found = 0;
-	struct list_head *head;
-	mcs_rwlock_lock_t *lock;
-	struct mcs_rwlock_node_irqsave mcs_rw_node;
-	struct sig_pending *next;
-	struct sig_pending *pending;
-	__sigset_t w;
-	__sigset_t x;
-	int sig = 0;
-	struct k_sigaction *k;
-	struct cpu_local_var *v;
-
-	v = get_this_cpu_local_var();
-	w = thread->sigmask.__val[0];
-
-	lock = &thread->sigcommon->lock;
-	head = &thread->sigcommon->sigpending;
-	for (;;) {
-		mcs_rwlock_reader_lock(lock, &mcs_rw_node);
-
-		list_for_each_entry_safe(pending, next, head, list) {
-			for (x = pending->sigmask.__val[0], sig = 0; x;
-			     sig++, x >>= 1)
-				;
-			k = thread->sigcommon->action + sig - 1;
-			if ((sig != SIGCHLD &&
-			     sig != SIGURG &&
-			     sig != SIGCONT) ||
-			    (k->sa.sa_handler != SIG_IGN &&
-			     k->sa.sa_handler != NULL)) {
-				if (!(pending->sigmask.__val[0] & w)) {
-					if (pending->interrupted == 0) {
-						pending->interrupted = 1;
-						found = 1;
-						if (sig != SIGCHLD &&
-						    sig != SIGURG &&
-						    sig != SIGCONT &&
-						    !k->sa.sa_handler) {
-							found = 2;
-							break;
-						}
-					}
-				}
-			}
-		}
-
-		mcs_rwlock_reader_unlock(lock, &mcs_rw_node);
-
-		if (found == 2) {
-			break;
-		}
-
-		if (lock == &thread->sigpendinglock) {
-			break;
-		}
-
-		lock = &thread->sigpendinglock;
-		head = &thread->sigpending;
-	}
-
-	if (found == 2) {
-		ihk_mc_spinlock_unlock(&v->runq_lock, v->runq_irqstate);
-		terminate_mcexec(0, sig);
-		return 1;
-	}
-	else if (found == 1) {
-		ihk_mc_spinlock_unlock(&v->runq_lock, v->runq_irqstate);
-		interrupt_syscall(thread, 0);
-		return 1;
-	}
-	return 0;
-}
-
-void
-check_sig_pending(void)
-{
-	struct thread *thread;
-	struct cpu_local_var *v;
-
-	if (clv == NULL)
-		return;
-
-	v = get_this_cpu_local_var();
-repeat:
-	v->runq_irqstate = ihk_mc_spinlock_lock(&v->runq_lock);
-	list_for_each_entry(thread, &(v->runq), sched_list) {
-
-		if (thread == NULL || thread == &cpu_local_var(idle)) {
-			continue;
-		}
-
-		if (thread->in_syscall_offload == 0) {
-			continue;
-		}
-
-		if (thread->proc->group_exit_status & 0x0000000100000000L) {
-			continue;
-		}
-
-		if (check_sig_pending_thread(thread))
-			goto repeat;
-	}
-	ihk_mc_spinlock_unlock(&v->runq_lock, v->runq_irqstate);
-}
-
 unsigned long
 do_kill(struct thread * thread, int pid, int tid, int sig, siginfo_t *info, int ptracecont)
 {
@@ -2010,9 +1726,18 @@ SYSCALL_DECLARE(mmap)
 
 	/* check arguments */
 	pgsize = PAGE_SIZE;
+#ifndef ENABLE_FUGAKU_HACKS
 	if (flags & MAP_HUGETLB) {
 		int hugeshift = flags & (0x3F << MAP_HUGE_SHIFT);
 
+		/* OpenMPI expects -EINVAL when trying to map
+		 * /dev/shm/ file with MAP_SHARED | MAP_HUGETLB
+		 */
+		if (!(flags & MAP_ANONYMOUS)) {
+			error = -EINVAL;
+			goto out;
+		}
+
 		if (hugeshift == 0) {
 			/* default hugepage size */
 			flags |= ihk_mc_get_linux_default_huge_page_shift() <<
@@ -2042,6 +1767,11 @@ SYSCALL_DECLARE(mmap)
 			goto out;
 		}
 	}
+#else
+	if (flags & MAP_HUGETLB) {
+		flags &= ~(MAP_HUGETLB);
+	}
+#endif
 
 #define	VALID_DUMMY_ADDR	((region->user_start + PTL3_SIZE - 1) & ~(PTL3_SIZE - 1))
 	addr = (flags & MAP_FIXED)? addr0: VALID_DUMMY_ADDR;

arch/arm64/kernel/traps.c

@@ -174,9 +174,14 @@ void bad_mode(struct pt_regs *regs, int reason, unsigned int esr)
 
 	arch_show_interrupt_context(regs);
 
+#ifdef ENABLE_TOFU
+	info.si_signo = SIGSTOP;
+	info.si_errno = 0;
+#else
 	info.si_signo = SIGILL;
 	info.si_errno = 0;
 	info.si_code  = ILL_ILLOPC;
+#endif
 	info._sifields._sigfault.si_addr  = (void*)regs->pc;
 
 	arm64_notify_die("Oops - bad mode", regs, &info, 0);

arch/x86_64/kernel/cpu.c

@@ -868,6 +868,49 @@ void show_context_stack(uintptr_t *rbp) {
 	return;
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+void __show_context_stack(struct thread *thread,
+        unsigned long pc, uintptr_t sp, int kprintf_locked)
+{
+    uintptr_t stack_top;
+    unsigned long irqflags = 0;
+
+    stack_top = ALIGN_UP(sp, (uintptr_t)KERNEL_STACK_SIZE);
+
+    if (!kprintf_locked)
+        irqflags = kprintf_lock();
+
+    __kprintf("TID: %d, call stack (most recent first):\n",
+        thread->tid);
+    __kprintf("PC: %016lx, SP: %016lx\n", pc, sp);
+    for (;;) {
+        extern char _head[], _end[];
+        uintptr_t *fp, *lr;
+        fp = (uintptr_t *)sp;
+        lr = (uintptr_t *)(sp + 8);
+
+        if ((*fp <= sp)) {
+            break;
+        }
+
+        if ((*fp > stack_top)) {
+            break;
+        }
+
+        if ((*lr < (unsigned long)_head) ||
+            (*lr > (unsigned long)_end)) {
+            break;
+        }
+
+        __kprintf("PC: %016lx, SP: %016lx, FP: %016lx\n", *lr - 4, sp, *fp);
+        sp = *fp;
+    }
+
+    if (!kprintf_locked)
+        kprintf_unlock(irqflags);
+}
+#endif
+
 void interrupt_exit(struct x86_user_context *regs)
 {
 	if (interrupt_from_user(regs)) {
@@ -1137,6 +1180,17 @@ void cpu_halt(void)
 	asm volatile("hlt");
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+/*@
+  @ assigns \nothing;
+  @ ensures \interrupt_disabled == 0;
+  @*/
+void cpu_halt_panic(void)
+{
+    cpu_halt();
+}
+#endif
+
 /*@
   @ assigns \nothing;
   @ ensures \interrupt_disabled == 0;
@@ -1521,6 +1575,16 @@ void arch_print_stack(void)
 	__print_stack(rbp, 0);
 }
 
+#ifdef ENABLE_FUGAKU_HACKS
+unsigned long arch_get_instruction_address(const void *reg)
+{
+	const struct x86_user_context *uctx = reg;
+	const struct x86_basic_regs *regs = &uctx->gpr;
+
+	return regs->rip;
+}
+#endif
+
 /*@
   @ requires \valid(reg);
   @ assigns \nothing;

arch/x86_64/kernel/include/arch-memory.h

@@ -451,4 +451,12 @@ extern unsigned long ap_trampoline;
 
 /* Local is cachable */
 #define IHK_IKC_QUEUE_PT_ATTR (PTATTR_NO_EXECUTE | PTATTR_WRITABLE)
+
+#ifdef ENABLE_FUGAKU_HACKS
+#ifndef __ASSEMBLY__
+# define ALIGN_UP(x, align)     ALIGN_DOWN((x) + (align) - 1, align)
+# define ALIGN_DOWN(x, align)   ((x) & ~((align) - 1))
+#endif /* !__ASSEMBLY__ */
+#endif
+
 #endif

arch/x86_64/kernel/include/cpulocal.h

@@ -53,5 +53,9 @@ struct x86_cpu_local_variables *get_x86_this_cpu_local(void);
 void *get_x86_cpu_local_kstack(int id);
 void *get_x86_this_cpu_kstack(void);
 
+#ifdef ENABLE_FUGAKU_HACKS
+#define LOCALS_SPAN (4 * PAGE_SIZE)
+#define KERNEL_STACK_SIZE LOCALS_SPAN
+#endif
 
 #endif

arch/x86_64/kernel/include/rlimit.h

@@ -84,7 +84,11 @@ enum __rlimit_resource
   __RLIMIT_RTPRIO = 14,
 #define RLIMIT_RTPRIO __RLIMIT_RTPRIO
 
-  __RLIMIT_NLIMITS = 15,
+  /* timeout for RT tasks in us */
+  __RLIMIT_RTTIME = 15,
+#define RLIMIT_RTTIME __RLIMIT_RTTIME
+
+  __RLIMIT_NLIMITS = 16,
   __RLIM_NLIMITS = __RLIMIT_NLIMITS
 #define RLIMIT_NLIMITS __RLIMIT_NLIMITS
 #define RLIM_NLIMITS __RLIM_NLIMITS

arch/x86_64/kernel/include/syscall_list.h

@@ -74,6 +74,7 @@ SYSCALL_DELEGATED(89, readlink)
 SYSCALL_HANDLED(96, gettimeofday)
 SYSCALL_HANDLED(97, getrlimit)
 SYSCALL_HANDLED(98, getrusage)
+SYSCALL_HANDLED(99, sysinfo)
 SYSCALL_HANDLED(100, times)
 SYSCALL_HANDLED(101, ptrace)
 SYSCALL_HANDLED(102, getuid)
@@ -157,6 +158,7 @@ SYSCALL_HANDLED(289, signalfd4)
 #ifdef ENABLE_PERF
 SYSCALL_HANDLED(298, perf_event_open)
 #endif
+SYSCALL_HANDLED(302, prlimit64)
 #ifdef DCFA_KMOD
 SYSCALL_HANDLED(303, mod_call)
 #endif
@@ -166,7 +168,7 @@ SYSCALL_HANDLED(311, process_vm_writev)
 SYSCALL_HANDLED(322, execveat)
 SYSCALL_HANDLED(700, get_cpu_id)
 #ifdef PROFILE_ENABLE
-SYSCALL_HANDLED(__NR_profile, profile)
+SYSCALL_HANDLED(PROFILE_EVENT_MAX, profile)
 #endif // PROFILE_ENABLE
 SYSCALL_HANDLED(730, util_migrate_inter_kernel)
 SYSCALL_HANDLED(731, util_indicate_clone)
@@ -179,4 +181,8 @@ SYSCALL_HANDLED(802, linux_mlock)
 SYSCALL_HANDLED(803, suspend_threads)
 SYSCALL_HANDLED(804, resume_threads)
 SYSCALL_HANDLED(811, linux_spawn)
-/**** End of File ****/
+
+/* Do not edit the lines including this comment and
+ * EOF just after it because those are used as a
+ * robust marker for the autotest patch.
+ */

arch/x86_64/kernel/local.c

@@ -21,7 +21,9 @@
 #include <registers.h>
 #include <string.h>
 
+#ifndef ENABLE_FUGAKU_HACKS
 #define LOCALS_SPAN (4 * PAGE_SIZE)
+#endif
 
 struct x86_cpu_local_variables *locals;
 size_t x86_cpu_local_variables_span = LOCALS_SPAN;	/* for debugger */

arch/x86_64/kernel/memory.c

@@ -1651,6 +1651,14 @@ static int clear_range(struct page_table *pt, struct process_vm *vm,
 	if (memobj && ((memobj->flags & MF_PREMAP))) {
 		args.free_physical = 0;
 	}
+
+	if (vm->proc->straight_va &&
+			(void *)start == vm->proc->straight_va &&
+			(void *)end == (vm->proc->straight_va +
+				vm->proc->straight_len)) {
+		args.free_physical = 0;
+	}
+
 	args.memobj = memobj;
 	args.vm = vm;
 

arch/x86_64/kernel/syscall.c

@@ -147,44 +147,6 @@ int obtain_clone_cpuid(cpu_set_t *cpu_set, int use_last) {
     return min_cpu;
 }
 
-int
-arch_clear_host_user_space()
-{
-	struct thread *th = cpu_local_var(current);
-
-	/* XXX: might be unnecessary */
-	clear_host_pte(th->vm->region.user_start,
-			(th->vm->region.user_end - th->vm->region.user_start), 0);
-	return 0;
-}
-
-SYSCALL_DECLARE(rt_sigaction)
-{
-	int sig = ihk_mc_syscall_arg0(ctx);
-	const struct sigaction *act = (const struct sigaction *)ihk_mc_syscall_arg1(ctx);
-	struct sigaction *oact = (struct sigaction *)ihk_mc_syscall_arg2(ctx);
-	size_t sigsetsize = ihk_mc_syscall_arg3(ctx);
-	struct k_sigaction new_sa, old_sa;
-	int rc;
-
-	if (sigsetsize != sizeof(sigset_t))
-		return -EINVAL;
-
-	if(act)
-		if(copy_from_user(&new_sa.sa, act, sizeof new_sa.sa)){
-			goto fault;
-		}
-	rc = do_sigaction(sig, act? &new_sa: NULL, oact? &old_sa: NULL);
-	if(rc == 0 && oact)
-		if(copy_to_user(oact, &old_sa.sa, sizeof old_sa.sa)){
-			goto fault;
-		}
-
-	return rc;
-fault:
-	return -EFAULT;
-}
-
 SYSCALL_DECLARE(prctl)
 {
 	struct process *proc = cpu_local_var(current)->proc;
@@ -1039,82 +1001,6 @@ out:
 	return restart;
 }
 
-static struct sig_pending *
-getsigpending(struct thread *thread, int delflag){
-	struct list_head *head;
-	mcs_rwlock_lock_t *lock;
-	struct mcs_rwlock_node_irqsave mcs_rw_node;
-	struct sig_pending *next;
-	struct sig_pending *pending;
-	__sigset_t w;
-	__sigset_t x;
-	int sig;
-	struct k_sigaction *k;
-
-	w = thread->sigmask.__val[0];
-
-	lock = &thread->sigcommon->lock;
-	head = &thread->sigcommon->sigpending;
-	for(;;) {
-		if (delflag) {
-			mcs_rwlock_writer_lock(lock, &mcs_rw_node);
-		}
-		else {
-			mcs_rwlock_reader_lock(lock, &mcs_rw_node);
-		}
-
-		list_for_each_entry_safe(pending, next, head, list){
-			for(x = pending->sigmask.__val[0], sig = 0; x; sig++, x >>= 1);
-			k = thread->sigcommon->action + sig - 1;
-			if(delflag ||
-			   (sig != SIGCHLD &&
-			    sig != SIGURG &&
-			    sig != SIGCONT) ||
-			   (k->sa.sa_handler != (void *)1 &&
-			    k->sa.sa_handler != NULL)){
-				if(!(pending->sigmask.__val[0] & w)){
-					if(delflag) 
-						list_del(&pending->list);
-
-					if (delflag) {
-						mcs_rwlock_writer_unlock(lock, &mcs_rw_node);
-					}
-					else {
-						mcs_rwlock_reader_unlock(lock, &mcs_rw_node);
-					}
-					return pending;
-				}
-			}
-		}
-
-		if (delflag) {
-			mcs_rwlock_writer_unlock(lock, &mcs_rw_node);
-		}
-		else {
-			mcs_rwlock_reader_unlock(lock, &mcs_rw_node);
-		}
-
-		if(lock == &thread->sigpendinglock)
-			return NULL;
-
-		lock = &thread->sigpendinglock;
-		head = &thread->sigpending;
-	}
-
-	return NULL;
-}
-
-struct sig_pending *
-hassigpending(struct thread *thread)
-{
-	if (list_empty(&thread->sigpending) &&
-	    list_empty(&thread->sigcommon->sigpending)) {
-		return NULL;
-	}
-
-	return getsigpending(thread, 0);
-}
-
 int
 interrupt_from_user(void *regs0)
 {
@@ -1129,175 +1015,6 @@ void save_syscall_return_value(int num, unsigned long rc)
 	return;
 }
 
-/** \brief check arrived signals and processing
- *
- * @param rc    return value of syscall
- * @param regs0 context
- * @param num   syscall number (-1: Not called on exiting system call)
- */
-void
-check_signal(unsigned long rc, void *regs0, int num)
-{
-	struct x86_user_context *regs = regs0;
-	struct thread *thread;
-	struct sig_pending *pending;
-	int	irqstate;
-
-	if(clv == NULL)
-		return;
-	thread = cpu_local_var(current);
-
-	if(thread == NULL || thread == &cpu_local_var(idle)){
-		struct thread *t;
-
-		irqstate = cpu_disable_interrupt_save();
-		ihk_mc_spinlock_lock_noirq(&(cpu_local_var(runq_lock)));
-		list_for_each_entry(t, &(cpu_local_var(runq)), sched_list){
-			if(t == &cpu_local_var(idle))
-				continue;
-			if(t->status == PS_INTERRUPTIBLE &&
-			   hassigpending(t)){
-				t->status = PS_RUNNING;
-				break;
-			}
-		}
-		ihk_mc_spinlock_unlock_noirq(&(cpu_local_var(runq_lock)));
-		cpu_restore_interrupt(irqstate);
-		goto out;
-	}
-
-	if(regs != NULL && !interrupt_from_user(regs)) {
-		goto out;
-	}
-
-	if (list_empty(&thread->sigpending) &&
-	    list_empty(&thread->sigcommon->sigpending)) {
-		goto out;
-	}
-
-	for(;;){
-		pending = getsigpending(thread, 1);
-		if(!pending) {
-			dkprintf("check_signal,queue is empty\n");
-			goto out;
-		}
-
-		if (do_signal(rc, regs, thread, pending, num)) {
-			num = -1;
-		}
-	}
-
-out:
-	return;
-}
-
-static int
-check_sig_pending_thread(struct thread *thread)
-{
-	int found = 0;
-	struct list_head *head;
-	mcs_rwlock_lock_t *lock;
-	struct mcs_rwlock_node_irqsave mcs_rw_node;
-	struct sig_pending *next;
-	struct sig_pending *pending;
-	__sigset_t w;
-	__sigset_t x;
-	int sig = 0;
-	struct k_sigaction *k;
-	struct cpu_local_var *v;
-
-	v = get_this_cpu_local_var();
-	w = thread->sigmask.__val[0];
-
-	lock = &thread->sigcommon->lock;
-	head = &thread->sigcommon->sigpending;
-	for (;;) {
-		mcs_rwlock_reader_lock(lock, &mcs_rw_node);
-
-		list_for_each_entry_safe(pending, next, head, list){
-			for (x = pending->sigmask.__val[0], sig = 0; x;
-			     sig++, x >>= 1);
-			k = thread->sigcommon->action + sig - 1;
-			if ((sig != SIGCHLD &&
-			     sig != SIGURG &&
-			     sig != SIGCONT) ||
-			    (k->sa.sa_handler != (void *)1 &&
-			     k->sa.sa_handler != NULL)) {
-				if (!(pending->sigmask.__val[0] & w)) {
-					if (pending->interrupted == 0) {
-						pending->interrupted = 1;
-						found = 1;
-						if (sig != SIGCHLD &&
-						    sig != SIGURG &&
-						    sig != SIGCONT &&
-						    !k->sa.sa_handler) {
-							found = 2;
-							break;
-						}
-					}
-				}
-			}
-		}
-
-		mcs_rwlock_reader_unlock(lock, &mcs_rw_node);
-
-		if (found == 2) {
-			break;
-		}
-
-		if (lock == &thread->sigpendinglock) {
-			break;
-		}
-
-		lock = &thread->sigpendinglock;
-		head = &thread->sigpending;
-	}
-
-	if (found == 2) {
-		ihk_mc_spinlock_unlock(&v->runq_lock, v->runq_irqstate);
-		terminate_mcexec(0, sig);
-		return 1;
-	}
-	else if (found == 1) {
-		ihk_mc_spinlock_unlock(&v->runq_lock, v->runq_irqstate);
-		interrupt_syscall(thread, 0);
-		return 1;
-	}
-	return 0;
-}
-
-void
-check_sig_pending(void)
-{
-	struct thread *thread;
-	struct cpu_local_var *v;
-
-	if (clv == NULL)
-		return;
-
-	v = get_this_cpu_local_var();
-repeat:
-	v->runq_irqstate = ihk_mc_spinlock_lock(&v->runq_lock);
-	list_for_each_entry(thread, &(v->runq), sched_list) {
-
-		if (thread == NULL || thread == &cpu_local_var(idle)) {
-			continue;
-		}
-
-		if (thread->in_syscall_offload == 0) {
-			continue;
-		}
-
-		if (thread->proc->group_exit_status & 0x0000000100000000L) {
-			continue;
-		}
-
-		if (check_sig_pending_thread(thread))
-			goto repeat;
-	}
-	ihk_mc_spinlock_unlock(&v->runq_lock, v->runq_irqstate);
-}
-
 unsigned long
 do_kill(struct thread *thread, int pid, int tid, int sig, siginfo_t *info,
         int ptracecont)
@@ -1713,6 +1430,14 @@ SYSCALL_DECLARE(mmap)
 	/* check arguments */
 	pgsize = PAGE_SIZE;
 	if (flags & MAP_HUGETLB) {
+		/* OpenMPI expects -EINVAL when trying to map
+		 * /dev/shm/ file with MAP_SHARED | MAP_HUGETLB
+		 */
+		if (!(flags & MAP_ANONYMOUS)) {
+			error = -EINVAL;
+			goto out;
+		}
+
 		switch (flags & (0x3F << MAP_HUGE_SHIFT)) {
 		case 0:
 			/* default hugepage size */

cmake/modules/Kbuild.cmake

@@ -30,6 +30,9 @@ endif ()
 
 if (NOT "${LINUX_ARCH}" STREQUAL "${CMAKE_HOST_SYSTEM_PROCESSOR}")
 	string(REGEX REPLACE "ld$" "" CROSS_COMPILE "${CMAKE_LINKER}")
+	if (CMAKE_CROSSCOMPILING)
+	   list(APPEND KBUILD_MAKE_FLAGS "QEMU_LD_PREFIX=${CMAKE_FIND_ROOT_PATH}")
+	endif()
 	list(APPEND KBUILD_MAKE_FLAGS "ARCH=${ARCH}")
 	list(APPEND KBUILD_MAKE_FLAGS "CROSS_COMPILE=${CROSS_COMPILE}")
 endif()

docs/Makefile

@@ -0,0 +1,29 @@
+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line.
+SPHINXOPTS    =
+SPHINXBUILD   = sphinx-build
+SOURCEDIR     = .
+BUILDDIR      = _build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+po:
+	sphinx-build -b gettext . _build/gettext
+	/opt/local/Library/Frameworks/Python.framework/Versions/3.8/bin/sphinx-intl update -p _build/gettext -l ja
+
+ja:
+	sphinx-build -b html -D language=ja . _build/html/ja
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	echo ".. figure:: docs/mckernel-logo.png" > ../README.rst; echo >> ../README.rst
+	cat summary.rst >> ../README.rst; echo >> ../README.rst
+	cat doc.rst >> ../README.rst
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)

docs/NEWS.rst

@@ -0,0 +1,446 @@
+=============================================
+Version 1.7.0 (Nov 25, 2020)
+=============================================
+
+----------------------
+IHK major updates
+----------------------
+#. ihklib: add ihk_create_os_str
+#. ihklib: ihk_reserve_mem: add capped best effort to balanced
+
+------------------------
+IHK major bug fixes
+------------------------
+#. make /dev/mcdN sharable
+#. acpi: compat: RHEL-8.2
+#. gic_chip_data: compat: RHEL-8.3
+
+----------------------
+McKernel major updates
+----------------------
+#. arm64: Contiguous PTE support
+#. arm64: Scalable Vector Extension (SVE) support
+#. arm64: PMU overflow interrupt support
+#. arm64 port: Direct access to Mckernel memory from Linux
+#. arm64 port: utility thread offloading, which spawns thread onto Linux CPU
+#. eclair: support for live debug
+#. Crash utility extension
+#. Replace mcoverlayfs with a soft userspace overlay
+#. Build system is switched to cmake
+#. Core dump includes thread information
+#. mcinspect and mcps: DWARF based LWK inspection
+
+------------------------
+McKernel major bug fixes
+------------------------
+#. shmobj: Fix rusage counting for large page
+#. mcctrl control: task start_time changed to u64 nsec
+#. mcctrl: add handling for one more level of page tables
+#. Add kernel argument to turn on/off time sharing
+#. flatten_string / process env: realign env and clear trailing bits
+#. madvise: Add MADV_HUGEPAGE support
+#. mcctrl: remove in-kernel calls to syscalls
+#. arch_cpu_read_write_register: error return fix.
+#. set_cputime(): interrupt enable/disable fix.
+#. set_mempolicy(): Add mode check.
+#. mbind(): Fix memory_range_lock deadlock.
+#. ihk_ikc_recv: Record channel to packet for release
+#. Add set_cputime() kernel to kernel case and mode enum.
+#. execve: Call preempt_enable() before error-exit
+#. memory/x86_64: fix linux safe_kernel_map
+#. do_kill(): fix pids table when nr of threads is larger than num_processors
+#. shmget: Use transparent huge pages when page size isn't specified
+#. prctl: Add support for PR_SET_THP_DISABLE and PR_GET_THP_DISABLE
+#. monitor_init: fix undetected hang on highest numbered core
+#. init_process_stack: change premapped stack size based on arch
+#. x86 syscalls: add a bunch of XXat() delegated syscalls
+#. do_pageout: fix direct kernel-user access
+#. stack: add hwcap auxval
+#. perf counters: add arch-specific perf counters
+#. Added check of nohost to terminate_host().
+#. kmalloc: Fix address order in free list
+#. sysfs: use nr_cpu_ids for cpumasks (fixes libnuma parsing error on ARM)
+#. monitor_init: Use ihk_mc_cpu_info()
+#. Fix ThunderX2 write-combined PTE flag insanity
+#. ARM: eliminate zero page mapping (i.e, init_low_area())
+#. eliminate futex_cmpxchg_enabled check (not used and dereffed a NULL pointer)
+#. page_table: Fix return value of lookup_pte when ptl4 is blank
+#. sysfs: add missing symlinks for cpu/node
+#. Make Linux handler run when mmap to procfs.
+#. Separate mmap area from program loading (relocation) area
+#. move rusage into kernel ELF image (avoid dynamic alloc before NUMA init)
+#. arm: turn off cpu on panic
+#. page fault handler: protect thread accesses
+#. Register PPD and release_handler at the same time.
+#. fix to missing exclusive processing between terminate() and finalize_process().
+#. perfctr_stop: add flags to no 'disable_intens'
+#. fileobj, shmobj: free pages in object destructor (as opposed to page_unmap())
+#. clear_range_l1, clear_range_middle: Fix handling contiguous PTE
+#. do_mmap: don't pre-populate the whole file when asked for smaller segment
+#. invalidate_one_page: Support shmobj and contiguous PTE
+#. ubsan: fix undefined shifts
+#. x86: disable zero mapping and add a boot pt for ap trampoline
+#. rusage: Don't count PF_PATCH change
+#. Fixed time processing.
+#. copy_user_pte: vmap area not owned by McKernel
+#. gencore: Zero-clear ELF header and memory range table
+#. rpm: ignore CMakeCache.txt in dist and relax BuildRequires on cross build
+#. gencore: Allocate ELF header to heap instead of stack
+#. nanosleep: add cpu_pause() in spinwait loop
+#. init_process: add missing initializations to proc struct
+#. rus_vm_fault: always use a packet on the stack
+#. process stack: use PAGE_SIZE in aux vector
+#. copy_user_pte: base memobj copy on range & VR_PRIVATE
+#. arm64: ptrace: Fix overwriting 1st argument with return value
+#. page fault: use cow for private device mappings
+#. reproductible builds: remove most install paths in c code
+#. page fault: clear writable bit for non-dirtying access to shared ranges
+#. mcreboot/mcstop+release: support for regular user execution
+#. irqbalance_mck: replace extra service with service drop-in
+#. do_mmap: give addr argument a chance even if not MAP_FIXED
+#. x86: fix xchg() and cmpxchg() macros
+#. IHK: support for using Linux work IRQ as IKC interrupt (optional)
+#. MCS: fix ARM64 issue by using smp_XXX() functions (i.e., barrier()s)
+#. procfs: add number of threads to stat and status
+#. memory_range_lock: Fix deadlock in procfs/sysfs handler
+#. flush instruction cache at context switch time if necessary
+#. arm64: Fix PMU related functions
+#. page_fault_process_memory_range: Disable COW for VM region with zeroobj
+#. extend_process_region: Fall back to demand paging when not contiguous
+#. munmap: fix deadlock with remote pagefault on vm range lock
+#. procfs: if memory_range_lock fails, process later
+#. migrate-cpu: Prevent migration target from calling schedule() twice
+#. sched_request_migrate(): fix race condition between migration req and IRQs
+#. get_one_cpu_topology: Renumber core_id (physical core id)
+#. bb7e140 procfs cpuinfo: use sequence number as processor
+#. set_host_vma(): do NOT read protect Linux VMA
+#. hugefileobj: rewrite page allocation/handling
+#. VM: use RW spinlock for vm_range_lock
+#. /dev/shm: use Linux PFNs and populate mappings
+#. Make struct ihk_os_rusage compatible with mckernel_rusage (workaround for Fugaku)
+#. Record pthread routine address in clone(), keep helper threads on caller CPU core (workaround for Fugaku)
+#. struct process: fix type of group_exit_status
+#. tgkill: Fix argument validatation
+#. set_robust_list: Add error check
+#. mcexec: Don't forward SIGTSTP SIGTTIN SIGTTOUT to mckernel
+#. syscall: add prlimit64
+#. stack: grow on page fault
+#. mcexec: use FLIB_NUM_PROCESS_ON_NODE when -n not specified (Fugaku specific)
+
+===========================================
+Version 1.6.0 (Nov 11, 2018)
+===========================================
+
+-----------------------------------------------
+McKernel major updates
+-----------------------------------------------
+#. McKernel and Linux share one unified kernel virtual address space.
+   That is, McKernel sections resides in Linux sections spared for
+   modules.  In this way, Linux can access the McKernel kernel memory area.
+#. hugetlbfs support
+#. IHK is now included as a git submodule
+#. Debug messages are turned on/off in per souce file basis at run-time.
+#. It's prohibited for McKernel to access physical memory ranges which Linux didn't give to McKernel.
+#. UTI (capability to spawn a thread on Linux CPU) improvement:
+
+   * System calls issued from the thread are hooked by modifying binary in memory.
+
+---------------------------
+McKernel major bug fixes
+---------------------------
+#<digits> below denotes the redmine issue number (https://postpeta.pccluster.org/redmine/).
+
+1. #926: shmget: Hide object with IPC_RMID from shmget
+2. #1028: init_process: Inherit parent cpu_set
+3. #995: Fix shebang recorded in argv[0]
+4. #1024: Fix VMAP virtual address leak
+5. #1109: init_process_stack: Support "ulimit -s unlimited"
+6. x86 mem init: do not map identity mapping
+7. mcexec_wait_syscall: requeue potential request on interrupted wait
+8. mcctrl_ikc_send_wait: fix interrupt with do_frees == NULL
+9. pager_req_read: handle short read
+10. kprintf: only call eventfd() if it is safe to interrupt
+11. process_procfs_request: Add Pid to /proc/<PID>/status
+12. terminate: fix oversubscribe hang when waiting for other threads on same CPU to die
+13. mcexec: Do not close fd returned to mckernel side
+14. #976: execve: Clear sigaltstack and fp_regs
+15. #1002: perf_event: Specify counter by bit_mask on start/stop
+16. #1027: schedule: Don't reschedule immediately when wake up on migrate
+17. #mcctrl: lookup unexported symbols at runtime
+18. __sched_wakeup_thread: Notify interrupt_exit() of re-schedule
+19. futex_wait_queue_me: Spin-sleep when timeout and idle_halt is specified
+20. #1167: ihk_os_getperfevent,setperfevent: Timeout IKC sent by mcctrl
+21. devobj: fix object size (POSTK_DEBUG_TEMP_FIX_36)
+22. mcctrl: remove rus page cache
+23. #1021: procfs: Support multiple reads of e.g. ``/proc/*/maps``
+24. #1006: wait: Delay wake-up parent within switch context
+25. #1164: mem: Check if phys-mem is within the range of McKernel memory
+26. #1039: page_fault_process_memory_range: Remove ihk_mc_map_virtual for CoW of device map
+27. partitioned execution: pass process rank to LWK
+28. process/vm: implement access_ok()
+29. spinlock: rewrite spinlock to use Linux ticket head/tail format
+30. #986: Fix deadlock involving mmap_sem and memory_range_lock
+31. Prevent one CPU from getting chosen by concurrent forks
+32. #1009: check_signal: system call restart is done only once
+33. #1176: syscall: the signal received during system call processing is not processed.
+34. #1036 syscall_time: Handle by McKernel
+35. #1165 do_syscall: Delegate system calls to the mcexec with the same pid
+36. #1194 execve: Fix calling ptrace_report_signal after preemption is disabled
+37. #1005 coredump: Exclude special areas
+38. #1018 procfs: Fix pread/pwrite to procfs fail when specified size is bigger than 4MB
+39. #1180 sched_setaffinity: Check migration after decrementing in_interrupt
+40. #771, #1179, #1143 ptrace supports threads
+41. #1189 procfs/do_fork: wait until procfs entries are registered
+42. #1114 procfs: add '/proc/pid/stat' to mckernel side and fix its comm
+43. #1116 mcctrl procfs: check entry was returned before using it
+44. #1167 ihk_os_getperfevent,setperfevent: Return -ETIME when IKC timeouts
+45. mcexec/execve: fix shebangs handling
+46. procfs: handle 'comm' on mckernel side
+47. ihk_os_setperfevent: Return number of registered events
+48. mcexec: fix terminating zero after readlink()
+
+===========================================
+Version 1.5.1 (July 9, 2018)
+===========================================
+
+-----------------------------------------------
+McKernel major updates
+-----------------------------------------------
+
+Watchdog timer to detect hang of McKernel
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+mcexec prints out the following line to its stderr when a hang of McKernel is detected.
+
+::
+   
+       mcexec detected hang of McKernel
+
+The watchdog timer is enabled by passing -i <timeout_in_sec> option to mcreboot.sh. <timeout_in_sec> specifies the interval of checking if McKernel is alive.
+
+For example, specify ``-i 600`` to detect the hang with 10 minutes interval:
+
+::
+
+   mcreboot.sh -i 600
+
+The detailed step of the hang detection is as follows.
+   #. mcexec acquires eventfd for notification from IHK and perform epoll() on it.
+   #. A daemon called ihkmond monitors the state of McKernel periodically with the interval specified by the -i option. It judges that McKernel is hanging and notifies mcexec by the eventfd if its state hasn't changed since the last check.
+
+---------------------------
+McKernel major bug fixes
+---------------------------
+1.  #1146: pager_req_map(): do not take mmap_sem if not needed
+2.  #1135: prepare_process_ranges_args_envs(): fix saving cmdline
+3.  #1144: fileobj/devobj: record path name
+4.  #1145: fileobj: use MCS locks for per-file page hash
+5.  #1076: mcctrl: refactor prepare_image into new generic ikc send&wait
+6.  #1072: execve: fix execve with oversubscribing
+7.  #1132: execve: use thread variable instead of cpu_local_var(current)
+8.  #1117: mprotect: do not set page table writable for cow pages
+9.  #1143: syscall wait4: add _WALL (POSTK_DEBUG_ARCH_DEP_44)
+10. #1064: rusage: Fix initialization of rusage->num_processors
+11. #1133: pager_req_unmap: Put per-process data at exit
+12. #731: do_fork: Propagate error code returned by mcexec
+13. #1149: execve: Reinitialize vm_regions's map area on execve
+14. #1065: procfs: Show file names in /proc/<PID>/maps
+15. #1112: mremap: Fix type of size arguments (from ssize_t to size_t)
+16. #1121: sched_getaffinity: Check arguments in the same order as in Linux
+17. #1137: mmap, mremap: Check arguments in the same order as in Linux
+18. #1122: fix return value of sched_getaffinity
+19. #732: fix: /proc/<PID>/maps outputs a unnecessary NULL character
+
+===================================
+Version 1.5.0 (Apr 5, 2018)
+===================================
+
+--------------------------------------
+McKernel major updates
+--------------------------------------
+1. Aid for Linux version migration: Detect /proc, /sys format change
+   between two kernel verions
+2. Swap out
+   * Only swap-out anonymous pages for now
+3. Improve support of /proc/maps
+4. mcstat: Linux tool to show resource usage
+
+---------------------------
+McKernel major bug fixes
+---------------------------
+#. #727: execve: Fix memory leak when receiving SIGKILL
+#. #829: perf_event_open: Support PERF_TYPE_HARDWARE and PERF_TYPE_HW_CACHE
+#. #906: mcexec: Check return code of fork()
+#. #1038: mcexec: Timeout when incorrect value is given to -n option
+#. #943 #945 #946 #960 #961: mcexec: Support strace
+#. #1029: struct thread is not released with stress-test involving signal and futex
+#. #863 #870 : Respond immediately to terminating signal when offloading system call
+#. #1119: translate_rva_to_rpa(): use 2MB blocks in 1GB pages on x86
+#. #898: Shutdown OS only after no in-flight IKC exist
+#. #882: release_handler: Destroy objects as the process which opened it
+#. #882: mcexec: Make child process exit if the parent is killed during fork()
+#. #925: XPMEM: Don't destroy per-process object of the parent
+#. #885: ptrace: Support the case where a process attaches its child
+#. #1031: sigaction: Support SA_RESETHAND
+#. #923: rus_vm_fault: Return error when a thread not performing system call offloading causes remote page fault
+#. #1032 #1033 #1034: getrusage: Fix ru_maxrss, RUSAGE_CHILDREN, ru_stime related bugs
+#. #1120: getrusage: Fix deadlock on thread->times_update
+#. #1123: Fix deadlock related to wait_queue_head_list_node
+#. #1124: Fix deadlock of calling terminate() from terminate()
+#. #1125: Fix deadlock related to thread status
+
+   * Related functions are: hold_thread(), do_kill() and terminate()
+
+#. #1126: uti: Fix uti thread on the McKernel side blocks others in do_syscall()
+#. #1066: procfs: Show Linux /proc/self/cgroup
+#. #1127: prepare_process_ranges_args_envs(): fix generating saved_cmdline to avoid PF in strlen()
+#. #1128: ihk_mc_map/unmap_virtual(): do proper TLB invalidation
+#. #1043: terminate(): fix update_lock and threads_lock order to avoid deadlock
+#. #1129: mcreboot.sh: Save ``/proc/irq/*/smp_affinity`` to ``/tmp/mcreboot``
+#. #1130: mcexec: drop READ_IMPLIES_EXEC from personality
+
+--------------------
+McKernel workarounds
+--------------------
+#. Forbid CPU oversubscription
+
+   * It can be turned on by mcreboot.sh -O option
+
+
+===================================
+Version 1.4.0 (Oct 30, 2017)
+===================================
+
+-----------------------------------------------------------
+Abstracted event type support in perf_event_open()
+-----------------------------------------------------------
+
+PERF_TYPE_HARDWARE and PERF_TYPE_CACHE types are supported.
+
+----------------------------------
+Direct user-space access
+----------------------------------
+Code lines using direct user-space access (e.g. passing user-space
+pointer to memcpy()) becomes more portable across processor
+architectures. The modification follows the following rules.
+
+1. Move the code section as it is to the architecture dependent
+   directory if it is a part of the critical-path.
+2. Otherwise, rewrite the code section by using the portable methods.
+   The methods include copy_from_user(), copy_to_user(),
+   pte_get_phys() and phys_to_virt().
+
+--------------------------------
+MPI and OpenMP micro-bench tests
+--------------------------------
+The performance figures of MPI and OpenMP primitives are compared with
+those of Linux by using Intel MPI Benchmarks and EPCC OpenMP Micro
+Benchmark.
+
+
+===================================
+Version 1.3.0 (Sep 30, 2017)
+===================================
+
+--------------------
+Kernel dump
+--------------------
+#. A dump level of "only kernel memory" is added.
+
+The following two levels are available now:
+
++--+-----------------------+
+| 0|Dump all               |
++--+-----------------------+
+|24|Dump only kernel memory|
++--+-----------------------+
+
+The dump level can be set by -d option in ihkosctl or the argument
+for ihk_os_makedumpfile(), as shown in the following examples:
+
+::
+
+   Command:		ihkosctl 0 dump -d 24
+   Function call:	ihk_os_makedumpfile(0, NULL, 24, 0);
+
+#. Dump file is created when Linux panics.
+
+The dump level can be set by dump_level kernel argument, as shown in the
+following example:
+
+::
+
+   ihkosctl 0 kargs "hidos dump_level=24"
+
+The IHK dump function is registered to panic_notifier_list when creating /dev/mcdX and called when Linux panics.
+
+-----------------------------
+Quick Process Launch
+-----------------------------
+
+MPI process launch time and some of the initialization time can be
+reduced in application consisting of multiple MPI programs which are
+launched in turn in the job script.
+
+The following two steps should be performed to use this feature:
+#. Replace mpiexec with ql_mpiexec_start and add some lines for ql_mpiexec_finalize in the job script
+#. Modify the app so that it can repeat calculations and wait for the instructions from ql_mpiexec_{start,finalize} at the end of the loop
+
+The first step is explained using an example. Assume the original job script looks like this:
+
+.. code-block:: none
+   
+   /* Execute ensamble simulation and then data assimilation, and repeat this ten times */
+   for i in {1..10}; do
+   
+      /* Each ensamble simulation execution uses 100 nodes, launch ten of them in parallel */
+      for j in {1..10}; do
+         mpiexec -n 100 -machinefile ./list1_$j p1.out a1 & pids[$i]=$!;
+      done
+
+      /* Wait until the ten ensamble simulation programs finish */
+      for j in {1..10}; do wait ${pids[$j]}; done
+      
+      /* Launch one data assimilation program using 1000 nodes */
+      mpiexec -n 1000 -machinefile ./list2 p2.out a2
+   done
+   
+The job script should be modified like this:
+
+.. code-block:: none
+
+   for i in {1..10}; do
+      for j in {1..10}; do
+         /*  Replace mpiexec with ql_mpiexec_start */
+         ql_mpiexec_start -n 100 -machinefile ./list1_$j p1.out a1 & pids[$j]=$!;
+      done
+      
+      for j in {1..10}; do wait ${pids[$j]}; done
+      
+      ql_mpiexec_start -n 1000 -machinefile ./list2 p2.out a2
+   done
+   
+   /* p1.out and p2.out don't exit but are waiting for the next calculation. So tell them to exit */
+   for j in {1..10}; do
+      ql_mpiexec_finalize -machinefile ./list1_$i p1.out a1;
+   done
+   
+   ql_mpiexec_finalize -machinefile ./list2 p2.out a2;
+
+The second step is explained using a pseudo-code.
+
+.. code-block:: none
+
+   MPI_Init();
+   Prepare data exchange with preceding / following MPI programs
+   loop:
+   foreach Fortran module
+      Initialize data using command-line argments, parameter files, environment variables
+      Input data from preceding MPI programs / Read snap-shot
+      Perform main calculation
+      Output data to following MPI programs / Write snap-shot
+      /* ql_client() waits for command of ql_mpiexec_{start,finish} */
+      if (ql_client() == QL_CONTINUE) { goto loop; }
+      MPI_Finalize();
+
+qlmpilib.h should be included in the code and libql{mpi,fort}.so should be linked to the executable file.

docs/archtecture.rst

@@ -0,0 +1,65 @@
+Architectural Overview
+======================
+
+At the heart of the stack is a low-level software infrastructure called
+Interface for Heterogeneous Kernels (IHK). IHK is a general framework
+that provides capabilities for partitioning resources in a many-core
+environment (e.g.,CPU cores and physical memory) and it enables
+management of lightweight kernels. IHK can allocate and release host
+resources dynamically and no reboot of the host machine is required when
+altering configuration. IHK also provides a low-level inter-kernel
+messaging infrastructure, called the Inter-Kernel Communication (IKC)
+layer. An architectural overview of the main system components is shown
+below.
+
+.. figure:: mckernel.png
+   :alt: arch
+
+McKernel is a lightweight kernel written from scratch. It is designed
+for HPC and is booted from IHK. McKernel retains a binary compatible ABI
+with Linux, however, it implements only a small set of performance
+sensitive system calls and the rest are offloaded to Linux.
+Specifically, McKernel has its own memory management, it supports
+processes and multi-threading with a simple round-robin cooperative
+(tick-less) scheduler, and it implements signaling. It also allows
+inter-process memory mappings and it provides interfaces to hardware
+performance counters.
+
+Functionality
+-------------
+
+An overview of some of the principal functionalities of the IHK/McKernel
+stack is provided below.
+
+System Call Offloading
+~~~~~~~~~~~~~~~~~~~~~~
+
+System call forwarding in McKernel is implemented as follows. When an
+offloaded system call occurs, McKernel marshals the system call number
+along with its arguments and sends a message to Linux via a dedicated
+IKC channel. The corresponding proxy process running on Linux is by
+default waiting for system call requests through an ioctl() call into
+IHK’s system call delegator kernel module. The delegator kernel module’s
+IKC interrupt handler wakes up the proxy process, which returns to
+userspace and simply invokes the requested system call. Once it obtains
+the return value, it instructs the delegator module to send the result
+back to McKernel, which subsequently passes the value to user-space.
+
+Unified Address Space
+~~~~~~~~~~~~~~~~~~~~~
+
+The unified address space model in IHK/McKernel ensures that offloaded
+system calls can seamlessly resolve arguments even in case of pointers.
+This mechanism is depicted below and is implemented as follows.
+
+.. figure:: unified_address_space_en.png
+   :alt: unified_ap
+
+First, the proxy process is compiled as a position independent binary,
+which enables us to map the code and data segments specific to the proxy
+process to an address range which is explicitly excluded from McKernel’s
+user space. The grey box on the right side of the figure demonstrates
+the excluded region. Second, the entire valid virtual address range of
+McKernel’s application user-space is covered by a special mapping in the
+proxy process for which we use a pseudo file mapping in Linux. This
+mapping is indicated by the blue box on the left side of the figure.

docs/background.rst

@@ -0,0 +1,47 @@
+Background and Motivation
+=========================
+
+With the growing complexity of high-end supercomputers, the current
+system software stack faces significant challenges as we move forward to
+exascale and beyond. The necessity to deal with extreme degree of
+parallelism, heterogeneous architectures, multiple levels of memory
+hierarchy, power constraints, etc., advocates operating systems that can
+rapidly adapt to new hardware requirements, and that can support novel
+programming paradigms and runtime systems. On the other hand, a new
+class of more dynamic and complex applications are also on the horizon,
+with an increasing demand for application constructs such as in-situ
+analysis, workflows, elaborate monitoring and performance tools. This
+complexity relies not only on the rich features of POSIX, but also on
+the Linux APIs (such as the */proc*, */sys* filesystems, etc.) in
+particular.
+
+Two Traditional HPC OS Approaches
+---------------------------------
+
+Traditionally, light-weight operating systems specialized for HPC
+followed two approaches to tackle scalable execution of large-scale
+applications. In the full weight kernel (FWK) approach, a full Linux
+environment is taken as the basis, and features that inhibit attaining
+HPC scalability are removed, i.e., making it light-weight. The pure
+light-weight kernel (LWK) approach, on the other hand, starts from
+scratch and effort is undertaken to add sufficient functionality so that
+it provides a familiar API, typically something close to that of a
+general purpose OS, while at the same time it retains the desired
+scalability and reliability attributes. Neither of these approaches
+yields a fully Linux compatible environment.
+
+The Multi-kernel Approach
+-------------------------
+
+A hybrid approach recognized recently by the system software community
+is to run Linux simultaneously with a lightweight kernel on compute
+nodes and multiple research projects are now pursuing this direction.
+The basic idea is that simulations run on an HPC tailored lightweight
+kernel, ensuring the necessary isolation for noiseless execution of
+parallel applications, but Linux is leveraged so that the full POSIX API
+is supported. Additionally, the small code base of the LWK can also
+facilitate rapid prototyping for new, exotic hardware features.
+Nevertheless, the questions of how to share node resources between the
+two types of kernels, where do device drivers execute, how exactly do
+the two kernels interact with each other and to what extent are they
+integrated, remain subjects of ongoing debate.

docs/boot_run_shutdown.rst

@@ -0,0 +1,60 @@
+Boot McKernel
+----------------
+
+A boot script called ``mcreboot.sh`` is provided under ``sbin`` in the install
+folder. To boot on logical CPU 1 with 512MB of memory, use the following
+invocation:
+
+::
+
+   export TOP=${HOME}/ihk+mckernel/
+   cd ${TOP}
+   sudo ./sbin/mcreboot.sh -c 1 -m 512m
+
+You should see something similar like this if you display the McKernel’s
+kernel message log:
+
+.. code-block:: none
+
+   ./sbin/ihkosctl 0 kmsg
+
+   IHK/McKernel started.
+   [ -1]: no_execute_available: 1
+   [ -1]: map_fixed: phys: 0xfee00000 => 0xffff860000009000 (1 pages)
+   [ -1]: setup_x86 done.
+   [ -1]: ns_per_tsc: 385
+   [ -1]: KCommand Line: hidos    dump_level=24
+   [ -1]: Physical memory: 0x1ad3000 - 0x21000000, 525520896 bytes, 128301 pages available @ NUMA: 0
+   [ -1]: NUMA: 0, Linux NUMA: 0, type: 1, available bytes: 525520896, pages: 128301
+   [ -1]: NUMA 0 distances: 0 (10),
+   [ -1]: map_fixed: phys: 0x28000 => 0xffff86000000a000 (2 pages)
+   [ -1]: Trampoline area: 0x28000
+   [ -1]: map_fixed: phys: 0x0 => 0xffff86000000c000 (1 pages)
+   [ -1]: # of cpus : 1
+   [ -1]: locals = ffff880001af6000
+   [  0]: BSP: 0 (HW ID: 1 @ NUMA 0)
+   [  0]: BSP: booted 0 AP CPUs
+   [  0]: Master channel init acked.
+   [  0]: vdso is enabled
+   IHK/McKernel booted.
+
+Run a simple program on McKernel
+-----------------------------------
+
+The mcexec command line tool (which is also the Linux proxy process) can
+be used for executing applications on McKernel:
+
+::
+
+   ./bin/mcexec hostname
+   centos-vm
+
+Shutdown McKernel
+--------------------
+
+Finally, to shutdown McKernel and release CPU/memory resources back to
+Linux use the following command:
+
+::
+
+   sudo ./sbin/mcstop+release.sh

docs/conf.py

@@ -0,0 +1,173 @@
+# -*- coding: utf-8 -*-
+#
+# Configuration file for the Sphinx documentation builder.
+#
+# This file does only contain a selection of the most common options. For a
+# full list see the documentation:
+# http://www.sphinx-doc.org/en/master/config
+
+# -- Path setup --------------------------------------------------------------
+
+# If extensions (or modules to document with autodoc) are in another directory,
+# add these directories to sys.path here. If the directory is relative to the
+# documentation root, use os.path.abspath to make it absolute, like shown here.
+#
+# import os
+# import sys
+# sys.path.insert(0, os.path.abspath('.'))
+
+
+# -- Project information -----------------------------------------------------
+
+project = u'IHK/McKernel'
+copyright = u'2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa'
+author = u'Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa'
+
+# The short X.Y version
+version = u''
+# The full version, including alpha/beta/rc tags
+release = u''
+
+
+# -- General configuration ---------------------------------------------------
+
+# If your documentation needs a minimal Sphinx version, state it here.
+#
+# needs_sphinx = '1.0'
+
+# Add any Sphinx extension module names here, as strings. They can be
+# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
+# ones.
+extensions = [
+    'recommonmark',
+]
+numfig = True
+
+# Add any paths that contain templates here, relative to this directory.
+templates_path = ['_templates']
+
+# The suffix(es) of source filenames.
+# You can specify multiple suffix as a list of string:
+source_suffix = ['.rst', '.md']
+
+# The master toctree document.
+master_doc = 'index'
+
+# The language for content autogenerated by Sphinx. Refer to documentation
+# for a list of supported languages.
+#
+# This is also used if you do content translation via gettext catalogs.
+# Usually you set "language" from the command line for these cases.
+language = None
+
+# List of patterns, relative to source directory, that match files and
+# directories to ignore when looking for source files.
+# This pattern also affects html_static_path and html_extra_path.
+exclude_patterns = [u'_build', 'Thumbs.db', '.DS_Store']
+
+# The name of the Pygments (syntax highlighting) style to use.
+pygments_style = None
+
+
+# -- Options for HTML output -------------------------------------------------
+
+# The theme to use for HTML and HTML Help pages.  See the documentation for
+# a list of builtin themes.
+#
+html_theme = 'sphinx_rtd_theme'
+
+# Theme options are theme-specific and customize the look and feel of a theme
+# further.  For a list of options available for each theme, see the
+# documentation.
+#
+# html_theme_options = {}
+
+# Add any paths that contain custom static files (such as style sheets) here,
+# relative to this directory. They are copied after the builtin static files,
+# so a file named "default.css" will overwrite the builtin "default.css".
+html_static_path = ['_static']
+
+# Custom sidebar templates, must be a dictionary that maps document names
+# to template names.
+#
+# The default sidebars (for documents that don't match any pattern) are
+# defined by theme itself.  Builtin themes are using these templates by
+# default: ``['localtoc.html', 'relations.html', 'sourcelink.html',
+# 'searchbox.html']``.
+#
+# html_sidebars = {}
+
+
+# -- Options for HTMLHelp output ---------------------------------------------
+
+# Output file base name for HTML help builder.
+htmlhelp_basename = 'IHKMcKerneldoc'
+
+
+# -- Options for LaTeX output ------------------------------------------------
+
+latex_elements = {
+    # The paper size ('letterpaper' or 'a4paper').
+    #
+    # 'papersize': 'letterpaper',
+
+    # The font size ('10pt', '11pt' or '12pt').
+    #
+    # 'pointsize': '10pt',
+
+    # Additional stuff for the LaTeX preamble.
+    #
+    # 'preamble': '',
+
+    # Latex figure (float) alignment
+    #
+    # 'figure_align': 'htbp',
+}
+
+# Grouping the document tree into LaTeX files. List of tuples
+# (source start file, target name, title,
+#  author, documentclass [howto, manual, or own class]).
+latex_documents = [
+    (master_doc, 'IHKMcKernel.tex', u'IHK/McKernel Documentation',
+     u'Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa', 'manual'),
+]
+
+
+# -- Options for manual page output ------------------------------------------
+
+# One entry per manual page. List of tuples
+# (source start file, name, description, authors, manual section).
+man_pages = [
+    (master_doc, 'ihkmckernel', u'IHK/McKernel Documentation',
+     [author], 1)
+]
+
+
+# -- Options for Texinfo output ----------------------------------------------
+
+# Grouping the document tree into Texinfo files. List of tuples
+# (source start file, target name, title, author,
+#  dir menu entry, description, category)
+texinfo_documents = [
+    (master_doc, 'IHKMcKernel', u'IHK/McKernel Documentation',
+     author, 'IHKMcKernel', 'One line description of project.',
+     'Miscellaneous'),
+]
+
+
+# -- Options for Epub output -------------------------------------------------
+
+# Bibliographic Dublin Core info.
+epub_title = project
+
+# The unique identifier of the text. This can be a ISBN number
+# or the project homepage.
+#
+# epub_identifier = ''
+
+# A unique identification for the text.
+#
+# epub_uid = ''
+
+# A list of files that should not be packed into the epub file.
+epub_exclude_files = ['search.html']

docs/contact.rst

@@ -0,0 +1,11 @@
+Contact
+=======
+
+Please give your feedback to us via one of the following mailing lists.
+Subscription via
+`www.pccluster.org <http://www.pccluster.org/mailman/listinfo/mckernel-users>`__
+is needed.
+
+-  English: mckernel-users@pccluster.org
+-  Japanese: mckernel-users-jp@pccluster.org
+

docs/cover.rst

@@ -0,0 +1,5 @@
+.. include:: logo.rst
+
+.. include:: summary.rst
+
+See `Quick Guide -- Installation <quick.html#installation>`__ for jump start.

docs/doc.rst

@@ -0,0 +1,5 @@
+Documentation
+=============
+
+Documentation is available
+`here <https://ihkmckernel.readthedocs.io>`__.

docs/ihk_developers.rst

@@ -0,0 +1,18 @@
+.. sectnum::
+   :suffix: .
+   :depth: 3
+
+External Specs
+==============
+
+Overview
+--------
+
+Function Specs
+--------------
+
+Command / Daemon Specs
+----------------------
+
+Booting LWK
+===========

docs/index.rst

@@ -0,0 +1,58 @@
+.. IHK/McKernel documentation master file, created by
+   sphinx-quickstart on Mon Jul 27 15:57:24 2020.
+   You can adapt this file completely to your liking, but it should at least
+   contain the root `toctree` directive.
+
+===================
+IHK/McKernel
+===================
+
+.. include:: cover.rst
+
+.. toctree::
+   :maxdepth: 2
+   :caption: Quick Guide
+
+   quick
+
+.. toctree::
+   :maxdepth: 2
+   :caption: Users' Guide
+
+   users
+
+.. toctree::
+   :maxdepth: 2
+   :caption: Operators' Guide
+
+   operators
+
+.. toctree::
+   :maxdepth: 2
+   :caption: IHK Developers' Guide
+
+   ihk_developers
+
+.. toctree::
+   :maxdepth: 2
+   :caption: McKernel Developers' Guide
+
+   mckernel_developers
+
+.. toctree::
+   :maxdepth: 2
+   :caption: IHK Specifications
+
+   spec/ihk
+
+.. toctree::
+   :maxdepth: 2
+   :caption: McKernel Specifications
+
+   spec/mckernel
+
+.. toctree::
+   :maxdepth: 2
+   :caption: What's New
+
+   NEWS

docs/install.rst

@@ -0,0 +1,220 @@
+.. highlight:: bash
+
+Installation
+============
+
+The following OS distributions and platforms are recommended:
+
+* OS distribution
+
+  * CentOS 7.3 or later
+  * RHEL 7.3 or later
+
+* Platform
+
+  * Intel Xeon
+  * Intel Xeon Phi
+  * Fujitsu A64FX
+
+Prepare files for building McKernel
+-----------------------------------
+
+Grant read permission to the System.map file of your kernel version on the build machine:
+
+::
+
+   sudo chmod a+r /boot/System.map-`uname -r`
+
+Install the following packages to the build machine:
+
+::
+
+   cmake kernel-devel binutils-devel systemd-devel numactl-devel gcc make nasm git libdwarf-devel
+
+When having access to repositories
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+On RHEL 8, enable the CodeReady Linux Builder (CLB) repository:
+
+::
+
+   sudo subscription-manager repos --enable codeready-builder-for-rhel-8-$(/bin/arch)-rpms
+
+On CentOS 8, enable the PowerTools repository:
+
+::
+
+   sudo dnf config-manager --set-enabled PowerTools
+
+Install with yum:
+
+::
+
+   sudo yum install cmake kernel-devel binutils-devel systemd-devel numactl-devel gcc make nasm git libdwarf-devel
+
+When not having access to repositories
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Ask the system administrator to install them. Note that ``libdwarf-devel`` is in the CodeReady Linux Builder repository on RHEL 8 or in the PowerTools repository on CentOS 8.
+
+Clone, compile, install
+--------------------------
+
+Clone the source code:
+
+::
+
+   mkdir -p ~/src/ihk+mckernel/
+   cd ~/src/ihk+mckernel/
+   git clone --recursive -b development https://github.com/RIKEN-SysSoft/mckernel.git
+
+(Optional) Checkout to the specific branch or version:
+
+::
+
+   cd mckernel
+   git checkout <pathspec>
+   git submodule update
+
+Foe example, if you want to try the development branch, use
+“development” as the pathspec. If you want to try the prerelease version
+1.7.0-0.2, use “1.7.0-0.2”.
+
+Move to build directory:
+
+::
+
+   mkdir -p ~/src/ihk+mckernel/build && cd ~/src/ihk+mckernel/build
+
+Run cmake:
+
+When not cross-compiling:
+~~~~~~~~~~~~~~~~~~~~~~~~~
+
+::
+
+   cmake -DCMAKE_INSTALL_PREFIX=${HOME}/ihk+mckernel ../mckernel
+
+When cross-compiling:
+~~~~~~~~~~~~~~~~~~~~~
+
+::
+
+   cmake -DCMAKE_INSTALL_PREFIX=${HOME}/ihk+mckernel \
+     -DUNAME_R=<target_uname_r> \
+     -DKERNEL_DIR=<kernnel_dir> \
+     -DBUILD_TARGET=smp-arm64 \
+     -DCMAKE_TOOLCHAIN_FILE=../mckernel/cmake/cross-aarch64.cmake \
+     ../mckernel
+
+Install with cmake
+~~~~~~~~~~~~~~~~~~~~~~
+
+Install with make:
+
+::
+
+   make -j install
+
+The kernel modules and McKernel kernel image should be installed
+under the **ihk+mckernel** folder in your home directory.
+
+Install with rpm
+~~~~~~~~~~~~~~~~~~~~
+
+Create the tarball and the spec file:
+
+::
+
+   make dist
+   cp mckernel-<version>.tar.gz <rpmbuild>/SOURCES
+
+(optional) Edit the following line in ``scripts/mckernel.spec`` to change
+cmake options. For example:
+
+::
+
+   %cmake -DCMAKE_BUILD_TYPE=Release \
+	-DUNAME_R=%{kernel_version} \
+	-DKERNEL_DIR=%{kernel_dir} \
+	%{?cmake_libdir:-DCMAKE_INSTALL_LIBDIR=%{cmake_libdir}} \
+	%{?build_target:-DBUILD_TARGET=%{build_target}} \
+	%{?toolchain_file:-DCMAKE_TOOLCHAIN_FILE=%{toolchain_file}} \
+	-DENABLE_TOFU=ON -DENABLE_FUGAKU_HACKS=ON \
+	-DENABLE_KRM_WORKAROUND=OFF -DWITH_KRM=ON \
+	-DENABLE_FUGAKU_DEBUG=OFF \
+	.
+
+Create the rpm package:
+
+When not cross-compiling:
+"""""""""""""""""""""""""
+
+Then build the rpm:
+
+::
+
+   rpmbuild -ba scripts/mckernel.spec
+
+When cross-compiling:
+"""""""""""""""""""""
+
+::
+
+   rpmbuild -ba scripts/mckernel.spec --target <target_uname_m> -D 'kernel_version <target_uname_r>' -D 'kernel_dir <kernel_source>'
+
+Install the rpm package:
+
+::
+
+   sudo rpm -ivh <rpmbuild>/RPMS/<arch>/mckernel-<version>-<release>_<linux_kernel_ver>_<dist>.<arch>.rpm
+
+The kernel modules and McKernel kernel image are installed under the
+standard system directories.
+
+Prepare files and change settings for installing McKernel
+---------------------------------------------------------
+
+Disable SELinux of the compute nodes:
+
+::
+
+   sudo vim /etc/selinux/config
+
+Change the file to SELINUX=disabled. And then reboot the compute nodes:
+
+::
+
+   sudo reboot
+
+Install the following packages to the compute nodes:
+
+::
+
+   systemd-libs numactl-libs libdwarf
+
+When having access to repositories
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+On RHEL 8, enable the CodeReady Linux Builder (CLB) repository:
+
+::
+
+   sudo subscription-manager repos --enable codeready-builder-for-rhel-8-$(/bin/arch)-rpms
+
+On CentOS 8, enable the PowerTools repository:
+
+::
+
+   sudo dnf config-manager --set-enabled PowerTools
+
+Install with yum:
+
+::
+
+   sudo yum install systemd-libs numactl-libs libdwarf
+
+When not having access to repositories
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Ask the system administrator to install them. Note that ``libdwarf`` is in the CodeReady Linux Builder repository on RHEL 8 or in the PowerTools repository on CentOS 8.

docs/license.rst

@@ -0,0 +1,5 @@
+License
+=======
+
+McKernel is GPL licensed, as found in the LICENSE file.
+

docs/locales/ja/LC_MESSAGES/README_.po

@@ -0,0 +1,469 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../README_.rst:2
+msgid "|McKernel Logo|"
+msgstr ""
+
+#: ../../README_.rst:4
+msgid ""
+"IHK/McKernel is a light-weight multi-kernel operating system designed for"
+" high-end supercomputing. It runs Linux and McKernel, a light-weight "
+"kernel (LWK), side-by-side inside compute nodes and aims at the "
+"following:"
+msgstr ""
+
+#: ../../README_.rst:9
+msgid ""
+"Provide scalable and consistent execution of large-scale parallel "
+"scientific applications, but at the same time maintain the ability to "
+"rapidly adapt to new hardware features and emerging programming models"
+msgstr ""
+
+#: ../../README_.rst:13
+msgid ""
+"Provide efficient memory and device management so that resource "
+"contention and data movement are minimized at the system level"
+msgstr ""
+
+#: ../../README_.rst:15
+msgid ""
+"Eliminate OS noise by isolating OS services in Linux and provide jitter "
+"free execution on the LWK"
+msgstr ""
+
+#: ../../README_.rst:17
+msgid ""
+"Support the full POSIX/Linux APIs by selectively offloading (slow-path) "
+"system calls to Linux"
+msgstr ""
+
+#: ../../README_.rst:21
+msgid "Documentation"
+msgstr ""
+
+#: ../../README_.rst:23
+msgid "Documentation is available `here <https://ihkmckernel.readthedocs.io>`__."
+msgstr ""
+
+#: ../../README_.rst:27
+msgid "Contents"
+msgstr ""
+
+#: ../../README_.rst:29
+msgid "`Background <#background-and-motivation>`__"
+msgstr ""
+
+#: ../../README_.rst:30
+msgid "`Architectural Overview <#architectural-overview>`__"
+msgstr ""
+
+#: ../../README_.rst:31
+msgid "`Installation <#installation>`__"
+msgstr ""
+
+#: ../../README_.rst:32
+msgid "`The Team <#the-team>`__"
+msgstr ""
+
+#: ../../README_.rst:35
+msgid "Background and Motivation"
+msgstr ""
+
+#: ../../README_.rst:37
+msgid ""
+"With the growing complexity of high-end supercomputers, the current "
+"system software stack faces significant challenges as we move forward to "
+"exascale and beyond. The necessity to deal with extreme degree of "
+"parallelism, heterogeneous architectures, multiple levels of memory "
+"hierarchy, power constraints, etc., advocates operating systems that can "
+"rapidly adapt to new hardware requirements, and that can support novel "
+"programming paradigms and runtime systems. On the other hand, a new class"
+" of more dynamic and complex applications are also on the horizon, with "
+"an increasing demand for application constructs such as in-situ analysis,"
+" workflows, elaborate monitoring and performance tools. This complexity "
+"relies not only on the rich features of POSIX, but also on the Linux APIs"
+" (such as the */proc*, */sys* filesystems, etc.) in particular."
+msgstr ""
+
+#: ../../README_.rst:52
+msgid "Two Traditional HPC OS Approaches"
+msgstr ""
+
+#: ../../README_.rst:54
+msgid ""
+"Traditionally, light-weight operating systems specialized for HPC "
+"followed two approaches to tackle scalable execution of large-scale "
+"applications. In the full weight kernel (FWK) approach, a full Linux "
+"environment is taken as the basis, and features that inhibit attaining "
+"HPC scalability are removed, i.e., making it light-weight. The pure "
+"light-weight kernel (LWK) approach, on the other hand, starts from "
+"scratch and effort is undertaken to add sufficient functionality so that "
+"it provides a familiar API, typically something close to that of a "
+"general purpose OS, while at the same time it retains the desired "
+"scalability and reliability attributes. Neither of these approaches "
+"yields a fully Linux compatible environment."
+msgstr ""
+
+#: ../../README_.rst:67
+msgid "The Multi-kernel Approach"
+msgstr ""
+
+#: ../../README_.rst:69
+msgid ""
+"A hybrid approach recognized recently by the system software community is"
+" to run Linux simultaneously with a lightweight kernel on compute nodes "
+"and multiple research projects are now pursuing this direction. The basic"
+" idea is that simulations run on an HPC tailored lightweight kernel, "
+"ensuring the necessary isolation for noiseless execution of parallel "
+"applications, but Linux is leveraged so that the full POSIX API is "
+"supported. Additionally, the small code base of the LWK can also "
+"facilitate rapid prototyping for new, exotic hardware features. "
+"Nevertheless, the questions of how to share node resources between the "
+"two types of kernels, where do device drivers execute, how exactly do the"
+" two kernels interact with each other and to what extent are they "
+"integrated, remain subjects of ongoing debate."
+msgstr ""
+
+#: ../../README_.rst:83
+msgid "Architectural Overview"
+msgstr ""
+
+#: ../../README_.rst:85
+msgid ""
+"At the heart of the stack is a low-level software infrastructure called "
+"Interface for Heterogeneous Kernels (IHK). IHK is a general framework "
+"that provides capabilities for partitioning resources in a many-core "
+"environment (e.g.,CPU cores and physical memory) and it enables "
+"management of lightweight kernels. IHK can allocate and release host "
+"resources dynamically and no reboot of the host machine is required when "
+"altering configuration. IHK also provides a low-level inter-kernel "
+"messaging infrastructure, called the Inter-Kernel Communication (IKC) "
+"layer. An architectural overview of the main system components is shown "
+"below."
+msgstr ""
+
+#: ../../README_.rst:99
+msgid "arch"
+msgstr ""
+
+#: ../../README_.rst:101
+msgid ""
+"McKernel is a lightweight kernel written from scratch. It is designed for"
+" HPC and is booted from IHK. McKernel retains a binary compatible ABI "
+"with Linux, however, it implements only a small set of performance "
+"sensitive system calls and the rest are offloaded to Linux. Specifically,"
+" McKernel has its own memory management, it supports processes and multi-"
+"threading with a simple round-robin cooperative (tick-less) scheduler, "
+"and it implements signaling. It also allows inter-process memory mappings"
+" and it provides interfaces to hardware performance counters."
+msgstr ""
+
+#: ../../README_.rst:112
+msgid "Functionality"
+msgstr ""
+
+#: ../../README_.rst:114
+msgid ""
+"An overview of some of the principal functionalities of the IHK/McKernel "
+"stack is provided below."
+msgstr ""
+
+#: ../../README_.rst:118
+msgid "System Call Offloading"
+msgstr ""
+
+#: ../../README_.rst:120
+msgid ""
+"System call forwarding in McKernel is implemented as follows. When an "
+"offloaded system call occurs, McKernel marshals the system call number "
+"along with its arguments and sends a message to Linux via a dedicated IKC"
+" channel. The corresponding proxy process running on Linux is by default "
+"waiting for system call requests through an ioctl() call into IHK’s "
+"system call delegator kernel module. The delegator kernel module’s IKC "
+"interrupt handler wakes up the proxy process, which returns to userspace "
+"and simply invokes the requested system call. Once it obtains the return "
+"value, it instructs the delegator module to send the result back to "
+"McKernel, which subsequently passes the value to user-space."
+msgstr ""
+
+#: ../../README_.rst:132
+msgid "Unified Address Space"
+msgstr ""
+
+#: ../../README_.rst:134
+msgid ""
+"The unified address space model in IHK/McKernel ensures that offloaded "
+"system calls can seamlessly resolve arguments even in case of pointers. "
+"This mechanism is depicted below and is implemented as follows."
+msgstr ""
+
+#: ../../README_.rst:141
+msgid "unified_ap"
+msgstr ""
+
+#: ../../README_.rst:143
+msgid ""
+"First, the proxy process is compiled as a position independent binary, "
+"which enables us to map the code and data segments specific to the proxy "
+"process to an address range which is explicitly excluded from McKernel’s "
+"user space. The grey box on the right side of the figure demonstrates the"
+" excluded region. Second, the entire valid virtual address range of "
+"McKernel’s application user-space is covered by a special mapping in the "
+"proxy process for which we use a pseudo file mapping in Linux. This "
+"mapping is indicated by the blue box on the left side of the figure."
+msgstr ""
+
+#: ../../README_.rst:153
+msgid "Installation"
+msgstr ""
+
+#: ../../README_.rst:155
+msgid ""
+"For a smooth experience, we recommend the following combination of OS "
+"distributions and platforms:"
+msgstr ""
+
+#: ../../README_.rst:158
+msgid "CentOS 7.3+ running on Intel Xeon, Xeon Phi, Fujitsu A64FX"
+msgstr ""
+
+#: ../../README_.rst:161
+msgid "1. Change SELinux settings"
+msgstr ""
+
+#: ../../README_.rst:163
+msgid "Log in as the root and disable SELinux:"
+msgstr ""
+
+#: ../../README_.rst:169
+msgid "Change the file to SELINUX=disabled"
+msgstr ""
+
+#: ../../README_.rst:172
+msgid "2. Reboot the host machine"
+msgstr ""
+
+#: ../../README_.rst:179
+msgid "3. Prepare packages, kernel symbol table file"
+msgstr ""
+
+#: ../../README_.rst:181
+msgid "You will need the following packages installed:"
+msgstr ""
+
+#: ../../README_.rst:187
+msgid ""
+"Note that to install libdwarf-devel to RHEL-8.2, you need to enable the "
+"CodeReady Linux Builder (CLB) repository and the EPEL repository with the"
+" following commands:"
+msgstr ""
+
+#: ../../README_.rst:195
+msgid "Grant read permission to the System.map file of your kernel version:"
+msgstr ""
+
+#: ../../README_.rst:202
+msgid "4. Obtain sources and compile the kernel"
+msgstr ""
+
+#: ../../README_.rst:204
+msgid "Clone the source code:"
+msgstr ""
+
+#: ../../README_.rst:212
+msgid "(Optional) Checkout to the specific branch or version:"
+msgstr ""
+
+#: ../../README_.rst:220
+msgid ""
+"Foe example, if you want to try the development branch, use “development”"
+" as the pathspec. If you want to try the prerelease version 1.7.0-0.2, "
+"use “1.7.0-0.2”."
+msgstr ""
+
+#: ../../README_.rst:225
+msgid "4.1 Install with cmake"
+msgstr ""
+
+#: ../../README_.rst:227
+msgid "Configure and compile:"
+msgstr ""
+
+#: ../../README_.rst:235
+msgid ""
+"The IHK kernel modules and McKernel kernel image should be installed "
+"under the **ihk+mckernel** folder in your home directory."
+msgstr ""
+
+#: ../../README_.rst:239
+msgid "4.2 Install with rpm"
+msgstr ""
+
+#: ../../README_.rst:241
+msgid "Build rpm:"
+msgstr ""
+
+#: ../../README_.rst:252
+msgid ""
+"The IHK kernel modules and McKernel kernel image are installed under the "
+"system directory."
+msgstr ""
+
+#: ../../README_.rst:256
+msgid "5. Boot McKernel"
+msgstr ""
+
+#: ../../README_.rst:258
+msgid ""
+"A boot script called mcreboot.sh is provided under sbin in the install "
+"folder. To boot on logical CPU 1 with 512MB of memory, use the following "
+"invocation:"
+msgstr ""
+
+#: ../../README_.rst:268
+msgid ""
+"You should see something similar like this if you display the McKernel’s "
+"kernel message log:"
+msgstr ""
+
+#: ../../README_.rst:296
+msgid "6. Run a simple program on McKernel"
+msgstr ""
+
+#: ../../README_.rst:298
+msgid ""
+"The mcexec command line tool (which is also the Linux proxy process) can "
+"be used for executing applications on McKernel:"
+msgstr ""
+
+#: ../../README_.rst:307
+msgid "7. Shutdown McKernel"
+msgstr ""
+
+#: ../../README_.rst:309
+msgid ""
+"Finally, to shutdown McKernel and release CPU/memory resources back to "
+"Linux use the following command:"
+msgstr ""
+
+#: ../../README_.rst:317
+msgid "8. Advanced: Enable Utility Thread offloading Interface (UTI)"
+msgstr ""
+
+#: ../../README_.rst:319
+msgid ""
+"UTI enables a runtime such as MPI runtime to spawn utility threads such "
+"as MPI asynchronous progress threads to Linux cores."
+msgstr ""
+
+#: ../../README_.rst:323
+msgid "8.1 Install capstone"
+msgstr ""
+
+#: ../../README_.rst:325
+msgid "Install EPEL capstone-devel:"
+msgstr ""
+
+#: ../../README_.rst:333
+msgid "8.2 Install syscall_intercept"
+msgstr ""
+
+#: ../../README_.rst:342
+msgid "8.3 Install UTI for McKernel"
+msgstr ""
+
+#: ../../README_.rst:344
+msgid "Install:"
+msgstr ""
+
+#: ../../README_.rst:354
+msgid "8.4 Install McKernel"
+msgstr ""
+
+#: ../../README_.rst:361
+msgid "8.5 Run executable"
+msgstr ""
+
+#: ../../README_.rst:368
+msgid "8.6 Install UTI for Linux for performance comparison"
+msgstr ""
+
+#: ../../README_.rst:370
+msgid "Install by make:"
+msgstr ""
+
+#: ../../README_.rst:379
+msgid "Install by rpm:"
+msgstr ""
+
+#: ../../README_.rst:391
+msgid "The Team"
+msgstr ""
+
+#: ../../README_.rst:393
+msgid ""
+"The McKernel project was started at The University of Tokyo and currently"
+" it is mainly developed at RIKEN. Some of our collaborators include:"
+msgstr ""
+
+#: ../../README_.rst:397
+msgid "Hitachi"
+msgstr ""
+
+#: ../../README_.rst:398
+msgid "Fujitsu"
+msgstr ""
+
+#: ../../README_.rst:399
+msgid "CEA (France)"
+msgstr ""
+
+#: ../../README_.rst:400
+msgid "NEC"
+msgstr ""
+
+#: ../../README_.rst:403
+msgid "License"
+msgstr ""
+
+#: ../../README_.rst:405
+msgid "McKernel is GPL licensed, as found in the LICENSE file."
+msgstr ""
+
+#: ../../README_.rst:408
+msgid "Contact"
+msgstr ""
+
+#: ../../README_.rst:410
+msgid ""
+"Please give your feedback to us via one of the following mailing lists. "
+"Subscription via `www.pccluster.org "
+"<http://www.pccluster.org/mailman/listinfo/mckernel-users>`__ is needed."
+msgstr ""
+
+#: ../../README_.rst:415
+msgid "English: mckernel-users@pccluster.org"
+msgstr ""
+
+#: ../../README_.rst:416
+msgid "Japanese: mckernel-users-jp@pccluster.org"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/archtecture.po

@@ -0,0 +1,101 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../archtecture.rst:2
+msgid "Architectural Overview"
+msgstr ""
+
+#: ../../archtecture.rst:4
+msgid ""
+"At the heart of the stack is a low-level software infrastructure called "
+"Interface for Heterogeneous Kernels (IHK). IHK is a general framework "
+"that provides capabilities for partitioning resources in a many-core "
+"environment (e.g.,CPU cores and physical memory) and it enables "
+"management of lightweight kernels. IHK can allocate and release host "
+"resources dynamically and no reboot of the host machine is required when "
+"altering configuration. IHK also provides a low-level inter-kernel "
+"messaging infrastructure, called the Inter-Kernel Communication (IKC) "
+"layer. An architectural overview of the main system components is shown "
+"below."
+msgstr ""
+
+#: ../../archtecture.rst:18
+msgid ""
+"McKernel is a lightweight kernel written from scratch. It is designed for"
+" HPC and is booted from IHK. McKernel retains a binary compatible ABI "
+"with Linux, however, it implements only a small set of performance "
+"sensitive system calls and the rest are offloaded to Linux. Specifically,"
+" McKernel has its own memory management, it supports processes and multi-"
+"threading with a simple round-robin cooperative (tick-less) scheduler, "
+"and it implements signaling. It also allows inter-process memory mappings"
+" and it provides interfaces to hardware performance counters."
+msgstr ""
+
+#: ../../archtecture.rst:29
+msgid "Functionality"
+msgstr ""
+
+#: ../../archtecture.rst:31
+msgid ""
+"An overview of some of the principal functionalities of the IHK/McKernel "
+"stack is provided below."
+msgstr ""
+
+#: ../../archtecture.rst:35
+msgid "System Call Offloading"
+msgstr ""
+
+#: ../../archtecture.rst:37
+msgid ""
+"System call forwarding in McKernel is implemented as follows. When an "
+"offloaded system call occurs, McKernel marshals the system call number "
+"along with its arguments and sends a message to Linux via a dedicated IKC"
+" channel. The corresponding proxy process running on Linux is by default "
+"waiting for system call requests through an ioctl() call into IHK’s "
+"system call delegator kernel module. The delegator kernel module’s IKC "
+"interrupt handler wakes up the proxy process, which returns to userspace "
+"and simply invokes the requested system call. Once it obtains the return "
+"value, it instructs the delegator module to send the result back to "
+"McKernel, which subsequently passes the value to user-space."
+msgstr ""
+
+#: ../../archtecture.rst:49
+msgid "Unified Address Space"
+msgstr ""
+
+#: ../../archtecture.rst:51
+msgid ""
+"The unified address space model in IHK/McKernel ensures that offloaded "
+"system calls can seamlessly resolve arguments even in case of pointers. "
+"This mechanism is depicted below and is implemented as follows."
+msgstr ""
+
+#: ../../archtecture.rst:58
+msgid ""
+"First, the proxy process is compiled as a position independent binary, "
+"which enables us to map the code and data segments specific to the proxy "
+"process to an address range which is explicitly excluded from McKernel’s "
+"user space. The grey box on the right side of the figure demonstrates the"
+" excluded region. Second, the entire valid virtual address range of "
+"McKernel’s application user-space is covered by a special mapping in the "
+"proxy process for which we use a pseudo file mapping in Linux. This "
+"mapping is indicated by the blue box on the left side of the figure."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/background.po

@@ -0,0 +1,79 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../background.rst:2
+msgid "Background and Motivation"
+msgstr ""
+
+#: ../../background.rst:4
+msgid ""
+"With the growing complexity of high-end supercomputers, the current "
+"system software stack faces significant challenges as we move forward to "
+"exascale and beyond. The necessity to deal with extreme degree of "
+"parallelism, heterogeneous architectures, multiple levels of memory "
+"hierarchy, power constraints, etc., advocates operating systems that can "
+"rapidly adapt to new hardware requirements, and that can support novel "
+"programming paradigms and runtime systems. On the other hand, a new class"
+" of more dynamic and complex applications are also on the horizon, with "
+"an increasing demand for application constructs such as in-situ analysis,"
+" workflows, elaborate monitoring and performance tools. This complexity "
+"relies not only on the rich features of POSIX, but also on the Linux APIs"
+" (such as the */proc*, */sys* filesystems, etc.) in particular."
+msgstr ""
+
+#: ../../background.rst:19
+msgid "Two Traditional HPC OS Approaches"
+msgstr ""
+
+#: ../../background.rst:21
+msgid ""
+"Traditionally, light-weight operating systems specialized for HPC "
+"followed two approaches to tackle scalable execution of large-scale "
+"applications. In the full weight kernel (FWK) approach, a full Linux "
+"environment is taken as the basis, and features that inhibit attaining "
+"HPC scalability are removed, i.e., making it light-weight. The pure "
+"light-weight kernel (LWK) approach, on the other hand, starts from "
+"scratch and effort is undertaken to add sufficient functionality so that "
+"it provides a familiar API, typically something close to that of a "
+"general purpose OS, while at the same time it retains the desired "
+"scalability and reliability attributes. Neither of these approaches "
+"yields a fully Linux compatible environment."
+msgstr ""
+
+#: ../../background.rst:34
+msgid "The Multi-kernel Approach"
+msgstr ""
+
+#: ../../background.rst:36
+msgid ""
+"A hybrid approach recognized recently by the system software community is"
+" to run Linux simultaneously with a lightweight kernel on compute nodes "
+"and multiple research projects are now pursuing this direction. The basic"
+" idea is that simulations run on an HPC tailored lightweight kernel, "
+"ensuring the necessary isolation for noiseless execution of parallel "
+"applications, but Linux is leveraged so that the full POSIX API is "
+"supported. Additionally, the small code base of the LWK can also "
+"facilitate rapid prototyping for new, exotic hardware features. "
+"Nevertheless, the questions of how to share node resources between the "
+"two types of kernels, where do device drivers execute, how exactly do the"
+" two kernels interact with each other and to what extent are they "
+"integrated, remain subjects of ongoing debate."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/boot_run_shutdown.po

@@ -0,0 +1,57 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../boot_run_shutdown.rst:2
+msgid "Boot McKernel"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:4
+msgid ""
+"A boot script called ``mcreboot.sh`` is provided under ``sbin`` in the "
+"install folder. To boot on logical CPU 1 with 512MB of memory, use the "
+"following invocation:"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:14
+msgid ""
+"You should see something similar like this if you display the McKernel’s "
+"kernel message log:"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:42
+msgid "Run a simple program on McKernel"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:44
+msgid ""
+"The mcexec command line tool (which is also the Linux proxy process) can "
+"be used for executing applications on McKernel:"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:53
+msgid "Shutdown McKernel"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:55
+msgid ""
+"Finally, to shutdown McKernel and release CPU/memory resources back to "
+"Linux use the following command:"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/contact.po

@@ -0,0 +1,39 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../contact.rst:2
+msgid "Contact"
+msgstr ""
+
+#: ../../contact.rst:4
+msgid ""
+"Please give your feedback to us via one of the following mailing lists. "
+"Subscription via `www.pccluster.org "
+"<http://www.pccluster.org/mailman/listinfo/mckernel-users>`__ is needed."
+msgstr ""
+
+#: ../../contact.rst:9
+msgid "English: mckernel-users@pccluster.org"
+msgstr ""
+
+#: ../../contact.rst:10
+msgid "Japanese: mckernel-users-jp@pccluster.org"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/cover.po

@@ -0,0 +1,59 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../summary.rst:1
+msgid ""
+"IHK/McKernel is a light-weight multi-kernel operating system designed for"
+" high-end supercomputing. It runs Linux and McKernel, a light-weight "
+"kernel (LWK), side-by-side inside compute nodes and aims at the "
+"following:"
+msgstr ""
+
+#: ../../summary.rst:6
+msgid ""
+"Provide scalable and consistent execution of large-scale parallel "
+"scientific applications, but at the same time maintain the ability to "
+"rapidly adapt to new hardware features and emerging programming models"
+msgstr ""
+
+#: ../../summary.rst:10
+msgid ""
+"Provide efficient memory and device management so that resource "
+"contention and data movement are minimized at the system level"
+msgstr ""
+
+#: ../../summary.rst:12
+msgid ""
+"Eliminate OS noise by isolating OS services in Linux and provide jitter "
+"free execution on the LWK"
+msgstr ""
+
+#: ../../summary.rst:14
+msgid ""
+"Support the full POSIX/Linux APIs by selectively offloading (slow-path) "
+"system calls to Linux"
+msgstr ""
+
+#: ../../cover.rst:5
+msgid ""
+"See `Quick Guide -- Installation <quick.html#installation>`__ for jump "
+"start."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/doc.po

@@ -0,0 +1,28 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../doc.rst:2
+msgid "Documentation"
+msgstr ""
+
+#: ../../doc.rst:4
+msgid "Documentation is available `here <https://ihkmckernel.readthedocs.io>`__."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/ihk_developers.po

@@ -0,0 +1,40 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../ihk_developers.rst:6
+msgid "External Specs"
+msgstr ""
+
+#: ../../ihk_developers.rst:9
+msgid "Overview"
+msgstr ""
+
+#: ../../ihk_developers.rst:12
+msgid "Function Specs"
+msgstr ""
+
+#: ../../ihk_developers.rst:15
+msgid "Command / Daemon Specs"
+msgstr ""
+
+#: ../../ihk_developers.rst:18
+msgid "Booting LWK"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/index.po

@@ -0,0 +1,95 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-07 10:00+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../index.rst:12
+msgid "Quick Guide"
+msgstr "クイックガイド"
+
+#: ../../index.rst:18
+msgid "Users' Guide"
+msgstr "ユーザガイド"
+
+#: ../../index.rst:24
+msgid "Operators' Guide"
+msgstr "運用ガイド"
+
+#: ../../index.rst:30
+msgid "IHK Developers' Guide"
+msgstr ""
+
+#: ../../index.rst:36
+msgid "McKernel Developers' Guide"
+msgstr ""
+
+#: ../../index.rst:42
+msgid "IHK Specifications"
+msgstr ""
+
+#: ../../index.rst:48
+msgid "McKernel Specifications"
+msgstr ""
+
+#: ../../index.rst:54
+msgid "What's New"
+msgstr ""
+
+#: ../../index.rst:8
+msgid "IHK/McKernel"
+msgstr ""
+
+#: ../../summary.rst:1
+msgid ""
+"IHK/McKernel is a light-weight multi-kernel operating system designed for"
+" high-end supercomputing. It runs Linux and McKernel, a light-weight "
+"kernel (LWK), side-by-side inside compute nodes and aims at the "
+"following:"
+msgstr ""
+
+#: ../../summary.rst:6
+msgid ""
+"Provide scalable and consistent execution of large-scale parallel "
+"scientific applications, but at the same time maintain the ability to "
+"rapidly adapt to new hardware features and emerging programming models"
+msgstr ""
+
+#: ../../summary.rst:10
+msgid ""
+"Provide efficient memory and device management so that resource "
+"contention and data movement are minimized at the system level"
+msgstr ""
+
+#: ../../summary.rst:12
+msgid ""
+"Eliminate OS noise by isolating OS services in Linux and provide jitter "
+"free execution on the LWK"
+msgstr ""
+
+#: ../../summary.rst:14
+msgid ""
+"Support the full POSIX/Linux APIs by selectively offloading (slow-path) "
+"system calls to Linux"
+msgstr ""
+
+#: ../../cover.rst:5
+msgid ""
+"See `Quick Guide -- Installation <quick.html#installation>`__ for jump "
+"start."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/install.po

@@ -0,0 +1,191 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../install.rst:4
+msgid "Installation"
+msgstr "インストール"
+
+#: ../../install.rst:6
+msgid "The following OS distributions and platforms are recommended:"
+msgstr "推奨OSディストリビューションとプロセッサは以下の通り。"
+
+#: ../../install.rst:8
+msgid "OS distribution"
+msgstr ""
+
+#: ../../install.rst:10
+msgid "CentOS 7.3 or later"
+msgstr ""
+
+#: ../../install.rst:11
+msgid "RHEL 7.3 or later"
+msgstr ""
+
+#: ../../install.rst:13
+msgid "Platform"
+msgstr ""
+
+#: ../../install.rst:15
+msgid "Intel Xeon"
+msgstr ""
+
+#: ../../install.rst:16
+msgid "Intel Xeon Phi"
+msgstr ""
+
+#: ../../install.rst:17
+msgid "Fujitsu A64FX"
+msgstr ""
+
+#: ../../install.rst:20
+msgid "Prepare files for building McKernel"
+msgstr ""
+
+#: ../../install.rst:22
+msgid ""
+"Grant read permission to the System.map file of your kernel version on "
+"the build machine:"
+msgstr ""
+
+#: ../../install.rst:28
+msgid "Install the following packages to the build machine:"
+msgstr ""
+
+#: ../../install.rst:35 ../../install.rst:179
+msgid "When having access to repositories"
+msgstr ""
+
+#: ../../install.rst:37 ../../install.rst:181
+msgid "On RHEL 8, enable the CodeReady Linux Builder (CLB) repository:"
+msgstr ""
+
+#: ../../install.rst:43 ../../install.rst:187
+msgid "On CentOS 8, enable the PowerTools repository:"
+msgstr ""
+
+#: ../../install.rst:49 ../../install.rst:193
+msgid "Install with yum:"
+msgstr ""
+
+#: ../../install.rst:56 ../../install.rst:200
+msgid "When not having access to repositories"
+msgstr ""
+
+#: ../../install.rst:58
+msgid ""
+"Ask the system administrator to install them. Note that ``libdwarf-"
+"devel`` is in the CodeReady Linux Builder repository on RHEL 8 or in the "
+"PowerTools repository on CentOS 8."
+msgstr ""
+
+#: ../../install.rst:61
+msgid "Clone, compile, install"
+msgstr ""
+
+#: ../../install.rst:63
+msgid "Clone the source code:"
+msgstr ""
+
+#: ../../install.rst:71
+msgid "(Optional) Checkout to the specific branch or version:"
+msgstr ""
+
+#: ../../install.rst:79
+msgid ""
+"Foe example, if you want to try the development branch, use “development”"
+" as the pathspec. If you want to try the prerelease version 1.7.0-0.2, "
+"use “1.7.0-0.2”."
+msgstr ""
+
+#: ../../install.rst:83
+msgid "Move to build directory:"
+msgstr ""
+
+#: ../../install.rst:89
+msgid "Run cmake:"
+msgstr ""
+
+#: ../../install.rst:92 ../../install.rst:135
+msgid "When not cross-compiling:"
+msgstr ""
+
+#: ../../install.rst:99 ../../install.rst:142
+msgid "When cross-compiling:"
+msgstr ""
+
+#: ../../install.rst:111
+msgid "Install with cmake"
+msgstr ""
+
+#: ../../install.rst:113
+msgid "Install with make:"
+msgstr ""
+
+#: ../../install.rst:119
+msgid ""
+"The kernel modules and McKernel kernel image should be installed under "
+"the **ihk+mckernel** folder in your home directory."
+msgstr ""
+
+#: ../../install.rst:123
+msgid "Install with rpm"
+msgstr ""
+
+#: ../../install.rst:125
+msgid "Create the tarball and the spec file:"
+msgstr ""
+
+#: ../../install.rst:132
+msgid "Create the rpm package:"
+msgstr ""
+
+#: ../../install.rst:148
+msgid "Install the rpm package:"
+msgstr ""
+
+#: ../../install.rst:154
+msgid ""
+"The kernel modules and McKernel kernel image are installed under the "
+"standard system directories."
+msgstr ""
+
+#: ../../install.rst:158
+msgid "Prepare files and change settings for installing McKernel"
+msgstr ""
+
+#: ../../install.rst:160
+msgid "Disable SELinux of the compute nodes:"
+msgstr ""
+
+#: ../../install.rst:166
+msgid "Change the file to SELINUX=disabled. And then reboot the compute nodes:"
+msgstr ""
+
+#: ../../install.rst:172
+msgid "Install the following packages to the compute nodes:"
+msgstr ""
+
+#: ../../install.rst:202
+msgid ""
+"Ask the system administrator to install them. Note that ``libdwarf`` is "
+"in the CodeReady Linux Builder repository on RHEL 8 or in the PowerTools "
+"repository on CentOS 8."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/license.po

@@ -0,0 +1,28 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../license.rst:2
+msgid "License"
+msgstr ""
+
+#: ../../license.rst:4
+msgid "McKernel is GPL licensed, as found in the LICENSE file."
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/logo.po

@@ -0,0 +1,20 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+

docs/locales/ja/LC_MESSAGES/mckernel_developers.po

@@ -0,0 +1,28 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../mckernel_developers.rst:6
+msgid "Interfaces"
+msgstr ""
+
+#: ../../mckernel_developers.rst:9
+msgid "Interface details"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/operators.po

@@ -0,0 +1,886 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-07 10:26+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../operators.rst:5
+#, fuzzy
+msgid "This document will explain how to operate system with McKernel."
+msgstr "McKernelを用いたシステムを運用するシステム管理者を対象として、運用手順を説明する。"
+
+#: ../../operators.rst:8
+msgid "Installation"
+msgstr "インストール"
+
+#: ../../operators.rst:10
+msgid "See `Quick Guide -- Installation <quick.html#installation>`__."
+msgstr "`クイックガイド ―  インストール <quick.html#installation>`__ に記載する。"
+
+#: ../../uti.rst:2
+msgid "Advanced: Enable Utility Thread offloading Interface (UTI)"
+msgstr ""
+
+#: ../../uti.rst:4
+msgid ""
+"UTI enables a runtime such as MPI runtime to spawn utility threads such "
+"as MPI asynchronous progress threads to Linux cores."
+msgstr ""
+
+#: ../../uti.rst:8
+msgid "Install capstone"
+msgstr ""
+
+#: ../../uti.rst:11 ../../uti.rst:22
+msgid "When compute nodes don't have access to repositories"
+msgstr ""
+
+#: ../../uti.rst:13
+msgid "Install EPEL capstone-devel:"
+msgstr ""
+
+#: ../../uti.rst:24
+msgid ""
+"Ask the system administrator to install ``capstone-devel``. Note that it "
+"is in the EPEL repository."
+msgstr ""
+
+#: ../../uti.rst:28
+msgid "Install syscall_intercept"
+msgstr ""
+
+#: ../../uti.rst:37
+msgid "Install UTI for McKernel"
+msgstr ""
+
+#: ../../uti.rst:39
+msgid "Install:"
+msgstr ""
+
+#: ../../uti.rst:49
+msgid "Install McKernel"
+msgstr ""
+
+#: ../../uti.rst:51
+msgid "Add ``-DENABLE_UTI=ON`` option to ``cmake``:"
+msgstr ""
+
+#: ../../uti.rst:58
+msgid "Run programs"
+msgstr ""
+
+#: ../../uti.rst:60
+msgid "Add ``--enable-uti`` option to ``mcexec``:"
+msgstr ""
+
+#: ../../uti.rst:67
+msgid "Install UTI for Linux"
+msgstr ""
+
+#: ../../uti.rst:69
+msgid ""
+"You should skip this step if it's already installed as with, for example,"
+" Fujitsu Technical Computing Suite."
+msgstr ""
+
+#: ../../uti.rst:72
+msgid "Install by make"
+msgstr ""
+
+#: ../../uti.rst:82
+msgid "Install by rpm"
+msgstr ""
+
+#: ../../operators.rst:15
+msgid "Boot and Shut-down"
+msgstr "起動停止"
+
+#: ../../operators.rst:18
+msgid "Related files"
+msgstr "関連ファイル"
+
+#: ../../operators.rst:20
+msgid ""
+"In the followings, the install directory of IHK/McKernel is shown as "
+"``<install>`` . The kernel modules and their locations are as follows."
+msgstr "以降、IHK/McKernelのインストールディレクトリを<install>とする。SMPプロセッサ向け、x86_64アーキ向けの関連ファイルの場所は以下の通り。"
+
+#: ../../operators.rst:24
+msgid "<install>/kmod/ihk.ko"
+msgstr ""
+
+#: ../../operators.rst:24
+msgid "IHK-master core"
+msgstr ""
+
+#: ../../operators.rst:26
+msgid "|ihk-smp|"
+msgstr ""
+
+#: ../../operators.rst:26
+msgid "IHK-master driver"
+msgstr ""
+
+#: ../../operators.rst:29
+msgid "|mcctrl|"
+msgstr ""
+
+#: ../../operators.rst:29
+msgid "Delegator module"
+msgstr ""
+
+#: ../../operators.rst:32
+msgid "|mckernel.img|"
+msgstr ""
+
+#: ../../operators.rst:32
+msgid "Kernel Image"
+msgstr "カーネルイメージ"
+
+#: ../../operators.rst:40
+msgid "The commands and daemons for operation and their locations are as follows."
+msgstr "運用向けコマンド・デーモンのファイルの場所は以下の通り。"
+
+#: ../../operators.rst:43
+msgid "|mcreboot|"
+msgstr ""
+
+#: ../../operators.rst:43
+msgid "Boot script"
+msgstr "起動スクリプト"
+
+#: ../../operators.rst:46
+msgid "|mcstop|"
+msgstr ""
+
+#: ../../operators.rst:46
+msgid "Shutdown script"
+msgstr "シャットダウンスクリプト"
+
+#: ../../operators.rst:49
+msgid "<install>/bin/mcexec"
+msgstr ""
+
+#: ../../operators.rst:49
+msgid "Process invocation command"
+msgstr "プロセス起動コマンド"
+
+#: ../../operators.rst:51
+msgid "<install>/bin/eclair"
+msgstr ""
+
+#: ../../operators.rst:51
+msgid "Kernel dump analysis tool"
+msgstr "カーネルダンプ解析ツール"
+
+#: ../../operators.rst:53
+msgid "|vmcore2mckdump|"
+msgstr ""
+
+#: ../../operators.rst:53
+msgid "Kernel dump format conversion tool"
+msgstr "カーネルダンプ形式変換ツール"
+
+#: ../../operators.rst:60
+msgid "以下、関連コマンドおよび関連関数のインターフェイスを説明する。"
+msgstr ""
+
+#: ../../operators.rst:63
+msgid "インターフェイス"
+msgstr ""
+
+#: ../../operators.rst:66
+msgid "カーネル引数"
+msgstr ""
+
+#: ../../operators.rst:68
+msgid "McKernelのカーネル引数を表 :numref:`tab-kargs` に示す。"
+msgstr ""
+
+#: ../../operators.rst:72
+msgid "McKernelのカーネル引数"
+msgstr ""
+
+#: ../../operators.rst:75
+msgid "hidos"
+msgstr ""
+
+#: ../../operators.rst:75
+msgid "IKCを有効にする。"
+msgstr ""
+
+#: ../../operators.rst:77
+msgid "|dlv|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "Linuxのpanicハンドラ経由でダンプを行った場合の、ダ"
+msgstr ""
+
+#: ../../operators.rst
+msgid "ンプ対象とするメモリ領域の種類を<level>に設定する。"
+msgstr ""
+
+#: ../../operators.rst
+msgid "設定可能な値は以下の通り。"
+msgstr ""
+
+#: ../../operators.rst:83 ../../operators.rst:164 ../../operators.rst:196
+#: ../../operators.rst:234
+msgid "0"
+msgstr ""
+
+#: ../../operators.rst:84 ../../operators.rst:165
+msgid "IHKがMcKernelに割り当てたメモリ領域を出力する"
+msgstr ""
+
+#: ../../operators.rst:85 ../../operators.rst:166
+msgid "24"
+msgstr ""
+
+#: ../../operators.rst:86 ../../operators.rst:167
+msgid "カーネルが使用しているメモリ領域を出力する"
+msgstr ""
+
+#: ../../operators.rst:88 ../../operators.rst:169
+msgid "指定がなかった場合は24が用いられる。"
+msgstr ""
+
+#: ../../operators.rst:90
+msgid "|allow|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "McKernelに割り当てられたCPU数より大きい数のスレッド"
+msgstr ""
+
+#: ../../operators.rst
+msgid "またはプロセスの生成を許可する。この引数が指定され"
+msgstr ""
+
+#: ../../operators.rst
+msgid "ない場合に、CPU数より大きい数のスレッドまたはプロセ"
+msgstr ""
+
+#: ../../operators.rst
+msgid "スをclone(), fork(), vfork()などで生成しようとする"
+msgstr ""
+
+#: ../../operators.rst
+msgid "と、当該システムコールがEINVALエラーを返す。"
+msgstr ""
+
+#: ../../operators.rst:102
+msgid "ブートスクリプト"
+msgstr "ブートスクリプト"
+
+#: ../../operators.rst:105 ../../operators.rst:207
+msgid "書式"
+msgstr ""
+
+#: ../../operators.rst:113 ../../operators.rst:217
+msgid "オプション"
+msgstr ""
+
+#: ../../operators.rst:116
+msgid "|opt-c|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "McKernelに割り当てるCPUのリストを指定する。フォー"
+msgstr ""
+
+#: ../../operators.rst
+msgid "マットは以下の通り。"
+msgstr ""
+
+#: ../../operators.rst
+msgid "<CPU logical id>,<CPU logical id>...または"
+msgstr ""
+
+#: ../../operators.rst
+msgid "<CPU logical id>-<CPU logical id>,<CPU logical id>"
+msgstr ""
+
+#: ../../operators.rst
+msgid "-<CPU logical id>...または両者の混合。"
+msgstr ""
+
+#: ../../operators.rst:122
+msgid "|opt-r|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "McKernelのCPUがIKCメッセージを送るLinux"
+msgstr ""
+
+#: ../../operators.rst
+msgid "CPUを指定する。フォーマットは以下の通り。"
+msgstr ""
+
+#: ../../operators.rst
+msgid "<CPU list>:<CPU id>+<CPU list>:<CPU id>..."
+msgstr ""
+
+#: ../../operators.rst
+msgid "<CPU list>のフォーマットは-cオプションにおけるもの"
+msgstr ""
+
+#: ../../operators.rst
+msgid "と同じである。"
+msgstr ""
+
+#: ../../operators.rst
+msgid "各<CPU list>:<CPU id>は<CPU list>で示されるMcKernel"
+msgstr ""
+
+#: ../../operators.rst
+msgid "のCPUが<CPU logical id>で示されるLinuxのCPUにIKC"
+msgstr ""
+
+#: ../../operators.rst
+msgid "メッセージを送信することを意味する。"
+msgstr ""
+
+#: ../../operators.rst:131
+msgid "|opt-m|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "McKernelに割り当てるメモリ領域を指定する。フォーマッ"
+msgstr ""
+
+#: ../../operators.rst
+msgid "トは以下の通り。"
+msgstr ""
+
+#: ../../operators.rst
+msgid "<size>@<NUMA-id>, <size>@<NUMA-id>..."
+msgstr ""
+
+#: ../../operators.rst:135
+msgid "|opt-f|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "ihkmondが使用するsyslogプロトコルのfacilityを指定す"
+msgstr ""
+
+#: ../../operators.rst
+msgid "る。デフォルトはLOG_LOCAL6。"
+msgstr ""
+
+#: ../../operators.rst:138
+msgid "|opt-o|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "IHKのデバイスファイル(/dev/mcd*, /dev/mcos*)のオー"
+msgstr ""
+
+#: ../../operators.rst
+msgid "ナーとグループの値を<user>[:<group>]の形式で指定す"
+msgstr ""
+
+#: ../../operators.rst
+msgid "る。デフォルトはmcreboot.shを実行したユーザ。"
+msgstr ""
+
+#: ../../operators.rst:142
+msgid "|opt-i|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "ihkmondがハングアップ検知のためにOS状態を確認する時"
+msgstr ""
+
+#: ../../operators.rst
+msgid "間間隔を秒単位で指定する。-1が指定された場合はハン"
+msgstr ""
+
+#: ../../operators.rst
+msgid "グアップ検知を行わない。指定がない場合はハングアッ"
+msgstr ""
+
+#: ../../operators.rst
+msgid "プ検知を行わない。"
+msgstr ""
+
+#: ../../operators.rst:147
+msgid "|opt-k|"
+msgstr ""
+
+#: ../../operators.rst
+msgid "カーネルメッセージの/dev/logへのリダイレクト有無を"
+msgstr ""
+
+#: ../../operators.rst
+msgid "指定する。0が指定された場合はリダイレクトを行わず、"
+msgstr ""
+
+#: ../../operators.rst
+msgid "0以外が指定された場合はリダイレクトを行う。指定がな"
+msgstr ""
+
+#: ../../operators.rst
+msgid "い場合はリダイレクトを行わない。"
+msgstr ""
+
+#: ../../operators.rst:152
+msgid "-q <irq>"
+msgstr ""
+
+#: ../../operators.rst
+msgid "IHKが使用するIRQ番号を指定する。指定がない場合は"
+msgstr ""
+
+#: ../../operators.rst
+msgid "64-255の範囲で空いているものを使用する。"
+msgstr ""
+
+#: ../../operators.rst:155
+msgid "-t"
+msgstr ""
+
+#: ../../operators.rst
+msgid "（x86_64アーキテクチャのみ）Turbo"
+msgstr ""
+
+#: ../../operators.rst
+msgid "Boostをオンにする。デフォルトはオフ。"
+msgstr ""
+
+#: ../../operators.rst:158
+msgid "-d <level>"
+msgstr ""
+
+#: ../../operators.rst:171
+msgid "-O"
+msgstr ""
+
+#: ../../operators.rst
+msgid "またはプロセスの生成を許可する。指定がない場合は許可"
+msgstr ""
+
+#: ../../operators.rst
+msgid "しない。すなわち、CPU数より大きい数のスレッドまたは"
+msgstr ""
+
+#: ../../operators.rst
+msgid "プロセスを生成しようとするとエラーとなる。"
+msgstr ""
+
+#: ../../operators.rst:187 ../../operators.rst:224
+msgid "説明"
+msgstr ""
+
+#: ../../operators.rst:189
+msgid ""
+"McKernel関連カーネルモジュールをinsmodし、<cpulist>で指定されたCPUと<memlist>で指定されたメモリ領域からなるパーティションを作成し、IKC"
+" mapを<ikcmap>に設定し、前記パーティションにMcKernelをブートする。"
+msgstr ""
+
+#: ../../operators.rst:193 ../../operators.rst:231
+msgid "戻り値"
+msgstr ""
+
+#: ../../operators.rst:196 ../../operators.rst:234
+msgid "正常終了"
+msgstr ""
+
+#: ../../operators.rst:198 ../../operators.rst:236
+msgid "0以外"
+msgstr ""
+
+#: ../../operators.rst:198 ../../operators.rst:236
+msgid "エラー"
+msgstr ""
+
+#: ../../operators.rst:202
+msgid "シャットダウンスクリプト"
+msgstr "シャットダウンスクリプト"
+
+#: ../../operators.rst:219
+msgid "なし"
+msgstr ""
+
+#: ../../operators.rst:226
+msgid "McKernelをシャットダウンし、McKernel用パーティションを削除し、関連カーネルモジュールをrmmodする。"
+msgstr ""
+
+#: ../../operators.rst:240
+msgid "プロセス起動コマンド"
+msgstr ""
+
+#: ../../operators.rst:245
+msgid "ダンプ解析コマンド"
+msgstr ""
+
+#: ../../operators.rst:250
+msgid "ダンプ形式変換コマンド"
+msgstr ""
+
+#: ../../operators.rst:255
+msgid "ブート手順"
+msgstr ""
+
+#: ../../operators.rst:257
+msgid "mcreboot.shを用いてブート手順を説明する。"
+msgstr ""
+
+#: ../../operators.rst:259 ../../operators.rst:906
+msgid "スクリプトは以下の通り。"
+msgstr ""
+
+#: ../../operators.rst:852 ../../operators.rst:1050
+msgid "手順は以下の通り。"
+msgstr ""
+
+#: ../../operators.rst:854
+msgid "ihkmondを起動する。ihkmondは任意のタイミングで起動してよい。これは、ihkmondはOSインスタンスの作成を検知して動作を開始するためである。（83行目）"
+msgstr ""
+
+#: ../../operators.rst:856
+msgid "Linuxのカーネルバージョンが、mcoverlayfsが動作するものであるかを確認する。（200–216行目）"
+msgstr ""
+
+#: ../../operators.rst:858
+msgid "irqbalanceを停止する。（251–257行目）"
+msgstr ""
+
+#: ../../operators.rst:860
+msgid ""
+"/proc/irq/[n]/affinityの設定を保存した上でMcKernel "
+"CPUを担当から外す。担当CPUが無くなる場合は、全てのLinux CPUを指定する。（269–303行目）"
+msgstr ""
+
+#: ../../operators.rst:864
+msgid "ihk.koをinsmodする。（307行目）"
+msgstr ""
+
+#: ../../operators.rst:866
+msgid "Linuxによるメモリフラグメンテーションを緩和するために以下を実施する。（313–320行目）"
+msgstr ""
+
+#: ../../operators.rst:868
+msgid "アクティブでないプロセスを積極的にスワップアウトするように設定する"
+msgstr ""
+
+#: ../../operators.rst:870
+msgid "クリーンなページキャッシュを無効化し、またdentriesやinodeのslabオブジェクトのうち可能なものを破棄する"
+msgstr ""
+
+#: ../../operators.rst:872
+msgid "連続する空き領域を結合してより大きな空き領域にまとめる"
+msgstr ""
+
+#: ../../operators.rst:874
+msgid ""
+"ihk-smp-x86.koをinsmodする。（340行目）ihk-smp-x86.koは関数をihk.koに登録する。このため、ihk-"
+"smp-x86.koはihk.koをinsmodした後にinsmodする必要がある。"
+msgstr ""
+
+#: ../../operators.rst:876
+msgid "メモリを予約する。（370行目）"
+msgstr ""
+
+#: ../../operators.rst:878
+msgid "CPUを予約する。（374行目）"
+msgstr ""
+
+#: ../../operators.rst:880
+msgid "McKernelのカーネルモジュールmcctrl.koをinsmodする。（382行目）mcctrl.koはMcKernelブート時に呼ばれる関数をihk.koに登録する。このため、mcctrl.koのinsmodはihk.koのinsmodの後に、またブートの前に行う必要がある。"
+msgstr ""
+
+#: ../../operators.rst:882
+msgid "OSインスタンスを作成する。（406行目）"
+msgstr ""
+
+#: ../../operators.rst:884
+msgid "OSインスタンスにCPUを割り当てる。（412行目）"
+msgstr ""
+
+#: ../../operators.rst:886
+msgid "McKernel CPUのIKCメッセージ送信先のLinux CPUを設定する。（419行目）"
+msgstr ""
+
+#: ../../operators.rst:888
+msgid "OSインスタンスにメモリを割り当てる。（426行目）"
+msgstr ""
+
+#: ../../operators.rst:890
+msgid "カーネルイメージをロードする。（432行目）"
+msgstr ""
+
+#: ../../operators.rst:892
+msgid "カーネル引数をカーネルに渡す。（438行目）"
+msgstr ""
+
+#: ../../operators.rst:894
+msgid "カーネルをブートする。（444行目）"
+msgstr ""
+
+#: ../../operators.rst:896
+msgid ""
+"/proc, "
+"/sysファイルの準備をする。また、その中でmcoverlayfs.koをinsmodする。mcoverlayfs.koは他モジュールとの依存関係を持たない。（454行目から567行目）なお、関数インターフェイスでの対応関数はihk_os_create_pseudofs()である。"
+msgstr ""
+
+#: ../../operators.rst:898
+msgid "irqbalanceを、Linux CPUのみを対象とする設定で開始する。（569–587行目）"
+msgstr ""
+
+#: ../../operators.rst:902
+msgid "シャットダウン手順"
+msgstr ""
+
+#: ../../operators.rst:904
+msgid "mcstop+release.shを用いてシャットダウン手順を説明する。"
+msgstr ""
+
+#: ../../operators.rst:1052
+msgid "ブート時にLinux CPUのみを対象とする設定で開始されたirqbalanceを停止する。（24–33行目）"
+msgstr ""
+
+#: ../../operators.rst:1055
+msgid "全てのOSインスタンスを破壊する。OSインスタンスに割り当てられていた資源はIHKがLWKのために予約した状態に移行する。（35–50行目）"
+msgstr ""
+
+#: ../../operators.rst:1057
+msgid "IHKがLWKのために予約していた資源を開放する。（52–77行目）"
+msgstr ""
+
+#: ../../operators.rst:1059
+msgid "mcctrl.koをrmmodする。（81行目）"
+msgstr ""
+
+#: ../../operators.rst:1061
+msgid ""
+"/proc, "
+"/sysファイルの準備をする。また、その中でmcoverlayfs.koをrmmodする。（87–100行目）なお、関数インターフェイスでの対応関数はihk_os_destroy_pseudofs()である。"
+msgstr ""
+
+#: ../../operators.rst:1063
+msgid "ihk-smp-x86.koをrmmodする。（104行目）"
+msgstr ""
+
+#: ../../operators.rst:1065
+msgid "ihk.koをrmmodする。（112行目）"
+msgstr ""
+
+#: ../../operators.rst:1067
+msgid "ihkmondを停止する。（121行目）"
+msgstr ""
+
+#: ../../operators.rst:1069
+msgid "/proc/irq/[n]/affinityの設定をブート時に保存しておいたものに戻し、ブート前の設定でirqbalanceを開始する。（124–135行目）"
+msgstr ""
+
+#: ../../operators.rst:1071
+msgid "Linuxカーネルのスワップアウト積極度の設定をデフォルトの値に戻す。（138行目）"
+msgstr ""
+
+#~ msgid "The following OS distributions and platforms are recommended:"
+#~ msgstr ""
+
+#~ msgid "OS distribution"
+#~ msgstr ""
+
+#~ msgid "CentOS 7.3 or later"
+#~ msgstr ""
+
+#~ msgid "RHEL 7.3 or later"
+#~ msgstr ""
+
+#~ msgid "Platform"
+#~ msgstr ""
+
+#~ msgid "Intel Xeon"
+#~ msgstr ""
+
+#~ msgid "Intel Xeon Phi"
+#~ msgstr ""
+
+#~ msgid "Fujitsu A64FX"
+#~ msgstr ""
+
+#~ msgid "Prepare files for building McKernel"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Grant read permission to the System.map"
+#~ " file of your kernel version on "
+#~ "the build machine:"
+#~ msgstr ""
+
+#~ msgid "Install the following packages to the build machine:"
+#~ msgstr ""
+
+#~ msgid "When having access to repositories"
+#~ msgstr ""
+
+#~ msgid "On RHEL 8, enable the CodeReady Linux Builder (CLB) repository:"
+#~ msgstr ""
+
+#~ msgid "On CentOS 8, enable the PowerTools repository:"
+#~ msgstr ""
+
+#~ msgid "Install with yum:"
+#~ msgstr ""
+
+#~ msgid "When not having access to repositories"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Ask the system administrator to install"
+#~ " them. Note that ``libdwarf-devel`` "
+#~ "is in the CodeReady Linux Builder "
+#~ "repository on RHEL 8 or in the "
+#~ "PowerTools repository on CentOS 8."
+#~ msgstr ""
+
+#~ msgid "Clone, compile, install"
+#~ msgstr ""
+
+#~ msgid "Clone the source code:"
+#~ msgstr ""
+
+#~ msgid "(Optional) Checkout to the specific branch or version:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Foe example, if you want to try"
+#~ " the development branch, use “development”"
+#~ " as the pathspec. If you want "
+#~ "to try the prerelease version 1.7.0-0.2,"
+#~ " use “1.7.0-0.2”."
+#~ msgstr ""
+
+#~ msgid "Move to build directory:"
+#~ msgstr ""
+
+#~ msgid "Run cmake:"
+#~ msgstr ""
+
+#~ msgid "When not cross-compiling:"
+#~ msgstr ""
+
+#~ msgid "When cross-compiling:"
+#~ msgstr ""
+
+#~ msgid "Install with cmake"
+#~ msgstr ""
+
+#~ msgid "Install with make:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "The kernel modules and McKernel kernel"
+#~ " image should be installed under the"
+#~ " **ihk+mckernel** folder in your home "
+#~ "directory."
+#~ msgstr ""
+
+#~ msgid "Install with rpm"
+#~ msgstr ""
+
+#~ msgid "Create the tarball and the spec file:"
+#~ msgstr ""
+
+#~ msgid "Create the rpm package:"
+#~ msgstr ""
+
+#~ msgid "Install the rpm package:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "The kernel modules and McKernel kernel"
+#~ " image are installed under the "
+#~ "standard system directories."
+#~ msgstr ""
+
+#~ msgid "Prepare files and change settings for installing McKernel"
+#~ msgstr ""
+
+#~ msgid "Disable SELinux of the compute nodes:"
+#~ msgstr ""
+
+#~ msgid "Change the file to SELINUX=disabled. And then reboot the compute nodes:"
+#~ msgstr ""
+
+#~ msgid "Install the following packages to the compute nodes:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Ask the system administrator to install"
+#~ " them. Note that ``libdwarf`` is in"
+#~ " the CodeReady Linux Builder repository "
+#~ "on RHEL 8 or in the PowerTools "
+#~ "repository on CentOS 8."
+#~ msgstr ""
+
+#~ msgid "起動停止"
+#~ msgstr "起動停止"
+
+#~ msgid "関連ファイル"
+#~ msgstr "関連ファイル"
+
+#~ msgid "McKernelを用いたシステムを運用するシステム管理者を対象として、運用手順を説明する。"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "SMPプロセッサ向け、x86_64アーキ向けの関連ファイルの場所は以下の通り。 "
+#~ "なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+#~ msgstr "運用向けコマンド・デーモンのファイルの場所は以下の通り。 なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+
+#~ msgid ""
+#~ "SMPプロセッサ向け、x86_64アーキ向けの関連ファイルの場所は以下の通り。 "
+#~ "なお、IHK/McKernelのインストールディレクトリを<install>とする。 The related "
+#~ "files and their locations are as "
+#~ "follows. Denote by ``<install>`` the "
+#~ "install directory of IHK/McKernel."
+#~ msgstr ""
+
+#~ msgid ""
+#~ "The related files and their locations"
+#~ " are as follows. Denote by "
+#~ "``<install>`` the install directory of "
+#~ "IHK/McKernel."
+#~ msgstr ""
+#~ "SMPプロセッサ向け、x86_64アーキ向けの関連ファイルの場所は以下の通り。 "
+#~ "なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+
+#~ msgid "運用向けコマンド・デーモンのファイルの場所は以下の通り。 なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+#~ msgstr "運用向けコマンド・デーモンのファイルの場所は以下の通り。 なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+
+#~ msgid "運用向けコマンド・デーモンのファイルの場所は以下の通り。"
+#~ msgstr ""
+#~ "The commands and daemons for operation"
+#~ " and their locations are as follows."
+
+#~ msgid "カーネルイメージ"
+#~ msgstr ""
+
+#~ msgid "ダンプ解析ツール"
+#~ msgstr ""
+
+#~ msgid "ダンプ形式変換ツール"
+#~ msgstr ""
+
+#~ msgid "tool"
+#~ msgstr "ツール"
+

docs/locales/ja/LC_MESSAGES/ops.po

@@ -0,0 +1,821 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-06 10:10+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../ops.rst:5
+msgid "McKernelを用いたシステムを運用するシステム管理者を対象として、運用手順を説明する。"
+msgstr ""
+
+#: ../../ops.rst:8
+msgid "Installation"
+msgstr "インストール"
+
+#: ../../ops.rst:10
+msgid "See `Quick Guide -- Installation <quick.html#installation>`__."
+msgstr ""
+
+#: ../../uti.rst:2
+msgid "Advanced: Enable Utility Thread offloading Interface (UTI)"
+msgstr ""
+
+#: ../../uti.rst:4
+msgid ""
+"UTI enables a runtime such as MPI runtime to spawn utility threads such "
+"as MPI asynchronous progress threads to Linux cores."
+msgstr ""
+
+#: ../../uti.rst:8
+msgid "Install capstone"
+msgstr ""
+
+#: ../../uti.rst:11 ../../uti.rst:22
+msgid "When compute nodes don't have access to repositories"
+msgstr ""
+
+#: ../../uti.rst:13
+msgid "Install EPEL capstone-devel:"
+msgstr ""
+
+#: ../../uti.rst:24
+msgid ""
+"Ask the system administrator to install ``capstone-devel``. Note that it "
+"is in the EPEL repository."
+msgstr ""
+
+#: ../../uti.rst:28
+msgid "Install syscall_intercept"
+msgstr ""
+
+#: ../../uti.rst:37
+msgid "Install UTI for McKernel"
+msgstr ""
+
+#: ../../uti.rst:39
+msgid "Install:"
+msgstr ""
+
+#: ../../uti.rst:49
+msgid "Install McKernel"
+msgstr ""
+
+#: ../../uti.rst:51
+msgid "Add ``-DENABLE_UTI=ON`` option to ``cmake``:"
+msgstr ""
+
+#: ../../uti.rst:58
+msgid "Run programs"
+msgstr ""
+
+#: ../../uti.rst:60
+msgid "Add ``--enable-uti`` option to ``mcexec``:"
+msgstr ""
+
+#: ../../uti.rst:67
+msgid "Install UTI for Linux"
+msgstr ""
+
+#: ../../uti.rst:69
+msgid ""
+"You should skip this step if it's already installed as with, for example,"
+" Fujitsu Technical Computing Suite."
+msgstr ""
+
+#: ../../uti.rst:72
+msgid "Install by make"
+msgstr ""
+
+#: ../../uti.rst:82
+msgid "Install by rpm"
+msgstr ""
+
+#: ../../ops.rst:15
+msgid "起動停止"
+msgstr ""
+
+#: ../../ops.rst:18
+msgid "関連ファイル"
+msgstr ""
+
+#: ../../ops.rst:20
+msgid ""
+"SMPプロセッサ向け、x86_64アーキ向けの関連ファイルの場所は以下の通り。 "
+"なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+msgstr ""
+
+#: ../../ops.rst:24
+msgid "<install>/kmod/ihk.ko"
+msgstr ""
+
+#: ../../ops.rst:24
+msgid "IHK-master core"
+msgstr ""
+
+#: ../../ops.rst:26
+msgid "|ihk-smp|"
+msgstr ""
+
+#: ../../ops.rst:26
+msgid "IHK-master driver"
+msgstr ""
+
+#: ../../ops.rst:29
+msgid "|mcctrl|"
+msgstr ""
+
+#: ../../ops.rst:29
+msgid "Delegator module"
+msgstr ""
+
+#: ../../ops.rst:32
+msgid "|mckernel.img|"
+msgstr ""
+
+#: ../../ops.rst:32
+msgid "カーネルイメージ"
+msgstr ""
+
+#: ../../ops.rst:40
+msgid "運用向けコマンド・デーモンのファイルの場所は以下の通り。 なお、IHK/McKernelのインストールディレクトリを<install>とする。"
+msgstr ""
+
+#: ../../ops.rst:44
+msgid "|mcreboot|"
+msgstr ""
+
+#: ../../ops.rst:44 ../../ops.rst:104
+msgid "ブートスクリプト"
+msgstr ""
+
+#: ../../ops.rst:47
+msgid "|mcstop|"
+msgstr ""
+
+#: ../../ops.rst:47 ../../ops.rst:204
+msgid "シャットダウンスクリプト"
+msgstr ""
+
+#: ../../ops.rst:50
+msgid "<install>/bin/mcexec"
+msgstr ""
+
+#: ../../ops.rst:50 ../../ops.rst:242
+msgid "プロセス起動コマンド"
+msgstr ""
+
+#: ../../ops.rst:52
+msgid "<install>/bin/eclair"
+msgstr ""
+
+#: ../../ops.rst:52
+msgid "ダンプ解析ツール"
+msgstr ""
+
+#: ../../ops.rst:54
+msgid "|vmcore2mckdump|"
+msgstr ""
+
+#: ../../ops.rst:54
+msgid "ダンプ形式変換ツール"
+msgstr ""
+
+#: ../../ops.rst:62
+msgid "以下、関連コマンドおよび関連関数のインターフェイスを説明する。"
+msgstr ""
+
+#: ../../ops.rst:65
+msgid "インターフェイス"
+msgstr ""
+
+#: ../../ops.rst:68
+msgid "カーネル引数"
+msgstr ""
+
+#: ../../ops.rst:70
+msgid "McKernelのカーネル引数を表 :numref:`tab-kargs` に示す。"
+msgstr ""
+
+#: ../../ops.rst:74
+msgid "McKernelのカーネル引数"
+msgstr ""
+
+#: ../../ops.rst:77
+msgid "hidos"
+msgstr ""
+
+#: ../../ops.rst:77
+msgid "IKCを有効にする。"
+msgstr ""
+
+#: ../../ops.rst:79
+msgid "|dlv|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "Linuxのpanicハンドラ経由でダンプを行った場合の、ダ"
+msgstr ""
+
+#: ../../ops.rst
+msgid "ンプ対象とするメモリ領域の種類を<level>に設定する。"
+msgstr ""
+
+#: ../../ops.rst
+msgid "設定可能な値は以下の通り。"
+msgstr ""
+
+#: ../../ops.rst:85 ../../ops.rst:166 ../../ops.rst:198 ../../ops.rst:236
+msgid "0"
+msgstr ""
+
+#: ../../ops.rst:86 ../../ops.rst:167
+msgid "IHKがMcKernelに割り当てたメモリ領域を出力する"
+msgstr ""
+
+#: ../../ops.rst:87 ../../ops.rst:168
+msgid "24"
+msgstr ""
+
+#: ../../ops.rst:88 ../../ops.rst:169
+msgid "カーネルが使用しているメモリ領域を出力する"
+msgstr ""
+
+#: ../../ops.rst:90 ../../ops.rst:171
+msgid "指定がなかった場合は24が用いられる。"
+msgstr ""
+
+#: ../../ops.rst:92
+msgid "|allow|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "McKernelに割り当てられたCPU数より大きい数のスレッド"
+msgstr ""
+
+#: ../../ops.rst
+msgid "またはプロセスの生成を許可する。この引数が指定され"
+msgstr ""
+
+#: ../../ops.rst
+msgid "ない場合に、CPU数より大きい数のスレッドまたはプロセ"
+msgstr ""
+
+#: ../../ops.rst
+msgid "スをclone(), fork(), vfork()などで生成しようとする"
+msgstr ""
+
+#: ../../ops.rst
+msgid "と、当該システムコールがEINVALエラーを返す。"
+msgstr ""
+
+#: ../../ops.rst:107 ../../ops.rst:209
+msgid "書式"
+msgstr ""
+
+#: ../../ops.rst:115 ../../ops.rst:219
+msgid "オプション"
+msgstr ""
+
+#: ../../ops.rst:118
+msgid "|opt-c|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "McKernelに割り当てるCPUのリストを指定する。フォー"
+msgstr ""
+
+#: ../../ops.rst
+msgid "マットは以下の通り。"
+msgstr ""
+
+#: ../../ops.rst
+msgid "<CPU logical id>,<CPU logical id>...または"
+msgstr ""
+
+#: ../../ops.rst
+msgid "<CPU logical id>-<CPU logical id>,<CPU logical id>"
+msgstr ""
+
+#: ../../ops.rst
+msgid "-<CPU logical id>...または両者の混合。"
+msgstr ""
+
+#: ../../ops.rst:124
+msgid "|opt-r|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "McKernelのCPUがIKCメッセージを送るLinux"
+msgstr ""
+
+#: ../../ops.rst
+msgid "CPUを指定する。フォーマットは以下の通り。"
+msgstr ""
+
+#: ../../ops.rst
+msgid "<CPU list>:<CPU id>+<CPU list>:<CPU id>..."
+msgstr ""
+
+#: ../../ops.rst
+msgid "<CPU list>のフォーマットは-cオプションにおけるもの"
+msgstr ""
+
+#: ../../ops.rst
+msgid "と同じである。"
+msgstr ""
+
+#: ../../ops.rst
+msgid "各<CPU list>:<CPU id>は<CPU list>で示されるMcKernel"
+msgstr ""
+
+#: ../../ops.rst
+msgid "のCPUが<CPU logical id>で示されるLinuxのCPUにIKC"
+msgstr ""
+
+#: ../../ops.rst
+msgid "メッセージを送信することを意味する。"
+msgstr ""
+
+#: ../../ops.rst:133
+msgid "|opt-m|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "McKernelに割り当てるメモリ領域を指定する。フォーマッ"
+msgstr ""
+
+#: ../../ops.rst
+msgid "トは以下の通り。"
+msgstr ""
+
+#: ../../ops.rst
+msgid "<size>@<NUMA-id>, <size>@<NUMA-id>..."
+msgstr ""
+
+#: ../../ops.rst:137
+msgid "|opt-f|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "ihkmondが使用するsyslogプロトコルのfacilityを指定す"
+msgstr ""
+
+#: ../../ops.rst
+msgid "る。デフォルトはLOG_LOCAL6。"
+msgstr ""
+
+#: ../../ops.rst:140
+msgid "|opt-o|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "IHKのデバイスファイル(/dev/mcd*, /dev/mcos*)のオー"
+msgstr ""
+
+#: ../../ops.rst
+msgid "ナーとグループの値を<user>[:<group>]の形式で指定す"
+msgstr ""
+
+#: ../../ops.rst
+msgid "る。デフォルトはmcreboot.shを実行したユーザ。"
+msgstr ""
+
+#: ../../ops.rst:144
+msgid "|opt-i|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "ihkmondがハングアップ検知のためにOS状態を確認する時"
+msgstr ""
+
+#: ../../ops.rst
+msgid "間間隔を秒単位で指定する。-1が指定された場合はハン"
+msgstr ""
+
+#: ../../ops.rst
+msgid "グアップ検知を行わない。指定がない場合はハングアッ"
+msgstr ""
+
+#: ../../ops.rst
+msgid "プ検知を行わない。"
+msgstr ""
+
+#: ../../ops.rst:149
+msgid "|opt-k|"
+msgstr ""
+
+#: ../../ops.rst
+msgid "カーネルメッセージの/dev/logへのリダイレクト有無を"
+msgstr ""
+
+#: ../../ops.rst
+msgid "指定する。0が指定された場合はリダイレクトを行わず、"
+msgstr ""
+
+#: ../../ops.rst
+msgid "0以外が指定された場合はリダイレクトを行う。指定がな"
+msgstr ""
+
+#: ../../ops.rst
+msgid "い場合はリダイレクトを行わない。"
+msgstr ""
+
+#: ../../ops.rst:154
+msgid "-q <irq>"
+msgstr ""
+
+#: ../../ops.rst
+msgid "IHKが使用するIRQ番号を指定する。指定がない場合は"
+msgstr ""
+
+#: ../../ops.rst
+msgid "64-255の範囲で空いているものを使用する。"
+msgstr ""
+
+#: ../../ops.rst:157
+msgid "-t"
+msgstr ""
+
+#: ../../ops.rst
+msgid "（x86_64アーキテクチャのみ）Turbo"
+msgstr ""
+
+#: ../../ops.rst
+msgid "Boostをオンにする。デフォルトはオフ。"
+msgstr ""
+
+#: ../../ops.rst:160
+msgid "-d <level>"
+msgstr ""
+
+#: ../../ops.rst:173
+msgid "-O"
+msgstr ""
+
+#: ../../ops.rst
+msgid "またはプロセスの生成を許可する。指定がない場合は許可"
+msgstr ""
+
+#: ../../ops.rst
+msgid "しない。すなわち、CPU数より大きい数のスレッドまたは"
+msgstr ""
+
+#: ../../ops.rst
+msgid "プロセスを生成しようとするとエラーとなる。"
+msgstr ""
+
+#: ../../ops.rst:189 ../../ops.rst:226
+msgid "説明"
+msgstr ""
+
+#: ../../ops.rst:191
+msgid ""
+"McKernel関連カーネルモジュールをinsmodし、<cpulist>で指定されたCPUと<memlist>で指定されたメモリ領域からなるパーティションを作成し、IKC"
+" mapを<ikcmap>に設定し、前記パーティションにMcKernelをブートする。"
+msgstr ""
+
+#: ../../ops.rst:195 ../../ops.rst:233
+msgid "戻り値"
+msgstr ""
+
+#: ../../ops.rst:198 ../../ops.rst:236
+msgid "正常終了"
+msgstr ""
+
+#: ../../ops.rst:200 ../../ops.rst:238
+msgid "0以外"
+msgstr ""
+
+#: ../../ops.rst:200 ../../ops.rst:238
+msgid "エラー"
+msgstr ""
+
+#: ../../ops.rst:221
+msgid "なし"
+msgstr ""
+
+#: ../../ops.rst:228
+msgid "McKernelをシャットダウンし、McKernel用パーティションを削除し、関連カーネルモジュールをrmmodする。"
+msgstr ""
+
+#: ../../ops.rst:247
+msgid "ダンプ解析コマンド"
+msgstr ""
+
+#: ../../ops.rst:252
+msgid "ダンプ形式変換コマンド"
+msgstr ""
+
+#: ../../ops.rst:257
+msgid "ブート手順"
+msgstr ""
+
+#: ../../ops.rst:259
+msgid "mcreboot.shを用いてブート手順を説明する。"
+msgstr ""
+
+#: ../../ops.rst:261 ../../ops.rst:908
+msgid "スクリプトは以下の通り。"
+msgstr ""
+
+#: ../../ops.rst:854 ../../ops.rst:1052
+msgid "手順は以下の通り。"
+msgstr ""
+
+#: ../../ops.rst:856
+msgid "ihkmondを起動する。ihkmondは任意のタイミングで起動してよい。これは、ihkmondはOSインスタンスの作成を検知して動作を開始するためである。（83行目）"
+msgstr ""
+
+#: ../../ops.rst:858
+msgid "Linuxのカーネルバージョンが、mcoverlayfsが動作するものであるかを確認する。（200–216行目）"
+msgstr ""
+
+#: ../../ops.rst:860
+msgid "irqbalanceを停止する。（251–257行目）"
+msgstr ""
+
+#: ../../ops.rst:862
+msgid ""
+"/proc/irq/[n]/affinityの設定を保存した上でMcKernel "
+"CPUを担当から外す。担当CPUが無くなる場合は、全てのLinux CPUを指定する。（269–303行目）"
+msgstr ""
+
+#: ../../ops.rst:866
+msgid "ihk.koをinsmodする。（307行目）"
+msgstr ""
+
+#: ../../ops.rst:868
+msgid "Linuxによるメモリフラグメンテーションを緩和するために以下を実施する。（313–320行目）"
+msgstr ""
+
+#: ../../ops.rst:870
+msgid "アクティブでないプロセスを積極的にスワップアウトするように設定する"
+msgstr ""
+
+#: ../../ops.rst:872
+msgid "クリーンなページキャッシュを無効化し、またdentriesやinodeのslabオブジェクトのうち可能なものを破棄する"
+msgstr ""
+
+#: ../../ops.rst:874
+msgid "連続する空き領域を結合してより大きな空き領域にまとめる"
+msgstr ""
+
+#: ../../ops.rst:876
+msgid ""
+"ihk-smp-x86.koをinsmodする。（340行目）ihk-smp-x86.koは関数をihk.koに登録する。このため、ihk-"
+"smp-x86.koはihk.koをinsmodした後にinsmodする必要がある。"
+msgstr ""
+
+#: ../../ops.rst:878
+msgid "メモリを予約する。（370行目）"
+msgstr ""
+
+#: ../../ops.rst:880
+msgid "CPUを予約する。（374行目）"
+msgstr ""
+
+#: ../../ops.rst:882
+msgid "McKernelのカーネルモジュールmcctrl.koをinsmodする。（382行目）mcctrl.koはMcKernelブート時に呼ばれる関数をihk.koに登録する。このため、mcctrl.koのinsmodはihk.koのinsmodの後に、またブートの前に行う必要がある。"
+msgstr ""
+
+#: ../../ops.rst:884
+msgid "OSインスタンスを作成する。（406行目）"
+msgstr ""
+
+#: ../../ops.rst:886
+msgid "OSインスタンスにCPUを割り当てる。（412行目）"
+msgstr ""
+
+#: ../../ops.rst:888
+msgid "McKernel CPUのIKCメッセージ送信先のLinux CPUを設定する。（419行目）"
+msgstr ""
+
+#: ../../ops.rst:890
+msgid "OSインスタンスにメモリを割り当てる。（426行目）"
+msgstr ""
+
+#: ../../ops.rst:892
+msgid "カーネルイメージをロードする。（432行目）"
+msgstr ""
+
+#: ../../ops.rst:894
+msgid "カーネル引数をカーネルに渡す。（438行目）"
+msgstr ""
+
+#: ../../ops.rst:896
+msgid "カーネルをブートする。（444行目）"
+msgstr ""
+
+#: ../../ops.rst:898
+msgid ""
+"/proc, "
+"/sysファイルの準備をする。また、その中でmcoverlayfs.koをinsmodする。mcoverlayfs.koは他モジュールとの依存関係を持たない。（454行目から567行目）なお、関数インターフェイスでの対応関数はihk_os_create_pseudofs()である。"
+msgstr ""
+
+#: ../../ops.rst:900
+msgid "irqbalanceを、Linux CPUのみを対象とする設定で開始する。（569–587行目）"
+msgstr ""
+
+#: ../../ops.rst:904
+msgid "シャットダウン手順"
+msgstr ""
+
+#: ../../ops.rst:906
+msgid "mcstop+release.shを用いてシャットダウン手順を説明する。"
+msgstr ""
+
+#: ../../ops.rst:1054
+msgid "ブート時にLinux CPUのみを対象とする設定で開始されたirqbalanceを停止する。（24–33行目）"
+msgstr ""
+
+#: ../../ops.rst:1057
+msgid "全てのOSインスタンスを破壊する。OSインスタンスに割り当てられていた資源はIHKがLWKのために予約した状態に移行する。（35–50行目）"
+msgstr ""
+
+#: ../../ops.rst:1059
+msgid "IHKがLWKのために予約していた資源を開放する。（52–77行目）"
+msgstr ""
+
+#: ../../ops.rst:1061
+msgid "mcctrl.koをrmmodする。（81行目）"
+msgstr ""
+
+#: ../../ops.rst:1063
+msgid ""
+"/proc, "
+"/sysファイルの準備をする。また、その中でmcoverlayfs.koをrmmodする。（87–100行目）なお、関数インターフェイスでの対応関数はihk_os_destroy_pseudofs()である。"
+msgstr ""
+
+#: ../../ops.rst:1065
+msgid "ihk-smp-x86.koをrmmodする。（104行目）"
+msgstr ""
+
+#: ../../ops.rst:1067
+msgid "ihk.koをrmmodする。（112行目）"
+msgstr ""
+
+#: ../../ops.rst:1069
+msgid "ihkmondを停止する。（121行目）"
+msgstr ""
+
+#: ../../ops.rst:1071
+msgid "/proc/irq/[n]/affinityの設定をブート時に保存しておいたものに戻し、ブート前の設定でirqbalanceを開始する。（124–135行目）"
+msgstr ""
+
+#: ../../ops.rst:1073
+msgid "Linuxカーネルのスワップアウト積極度の設定をデフォルトの値に戻す。（138行目）"
+msgstr ""
+
+#~ msgid "The following OS distributions and platforms are recommended:"
+#~ msgstr "推奨OSディストリビューションとプロセッサは以下の通り。"
+
+#~ msgid "OS distribution"
+#~ msgstr ""
+
+#~ msgid "CentOS 7.3 or later"
+#~ msgstr ""
+
+#~ msgid "RHEL 7.3 or later"
+#~ msgstr ""
+
+#~ msgid "Platform"
+#~ msgstr ""
+
+#~ msgid "Intel Xeon"
+#~ msgstr ""
+
+#~ msgid "Intel Xeon Phi"
+#~ msgstr ""
+
+#~ msgid "Fujitsu A64FX"
+#~ msgstr ""
+
+#~ msgid "Prepare files for building McKernel"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Grant read permission to the System.map"
+#~ " file of your kernel version on "
+#~ "the build machine:"
+#~ msgstr ""
+
+#~ msgid "Install the following packages to the build machine:"
+#~ msgstr ""
+
+#~ msgid "When having access to repositories"
+#~ msgstr ""
+
+#~ msgid "On RHEL 8, enable the CodeReady Linux Builder (CLB) repository:"
+#~ msgstr ""
+
+#~ msgid "On CentOS 8, enable the PowerTools repository:"
+#~ msgstr ""
+
+#~ msgid "Install with yum:"
+#~ msgstr ""
+
+#~ msgid "When not having access to repositories"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Ask the system administrator to install"
+#~ " them. Note that ``libdwarf-devel`` "
+#~ "is in the CodeReady Linux Builder "
+#~ "repository on RHEL 8 or in the "
+#~ "PowerTools repository on CentOS 8."
+#~ msgstr ""
+
+#~ msgid "Clone, compile, install"
+#~ msgstr ""
+
+#~ msgid "Clone the source code:"
+#~ msgstr ""
+
+#~ msgid "(Optional) Checkout to the specific branch or version:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Foe example, if you want to try"
+#~ " the development branch, use “development”"
+#~ " as the pathspec. If you want "
+#~ "to try the prerelease version 1.7.0-0.2,"
+#~ " use “1.7.0-0.2”."
+#~ msgstr ""
+
+#~ msgid "Move to build directory:"
+#~ msgstr ""
+
+#~ msgid "Run cmake:"
+#~ msgstr ""
+
+#~ msgid "When not cross-compiling:"
+#~ msgstr ""
+
+#~ msgid "When cross-compiling:"
+#~ msgstr ""
+
+#~ msgid "Install with cmake"
+#~ msgstr ""
+
+#~ msgid "Install with make:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "The kernel modules and McKernel kernel"
+#~ " image should be installed under the"
+#~ " **ihk+mckernel** folder in your home "
+#~ "directory."
+#~ msgstr ""
+
+#~ msgid "Install with rpm"
+#~ msgstr ""
+
+#~ msgid "Create the tarball and the spec file:"
+#~ msgstr ""
+
+#~ msgid "Create the rpm package:"
+#~ msgstr ""
+
+#~ msgid "Install the rpm package:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "The kernel modules and McKernel kernel"
+#~ " image are installed under the "
+#~ "standard system directories."
+#~ msgstr ""
+
+#~ msgid "Prepare files and change settings for installing McKernel"
+#~ msgstr ""
+
+#~ msgid "Disable SELinux of the compute nodes:"
+#~ msgstr ""
+
+#~ msgid "Change the file to SELINUX=disabled. And then reboot the compute nodes:"
+#~ msgstr ""
+
+#~ msgid "Install the following packages to the compute nodes:"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "Ask the system administrator to install"
+#~ " them. Note that ``libdwarf`` is in"
+#~ " the CodeReady Linux Builder repository "
+#~ "on RHEL 8 or in the PowerTools "
+#~ "repository on CentOS 8."
+#~ msgstr ""
+

docs/locales/ja/LC_MESSAGES/quick.po

@@ -0,0 +1,458 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../quick.rst:6
+msgid "Introduction"
+msgstr ""
+
+#: ../../summary.rst:1
+msgid ""
+"IHK/McKernel is a light-weight multi-kernel operating system designed for"
+" high-end supercomputing. It runs Linux and McKernel, a light-weight "
+"kernel (LWK), side-by-side inside compute nodes and aims at the "
+"following:"
+msgstr ""
+
+#: ../../summary.rst:6
+msgid ""
+"Provide scalable and consistent execution of large-scale parallel "
+"scientific applications, but at the same time maintain the ability to "
+"rapidly adapt to new hardware features and emerging programming models"
+msgstr ""
+
+#: ../../summary.rst:10
+msgid ""
+"Provide efficient memory and device management so that resource "
+"contention and data movement are minimized at the system level"
+msgstr ""
+
+#: ../../summary.rst:12
+msgid ""
+"Eliminate OS noise by isolating OS services in Linux and provide jitter "
+"free execution on the LWK"
+msgstr ""
+
+#: ../../summary.rst:14
+msgid ""
+"Support the full POSIX/Linux APIs by selectively offloading (slow-path) "
+"system calls to Linux"
+msgstr ""
+
+#: ../../background.rst:2
+msgid "Background and Motivation"
+msgstr ""
+
+#: ../../background.rst:4
+msgid ""
+"With the growing complexity of high-end supercomputers, the current "
+"system software stack faces significant challenges as we move forward to "
+"exascale and beyond. The necessity to deal with extreme degree of "
+"parallelism, heterogeneous architectures, multiple levels of memory "
+"hierarchy, power constraints, etc., advocates operating systems that can "
+"rapidly adapt to new hardware requirements, and that can support novel "
+"programming paradigms and runtime systems. On the other hand, a new class"
+" of more dynamic and complex applications are also on the horizon, with "
+"an increasing demand for application constructs such as in-situ analysis,"
+" workflows, elaborate monitoring and performance tools. This complexity "
+"relies not only on the rich features of POSIX, but also on the Linux APIs"
+" (such as the */proc*, */sys* filesystems, etc.) in particular."
+msgstr ""
+
+#: ../../background.rst:19
+msgid "Two Traditional HPC OS Approaches"
+msgstr ""
+
+#: ../../background.rst:21
+msgid ""
+"Traditionally, light-weight operating systems specialized for HPC "
+"followed two approaches to tackle scalable execution of large-scale "
+"applications. In the full weight kernel (FWK) approach, a full Linux "
+"environment is taken as the basis, and features that inhibit attaining "
+"HPC scalability are removed, i.e., making it light-weight. The pure "
+"light-weight kernel (LWK) approach, on the other hand, starts from "
+"scratch and effort is undertaken to add sufficient functionality so that "
+"it provides a familiar API, typically something close to that of a "
+"general purpose OS, while at the same time it retains the desired "
+"scalability and reliability attributes. Neither of these approaches "
+"yields a fully Linux compatible environment."
+msgstr ""
+
+#: ../../background.rst:34
+msgid "The Multi-kernel Approach"
+msgstr ""
+
+#: ../../background.rst:36
+msgid ""
+"A hybrid approach recognized recently by the system software community is"
+" to run Linux simultaneously with a lightweight kernel on compute nodes "
+"and multiple research projects are now pursuing this direction. The basic"
+" idea is that simulations run on an HPC tailored lightweight kernel, "
+"ensuring the necessary isolation for noiseless execution of parallel "
+"applications, but Linux is leveraged so that the full POSIX API is "
+"supported. Additionally, the small code base of the LWK can also "
+"facilitate rapid prototyping for new, exotic hardware features. "
+"Nevertheless, the questions of how to share node resources between the "
+"two types of kernels, where do device drivers execute, how exactly do the"
+" two kernels interact with each other and to what extent are they "
+"integrated, remain subjects of ongoing debate."
+msgstr ""
+
+#: ../../archtecture.rst:2
+msgid "Architectural Overview"
+msgstr ""
+
+#: ../../archtecture.rst:4
+msgid ""
+"At the heart of the stack is a low-level software infrastructure called "
+"Interface for Heterogeneous Kernels (IHK). IHK is a general framework "
+"that provides capabilities for partitioning resources in a many-core "
+"environment (e.g.,CPU cores and physical memory) and it enables "
+"management of lightweight kernels. IHK can allocate and release host "
+"resources dynamically and no reboot of the host machine is required when "
+"altering configuration. IHK also provides a low-level inter-kernel "
+"messaging infrastructure, called the Inter-Kernel Communication (IKC) "
+"layer. An architectural overview of the main system components is shown "
+"below."
+msgstr ""
+
+#: ../../archtecture.rst:18
+msgid ""
+"McKernel is a lightweight kernel written from scratch. It is designed for"
+" HPC and is booted from IHK. McKernel retains a binary compatible ABI "
+"with Linux, however, it implements only a small set of performance "
+"sensitive system calls and the rest are offloaded to Linux. Specifically,"
+" McKernel has its own memory management, it supports processes and multi-"
+"threading with a simple round-robin cooperative (tick-less) scheduler, "
+"and it implements signaling. It also allows inter-process memory mappings"
+" and it provides interfaces to hardware performance counters."
+msgstr ""
+
+#: ../../archtecture.rst:29
+msgid "Functionality"
+msgstr ""
+
+#: ../../archtecture.rst:31
+msgid ""
+"An overview of some of the principal functionalities of the IHK/McKernel "
+"stack is provided below."
+msgstr ""
+
+#: ../../archtecture.rst:35
+msgid "System Call Offloading"
+msgstr ""
+
+#: ../../archtecture.rst:37
+msgid ""
+"System call forwarding in McKernel is implemented as follows. When an "
+"offloaded system call occurs, McKernel marshals the system call number "
+"along with its arguments and sends a message to Linux via a dedicated IKC"
+" channel. The corresponding proxy process running on Linux is by default "
+"waiting for system call requests through an ioctl() call into IHK’s "
+"system call delegator kernel module. The delegator kernel module’s IKC "
+"interrupt handler wakes up the proxy process, which returns to userspace "
+"and simply invokes the requested system call. Once it obtains the return "
+"value, it instructs the delegator module to send the result back to "
+"McKernel, which subsequently passes the value to user-space."
+msgstr ""
+
+#: ../../archtecture.rst:49
+msgid "Unified Address Space"
+msgstr ""
+
+#: ../../archtecture.rst:51
+msgid ""
+"The unified address space model in IHK/McKernel ensures that offloaded "
+"system calls can seamlessly resolve arguments even in case of pointers. "
+"This mechanism is depicted below and is implemented as follows."
+msgstr ""
+
+#: ../../archtecture.rst:58
+msgid ""
+"First, the proxy process is compiled as a position independent binary, "
+"which enables us to map the code and data segments specific to the proxy "
+"process to an address range which is explicitly excluded from McKernel’s "
+"user space. The grey box on the right side of the figure demonstrates the"
+" excluded region. Second, the entire valid virtual address range of "
+"McKernel’s application user-space is covered by a special mapping in the "
+"proxy process for which we use a pseudo file mapping in Linux. This "
+"mapping is indicated by the blue box on the left side of the figure."
+msgstr ""
+
+#: ../../install.rst:4
+msgid "Installation"
+msgstr "インストール"
+
+#: ../../install.rst:6
+msgid "The following OS distributions and platforms are recommended:"
+msgstr "推奨OSディストリビューションとプロセッサは以下の通り。"
+
+#: ../../install.rst:8
+msgid "OS distribution"
+msgstr ""
+
+#: ../../install.rst:10
+msgid "CentOS 7.3 or later"
+msgstr ""
+
+#: ../../install.rst:11
+msgid "RHEL 7.3 or later"
+msgstr ""
+
+#: ../../install.rst:13
+msgid "Platform"
+msgstr ""
+
+#: ../../install.rst:15
+msgid "Intel Xeon"
+msgstr ""
+
+#: ../../install.rst:16
+msgid "Intel Xeon Phi"
+msgstr ""
+
+#: ../../install.rst:17
+msgid "Fujitsu A64FX"
+msgstr ""
+
+#: ../../install.rst:20
+msgid "Prepare files for building McKernel"
+msgstr ""
+
+#: ../../install.rst:22
+msgid ""
+"Grant read permission to the System.map file of your kernel version on "
+"the build machine:"
+msgstr ""
+
+#: ../../install.rst:28
+msgid "Install the following packages to the build machine:"
+msgstr ""
+
+#: ../../install.rst:35 ../../install.rst:179
+msgid "When having access to repositories"
+msgstr ""
+
+#: ../../install.rst:37 ../../install.rst:181
+msgid "On RHEL 8, enable the CodeReady Linux Builder (CLB) repository:"
+msgstr ""
+
+#: ../../install.rst:43 ../../install.rst:187
+msgid "On CentOS 8, enable the PowerTools repository:"
+msgstr ""
+
+#: ../../install.rst:49 ../../install.rst:193
+msgid "Install with yum:"
+msgstr ""
+
+#: ../../install.rst:56 ../../install.rst:200
+msgid "When not having access to repositories"
+msgstr ""
+
+#: ../../install.rst:58
+msgid ""
+"Ask the system administrator to install them. Note that ``libdwarf-"
+"devel`` is in the CodeReady Linux Builder repository on RHEL 8 or in the "
+"PowerTools repository on CentOS 8."
+msgstr ""
+
+#: ../../install.rst:61
+msgid "Clone, compile, install"
+msgstr ""
+
+#: ../../install.rst:63
+msgid "Clone the source code:"
+msgstr ""
+
+#: ../../install.rst:71
+msgid "(Optional) Checkout to the specific branch or version:"
+msgstr ""
+
+#: ../../install.rst:79
+msgid ""
+"Foe example, if you want to try the development branch, use “development”"
+" as the pathspec. If you want to try the prerelease version 1.7.0-0.2, "
+"use “1.7.0-0.2”."
+msgstr ""
+
+#: ../../install.rst:83
+msgid "Move to build directory:"
+msgstr ""
+
+#: ../../install.rst:89
+msgid "Run cmake:"
+msgstr ""
+
+#: ../../install.rst:92 ../../install.rst:135
+msgid "When not cross-compiling:"
+msgstr ""
+
+#: ../../install.rst:99 ../../install.rst:142
+msgid "When cross-compiling:"
+msgstr ""
+
+#: ../../install.rst:111
+msgid "Install with cmake"
+msgstr ""
+
+#: ../../install.rst:113
+msgid "Install with make:"
+msgstr ""
+
+#: ../../install.rst:119
+msgid ""
+"The kernel modules and McKernel kernel image should be installed under "
+"the **ihk+mckernel** folder in your home directory."
+msgstr ""
+
+#: ../../install.rst:123
+msgid "Install with rpm"
+msgstr ""
+
+#: ../../install.rst:125
+msgid "Create the tarball and the spec file:"
+msgstr ""
+
+#: ../../install.rst:132
+msgid "Create the rpm package:"
+msgstr ""
+
+#: ../../install.rst:148
+msgid "Install the rpm package:"
+msgstr ""
+
+#: ../../install.rst:154
+msgid ""
+"The kernel modules and McKernel kernel image are installed under the "
+"standard system directories."
+msgstr ""
+
+#: ../../install.rst:158
+msgid "Prepare files and change settings for installing McKernel"
+msgstr ""
+
+#: ../../install.rst:160
+msgid "Disable SELinux of the compute nodes:"
+msgstr ""
+
+#: ../../install.rst:166
+msgid "Change the file to SELINUX=disabled. And then reboot the compute nodes:"
+msgstr ""
+
+#: ../../install.rst:172
+msgid "Install the following packages to the compute nodes:"
+msgstr ""
+
+#: ../../install.rst:202
+msgid ""
+"Ask the system administrator to install them. Note that ``libdwarf`` is "
+"in the CodeReady Linux Builder repository on RHEL 8 or in the PowerTools "
+"repository on CentOS 8."
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:2
+msgid "Boot McKernel"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:4
+msgid ""
+"A boot script called ``mcreboot.sh`` is provided under ``sbin`` in the "
+"install folder. To boot on logical CPU 1 with 512MB of memory, use the "
+"following invocation:"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:14
+msgid ""
+"You should see something similar like this if you display the McKernel’s "
+"kernel message log:"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:42
+msgid "Run a simple program on McKernel"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:44
+msgid ""
+"The mcexec command line tool (which is also the Linux proxy process) can "
+"be used for executing applications on McKernel:"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:53
+msgid "Shutdown McKernel"
+msgstr ""
+
+#: ../../boot_run_shutdown.rst:55
+msgid ""
+"Finally, to shutdown McKernel and release CPU/memory resources back to "
+"Linux use the following command:"
+msgstr ""
+
+#: ../../team.rst:2
+msgid "The Team"
+msgstr ""
+
+#: ../../team.rst:4
+msgid ""
+"The McKernel project was started at The University of Tokyo and currently"
+" it is mainly developed at RIKEN. Some of our collaborators include:"
+msgstr ""
+
+#: ../../team.rst:8
+msgid "Hitachi"
+msgstr ""
+
+#: ../../team.rst:9
+msgid "Fujitsu"
+msgstr ""
+
+#: ../../team.rst:10
+msgid "CEA (France)"
+msgstr ""
+
+#: ../../team.rst:11
+msgid "NEC"
+msgstr ""
+
+#: ../../license.rst:2
+msgid "License"
+msgstr ""
+
+#: ../../license.rst:4
+msgid "McKernel is GPL licensed, as found in the LICENSE file."
+msgstr ""
+
+#: ../../contact.rst:2
+msgid "Contact"
+msgstr ""
+
+#: ../../contact.rst:4
+msgid ""
+"Please give your feedback to us via one of the following mailing lists. "
+"Subscription via `www.pccluster.org "
+"<http://www.pccluster.org/mailman/listinfo/mckernel-users>`__ is needed."
+msgstr ""
+
+#: ../../contact.rst:9
+msgid "English: mckernel-users@pccluster.org"
+msgstr ""
+
+#: ../../contact.rst:10
+msgid "Japanese: mckernel-users-jp@pccluster.org"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/spec-ihk.po

@@ -0,0 +1,28 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../spec-ihk.md:1
+msgid "hi"
+msgstr ""
+
+#: ../../spec-ihk.md:3
+msgid ":download:IHK Spec <ihk.pdf>"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/spec-mckernel.po

@@ -0,0 +1,24 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../spec-mckernel.md:1
+msgid "Hello"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/spec.po

@@ -0,0 +1,48 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-07 10:00+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../spec/ihk.rst:6
+msgid "External Specs"
+msgstr ""
+
+#: ../../spec/ihk.rst:9
+msgid "Overview"
+msgstr ""
+
+#: ../../spec/ihk.rst:12
+msgid "Function Specs"
+msgstr ""
+
+#: ../../spec/ihk.rst:15
+msgid "Command / Daemon Specs"
+msgstr ""
+
+#: ../../spec/ihk.rst:18
+msgid "Booting LWK"
+msgstr ""
+
+#: ../../spec/mckernel.rst:6
+msgid "Interfaces"
+msgstr ""
+
+#: ../../spec/mckernel.rst:9
+msgid "Interface details"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/summary.po

@@ -0,0 +1,53 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../summary.rst:1
+msgid ""
+"IHK/McKernel is a light-weight multi-kernel operating system designed for"
+" high-end supercomputing. It runs Linux and McKernel, a light-weight "
+"kernel (LWK), side-by-side inside compute nodes and aims at the "
+"following:"
+msgstr ""
+
+#: ../../summary.rst:6
+msgid ""
+"Provide scalable and consistent execution of large-scale parallel "
+"scientific applications, but at the same time maintain the ability to "
+"rapidly adapt to new hardware features and emerging programming models"
+msgstr ""
+
+#: ../../summary.rst:10
+msgid ""
+"Provide efficient memory and device management so that resource "
+"contention and data movement are minimized at the system level"
+msgstr ""
+
+#: ../../summary.rst:12
+msgid ""
+"Eliminate OS noise by isolating OS services in Linux and provide jitter "
+"free execution on the LWK"
+msgstr ""
+
+#: ../../summary.rst:14
+msgid ""
+"Support the full POSIX/Linux APIs by selectively offloading (slow-path) "
+"system calls to Linux"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/team.po

@@ -0,0 +1,46 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../team.rst:2
+msgid "The Team"
+msgstr ""
+
+#: ../../team.rst:4
+msgid ""
+"The McKernel project was started at The University of Tokyo and currently"
+" it is mainly developed at RIKEN. Some of our collaborators include:"
+msgstr ""
+
+#: ../../team.rst:8
+msgid "Hitachi"
+msgstr ""
+
+#: ../../team.rst:9
+msgid "Fujitsu"
+msgstr ""
+
+#: ../../team.rst:10
+msgid "CEA (France)"
+msgstr ""
+
+#: ../../team.rst:11
+msgid "NEC"
+msgstr ""
+

docs/locales/ja/LC_MESSAGES/users.po

@@ -0,0 +1,454 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-07 10:00+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../users.rst:6
+msgid "Architectural Overview"
+msgstr ""
+
+#: ../../users.rst:8
+msgid ""
+"See `Quick Guide -- Architectural Overview <quick.html#architectural-"
+"overview>`__."
+msgstr ""
+
+#: ../../users.rst:11
+msgid "Running Programs"
+msgstr ""
+
+#: ../../users.rst:13
+msgid ""
+"You need to check if your application and pre-/post-processing programs "
+"are suited to run with McKernel. Follow the guide below to choose to run "
+"the whole on McKernel, or run the whole on Linux, or run pre-/post-"
+"processing on Linux and the application on McKernel:"
+msgstr ""
+
+#: ../../users.rst:16
+msgid "Application"
+msgstr ""
+
+#: ../../users.rst:18
+msgid ""
+"Run the whole on Linux if it issues system calls frequently and becoming "
+"the bottleneck with McKernel. For example, it's better to run on Linux "
+"those programs performing many file I/O operations."
+msgstr ""
+
+#: ../../users.rst:19
+msgid "Otherwise, run it on McKernel."
+msgstr ""
+
+#: ../../users.rst:21
+msgid "Pre-/Post-processing"
+msgstr ""
+
+#: ../../users.rst:23
+msgid ""
+"Run it on McKernel if it consumes a large amount of memory or the "
+"execution time isn't prolonged prohivitively with McKernel. The reason "
+"for the first condition is that the resource could be limited for Linux "
+"CPUs in the nodes for McKernel."
+msgstr ""
+
+#: ../../users.rst:24
+msgid "Otherwise, run it on Linux."
+msgstr ""
+
+#: ../../users.rst:28
+msgid "Modify job script"
+msgstr ""
+
+#: ../../users.rst:30
+msgid ""
+"When using job submission system, you need to modify the job scripts so "
+"that the job script itself is going to run on Linux. For example, with "
+"Fujitsu Technical Computing Suite (TCS), you need to specify "
+"``jobenv=linux`` by inserting the following line into the job script:"
+msgstr ""
+
+#: ../../users.rst:38
+msgid "Insert ``mcexec`` into the command line"
+msgstr ""
+
+#: ../../users.rst:40
+msgid ""
+"You need to insert ``mcexec`` into the command lines invoking the "
+"programs that you chose to run on McKernel:"
+msgstr ""
+
+#: ../../users.rst:43
+msgid "Non-MPI programs"
+msgstr ""
+
+#: ../../users.rst:45
+msgid "Insert ``mcexec`` before an executable:"
+msgstr ""
+
+#: ../../users.rst:52
+msgid "MPI programs"
+msgstr ""
+
+#: ../../users.rst:54
+msgid ""
+"Insert ``mcexec -n <processes-per-node>`` **after mpirun** and before an "
+"executable:"
+msgstr ""
+
+#: ../../users.rst:61
+msgid ""
+"``<processes-per-node>`` is the number of the processes per node and "
+"calculated by (number of MPI processes) / (number of nodes)."
+msgstr ""
+
+#: ../../users.rst:64
+msgid ""
+"For example, ``<processes-per-node>`` equals to 4 (=32/8) when specifying"
+" the number of processes and nodes as follows with Fujitsu Technical "
+"Computing Suite."
+msgstr ""
+
+#: ../../users.rst:74
+msgid "Limitations"
+msgstr ""
+
+#: ../../users.rst:76
+msgid ""
+"Pseudo devices such as /dev/mem and /dev/zero are not mmap()ed correctly "
+"even if the mmap() returns a success. An access of their mapping receives"
+" the SIGSEGV signal."
+msgstr ""
+
+#: ../../users.rst:80
+msgid ""
+"clone() supports only the following flags. All the other flags cause "
+"clone() to return error or are simply ignored."
+msgstr ""
+
+#: ../../users.rst:83
+msgid "CLONE_CHILD_CLEARTID"
+msgstr ""
+
+#: ../../users.rst:84
+msgid "CLONE_CHILD_SETTID"
+msgstr ""
+
+#: ../../users.rst:85
+msgid "CLONE_PARENT_SETTID"
+msgstr ""
+
+#: ../../users.rst:86
+msgid "CLONE_SETTLS"
+msgstr ""
+
+#: ../../users.rst:87
+msgid "CLONE_SIGHAND"
+msgstr ""
+
+#: ../../users.rst:88
+msgid "CLONE_VM"
+msgstr ""
+
+#: ../../users.rst:90
+msgid "PAPI has the following restriction."
+msgstr ""
+
+#: ../../users.rst:92
+msgid ""
+"Number of counters a user can use at the same time is up to the number of"
+" the physical counters in the processor."
+msgstr ""
+
+#: ../../users.rst:95
+msgid "msync writes back only the modified pages mapped by the calling process."
+msgstr ""
+
+#: ../../users.rst:98
+msgid "The following syscalls always return the ENOSYS error."
+msgstr ""
+
+#: ../../users.rst:100
+msgid "migrate_pages()"
+msgstr ""
+
+#: ../../users.rst:101
+msgid "move_pages()"
+msgstr ""
+
+#: ../../users.rst:102
+msgid "set_robust_list()"
+msgstr ""
+
+#: ../../users.rst:104
+msgid "The following syscalls always return the EOPNOTSUPP error."
+msgstr ""
+
+#: ../../users.rst:106
+msgid "arch_prctl(ARCH_SET_GS)"
+msgstr ""
+
+#: ../../users.rst:107
+msgid "signalfd()"
+msgstr ""
+
+#: ../../users.rst:109
+msgid "signalfd4() returns a fd, but signal is not notified through the fd."
+msgstr ""
+
+#: ../../users.rst:111
+msgid "set_rlimit sets the limit values but they are not enforced."
+msgstr ""
+
+#: ../../users.rst:113
+msgid "Address randomization is not supported."
+msgstr ""
+
+#: ../../users.rst:115
+msgid ""
+"brk() extends the heap more than requestd when -h (–extend-heap-by=) "
+"option of mcexec is used with the value larger than 4 KiB. "
+"syscall_pwrite02 of LTP would fail for this reason. This is because the "
+"test expects that the end of the heap is set to the same address as the "
+"argument of sbrk() and expects a segmentation violation occurs when it "
+"tries to access the memory area right next to the boundary. However, the "
+"optimization sets the end to a value larger than the requested. "
+"Therefore, the expected segmentation violation doesn’t occur."
+msgstr ""
+
+#: ../../users.rst:125
+msgid ""
+"setpriority()/getpriority() won’t work. They might set/get the priority "
+"of a random mcexec thread. This is because there’s no fixed "
+"correspondence between a McKernel thread which issues the system call and"
+" a mcexec thread which handles the offload request."
+msgstr ""
+
+#: ../../users.rst:130
+msgid ""
+"mbind() can set the policy but it is not used when allocating physical "
+"pages."
+msgstr ""
+
+#: ../../users.rst:133
+msgid ""
+"MPOL_F_RELATIVE_NODES and MPOL_INTERLEAVE flags for "
+"set_mempolicy()/mbind() are not supported."
+msgstr ""
+
+#: ../../users.rst:136
+msgid ""
+"The MPOL_BIND policy for set_mempolicy()/mbind() works as the same as the"
+" MPOL_PREFERRED policy. That is, the physical page allocator doesn’t give"
+" up the allocation when the specified nodes are running out of pages but "
+"continues to search pages in the other nodes."
+msgstr ""
+
+#: ../../users.rst:141
+msgid ""
+"Kernel dump on Linux panic requires Linux kernel CentOS-7.4 and later. In"
+" addition, crash_kexec_post_notifiers kernel argument must be given to "
+"Linux kernel."
+msgstr ""
+
+#: ../../users.rst:145
+msgid ""
+"setfsuid()/setfsgid() cannot change the id of the calling thread. "
+"Instead, it changes that of the mcexec worker thread which takes the "
+"system-call offload request."
+msgstr ""
+
+#: ../../users.rst:149
+msgid ""
+"mmap (hugeTLBfs): The physical pages corresponding to a map are released "
+"when no McKernel process exist. The next map gets fresh physical pages."
+msgstr ""
+
+#: ../../users.rst:153
+msgid "Sticky bit on executable file has no effect."
+msgstr ""
+
+#: ../../users.rst:155
+msgid ""
+"Linux (RHEL-7 for x86_64) could hang when offlining CPUs in the process "
+"of booting McKernel due to the Linux bug, found in Linux-3.10 and fixed "
+"in the later version. One way to circumvent this is to always assign the "
+"same CPU set to McKernel."
+msgstr ""
+
+#: ../../users.rst:160
+msgid "madvise:"
+msgstr ""
+
+#: ../../users.rst:162
+msgid "MADV_HWPOISON and MADV_SOFT_OFFLINE always returns -EPERM."
+msgstr ""
+
+#: ../../users.rst:163
+msgid "MADV_MERGEABLE and MADV_UNMERGEABLE always returns -EINVAL."
+msgstr ""
+
+#: ../../users.rst:164
+msgid ""
+"MADV_HUGEPAGE and MADV_NOHUGEPAGE on file map returns -EINVAL except on "
+"RHEL-8 for aarch64."
+msgstr ""
+
+#: ../../users.rst:167
+msgid ""
+"brk() and mmap() doesn’t report out-of-memory through its return value. "
+"Instead, page-fault reports the error."
+msgstr ""
+
+#: ../../users.rst:170
+msgid ""
+"Anonymous mmap pre-maps requested number of pages when contiguous pages "
+"are available. Demand paging is used when not available."
+msgstr ""
+
+#: ../../users.rst:173
+msgid ""
+"Mixing page sizes in anonymous shared mapping is not allowed. mmap "
+"creates vm_range with one page size. And munmap or mremap that needs the "
+"reduced page size changes the sizes of all the pages of the vm_range."
+msgstr ""
+
+#: ../../users.rst:178
+msgid ""
+"ihk_os_getperfevent() could time-out when invoked from Fujitsu TCS (job-"
+"scheduler)."
+msgstr ""
+
+#: ../../users.rst:181
+msgid ""
+"The behaviors of madvise and mbind are changed to do nothing and report "
+"success as a workaround for Fugaku."
+msgstr ""
+
+#: ../../users.rst:184
+msgid ""
+"mmap() allows unlimited overcommit. Note that it corresponds to setting "
+"sysctl ``vm.overcommit_memory`` to 1."
+msgstr ""
+
+#~ msgid ""
+#~ "At the heart of the stack is "
+#~ "a low-level software infrastructure "
+#~ "called Interface for Heterogeneous Kernels "
+#~ "(IHK). IHK is a general framework "
+#~ "that provides capabilities for partitioning"
+#~ " resources in a many-core environment"
+#~ " (e.g.,CPU cores and physical memory) "
+#~ "and it enables management of lightweight"
+#~ " kernels. IHK can allocate and "
+#~ "release host resources dynamically and "
+#~ "no reboot of the host machine is"
+#~ " required when altering configuration. IHK"
+#~ " also provides a low-level inter-"
+#~ "kernel messaging infrastructure, called the"
+#~ " Inter-Kernel Communication (IKC) layer."
+#~ " An architectural overview of the "
+#~ "main system components is shown below."
+#~ msgstr ""
+
+#~ msgid ""
+#~ "McKernel is a lightweight kernel written"
+#~ " from scratch. It is designed for "
+#~ "HPC and is booted from IHK. "
+#~ "McKernel retains a binary compatible ABI"
+#~ " with Linux, however, it implements "
+#~ "only a small set of performance "
+#~ "sensitive system calls and the rest "
+#~ "are offloaded to Linux. Specifically, "
+#~ "McKernel has its own memory management,"
+#~ " it supports processes and multi-"
+#~ "threading with a simple round-robin "
+#~ "cooperative (tick-less) scheduler, and "
+#~ "it implements signaling. It also allows"
+#~ " inter-process memory mappings and it"
+#~ " provides interfaces to hardware "
+#~ "performance counters."
+#~ msgstr ""
+
+#~ msgid "Functionality"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "An overview of some of the "
+#~ "principal functionalities of the IHK/McKernel"
+#~ " stack is provided below."
+#~ msgstr ""
+
+#~ msgid "System Call Offloading"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "System call forwarding in McKernel is"
+#~ " implemented as follows. When an "
+#~ "offloaded system call occurs, McKernel "
+#~ "marshals the system call number along"
+#~ " with its arguments and sends a "
+#~ "message to Linux via a dedicated "
+#~ "IKC channel. The corresponding proxy "
+#~ "process running on Linux is by "
+#~ "default waiting for system call requests"
+#~ " through an ioctl() call into IHK’s"
+#~ " system call delegator kernel module. "
+#~ "The delegator kernel module’s IKC "
+#~ "interrupt handler wakes up the proxy "
+#~ "process, which returns to userspace and"
+#~ " simply invokes the requested system "
+#~ "call. Once it obtains the return "
+#~ "value, it instructs the delegator module"
+#~ " to send the result back to "
+#~ "McKernel, which subsequently passes the "
+#~ "value to user-space."
+#~ msgstr ""
+
+#~ msgid "Unified Address Space"
+#~ msgstr ""
+
+#~ msgid ""
+#~ "The unified address space model in "
+#~ "IHK/McKernel ensures that offloaded system "
+#~ "calls can seamlessly resolve arguments "
+#~ "even in case of pointers. This "
+#~ "mechanism is depicted below and is "
+#~ "implemented as follows."
+#~ msgstr ""
+
+#~ msgid ""
+#~ "First, the proxy process is compiled "
+#~ "as a position independent binary, which"
+#~ " enables us to map the code and"
+#~ " data segments specific to the proxy"
+#~ " process to an address range which"
+#~ " is explicitly excluded from McKernel’s "
+#~ "user space. The grey box on the"
+#~ " right side of the figure "
+#~ "demonstrates the excluded region. Second, "
+#~ "the entire valid virtual address range"
+#~ " of McKernel’s application user-space "
+#~ "is covered by a special mapping in"
+#~ " the proxy process for which we "
+#~ "use a pseudo file mapping in "
+#~ "Linux. This mapping is indicated by "
+#~ "the blue box on the left side "
+#~ "of the figure."
+#~ msgstr ""
+

docs/locales/ja/LC_MESSAGES/uti.po

@@ -0,0 +1,94 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) 2020, Masamichi Takagi, Balazs Gerofi, Yutaka Ishikawa
+# This file is distributed under the same license as the IHK/McKernel
+# package.
+# FIRST AUTHOR <EMAIL@ADDRESS>, 2020.
+#
+#, fuzzy
+msgid ""
+msgstr ""
+"Project-Id-Version: IHK/McKernel \n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2020-08-04 16:40+0900\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: Babel 2.7.0\n"
+
+#: ../../uti.rst:2
+msgid "Advanced: Enable Utility Thread offloading Interface (UTI)"
+msgstr ""
+
+#: ../../uti.rst:4
+msgid ""
+"UTI enables a runtime such as MPI runtime to spawn utility threads such "
+"as MPI asynchronous progress threads to Linux cores."
+msgstr ""
+
+#: ../../uti.rst:8
+msgid "Install capstone"
+msgstr ""
+
+#: ../../uti.rst:11 ../../uti.rst:22
+msgid "When compute nodes don't have access to repositories"
+msgstr ""
+
+#: ../../uti.rst:13
+msgid "Install EPEL capstone-devel:"
+msgstr ""
+
+#: ../../uti.rst:24
+msgid ""
+"Ask the system administrator to install ``capstone-devel``. Note that it "
+"is in the EPEL repository."
+msgstr ""
+
+#: ../../uti.rst:28
+msgid "Install syscall_intercept"
+msgstr ""
+
+#: ../../uti.rst:37
+msgid "Install UTI for McKernel"
+msgstr ""
+
+#: ../../uti.rst:39
+msgid "Install:"
+msgstr ""
+
+#: ../../uti.rst:49
+msgid "Install McKernel"
+msgstr ""
+
+#: ../../uti.rst:51
+msgid "Add ``-DENABLE_UTI=ON`` option to ``cmake``:"
+msgstr ""
+
+#: ../../uti.rst:58
+msgid "Run programs"
+msgstr ""
+
+#: ../../uti.rst:60
+msgid "Add ``--enable-uti`` option to ``mcexec``:"
+msgstr ""
+
+#: ../../uti.rst:67
+msgid "Install UTI for Linux"
+msgstr ""
+
+#: ../../uti.rst:69
+msgid ""
+"You should skip this step if it's already installed as with, for example,"
+" Fujitsu Technical Computing Suite."
+msgstr ""
+
+#: ../../uti.rst:72
+msgid "Install by make"
+msgstr ""
+
+#: ../../uti.rst:82
+msgid "Install by rpm"
+msgstr ""
+

docs/logo.rst

@@ -0,0 +1 @@
+.. figure:: mckernel-logo.png

docs/mckernel_developers.rst

@@ -0,0 +1,9 @@
+.. sectnum::
+   :suffix: .
+   :depth: 3
+
+Interfaces
+==========
+
+Interface details
+=================

docs/quick.rst

@@ -0,0 +1,22 @@
+.. sectnum::
+   :suffix: .
+   :depth: 3
+
+Introduction
+============
+
+.. include:: summary.rst
+
+.. include:: background.rst
+
+.. include:: archtecture.rst
+
+.. include:: install.rst
+
+.. include:: boot_run_shutdown.rst
+
+.. include:: team.rst
+
+.. include:: license.rst
+
+.. include:: contact.rst

docs/requirements.txt

@@ -0,0 +1,30 @@
+alabaster==0.7.12
+attrs==19.3.0
+Babel==2.7.0
+certifi==2019.11.28
+chardet==3.0.4
+commonmark==0.9.1
+docutils==0.15
+idna==2.8
+imagesize==1.1.0
+Jinja2==2.10.3
+MarkupSafe==1.1.1
+mistune==0.8.4
+packaging==19.2
+Pygments==2.5.2
+pyparsing==2.4.5
+pytz==2019.3
+recommonmark==0.6.0
+requests==2.22.0
+roman==3.2
+six==1.13.0
+snowballstemmer==2.0.0
+Sphinx==2.2.2
+sphinx-rtd-theme==0.5.0
+sphinxcontrib-applehelp==1.0.1
+sphinxcontrib-devhelp==1.0.1
+sphinxcontrib-htmlhelp==1.0.2
+sphinxcontrib-jsmath==1.0.1
+sphinxcontrib-qthelp==1.0.2
+sphinxcontrib-serializinghtml==1.1.3
+urllib3==1.25.7

docs/spec-ihk.md

@@ -0,0 +1,3 @@
+## hi
+
+:download:`IHK Spec <ihk.pdf>`

docs/spec-mckernel.md

@@ -0,0 +1 @@
+## Hello

docs/spec/ihk.rst

@@ -0,0 +1,7 @@
+.. sectnum::
+   :suffix: .
+   :depth: 3
+
+Specifications
+==============
+The specifications pdf is :download:`here <ihk.pdf>`

docs/spec/mckernel.rst

@@ -0,0 +1,7 @@
+.. sectnum::
+   :suffix: .
+   :depth: 3
+
+Specifications
+==============
+The specifications pdf is :download:`here <mckernel.pdf>`

docs/summary.rst

@@ -0,0 +1,15 @@
+IHK/McKernel is a light-weight multi-kernel operating system designed
+for high-end supercomputing. It runs Linux and McKernel, a light-weight
+kernel (LWK), side-by-side inside compute nodes and aims at the
+following:
+
+-  Provide scalable and consistent execution of large-scale parallel
+   scientific applications, but at the same time maintain the ability to
+   rapidly adapt to new hardware features and emerging programming
+   models
+-  Provide efficient memory and device management so that resource
+   contention and data movement are minimized at the system level
+-  Eliminate OS noise by isolating OS services in Linux and provide
+   jitter free execution on the LWK
+-  Support the full POSIX/Linux APIs by selectively offloading
+   (slow-path) system calls to Linux

docs/team.rst

@@ -0,0 +1,12 @@
+The Team
+========
+
+The McKernel project was started at The University of Tokyo and
+currently it is mainly developed at RIKEN. Some of our collaborators
+include:
+
+-  Hitachi
+-  Fujitsu
+-  CEA (France)
+-  NEC
+

docs/users.rst

@@ -0,0 +1,204 @@
+.. sectnum::
+   :suffix: .
+   :depth: 3
+
+Architectural Overview
+======================
+
+See `Quick Guide -- Architectural Overview <quick.html#architectural-overview>`__.
+
+Running Programs
+================
+
+You need to check if your application and pre-/post-processing programs are suited to run with McKernel.
+Follow the guide below to choose to run the whole on McKernel, or run the whole on Linux, or run pre-/post-processing on Linux and the application on McKernel:
+
+* Application
+
+  - Run the whole on Linux if it issues system calls frequently and becoming the bottleneck with McKernel, e.g., those performing many file I/O operations.
+  - Otherwise, run it on McKernel.
+
+* Pre-/Post-processing
+
+  - Run it on McKernel if it consumes a large amount of memory or the execution time isn't prolonged prohivitively with McKernel. The reason for the first condition is that the resource could be limited for Linux CPUs in the nodes for McKernel.
+  - Otherwise, run it on Linux.
+
+
+Modify job script
+-----------------
+
+When using job submission system, you need to modify the job scripts so that the job script itself is going to run on Linux.
+For example, with Fujitsu Technical Computing Suite (TCS), you need to specify ``jobenv=mck1`` by inserting the following line into the job script:
+
+.. code-block:: none
+
+   #PJM -L jobenv=mck1
+
+(Optional, Fujitsu TCS only) Specify boot parameters
+----------------------------------------------------
+
+You can specify the boot parameters by defining environmental variables and pass them to Fujitsu TCS.
+The parameters include the resource reservation settings, resource reservation amount, kernel arguments and routing of message channels between McKernel CPUs and Linux CPUs.
+See `IHK Specifications - ihk_create_os_str() <spec/ihk.html>`__ for the parameter names and allowed values.
+The example of setting the memory amount is shown below.
+
+.. code-block:: none
+
+   export IHK_MEM="7G@4,7G@5,7G@6,7G@7"
+   pjsub -X run.sh
+
+Insert ``mcexec`` into the command line
+---------------------------------------
+
+You need to insert ``mcexec`` into the command lines invoking the programs that you chose to run on McKernel:
+
+Non-MPI programs
+~~~~~~~~~~~~~~~~
+
+Insert ``mcexec`` before an executable:
+
+::
+
+   mcexec ./a.out
+
+MPI programs
+~~~~~~~~~~~~
+
+Insert ``mcexec -n <processes-per-node>`` **after mpirun** and before an
+executable:
+
+::
+
+   mpirun -n <number-of-MPI-processes> mcexec -n <processes-per-node> ./a.out
+
+``<processes-per-node>`` is the number of the processes per node and
+calculated by (number of MPI processes) / (number of nodes).
+
+For example, ``<processes-per-node>`` equals to 4 (=32/8) when
+specifying the number of processes and nodes as follows with
+Fujitsu Technical Computing Suite.
+
+.. code-block:: none
+
+   #PJM --mpi "proc=32"
+   #PJM -L "node=8"
+
+Limitations
+===========
+
+1.  Pseudo devices such as /dev/mem and /dev/zero are not mmap()ed
+    correctly even if the mmap() returns a success. An access of their
+    mapping receives the SIGSEGV signal.
+
+2.  clone() supports only the following flags. All the other flags cause
+    clone() to return error or are simply ignored.
+
+    -  CLONE_CHILD_CLEARTID
+    -  CLONE_CHILD_SETTID
+    -  CLONE_PARENT_SETTID
+    -  CLONE_SETTLS
+    -  CLONE_SIGHAND
+    -  CLONE_VM
+
+3.  PAPI has the following restriction.
+
+    -  Number of counters a user can use at the same time is up to the
+       number of the physical counters in the processor.
+
+4.  msync writes back only the modified pages mapped by the calling
+    process.
+
+5.  The following syscalls always return the ENOSYS error.
+
+    -  migrate_pages()
+    -  move_pages()
+    -  set_robust_list()
+
+6.  The following syscalls always return the EOPNOTSUPP error.
+
+    -  arch_prctl(ARCH_SET_GS)
+    -  signalfd()
+
+7.  signalfd4() returns a fd, but signal is not notified through the fd.
+
+8.  set_rlimit sets the limit values but they are not enforced.
+
+9.  Address randomization is not supported.
+
+10. brk() extends the heap more than requestd when -h (–extend-heap-by=)
+    option of mcexec is used with the value larger than 4 KiB.
+    syscall_pwrite02 of LTP would fail for this reason. This is because
+    the test expects that the end of the heap is set to the same address
+    as the argument of sbrk() and expects a segmentation violation
+    occurs when it tries to access the memory area right next to the
+    boundary. However, the optimization sets the end to a value larger
+    than the requested. Therefore, the expected segmentation violation
+    doesn’t occur.
+
+11. setpriority()/getpriority() won’t work. They might set/get the
+    priority of a random mcexec thread. This is because there’s no fixed
+    correspondence between a McKernel thread which issues the system
+    call and a mcexec thread which handles the offload request.
+
+12. mbind() can set the policy but it is not used when allocating
+    physical pages.
+
+13. MPOL_F_RELATIVE_NODES and MPOL_INTERLEAVE flags for
+    set_mempolicy()/mbind() are not supported.
+
+14. The MPOL_BIND policy for set_mempolicy()/mbind() works as the same
+    as the MPOL_PREFERRED policy. That is, the physical page allocator
+    doesn’t give up the allocation when the specified nodes are running
+    out of pages but continues to search pages in the other nodes.
+
+15. Kernel dump on Linux panic requires Linux kernel CentOS-7.4 and
+    later. In addition, crash_kexec_post_notifiers kernel argument must
+    be given to Linux kernel.
+
+16. setfsuid()/setfsgid() cannot change the id of the calling thread.
+    Instead, it changes that of the mcexec worker thread which takes the
+    system-call offload request.
+
+17. mmap (hugeTLBfs): The physical pages corresponding to a map are
+    released when no McKernel process exist. The next map gets fresh
+    physical pages.
+
+18. Sticky bit on executable file has no effect.
+
+19. Linux (RHEL-7 for x86_64) could hang when offlining CPUs in the
+    process of booting McKernel due to the Linux bug, found in
+    Linux-3.10 and fixed in the later version. One way to circumvent
+    this is to always assign the same CPU set to McKernel.
+
+20. madvise:
+
+    -  MADV_HWPOISON and MADV_SOFT_OFFLINE always returns -EPERM.
+    -  MADV_MERGEABLE and MADV_UNMERGEABLE always returns -EINVAL.
+    -  MADV_HUGEPAGE and MADV_NOHUGEPAGE on file map returns -EINVAL
+       except on RHEL-8 for aarch64.
+
+21. brk() and mmap() doesn’t report out-of-memory through its return
+    value. Instead, page-fault reports the error.
+
+22. Anonymous mmap pre-maps requested number of pages when contiguous
+    pages are available. Demand paging is used when not available.
+
+23. Mixing page sizes in anonymous shared mapping is not allowed. mmap
+    creates vm_range with one page size. And munmap or mremap that needs
+    the reduced page size changes the sizes of all the pages of the
+    vm_range.
+
+24. ihk_os_getperfevent() could time-out when invoked from Fujitsu TCS
+    (job-scheduler).
+
+25. The behaviors of madvise and mbind are changed to do nothing and
+    report success as a workaround for Fugaku.
+
+26. mmap() allows unlimited overcommit. Note that it corresponds to
+    setting sysctl ``vm.overcommit_memory`` to 1.
+
+27. mlockall() is not supported and returns -EPERM.
+
+28. munlockall() is not supported and returns zero.
+
+29. scheduling behavior is not Linux compatible. For example, sometimes one of the two processes on the same CPU continues to run after yielding.

docs/uti.rst

@@ -0,0 +1,126 @@
+Advanced: Enable Utility Thread offloading Interface (UTI)
+-------------------------------------------------------------
+
+UTI enables a runtime such as MPI runtime to spawn utility threads such
+as MPI asynchronous progress threads to Linux cores.
+
+Install ``capstone`` and ``capstone-devel``
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+When compute nodes don't have access to EPEL repository
+"""""""""""""""""""""""""""""""""""""""""""""""""""""""
+
+Install EPEL ``capstone`` and ``capstone-devel``:
+
+::
+
+   sudo yum install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm
+   sudo yum install capstone capstone-devel
+
+
+When compute nodes don't have access to EPEL repository
+"""""""""""""""""""""""""""""""""""""""""""""""""""""""
+
+A. Ask the system administrator to install ``capstone`` and ``capstone-devel``. Note that it is in the EPEL repository.
+
+B. Download the rpm with the machine in which you are the administrator:
+
+::
+
+   sudo yum install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm
+   sudo yum install yum-utils
+   yumdownloader capstone capstone-devel
+
+and then install it to your home directory of the login node:
+
+::
+
+   cd $HOME/$(uname -p)
+   rpm2cpio capstone-4.0.1-9.el8.aarch64.rpm | cpio -idv
+   rpm2cpio capstone-devel-4.0.1-9.el8.aarch64.rpm | cpio -idv
+   sed -i 's#/usr/#'"$HOME"'/'"$(uname -p)"'/usr/#' $HOME/$(uname -p)/usr/lib64/pkgconfig/capstone.pc
+
+
+Install syscall_intercept
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+::
+
+   git clone https://github.com/RIKEN-SysSoft/syscall_intercept.git
+   mkdir build && cd build
+
+When ``capstone`` and ``capstone-devel`` are installed into the system directory:
+
+::
+
+   cmake ../syscall_intercept/arch/aarch64 -DCMAKE_INSTALL_PREFIX=${HOME}/$(uname -p)/usr -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=gcc -DTREAT_WARNINGS_AS_ERRORS=OFF
+
+When ``capstone`` and ``capstone-devel`` are installed into your home directory:
+
+::
+
+   CMAKE_PREFIX_PATH=${HOME}/$(uname -p)/usr cmake ../syscall_intercept/arch/aarch64 -DCMAKE_INSTALL_PREFIX=${HOME}/$(uname -p)/usr -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=gcc -DTREAT_WARNINGS_AS_ERRORS=OFF
+
+Install:
+
+::
+
+   make && make install && make test
+
+Install UTI for McKernel
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Install:
+
+.. code-block:: none
+
+   git clone https://github.com/RIKEN-SysSoft/uti.git
+   mkdir build && cd build
+   ../uti/configure --prefix=<mckernel-install> --with-rm=mckernel
+   make && make install
+
+Install McKernel
+~~~~~~~~~~~~~~~~~~~~
+
+``cmake`` with the additional options:
+
+::
+
+   CMAKE_PREFIX_PATH=${HOME}/$(uname -p)/usr cmake -DCMAKE_INSTALL_PREFIX=${HOME}/ihk+mckernel -DENABLE_UTI=ON $HOME/src/ihk+mckernel/mckernel
+   make -j install
+
+Run programs
+~~~~~~~~~~~~~~~~
+
+``mcexec`` with ``--enable-uti`` option:
+
+::
+
+   mcexec --enable-uti <command>
+
+Install UTI for Linux
+~~~~~~~~~~~~~~~~~~~~~~~~~
+
+You should skip this step if it's already installed as with, for example, Fujitsu Technical Computing Suite.
+
+Install by make
+"""""""""""""""
+
+.. code-block:: none
+
+   git clone https://github.com/RIKEN-SysSoft/uti.git
+   mkdir build && cd build
+   ../uti/configure --prefix=<uti-install> --with-rm=linux
+   make && make install
+
+Install by rpm
+""""""""""""""
+
+.. code-block:: none
+
+   git clone https://github.com/RIKEN-SysSoft/uti.git
+   mkdir build && cd build
+   ../uti/configure --prefix=<uti-install> --with-rm=linux
+   rm -f ~/rpmbuild/SOURCES/<version>.tar.gz
+   rpmbuild -ba ./scripts/uti.spec
+   rpm -Uvh uti-<version>-<release>-<arch>.rpm

executer/include/defs.h

@@ -4671,7 +4671,7 @@ void cmd_ipcs(void);         /* ipcs.c */
 /*
  *  main.c
  */
-void main_loop(void);
+//void main_loop(void);
 void exec_command(void);
 struct command_table_entry *get_command_table_entry(char *);
 void program_usage(int);

executer/include/uprotocol.h

@@ -91,7 +91,10 @@ struct program_image_section {
 
 struct get_cpu_set_arg {
 	int nr_processes;
+	char *req_cpu_list;   // Requested by user-space
+	int req_cpu_list_len; // Lenght of request string
 	int *process_rank;
+	pid_t ppid;
 	void *cpu_set;
 	size_t cpu_set_size;	// Size in bytes
 	int *target_core;
@@ -110,6 +113,18 @@ typedef unsigned long __cpu_set_unit;
 #define MPOL_NO_BSS               0x04
 #define MPOL_SHM_PREMAP           0x08
 
+/* should be the same as process.h */
+#define PLD_PROCESS_NUMA_MASK_BITS 256
+
+enum {
+	PLD_MPOL_DEFAULT,
+	PLD_MPOL_PREFERRED,
+	PLD_MPOL_BIND,
+	PLD_MPOL_INTERLEAVE,
+	PLD_MPOL_LOCAL,
+	PLD_MPOL_MAX,	/* always last member of enum */
+};
+
 #define PLD_MAGIC 0xcafecafe44332211UL
 
 struct program_load_desc {
@@ -144,9 +159,18 @@ struct program_load_desc {
 	unsigned long heap_extension;
 	long stack_premap;
 	unsigned long mpol_bind_mask;
+	int mpol_mode;
+	unsigned long mpol_nodemask[PLD_PROCESS_NUMA_MASK_BITS /
+				    (sizeof(unsigned long) * 8)];
+
 	int thp_disable;
 	int uti_thread_rank; /* N-th clone() spawns a thread on Linux CPU */
 	int uti_use_last_cpu; /* Work-around not to share CPU with OpenMP thread */
+	int straight_map;
+	size_t straight_map_threshold;
+#ifdef ENABLE_TOFU
+	int enable_tofu;
+#endif
 	int nr_processes;
 	int process_rank;
 	__cpu_set_unit cpu_set[PLD_CPU_SET_SIZE];
@@ -193,6 +217,9 @@ struct syscall_response {
 	unsigned long req_thread_status;
 	long ret;
 	unsigned long fault_address;
+#ifdef ENABLE_TOFU
+	void *pde_data;
+#endif
 };
 
 struct syscall_ret_desc {

executer/kernel/mcctrl/arch/arm64/archdeps.c

@@ -2,11 +2,16 @@
 #include <linux/version.h>
 #include <linux/mm_types.h>
 #include <linux/kallsyms.h>
+#include <linux/delay.h>
 #if KERNEL_VERSION(4, 11, 0) <= LINUX_VERSION_CODE
 #include <linux/sched/task_stack.h>
 #endif /* LINUX_VERSION_CODE >= KERNEL_VERSION(4, 11, 0) */
 #include <linux/ptrace.h>
 #include <linux/uaccess.h>
+#include <linux/mmu_notifier.h>
+#include <linux/kref.h>
+#include <linux/file.h>
+#include <linux/proc_fs.h>
 #include <asm/vdso.h>
 #include "config.h"
 #include "../../mcctrl.h"
@@ -27,6 +32,39 @@ void *vdso_end;
 static struct vm_special_mapping (*vdso_spec)[2];
 #endif
 
+#ifdef ENABLE_TOFU
+/* Tofu CQ and barrier gate release functions */
+struct file_operations *mcctrl_tof_utofu_procfs_ops_cq;
+int (*mcctrl_tof_utofu_release_cq)(struct inode *inode,
+		struct file *filp);
+struct file_operations *mcctrl_tof_utofu_procfs_ops_bch;
+int (*mcctrl_tof_utofu_release_bch)(struct inode *inode,
+		struct file *filp);
+int (*mcctrl_tof_core_cq_cacheflush)(int tni, int cqid);
+int (*mcctrl_tof_core_disable_bch)(int tni, int bgid);
+int (*mcctrl_tof_core_unset_bg)(int tni, int bgid);
+typedef void (*tof_core_signal_handler)(int, int, uint64_t, uint64_t);
+void (*mcctrl_tof_core_register_signal_bg)(int tni, int bgid,
+			tof_core_signal_handler handler);
+struct tof_utofu_bg;
+struct tof_utofu_bg *mcctrl_tof_utofu_bg;
+
+
+/* Tofu MMU notifier */
+struct mmu_notifier_ops *mcctrl_tof_utofu_mn_ops;
+struct mmu_notifier_ops __mcctrl_tof_utofu_mn_ops;
+static void (*mcctrl_tof_utofu_mn_invalidate_range_end)(
+	struct mmu_notifier *mn,
+	struct mm_struct *mm,
+	unsigned long start,
+	unsigned long end);
+void __mcctrl_tof_utofu_mn_invalidate_range_end(
+	struct mmu_notifier *mn,
+	struct mm_struct *mm,
+	unsigned long start,
+	unsigned long end);
+#endif
+
 int arch_symbols_init(void)
 {
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 0, 0)
@@ -43,6 +81,71 @@ int arch_symbols_init(void)
 		return -EFAULT;
 #endif
 
+#ifdef ENABLE_TOFU
+	mcctrl_tof_utofu_procfs_ops_cq =
+		(void *)kallsyms_lookup_name("tof_utofu_procfs_ops_cq");
+	if (WARN_ON(!mcctrl_tof_utofu_procfs_ops_cq))
+		return -EFAULT;
+
+	mcctrl_tof_utofu_procfs_ops_bch =
+		(void *)kallsyms_lookup_name("tof_utofu_procfs_ops_bch");
+	if (WARN_ON(!mcctrl_tof_utofu_procfs_ops_bch))
+		return -EFAULT;
+
+	mcctrl_tof_utofu_release_cq =
+		(void *)kallsyms_lookup_name("tof_utofu_release_cq");
+	if (WARN_ON(!mcctrl_tof_utofu_release_cq))
+		return -EFAULT;
+
+	mcctrl_tof_utofu_release_bch =
+		(void *)kallsyms_lookup_name("tof_utofu_release_bch");
+	if (WARN_ON(!mcctrl_tof_utofu_release_bch))
+		return -EFAULT;
+
+	mcctrl_tof_core_cq_cacheflush =
+		(void *)kallsyms_lookup_name("tof_core_cq_cacheflush");
+	if (WARN_ON(!mcctrl_tof_core_cq_cacheflush))
+		return -EFAULT;
+
+	mcctrl_tof_core_disable_bch =
+		(void *)kallsyms_lookup_name("tof_core_disable_bch");
+	if (WARN_ON(!mcctrl_tof_core_disable_bch))
+		return -EFAULT;
+
+	mcctrl_tof_core_unset_bg =
+		(void *)kallsyms_lookup_name("tof_core_unset_bg");
+	if (WARN_ON(!mcctrl_tof_core_unset_bg))
+		return -EFAULT;
+
+	mcctrl_tof_core_register_signal_bg =
+		(void *)kallsyms_lookup_name("tof_core_register_signal_bg");
+	if (WARN_ON(!mcctrl_tof_core_register_signal_bg))
+		return -EFAULT;
+
+	mcctrl_tof_utofu_bg =
+		(void *)kallsyms_lookup_name("tof_utofu_bg");
+	if (WARN_ON(!mcctrl_tof_utofu_bg))
+		return -EFAULT;
+
+	mcctrl_tof_utofu_mn_ops =
+		(void *)kallsyms_lookup_name("tof_utofu_mn_ops");
+	if (WARN_ON(!mcctrl_tof_utofu_mn_ops))
+		return -EFAULT;
+	/*
+	 * Copy original content and update redirected function,
+	 * CQ will be pointed to this structure after init ioctl()
+	 */
+	memcpy(&__mcctrl_tof_utofu_mn_ops, mcctrl_tof_utofu_mn_ops,
+			sizeof(*mcctrl_tof_utofu_mn_ops));
+	__mcctrl_tof_utofu_mn_ops.invalidate_range =
+		__mcctrl_tof_utofu_mn_invalidate_range_end;
+
+	mcctrl_tof_utofu_mn_invalidate_range_end =
+		(void *)kallsyms_lookup_name("tof_utofu_mn_invalidate_range_end");
+	if (WARN_ON(!mcctrl_tof_utofu_mn_invalidate_range_end))
+		return -EFAULT;
+#endif
+
 	return 0;
 }
 
@@ -331,6 +434,15 @@ int translate_rva_to_rpa(ihk_os_t os, unsigned long rpt, unsigned long rva,
 
 	// page table to translation_table.
 	phys = ihk_device_map_memory(ihk_os_to_dev(os), rpt, PAGE_SIZE);
+#ifdef ENABLE_FUGAKU_HACKS
+	if (!phys) {
+		pr_err("%s(): ERROR: VA: 0x%lx, rpt is NULL for PID %d\n",
+			__func__, rva, task_tgid_vnr(current));
+		error = -EFAULT;
+		goto out;
+	}
+#endif
+
 	tbl = ihk_device_map_virtual(ihk_os_to_dev(os), phys, PAGE_SIZE, NULL, 0);
 	rpa = (unsigned long)tbl->tt_pa;
 
@@ -417,3 +529,488 @@ long arch_switch_ctx(struct uti_switch_ctx_desc *desc)
 out:
 	return rc;
 }
+
+
+#ifdef ENABLE_TOFU
+/*
+ * Tofu CQ and BCH release handlers
+ */
+int __mcctrl_tof_utofu_release_cq(struct inode *inode, struct file *filp);
+int __mcctrl_tof_utofu_release_bch(struct inode *inode, struct file *filp);
+
+void mcctrl_tofu_hijack_release_handlers(void)
+{
+	mcctrl_tof_utofu_procfs_ops_cq->release =
+		__mcctrl_tof_utofu_release_cq;
+	mcctrl_tof_utofu_procfs_ops_bch->release =
+		__mcctrl_tof_utofu_release_bch;
+	wmb();
+}
+
+void mcctrl_tofu_restore_release_handlers(void)
+{
+	mcctrl_tof_utofu_procfs_ops_cq->release =
+		mcctrl_tof_utofu_release_cq;
+	mcctrl_tof_utofu_procfs_ops_bch->release =
+		mcctrl_tof_utofu_release_bch;
+	wmb();
+}
+
+/*
+ * Tofu cleanup functions
+ */
+#include <tofu/tof_uapi.h>
+#include <tofu/tof_icc.h>
+#include <tofu/tofu_generated-tof_core_cq.h>
+#include <tofu/tofu_generated-tof_utofu_device.h>
+#include <tofu/tofu_generated-tof_utofu_cq.h>
+#include <tofu/tofu_generated-tof_utofu_mbpt.h>
+#include <tofu/tofu_generated-tof_utofu_bg.h>
+
+#define TOF_UTOFU_VERSION TOF_UAPI_VERSION
+#define TOF_UTOFU_NUM_STAG_NTYPES 3
+#define TOF_UTOFU_NUM_STAG_BITS(size) ((size) + 13)
+#define TOF_UTOFU_NUM_STAG(size) ((uint64_t)1 << TOF_UTOFU_NUM_STAG_BITS(size))
+#define TOF_UTOFU_STAG_TRANS_BITS 3
+#define TOF_UTOFU_STAG_TRANS_SIZE ((uint64_t)1 << TOF_UTOFU_STAG_TRANS_BITS)
+#define TOF_UTOFU_STAG_TRANS_TABLE_LEN(size) (TOF_UTOFU_NUM_STAG(size) * TOF_UTOFU_STAG_TRANS_SIZE)
+#define TOF_UTOFU_STEERING_TABLE_LEN(size) (TOF_UTOFU_NUM_STAG(size) * TOF_ICC_STEERING_SIZE)
+#define TOF_UTOFU_MB_TABLE_LEN(size) (TOF_UTOFU_NUM_STAG(size) * TOF_ICC_MB_SIZE)
+#define TOF_UTOFU_STAG_MEM_LEN(size) (TOF_UTOFU_STEERING_TABLE_LEN(size) * 4)
+#define TOF_UTOFU_SPECIAL_STAG 4096
+
+#define TOF_UTOFU_ICC_COMMON_REGISTER (tof_icc_reg_pa + 0x0B000000)
+#define TOF_UTOFU_REG_START tof_icc_reg_pa
+#define TOF_UTOFU_REG_END (TOF_UTOFU_ICC_COMMON_REGISTER + 0x000FFFFF)
+
+#define TOF_UTOFU_SET_SUBNET_TNI	0	/* This number is kernel TNIs number in setting subnet */
+#define TOF_UTOFU_KCQ			11
+#define TOF_UTOFU_LINKDOWN_PORT_MASK	0x000003FF
+
+#define TOF_UTOFU_ALLOC_STAG_LPG	0x2
+#define TOF_UTOFU_BLANK_MBVA (-1)
+
+#define TOF_UTOFU_MRU_EMPTY (-1)
+
+struct tof_utofu_trans_list {
+	int16_t prev;
+	int16_t next;
+	uint8_t pgszbits;
+	struct tof_utofu_mbpt *mbpt;
+};
+
+
+/*
+ * Bit 30 marks a kref as McKernel internal.
+ * This can be used to distinguish krefs from Linux and
+ * it also ensures that a non deallocated kref will not
+ * crash the Linux allocator.
+ */
+#define MCKERNEL_KREF_MARK	(1U << 30)
+static inline unsigned int mcctrl_kref_is_mckernel(const struct kref *kref)
+{
+	return (refcount_read(&kref->refcount) & (MCKERNEL_KREF_MARK));
+}
+
+/**
+ * kref_put - decrement refcount for object.
+ * @kref: object.
+ * @release: pointer to the function that will clean up the object when the
+ *	     last reference to the object is released.
+ *	     This pointer is required, and it is not acceptable to pass kfree
+ *	     in as this function.  If the caller does pass kfree to this
+ *	     function, you will be publicly mocked mercilessly by the kref
+ *	     maintainer, and anyone else who happens to notice it.  You have
+ *	     been warned.
+ *
+ * Decrement the refcount, and if 0, call release().
+ * Return 1 if the object was removed, otherwise return 0.  Beware, if this
+ * function returns 0, you still can not count on the kref from remaining in
+ * memory.  Only use the return value if you want to see if the kref is now
+ * gone, not present.
+ */
+static inline int mcctrl_kref_put(struct kref *kref, void (*release)(struct kref *kref))
+{
+	if (atomic_dec_return(&kref->refcount.refs) == MCKERNEL_KREF_MARK) {
+		release(kref);
+		return 1;
+	}
+	return 0;
+}
+
+static int tof_utofu_cq_cacheflush(struct tof_utofu_cq *ucq){
+	return mcctrl_tof_core_cq_cacheflush(ucq->tni, ucq->cqid);
+}
+
+static void tof_utofu_trans_mru_delete(struct tof_utofu_cq *ucq, int stag){
+	struct tof_utofu_trans_list *mru = ucq->trans.mru;
+	int prev = mru[stag].prev;
+	int next = mru[stag].next;
+	if(prev == TOF_UTOFU_MRU_EMPTY || next == TOF_UTOFU_MRU_EMPTY){ /* already deleted */
+		return;
+	}
+	if(prev == stag){  /* a single entry */
+		ucq->trans.mruhead = TOF_UTOFU_MRU_EMPTY;
+	}else{
+		if(ucq->trans.mruhead == stag){
+			ucq->trans.mruhead = next;
+		}
+		mru[prev].next = next;
+		mru[next].prev = prev;
+	}
+	mru[stag].prev = TOF_UTOFU_MRU_EMPTY;
+	mru[stag].next = TOF_UTOFU_MRU_EMPTY;
+}
+
+static void tof_utofu_trans_disable(struct tof_utofu_cq *ucq, int stag){
+	struct tof_trans_table *table = ucq->trans.table;
+	atomic64_set((atomic64_t *)&table[stag], 0);
+	tof_utofu_trans_mru_delete(ucq, stag);
+}
+
+/* McKernel scatterlist is simply a contiguous buffer. */
+struct scatterlist {
+	void *pages;
+	unsigned int	offset;
+	unsigned int	length;
+	unsigned long	dma_address;
+	unsigned int	dma_length;
+};
+
+static uintptr_t tof_utofu_disable_mbpt(struct tof_utofu_mbpt *mbpt, int idx){
+	int i0, i1;
+	struct tof_icc_mbpt_entry *ent;
+	uintptr_t ipa;
+	i0 = idx / (PAGE_SIZE / TOF_ICC_MBPT_SIZE);
+	i1 = idx - i0 * (PAGE_SIZE / TOF_ICC_MBPT_SIZE);
+	//ent = sg_virt(&mbpt->sg[i0]);
+	ent = mbpt->sg->pages + (i0 * PAGE_SIZE);
+	if(!ent[i1].enable){
+		return 0;
+	}
+	ent[i1].enable = 0;
+	ipa = (uint64_t)ent[i1].ipa << 12;
+	ent[i1].ipa = 0;
+	return ipa;
+}
+
+static void tof_utofu_free_mbpt(struct tof_utofu_cq *ucq, struct tof_utofu_mbpt *mbpt){
+	int i;
+
+	for(i = 0; i < mbpt->nsgents * PAGE_SIZE / sizeof(struct tof_icc_mbpt_entry); i++){
+		uintptr_t iova;
+		iova = tof_utofu_disable_mbpt(mbpt, i);
+#if 0
+		/*
+		 * NOTE: Not performed for McKernel managed stags.
+		 */
+		if(iova){
+			tof_smmu_release_ipa_cq(ucq->tni, ucq->cqid, iova, mbpt->pgsz);
+		}
+#endif
+	}
+
+#if 0
+	/*
+	 * NOTE: Everyhing below has been allocated in McKernel, do nothing here!!
+	 * This leaks memory in McKernel, but it doesn't crash Linux.
+	 * Memory will be released once McKernel is unbooted.
+	 */
+	tof_smmu_iova_unmap_sg(ucq->tni, ucq->cqid, mbpt->sg, mbpt->nsgents);
+
+	for(i = 0; i < mbpt->nsgents; i++){
+		tof_util_free_pages((unsigned long)sg_virt(&mbpt->sg[i]), 0);
+	}
+	tof_util_free(mbpt->sg);
+	tof_util_free(mbpt);
+#endif
+}
+
+static void tof_utofu_mbpt_release(struct kref *kref)
+{
+	struct tof_utofu_mbpt *mbpt = container_of(kref, struct tof_utofu_mbpt, kref);
+	//atomic64_inc((atomic64_t *)&kref_free_count);
+	tof_utofu_free_mbpt(mbpt->ucq, mbpt);
+}
+
+static int tof_utofu_free_stag(struct tof_utofu_cq *ucq, int stag){
+	if(stag < 0 || stag >= TOF_UTOFU_NUM_STAG(ucq->num_stag) ||
+	   ucq->steering == NULL){
+		return -EINVAL;
+	}
+	if(!(ucq->steering[stag].enable)){
+		return -ENOENT;
+	}
+	if (!mcctrl_kref_is_mckernel(&ucq->trans.mru[stag].mbpt->kref)) {
+		printk("%s: stag: %d is not an McKernel kref\n", __func__, stag);
+		return -EINVAL;
+	}
+	ucq->steering[stag].enable = 0;
+	ucq->mb[stag].enable = 0;
+	tof_utofu_trans_disable(ucq, stag);
+	dma_wmb();
+	tof_utofu_cq_cacheflush(ucq);
+	mcctrl_kref_put(&ucq->trans.mru[stag].mbpt->kref, tof_utofu_mbpt_release);
+	ucq->trans.mru[stag].mbpt = NULL;
+	dprintk("%s: TNI: %d, CQ: %d: stag %d deallocated\n",
+			__func__, ucq->tni, ucq->cqid, stag);
+	return 0;
+}
+
+void mcctrl_mckernel_tof_utofu_release_cq(void *pde_data)
+{
+	struct tof_utofu_cq *ucq;
+	struct tof_utofu_device *dev;
+	unsigned long irqflags;
+	int stag;
+
+	dev = (struct tof_utofu_device *)pde_data;
+	ucq = container_of(dev, struct tof_utofu_cq, common);
+
+	if (!ucq->common.enabled) {
+		return;
+	}
+
+	dprintk("%s: UCQ (PDE: 0x%lx) TNI %d CQ %d\n",
+		__func__, (unsigned long)pde_data, ucq->tni, ucq->cqid);
+
+	/*
+	 * Only release stags here, actual cleanup is still performed
+	 * in the Tofu driver
+	 */
+	for (stag = 0; stag < TOF_UTOFU_NUM_STAG(ucq->num_stag); stag++) {
+		spin_lock_irqsave(&ucq->trans.mru_lock, irqflags);
+		tof_utofu_free_stag(ucq, stag);
+		spin_unlock_irqrestore(&ucq->trans.mru_lock, irqflags);
+	}
+}
+
+static inline void tof_core_unregister_signal_bg(int tni, int bgid)
+{
+	return mcctrl_tof_core_register_signal_bg(tni, bgid, NULL);
+}
+
+static struct tof_utofu_bg *tof_utofu_bg_get(int tni, int bgid){
+	if((unsigned int)tni >= TOF_ICC_NTNIS ||
+	   (unsigned int)bgid >= TOF_ICC_NBGS){
+		return NULL;
+	}
+	//return &tof_utofu_bg[tni][bgid];
+
+	// Convert [][] notion into pointer aritmethic
+	return mcctrl_tof_utofu_bg + (tni * TOF_ICC_NBGS) + bgid;
+}
+
+static int __tof_utofu_unset_bg(struct tof_utofu_bg *ubg){
+	if(ubg->common.enabled){
+		mcctrl_tof_core_unset_bg(ubg->tni, ubg->bgid);
+		ubg->common.enabled = false;
+		tof_core_unregister_signal_bg(ubg->tni, ubg->bgid);
+	}
+	return 0;
+}
+
+static int mcctrl_tof_utofu_disable_bch(struct tof_utofu_bg *ubg){
+	int ret;
+	int tni, bgid;
+
+	if(!ubg->bch.enabled){
+		return -EPERM;
+	}
+
+	ret = mcctrl_tof_core_disable_bch(ubg->tni, ubg->bgid);
+	if(ret < 0){
+		return ret;
+	}
+
+	for(tni = 0; tni < TOF_ICC_NTNIS; tni++){
+		uint64_t mask = ubg->bch.bgmask[tni];
+		for(bgid = 0; bgid < TOF_ICC_NBGS; bgid++){
+			if((mask >> bgid) & 1){
+				ret = __tof_utofu_unset_bg(tof_utofu_bg_get(tni, bgid));
+				if(ret < 0){
+					/* OK? */
+					//BUG();
+					return ret;
+				}
+			}
+		}
+	}
+
+	/* Not performed in McKernel handler */
+	//tof_smmu_release_ipa_bg(ubg->tni, ubg->bgid, ubg->bch.iova, TOF_ICC_BCH_DMA_ALIGN);
+	//put_page(ubg->bch.page);
+	ubg->bch.enabled = false;
+	smp_mb();
+	dprintk("%s: tni=%d bgid=%d\n", __func__, ubg->tni, ubg->bgid);
+	return 0;
+}
+
+void mcctrl_mckernel_tof_utofu_release_bch(void *pde_data)
+{
+	struct tof_utofu_bg *ubg;
+	struct tof_utofu_device *dev = (struct tof_utofu_device *)pde_data;
+
+	ubg = container_of(dev, struct tof_utofu_bg, common);
+	//tof_log_if("tni=%d bgid=%d\n", ubg->tni, ubg->bgid);
+	dprintk("%s: tni=%d bgid=%d\n", __func__, ubg->tni, ubg->bgid);
+	mcctrl_tof_utofu_disable_bch(ubg);
+}
+
+void mcctrl_tofu_cleanup_file(struct mcctrl_file_to_pidfd *f2pfd)
+{
+	/* Figure out whether CQ or BCH */
+	if (strstr(f2pfd->tofu_dev_path, "cq")) {
+		dprintk("%s: PID: %d, fd: %d (%s) -> release CQ\n",
+			__func__, f2pfd->pid, f2pfd->fd, f2pfd->tofu_dev_path);
+		mcctrl_mckernel_tof_utofu_release_cq(f2pfd->pde_data);
+	}
+
+	else if (strstr(f2pfd->tofu_dev_path, "bch")) {
+		dprintk("%s: PID: %d, fd: %d (%s) -> release BCH\n",
+			__func__, f2pfd->pid, f2pfd->fd, f2pfd->tofu_dev_path);
+		mcctrl_mckernel_tof_utofu_release_bch(f2pfd->pde_data);
+	}
+}
+
+
+int __mcctrl_tof_utofu_release_handler(struct inode *inode, struct file *filp,
+		int (*__release_func)(struct inode *inode, struct file *filp))
+{
+	struct mcctrl_usrdata *usrdata;
+	struct mcctrl_file_to_pidfd *f2pfd;
+	struct mcctrl_per_proc_data *ppd;
+	struct ikc_scd_packet isp;
+	int ret;
+
+	dprintk("%s: current PID: %d, comm: %s \n",
+			__func__, task_tgid_vnr(current), current->comm);
+
+	f2pfd = mcctrl_file_to_pidfd_hash_lookup(filp, current->group_leader);
+	if (!f2pfd) {
+		goto out;
+	}
+
+	dprintk("%s: current PID: %d, PID: %d, fd: %d ...\n",
+			__func__, task_tgid_vnr(current), f2pfd->pid, f2pfd->fd);
+	usrdata = ihk_host_os_get_usrdata(f2pfd->os);
+
+	/* Look up per-process structure */
+	ppd = mcctrl_get_per_proc_data(usrdata, f2pfd->pid);
+	if (!ppd) {
+		pr_err("%s: PID: %d, fd: %d no PPD\n",
+				__func__, f2pfd->pid, f2pfd->fd);
+		goto out;
+	}
+
+	dprintk("%s: PID: %d, fd: %d PPD OK\n",
+			__func__, f2pfd->pid, f2pfd->fd);
+
+	/*
+	 * We are in release() due to the process being killed,
+	 * or because the application didn't close the file properly.
+	 * Ask McKernel to clean up this fd.
+	 */
+	isp.msg = SCD_MSG_CLEANUP_FD;
+	isp.pid = f2pfd->pid;
+	isp.arg = f2pfd->fd;
+
+	ret = mcctrl_ikc_send_wait(f2pfd->os, ppd->ikc_target_cpu,
+			&isp, -20, NULL, NULL, 0);
+	if (ret != 0) {
+		pr_err("%s: WARNING: IKC req for PID: %d, fd: %d failed\n",
+				__func__, f2pfd->pid, f2pfd->fd);
+	}
+
+	/* Disable any remaining STAGs/BCH in mcctrl anyway */
+	mcctrl_tofu_cleanup_file(f2pfd);
+
+	mcctrl_file_to_pidfd_hash_remove(filp, f2pfd->os,
+			current->group_leader, f2pfd->fd);
+
+	mcctrl_put_per_proc_data(ppd);
+
+out:
+	dprintk("%s: current PID: %d, comm: %s -> calling release\n",
+			__func__, task_tgid_vnr(current), current->comm);
+	return __release_func(inode, filp);
+}
+
+int __mcctrl_tof_utofu_release_cq(struct inode *inode, struct file *filp)
+{
+	return __mcctrl_tof_utofu_release_handler(inode, filp,
+			mcctrl_tof_utofu_release_cq);
+}
+
+int __mcctrl_tof_utofu_release_bch(struct inode *inode, struct file *filp)
+{
+	return __mcctrl_tof_utofu_release_handler(inode, filp,
+			mcctrl_tof_utofu_release_bch);
+}
+
+/*
+ * Tofu MMU notifier functions
+ */
+void __mcctrl_tof_utofu_mn_invalidate_range_end(
+	struct mmu_notifier *mn,
+	struct mm_struct *mm,
+	unsigned long start,
+	unsigned long end)
+{
+	char tmpname[TASK_COMM_LEN];
+
+	/* Not an offloaded syscall? */
+	if (current->mm != mm) {
+		goto out_call_real;
+	}
+
+	/* Not mcexec? Just in case.. */
+	get_task_comm(tmpname, current);
+	if (strncmp(tmpname, "mcexec", TASK_COMM_LEN)) {
+		goto out_call_real;
+	}
+
+	/* This is only called for Tofu enabled mcexec processes */
+	dprintk("%s: skipping tof_utofu_mn_invalidate_range_end() "
+			"for mcexec PID %d\n",
+			__func__, task_tgid_vnr(current));
+	return;
+
+out_call_real:
+	return mcctrl_tof_utofu_mn_invalidate_range_end(mn, mm, start, end);
+}
+
+int __mcctrl_tof_utofu_ioctl_init_cq(struct tof_utofu_device *dev,
+		unsigned long arg) {
+	struct tof_utofu_cq *ucq;
+
+	ucq = container_of(dev, struct tof_utofu_cq, common);
+	if (!ucq->common.enabled) {
+		return -EINVAL;
+	}
+
+	dprintk("%s: Tofu TNI %d CQ %d (PDE: 0x%lx) MMU notifier to be hijacked\n",
+		__func__, ucq->tni, ucq->cqid, (unsigned long)dev);
+	/* Override the MMU notifier */
+	ucq->mn.ops = &__mcctrl_tof_utofu_mn_ops;
+
+	return 0;
+}
+
+long __mcctrl_tof_utofu_unlocked_ioctl_cq(void *pde_data, unsigned int cmd,
+					unsigned long arg) {
+	struct tof_utofu_device *dev = (struct tof_utofu_device *)pde_data;
+	int ret;
+
+	switch (cmd) {
+		/* We only care about init, where we hijack the MMU notifier */
+		case TOF_IOCTL_INIT_CQ:
+			ret = __mcctrl_tof_utofu_ioctl_init_cq(dev, arg);
+			break;
+		default:
+			ret = 0;
+	}
+
+	return ret;
+}
+#endif

executer/kernel/mcctrl/arch/arm64/include/tofu/generate_headers.sh

@@ -0,0 +1,41 @@
+#!/bin/bash
+
+SCRIPT="`readlink -f ${BASH_SOURCE[0]:-}`"
+SCRIPT_DIR=$(dirname ${SCRIPT})
+CURRENT_DIR=`pwd`
+
+cd ${SCRIPT_DIR}
+
+DWARF_TOOL=${SCRIPT_DIR}/../../../../../../../tools/dwarf-extract-struct/dwarf-extract-struct
+if [ ! -x ${DWARF_TOOL} ]; then
+	echo "error: couldn't find DWARF extractor executable (${DWARF_TOOL}), have you compiled it?"
+	cd -
+	exit 1
+fi
+
+echo "Looking for Tofu driver debug symbols..."
+if [ "`find /lib/modules/ -name "tof_module.tar.gz" | xargs -r ls -t | head -n 1 | wc -l`" == "0" ]; then
+	echo "error: couldn't find Tofu modules with debug symbols"
+	cd -
+	exit 1
+fi
+
+MODULE_TAR_GZ=`find /lib/modules/ -name "tof_module.tar.gz" | xargs ls -t | head -n 1`
+echo "Using Tofu driver debug symbols: ${MODULE_TAR_GZ}"
+
+KMODULE=tof_utofu.ko
+if ! tar zxvf ${MODULE_TAR_GZ} ${KMODULE} 2>&1 > /dev/null; then
+	echo "error: uncompressing kernel module with debug symbols"
+	cd -
+	exit 1
+fi
+
+${DWARF_TOOL} ${KMODULE} tof_utofu_device enabled subnet gpid > tofu_generated-tof_utofu_device.h
+${DWARF_TOOL} ${KMODULE} tof_utofu_cq common tni cqid mn trans steering mb num_stag | sed "s/struct FILL_IN_MANUALLY trans;/#include \"tof_utofu_cq_trans.h\"/g" > tofu_generated-tof_utofu_cq.h
+${DWARF_TOOL} ${KMODULE} tof_utofu_mbpt ucq iova sg nsgents mbptstart pgsz kref > tofu_generated-tof_utofu_mbpt.h
+${DWARF_TOOL} ${KMODULE} tof_utofu_bg common tni bgid bch | sed "s/struct FILL_IN_MANUALLY bch;/#include \"tof_utofu_bg_bch.h\"/g" > tofu_generated-tof_utofu_bg.h
+rm ${KMODULE}
+
+
+#cat tofu_generated*.h
+cd - > /dev/null

executer/kernel/mcctrl/arch/arm64/include/tofu/tof_icc.h

@@ -0,0 +1,831 @@
+#ifndef _TOF_ICC_H_
+#define _TOF_ICC_H_
+
+#include <linux/types.h>
+#ifdef __KERNEL__
+#include <linux/bitops.h>
+#else
+#include <stdint.h>
+typedef uint64_t phys_addr_t;
+#endif
+
+/* constants related to the Tofu Interconnect D */
+
+#define TOF_ICC_NTNIS 6
+#define TOF_ICC_NCQS 12
+#define TOF_ICC_NBGS 48
+#define TOF_ICC_NBCHS 16
+#define TOF_ICC_NPORTS 10
+#define TOF_ICC_NVMSIDS 16
+
+#define TOF_ICC_RH_LEN 8
+#define TOF_ICC_ECRC_LEN 4
+#define TOF_ICC_FRAME_ALIGN 32
+#define TOF_ICC_TLP_LEN(len) (((len) + 1) * TOF_ICC_FRAME_ALIGN)
+#define TOF_ICC_TLP_PAYLOAD_MAX (TOF_ICC_TLP_LEN(61) - TOF_ICC_ECRC_LEN)
+#define TOF_ICC_FRAME_LEN(len) (TOF_ICC_RH_LEN + TOF_ICC_TLP_LEN(len))
+#define TOF_ICC_FRAME_LEN_MIN TOF_ICC_FRAME_LEN(2)
+#define TOF_ICC_FRAME_LEN_MAX TOF_ICC_FRAME_LEN(61)
+#define TOF_ICC_FRAME_BUF_SIZE_BITS 11
+#define TOF_ICC_FRAME_BUF_SIZE (1 << TOF_ICC_FRAME_BUF_SIZE_BITS)
+#define TOF_ICC_FRAME_BUF_ALIGN_BITS 8
+#define TOF_ICC_FRAME_BUF_ALIGN (1 << TOF_ICC_FRAME_BUF_ALIGN_BITS)
+#define TOF_ICC_PB_SIZE_BITS 11
+#define TOF_ICC_PB_SIZE (1 << TOF_ICC_PB_SIZE_BITS)
+#define TOF_ICC_PB_ALIGN_BITS 11
+#define TOF_ICC_PB_ALIGN (1 << TOF_ICC_PB_ALIGN_BITS)
+
+#define TOF_ICC_ST_ALIGN_BITS 8
+#define TOF_ICC_ST_ALIGN (1 << TOF_ICC_ST_ALIGN_BITS)
+
+#define TOF_ICC_MBT_ALIGN_BITS 8
+#define TOF_ICC_MBT_ALIGN (1 << TOF_ICC_MBT_ALIGN_BITS)
+
+#define TOF_ICC_MBPT_ALIGN_BITS 8
+#define TOF_ICC_MBPT_ALIGN (1 << TOF_ICC_MBPT_ALIGN_BITS)
+
+#define TOF_ICC_BG_BSEQ_SIZE_BITS 24
+#define TOF_ICC_BG_BSEQ_SIZE (1 << TOF_ICC_BG_BSEQ_SIZE_BITS)
+
+#define TOF_ICC_BCH_DMA_ALIGN_BITS 8
+#define TOF_ICC_BCH_DMA_ALIGN (1 << TOF_ICC_BCH_DMA_ALIGN_BITS)
+
+/* this is a CPU-specific constant, but referred in the ICC spec. */
+#define TOF_ICC_CACHE_LINE_SIZE_BITS 8
+#define TOF_ICC_CACHE_LINE_SIZE (1 << TOF_ICC_CACHE_LINE_SIZE_BITS)
+
+#define TOF_ICC_TOQ_DESC_SIZE_BITS 5
+#define TOF_ICC_TOQ_DESC_SIZE (1 << TOF_ICC_TOQ_DESC_SIZE_BITS)
+#define TOF_ICC_TCQ_DESC_SIZE_BITS 3
+#define TOF_ICC_TCQ_DESC_SIZE (1 << TOF_ICC_TCQ_DESC_SIZE_BITS)
+#define TOF_ICC_TCQ_NLINE_BITS (TOF_ICC_CACHE_LINE_SIZE_BITS - TOF_ICC_TCQ_DESC_SIZE_BITS)
+#define TOF_ICC_MRQ_DESC_SIZE_BITS 5
+#define TOF_ICC_MRQ_DESC_SIZE (1 << TOF_ICC_MRQ_DESC_SIZE_BITS)
+#define TOF_ICC_PBQ_DESC_SIZE_BITS 3
+#define TOF_ICC_PBQ_DESC_SIZE (1 << TOF_ICC_PBQ_DESC_SIZE_BITS)
+#define TOF_ICC_PRQ_DESC_SIZE_BITS 3
+#define TOF_ICC_PRQ_DESC_SIZE (1 << TOF_ICC_PRQ_DESC_SIZE_BITS)
+#define TOF_ICC_PRQ_NLINE_BITS (TOF_ICC_CACHE_LINE_SIZE_BITS - TOF_ICC_PBQ_DESC_SIZE_BITS)
+
+#define TOF_ICC_TOQ_SIZE_NTYPES 6
+#define TOF_ICC_TOQ_SIZE_BITS(size) ((size) * 2 + 11)
+#define TOF_ICC_TOQ_SIZE(size) (1 << TOF_ICC_TOQ_SIZE_BITS(size))
+#define TOF_ICC_TOQ_LEN(size) (TOF_ICC_TOQ_SIZE(size) * TOF_ICC_TOQ_DESC_SIZE)
+#define TOF_ICC_TCQ_LEN(size) (TOF_ICC_TOQ_SIZE(size) * TOF_ICC_TCQ_DESC_SIZE)
+
+#define TOF_ICC_MRQ_SIZE_NTYPES 6
+#define TOF_ICC_MRQ_SIZE_BITS(size) ((size) * 2 + 11)
+#define TOF_ICC_MRQ_SIZE(size) (1 << TOF_ICC_MRQ_SIZE_BITS(size))
+#define TOF_ICC_MRQ_LEN(size) (TOF_ICC_MRQ_SIZE(size) * TOF_ICC_MRQ_DESC_SIZE)
+
+#define TOF_ICC_PBQ_SIZE_NTYPES 6
+#define TOF_ICC_PBQ_SIZE_BITS(size) ((size) * 2 + 11)
+#define TOF_ICC_PBQ_SIZE(size) (1 << TOF_ICC_PBQ_SIZE_BITS(size))
+#define TOF_ICC_PBQ_LEN(size) (TOF_ICC_PBQ_SIZE(size) * TOF_ICC_PBQ_DESC_SIZE)
+
+#define TOF_ICC_PRQ_SIZE_NTYPES 6
+#define TOF_ICC_PRQ_SIZE_BITS(size) ((size) * 2 + 11)
+#define TOF_ICC_PRQ_SIZE(size) (1 << TOF_ICC_PRQ_SIZE_BITS(size))
+#define TOF_ICC_PRQ_LEN(size) (TOF_ICC_PRQ_SIZE(size) * TOF_ICC_PRQ_DESC_SIZE)
+
+#define TOF_ICC_STEERING_TABLE_ALIGN_BITS 8
+#define TOF_ICC_STEERING_TABLE_ALIGN (1 << TOF_ICC_STEERING_TABLE_ALIGN_BITS)
+#define TOF_ICC_STEERING_SIZE_BITS 4
+#define TOF_ICC_STEERING_SIZE (1 << TOF_ICC_STEERING_SIZE_BITS)
+
+#define TOF_ICC_MB_TABLE_ALIGN_BITS 8
+#define TOF_ICC_MB_TABLE_ALIGN (1 << TOF_ICC_MB_TABLE_ALIGN_BITS)
+#define TOF_ICC_MB_SIZE_BITS 4
+#define TOF_ICC_MB_SIZE (1 << TOF_ICC_MB_SIZE_BITS)
+#define TOF_ICC_MB_PS_ENCODE(bits) ((bits) % 9 == 3 ? (bits) / 9 - 1 : (bits) / 13 + 3)
+
+#define TOF_ICC_MBPT_ALIGN_BITS 8
+#define TOF_ICC_MBPT_ALIGN (1 << TOF_ICC_MBPT_ALIGN_BITS)
+#define TOF_ICC_MBPT_SIZE_BITS 3
+#define TOF_ICC_MBPT_SIZE (1 << TOF_ICC_MBPT_SIZE_BITS)
+
+#define TOF_ICC_X_BITS 5
+#define TOF_ICC_Y_BITS 5
+#define TOF_ICC_Z_BITS 5
+#define TOF_ICC_A_BITS 1
+#define TOF_ICC_B_BITS 2
+#define TOF_ICC_C_BITS 1
+#define TOF_ICC_MAX_X_SIZE (1 << TOF_ICC_X_BITS)
+#define TOF_ICC_MAX_Y_SIZE (1 << TOF_ICC_Y_BITS)
+#define TOF_ICC_MAX_Z_SIZE (1 << TOF_ICC_Z_BITS)
+#define TOF_ICC_A_SIZE 2
+#define TOF_ICC_B_SIZE 3
+#define TOF_ICC_C_SIZE 2
+#define TOF_ICC_X_MASK ((1 << TOF_ICC_X_BITS) - 1)
+#define TOF_ICC_Y_MASK ((1 << TOF_ICC_Y_BITS) - 1)
+#define TOF_ICC_Z_MASK ((1 << TOF_ICC_Z_BITS) - 1)
+#define TOF_ICC_A_MASK ((1 << TOF_ICC_A_BITS) - 1)
+#define TOF_ICC_B_MASK ((1 << TOF_ICC_B_BITS) - 1)
+#define TOF_ICC_C_MASK ((1 << TOF_ICC_C_BITS) - 1)
+#define TOF_ICC_ABC_SIZE (TOF_ICC_A_SIZE * TOF_ICC_B_SIZE * TOF_ICC_C_SIZE)
+
+static inline int tof_icc_get_framelen(int len){
+	len = TOF_ICC_RH_LEN + round_up(len + TOF_ICC_ECRC_LEN, TOF_ICC_FRAME_ALIGN);
+	if(len < TOF_ICC_FRAME_LEN_MIN){
+		len = TOF_ICC_FRAME_LEN_MIN;
+	}
+	return len;
+}
+
+/** Descriptors **/
+/** commands and rcodes **/
+enum {
+	TOF_ICC_TOQ_NOP,
+	TOF_ICC_TOQ_PUT,
+	TOF_ICC_TOQ_WRITE_PIGGYBACK_BUFFER,
+	TOF_ICC_TOQ_PUT_PIGGYBACK,
+	TOF_ICC_TOQ_GET,
+	TOF_ICC_TOQ_GETL,
+	TOF_ICC_TOQ_ATOMIC_READ_MODIFY_WRITE = 0xe,
+	TOF_ICC_TOQ_TRANSMIT_RAW_PACKET1 = 0x10,
+	TOF_ICC_TOQ_TRANSMIT_RAW_PACKET2,
+	TOF_ICC_TOQ_TRANSMIT_SYSTEM_PACKET1,
+	TOF_ICC_TOQ_TRANSMIT_SYSTEM_PACKET2,
+
+	TOF_ICC_TOQ_NCOMMANDS,
+};
+
+enum {
+	TOF_ICC_MRQ_ATOMIC_READ_MODIFY_WRITE_HALFWAY_NOTICE = 0x1,
+	TOF_ICC_MRQ_ATOMIC_READ_MODIFY_WRITE_NOTICE,
+	TOF_ICC_MRQ_ATOMIC_READ_MODIFY_WRITE_REMOTE_ERROR,
+	TOF_ICC_MRQ_PUT_HALFWAY_NOTICE,
+	TOF_ICC_MRQ_PUT_LAST_HALFWAY_NOTICE,
+	TOF_ICC_MRQ_GET_HALFWAY_NOTICE,
+	TOF_ICC_MRQ_GET_LAST_HALFWAY_NOTICE,
+	TOF_ICC_MRQ_PUT_NOTICE,
+	TOF_ICC_MRQ_PUT_LAST_NOTICE,
+	TOF_ICC_MRQ_GET_NOTICE,
+	TOF_ICC_MRQ_GET_LAST_NOTICE,
+	TOF_ICC_MRQ_PUT_REMOTE_ERROR,
+	TOF_ICC_MRQ_PUT_LAST_REMOTE_ERROR,
+	TOF_ICC_MRQ_GET_REMOTE_ERROR,
+	TOF_ICC_MRQ_GET_LAST_REMOTE_ERROR,
+
+	TOF_ICC_MRQ_NCOMMANDS,
+};
+
+enum {
+	TOF_ICC_PRQ_UNKNOWN_TLP,
+	TOF_ICC_PRQ_SYSTEM_TLP,
+	TOF_ICC_PRQ_ADDRESS_RANGE_EXCEPTION = 0x6,
+	TOF_ICC_PRQ_CQ_EXCEPTION = 0x8,
+	TOF_ICC_PRQ_ILLEGAL_TLP_FLAGS,
+	TOF_ICC_PRQ_ILLEGAL_TLP_LENGTH,
+	TOF_ICC_PRQ_CQ_ERROR = 0xc,
+};
+
+/** structures **/
+struct tof_icc_steering_entry {
+	uint64_t res1:6;
+	uint64_t readonly:1;
+	uint64_t enable:1;
+	uint64_t mbva:32;
+	uint64_t res2:8;
+	uint64_t mbid:16;
+	uint64_t length;  /* for optimization */
+};
+
+struct tof_icc_mb_entry {
+	uint64_t ps:3;
+	uint64_t res1:4;
+	uint64_t enable:1;
+	uint64_t ipa:32;
+	uint64_t res2:24;
+	uint64_t npage;  /* for optimization */
+};
+
+struct tof_icc_mbpt_entry {
+	uint64_t res1:7;
+	uint64_t enable:1;
+	uint64_t res2:4;
+	uint64_t ipa:28;
+	uint64_t res3:24;
+};
+
+struct tof_icc_cq_stag_offset {
+	uint64_t offset:40;
+	uint64_t stag:18;
+	uint64_t cqid:6;
+};
+
+struct tof_icc_toq_common_header1 {
+	uint8_t interrupt:1;
+	uint8_t res1:4;
+	uint8_t source_type:2;
+	uint8_t flip:1;
+	uint8_t command;
+	union {
+		uint8_t mtu;
+		struct {
+			uint8_t res:4;
+			uint8_t op:4;
+		} armw;
+	} mtuop;
+	uint8_t sps:4;
+	uint8_t pa:1;
+	uint8_t pb:2;
+	uint8_t pc:1;
+	uint8_t rx;
+	uint8_t ry;
+	uint8_t rz;
+	uint8_t ra:1;
+	uint8_t rb:2;
+	uint8_t rc:1;
+	uint8_t res3:1;
+	uint8_t ri:3;
+};
+
+struct tof_icc_toq_common_header2 {
+	uint8_t gap;
+	uint8_t s:1;
+	uint8_t r:1;
+	uint8_t q:1;
+	uint8_t p:1;
+	uint8_t res1:1;
+	uint8_t j:1;
+	uint8_t res2:2;
+	uint16_t edata;
+	union{
+		struct {
+			uint32_t length:24;
+			uint32_t res:8;
+		} normal;
+		struct {
+			uint32_t length:6;
+			uint32_t res:26;
+		} piggyback;
+	} len;
+};
+
+struct tof_icc_toq_descriptor {
+	struct tof_icc_toq_common_header1 head1;
+	uint64_t res[3];
+};
+
+struct tof_icc_toq_nop {
+	struct tof_icc_toq_common_header1 head1;
+	uint64_t res[3];
+};
+
+struct tof_icc_toq_put {
+	struct tof_icc_toq_common_header1 head1;
+	struct tof_icc_toq_common_header2 head2;
+	struct tof_icc_cq_stag_offset remote;
+	struct tof_icc_cq_stag_offset local;
+};
+
+struct tof_icc_toq_write_piggyback_buffer {
+	struct tof_icc_toq_common_header1 head1;
+	uint64_t data[3];
+};
+
+struct tof_icc_toq_put_piggyback {
+	struct tof_icc_toq_common_header1 head1;
+	struct tof_icc_toq_common_header2 head2;
+	struct tof_icc_cq_stag_offset remote;
+	uint64_t data;
+};
+
+struct tof_icc_toq_get {
+	struct tof_icc_toq_common_header1 head1;
+	struct tof_icc_toq_common_header2 head2;
+	struct tof_icc_cq_stag_offset remote;
+	struct tof_icc_cq_stag_offset local;
+};
+
+struct tof_icc_toq_atomic_read_modify_write {
+	struct tof_icc_toq_common_header1 head1;
+	struct tof_icc_toq_common_header2 head2;
+	struct tof_icc_cq_stag_offset remote;
+	uint64_t data;
+};
+
+struct tof_icc_toq_transmit_raw_packet1 {
+	struct tof_icc_toq_common_header1 head1;
+	uint8_t gap;
+	uint8_t res4[3];
+	uint32_t length:12;
+	uint32_t res5:20;
+	uint64_t res6;
+	uint64_t pa:48;  /* for optimization */
+	uint64_t res7:16;
+};
+
+struct tof_icc_toq_transmit_raw_packet2 {
+	uint8_t interrupt:1;
+	uint8_t res1:4;
+	uint8_t source_type:2;
+	uint8_t flip:1;
+	uint8_t command;
+	uint8_t res2:7;
+	uint8_t e:1;
+	uint8_t res3[4];
+	uint8_t port:5;
+	uint8_t res4:1;
+	uint8_t vc:2;
+	uint8_t gap;
+	uint8_t res5[3];
+	uint32_t length:12;
+	uint32_t res6:20;
+	uint64_t res7;
+	uint64_t pa:48;  /* for optimization */
+	uint64_t res8:16;
+};
+
+struct tof_icc_toq_transmit_system_packet {
+	struct tof_icc_toq_common_header1 head1;  /* rx, ry, rz should be rdx, rdy, rdz */
+	uint8_t gap;
+	uint8_t res4[3];
+	uint32_t length:12;
+	uint32_t res5:20;
+	uint64_t res6;
+	uint64_t pa:48;  /* for optimization */
+	uint64_t res7:16;
+};
+
+struct tof_icc_tcq_descriptor {
+	uint8_t res1:5;
+	uint8_t counter_unmatch:1;
+	uint8_t res2:1;
+	uint8_t flip:1;
+	uint8_t rcode;
+	uint8_t res3[2];
+	union{
+		struct {
+			uint32_t length:24;
+			uint32_t res:8;
+		} normal;
+		struct {
+			uint32_t length:6;
+			uint32_t res:26;
+		} piggyback;
+	} len;
+};
+
+struct tof_icc_mrq_common_header1 {
+	uint8_t res1:7;
+	uint8_t flip:1;
+	uint8_t id;
+	uint8_t rcode;
+	uint8_t res2:4;
+	uint8_t pa:1;
+	uint8_t pb:2;
+	uint8_t pc:1;
+	uint8_t x;
+	uint8_t y;
+	uint8_t z;
+	uint8_t a:1;
+	uint8_t b:2;
+	uint8_t c:1;
+	uint8_t res3:1;
+	uint8_t i:3;
+};
+
+struct tof_icc_mrq_common_header2 {
+	uint8_t res1;
+	uint8_t res2:4;
+	uint8_t initial:1;
+	uint8_t res3:3;
+	uint16_t edata;
+	union {
+		struct {
+			uint32_t length:11;
+			uint32_t res:21;
+		} normal;
+		struct {
+			uint32_t op:4;
+			uint32_t res:28;
+		} armw;
+	} lenop;
+};
+
+struct tof_icc_mrq_atomic_read_modify_write_halfway_notice {
+	struct tof_icc_mrq_common_header1 head1;
+	struct tof_icc_mrq_common_header2 head2;
+	struct tof_icc_cq_stag_offset local;
+	struct tof_icc_cq_stag_offset remote;
+};
+
+struct tof_icc_mrq_descriptor {
+	struct tof_icc_mrq_common_header1 head1;
+	struct tof_icc_mrq_common_header2 head2;
+	struct tof_icc_cq_stag_offset cso1;
+	struct tof_icc_cq_stag_offset cso2;
+};
+
+struct tof_icc_pbq_descriptor {
+	uint64_t res1:7;
+	uint64_t f:1;
+	uint64_t res2:3;
+	uint64_t pa:29;
+	uint64_t res3:24;
+};
+
+struct tof_icc_prq_descriptor {
+	uint64_t rcode:7;
+	uint64_t f:1;
+	uint64_t res1:3;
+	uint64_t pa:29;
+	uint64_t res2:8;
+	uint64_t w:1;
+	uint64_t res3:5;
+	uint64_t l:1;
+	uint64_t e:1;
+	uint64_t res4:8;
+};
+
+
+/** Registers **/
+/* useful packed structures */
+struct tof_icc_reg_subnet {
+	uint64_t lz:6;
+	uint64_t sz:6;
+	uint64_t nz:6;
+	uint64_t ly:6;
+	uint64_t sy:6;
+	uint64_t ny:6;
+	uint64_t lx:6;
+	uint64_t sx:6;
+	uint64_t nx:6;
+	uint64_t res:10;
+};
+
+struct tof_icc_reg_bg_address {
+	uint32_t bgid:6;
+	uint32_t tni:3;
+	uint32_t c:1;
+	uint32_t b:2;
+	uint32_t a:1;
+	uint32_t z:5;
+	uint32_t y:5;
+	uint32_t x:5;
+	uint32_t pc:1;
+	uint32_t pb:2;
+	uint32_t pa:1;
+};
+
+/* relative offset of interrupt controller registers */
+#define TOF_ICC_IRQREG_IRR 0x0
+#define TOF_ICC_IRQREG_IMR 0x8
+#define TOF_ICC_IRQREG_IRC 0x10
+#define TOF_ICC_IRQREG_IMC 0x18
+#define TOF_ICC_IRQREG_ICL 0x20
+
+/* TOFU REGISTERS */
+#define  tof_icc_reg_pa 0x40000000
+
+/* CQ */
+#define TOF_ICC_REG_CQ_PA(tni, cqid) (tof_icc_reg_pa + 0 + (tni) * 0x1000000 + (cqid) * 0x10000)
+#define TOF_ICC_REG_CQ_TOQ_DIRECT_DESCRIPTOR 0x0
+#define TOF_ICC_REG_CQ_TOQ_FETCH_START 0x40
+#define TOF_ICC_REG_CQ_MRQ_FULL_POINTER 0x48
+#define TOF_ICC_REG_CQ_TOQ_PIGGYBACK_BUFFER0 0x50
+#define TOF_ICC_REG_CQ_TOQ_PIGGYBACK_BUFFER1 0x58
+#define TOF_ICC_REG_CQ_TOQ_PIGGYBACK_BUFFER2 0x60
+#define TOF_ICC_REG_CQ_TCQ_NUM_NOTICE 0x68
+#define TOF_ICC_REG_CQ_MRQ_NUM_NOTICE 0x70
+#define TOF_ICC_REG_CQ_TX_PAYLOAD_BYTE 0x78
+#define TOF_ICC_REG_CQ_RX_PAYLOAD_BYTE 0x80
+#define TOF_ICC_REG_CQ_DUMP_START 0x0
+#define TOF_ICC_REG_CQ_DUMP_END 0x88
+
+/* BCH */
+#define TOF_ICC_REG_BCH_PA(tni, bgid) (tof_icc_reg_pa + 0x0000e00000 + (tni) * 0x1000000 + (bgid) * 0x10000)
+#define TOF_ICC_REG_BCH_IDATA 0x800
+#define TOF_ICC_REG_BCH_READY 0x840
+#define TOF_ICC_REG_BCH_READY_STATE BIT(63)
+#define TOF_ICC_REG_BCH_IGNORED_SIGNAL_COUNT 0x848
+#define TOF_ICC_REG_BCH_DUMP_START 0x800
+#define TOF_ICC_REG_BCH_DUMP_END 0x850
+
+/* CQS */
+#define TOF_ICC_REG_CQS_PA(tni, cqid) (tof_icc_reg_pa + 0x0000400000 + (tni) * 0x1000000 + (cqid) * 0x10000)
+#define TOF_ICC_REG_CQS_STATUS 0x0
+#define TOF_ICC_REG_CQS_STATUS_DESCRIPTOR_PROCESS_STOP BIT(63)
+#define TOF_ICC_REG_CQS_STATUS_DESCRIPTOR_FETCH_STOP BIT(62)
+#define TOF_ICC_REG_CQS_STATUS_BLANK_ENTRY_FLIP_BIT BIT(61)
+#define TOF_ICC_REG_CQS_STATUS_CACHE_FLUSH_BUSY BIT(60)
+#define TOF_ICC_REG_CQS_STATUS_CQ_ENABLE BIT(59)
+#define TOF_ICC_REG_CQS_STATUS_SESSION_DEAD BIT(58)
+#define TOF_ICC_REG_CQS_STATUS_SESSION_OFFSET_OVERFLOW BIT(57)
+#define TOF_ICC_REG_CQS_STATUS_SESSION_OFFSET GENMASK(56, 32)
+#define TOF_ICC_REG_CQS_STATUS_NEXT_DESCRIPTOR_OFFSET GENMASK(29, 5)
+#define TOF_ICC_REG_CQS_ENABLE 0x8
+#define TOF_ICC_REG_CQS_CACHE_FLUSH 0x10
+#define TOF_ICC_REG_CQS_FETCH_STOP 0x18
+#define TOF_ICC_REG_CQS_MODE 0x20
+#define TOF_ICC_REG_CQS_MODE_SYSTEM BIT(63)
+#define TOF_ICC_REG_CQS_MODE_TRP2_ENABLE BIT(62)
+#define TOF_ICC_REG_CQS_MODE_TRP1_ENABLE BIT(61)
+#define TOF_ICC_REG_CQS_MODE_SESSION BIT(60)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_NX GENMASK(53, 48)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_SX GENMASK(47, 42)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_LX GENMASK(41, 36)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_NY GENMASK(35, 30)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_SY GENMASK(29, 24)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_LY GENMASK(23, 18)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_NZ GENMASK(17, 12)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_SZ GENMASK(11, 6)
+#define TOF_ICC_REG_CQS_MODE_SUBNET_LZ GENMASK(5, 0)
+#define TOF_ICC_REG_CQS_GPID 0x28
+#define TOF_ICC_REG_CQS_TOQ_IPA 0x30
+#define TOF_ICC_REG_CQS_TOQ_SIZE 0x38
+#define TOF_ICC_REG_CQS_TCQ_IPA 0x40
+#define TOF_ICC_REG_CQS_TCQ_IPA_CACHE_INJECTION BIT(63)
+#define TOF_ICC_REG_CQS_MRQ_IPA 0x48
+#define TOF_ICC_REG_CQS_MRQ_IPA_CACHE_INJECTION BIT(63)
+#define TOF_ICC_REG_CQS_MRQ_SIZE 0x50
+#define TOF_ICC_REG_CQS_MRQ_MASK 0x58
+#define TOF_ICC_REG_CQS_TCQ_DESCRIPTOR_COALESCING_TIMER 0x60
+#define TOF_ICC_REG_CQS_MRQ_DESCRIPTOR_COALESCING_TIMER 0x68
+#define TOF_ICC_REG_CQS_MRQ_INTERRUPT_COALESCING_TIMER 0x70
+#define TOF_ICC_REG_CQS_MRQ_INTERRUPT_COALESCING_COUNT 0x78
+#define TOF_ICC_REG_CQS_TOQ_DIRECT_SOURCE_COUNT 0x80
+#define TOF_ICC_REG_CQS_TOQ_DIRECT_DESCRIPTOR_COUNT 0x88
+#define TOF_ICC_REG_CQS_MEMORY_BLOCK_TABLE_ENABLE 0x90
+#define TOF_ICC_REG_CQS_MEMORY_BLOCK_TABLE_IPA 0x98
+#define TOF_ICC_REG_CQS_MEMORY_BLOCK_TABLE_SIZE 0xa0
+#define TOF_ICC_REG_CQS_STEERING_TABLE_ENABLE 0xa8
+#define TOF_ICC_REG_CQS_STEERING_TABLE_IPA 0xb0
+#define TOF_ICC_REG_CQS_STEERING_TABLE_SIZE 0xb8
+#define TOF_ICC_REG_CQS_MRQ_INTERRUPT_MASK 0xc0
+#define TOF_ICC_REG_CQS_IRR 0xc8
+#define TOF_ICC_REG_CQS_IMR 0xd0
+#define TOF_ICC_REG_CQS_IRC 0xd8
+#define TOF_ICC_REG_CQS_IMC 0xe0
+#define TOF_ICC_REG_CQS_ICL 0xe8
+#define TOF_ICC_REG_CQS_DUMP_START 0x0
+#define TOF_ICC_REG_CQS_DUMP_END 0xf0
+
+/* BGS */
+#define TOF_ICC_REG_BGS_PA(tni, bgid) (tof_icc_reg_pa + 0x0000800000 + (tni) * 0x1000000 + (bgid) * 0x10000)
+#define TOF_ICC_REG_BGS_ENABLE 0x0
+#define TOF_ICC_REG_BGS_IRR 0x8
+#define TOF_ICC_REG_BGS_IMR 0x10
+#define TOF_ICC_REG_BGS_IRC 0x18
+#define TOF_ICC_REG_BGS_IMC 0x20
+#define TOF_ICC_REG_BGS_ICL 0x28
+#define TOF_ICC_REG_BGS_STATE 0x30
+#define TOF_ICC_REG_BGS_STATE_ENABLE BIT(0)
+#define TOF_ICC_REG_BGS_EXCEPTION_INFO_GPID_UNMATCH 0x38
+#define TOF_ICC_REG_BGS_EXCEPTION_INFO_GPID_UNMATCH_BG_ADDRESS GENMASK(27, 0)
+#define TOF_ICC_REG_BGS_EXCEPTION_INFO_ADDRESS_UNMATCH 0x40
+#define TOF_ICC_REG_BGS_EXCEPTION_INFO_ADDRESS_UNMATCH_BG_ADDRESS GENMASK(27, 0)
+#define TOF_ICC_REG_BGS_SIGNAL_A 0x48
+#define TOF_ICC_REG_BGS_SIGNAL_A_SIG_RECV BIT(63)
+#define TOF_ICC_REG_BGS_SIGNAL_A_TLP_RECV BIT(62)
+#define TOF_ICC_REG_BGS_SIGNAL_A_SIG_SEND BIT(61)
+#define TOF_ICC_REG_BGS_SIGNAL_A_OP_TYPE GENMASK(3, 0)
+#define TOF_ICC_REG_BGS_SIGNAL_B 0x50
+#define TOF_ICC_REG_BGS_SIGNAL_B_SIG_RECV BIT(63)
+#define TOF_ICC_REG_BGS_SIGNAL_B_TLP_RECV BIT(62)
+#define TOF_ICC_REG_BGS_SIGNAL_B_SIG_SEND BIT(61)
+#define TOF_ICC_REG_BGS_SIGNAL_B_OP_TYPE GENMASK(3, 0)
+#define TOF_ICC_REG_BGS_SIGNAL_MASK 0x58
+#define TOF_ICC_REG_BGS_SIGNAL_MASK_SIG_RECV BIT(63)
+#define TOF_ICC_REG_BGS_SIGNAL_MASK_TLP_RECV BIT(62)
+#define TOF_ICC_REG_BGS_SIGNAL_MASK_SIG_SEND BIT(61)
+#define TOF_ICC_REG_BGS_SIGNAL_MASK_TLP_SEND BIT(60)
+#define TOF_ICC_REG_BGS_LOCAL_LINK 0x60
+#define TOF_ICC_REG_BGS_LOCAL_LINK_BGID_RECV GENMASK(37, 32)
+#define TOF_ICC_REG_BGS_LOCAL_LINK_BGID_SEND GENMASK(5, 0)
+#define TOF_ICC_REG_BGS_REMOTE_LINK 0x68
+#define TOF_ICC_REG_BGS_REMOTE_LINK_BG_ADDRESS_RECV GENMASK(59, 32)
+#define TOF_ICC_REG_BGS_REMOTE_LINK_BG_ADDRESS_SEND GENMASK(31, 0)
+#define TOF_ICC_REG_BGS_SUBNET_SIZE 0x70
+#define TOF_ICC_REG_BGS_GPID_BSEQ 0x78
+#define TOF_ICC_REG_BGS_DATA_A0 0x108
+#define TOF_ICC_REG_BGS_DATA_AE 0x178
+#define TOF_ICC_REG_BGS_DATA_B0 0x188
+#define TOF_ICC_REG_BGS_DATA_BE 0x1f8
+#define TOF_ICC_REG_BGS_BCH_MASK 0x800
+#define TOF_ICC_REG_BGS_BCH_MASK_MASK BIT(63)
+#define TOF_ICC_REG_BGS_BCH_MASK_STATUS 0x808
+#define TOF_ICC_REG_BGS_BCH_MASK_STATUS_RUN BIT(63)
+#define TOF_ICC_REG_BGS_BCH_NOTICE_IPA 0x810
+#define TOF_ICC_REG_BGS_DUMP_START 0x0
+#define TOF_ICC_REG_BGS_DUMP_END 0x818
+
+/* TNI */
+#define TOF_ICC_REG_TNI_PA(tni) (tof_icc_reg_pa + 0x0000c00000 + (tni) * 0x1000000)
+#define TOF_ICC_REG_TNI_IRR 0x8
+#define TOF_ICC_REG_TNI_IMR 0x10
+#define TOF_ICC_REG_TNI_IRC 0x18
+#define TOF_ICC_REG_TNI_IMC 0x20
+#define TOF_ICC_REG_TNI_ICL 0x28
+#define TOF_ICC_REG_TNI_STATE 0x30
+#define TOF_ICC_REG_TNI_STATE_MASK GENMASK(1, 0)
+#define TOF_ICC_REG_TNI_STATE_DISABLE 0
+#define TOF_ICC_REG_TNI_STATE_NORMAL 2
+#define TOF_ICC_REG_TNI_STATE_ERROR 3
+#define TOF_ICC_REG_TNI_ENABLE 0x38
+#define TOF_ICC_REG_TNI_CQ_PRESENT 0x40
+#define TOF_ICC_REG_TNI_EXCEPTION_INFO_INACTIVE_BG 0x48
+#define TOF_ICC_REG_TNI_EXCEPTION_INFO_INACTIVE_BG_DEST_BG GENMASK(37, 32)
+#define TOF_ICC_REG_TNI_EXCEPTION_INFO_INACTIVE_BG_SOURCE_BG_ADDRESS GENMASK(27, 0)
+#define TOF_ICC_REG_TNI_PRQ_FULL_POINTER 0x100
+#define TOF_ICC_REG_TNI_PBQ_PA 0x108
+#define TOF_ICC_REG_TNI_PBQ_SIZE 0x110
+#define TOF_ICC_REG_TNI_PRQ_PA 0x118
+#define TOF_ICC_REG_TNI_PRQ_PA_CACHE_INJECTION BIT(63)
+#define TOF_ICC_REG_TNI_PRQ_SIZE 0x120
+#define TOF_ICC_REG_TNI_PRQ_MASK 0x128
+#define TOF_ICC_REG_TNI_PRQ_ENTRY_COALESCING_TIMER 0x130
+#define TOF_ICC_REG_TNI_PRQ_INTERRUPT_COALESCING_TIMER 0x138
+#define TOF_ICC_REG_TNI_PRQ_INTERRUPT_COALESCING_COUNT 0x140
+#define TOF_ICC_REG_TNI_SEND_COUNT 0x148
+#define TOF_ICC_REG_TNI_NO_SEND_COUNT 0x150
+#define TOF_ICC_REG_TNI_BLOCK_SEND_COUNT 0x158
+#define TOF_ICC_REG_TNI_RECEIVE_COUNT 0x160
+#define TOF_ICC_REG_TNI_NO_RECEIVE_COUNT 0x168
+#define TOF_ICC_REG_TNI_NUM_SEND_TLP 0x170
+#define TOF_ICC_REG_TNI_BYTE_SEND_TLP 0x178
+#define TOF_ICC_REG_TNI_NUM_SEND_SYSTEM_TLP 0x180
+#define TOF_ICC_REG_TNI_NUM_RECEIVE_TLP 0x188
+#define TOF_ICC_REG_TNI_BYTE_RECEIVE_TLP 0x190
+#define TOF_ICC_REG_TNI_NUM_RECEIVE_NULLIFIED_TLP 0x198
+#define TOF_ICC_REG_TNI_RX_NUM_UNKNOWN_TLP 0x1a0
+#define TOF_ICC_REG_TNI_RX_NUM_SYSTEM_TLP 0x1a8
+#define TOF_ICC_REG_TNI_RX_NUM_EXCEPTION_TLP 0x1b0
+#define TOF_ICC_REG_TNI_RX_NUM_DISCARD_UNKNOWN_TLP 0x1b8
+#define TOF_ICC_REG_TNI_RX_NUM_DISCARD_SYSTEM_TLP 0x1c0
+#define TOF_ICC_REG_TNI_RX_NUM_DISCARD_EXCEPTION_TLP 0x1c8
+#define TOF_ICC_REG_TNI_DUMP_START 0x8
+#define TOF_ICC_REG_TNI_DUMP_END 0x1d0
+
+/* Port */
+#define TOF_ICC_REG_PORT_PA(port) (tof_icc_reg_pa + 0x0006000000 + (port) * 0x1000)
+#define TOF_ICC_REG_PORT_TX_VC0_ZERO_CREDIT_COUNT 0x0
+#define TOF_ICC_REG_PORT_TX_VC1_ZERO_CREDIT_COUNT 0x8
+#define TOF_ICC_REG_PORT_TX_VC2_ZERO_CREDIT_COUNT 0x10
+#define TOF_ICC_REG_PORT_TX_VC3_ZERO_CREDIT_COUNT 0x18
+#define TOF_ICC_REG_PORT_FREE_RUN_COUNT 0x80
+#define TOF_ICC_REG_PORT_NUM_SEND_DLLP 0xc0
+#define TOF_ICC_REG_PORT_NUM_SEND_TLP 0xc8
+#define TOF_ICC_REG_PORT_BYTE_SEND_TLP 0xd0
+#define TOF_ICC_REG_PORT_NUM_SEND_SYSTEM_TLP 0xd8
+#define TOF_ICC_REG_PORT_NUM_SEND_NULLIFIED_TLP 0xe0
+#define TOF_ICC_REG_PORT_NUM_TX_DISCARD_SYSTEM_TLP 0xe8
+#define TOF_ICC_REG_PORT_NUM_TX_DISCARD_NORMAL_TLP 0xf0
+#define TOF_ICC_REG_PORT_NUM_TX_FILTERED_NORMAL_TLP 0xf8
+#define TOF_ICC_REG_PORT_NUM_VIRTUAL_CUT_THROUGH_TLP 0x100
+#define TOF_ICC_REG_PORT_NUM_GENERATE_NULLIFIED_TLP 0x108
+#define TOF_ICC_REG_PORT_NUM_RECEIVE_DLLP 0x110
+#define TOF_ICC_REG_PORT_NUM_RECEIVE_TLP 0x118
+#define TOF_ICC_REG_PORT_BYTE_RECEIVE_TLP 0x120
+#define TOF_ICC_REG_PORT_NUM_RECEIVE_SYSTEM_TLP 0x128
+#define TOF_ICC_REG_PORT_NUM_RECEIVE_NULLIFIED_TLP 0x130
+#define TOF_ICC_REG_PORT_NUM_RX_DISCARD_SYSTEM_TLP 0x138
+#define TOF_ICC_REG_PORT_NUM_RX_DISCARD_NORMAL_TLP 0x140
+#define TOF_ICC_REG_PORT_NUM_RX_FILTERED_NORMAL_TLP 0x158
+#define TOF_ICC_REG_PORT_NUM_RX_DISCARD_NULLIFIED_TLP 0x160
+#define TOF_ICC_REG_PORT_FRAME_LCRC_ERROR_COUNT 0x170
+#define TOF_ICC_REG_PORT_TX_RETRY_BUFFER_CE_COUNT 0x180
+#define TOF_ICC_REG_PORT_RX_VC_BUFFER_CE_COUNT 0x188
+#define TOF_ICC_REG_PORT_XB_CE_COUNT 0x190
+#define TOF_ICC_REG_PORT_ACK_NACK_TIME_OUT_COUNT 0x198
+#define TOF_ICC_REG_PORT_SLICE0_FCS_ERROR_COUNT 0x1a0
+#define TOF_ICC_REG_PORT_SLICE1_FCS_ERROR_COUNT 0x1a8
+#define TOF_ICC_REG_PORT_DUMP_START 0x0
+#define TOF_ICC_REG_PORT_DUMP_END 0x1b0
+
+/* XB */
+#define TOF_ICC_REG_XB_PA (tof_icc_reg_pa + 0x000600f000)
+#define TOF_ICC_REG_XB_STQ_ENABLE 0x0
+#define TOF_ICC_REG_XB_STQ_UPDATE_INTERVAL 0x8
+#define TOF_ICC_REG_XB_STQ_PA 0x10
+#define TOF_ICC_REG_XB_STQ_SIZE 0x18
+#define TOF_ICC_REG_XB_STQ_NEXT_OFFSET 0x20
+#define TOF_ICC_REG_XB_DUMP_START 0x0
+#define TOF_ICC_REG_XB_DUMP_END 0x28
+
+#define TOF_ICC_XB_TC_DATA_CYCLE_COUNT(tni) ((tni) * 0x10 + 0x0)
+#define TOF_ICC_XB_TC_WAIT_CYCLE_COUNT(tni) ((tni) * 0x10 + 0x8)
+#define TOF_ICC_XB_TD_DATA_CYCLE_COUNT(tnr) ((tnr) * 0x10 + 0x60)
+#define TOF_ICC_XB_TD_WAIT_CYCLE_COUNT(tnr) ((tnr) * 0x10 + 0x68)
+
+/* Tofu */
+#define TOF_ICC_REG_TOFU_PA (tof_icc_reg_pa + 0x0007000000)
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS 0x0
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS_X GENMASK(22, 18)
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS_Y GENMASK(17, 13)
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS_Z GENMASK(12, 8)
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS_A BIT(7)
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS_B GENMASK(6, 5)
+#define TOF_ICC_REG_TOFU_NODE_ADDRESS_C BIT(4)
+#define TOF_ICC_REG_TOFU_PORT_SETTING 0x8
+#define TOF_ICC_REG_TOFU_TD_TLP_FILTER(tnr) ((tnr) * 0x10 + 0x10)
+#define TOF_ICC_REG_TOFU_TD_SETTINGS(tnr) ((tnr) * 0x10 + 0x18)
+#define TOF_ICC_REG_TOFU_TNR_MSI_BASE 0xc0
+#define TOF_ICC_REG_TOFU_TNR_IRR 0xc8
+#define TOF_ICC_REG_TOFU_TNR_IMR 0xd0
+#define TOF_ICC_REG_TOFU_TNR_IRC 0xd8
+#define TOF_ICC_REG_TOFU_TNR_IMC 0xe0
+#define TOF_ICC_REG_TOFU_TNR_ICL 0xe8
+#define TOF_ICC_REG_TOFU_TNI_VMS(tni, vmsid) ((tni) * 0x100 + (vmsid) * 0x8 + 0x100)
+#define TOF_ICC_REG_TOFU_TNI_VMS_CQ00(tni) ((tni) * 0x100 + 0x180)
+#define TOF_ICC_REG_TOFU_TNI_VMS_BG00(tni) ((tni) * 0x100 + 0x1a0)
+#define TOF_ICC_REG_TOFU_TNI_VMS_BG16(tni) ((tni) * 0x100 + 0x1a8)
+#define TOF_ICC_REG_TOFU_TNI_VMS_BG32(tni) ((tni) * 0x100 + 0x1b0)
+#define TOF_ICC_REG_TOFU_TNI_MSI_BASE(tni) ((tni) * 0x100 + 0x1c0)
+#define TOF_ICC_REG_TOFU_DUMP_START 0x0
+#define TOF_ICC_REG_TOFU_DUMP_END 0x6c8
+
+/** Interrupts **/
+#define TOF_ICC_IRQ_CQS_TOQ_READ_EXCEPTION BIT(0)
+#define TOF_ICC_IRQ_CQS_TOQ_DIRECT_DESCRIPTOR_EXCEPTION BIT(1)
+#define TOF_ICC_IRQ_CQS_TOQ_MARKED_UE BIT(2)
+#define TOF_ICC_IRQ_CQS_TCQ_WRITE_EXCEPTION BIT(3)
+#define TOF_ICC_IRQ_CQS_TOQ_SOURCE_TYPE_EXCEPTION BIT(4)
+#define TOF_ICC_IRQ_CQS_TCQ_WRITE_ACKNOWLEDGE BIT(5)
+#define TOF_ICC_IRQ_CQS_MRQ_WRITE_ACKNOWLEDGE BIT(7)
+#define TOF_ICC_IRQ_CQS_MRQ_WRITE_EXCEPTION BIT(8)
+#define TOF_ICC_IRQ_CQS_MRQ_OVERFLOW BIT(9)
+#define TOF_ICC_IRQ_CQS_STEERING_READ_EXCEPTION BIT(36)
+#define TOF_ICC_IRQ_CQS_MB_READ_EXCEPTION BIT(38)
+#define TOF_ICC_IRQ_CQS_PAYLOAD_READ_EXCEPTION BIT(39)
+#define TOF_ICC_IRQ_CQS_PAYLOAD_WRITE_EXCEPTION BIT(40)
+/* Just for convinience of irr value, no exists CQS CACHEFLUSH_TIMEOUT interrupt  */
+#define TOF_ICC_DUMMY_IRQ_CQS_CACHEFLUSH_TIMEOUT BIT(63)
+
+#define TOF_ICC_IRQ_BGS_NODE_ADDRESS_UNMATCH BIT(0)
+#define TOF_ICC_IRQ_BGS_BG_RECV_ADDRESS_EXCEPTION BIT(1)
+#define TOF_ICC_IRQ_BGS_BG_SEND_ADDRESS_EXCEPTION BIT(2)
+#define TOF_ICC_IRQ_BGS_GPID_UNMATCH BIT(3)
+#define TOF_ICC_IRQ_BGS_BSEQ_UNMATCH BIT(4)
+#define TOF_ICC_IRQ_BGS_SIGNAL_STATE_ERROR BIT(5)
+#define TOF_ICC_IRQ_BGS_SYNCHRONIZATION_ACKNOWLEDGE BIT(24)
+#define TOF_ICC_IRQ_BGS_ERROR_SYNCHRONIZATION_ACKNOWLEDGE BIT(25)
+#define TOF_ICC_IRQ_BGS_DMA_COMPLETION_EXCEPTION BIT(26)
+
+#define TOF_ICC_IRQ_TNI_PBQ_READ_EXCEPTION BIT(0)
+#define TOF_ICC_IRQ_TNI_PBQ_MARKED_UE BIT(1)
+#define TOF_ICC_IRQ_TNI_PBQ_UNDERFLOW BIT(2)
+#define TOF_ICC_IRQ_TNI_PRQ_PACKET_DISCARD BIT(3)
+#define TOF_ICC_IRQ_TNI_PRQ_WRITE_ACKNOWLEDGE BIT(4)
+#define TOF_ICC_IRQ_TNI_PRQ_WRITE_EXCEPTION BIT(5)
+#define TOF_ICC_IRQ_TNI_PRQ_OVERFLOW BIT(6)
+#define TOF_ICC_IRQ_TNI_INACTIVE_BG BIT(16)
+#define TOF_ICC_IRQ_TNI_STAGE2_TRANSLATION_FAULT BIT(32)
+
+#define TOF_ICC_IRQ_TNR_TNR0_RX_FILTER_OUT BIT(0)
+#define TOF_ICC_IRQ_TNR_TNR0_TX_FILTER_OUT BIT(1)
+#define TOF_ICC_IRQ_TNR_TNR0_PORT_ERROR BIT(2)
+#define TOF_ICC_IRQ_TNR_TNR0_DATELINE_ERROR BIT(3)
+#define TOF_ICC_IRQ_TNR_TNR0_ROUTING_ERROR BIT(4)
+#define TOF_ICC_IRQ_TNR_TNR1_RX_FILTER_OUT BIT(6)
+#define TOF_ICC_IRQ_TNR_TNR1_TX_FILTER_OUT BIT(7)
+#define TOF_ICC_IRQ_TNR_TNR1_PORT_ERROR BIT(8)
+#define TOF_ICC_IRQ_TNR_TNR1_DATELINE_ERROR BIT(9)
+#define TOF_ICC_IRQ_TNR_TNR1_ROUTING_ERROR BIT(10)
+#define TOF_ICC_IRQ_TNR_TNR2_RX_FILTER_OUT BIT(12)
+#define TOF_ICC_IRQ_TNR_TNR2_TX_FILTER_OUT BIT(13)
+#define TOF_ICC_IRQ_TNR_TNR2_PORT_ERROR BIT(14)
+#define TOF_ICC_IRQ_TNR_TNR2_DATELINE_ERROR BIT(15)
+#define TOF_ICC_IRQ_TNR_TNR2_ROUTING_ERROR BIT(16)
+#define TOF_ICC_IRQ_TNR_TNR3_RX_FILTER_OUT BIT(18)
+#define TOF_ICC_IRQ_TNR_TNR3_TX_FILTER_OUT BIT(19)
+#define TOF_ICC_IRQ_TNR_TNR3_PORT_ERROR BIT(20)
+#define TOF_ICC_IRQ_TNR_TNR3_DATELINE_ERROR BIT(21)
+#define TOF_ICC_IRQ_TNR_TNR3_ROUTING_ERROR BIT(22)
+#define TOF_ICC_IRQ_TNR_TNR4_RX_FILTER_OUT BIT(24)
+#define TOF_ICC_IRQ_TNR_TNR4_TX_FILTER_OUT BIT(25)
+#define TOF_ICC_IRQ_TNR_TNR4_PORT_ERROR BIT(26)
+#define TOF_ICC_IRQ_TNR_TNR4_DATELINE_ERROR BIT(27)
+#define TOF_ICC_IRQ_TNR_TNR4_ROUTING_ERROR BIT(28)
+#define TOF_ICC_IRQ_TNR_TNR5_RX_FILTER_OUT BIT(30)
+#define TOF_ICC_IRQ_TNR_TNR5_TX_FILTER_OUT BIT(31)
+#define TOF_ICC_IRQ_TNR_TNR5_PORT_ERROR BIT(32)
+#define TOF_ICC_IRQ_TNR_TNR5_DATELINE_ERROR BIT(33)
+#define TOF_ICC_IRQ_TNR_TNR5_ROUTING_ERROR BIT(34)
+#define TOF_ICC_IRQ_TNR_TNR6_RX_FILTER_OUT BIT(36)
+#define TOF_ICC_IRQ_TNR_TNR6_TX_FILTER_OUT BIT(37)
+#define TOF_ICC_IRQ_TNR_TNR6_PORT_ERROR BIT(38)
+#define TOF_ICC_IRQ_TNR_TNR6_DATELINE_ERROR BIT(39)
+#define TOF_ICC_IRQ_TNR_TNR6_ROUTING_ERROR BIT(40)
+#define TOF_ICC_IRQ_TNR_TNR7_RX_FILTER_OUT BIT(42)
+#define TOF_ICC_IRQ_TNR_TNR7_TX_FILTER_OUT BIT(43)
+#define TOF_ICC_IRQ_TNR_TNR7_PORT_ERROR BIT(44)
+#define TOF_ICC_IRQ_TNR_TNR7_DATELINE_ERROR BIT(45)
+#define TOF_ICC_IRQ_TNR_TNR7_ROUTING_ERROR BIT(46)
+#define TOF_ICC_IRQ_TNR_TNR8_RX_FILTER_OUT BIT(48)
+#define TOF_ICC_IRQ_TNR_TNR8_TX_FILTER_OUT BIT(49)
+#define TOF_ICC_IRQ_TNR_TNR8_PORT_ERROR BIT(50)
+#define TOF_ICC_IRQ_TNR_TNR8_DATELINE_ERROR BIT(51)
+#define TOF_ICC_IRQ_TNR_TNR8_ROUTING_ERROR BIT(52)
+#define TOF_ICC_IRQ_TNR_TNR9_RX_FILTER_OUT BIT(54)
+#define TOF_ICC_IRQ_TNR_TNR9_TX_FILTER_OUT BIT(55)
+#define TOF_ICC_IRQ_TNR_TNR9_PORT_ERROR BIT(56)
+#define TOF_ICC_IRQ_TNR_TNR9_DATELINE_ERROR BIT(57)
+#define TOF_ICC_IRQ_TNR_TNR9_ROUTING_ERROR BIT(58)
+
+#endif
+
+/* vim: set noet ts=8 sw=8 sts=0 tw=0 : */
+

executer/kernel/mcctrl/arch/arm64/include/tofu/tof_uapi.h

@@ -0,0 +1,345 @@
+#ifndef _TOF_UAPI_H_
+#define _TOF_UAPI_H_
+
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#include <unistd.h>
+#include <sys/ioctl.h>
+#include <sys/types.h>
+#endif
+
+enum tof_sig_errno_cq {
+	TOF_TOQ_DIRECT_DESCRIPTOR_EXCEPTION,
+	TOF_TOQ_SOURCE_TYPE_EXCEPTION,
+	TOF_MRQ_OVERFLOW,
+	TOF_CQS_CACHEFLUSH_TIMEOUT,
+};
+
+enum tof_sig_errno_bg {
+	TOF_NODE_ADDRESS_UNMATCH,
+	TOF_BSEQ_UNMATCH,
+	TOF_SIGNAL_STATE_ERROR,
+	TOF_ERROR_SYNCHRONIZATION_ACKNOWLEDGE,
+};
+
+#define TOF_UAPI_VERSION 0x2a00
+
+struct tof_init_cq {
+	uint16_t version;
+	uint8_t session_mode;
+	uint8_t toq_size;
+	uint8_t mrq_size;
+	uint8_t num_stag;
+	uint8_t tcq_cinj;
+	uint8_t mrq_cinj;
+	void *toq_mem;
+	void *tcq_mem;
+	void *mrq_mem;
+};
+
+struct tof_alloc_stag {
+	uint32_t flags;
+	int stag;
+	uint64_t offset;
+	void *va;
+	uint64_t len;
+};
+
+struct tof_free_stags {
+	uint16_t num;
+	int *stags;
+};
+
+struct tof_addr {
+	uint8_t pa;
+	uint8_t pb;
+	uint8_t pc;
+	uint8_t x;
+	uint8_t y;
+	uint8_t z;
+	uint8_t a;
+	uint8_t b;
+	uint8_t c;
+};
+
+struct tof_set_bg {
+	int tni;
+	int gate;
+	int source_lgate;
+	struct tof_addr source_raddr;
+	int source_rtni;
+	int source_rgate;
+	int dest_lgate;
+	struct tof_addr dest_raddr;
+	int dest_rtni;
+	int dest_rgate;
+};
+
+struct tof_enable_bch {
+	void *addr;
+	int bseq;
+	int num;
+	struct tof_set_bg *bgs;
+};
+
+struct tof_set_subnet {
+	int res0;
+	int res1;
+	uint8_t nx;
+	uint8_t sx;
+	uint8_t lx;
+	uint8_t ny;
+	uint8_t sy;
+	uint8_t ly;
+	uint8_t nz;
+	uint8_t sz;
+	uint8_t lz;
+};
+
+struct tof_reg_user {
+	uid_t uid;
+	uint32_t gpid;
+	struct tof_set_subnet subnet;
+	uint64_t *cqmask;
+	uint64_t *bgmask;
+};
+
+struct tof_notify_linkdown {
+	int num;
+	struct {
+		uint8_t x;
+		uint8_t y;
+		uint8_t z;
+		uint8_t a;
+		uint8_t b;
+		uint8_t c;
+		uint16_t ports;
+	} *items;
+};
+
+struct tof_get_port_stat {
+	int port_no;
+	uint64_t mask;
+	uint64_t pa[31];
+};
+
+struct tof_get_cq_stat {
+	int tni;
+	int cqid;
+	uint64_t txbyte;
+	uint64_t rxbyte;
+};
+
+struct tof_load_register {
+	uint64_t pa;
+	uint64_t len;
+	void *buf;
+};
+
+struct tof_load_resource {
+	uint64_t rsc_id;
+	uint64_t offset;
+	uint64_t len;
+	void *buf;
+};
+
+union tof_trans_table_bitfield {
+	struct {
+		uint64_t start:36;
+		uint64_t len:27;
+		uint64_t ps_code:1;
+	} bits;
+	uint64_t atomic;
+};
+
+struct tof_trans_table {
+	union tof_trans_table_bitfield steering;
+	union tof_trans_table_bitfield mbpt;
+};
+
+void tof_utofu_set_linkdown_callback(void (*callback)(int, const void *));
+void tof_utofu_unset_linkdown_callback(void);
+
+#define TOF_MMAP_CQ_REGISTER 0
+#ifdef __KERNEL__
+#define TOF_MMAP_CQ_TRANSTABLE (PAGE_SIZE)
+#else
+#define TOF_MMAP_CQ_TRANSTABLE (sysconf(_SC_PAGESIZE))
+#endif
+#define TOF_MMAP_BCH_REGISTER 0
+#define TOF_MMAP_XB_STQ 0
+
+#define TOF_ST_RDWR 0x0
+#define TOF_ST_RDONLY 0x1
+#define TOF_ST_LPG 0x2
+
+#define TOF_STAG_TRANS_PS_CODE_64KB 0
+#define TOF_STAG_TRANS_PS_CODE_2MB  1
+
+#define TOF_IOC_MAGIC 'd'
+#define TOF_IOCTL_INIT_CQ _IOWR(TOF_IOC_MAGIC, 0, long)
+#define TOF_IOCTL_ALLOC_STAG _IOWR(TOF_IOC_MAGIC, 1, long)
+#define TOF_IOCTL_FREE_STAGS _IOWR(TOF_IOC_MAGIC, 2, long)
+#define TOF_IOCTL_ENABLE_BCH _IOWR(TOF_IOC_MAGIC, 3, long)
+#define TOF_IOCTL_DISABLE_BCH _IOWR(TOF_IOC_MAGIC, 4, long)
+#define TOF_IOCTL_SET_RT_SIGNAL _IOWR(TOF_IOC_MAGIC, 5, long)
+#define TOF_IOCTL_SET_SUBNET _IOWR(TOF_IOC_MAGIC, 6, long)
+#define TOF_IOCTL_REG_USER _IOWR(TOF_IOC_MAGIC, 7, long)
+#define TOF_IOCTL_NOTIFY_LINKDOWN _IOWR(TOF_IOC_MAGIC, 8, long)
+#define TOF_IOCTL_GET_PORT_STAT _IOWR(TOF_IOC_MAGIC, 9, long)
+#define TOF_IOCTL_GET_CQ_STAT _IOWR(TOF_IOC_MAGIC, 10, long)
+#define TOF_IOCTL_LOAD_REGISTER _IOWR(TOF_IOC_MAGIC, 11, long)
+#define TOF_IOCTL_LOAD_RESOURCE _IOWR(TOF_IOC_MAGIC, 12, long)
+
+#ifndef __KERNEL__
+#define TOF_INIT_CQ TOF_IOCTL_INIT_CQ
+#define TOF_ALLOC_STAG TOF_IOCTL_ALLOC_STAG
+#define TOF_FREE_STAGS TOF_IOCTL_FREE_STAGS
+#define TOF_ENABLE_BCH TOF_IOCTL_ENABLE_BCH
+#define TOF_DISABLE_BCH TOF_IOCTL_DISABLE_BCH
+#define TOF_SET_RT_SIGNAL TOF_IOCTL_SET_RT_SIGNAL
+#define TOF_SET_SUBNET TOF_IOCTL_SET_SUBNET
+#define TOF_REG_USER TOF_IOCTL_REG_USER
+#define TOF_NOTIFY_LINKDOWN TOF_IOCTL_NOTIFY_LINKDOWN
+#define TOF_GET_PORT_STAT TOF_IOCTL_GET_PORT_STAT
+#define TOF_GET_CQ_STAT TOF_IOCTL_GET_CQ_STAT
+#define TOF_LOAD_REGISTER TOF_IOCTL_LOAD_REGISTER
+#define TOF_LOAD_RESOURCE TOF_IOCTL_LOAD_RESOURCE
+#endif
+
+enum {
+	/* TOQ (0 - 71) */
+	TOF_RSC_TNI0_TOQ0 = 0, TOF_RSC_TNI0_TOQ1, TOF_RSC_TNI0_TOQ2, TOF_RSC_TNI0_TOQ3,
+	TOF_RSC_TNI0_TOQ4, TOF_RSC_TNI0_TOQ5, TOF_RSC_TNI0_TOQ6, TOF_RSC_TNI0_TOQ7,
+	TOF_RSC_TNI0_TOQ8, TOF_RSC_TNI0_TOQ9, TOF_RSC_TNI0_TOQ10, TOF_RSC_TNI0_TOQ11,
+	TOF_RSC_TNI1_TOQ0, TOF_RSC_TNI1_TOQ1, TOF_RSC_TNI1_TOQ2, TOF_RSC_TNI1_TOQ3,
+	TOF_RSC_TNI1_TOQ4, TOF_RSC_TNI1_TOQ5, TOF_RSC_TNI1_TOQ6, TOF_RSC_TNI1_TOQ7,
+	TOF_RSC_TNI1_TOQ8, TOF_RSC_TNI1_TOQ9, TOF_RSC_TNI1_TOQ10, TOF_RSC_TNI1_TOQ11,
+	TOF_RSC_TNI2_TOQ0, TOF_RSC_TNI2_TOQ1, TOF_RSC_TNI2_TOQ2, TOF_RSC_TNI2_TOQ3,
+	TOF_RSC_TNI2_TOQ4, TOF_RSC_TNI2_TOQ5, TOF_RSC_TNI2_TOQ6, TOF_RSC_TNI2_TOQ7,
+	TOF_RSC_TNI2_TOQ8, TOF_RSC_TNI2_TOQ9, TOF_RSC_TNI2_TOQ10, TOF_RSC_TNI2_TOQ11,
+	TOF_RSC_TNI3_TOQ0, TOF_RSC_TNI3_TOQ1, TOF_RSC_TNI3_TOQ2, TOF_RSC_TNI3_TOQ3,
+	TOF_RSC_TNI3_TOQ4, TOF_RSC_TNI3_TOQ5, TOF_RSC_TNI3_TOQ6, TOF_RSC_TNI3_TOQ7,
+	TOF_RSC_TNI3_TOQ8, TOF_RSC_TNI3_TOQ9, TOF_RSC_TNI3_TOQ10, TOF_RSC_TNI3_TOQ11,
+	TOF_RSC_TNI4_TOQ0, TOF_RSC_TNI4_TOQ1, TOF_RSC_TNI4_TOQ2, TOF_RSC_TNI4_TOQ3,
+	TOF_RSC_TNI4_TOQ4, TOF_RSC_TNI4_TOQ5, TOF_RSC_TNI4_TOQ6, TOF_RSC_TNI4_TOQ7,
+	TOF_RSC_TNI4_TOQ8, TOF_RSC_TNI4_TOQ9, TOF_RSC_TNI4_TOQ10, TOF_RSC_TNI4_TOQ11,
+	TOF_RSC_TNI5_TOQ0, TOF_RSC_TNI5_TOQ1, TOF_RSC_TNI5_TOQ2, TOF_RSC_TNI5_TOQ3,
+	TOF_RSC_TNI5_TOQ4, TOF_RSC_TNI5_TOQ5, TOF_RSC_TNI5_TOQ6, TOF_RSC_TNI5_TOQ7,
+	TOF_RSC_TNI5_TOQ8, TOF_RSC_TNI5_TOQ9, TOF_RSC_TNI5_TOQ10, TOF_RSC_TNI5_TOQ11,
+
+	/* TOQ (72 - 143) */
+	TOF_RSC_TNI0_TCQ0, TOF_RSC_TNI0_TCQ1, TOF_RSC_TNI0_TCQ2, TOF_RSC_TNI0_TCQ3,
+	TOF_RSC_TNI0_TCQ4, TOF_RSC_TNI0_TCQ5, TOF_RSC_TNI0_TCQ6, TOF_RSC_TNI0_TCQ7,
+	TOF_RSC_TNI0_TCQ8, TOF_RSC_TNI0_TCQ9, TOF_RSC_TNI0_TCQ10, TOF_RSC_TNI0_TCQ11,
+	TOF_RSC_TNI1_TCQ0, TOF_RSC_TNI1_TCQ1, TOF_RSC_TNI1_TCQ2, TOF_RSC_TNI1_TCQ3,
+	TOF_RSC_TNI1_TCQ4, TOF_RSC_TNI1_TCQ5, TOF_RSC_TNI1_TCQ6, TOF_RSC_TNI1_TCQ7,
+	TOF_RSC_TNI1_TCQ8, TOF_RSC_TNI1_TCQ9, TOF_RSC_TNI1_TCQ10, TOF_RSC_TNI1_TCQ11,
+	TOF_RSC_TNI2_TCQ0, TOF_RSC_TNI2_TCQ1, TOF_RSC_TNI2_TCQ2, TOF_RSC_TNI2_TCQ3,
+	TOF_RSC_TNI2_TCQ4, TOF_RSC_TNI2_TCQ5, TOF_RSC_TNI2_TCQ6, TOF_RSC_TNI2_TCQ7,
+	TOF_RSC_TNI2_TCQ8, TOF_RSC_TNI2_TCQ9, TOF_RSC_TNI2_TCQ10, TOF_RSC_TNI2_TCQ11,
+	TOF_RSC_TNI3_TCQ0, TOF_RSC_TNI3_TCQ1, TOF_RSC_TNI3_TCQ2, TOF_RSC_TNI3_TCQ3,
+	TOF_RSC_TNI3_TCQ4, TOF_RSC_TNI3_TCQ5, TOF_RSC_TNI3_TCQ6, TOF_RSC_TNI3_TCQ7,
+	TOF_RSC_TNI3_TCQ8, TOF_RSC_TNI3_TCQ9, TOF_RSC_TNI3_TCQ10, TOF_RSC_TNI3_TCQ11,
+	TOF_RSC_TNI4_TCQ0, TOF_RSC_TNI4_TCQ1, TOF_RSC_TNI4_TCQ2, TOF_RSC_TNI4_TCQ3,
+	TOF_RSC_TNI4_TCQ4, TOF_RSC_TNI4_TCQ5, TOF_RSC_TNI4_TCQ6, TOF_RSC_TNI4_TCQ7,
+	TOF_RSC_TNI4_TCQ8, TOF_RSC_TNI4_TCQ9, TOF_RSC_TNI4_TCQ10, TOF_RSC_TNI4_TCQ11,
+	TOF_RSC_TNI5_TCQ0, TOF_RSC_TNI5_TCQ1, TOF_RSC_TNI5_TCQ2, TOF_RSC_TNI5_TCQ3,
+	TOF_RSC_TNI5_TCQ4, TOF_RSC_TNI5_TCQ5, TOF_RSC_TNI5_TCQ6, TOF_RSC_TNI5_TCQ7,
+	TOF_RSC_TNI5_TCQ8, TOF_RSC_TNI5_TCQ9, TOF_RSC_TNI5_TCQ10, TOF_RSC_TNI5_TCQ11,
+
+	/* MRQ (144 - 215) */
+	TOF_RSC_TNI0_MRQ0, TOF_RSC_TNI0_MRQ1, TOF_RSC_TNI0_MRQ2, TOF_RSC_TNI0_MRQ3,
+	TOF_RSC_TNI0_MRQ4, TOF_RSC_TNI0_MRQ5, TOF_RSC_TNI0_MRQ6, TOF_RSC_TNI0_MRQ7,
+	TOF_RSC_TNI0_MRQ8, TOF_RSC_TNI0_MRQ9, TOF_RSC_TNI0_MRQ10, TOF_RSC_TNI0_MRQ11,
+	TOF_RSC_TNI1_MRQ0, TOF_RSC_TNI1_MRQ1, TOF_RSC_TNI1_MRQ2, TOF_RSC_TNI1_MRQ3,
+	TOF_RSC_TNI1_MRQ4, TOF_RSC_TNI1_MRQ5, TOF_RSC_TNI1_MRQ6, TOF_RSC_TNI1_MRQ7,
+	TOF_RSC_TNI1_MRQ8, TOF_RSC_TNI1_MRQ9, TOF_RSC_TNI1_MRQ10, TOF_RSC_TNI1_MRQ11,
+	TOF_RSC_TNI2_MRQ0, TOF_RSC_TNI2_MRQ1, TOF_RSC_TNI2_MRQ2, TOF_RSC_TNI2_MRQ3,
+	TOF_RSC_TNI2_MRQ4, TOF_RSC_TNI2_MRQ5, TOF_RSC_TNI2_MRQ6, TOF_RSC_TNI2_MRQ7,
+	TOF_RSC_TNI2_MRQ8, TOF_RSC_TNI2_MRQ9, TOF_RSC_TNI2_MRQ10, TOF_RSC_TNI2_MRQ11,
+	TOF_RSC_TNI3_MRQ0, TOF_RSC_TNI3_MRQ1, TOF_RSC_TNI3_MRQ2, TOF_RSC_TNI3_MRQ3,
+	TOF_RSC_TNI3_MRQ4, TOF_RSC_TNI3_MRQ5, TOF_RSC_TNI3_MRQ6, TOF_RSC_TNI3_MRQ7,
+	TOF_RSC_TNI3_MRQ8, TOF_RSC_TNI3_MRQ9, TOF_RSC_TNI3_MRQ10, TOF_RSC_TNI3_MRQ11,
+	TOF_RSC_TNI4_MRQ0, TOF_RSC_TNI4_MRQ1, TOF_RSC_TNI4_MRQ2, TOF_RSC_TNI4_MRQ3,
+	TOF_RSC_TNI4_MRQ4, TOF_RSC_TNI4_MRQ5, TOF_RSC_TNI4_MRQ6, TOF_RSC_TNI4_MRQ7,
+	TOF_RSC_TNI4_MRQ8, TOF_RSC_TNI4_MRQ9, TOF_RSC_TNI4_MRQ10, TOF_RSC_TNI4_MRQ11,
+	TOF_RSC_TNI5_MRQ0, TOF_RSC_TNI5_MRQ1, TOF_RSC_TNI5_MRQ2, TOF_RSC_TNI5_MRQ3,
+	TOF_RSC_TNI5_MRQ4, TOF_RSC_TNI5_MRQ5, TOF_RSC_TNI5_MRQ6, TOF_RSC_TNI5_MRQ7,
+	TOF_RSC_TNI5_MRQ8, TOF_RSC_TNI5_MRQ9, TOF_RSC_TNI5_MRQ10, TOF_RSC_TNI5_MRQ11,
+
+	/* PBQ (216 - 221) */
+	TOF_RSC_TNI0_PBQ, TOF_RSC_TNI1_PBQ, TOF_RSC_TNI2_PBQ, TOF_RSC_TNI3_PBQ,
+	TOF_RSC_TNI4_PBQ, TOF_RSC_TNI5_PBQ,
+
+	/* PRQ (222 - 227) */
+	TOF_RSC_TNI0_PRQ, TOF_RSC_TNI1_PRQ, TOF_RSC_TNI2_PRQ, TOF_RSC_TNI3_PRQ,
+	TOF_RSC_TNI4_PRQ, TOF_RSC_TNI5_PRQ,
+
+	/* STEERINGTABLE (228 - 299) */
+	TOF_RSC_TNI0_STEERINGTABLE0, TOF_RSC_TNI0_STEERINGTABLE1, TOF_RSC_TNI0_STEERINGTABLE2,
+	TOF_RSC_TNI0_STEERINGTABLE3, TOF_RSC_TNI0_STEERINGTABLE4, TOF_RSC_TNI0_STEERINGTABLE5,
+	TOF_RSC_TNI0_STEERINGTABLE6, TOF_RSC_TNI0_STEERINGTABLE7, TOF_RSC_TNI0_STEERINGTABLE8,
+	TOF_RSC_TNI0_STEERINGTABLE9, TOF_RSC_TNI0_STEERINGTABLE10, TOF_RSC_TNI0_STEERINGTABLE11,
+	TOF_RSC_TNI1_STEERINGTABLE0, TOF_RSC_TNI1_STEERINGTABLE1, TOF_RSC_TNI1_STEERINGTABLE2,
+	TOF_RSC_TNI1_STEERINGTABLE3, TOF_RSC_TNI1_STEERINGTABLE4, TOF_RSC_TNI1_STEERINGTABLE5,
+	TOF_RSC_TNI1_STEERINGTABLE6, TOF_RSC_TNI1_STEERINGTABLE7, TOF_RSC_TNI1_STEERINGTABLE8,
+	TOF_RSC_TNI1_STEERINGTABLE9, TOF_RSC_TNI1_STEERINGTABLE10, TOF_RSC_TNI1_STEERINGTABLE11,
+	TOF_RSC_TNI2_STEERINGTABLE0, TOF_RSC_TNI2_STEERINGTABLE1, TOF_RSC_TNI2_STEERINGTABLE2,
+	TOF_RSC_TNI2_STEERINGTABLE3, TOF_RSC_TNI2_STEERINGTABLE4, TOF_RSC_TNI2_STEERINGTABLE5,
+	TOF_RSC_TNI2_STEERINGTABLE6, TOF_RSC_TNI2_STEERINGTABLE7, TOF_RSC_TNI2_STEERINGTABLE8,
+	TOF_RSC_TNI2_STEERINGTABLE9, TOF_RSC_TNI2_STEERINGTABLE10, TOF_RSC_TNI2_STEERINGTABLE11,
+	TOF_RSC_TNI3_STEERINGTABLE0, TOF_RSC_TNI3_STEERINGTABLE1, TOF_RSC_TNI3_STEERINGTABLE2,
+	TOF_RSC_TNI3_STEERINGTABLE3, TOF_RSC_TNI3_STEERINGTABLE4, TOF_RSC_TNI3_STEERINGTABLE5,
+	TOF_RSC_TNI3_STEERINGTABLE6, TOF_RSC_TNI3_STEERINGTABLE7, TOF_RSC_TNI3_STEERINGTABLE8,
+        TOF_RSC_TNI3_STEERINGTABLE9, TOF_RSC_TNI3_STEERINGTABLE10, TOF_RSC_TNI3_STEERINGTABLE11,
+        TOF_RSC_TNI4_STEERINGTABLE0, TOF_RSC_TNI4_STEERINGTABLE1, TOF_RSC_TNI4_STEERINGTABLE2,
+        TOF_RSC_TNI4_STEERINGTABLE3, TOF_RSC_TNI4_STEERINGTABLE4, TOF_RSC_TNI4_STEERINGTABLE5,
+        TOF_RSC_TNI4_STEERINGTABLE6, TOF_RSC_TNI4_STEERINGTABLE7, TOF_RSC_TNI4_STEERINGTABLE8,
+        TOF_RSC_TNI4_STEERINGTABLE9, TOF_RSC_TNI4_STEERINGTABLE10, TOF_RSC_TNI4_STEERINGTABLE11,
+        TOF_RSC_TNI5_STEERINGTABLE3, TOF_RSC_TNI5_STEERINGTABLE4, TOF_RSC_TNI5_STEERINGTABLE5,
+        TOF_RSC_TNI5_STEERINGTABLE6, TOF_RSC_TNI5_STEERINGTABLE7, TOF_RSC_TNI5_STEERINGTABLE8,
+        TOF_RSC_TNI5_STEERINGTABLE9, TOF_RSC_TNI5_STEERINGTABLE10, TOF_RSC_TNI5_STEERINGTABLE11,
+
+        /* MBTABLE (300 - 371) */
+        TOF_RSC_TNI0_MBTABLE0, TOF_RSC_TNI0_MBTABLE1, TOF_RSC_TNI0_MBTABLE2,
+        TOF_RSC_TNI0_MBTABLE3, TOF_RSC_TNI0_MBTABLE4, TOF_RSC_TNI0_MBTABLE5,
+        TOF_RSC_TNI0_MBTABLE6, TOF_RSC_TNI0_MBTABLE7, TOF_RSC_TNI0_MBTABLE8,
+        TOF_RSC_TNI0_MBTABLE9, TOF_RSC_TNI0_MBTABLE10, TOF_RSC_TNI0_MBTABLE11,
+        TOF_RSC_TNI1_MBTABLE0, TOF_RSC_TNI1_MBTABLE1, TOF_RSC_TNI1_MBTABLE2,
+        TOF_RSC_TNI1_MBTABLE3, TOF_RSC_TNI1_MBTABLE4, TOF_RSC_TNI1_MBTABLE5,
+        TOF_RSC_TNI1_MBTABLE6, TOF_RSC_TNI1_MBTABLE7, TOF_RSC_TNI1_MBTABLE8,
+        TOF_RSC_TNI1_MBTABLE9, TOF_RSC_TNI1_MBTABLE10, TOF_RSC_TNI1_MBTABLE11,
+        TOF_RSC_TNI2_MBTABLE0, TOF_RSC_TNI2_MBTABLE1, TOF_RSC_TNI2_MBTABLE2,
+        TOF_RSC_TNI2_MBTABLE3, TOF_RSC_TNI2_MBTABLE4, TOF_RSC_TNI2_MBTABLE5,
+        TOF_RSC_TNI2_MBTABLE6, TOF_RSC_TNI2_MBTABLE7, TOF_RSC_TNI2_MBTABLE8,
+        TOF_RSC_TNI2_MBTABLE9, TOF_RSC_TNI2_MBTABLE10, TOF_RSC_TNI2_MBTABLE11,
+        TOF_RSC_TNI3_MBTABLE0, TOF_RSC_TNI3_MBTABLE1, TOF_RSC_TNI3_MBTABLE2,
+        TOF_RSC_TNI3_MBTABLE3, TOF_RSC_TNI3_MBTABLE4, TOF_RSC_TNI3_MBTABLE5,
+        TOF_RSC_TNI3_MBTABLE6, TOF_RSC_TNI3_MBTABLE7, TOF_RSC_TNI3_MBTABLE8,
+        TOF_RSC_TNI3_MBTABLE9, TOF_RSC_TNI3_MBTABLE10, TOF_RSC_TNI3_MBTABLE11,
+        TOF_RSC_TNI4_MBTABLE0, TOF_RSC_TNI4_MBTABLE1, TOF_RSC_TNI4_MBTABLE2,
+        TOF_RSC_TNI4_MBTABLE3, TOF_RSC_TNI4_MBTABLE4, TOF_RSC_TNI4_MBTABLE5,
+        TOF_RSC_TNI4_MBTABLE6, TOF_RSC_TNI4_MBTABLE7, TOF_RSC_TNI4_MBTABLE8,
+        TOF_RSC_TNI4_MBTABLE9, TOF_RSC_TNI4_MBTABLE10, TOF_RSC_TNI4_MBTABLE11,
+        TOF_RSC_TNI5_MBTABLE0, TOF_RSC_TNI5_MBTABLE1, TOF_RSC_TNI5_MBTABLE2,
+        TOF_RSC_TNI5_MBTABLE3, TOF_RSC_TNI5_MBTABLE4, TOF_RSC_TNI5_MBTABLE5,
+        TOF_RSC_TNI5_MBTABLE6, TOF_RSC_TNI5_MBTABLE7, TOF_RSC_TNI5_MBTABLE8,
+        TOF_RSC_TNI5_MBTABLE9, TOF_RSC_TNI5_MBTABLE10, TOF_RSC_TNI5_MBTABLE11,
+
+        TOF_RSC_NUM     /* 372 */
+};
+#define TOF_RSC_TOQ(TNI, CQID)  (TOF_RSC_TNI0_TOQ0 + (TNI * 12) + CQID)
+#define TOF_RSC_TCQ(TNI, CQID)  (TOF_RSC_TNI0_TCQ0 + (TNI * 12) + CQID)
+#define TOF_RSC_MRQ(TNI, CQID)  (TOF_RSC_TNI0_MRQ0 + (TNI * 12) + CQID)
+#define TOF_RSC_PBQ(TNI)        (TOF_RSC_TNI0_PBQ + TNI)
+#define TOF_RSC_PRQ(TNI)        (TOF_RSC_TNI0_PRQ + TNI)
+#define TOF_RSC_STT(TNI, CQID)  (TOF_RSC_TNI0_STEERINGTABLE0 + (TNI * 12) + CQID)
+#define TOF_RSC_MBT(TNI, CQID)  (TOF_RSC_TNI0_MBTABLE0 + (TNI * 12) + CQID)
+
+#endif
+
+/* vim: set noet ts=8 sw=8 sts=0 tw=0 : */
+

executer/kernel/mcctrl/arch/arm64/include/tofu/tof_utofu_bg_bch.h

@@ -0,0 +1,6 @@
+	struct {
+		bool enabled;
+		uint64_t bgmask[TOF_ICC_NTNIS];
+		uintptr_t iova;
+		void *kaddr;
+	} bch;

executer/kernel/mcctrl/arch/arm64/include/tofu/tof_utofu_cq_trans.h

@@ -0,0 +1,6 @@
+			struct {
+				struct tof_utofu_trans_list *mru;
+				struct tof_trans_table *table;
+				int mruhead;
+				ihk_spinlock_t mru_lock;
+			} trans;

executer/kernel/mcctrl/arch/arm64/include/tofu/tofu_generated-tof_utofu_bg.h

@@ -0,0 +1,21 @@
+struct tof_utofu_bg {
+	union {
+		char whole_struct[160];
+		struct {
+			char padding0[0];
+			struct tof_utofu_device common;
+		};
+		struct {
+			char padding1[80];
+			uint8_t tni;
+		};
+		struct {
+			char padding2[81];
+			uint8_t bgid;
+		};
+		struct {
+			char padding3[88];
+			#include "tof_utofu_bg_bch.h"
+		};
+	};
+};

executer/kernel/mcctrl/arch/arm64/include/tofu/tofu_generated-tof_utofu_cq.h

@@ -0,0 +1,37 @@
+struct tof_utofu_cq {
+	union {
+		char whole_struct[384];
+		struct {
+			char padding0[0];
+			struct tof_utofu_device common;
+		};
+		struct {
+			char padding1[80];
+			uint8_t tni;
+		};
+		struct {
+			char padding2[81];
+			uint8_t cqid;
+		};
+		struct {
+			char padding3[104];
+			#include "tof_utofu_cq_trans.h"
+		};
+		struct {
+			char padding4[128];
+			struct tof_icc_steering_entry *steering;
+		};
+		struct {
+			char padding5[136];
+			struct tof_icc_mb_entry *mb;
+		};
+		struct {
+			char padding6[186];
+			uint8_t num_stag;
+		};
+		struct {
+			char padding7[336];
+			struct mmu_notifier mn;
+		};
+	};
+};

executer/kernel/mcctrl/arch/arm64/include/tofu/tofu_generated-tof_utofu_device.h

@@ -0,0 +1,17 @@
+struct tof_utofu_device {
+	union {
+		char whole_struct[80];
+		struct {
+			char padding0[0];
+			bool enabled;
+		};
+		struct {
+			char padding1[12];
+			uint32_t gpid;
+		};
+		struct {
+			char padding2[24];
+			uint64_t subnet;
+		};
+	};
+};

executer/kernel/mcctrl/arch/arm64/include/tofu/tofu_generated-tof_utofu_mbpt.h

@@ -0,0 +1,33 @@
+struct tof_utofu_mbpt {
+	union {
+		char whole_struct[56];
+		struct {
+			char padding0[0];
+			struct kref kref;
+		};
+		struct {
+			char padding1[8];
+			struct tof_utofu_cq *ucq;
+		};
+		struct {
+			char padding2[16];
+			uintptr_t iova;
+		};
+		struct {
+			char padding3[24];
+			struct scatterlist *sg;
+		};
+		struct {
+			char padding4[32];
+			size_t nsgents;
+		};
+		struct {
+			char padding5[40];
+			uintptr_t mbptstart;
+		};
+		struct {
+			char padding6[48];
+			size_t pgsz;
+		};
+	};
+};

executer/kernel/mcctrl/control.c

@@ -36,6 +36,7 @@
 #include <linux/semaphore.h>
 #include <linux/interrupt.h>
 #include <linux/cpumask.h>
+#include <linux/delay.h>
 #include <asm/uaccess.h>
 #include <asm/delay.h>
 #include <asm/io.h>
@@ -229,6 +230,9 @@ static long mcexec_prepare_image(ihk_os_t os,
 
 	dprintk("%s: pid %d, rpgtable: 0x%lx added\n",
 		__FUNCTION__, ppd->pid, ppd->rpgtable);
+#ifdef ENABLE_TOFU
+	ppd->enable_tofu = pdesc->enable_tofu;
+#endif
 
 	ret = 0;
 
@@ -266,18 +270,24 @@ int mcexec_transfer_image(ihk_os_t os, struct remote_transfer *__user upt)
 		return -EFAULT;
 	}
 
+#ifdef CONFIG_MIC
 	if (pt.size > PAGE_SIZE) {
 		printk("mcexec_transfer_image(): ERROR: size exceeds PAGE_SIZE\n");
 		return -EFAULT;
 	}
-	
+
 	phys = ihk_device_map_memory(ihk_os_to_dev(os), pt.rphys, PAGE_SIZE);
-#ifdef CONFIG_MIC
 	rpm = ioremap_wc(phys, PAGE_SIZE);
 #else
-	rpm = ihk_device_map_virtual(ihk_os_to_dev(os), phys, PAGE_SIZE, NULL, 0);
+	phys = ihk_device_map_memory(ihk_os_to_dev(os), pt.rphys, pt.size);
+	rpm = ihk_device_map_virtual(ihk_os_to_dev(os), phys, pt.size, NULL, 0);
 #endif
-	
+
+	if (!rpm) {
+		pr_err("%s(): error: invalid remote address\n", __func__);
+		return -EFAULT;
+	}
+
 	if (pt.direction == MCEXEC_UP_TRANSFER_TO_REMOTE) {
 		if (copy_from_user(rpm, pt.userp, pt.size)) {
 			ret = -EFAULT;
@@ -295,10 +305,11 @@ int mcexec_transfer_image(ihk_os_t os, struct remote_transfer *__user upt)
 
 #ifdef CONFIG_MIC
 	iounmap(rpm);
+	ihk_device_unmap_memory(ihk_os_to_dev(os), phys, PAGE_SIZE);
 #else
-	ihk_device_unmap_virtual(ihk_os_to_dev(os), rpm, PAGE_SIZE);
+	ihk_device_unmap_virtual(ihk_os_to_dev(os), rpm, pt.size);
+	ihk_device_unmap_memory(ihk_os_to_dev(os), phys, pt.size);
 #endif
-	ihk_device_unmap_memory(ihk_os_to_dev(os), phys, PAGE_SIZE);	
 
 	return ret;
 
@@ -378,6 +389,7 @@ static void release_handler(ihk_os_t os, void *param)
 	int os_ind = ihk_host_os_get_index(os);
 	unsigned long flags;
 	struct host_thread *thread;
+	int ret;
 
 	/* Finalize FS switch for uti threads */ 
 	write_lock_irqsave(&host_thread_lock, flags);
@@ -399,7 +411,13 @@ static void release_handler(ihk_os_t os, void *param)
 
 	dprintk("%s: SCD_MSG_CLEANUP_PROCESS, info: %p, cpu: %d\n",
 			__FUNCTION__, info, info->cpu);
-	mcctrl_ikc_send(os, info->cpu, &isp);
+	ret = mcctrl_ikc_send_wait(os, info->cpu,
+			&isp, -20, NULL, NULL, 0);
+	if (ret != 0) {
+		printk("%s: WARNING: failed to send IKC msg: %d\n",
+				__func__, ret);
+	}
+
 	if (os_ind >= 0) {
 		delete_pid_entry(os_ind, info->pid);
 	}
@@ -587,13 +605,14 @@ extern int mckernel_cpu_2_linux_cpu(struct mcctrl_usrdata *udp, int cpu_id);
 static long mcexec_get_cpuset(ihk_os_t os, unsigned long arg)
 {
 	struct mcctrl_usrdata *udp = ihk_host_os_get_usrdata(os);
-	struct mcctrl_part_exec *pe;
+	struct mcctrl_part_exec *pe = NULL, *pe_itr;
 	struct get_cpu_set_arg req;
 	struct mcctrl_cpu_topology *cpu_top, *cpu_top_i;
 	struct cache_topology *cache_top;
 	int cpu, cpus_assigned, cpus_to_assign, cpu_prev;
 	int ret = 0;
 	int mcexec_linux_numa;
+	int pe_list_len = 0;
 	cpumask_t *mcexec_cpu_set = NULL;
 	cpumask_t *cpus_used = NULL;
 	cpumask_t *cpus_to_use = NULL;
@@ -613,24 +632,126 @@ static long mcexec_get_cpuset(ihk_os_t os, unsigned long arg)
 		return -EINVAL;
 	}
 
-	pe = &udp->part_exec;
-
-	mutex_lock(&pe->lock);
-
 	if (copy_from_user(&req, (void *)arg, sizeof(req))) {
-		printk("%s: error copying user request\n", __FUNCTION__);
+		pr_err("%s: error copying user request\n", __func__);
 		ret = -EINVAL;
-		goto put_and_unlock_out;
+		goto put_out;
 	}
 
-	/* First process to enter CPU partitioning */
-	if (pe->nr_processes == -1) {
+	/* User requested CPU mask? */
+	if (req.req_cpu_list && req.req_cpu_list_len) {
+		char *cpu_list = NULL;
+
+		cpu_list = kmalloc(req.req_cpu_list_len, GFP_KERNEL);
+		if (!cpu_list) {
+			printk("%s: error: allocating CPU list\n", __FUNCTION__);
+			ret = -ENOMEM;
+			goto put_out;
+		}
+
+		if (copy_from_user(cpu_list,
+					req.req_cpu_list, req.req_cpu_list_len)) {
+			printk("%s: error copying CPU list request\n", __FUNCTION__);
+			kfree(cpu_list);
+			ret = -EINVAL;
+			goto put_out;
+		}
+
+		cpus_used = kmalloc(sizeof(cpumask_t), GFP_KERNEL);
+		cpus_to_use = kmalloc(sizeof(cpumask_t), GFP_KERNEL);
+		if (!cpus_to_use || !cpus_used) {
+			printk("%s: error: allocating CPU mask\n", __FUNCTION__);
+			ret = -ENOMEM;
+			kfree(cpu_list);
+			goto put_out;
+		}
+		memset(cpus_used, 0, sizeof(cpumask_t));
+		memset(cpus_to_use, 0, sizeof(cpumask_t));
+
+		/* Parse CPU list */
+		if (cpulist_parse(cpu_list, cpus_to_use) < 0) {
+			printk("%s: invalid CPUs requested: %s\n",
+				__FUNCTION__, cpu_list);
+			ret = -EINVAL;
+			kfree(cpu_list);
+			goto put_out;
+		}
+
+		memcpy(cpus_used, cpus_to_use, sizeof(cpumask_t));
+
+		/* Copy mask to user-space */
+		if (copy_to_user(req.cpu_set, cpus_used,
+					(req.cpu_set_size < sizeof(cpumask_t) ?
+					 req.cpu_set_size : sizeof(cpumask_t)))) {
+			printk("%s: error copying mask to user\n", __FUNCTION__);
+			ret = -EINVAL;
+			kfree(cpu_list);
+			goto put_out;
+		}
+
+		/* Copy IKC target core */
+		cpu = cpumask_next(-1, cpus_used);
+		if (copy_to_user(req.target_core, &cpu, sizeof(cpu))) {
+			printk("%s: error copying target core to user\n",
+					__FUNCTION__);
+			ret = -EINVAL;
+			kfree(cpu_list);
+			goto put_out;
+		}
+
+		/* Save in per-process structure */
+		memcpy(&ppd->cpu_set, cpus_used, sizeof(cpumask_t));
+		ppd->ikc_target_cpu = cpu;
+		printk("%s: %s -> target McKernel CPU: %d\n",
+			__func__, cpu_list, cpu);
+
+		ret = 0;
+		kfree(cpu_list);
+		goto put_out;
+	}
+
+	mutex_lock(&udp->part_exec_lock);
+	/* Find part_exec having same node_proxy */
+	list_for_each_entry_reverse(pe_itr, &udp->part_exec_list, chain) {
+		pe_list_len++;
+		if (pe_itr->node_proxy_pid == req.ppid) {
+			pe = pe_itr;
+			break;
+		}
+	}
+
+	if (!pe) {
+		/* First process to enter CPU partitioning */
+		pr_debug("%s: pe_list_len:%d\n", __func__, pe_list_len);
+		if (pe_list_len >= PE_LIST_MAXLEN) {
+			/* delete head entry of pe_list */
+			pe_itr = list_first_entry(&udp->part_exec_list,
+					struct mcctrl_part_exec, chain);
+			list_del(&pe_itr->chain);
+			kfree(pe_itr);
+		}
+
+		pe = kzalloc(sizeof(struct mcctrl_part_exec), GFP_KERNEL);
+		if (!pe) {
+			mutex_unlock(&udp->part_exec_lock);
+			ret = -ENOMEM;
+			goto put_out;
+		}
+		/* Init part_exec */
+		mutex_init(&pe->lock);
+		INIT_LIST_HEAD(&pe->pli_list);
 		pe->nr_processes = req.nr_processes;
 		pe->nr_processes_left = req.nr_processes;
+		pe->nr_processes_joined = 0;
+		pe->node_proxy_pid = req.ppid;
+
+		list_add_tail(&pe->chain, &udp->part_exec_list);
 		dprintk("%s: nr_processes: %d (partitioned exec starts)\n",
-				__FUNCTION__,
-				pe->nr_processes);
+				__func__, pe->nr_processes);
 	}
+	mutex_unlock(&udp->part_exec_lock);
+
+	mutex_lock(&pe->lock);
 
 	if (pe->nr_processes != req.nr_processes) {
 		printk("%s: error: requested number of processes"
@@ -640,7 +761,15 @@ static long mcexec_get_cpuset(ihk_os_t os, unsigned long arg)
 		goto put_and_unlock_out;
 	}
 
+	if (pe->nr_processes_joined >= pe->nr_processes) {
+		printk("%s: too many processes have joined to the group of %d\n",
+				__func__, req.ppid);
+		ret = -EINVAL;
+		goto put_and_unlock_out;
+	}
+
 	--pe->nr_processes_left;
+	++pe->nr_processes_joined;
 	dprintk("%s: nr_processes: %d, nr_processes_left: %d\n",
 			__FUNCTION__,
 			pe->nr_processes,
@@ -726,8 +855,6 @@ static long mcexec_get_cpuset(ihk_os_t os, unsigned long arg)
 				wake_up_interruptible(&pli_next->pli_wq);
 			}
 
-			/* Reset process counter to start state */
-			pe->nr_processes = -1;
 			ret = -ETIMEDOUT;
 			goto put_and_unlock_out;
 		}
@@ -975,16 +1102,8 @@ next_cpu:
 	/* Commit used cores to OS structure */
 	memcpy(&pe->cpus_used, cpus_used, sizeof(*cpus_used));
 
-	/* Reset if last process */
-	if (pe->nr_processes_left == 0) {
-		dprintk("%s: nr_processes: %d (partitioned exec ends)\n",
-				__FUNCTION__,
-				pe->nr_processes);
-		pe->nr_processes = -1;
-		memset(&pe->cpus_used, 0, sizeof(pe->cpus_used));
-	}
-	/* Otherwise wake up next process in list */
-	else {
+	/* If not last process, wake up next process in list */
+	if (pe->nr_processes_left != 0) {
 		++pe->process_rank;
 		pli_next = list_first_entry(&pe->pli_list,
 			struct process_list_item, list);
@@ -997,11 +1116,14 @@ next_cpu:
 	ret = 0;
 
 put_and_unlock_out:
+	mutex_unlock(&pe->lock);
+
+put_out:
+	mcctrl_put_per_proc_data(ppd);
+
 	kfree(cpus_to_use);
 	kfree(cpus_used);
 	kfree(mcexec_cpu_set);
-	mcctrl_put_per_proc_data(ppd);
-	mutex_unlock(&pe->lock);
 
 	return ret;
 }
@@ -1146,7 +1268,7 @@ void mcctrl_put_per_proc_data(struct mcctrl_per_proc_data *ppd)
 			   process is gone and the application should be terminated. */
 			packet = (struct ikc_scd_packet *)ptd->data;
 			dprintk("%s: calling __return_syscall (hash),target pid=%d,tid=%d\n", __FUNCTION__, ppd->pid, packet->req.rtid);
-			__return_syscall(ppd->ud->os, packet, -ERESTARTSYS,
+			__return_syscall(ppd->ud->os, ppd, packet, -ERESTARTSYS,
 					 packet->req.rtid);
 			ihk_ikc_release_packet((struct ihk_ikc_free_packet *)packet);
 
@@ -1170,7 +1292,7 @@ void mcctrl_put_per_proc_data(struct mcctrl_per_proc_data *ppd)
 
 		/* We use ERESTARTSYS to tell the LWK that the proxy
 		 * process is gone and the application should be terminated */
-		__return_syscall(ppd->ud->os, packet, -ERESTARTSYS,
+		__return_syscall(ppd->ud->os, ppd, packet, -ERESTARTSYS,
 				packet->req.rtid);
 		ihk_ikc_release_packet((struct ihk_ikc_free_packet *)packet);
 	}
@@ -1211,7 +1333,7 @@ int mcexec_syscall(struct mcctrl_usrdata *ud, struct ikc_scd_packet *packet)
 
 		/* We use ERESTARTSYS to tell the LWK that the proxy
 		 * process is gone and the application should be terminated */
-		__return_syscall(ud->os, packet, -ERESTARTSYS,
+		__return_syscall(ud->os, NULL, packet, -ERESTARTSYS,
 				packet->req.rtid);
 		ihk_ikc_release_packet((struct ihk_ikc_free_packet *)packet);
 
@@ -1654,7 +1776,7 @@ long mcexec_ret_syscall(ihk_os_t os, struct syscall_ret_desc *__user arg)
 		ihk_device_unmap_memory(ihk_os_to_dev(os), phys, ret.size);
 	} 
 
-	__return_syscall(os, packet, ret.ret, task_pid_vnr(current));
+	__return_syscall(os, ppd, packet, ret.ret, task_pid_vnr(current));
 
 	error = 0;
 out:
@@ -2097,7 +2219,13 @@ static DECLARE_WAIT_QUEUE_HEAD(perfctrlq);
 
 long mcctrl_perf_num(ihk_os_t os, unsigned long arg)
 {
-	struct mcctrl_usrdata *usrdata = ihk_host_os_get_usrdata(os);
+	struct mcctrl_usrdata *usrdata;
+
+	if (!os || ihk_host_validate_os(os)) {
+		return -EINVAL;
+	}
+
+	usrdata = ihk_host_os_get_usrdata(os);
 
 	if (!usrdata) {
 		pr_err("%s: error: mcctrl_usrdata not found\n", __func__);
@@ -2122,22 +2250,34 @@ struct mcctrl_perf_ctrl_desc {
  */
 long mcctrl_perf_set(ihk_os_t os, struct ihk_perf_event_attr *__user arg)
 {
-	struct mcctrl_usrdata *usrdata = ihk_host_os_get_usrdata(os);
+	struct mcctrl_usrdata *usrdata = NULL;
 	struct ikc_scd_packet isp;
 	struct perf_ctrl_desc *perf_desc;
 	struct ihk_perf_event_attr attr;
-	struct ihk_cpu_info *info = ihk_os_get_cpu_info(os);
+	struct ihk_cpu_info *info = NULL;
 	int ret = 0;
 	int i = 0, j = 0;
 	int need_free;
 	int num_registered = 0;
 	int err = 0;
 
+	if (!os || ihk_host_validate_os(os)) {
+		return -EINVAL;
+	}
+
+	usrdata = ihk_host_os_get_usrdata(os);
+
 	if (!usrdata) {
 		pr_err("%s: error: mcctrl_usrdata not found\n", __func__);
 		return -EINVAL;
 	}
 
+	info = ihk_os_get_cpu_info(os);
+	if (!info) {
+		pr_err("%s: error: cannot get cpu info\n", __func__);
+		return -EINVAL;
+	}
+
 	for (i = 0; i < usrdata->perf_event_num; i++) {
 		ret = copy_from_user(&attr, &arg[i],
 				     sizeof(struct ihk_perf_event_attr));
@@ -2197,20 +2337,30 @@ long mcctrl_perf_set(ihk_os_t os, struct ihk_perf_event_attr *__user arg)
 
 long mcctrl_perf_get(ihk_os_t os, unsigned long *__user arg)
 {
-	struct mcctrl_usrdata *usrdata = ihk_host_os_get_usrdata(os);
+	struct mcctrl_usrdata *usrdata = NULL;
 	struct ikc_scd_packet isp;
 	struct perf_ctrl_desc *perf_desc;
-	struct ihk_cpu_info *info = ihk_os_get_cpu_info(os);
+	struct ihk_cpu_info *info = NULL;
 	unsigned long value_sum = 0;
 	int ret = 0;
 	int i = 0, j = 0;
 	int need_free;
 
+	if (!os || ihk_host_validate_os(os)) {
+		return -EINVAL;
+	}
+
+	usrdata = ihk_host_os_get_usrdata(os);
 	if (!usrdata) {
 		pr_err("%s: error: mcctrl_usrdata not found\n", __func__);
 		return -EINVAL;
 	}
 
+	info = ihk_os_get_cpu_info(os);
+	if (!info || info->n_cpus < 1) {
+		return -EINVAL;
+	}
+
 	for (i = 0; i < usrdata->perf_event_num; i++) {
 		perf_desc = kmalloc(sizeof(struct mcctrl_perf_ctrl_desc),
 				    GFP_KERNEL);
@@ -2258,15 +2408,20 @@ long mcctrl_perf_get(ihk_os_t os, unsigned long *__user arg)
 
 long mcctrl_perf_enable(ihk_os_t os)
 {
-	struct mcctrl_usrdata *usrdata = ihk_host_os_get_usrdata(os);
+	struct mcctrl_usrdata *usrdata = NULL;
 	struct ikc_scd_packet isp;
 	struct perf_ctrl_desc *perf_desc;
-	struct ihk_cpu_info *info = ihk_os_get_cpu_info(os);
+	struct ihk_cpu_info *info = NULL;
 	unsigned long cntr_mask = 0;
 	int ret = 0;
 	int i = 0, j = 0;
 	int need_free;
 
+	if (!os || ihk_host_validate_os(os)) {
+		return -EINVAL;
+	}
+
+	usrdata = ihk_host_os_get_usrdata(os);
 	if (!usrdata) {
 		pr_err("%s: error: mcctrl_usrdata not found\n", __func__);
 		return -EINVAL;
@@ -2289,6 +2444,11 @@ long mcctrl_perf_enable(ihk_os_t os)
 	isp.msg = SCD_MSG_PERF_CTRL;
 	isp.arg = virt_to_phys(perf_desc);
 
+	info = ihk_os_get_cpu_info(os);
+	if (!info || info->n_cpus < 1) {
+		kfree(perf_desc);
+		return -EINVAL;
+	}
 	for (j = 0; j < info->n_cpus; j++) {
 		ret = mcctrl_ikc_send_wait(os, j, &isp, 0,
 					   wakeup_desc_of_perf_desc(perf_desc),
@@ -2316,15 +2476,20 @@ long mcctrl_perf_enable(ihk_os_t os)
 
 long mcctrl_perf_disable(ihk_os_t os)
 {
-	struct mcctrl_usrdata *usrdata = ihk_host_os_get_usrdata(os);
+	struct mcctrl_usrdata *usrdata = NULL;
 	struct ikc_scd_packet isp;
 	struct perf_ctrl_desc *perf_desc;
-	struct ihk_cpu_info *info = ihk_os_get_cpu_info(os);
+	struct ihk_cpu_info *info = NULL;
 	unsigned long cntr_mask = 0;
 	int ret = 0;
 	int i = 0, j = 0;
 	int need_free;
 
+	if (!os || ihk_host_validate_os(os)) {
+		return -EINVAL;
+	}
+
+	usrdata = ihk_host_os_get_usrdata(os);
 	if (!usrdata) {
 		pr_err("%s: error: mcctrl_usrdata not found\n", __func__);
 		return -EINVAL;
@@ -2347,6 +2512,11 @@ long mcctrl_perf_disable(ihk_os_t os)
 	isp.msg = SCD_MSG_PERF_CTRL;
 	isp.arg = virt_to_phys(perf_desc);
 
+	info = ihk_os_get_cpu_info(os);
+	if (!info || info->n_cpus < 1) {
+		kfree(perf_desc);
+		return -EINVAL;
+	}
 	for (j = 0; j < info->n_cpus; j++) {
 		ret = mcctrl_ikc_send_wait(os, j, &isp, 0,
 				wakeup_desc_of_perf_desc(perf_desc),
@@ -2388,6 +2558,10 @@ long mcctrl_getrusage(ihk_os_t ihk_os, struct mcctrl_ioctl_getrusage_desc *__use
 	unsigned long ut;
 	unsigned long st;
 
+	if (!ihk_os || ihk_host_validate_os(ihk_os)) {
+		return -EINVAL;
+	}
+
 	ret = copy_from_user(&desc, _desc, sizeof(struct mcctrl_ioctl_getrusage_desc));
 	if (ret != 0) {
 		printk("%s: copy_from_user failed\n", __FUNCTION__);
@@ -2630,7 +2804,7 @@ static long mcexec_terminate_thread_unsafe(ihk_os_t os, int pid, int tid, long c
 				__FUNCTION__, tid);
 		goto no_ptd;
 	}
-	__return_syscall(usrdata->os, packet, code, tid);
+	__return_syscall(usrdata->os, ppd, packet, code, tid);
 	ihk_ikc_release_packet((struct ihk_ikc_free_packet *)packet);
 
 	/* Drop reference for this function */
@@ -3376,7 +3550,7 @@ int mcctrl_get_request_os_cpu(ihk_os_t os, int *ret_cpu)
 	struct ihk_ikc_channel_desc *ch;
 	int ret = 0;
 
-	if (!os) {
+	if (!os || ihk_host_validate_os(os) || !ret_cpu) {
 		return -EINVAL;
 	}
 
@@ -3418,7 +3592,11 @@ int mcctrl_get_request_os_cpu(ihk_os_t os, int *ret_cpu)
 	*ret_cpu = ch->send.queue->read_cpu;
 	ret = 0;
 
+#ifndef ENABLE_FUGAKU_HACKS
 	pr_info("%s: OS: %lx, CPU: %d\n",
+#else
+	dprintk("%s: OS: %lx, CPU: %d\n",
+#endif
 		__func__, (unsigned long)os, *ret_cpu);
 
 out_put_ppd:
@@ -3468,7 +3646,8 @@ int __mcctrl_os_read_write_cpu_register(ihk_os_t os, int cpu,
 	isp.op = op;
 	isp.pdesc = virt_to_phys(ldesc);
 
-	ret = mcctrl_ikc_send_wait(os, cpu, &isp, 0, NULL, &do_free, 1, ldesc);
+	/* 1 sec timeout for the case where McKernel can't respond */
+	ret = mcctrl_ikc_send_wait(os, cpu, &isp, 1000, NULL, &do_free, 1, ldesc);
 	if (ret != 0) {
 		printk("%s: ERROR sending IKC msg: %d\n", __FUNCTION__, ret);
 		goto out;
@@ -3482,7 +3661,11 @@ int __mcctrl_os_read_write_cpu_register(ihk_os_t os, int cpu,
 	/* Notify caller (for future async implementation) */
 	atomic_set(&desc->sync, 1);
 
+#ifndef ENABLE_FUGAKU_HACKS
 	dprintk("%s: MCCTRL_OS_CPU_%s_REGISTER: CPU: %d, addr_ext: 0x%lx, val: 0x%lx\n",
+#else
+	printk("%s: MCCTRL_OS_CPU_%s_REGISTER: CPU: %d, addr_ext: 0x%lx, val: 0x%lx\n",
+#endif
 		__FUNCTION__,
 		(op == MCCTRL_OS_CPU_READ_REGISTER ? "READ" : "WRITE"), cpu,
 		desc->addr_ext, desc->val);

executer/kernel/mcctrl/driver.c

@@ -50,6 +50,9 @@ extern void procfs_exit(int);
 extern void uti_attr_finalize(void);
 extern void binfmt_mcexec_init(void);
 extern void binfmt_mcexec_exit(void);
+#ifdef ENABLE_TOFU
+extern void mcctrl_file_to_pidfd_hash_init(void);
+#endif
 
 extern int mcctrl_os_read_cpu_register(ihk_os_t os, int cpu,
 		struct ihk_os_cpu_register *desc);
@@ -57,6 +60,11 @@ extern int mcctrl_os_write_cpu_register(ihk_os_t os, int cpu,
 		struct ihk_os_cpu_register *desc);
 extern int mcctrl_get_request_os_cpu(ihk_os_t os, int *cpu);
 
+#ifdef ENABLE_TOFU
+extern void mcctrl_tofu_hijack_release_handlers(void);
+extern void mcctrl_tofu_restore_release_handlers(void);
+#endif
+
 static long mcctrl_ioctl(ihk_os_t os, unsigned int request, void *priv,
                          unsigned long arg, struct file *file)
 {
@@ -227,7 +235,6 @@ void (*mcctrl_zap_page_range)(struct vm_area_struct *vma,
 
 struct inode_operations *mcctrl_hugetlbfs_inode_operations;
 
-
 static int symbols_init(void)
 {
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(4,17,0)
@@ -319,10 +326,17 @@ static int __init mcctrl_init(void)
 	}
 
 	binfmt_mcexec_init();
+#ifdef ENABLE_TOFU
+	mcctrl_file_to_pidfd_hash_init();
+#endif
 
 	if ((ret = symbols_init()))
 		goto error;
 
+#ifdef ENABLE_TOFU
+	mcctrl_tofu_hijack_release_handlers();
+#endif
+
 	if ((ret = ihk_host_register_os_notifier(&mcctrl_os_notifier)) != 0) {
 		printk("mcctrl: error: registering OS notifier\n");
 		goto error;
@@ -345,6 +359,9 @@ static void __exit mcctrl_exit(void)
 
 	binfmt_mcexec_exit();
 	uti_attr_finalize();
+#ifdef ENABLE_TOFU
+	mcctrl_tofu_restore_release_handlers();
+#endif
 
 	printk("mcctrl: unregistered.\n");
 }

executer/kernel/mcctrl/ikc.c

@@ -142,13 +142,35 @@ int mcctrl_ikc_send_wait(ihk_os_t os, int cpu, struct ikc_scd_packet *pisp,
 	ret = mcctrl_ikc_send(os, cpu, pisp);
 	if (ret < 0) {
 		pr_warn("%s: mcctrl_ikc_send failed: %d\n", __func__, ret);
-		kfree(desc);
+		if (alloc_desc)
+			kfree(desc);
 		return ret;
 	}
 
 	if (timeout) {
-		ret = wait_event_interruptible_timeout(desc->wq,
-			desc->status, timeout);
+		/*
+		 * Negative timeout indicates busy waiting, which can be used
+		 * in situations where wait_event_interruptible_XXX() would
+		 * fail, e.g., in a signal handler, at the time the process
+		 * is being killed, etc.
+		 */
+		if (timeout < 0) {
+			unsigned long timeout_jiffies =
+				jiffies + msecs_to_jiffies(timeout * -1);
+			ret = -ETIME;
+
+			while (time_before(jiffies, timeout_jiffies)) {
+				schedule();
+				if (READ_ONCE(desc->status)) {
+					ret = 0;
+					break;
+				}
+			}
+		}
+		else {
+			ret = wait_event_interruptible_timeout(desc->wq,
+					desc->status, msecs_to_jiffies(timeout));
+		}
 	} else {
 		ret = wait_event_interruptible(desc->wq, desc->status);
 	}
@@ -210,6 +232,8 @@ static int syscall_packet_handler(struct ihk_ikc_channel_desc *c,
 	case SCD_MSG_PROCFS_ANSWER:
 	case SCD_MSG_REMOTE_PAGE_FAULT_ANSWER:
 	case SCD_MSG_CPU_RW_REG_RESP:
+	case SCD_MSG_CLEANUP_PROCESS_RESP:
+	case SCD_MSG_CLEANUP_FD_RESP:
 		mcctrl_wakeup_cb(__os, pisp);
 		break;
 
@@ -280,7 +304,11 @@ int mcctrl_ikc_send(ihk_os_t os, int cpu, struct ikc_scd_packet *pisp)
 {
 	struct mcctrl_usrdata *usrdata;
 
-	if (!os || cpu < 0) {
+	if (!os || ihk_host_validate_os(os) || !pisp) {
+		return -EINVAL;
+	}
+
+	if (cpu < 0) {
 		return -EINVAL;
 	}
 
@@ -508,11 +536,9 @@ int prepare_ikc_channels(ihk_os_t os)
 	usrdata->os = os;
 	ihk_host_os_set_usrdata(os, usrdata);
 
-	ihk_ikc_listen_port(os, &lp_ikc2linux);
-	ihk_ikc_listen_port(os, &lp_ikc2mckernel);
-
 	init_waitqueue_head(&usrdata->wq_procfs);
 	mutex_init(&usrdata->reserve_lock);
+	mutex_init(&usrdata->part_exec_lock);
 
 	for (i = 0; i < MCCTRL_PER_PROC_DATA_HASH_SIZE; ++i) {
 		INIT_LIST_HEAD(&usrdata->per_proc_data_hash[i]);
@@ -521,13 +547,21 @@ int prepare_ikc_channels(ihk_os_t os)
 
 	INIT_LIST_HEAD(&usrdata->cpu_topology_list);
 	INIT_LIST_HEAD(&usrdata->node_topology_list);
+	INIT_LIST_HEAD(&usrdata->part_exec_list);
 
-	mutex_init(&usrdata->part_exec.lock);
-	INIT_LIST_HEAD(&usrdata->part_exec.pli_list);
-	usrdata->part_exec.nr_processes = -1;
 	INIT_LIST_HEAD(&usrdata->wakeup_descs_list);
 	spin_lock_init(&usrdata->wakeup_descs_lock);
 
+	/* ihk_ikc_listen_port should be performed after
+	 * usrdata->cpu_topology_list is initialized because the
+	 * function enables syscall_packet_handler which accesses
+	 * the list (the call path is sysfsm_packet_handler -->
+	 * sysfsm_work_main --> sysfsm_setup --> setup_sysfs_files
+	 * --> setup_cpus_sysfs_files).
+	 */
+	ihk_ikc_listen_port(os, &lp_ikc2linux);
+	ihk_ikc_listen_port(os, &lp_ikc2mckernel);
+
 	return 0;
 
 error:
@@ -580,6 +614,18 @@ void destroy_ikc_channels(ihk_os_t os)
 
 	kfree(usrdata->channels);
 	kfree(usrdata->ikc2linux);
+
+	mutex_lock(&usrdata->part_exec_lock);
+	while (!list_empty(&usrdata->part_exec_list)) {
+		struct mcctrl_part_exec *pe;
+
+		pe = list_first_entry(&usrdata->part_exec_list,
+				struct mcctrl_part_exec, chain);
+		list_del(&pe->chain);
+		kfree(pe);
+	}
+	mutex_unlock(&usrdata->part_exec_lock);
+
 	kfree(usrdata);
 }
 

executer/kernel/mcctrl/mcctrl.h

@@ -58,7 +58,8 @@
 #define SCD_MSG_SEND_SIGNAL             0x7
 #define SCD_MSG_SEND_SIGNAL_ACK         0x8
 #define SCD_MSG_CLEANUP_PROCESS         0x9
-#define SCD_MSG_GET_VDSO_INFO           0xa
+#define SCD_MSG_CLEANUP_PROCESS_RESP    0xa
+#define SCD_MSG_GET_VDSO_INFO           0xb
 
 //#define SCD_MSG_GET_CPU_MAPPING         0xc
 //#define SCD_MSG_REPLY_GET_CPU_MAPPING   0xd
@@ -104,6 +105,8 @@
 
 #define SCD_MSG_CPU_RW_REG              0x52
 #define SCD_MSG_CPU_RW_REG_RESP         0x53
+#define SCD_MSG_CLEANUP_FD              0x54
+#define SCD_MSG_CLEANUP_FD_RESP         0x55
 
 #define SCD_MSG_FUTEX_WAKE              0x60
 
@@ -260,6 +263,7 @@ struct mcctrl_per_proc_data {
 
 	struct list_head devobj_pager_list;
 	struct semaphore devobj_pager_lock;
+	int enable_tofu;
 };
 
 struct sysfsm_req {
@@ -324,13 +328,20 @@ struct process_list_item {
 	wait_queue_head_t pli_wq;
 };
 
+#define PE_LIST_MAXLEN 5
+
 struct mcctrl_part_exec {
 	struct mutex lock;	
 	int nr_processes;
+	/* number of processes to let in / out the synchronization point */
 	int nr_processes_left;
+	/* number of processes which have joined the partition */
+	int nr_processes_joined;
 	int process_rank;
+	pid_t node_proxy_pid;
 	cpumask_t cpus_used;
 	struct list_head pli_list;
+	struct list_head chain;
 };
 
 #define CPU_LONGS (((NR_CPUS) + (BITS_PER_LONG) - 1) / (BITS_PER_LONG))
@@ -353,6 +364,7 @@ struct mcctrl_usrdata {
 	int	job_pos;
 	int	mcctrl_dma_abort;
 	struct mutex reserve_lock;
+	struct mutex part_exec_lock;
 	unsigned long	last_thread_exec;
 	wait_queue_head_t wq_procfs;
 	struct list_head per_proc_data_hash[MCCTRL_PER_PROC_DATA_HASH_SIZE];
@@ -368,7 +380,7 @@ struct mcctrl_usrdata {
 	nodemask_t numa_online;
 	struct list_head cpu_topology_list;
 	struct list_head node_topology_list;
-	struct mcctrl_part_exec part_exec;
+	struct list_head part_exec_list;
 	int perf_event_num;
 };
 
@@ -453,7 +465,8 @@ struct mcctrl_per_thread_data *mcctrl_get_per_thread_data(struct mcctrl_per_proc
 							  struct task_struct *task);
 int mcctrl_clear_pte_range(uintptr_t start, uintptr_t len);
 
-void __return_syscall(ihk_os_t os, struct ikc_scd_packet *packet, 
+void __return_syscall(ihk_os_t os, struct mcctrl_per_proc_data *ppd,
+		struct ikc_scd_packet *packet,
 		long ret, int stid);
 int clear_pte_range(uintptr_t start, uintptr_t len);
 
@@ -548,4 +561,34 @@ struct uti_futex_resp {
 	int done;
 	wait_queue_head_t wq;
 };
+
+#ifdef ENABLE_TOFU
+/*
+ * Hash table to keep track of files and related processes
+ * and file descriptors.
+ * NOTE: Used for Tofu driver release handlers.
+ */
+#define MCCTRL_FILE_2_PIDFD_HASH_SHIFT 4
+#define MCCTRL_FILE_2_PIDFD_HASH_SIZE (1 << MCCTRL_FILE_2_PIDFD_HASH_SHIFT)
+#define MCCTRL_FILE_2_PIDFD_HASH_MASK (MCCTRL_FILE_2_PIDFD_HASH_SIZE - 1)
+
+struct mcctrl_file_to_pidfd {
+	struct file *filp;
+	ihk_os_t os;
+	struct task_struct *group_leader;
+	int pid;
+	int fd;
+	struct list_head hash;
+	char tofu_dev_path[128];
+	void *pde_data;
+};
+
+int mcctrl_file_to_pidfd_hash_insert(struct file *filp,
+	ihk_os_t os, int pid, struct task_struct *group_leader, int fd,
+	char *path, void *pde_data);
+struct mcctrl_file_to_pidfd *mcctrl_file_to_pidfd_hash_lookup(
+	struct file *filp, struct task_struct *group_leader);
+int mcctrl_file_to_pidfd_hash_remove(struct file *filp,
+	ihk_os_t os, struct task_struct *group_leader, int fd);
+#endif
 #endif

executer/kernel/mcctrl/procfs.c

@@ -126,7 +126,7 @@ find_procfs_entry(struct procfs_list_entry *parent, const char *name)
 static void
 delete_procfs_entries(struct procfs_list_entry *top)
 {
-	struct procfs_list_entry *e;
+	struct procfs_list_entry *e = NULL;
 	struct procfs_list_entry *n;
 
 	list_del(&top->list);
@@ -136,8 +136,10 @@ delete_procfs_entries(struct procfs_list_entry *top)
 	}
 
 #if LINUX_VERSION_CODE < KERNEL_VERSION(3,10,0)
-	e->entry->read_proc = NULL;
-	e->entry->data = NULL;
+	if (e) {
+		e->entry->read_proc = NULL;
+		e->entry->data = NULL;
+	}
 #endif
 	remove_proc_entry(top->name, top->parent? top->parent->entry: NULL);
 	if(top->data)

executer/kernel/mcctrl/syscall.c

@@ -45,6 +45,9 @@
 #include <linux/mount.h>
 #include <linux/kdev_t.h>
 #include <linux/hugetlb.h>
+#include <linux/proc_fs.h>
+#include <linux/rbtree.h>
+#include <linux/llist.h>
 #include <asm/uaccess.h>
 #include <asm/delay.h>
 #include <asm/io.h>
@@ -52,6 +55,7 @@
 #include "mcctrl.h"
 #include <linux/version.h>
 #include <archdeps.h>
+#include <asm/pgtable.h>
 
 #define ALIGN_WAIT_BUF(z)   (((z + 63) >> 6) << 6)
 
@@ -539,7 +543,11 @@ retry_alloc:
 #define	USE_VM_INSERT_PFN	1
 
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 11, 0)
+#if defined(RHEL_RELEASE_CODE) && RHEL_RELEASE_CODE >= RHEL_RELEASE_VERSION(8, 2)
+static vm_fault_t rus_vm_fault(struct vm_fault *vmf)
+#else
 static int rus_vm_fault(struct vm_fault *vmf)
+#endif
 {
 	struct vm_area_struct *vma = vmf->vma;
 #else
@@ -651,6 +659,9 @@ static int rus_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 		goto put_and_out;
 	}
 
+	// Force regular page size
+	pgsize = PAGE_SIZE;
+
 	rva = (unsigned long)addr & ~(pgsize - 1);
 	rpa = rpa & ~(pgsize - 1);
 
@@ -662,7 +673,8 @@ static int rus_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 
 		/* LWK may hold large page based mappings that align rva outside
 		 * Linux' VMA, make sure we don't try to map to those pages */
-		if (rva + (pix * PAGE_SIZE) < vma->vm_start) {
+		if (rva + (pix * PAGE_SIZE) < vma->vm_start ||
+			rva + (pix * PAGE_SIZE) > vma->vm_end) {
 			continue;
 		}
 
@@ -673,21 +685,27 @@ static int rus_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 			if (error) {
 				pr_err("%s: error inserting mapping for 0x%#lx "
 					"(req: TID: %d, syscall: %lu) error: %d,"
-					" vm_start: 0x%lx, vm_end: 0x%lx\n",
+					" vm_start: 0x%lx, vm_end: 0x%lx, pgsize: %lu, ind: %lu\n",
 					__func__,
 					(unsigned long)addr, packet.fault_tid,
 					rsysnum, error,
-					vma->vm_start, vma->vm_end);
+					vma->vm_start, vma->vm_end, pgsize, pix);
 			}
 		}
-		else
+		else {
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 18, 0)
 			error = vmf_insert_pfn(vma, rva+(pix*PAGE_SIZE),
 					       pfn+pix);
+			if (error == VM_FAULT_NOPAGE) {
+				dprintk("%s: vmf_insert_pfn returned %d\n",
+					__func__, error);
+				error = 0;
+			}
 #else
 			error = vm_insert_pfn(vma, rva+(pix*PAGE_SIZE),
 					      pfn+pix);
 #endif
+		}
 		if (error) {
 			pr_err("%s: vm_insert_pfn returned %d\n",
 			       __func__, error);
@@ -1831,20 +1849,165 @@ static long pager_call(ihk_os_t os, struct syscall_request *req)
 	return ret;
 }
 
-void __return_syscall(ihk_os_t os, struct ikc_scd_packet *packet,
+#ifdef ENABLE_TOFU
+struct list_head mcctrl_file_to_pidfd_hash[MCCTRL_FILE_2_PIDFD_HASH_SIZE];
+spinlock_t mcctrl_file_to_pidfd_hash_lock;
+
+void mcctrl_file_to_pidfd_hash_init(void)
+{
+	int hash;
+	spin_lock_init(&mcctrl_file_to_pidfd_hash_lock);
+
+	for (hash = 0; hash < MCCTRL_FILE_2_PIDFD_HASH_SIZE; ++hash) {
+		INIT_LIST_HEAD(&mcctrl_file_to_pidfd_hash[hash]);
+	}
+}
+
+int mcctrl_file_to_pidfd_hash_insert(struct file *filp,
+	ihk_os_t os, int pid, struct task_struct *group_leader, int fd,
+	char *path, void *pde_data)
+{
+	unsigned long irqflags;
+	struct mcctrl_file_to_pidfd *file2pidfd_iter;
+	struct mcctrl_file_to_pidfd *file2pidfd;
+	int hash = (int)((unsigned long)filp &
+			(unsigned long)MCCTRL_FILE_2_PIDFD_HASH_MASK);
+	int ret = 0;
+
+	file2pidfd = kmalloc(sizeof(*file2pidfd), GFP_ATOMIC);
+	if (!file2pidfd)
+		return -ENOMEM;
+
+	file2pidfd->filp = filp;
+	file2pidfd->os = os;
+	file2pidfd->pid = pid;
+	file2pidfd->group_leader = group_leader;
+	file2pidfd->fd = fd;
+	/* Only copy the name under /proc/tofu/dev/ */
+	strncpy(file2pidfd->tofu_dev_path, path + 15, 128);
+	file2pidfd->pde_data = pde_data;
+
+	spin_lock_irqsave(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+	list_for_each_entry(file2pidfd_iter,
+			&mcctrl_file_to_pidfd_hash[hash], hash) {
+		if (file2pidfd_iter->filp == filp) {
+			printk("%s: WARNING: filp: %p, pid: %d, fd: %d exists\n",
+					__func__, filp, pid, fd);
+			ret = -EBUSY;
+			goto free_out;
+		}
+	}
+
+	list_add_tail(&file2pidfd->hash,
+			&mcctrl_file_to_pidfd_hash[hash]);
+	dprintk("%s: filp: %p, pid: %d, fd: %d added\n",
+			__func__, filp, pid, fd);
+
+	spin_unlock_irqrestore(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+	return ret;
+
+free_out:
+	kfree(file2pidfd);
+	spin_unlock_irqrestore(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+	return ret;
+}
+
+/*
+ * XXX: lookup relies on group_leader to identify the process
+ * because PIDs might be different across name spaces (e.g.,
+ * when using Docker)
+ */
+struct mcctrl_file_to_pidfd *mcctrl_file_to_pidfd_hash_lookup(
+	struct file *filp, struct task_struct *group_leader)
+{
+	unsigned long irqflags;
+	struct mcctrl_file_to_pidfd *file2pidfd_iter;
+	struct mcctrl_file_to_pidfd *file2pidfd = NULL;
+	int hash = (int)((unsigned long)filp &
+			(unsigned long)MCCTRL_FILE_2_PIDFD_HASH_MASK);
+
+	spin_lock_irqsave(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+	list_for_each_entry(file2pidfd_iter,
+			&mcctrl_file_to_pidfd_hash[hash], hash) {
+		if (file2pidfd_iter->filp == filp && 
+				file2pidfd_iter->group_leader == group_leader) {
+			file2pidfd = file2pidfd_iter;
+			dprintk("%s: filp: %p, pid: %d, fd: %d found\n",
+					__func__, filp, file2pidfd->pid, file2pidfd->fd);
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+
+	return file2pidfd;
+}
+
+int mcctrl_file_to_pidfd_hash_remove(struct file *filp,
+	ihk_os_t os, struct task_struct *group_leader, int fd)
+{
+	unsigned long irqflags;
+	struct mcctrl_file_to_pidfd *file2pidfd_iter;
+	int hash = (int)((unsigned long)filp &
+			(unsigned long)MCCTRL_FILE_2_PIDFD_HASH_MASK);
+	int ret = 0;
+
+	spin_lock_irqsave(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+	list_for_each_entry(file2pidfd_iter,
+			&mcctrl_file_to_pidfd_hash[hash], hash) {
+		if (file2pidfd_iter->filp != filp)
+			continue;
+
+		if (file2pidfd_iter->os != os)
+			continue;
+
+		if (file2pidfd_iter->group_leader != group_leader)
+			continue;
+
+		if (file2pidfd_iter->fd != fd)
+			continue;
+
+		list_del(&file2pidfd_iter->hash);
+		dprintk("%s: filp: %p, pid: %d, fd: %d removed\n",
+				__func__, filp, file2pidfd_iter->pid, fd);
+		kfree(file2pidfd_iter);
+		goto unlock_out;
+	}
+
+	dprintk("%s: filp: %p, pid: %d, fd: %d couldn't be found\n",
+			__func__, filp, pid, fd);
+	ret = -ENOENT;
+
+unlock_out:
+	spin_unlock_irqrestore(&mcctrl_file_to_pidfd_hash_lock, irqflags);
+	return ret;
+}
+#endif
+
+void __return_syscall(ihk_os_t os, struct mcctrl_per_proc_data *ppd,
+		struct ikc_scd_packet *packet,
 		long ret, int stid)
 {
 	unsigned long phys;
 	struct syscall_response *res;
 
+	if (!os || ihk_host_validate_os(os) || !packet) {
+		return;
+	}
+
 	phys = ihk_device_map_memory(ihk_os_to_dev(os),
 			packet->resp_pa, sizeof(*res));
+	if (!phys) {
+		return;
+	}
+
 	res = ihk_device_map_virtual(ihk_os_to_dev(os),
 			phys, sizeof(*res), NULL, 0);
 
 	if (!res) {
 		printk("%s: ERROR: invalid response structure address\n",
 			__FUNCTION__);
+		ihk_device_unmap_memory(ihk_os_to_dev(os), phys, sizeof(*res));
 		return;
 	}
 
@@ -1852,6 +2015,109 @@ void __return_syscall(ihk_os_t os, struct ikc_scd_packet *packet,
 	res->ret = ret;
 	res->stid = stid;
 
+#ifdef ENABLE_TOFU
+	/* Tofu enabled process? */
+	if (ppd && ppd->enable_tofu) {
+		char *pathbuf, *fullpath;
+
+		/* Record PDE_DATA after open() calls for Tofu driver */
+		if (packet->req.number == __NR_openat && ret > 1) {
+			struct fd f;
+			int fd;
+
+			fd = ret;
+			f = fdget(fd);
+
+			if (!f.file) {
+				goto out_notify;
+			}
+
+			pathbuf = (char *)__get_free_page(GFP_ATOMIC);
+			if (!pathbuf) {
+				goto out_fdput_open;
+			}
+
+			fullpath = d_path(&f.file->f_path, pathbuf, PAGE_SIZE);
+			if (IS_ERR(fullpath)) {
+				goto out_free_open;
+			}
+
+			if (!strncmp("/proc/tofu/dev/", fullpath, 15)) {
+				res->pde_data = PDE_DATA(file_inode(f.file));
+				dprintk("%s: fd: %d, path: %s, PDE_DATA: 0x%lx\n",
+						__func__,
+						fd,
+						fullpath,
+						(unsigned long)res->pde_data);
+				dprintk("%s: pgd_index: %ld, pmd_index: %ld, pte_index: %ld\n",
+						__func__,
+						pgd_index((unsigned long)res->pde_data),
+						pmd_index((unsigned long)res->pde_data),
+						pte_index((unsigned long)res->pde_data));
+				dprintk("MAX_USER_VA_BITS: %d, PGDIR_SHIFT: %d\n",
+						MAX_USER_VA_BITS, PGDIR_SHIFT);
+				mcctrl_file_to_pidfd_hash_insert(f.file, os,
+						task_tgid_vnr(current),
+						current->group_leader, fd,
+						fullpath, res->pde_data);
+			}
+
+out_free_open:
+			free_page((unsigned long)pathbuf);
+out_fdput_open:
+			fdput(f);
+		}
+
+		/* Ioctl on Tofu CQ? */
+		else if (packet->req.number == __NR_ioctl &&
+				packet->req.args[0] > 0 && ret == 0) {
+			struct fd f;
+			int fd;
+			int tni, cq;
+			long __ret;
+
+			fd = packet->req.args[0];
+			f = fdget(fd);
+
+			if (!f.file) {
+				goto out_notify;
+			}
+
+			pathbuf = (char *)__get_free_page(GFP_ATOMIC);
+			if (!pathbuf) {
+				goto out_fdput_ioctl;
+			}
+
+			fullpath = d_path(&f.file->f_path, pathbuf, PAGE_SIZE);
+			if (IS_ERR(fullpath)) {
+				goto out_free_ioctl;
+			}
+
+			/* Looking for /proc/tofu/dev/tniXcqY pattern */
+			__ret = sscanf(fullpath, "/proc/tofu/dev/tni%dcq%d", &tni, &cq);
+			if (__ret == 2) {
+				extern long __mcctrl_tof_utofu_unlocked_ioctl_cq(void *pde_data,
+						unsigned int cmd, unsigned long arg);
+
+				dprintk("%s: ioctl(): fd: %d, path: %s\n",
+						__func__,
+						fd,
+						fullpath);
+
+				__ret = __mcctrl_tof_utofu_unlocked_ioctl_cq(
+						PDE_DATA(file_inode(f.file)),
+						packet->req.args[1], packet->req.args[2]);
+			}
+
+out_free_ioctl:
+			free_page((unsigned long)pathbuf);
+out_fdput_ioctl:
+			fdput(f);
+		}
+	}
+
+out_notify:
+#endif
 	if (__notify_syscall_requester(os, packet, res) < 0) {
 		printk("%s: WARNING: failed to notify PID %d\n",
 			__FUNCTION__, packet->pid);
@@ -2154,11 +2420,98 @@ int __do_in_kernel_irq_syscall(ihk_os_t os, struct ikc_scd_packet *packet)
 	if (ret == -ENOSYS)
 		return -ENOSYS;
 
-	__return_syscall(os, packet, ret, 0);
+	__return_syscall(os, NULL, packet, ret, 0);
 
 	return 0;
 }
 
+/*
+ * Memory clearing helpers.
+ */
+struct node_distance;
+
+#define IHK_RBTREE_ALLOCATOR
+
+#ifdef IHK_RBTREE_ALLOCATOR
+struct free_chunk {
+	unsigned long addr, size;
+	struct rb_node node;
+	struct llist_node list;
+};
+#endif
+
+typedef struct mcs_lock_node {
+#ifndef SPIN_LOCK_IN_MCS
+	unsigned long locked;
+	struct mcs_lock_node *next;
+#endif
+	unsigned long irqsave;
+#ifdef SPIN_LOCK_IN_MCS
+	ihk_spinlock_t spinlock;
+#endif
+#ifndef ENABLE_UBSAN
+} __aligned(64) mcs_lock_node_t;
+#else
+} mcs_lock_node_t;
+#endif
+
+struct ihk_mc_numa_node {
+	int id;
+	int linux_numa_id;
+	int type;
+	struct list_head allocators;
+	struct node_distance *nodes_by_distance;
+#ifdef IHK_RBTREE_ALLOCATOR
+	atomic_t zeroing_workers;
+	atomic_t nr_to_zero_pages;
+	struct llist_head zeroed_list;
+	struct llist_head to_zero_list;
+	struct rb_root free_chunks;
+	mcs_lock_node_t lock;
+
+	unsigned long nr_pages;
+	/*
+	 * nr_free_pages: all freed pages, zeroed if zero_at_free
+	 */
+	unsigned long nr_free_pages;
+	unsigned long min_addr;
+	unsigned long max_addr;
+#endif
+};
+
+void mcctrl_zero_mckernel_pages(unsigned long arg)
+{
+	struct llist_node *llnode;
+	struct ihk_mc_numa_node *node =
+		(struct ihk_mc_numa_node *)arg;
+
+	/* Iterate free chunks */
+	while ((llnode = llist_del_first(&node->to_zero_list))) {
+		unsigned long addr;
+		unsigned long size;
+		struct free_chunk *chunk =
+			container_of(llnode, struct free_chunk, list);
+
+		addr = chunk->addr;
+		size = chunk->size;
+
+		memset(phys_to_virt(addr) + sizeof(*chunk), 0,
+				chunk->size - sizeof(*chunk));
+		llist_add(&chunk->list, &node->zeroed_list);
+
+		dprintk("%s: zeroed %lu pages @ McKernel NUMA %d (chunk: 0x%lx:%lu)\n",
+				__func__,
+				size >> PAGE_SHIFT,
+				node->id,
+				addr, size);
+		barrier();
+		atomic_sub((int)(size >> PAGE_SHIFT), &node->nr_to_zero_pages);
+	}
+
+	atomic_dec(&node->zeroing_workers);
+}
+
+
 int __do_in_kernel_syscall(ihk_os_t os, struct ikc_scd_packet *packet)
 {
 	struct syscall_request *sc = &packet->req;
@@ -2167,6 +2520,28 @@ int __do_in_kernel_syscall(ihk_os_t os, struct ikc_scd_packet *packet)
 
 	dprintk("%s: system call: %lx\n", __FUNCTION__, sc->args[0]);
 	switch (sc->number) {
+#ifdef ENABLE_TOFU
+	case __NR_close: {
+		struct fd f;
+		int fd;
+
+		fd = (int)sc->args[0];
+		if (fd > 2) {
+			f = fdget(fd);
+
+			if (f.file) {
+				mcctrl_file_to_pidfd_hash_remove(f.file, os,
+						current->group_leader, fd);
+				fdput(f);
+			}
+		}
+
+		error = -ENOSYS;
+		goto out;
+
+		break;
+	}
+#endif
 	case __NR_mmap:
 		ret = pager_call(os, sc);
 		break;
@@ -2179,6 +2554,14 @@ int __do_in_kernel_syscall(ihk_os_t os, struct ikc_scd_packet *packet)
 		ret = remap_user_space(sc->args[0], sc->args[1], sc->args[2]);
 		break;
 
+	case __NR_move_pages:
+		/*
+		 * move pages is used for zeroing McKernel side memory,
+		 * this call is NOT offloaded by applications.
+		 */
+		mcctrl_zero_mckernel_pages(sc->args[0]);
+		goto out_no_syscall_return;
+
 	case __NR_exit_group: {
 	
 		/* Make sure the user space handler will be called as well */
@@ -2262,7 +2645,9 @@ sched_setparam_out:
 		break;
 	}
 
-	__return_syscall(os, packet, ret, 0);
+	__return_syscall(os, NULL, packet, ret, 0);
+
+out_no_syscall_return:
 	ihk_ikc_release_packet((struct ihk_ikc_free_packet *)packet);
 
 	error = 0;