f7b277a623
HFI1: use ihk_mc_pt_lookup_fault_pte() in SDMA/exp receive
2018-06-13 00:31:43 +09:00
8164b63fc2
HFI1: port to IFS 10.7 rpv1 and support queued_spin_lock in Linux 3.10.0-693.11.6
2018-06-13 00:31:42 +09:00
af22ce62d2
HFI1: clean up and eliminate dead code in user SDMA
2018-06-13 00:31:42 +09:00
238e346586
HFI1: use DWARF generated headers for user_sdma_request and user_sdma_txreq
2018-06-13 00:31:41 +09:00
32fbc015f5
HFI1: eliminate lots of dead code
2018-06-13 00:31:40 +09:00
fcc8310454
HFI1: track receive TIDs in a tree
2018-06-13 00:31:40 +09:00
521e0dc707
HFI1: add a bunch of fields to hfi1_devdata and hfi1_filedata for receive TID handling, do necessary mappings in hfi1_map_device_addresses()
2018-06-13 00:31:40 +09:00
5bea237581
HFI1: make kmalloc caches per-CPU and pre-allocate at boot time
2018-06-13 00:31:39 +09:00
2900ce20f7
HFI1: hfi1_unmap_device_addresses() at process terminate time
2018-06-13 00:31:39 +09:00
03fed4d1c8
automatically generate hfi structs from dwarf info
2018-06-13 00:31:38 +09:00
6959d5ead4
HFI: port to SFI driver version 10.5.1.0.2
2018-06-13 00:31:38 +09:00
7366da4390
Fix other warnings
...
Most were harmless, but the change to ACCESS_ONCE from volatile
cast is probably useful.
Expanding macro, we basically went from:
m = (volatile struct sdma_vl_map *)dd->sdma_map;
to
m = *(volatile struct sdma_vl_map **)&(dd->sdma_map);
i.e. the explicit lookup is at a different level.
2018-06-13 00:31:35 +09:00
2dc85ee417
user_sdma: fix use of uninitialized variable (vl)
...
This defines a single field in hfi1_pportdata, getting offset
from dwarf headers -- need to compute that at configure time
2018-06-13 00:31:35 +09:00
815e2244ca
HFI1: minor change of declarations
2018-06-13 00:31:34 +09:00
163af73554
HFI1: properly iterate iovecs according to underlying page sizes
2018-06-13 00:31:34 +09:00
fd316f3ca3
HFI1: pass per-CPU txreq_cache to user_sdma_send_pkts()
2018-06-13 00:31:34 +09:00
70238982c2
HFI1: use embedded kmalloc cache for req->tids (fixes AllReduce hang)
2018-06-13 00:31:34 +09:00
5b5191ef64
HFI1: move txreq kmalloc cache header into CPU local variable
2018-06-13 00:31:34 +09:00
5353b11f90
HFI1: disable kmalloc cache for req->tids (AllReduce fails otherwise)
2018-06-13 00:31:34 +09:00
bd170e63ba
kmalloc cache refactor and pre-alloc in HFI1 open()
2018-06-13 00:31:33 +09:00
d35fa16417
HFI1: more detailed profiling (disabled by default)
2018-06-13 00:31:33 +09:00
6406a0df6b
HFI1: compute SDMA pkt length taking large pages into account
2018-06-13 00:31:33 +09:00
52e8f03b4b
HFI1: store base physical address in iovec if physically contiguous
2018-06-13 00:31:33 +09:00
b071a3f32c
HFI1: use fast_memcpy() in header fillings
...
Conflicts:
kernel/user_sdma.c
2018-06-13 00:31:33 +09:00
90258f00bd
HFI1: use generic kmalloc cache for user_sdma_txreqs and req tids
2018-06-13 00:31:33 +09:00
744ebacf65
HFI1: more pre-allocation in txreq cache
2018-06-13 00:31:33 +09:00
62e438a0aa
HFI1: do device ioremap() mappings in per-process fashion
2018-06-13 00:31:32 +09:00
5ac582a678
user_sdma_send_pkts(): unlikely() around slow path condition
2018-06-13 00:31:32 +09:00
c43654d69b
user_sdma_send_pkts(): handle page sizes correctly
2018-06-13 00:31:32 +09:00
6e289e8d9f
HFI1: txreq cache and profiling
2018-06-13 00:31:32 +09:00
3b5363c533
HFI1: use original length calculation in sdma_send_pkts()
...
Conflicts:
kernel/include/hfi1/sdma.h
2018-06-13 00:31:32 +09:00
60f6862db2
HFI1: use local write if private data is present; fix lenght alignment
2018-06-13 00:31:31 +09:00
39deff4e10
HFI1: working but a bit slow
2018-06-13 00:31:31 +09:00
7f03c18d4d
Real run test version (update_tail, kregbase+offset crash)
2018-06-13 00:31:31 +09:00
640dba627f
Added debugging output. Bugfixes in user_sdma_send_pkts() and sdma_send_txreq().
2018-06-13 00:31:31 +09:00
ae368d97d4
Implemented a replacement for sdma_txadd_page()
...
Conflicts:
kernel/user_sdma.c
2018-06-13 00:31:31 +09:00
99c216d91e
HFI1: fix kregbase/piobase types to avoid warnings
2018-06-13 00:31:31 +09:00
3c357dc30a
HFI1: fix completion mapping
2018-06-13 00:31:31 +09:00
37866e61ab
HFI1: map completion queues
2018-06-13 00:31:31 +09:00
076e6b9b12
Enabled _sdma_txadd_daddr()
2018-06-13 00:31:30 +09:00
7d2e2f93b0
HFI1: map piobase and rcvarray_wc
2018-06-13 00:31:30 +09:00
fe4c461f2f
Updated kcalloc/kmalloc calls and enabled sdma_select_user_engine dependencies
...
Conflicts:
kernel/include/hfi1/ihk_hfi1_common.h
2018-06-13 00:31:29 +09:00
b60a980088
hfi1_user_sdma_process_request(): map HFI1 kregbase
2018-06-13 00:31:29 +09:00
ec66229063
HFI1: adjust sdma_select_user_engine()
...
Conflicts:
kernel/user_sdma.c
2018-06-13 00:31:29 +09:00
5cf884ef41
Updated TODO tags and struct hfi1_user_sdma_pkt_q
2018-06-13 00:31:29 +09:00
64e2639adc
* The relevant files have been modified in order to compile with McKernel.
...
Conflicts:
kernel/Makefile.build.in
2018-06-13 00:31:29 +09:00
14b360e867
* Added the original files of the driver as a basis for comparison
...
Conflicts:
kernel/include/hfi1/sdma.h
kernel/sdma.c
kernel/user_sdma.c
2018-06-13 00:31:29 +09:00
4a0e389953
HFI1: comments to keep in mind
...
Conflicts:
kernel/include/hfi1/sdma.h
kernel/sdma.c
kernel/user_sdma.c
2018-06-13 00:31:28 +09:00
