gh0s7/mckernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,628 Commits 1 Branch 61 Tags
88fca2c0df9655fb9559bce6b65ef7dc8479cabd
Commit Graph

530 Commits

Author SHA1 Message Date
Ken Sato
ce3af4734a fix: dual hold_thread() in do_kill() 2018-03-19 11:12:50 +09:00
Masamichi Takagi
fe42481d6f Add allow_oversubscribe kernel argument 
It's not allowed in the default setting.
Execute mcreboot.sh with -O option to allow it.

refs #1072
 2018-03-10 13:08:38 +09:00
Masamichi Takagi
65667709a8 Fix thread status race-condition caused by hold_thread() in do_kill() and terminate() 
Conflicts:
	arch/x86_64/kernel/syscall.c
	kernel/syscall.c
 2018-03-09 17:53:17 +09:00
Masamichi Takagi
3b277b2354 uti: Fix dead-lock of calling terminate() from terminate() 
Conflicts:
	arch/x86_64/kernel/syscall.c
	kernel/syscall.c
 2018-03-09 17:38:55 +09:00
Masamichi Takagi
06b1b4f8ab Fix deadlock on thread->times_update in getrusage() 
Set thread->in_kernel properly on exiting interrupt handler when entering
it from kernel mode.

Conflicts:
	arch/x86_64/kernel/cpu.c
	kernel/mem.c
 2018-03-09 17:26:31 +09:00
Ken Sato
a224bf648a fix: Bug for getrusage return incorrect ru_maxrss 
refs #1032
 2018-03-07 13:09:24 +09:00
Ken Sato
140f813d77 fix: differences in behavior of sigaction between Linux and Mckernel 2018-03-01 09:44:44 +09:00
Ken Sato
523a066245 sigaction: support for SA_RESETHAND on x86_64 
refs #1031
 2018-02-22 11:55:32 +09:00
Ken Sato
228f8f8533 Wait for LWK to run at shutdown. 
refs #898
refs #928
 2018-02-06 10:40:12 +09:00
Tomoki Shirasawa
559fc9746c signal: check_signal must be called after check_need_resched. 2018-01-28 13:38:51 +09:00
Balazs Gerofi
86efc86945 save_syscall_return_value(): separate from check_signal() and call from syscall() (for ARM64) 2018-01-26 14:43:18 +09:00
Balazs Gerofi
722ae0e7d5 ARM64 arch_clone_thread(): eliminate extra save_fp_regs() 2018-01-26 13:51:38 +09:00
Tomoki Shirasawa
f55f01cc11 signal: If the thread receiving the signal is not current, the signal is not processed. 2018-01-25 22:27:34 +09:00
Tomoki Shirasawa
1fa398cfab do_kill: fix to initialization leakage 2018-01-24 23:11:18 +09:00
Masamichi Takagi
d4459cf9f3 Add check to confirm IHK and McKernel with the same version are used 2018-01-24 00:20:57 +09:00
Tomoki Shirasawa
4bb65494e9 signal: When the process receives a termination signal, it first terminates mcexec. 
refs #863
refs #870
 2018-01-23 14:40:38 +09:00
Tomoki Shirasawa
2f2b3cdc6f signal: interrupt_syscall is called by the core executing the thread that recieved the signal. 
refs #999
 2018-01-23 14:31:04 +09:00
Tomoki Shirasawa
d209c00a30 part of Issue#994 
mcexec: open syscall moves to arch_dep
do_fork: don't use __NR_fork. use __NR_clone
vfork: moves to arch_dep
 2017-12-26 10:30:33 +09:00
Masamichi Takagi
9a5d5feb9c time(): Split into architecture dependent functions 
This fixes the bug reported as POSTK_ARCH_DEP_13 and POSTK_DEBUG_ARCH_DEP_13.
 2017-12-23 11:36:52 +09:00
Tomoki Shirasawa
0cda763f95 fix /proc/*/pagemap 
refs #387
 2017-12-25 16:08:51 +09:00
Balazs Gerofi
589504dc33 mcreboot: -h to indicate halting CPU in idle threads (e.g., in futex_wait()) 2017-12-18 11:22:15 +09:00
Balazs Gerofi
bf2f38051b mcreboot-smp: offline/online MCDRAM in one go 2017-12-06 14:41:25 +09:00
Tomoki Shirasawa
c9fa445f54 Merge branch 'development' of pccluster.org:mckernel into development 2017-11-22 10:53:33 +09:00
Tomoki Shirasawa
d273a2f58b add strace bundled test cases 2017-11-22 10:52:30 +09:00
Masamichi Takagi
66f44e77af mcstop+release.sh: Allow ihkmond to flush kmsg buffer 2017-11-20 18:28:48 +09:00
Balazs Gerofi
2f0089dfb9 mcstop+release: use ihkconfig release mem all 2017-11-20 17:06:01 +09:00
Ken Sato
2af6d5115a fix: depending arch futex_atomic_op_inuser() (a part of ARCH_DEP_8) 2017-11-20 16:42:47 +09:00
Masamichi Takagi
6c33e236d7 mcreboot: Fix umask for /proc and /sys files 2017-10-27 04:57:44 +09:00
Hitoshi Iizuka
0ecf31d896 modify:User space memory access(arm64) 2017-10-24 10:29:11 +09:00
Hitoshi Iizuka
08a625cc0d modify:User space memory access 
perf_event_open,futex,process_vm_readv,process_vm_writev,move_pages
 2017-10-23 20:27:56 +09:00
Ken Sato
12840601e1 support PERF_TYPE_{HARDWARE|HW_CACHE} in perf_event_open 
refs #829
 2017-10-20 23:10:20 +09:00
Masamichi Takagi
2ae6883a8b mcreboot.sh, mcstop+release.sh: Fix retry loop of shutdown 2017-10-19 01:54:46 +09:00
Hannes Weisbach
5b6d0a887c Add ARM64 arch_rusage header 2017-10-18 09:23:08 +09:00
Hannes Weisbach
5753db5846 Add ihk_mc_syscall_number() for ARM by reading x8 2017-10-18 09:20:52 +09:00
Hannes Weisbach
2d7cb0af89 Add copy_fp_regs to ARM (same as for x86_64) 2017-10-18 09:20:52 +09:00
Balazs Gerofi
43ecf06e83 arch: x86 -> x86_64 and build system changes 2017-10-18 09:20:52 +09:00
Takayuki Okamoto
e664ffba18 Show context registers at the interrupt by SGI 6 2017-10-18 09:20:51 +09:00
Takayuki Okamoto
3bd0137c25 Fix some race condition on arm64 
* move barrier() to architecture depended region
* add barrier() in issue_ipi, kprintf, map_virtual
* enable the workaround for cavium thunderx
 2017-10-18 09:20:51 +09:00
Hannes Weisbach
4f2b4aa402 Round the allocation for cpu-local variables up PAGE_SIZE 
Previously, this resulted in 0 pages being allocated.
 2017-10-18 09:20:51 +09:00
Hannes Weisbach
682cd34b74 Make mcstop+release architecture independent 2017-10-18 09:20:51 +09:00
Hannes Weisbach
2bc4d06a48 Add empty definition of visit_pte_range_safe() 
This is for linking only. visit_pte_range_safe() is required only
for memdump, as far as i can tell. Since memdump is disabled anyway
I think it's ok to leave this function empty for now.
 2017-10-18 09:20:51 +09:00
Hannes Weisbach
77bb3038d3 Add PT_ENTRIES macro 2017-10-18 09:20:50 +09:00
Hannes Weisbach
931448a94d Fix typo in page_align_up 2017-10-18 09:20:50 +09:00
Hannes Weisbach
c51bbbabc6 Change x86 to @ARCH@ in mcreboot-smp-x86.sh.in 
since it is used for smp-x86 and smp-arm64
 2017-10-18 09:20:50 +09:00
Dominique Martinet
3bf77446cc mcreboot-smp-x86.sh: add extra_kopts param 
This lets one specify arbitrary kernel parameters, instead of manually
fiddling with the script.
Could ultimately replace params like -t (turbo) and -d (dump_level) that
do not have any side effect (logmode starts a userland daemon)
 2017-10-13 10:02:11 +09:00
Dominique Martinet
217dd9c1e5 x86 set_signal: panic if interrupt came from kernel 
This makes debugging errors e.g. FPE from kernel much easier,
we really shouldn't be taking a user level coredump blaming user
in that case anyway
 2017-10-13 10:02:11 +09:00
Dominique Martinet
d4cd756a91 x86/cpu.c: unhandled page fault: print pre-fault stack 
Do basic manual unwinding and print raw stack addresses, with a
suggested invocation of addr2line to pretty-print the result.
 2017-10-13 10:02:11 +09:00
Dominique Martinet
b962da700b do_signal: ignore SIGWINCH 
McKernel would terminate() running program on terminal resizing
It actually looks like there is nothing for us to do when we
get that anyway (tested with `dialog`)
 2017-10-13 10:02:11 +09:00
Dominique Martinet
196379854b Fix a few more harmless compiler warnings: 
- myfree in pager.c was called with an argument, so add one to the
dummy definition
- pgoff is offset_t (unsigned) and doesn't need to be compared to 0
- clang says '*(int *)0 = 0' will be optimized away instead of keeping
the segfault without a volatile hint (?! that is wrong!), but it causes
no harm to add anyway.
 2017-10-13 10:02:11 +09:00
Dominique Martinet
4d4279121b process/vm; replace vm_range list by a rbtree 
This replaces the chained list used to keep track of all memory ranges
of a process by a standard rbtree (no need of interval tree here
because there is no overlap)

Accesses that were done directly through vm_range_list before were
replaced by lookup_process_memory_range, even full list scan (e.g.
coredump).
The full scans will thus be less efficient because calls to rb_next()
will not be inlined, but these are rarer calls that can probably afford
this compared to code simplicity.

The only reference to the actual backing structure left outside of
process.c is a call to rb_erase in xpmem_free_process_memory_range.

v2: fix lookup_process_memory_range with small start address

v3: make vm_range_insert error out properly

Panic does not lead to easy debug, all error paths
are handled to just return someting on error

v4: fix lookup_process_memory_range (again)

That optimistically going left was a more serious bug than just
last iteration, we could just pass by a match and continue down
the tree if the match was not a leaf.

v5: some users actually needed leftmost match, so restore behavior
without the breakage (hopefully)
 2017-10-13 10:00:27 +09:00