gh0s7/mckernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
4214441ac30e5fc67f85df85f66fc587733b7dd0
mckernel/kernel/syscall.c
Masamichi Takagi m-takagi@ab.jp.nec.com 4214441ac3 Make clone system-call create a new thread on the next available physical core 
and skip the physical core with a system process so that
programs using Intel OpenMP bind threads to cores in that manner.
2013-07-05 18:55:03 +09:00

952 lines
26 KiB
C
Raw Blame History

#include <types.h>
#include <kmsg.h>
#include <ihk/cpu.h>
#include <cpulocal.h>
#include <ihk/mm.h>
#include <ihk/debug.h>
#include <ihk/ikc.h>
#include <errno.h>
#include <cls.h>
#include <syscall.h>
#include <page.h>
#include <amemcpy.h>
#include <uio.h>
#include <ihk/lock.h>
#include <ctype.h>
#include <waitq.h>
#include <rlimit.h>
#include <affinity.h>
#include <time.h>
#include <ihk/perfctr.h>
/* Headers taken from kitten LWK */
#include <lwk/stddef.h>
#include <futex.h>
#define SYSCALL_BY_IKC
//#define DEBUG_PRINT_SC
#ifdef DEBUG_PRINT_SC
#define dkprintf kprintf
#else
#define dkprintf(...)
#endif
static ihk_atomic_t pid_cnt = IHK_ATOMIC_INIT(1024);
/* generate system call handler's prototypes */
#define SYSCALL_HANDLED(number,name) extern long sys_##name(int n, ihk_mc_user_context_t *ctx);
#define SYSCALL_DELEGATED(number,name)
#include <syscall_list.h>
#undef SYSCALL_HANDLED
#undef SYSCALL_DELEGATED
/* generate syscall_table[] */
static long (*syscall_table[])(int, ihk_mc_user_context_t *) = {
#define SYSCALL_HANDLED(number,name) [number] = &sys_##name,
#define SYSCALL_DELEGATED(number,name)
#include <syscall_list.h>
#undef SYSCALL_HANDLED
#undef SYSCALL_DELEGATED
};
/* generate syscall_name[] */
#define MCKERNEL_UNUSED __attribute__ ((unused))
static char *syscall_name[] MCKERNEL_UNUSED = {
#define DECLARATOR(number,name) [number] = #name,
#define SYSCALL_HANDLED(number,name) DECLARATOR(number,sys_##name)
#define SYSCALL_DELEGATED(number,name) DECLARATOR(number,sys_##name)
#include <syscall_list.h>
#undef DECLARATOR
#undef SYSCALL_HANDLED
#undef SYSCALL_DELEGATED
};
#ifdef DCFA_KMOD
static void do_mod_exit(int status);
#endif
static void send_syscall(struct syscall_request *req)
{
struct ikc_scd_packet packet;
struct syscall_response *res = cpu_local_var(scp).response_va;
unsigned long fin;
int w;
res->status = 0;
req->valid = 0;
memcpy_async(cpu_local_var(scp).request_pa,
virt_to_phys(req), sizeof(*req), 0, &fin);
memcpy_async_wait(&cpu_local_var(scp).post_fin);
cpu_local_var(scp).post_va->v[0] = cpu_local_var(scp).post_idx;
w = ihk_mc_get_processor_id() + 1;
memcpy_async_wait(&fin);
cpu_local_var(scp).request_va->valid = 1;
*(unsigned int *)cpu_local_var(scp).doorbell_va = w;
#ifdef SYSCALL_BY_IKC
packet.msg = SCD_MSG_SYSCALL_ONESIDE;
packet.ref = ihk_mc_get_processor_id();
packet.arg = cpu_local_var(scp).request_rpa;
ihk_ikc_send(cpu_local_var(syscall_channel), &packet, 0);
#endif
}
int do_syscall(struct syscall_request *req, ihk_mc_user_context_t *ctx)
{
struct syscall_response *res = cpu_local_var(scp).response_va;
dkprintf("SC(%d)[%3d] sending syscall\n",
ihk_mc_get_processor_id(),
req->number);
send_syscall(req);
dkprintf("SC(%d)[%3d] waiting for host.. \n",
ihk_mc_get_processor_id(),
req->number);
while (!res->status) {
cpu_pause();
}
dkprintf("SC(%d)[%3d] got host reply: %d \n",
ihk_mc_get_processor_id(),
req->number, res->ret);
return res->ret;
}
SYSCALL_DECLARE(open)
{
SYSCALL_HEADER;
dkprintf("open: %s\n", (char*)ihk_mc_syscall_arg0(ctx));
SYSCALL_ARGS_3(MI, D, D);
SYSCALL_FOOTER;
}
SYSCALL_DECLARE(exit_group)
{
SYSCALL_HEADER;
struct process *proc = cpu_local_var(current);
#ifdef DCFA_KMOD
do_mod_exit((int)ihk_mc_syscall_arg0(ctx));
#endif
/* XXX: send SIGKILL to all threads in this process */
do_syscall(&request, ctx);
#define IS_DETACHED_PROCESS(proc) (1) /* should be implemented in the future */
proc->status = PS_ZOMBIE;
if (IS_DETACHED_PROCESS(proc)) {
/* release a reference for wait(2) */
proc->status = PS_EXITED;
free_process(proc);
}
schedule();
return 0;
}
// MIC:9 linux:90
SYSCALL_DECLARE(mmap)
{
struct vm_regions *region = &cpu_local_var(current)->vm->region;
void *va;
dkprintf("syscall.c,mmap,addr=%lx,len=%lx,prot=%lx,flags=%x,fd=%x,offset=%lx\n",
ihk_mc_syscall_arg0(ctx), ihk_mc_syscall_arg1(ctx),
ihk_mc_syscall_arg2(ctx), ihk_mc_syscall_arg3(ctx),
ihk_mc_syscall_arg4(ctx), ihk_mc_syscall_arg5(ctx)
);
//kprintf("syscall.c,mmap,dumping kmsg...\n");
// send_kmsg(ctx);
// return -EINVAL; // debug
if((ihk_mc_syscall_arg3(ctx) & 0x10) == 0x10) {
// libc/sysdeps/unix/sysv/linux/x86_64/bits/mman.h
// #define MAP_FIXED 0x10
// use the given vaddr as is
struct syscall_request request IHK_DMA_ALIGN;
request.number = n;
// do a job similar to mcos/kernel/host.c:process_msg_prepare_process
unsigned long s = (ihk_mc_syscall_arg0(ctx)) & PAGE_MASK;
unsigned long e = (s + ihk_mc_syscall_arg1(ctx)
+ PAGE_SIZE - 1) & PAGE_MASK;
int range_npages = (e - s) >> PAGE_SHIFT;
unsigned long pa;
int r = ihk_mc_pt_virt_to_phys(cpu_local_var(current)->vm->page_table, (void *)s, &pa);
// va range is not overwrapped with existing mmap
if(r != 0) {
#ifdef USE_LARGE_PAGES
// use large pages if mapping is big enough
if (e - s >= LARGE_PAGE_SIZE) {
unsigned long p;
unsigned long p_aligned;
unsigned long s_orig = s;
unsigned long head_space = 0;
// compute head space before the first large page aligned
// virtual address
if ((s & (LARGE_PAGE_SIZE - 1)) != 0) {
s = (s + (LARGE_PAGE_SIZE - 1)) & LARGE_PAGE_MASK;
head_space = (s - s_orig);
}
e = (e + (LARGE_PAGE_SIZE - 1)) & LARGE_PAGE_MASK;
if((p = (unsigned long)ihk_mc_alloc_pages(
(e - s + 2 * LARGE_PAGE_SIZE) >> PAGE_SHIFT, IHK_MC_AP_NOWAIT)) == (unsigned long)NULL) {
return -ENOMEM;
}
p_aligned = (p + LARGE_PAGE_SIZE + (LARGE_PAGE_SIZE - 1))
& LARGE_PAGE_MASK;
// add range, mapping
if(add_process_memory_range(cpu_local_var(current), s_orig, e,
virt_to_phys((void *)(p_aligned - head_space)), VR_NONE) != 0){
ihk_mc_free_pages((void *)p, range_npages);
return -ENOMEM;
}
dkprintf("largePTE area: 0x%lX - 0x%lX (s: %lu) -> 0x%lX -\n",
s_orig, e, (e - s_orig),
virt_to_phys((void *)(p_aligned - head_space)));
}
else {
#endif
// allocate physical address
if((va = ihk_mc_alloc_pages(range_npages, IHK_MC_AP_NOWAIT)) == NULL){
return -ENOMEM;
}
pa = virt_to_phys(va);
// add page_table, add memory-range
if(add_process_memory_range(cpu_local_var(current), s, e, pa, VR_NONE) != 0){
ihk_mc_free_pages(va, range_npages);
return -ENOMEM;
}
dkprintf("syscall.c,pa allocated=%lx\n", pa);
#ifdef USE_LARGE_PAGES
}
#endif
} else {
dkprintf("syscall.c,pa found=%lx\n", pa);
// we need to clear to avoid BSS contamination, even when reusing physical memory range
// because ld.so performs mmap (va:0, size:va of second section including BSS, FIXED, prot:RX, offset:0)
// this causes contamination of BSS section when libc.so is large enough to reach BSS section
// then performs mmap (va:second section including BSS, FIXED, prot:RW, offset:second section in file)
dkprintf("syscall.c,clearing from %lx to %lx\n", s, e);
memset((void*)phys_to_virt(pa), 0, e - s);
}
if ((ihk_mc_syscall_arg3(ctx) & 0x20) == 0x20) {
// #define MAP_ANONYMOUS 0x20
dkprintf("syscall.c,MAP_FIXED,MAP_ANONYMOUS\n");
return ihk_mc_syscall_arg0(ctx); // maybe we should return zero
} else {
dkprintf("syscall.c,MAP_FIXED,!MAP_ANONYMOUS\n");
// lseek(mmap_fd, mmap_off, SEEK_SET);
// read(mmap_fd, mmap_addr, mmap_len);
SYSCALL_ARGS_6(MO, D, D, D, D, D);
int r = do_syscall(&request, ctx);
if(r == 0) { return ihk_mc_syscall_arg0(ctx); } else { return -EINVAL; }
}
} else if ((ihk_mc_syscall_arg3(ctx) & 0x20) == 0x20) {
// #define MAP_ANONYMOUS 0x20
dkprintf("syscall.c,!MAP_FIXED,MAP_ANONYMOUS\n");
ihk_mc_spinlock_lock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
unsigned long s = (region->map_end + PAGE_SIZE - 1) & PAGE_MASK;
unsigned long map_end_aligned = region->map_end;
unsigned long len = (ihk_mc_syscall_arg1(ctx) + PAGE_SIZE - 1) & PAGE_MASK;
dkprintf("syscall.c,mmap,len=%lx\n", len);
unsigned long flag = 0; /* eager paging */
#if 1
/* Intel OpenMP hack: it requests 128MB and munmap tail and head
to create 64MB-aligned 64MB memory area
and then it tries to touch memory-block
with offset of 0x10, 0x101008 */
if(ihk_mc_syscall_arg1(ctx) == 1024*1024*64 || ihk_mc_syscall_arg1(ctx) == 1024*1024*128) {
flag |= VR_DEMAND_PAGING; /* demand paging */
kprintf("SC(%d)[mmap],!MAP_FIXED,MAP_ANONYMOUS,sz=%lx\n", ihk_mc_get_processor_id(), ihk_mc_syscall_arg1(ctx));
}
//if(ihk_mc_syscall_arg3(ctx) & 0x100) { flag |= 0x1000; };
#endif
#ifdef USE_NOCACHE_MMAP
if ((ihk_mc_syscall_arg3(ctx) & 0x40) == 0x40) {
dkprintf("syscall.c,mmap,nocache,len=%lx\n", len);
region->map_end = extend_process_region(
cpu_local_var(current), region->map_start, map_end_aligned,
s + len, VR_IO_NOCACHE);
}
else
#endif
{
region->map_end =
extend_process_region(cpu_local_var(current),
region->map_start,
map_end_aligned,
s + len, flag);
}
if(ihk_mc_syscall_arg1(ctx) == 1024*1024*64 || ihk_mc_syscall_arg1(ctx) == 1024*1024*128) {
kprintf("syscall.c,mmap,cpuid=%d,map_end-len=%lx,s+len=%lx,map_end=%lx\n", ihk_mc_get_processor_id(), region->map_end-len, s+len, region->map_end);
}
ihk_mc_spinlock_unlock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
#ifdef USE_LARGE_PAGES
if (region->map_end >= s + len) {
/* NOTE: extend_process_region() might have large page aligned */
return region->map_end - len;
}
#else
if (region->map_end == s + len) return s;
#endif
else {
return -EINVAL;
}
} else if ((ihk_mc_syscall_arg3(ctx) & 0x02) == 0x02) {
// #define MAP_PRIVATE 0x02
ihk_mc_spinlock_lock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
#if 1 /* takagidebug*/
unsigned long amt_align = 0x100000; /* takagi */
unsigned long s = ((region->map_end + amt_align - 1) & ~(amt_align - 1));
unsigned long len = (ihk_mc_syscall_arg1(ctx) + PAGE_SIZE - 1) & PAGE_MASK;
dkprintf("(%d),syscall.c,!MAP_FIXED,!MAP_ANONYMOUS,amt_align=%lx,s=%lx,len=%lx\n", ihk_mc_get_processor_id(), amt_align, s, len);
region->map_end =
extend_process_region(cpu_local_var(current),
region->map_start,
s,
s + len, 0);
#else
unsigned long s = (region->map_end + PAGE_SIZE - 1) & PAGE_MASK;
unsigned long len = (ihk_mc_syscall_arg1(ctx) + PAGE_SIZE - 1) & PAGE_MASK;
region->map_end =
extend_process_region(cpu_local_var(current),
region->map_start,
region->map_end,
s + len, 0);
#endif
ihk_mc_spinlock_unlock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
if (region->map_end < s + len) { return -EINVAL; }
s = region->map_end - len;
struct syscall_request request IHK_DMA_ALIGN;
request.number = n;
dkprintf("syscall.c,!MAP_FIXED,!MAP_ANONYMOUS,MAP_PRIVATE\n");
// lseek(mmap_fd, mmap_off, SEEK_SET);
// read(mmap_fd, mmap_addr, mmap_len);
SYSCALL_ARGS_6(D, D, D, D, D, D);
// overwriting request.args[0]
unsigned long __phys;
if (ihk_mc_pt_virt_to_phys(cpu_local_var(current)->vm->page_table, (void *)s, &__phys)) {
return -EFAULT;
}
request.args[0] = __phys;
int r = do_syscall(&request, ctx);
if(r == 0) { return s; } else { return -EINVAL; }
}
dkprintf("mmap flags not supported: fd = %lx, %lx\n",
ihk_mc_syscall_arg4(ctx), ihk_mc_syscall_arg5(ctx));
while(1);
}
SYSCALL_DECLARE(munmap)
{
unsigned long address, len;
int error;
address = ihk_mc_syscall_arg0(ctx);
len = ihk_mc_syscall_arg1(ctx);
ihk_mc_spinlock_lock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
error = remove_process_memory_range(cpu_local_var(current), address, address+len);
ihk_mc_spinlock_unlock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
return error;
}
SYSCALL_DECLARE(mprotect)
{
dkprintf("mprotect returns 0\n");
return 0;
}
SYSCALL_DECLARE(brk)
{
unsigned long address = ihk_mc_syscall_arg0(ctx);
struct vm_regions *region = &cpu_local_var(current)->vm->region;
unsigned long r;
dkprintf("SC(%d)[sys_brk] brk_start=%lx,end=%lx\n",
ihk_mc_get_processor_id(), region->brk_start, region->brk_end);
/* brk change fail, including glibc trick brk(0) to obtain current brk */
if(address < region->brk_start) {
r = region->brk_end;
goto out;
}
/* brk change fail, because we don't shrink memory region */
if(address < region->brk_end) {
r = region->brk_end;
goto out;
}
/* try to extend memory region */
ihk_mc_spinlock_lock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
region->brk_end = extend_process_region(cpu_local_var(current),
region->brk_start, region->brk_end, address, 0);
ihk_mc_spinlock_unlock_noirq(&cpu_local_var(current)->vm->memory_range_lock);
dkprintf("SC(%d)[sys_brk] brk_end set to %lx\n",
ihk_mc_get_processor_id(), region->brk_end);
r = region->brk_end;
out:
return r;
}
SYSCALL_DECLARE(getpid)
{
return cpu_local_var(current)->pid;
}
long do_arch_prctl(unsigned long code, unsigned long address)
{
int err = 0;
enum ihk_asr_type type;
switch (code) {
case ARCH_SET_FS:
case ARCH_GET_FS:
type = IHK_ASR_X86_FS;
break;
case ARCH_GET_GS:
type = IHK_ASR_X86_GS;
break;
case ARCH_SET_GS:
return -ENOTSUPP;
default:
return -EINVAL;
}
switch (code) {
case ARCH_SET_FS:
dkprintf("[%d] arch_prctl: ARCH_SET_FS: 0x%lX\n",
ihk_mc_get_processor_id(), address);
cpu_local_var(current)->thread.tlsblock_base = address;
err = ihk_mc_arch_set_special_register(type, address);
break;
case ARCH_SET_GS:
err = ihk_mc_arch_set_special_register(type, address);
break;
case ARCH_GET_FS:
case ARCH_GET_GS:
err = ihk_mc_arch_get_special_register(type,
(unsigned long*)address);
break;
default:
break;
}
return err;
}
SYSCALL_DECLARE(arch_prctl)
{
return do_arch_prctl(ihk_mc_syscall_arg0(ctx),
ihk_mc_syscall_arg1(ctx));
}
SYSCALL_DECLARE(clone)
{
int cpuid;
int clone_flags = ihk_mc_syscall_arg0(ctx);
struct process *new;
dkprintf("[%d] clone(): stack_pointr: 0x%lX\n",
ihk_mc_get_processor_id(),
(unsigned long)ihk_mc_syscall_arg1(ctx));
cpuid = obtain_clone_cpuid();
new = clone_process(cpu_local_var(current), ihk_mc_syscall_pc(ctx),
ihk_mc_syscall_arg1(ctx));
if (!new) {
return -ENOMEM;
}
/* Allocate new pid */
new->pid = ihk_atomic_inc_return(&pid_cnt);
if (clone_flags & CLONE_PARENT_SETTID) {
dkprintf("clone_flags & CLONE_PARENT_SETTID: 0x%lX\n",
(unsigned long)ihk_mc_syscall_arg2(ctx));
*(int*)ihk_mc_syscall_arg2(ctx) = new->pid;
}
if (clone_flags & CLONE_CHILD_CLEARTID) {
dkprintf("clone_flags & CLONE_CHILD_CLEARTID: 0x%lX\n",
(unsigned long)ihk_mc_syscall_arg3(ctx));
new->thread.clear_child_tid = (int*)ihk_mc_syscall_arg3(ctx);
}
if (clone_flags & CLONE_SETTLS) {
dkprintf("clone_flags & CLONE_SETTLS: 0x%lX\n",
(unsigned long)ihk_mc_syscall_arg4(ctx));
new->thread.tlsblock_base =
(unsigned long)ihk_mc_syscall_arg4(ctx);
}
else {
new->thread.tlsblock_base =
cpu_local_var(current)->thread.tlsblock_base;
}
ihk_mc_syscall_ret(new->uctx) = 0;
dkprintf("clone: kicking scheduler!,cpuid=%d\n", cpuid);
runq_add_proc(new, cpuid);
return new->pid;
}
SYSCALL_DECLARE(set_tid_address)
{
cpu_local_var(current)->thread.clear_child_tid =
(int*)ihk_mc_syscall_arg2(ctx);
return cpu_local_var(current)->pid;
}
// see linux-2.6.34.13/kernel/signal.c
SYSCALL_DECLARE(tgkill)
{
int tgid = ihk_mc_syscall_arg0(ctx);
int pid = ihk_mc_syscall_arg1(ctx);
int sig = ihk_mc_syscall_arg2(ctx);
if(pid <= 0 || tgid <= 0) { return -EINVAL; }
// search pid
// check kill permission
if(sig == 0) {
return 0;
} else {
return -EPERM;
}
}
SYSCALL_DECLARE(set_robust_list)
{
return -ENOSYS;
}
SYSCALL_DECLARE(rt_sigaction)
{
// kprintf("sys_rt_sigaction called. returning zero...\n");
return 0;
}
SYSCALL_DECLARE(rt_sigprocmask)
{
// kprintf("sys_rt_sigprocmask called. returning zero...\n");
return 0;
}
SYSCALL_DECLARE(madvise)
{
// kprintf("sys_madvise called. returning zero...\n");
return 0;
}
SYSCALL_DECLARE(futex)
{
uint64_t timeout = 0; // No timeout
uint32_t val2 = 0;
uint32_t *uaddr = (uint32_t *)ihk_mc_syscall_arg0(ctx);
int op = (int)ihk_mc_syscall_arg1(ctx);
uint32_t val = (uint32_t)ihk_mc_syscall_arg2(ctx);
struct timespec *utime = (struct timespec*)ihk_mc_syscall_arg3(ctx);
uint32_t *uaddr2 = (uint32_t *)ihk_mc_syscall_arg4(ctx);
uint32_t val3 = (uint32_t)ihk_mc_syscall_arg5(ctx);
/* Mask off the FUTEX_PRIVATE_FLAG,
* assume all futexes are address space private */
op = (op & FUTEX_CMD_MASK);
dkprintf("futex op=[%x, %s],uaddr=%lx, val=%x, utime=%lx, uaddr2=%lx, val3=%x, []=%x\n",
op,
(op == FUTEX_WAIT) ? "FUTEX_WAIT" :
(op == FUTEX_WAIT_BITSET) ? "FUTEX_WAIT_BITSET" :
(op == FUTEX_WAKE) ? "FUTEX_WAKE" :
(op == FUTEX_WAKE_OP) ? "FUTEX_WAKE_OP" :
(op == FUTEX_WAKE_BITSET) ? "FUTEX_WAKE_BITSET" :
(op == FUTEX_CMP_REQUEUE) ? "FUTEX_CMP_REQUEUE" :
(op == FUTEX_REQUEUE) ? "FUTEX_REQUEUE (NOT IMPL!)" : "unknown",
(unsigned long)uaddr, op, val, utime, uaddr2, val3, *uaddr);
if (utime && (op == FUTEX_WAIT_BITSET || op == FUTEX_WAIT)) {
struct syscall_request request IHK_DMA_ALIGN;
struct timeval tv_now;
request.number = n;
unsigned long __phys;
dkprintf("futex,utime and FUTEX_WAIT_*, uaddr=%lx, []=%x\n", (unsigned long)uaddr, *uaddr);
if (ihk_mc_pt_virt_to_phys(cpu_local_var(current)->vm->page_table,
(void *)&tv_now, &__phys)) {
return -EFAULT;
}
request.args[0] = __phys;
int r = do_syscall(&request, ctx);
if (r < 0) {
return -EFAULT;
}
dkprintf("futex, FUTEX_WAIT_*, arg3 != NULL, pc=%lx\n", (unsigned long)ihk_mc_syscall_pc(ctx));
dkprintf("now->tv_sec=%016ld,tv_nsec=%016ld\n", tv_now.tv_sec, tv_now.tv_usec * 1000);
dkprintf("utime->tv_sec=%016ld,tv_nsec=%016ld\n", utime->tv_sec, utime->tv_nsec);
long nsec_now = ((long)tv_now.tv_sec * 1000000000ULL) +
tv_now.tv_usec * 1000;
long nsec_timeout = ((long)utime->tv_sec * 1000000000ULL) +
utime->tv_nsec * 1;
long diff_nsec = nsec_timeout - nsec_now;
timeout = (diff_nsec / 1000) * 1100; // (usec * 1.1GHz)
dkprintf("futex timeout: %lu\n", timeout);
}
/* Requeue parameter in 'utime' if op == FUTEX_CMP_REQUEUE.
* number of waiters to wake in 'utime' if op == FUTEX_WAKE_OP. */
if (op == FUTEX_CMP_REQUEUE || op == FUTEX_WAKE_OP)
val2 = (uint32_t) (unsigned long) ihk_mc_syscall_arg3(ctx);
return futex(uaddr, op, val, timeout, uaddr2, val2, val3);
}
SYSCALL_DECLARE(exit)
{
struct process *proc = cpu_local_var(current);
#ifdef DCFA_KMOD
do_mod_exit((int)ihk_mc_syscall_arg0(ctx));
#endif
/* XXX: for if all threads issued the exit(2) rather than exit_group(2),
* exit(2) also should delegate.
*/
/* If there is a clear_child_tid address set, clear it and wake it.
* This unblocks any pthread_join() waiters. */
if (proc->thread.clear_child_tid) {
dkprintf("exit clear_child!\n");
*proc->thread.clear_child_tid = 0;
barrier();
futex((uint32_t *)proc->thread.clear_child_tid,
FUTEX_WAKE, 1, 0, NULL, 0, 0);
}
proc->status = PS_ZOMBIE;
if (IS_DETACHED_PROCESS(proc)) {
/* release a reference for wait(2) */
proc->status = PS_EXITED;
free_process(proc);
}
schedule();
return 0;
}
SYSCALL_DECLARE(getrlimit)
{
int ret;
int resource = ihk_mc_syscall_arg0(ctx);
struct rlimit *rlm = (struct rlimit *)ihk_mc_syscall_arg1(ctx);
switch (resource) {
case RLIMIT_STACK:
dkprintf("[%d] getrlimit() RLIMIT_STACK\n", ihk_mc_get_processor_id());
rlm->rlim_cur = (512*4096); /* Linux provides 8MB */
rlm->rlim_max = (1024*1024*1024);
ret = 0;
break;
default:
return -ENOSYS;
}
return ret;
}
SYSCALL_DECLARE(sched_setaffinity)
{
#if 0
int pid = (int)ihk_mc_syscall_arg0(ctx);
unsigned int len = (unsigned int)ihk_mc_syscall_arg1(ctx);
#endif
cpu_set_t *mask = (cpu_set_t *)ihk_mc_syscall_arg2(ctx);
unsigned long __phys;
#if 0
int i;
#endif
/* TODO: check mask is in user's page table */
if(!mask) { return -EFAULT; }
if (ihk_mc_pt_virt_to_phys(cpu_local_var(current)->vm->page_table,
(void *)mask,
&__phys)) {
return -EFAULT;
}
#if 0
dkprintf("sched_setaffinity,\n");
for(i = 0; i < len/sizeof(__cpu_mask); i++) {
dkprintf("mask[%d]=%lx,", i, mask->__bits[i]);
}
#endif
return 0;
}
#define MIN2(x,y) (x) < (y) ? (x) : (y)
#define MIN3(x,y,z) MIN2(MIN2((x),(y)),MIN2((y),(z)))
// see linux-2.6.34.13/kernel/sched.c
SYSCALL_DECLARE(sched_getaffinity)
{
//int pid = (int)ihk_mc_syscall_arg0(ctx);
unsigned int len = (int)ihk_mc_syscall_arg1(ctx);
//int cpu_id;
cpu_set_t *mask = (cpu_set_t *)ihk_mc_syscall_arg2(ctx);
struct ihk_mc_cpu_info *cpu_info = ihk_mc_get_cpu_info();
if(len*8 < cpu_info->ncpus) { return -EINVAL; }
if(len & (sizeof(unsigned long)-1)) { return -EINVAL; }
int min_len = MIN2(len, sizeof(cpu_set_t));
//int min_ncpus = MIN2(min_len*8, cpu_info->ncpus);
CPU_ZERO_S(min_len, mask);
CPU_SET_S(ihk_mc_get_hardware_processor_id(), min_len, mask);
//for (cpu_id = 0; cpu_id < min_ncpus; ++cpu_id)
// CPU_SET_S(cpu_info->hw_ids[cpu_id], min_len, mask);
// dkprintf("sched_getaffinity returns full mask\n");
return min_len;
}
SYSCALL_DECLARE(sched_yield)
{
return -ENOSYS;
}
#ifdef DCFA_KMOD
#ifdef CMD_DCFA
extern int ibmic_cmd_syscall(char *uargs);
extern void ibmic_cmd_exit(int status);
#endif
#ifdef CMD_DCFAMPI
extern int dcfampi_cmd_syscall(char *uargs);
#endif
static int (*mod_call_table[]) (char *) = {
#ifdef CMD_DCFA
[1] = ibmic_cmd_syscall,
#endif
#ifdef CMD_DCFAMPI
[2] = dcfampi_cmd_syscall,
#endif
};
static void (*mod_exit_table[]) (int) = {
#ifdef CMD_DCFA
[1] = ibmic_cmd_exit,
#endif
#ifdef CMD_DCFAMPI
[2] = NULL,
#endif
};
SYSCALL_DECLARE(mod_call) {
int mod_id;
unsigned long long uargs;
mod_id = ihk_mc_syscall_arg0(ctx);
uargs = ihk_mc_syscall_arg1(ctx);
dkprintf("mod_call id:%d, uargs=0x%llx, type=%s, command=%x\n", mod_id, uargs, mod_id==1?"ibmic":"dcfampi", *((uint32_t*)(((char*)uargs)+0)));
if(mod_call_table[mod_id])
return mod_call_table[mod_id]((char*)uargs);
kprintf("ERROR! undefined mod_call id:%d\n", mod_id);
return -ENOSYS;
}
static void do_mod_exit(int status){
int i;
for(i=1; i<=2; i++){
if(mod_exit_table[i])
mod_exit_table[i](status);
}
}
#endif
/* select counter type */
SYSCALL_DECLARE(pmc_init)
{
int counter = ihk_mc_syscall_arg0(ctx);
enum ihk_perfctr_type type = (enum ihk_perfctr_type)ihk_mc_syscall_arg1(ctx);
/* see ihk/manycore/generic/include/ihk/perfctr.h */
int mode = PERFCTR_USER_MODE;
return ihk_mc_perfctr_init(counter, type, mode);
}
SYSCALL_DECLARE(pmc_start)
{
unsigned long counter = ihk_mc_syscall_arg0(ctx);
return ihk_mc_perfctr_start(1 << counter);
}
SYSCALL_DECLARE(pmc_stop)
{
unsigned long counter = ihk_mc_syscall_arg0(ctx);
return ihk_mc_perfctr_stop(1 << counter);
}
SYSCALL_DECLARE(pmc_reset)
{
int counter = ihk_mc_syscall_arg0(ctx);
return ihk_mc_perfctr_reset(counter);
}
long syscall_generic_forwarding(int n, ihk_mc_user_context_t *ctx)
{
SYSCALL_HEADER;
dkprintf("syscall_generic_forwarding(%d)\n", n);
SYSCALL_ARGS_6(D,D,D,D,D,D);
SYSCALL_FOOTER;
}
long syscall(int num, ihk_mc_user_context_t *ctx)
{
long l;
cpu_enable_interrupt();
#if 0
if(num != 24) // if not sched_yield
#endif
dkprintf("SC(%d:%d)[%3d=%s](%lx, %lx,%lx, %lx, %lx, %lx)@%lx,sp:%lx",
ihk_mc_get_processor_id(),
ihk_mc_get_hardware_processor_id(),
num, syscall_name[num],
ihk_mc_syscall_arg0(ctx), ihk_mc_syscall_arg1(ctx),
ihk_mc_syscall_arg2(ctx), ihk_mc_syscall_arg3(ctx),
ihk_mc_syscall_arg4(ctx), ihk_mc_syscall_arg5(ctx),
ihk_mc_syscall_pc(ctx), ihk_mc_syscall_sp(ctx));
#if 1
#if 0
if(num != 24) // if not sched_yield
#endif
dkprintf(",*sp:%lx,*(sp+8):%lx,*(sp+16):%lx,*(sp+24):%lx",
*((unsigned long*)ihk_mc_syscall_sp(ctx)),
*((unsigned long*)(ihk_mc_syscall_sp(ctx)+8)),
*((unsigned long*)(ihk_mc_syscall_sp(ctx)+16)),
*((unsigned long*)(ihk_mc_syscall_sp(ctx)+24)));
#endif
#if 0
if(num != 24) // if not sched_yield
#endif
dkprintf("\n");
if ((0 <= num) && (num < (sizeof(syscall_table) / sizeof(syscall_table[0])))
&& (syscall_table[num] != NULL)) {
l = syscall_table[num](num, ctx);
dkprintf("SC(%d)[%3d] ret: %d\n",
ihk_mc_get_processor_id(), num, l);
} else {
dkprintf("USC[%3d](%lx, %lx, %lx, %lx, %lx) @ %lx | %lx\n", num,
ihk_mc_syscall_arg0(ctx), ihk_mc_syscall_arg1(ctx),
ihk_mc_syscall_arg2(ctx), ihk_mc_syscall_arg3(ctx),
ihk_mc_syscall_arg4(ctx), ihk_mc_syscall_pc(ctx),
ihk_mc_syscall_sp(ctx));
l = syscall_generic_forwarding(num, ctx);
}
return l;
}
void __host_update_process_range(struct process *process,
struct vm_range *range)
{
struct syscall_post *post;
int idx;
memcpy_async_wait(&cpu_local_var(scp).post_fin);
post = &cpu_local_var(scp).post_buf;
post->v[0] = 1;
post->v[1] = range->start;
post->v[2] = range->end;
post->v[3] = range->phys;
cpu_disable_interrupt();
if (cpu_local_var(scp).post_idx >=
PAGE_SIZE / sizeof(struct syscall_post)) {
/* XXX: Wait until it is consumed */
} else {
idx = ++(cpu_local_var(scp).post_idx);
cpu_local_var(scp).post_fin = 0;
memcpy_async(cpu_local_var(scp).post_pa +
idx * sizeof(*post),
virt_to_phys(post), sizeof(*post), 0,
&cpu_local_var(scp).post_fin);
}
cpu_enable_interrupt();
}
Reference in New Issue View Git Blame Copy Permalink