| created | 2018-11-29T22:08:59Z |
|---|---|
| begin | 2018-06-15T00:00:00Z |
| end | 2018-06-16T00:00:00Z |
| path | src/sys |
| commits | 7 |
| date | 2018-06-15T01:09:49Z | |||
|---|---|---|---|---|
| author | cheloha | |||
| files | src/sys/kern/kern_event.c | log | diff | annotate |
| message |
kevent: don't reject timeouts greater than 100 million seconds We already clamp timeouts to 24 hours anyway, so there's no reason to set EINVAL because the caller's timeout passed an arbitrary threshold. In particular, this keeps www/chromium v67 from busy-waiting and consuming a ton of cpu as described in the bug report here: https://bugs.chromium.org/p/chromium/issues/detail?id=850450 Prompted by deraadt@, bug report found by pirofti@, testing by brynet@ and deraadt@. ok kettenis@ |
|||
| date | 2018-06-15T12:21:41Z | |||
|---|---|---|---|---|
| author | yasuoka | |||
| files | src/sys/dev/ipmi.c | log | diff | annotate |
| message |
Use IPL_MPFLOOR for ipmi mutex to prevent a interrupt which requires KERNEL_LOCK(). This fixes the panic with with witness. diff from fukaumi at soum.co.jp. input and ok mpi |
|||
| date | 2018-06-15T12:39:38Z | |||
|---|---|---|---|---|
| author | yasuoka | |||
| files | src/sys/net/if_pppx.c | log | diff | annotate |
| message |
Fix the lock order problem found by witness. Release the lock for pppx_ifs earlier not to overlap the protected section for ifnet. Also introduce pxi_ready flag to protect pxi not to be used before initialization. ok bluhm |
|||
| date | 2018-06-15T13:59:53Z | |||
|---|---|---|---|---|
| author | visa | |||
| files | src/sys/kern/kern_lock.c | log | diff | annotate |
| message |
Simplify #ifdefs. The kernel_lock symbol is no longer needed when building a uniprocessor kernel with WITNESS. OK mpi@ |
|||
| date | 2018-06-15T14:07:36Z | |||
|---|---|---|---|---|
| author | visa | |||
| files | src/sys/kern/subr_witness.c | log | diff | annotate |
| message |
Remove the witness_skipspin setting to simplify the code. OK mpi@ |
|||
| date | 2018-06-15T17:58:41Z | |||
|---|---|---|---|---|
| author | bluhm | |||
| files | src/sys/arch/i386/i386/db_trace.c | log | diff | annotate |
| src/sys/arch/i386/i386/locore.s | log | diff | annotate | |
| src/sys/arch/i386/include/frame.h | log | diff | annotate | |
| message |
Reorder trapframe/intrframe to put %ebp next to %eip and make it behave like a real call frame, thus vastly simplifying the ddb back trace logic. from hshoexer@; initially from guenther@; OK deraadt@ |
|||
| date | 2018-06-15T20:41:15Z | |||
|---|---|---|---|---|
| author | guenther | |||
| files | src/sys/arch/amd64/amd64/trap.c | log | diff | annotate |
| message |
Split out the T_PAGEFLT handling into its own function, pageflttrap(). Delete pcb_onfault check in the #GP, #NP, #AC, and #TS handlers: they can't be triggered by copyin/copyout in 64bit mode. ok kettenis@ |
|||