cvs commit: src/sys/kern kern_thread.c subr_witness.c src/sys/sys lock.h

[Marius Strobl]

On Wed, Aug 13, 2008 at 06:24:22PM +0000, Attilio Rao wrote:
> attilio     2008-08-13 18:24:22 UTC
> 
>   FreeBSD src repository
> 
>   Modified files:
>     sys/kern             kern_thread.c subr_witness.c 
>     sys/sys              lock.h 
>   Log:
>   SVN rev 181695 on 2008-08-13 18:24:22Z by attilio
>   
>   Introduce some WITNESS improvements:
>   - Speedup the lock orderings lookup modifying the witness graph from a
>     linked tree to a matrix. A table lookup caches the lock orderings in
>     order to make a O(1) access for them. Any witness object has an unique
>     index withing this lookup cache table.
>   - Reduce the lock contention on w_mtx acquiring it only when the LOR
>     actually happens and not in a sane case. In order to do this don't totally
>     flush lock lists (per-CPU spinlocks list and per-thread sleeplocks list)
>     but check for ll_count anytime we need to have to verify allocations sanity.
>   - Introduce the function witness_thread_exit() in the witness namespace which
>     should verify a thread doesn't hold any witness occurrence why exiting.
>   - Rename the sysctl debug.witness.graphs into debug.witness.fullgraph and
>     add debug.witness.badstacks which prints out stacks for LOR revealed.
>     This is implemented using the stack(9) support, which makes WITNESS to be
>     dependent by the STACK option or by the DDB (including STACK) option.
>   - Fix style(9) for src/sys/kern/subr_witness.c
>   
>   The hash table approach has been developed by Ilya Maykov on the behalf of
>   Isilon Systems which kindly released the patch.
>   Jeff Roberson, ported the patch to -CURRENT and fixed w_mtx contention, on the
>   behalf of Nokia.
>   
>   Submitted by:   Ilya Maykov <ivmaykov at gmail dot com> (Isilon Systems), jeff
>   Sponsored by:   Nokia
>   
>   Revision  Changes     Path
>   1.276     +5 -0       src/sys/kern/kern_thread.c
>   1.251     +1061 -491  src/sys/kern/subr_witness.c
>   1.74      +1 -0       src/sys/sys/lock.h

Hi Attilio,

this broke witness(4) on sparc64 (and probably all other platforms with
strict alignment requirement):

OK boot /boot/kernel/test2
/boot/kernel/test2 data=0x7cdb48+0x215c68 syms=[0x8+0x8a2b8+0x8+0x7a932]
nothing to autoload yet.
jumping to kernel entry at 0xc0078000.
GDB: no debug ports present
KDB: debugger backends: ddb
KDB: current backend: ddb
Copyright (c) 1992-2008 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
        The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 8.0-CURRENT #1: Tue Aug 19 19:49:27 CEST 2008
    marius at alchemy.franken.de:/usr/home/marius/co/build/ref/sparc64/usr/home/mar
WARNING: WITNESS option enabled, expect reduced performance.
panic: trap: memory address not aligned
cpuid = 0
KDB: enter: panic
[thread pid 0 tid 0 ]
Stopped at      0xc0300d80:     ta              %xcc, 1
db> bt
Tracing pid 0 tid 0 td 0xc07e9a80
(null)() at 0xc02c9208
(null)() at 0xc05e31d0
(null)() at 0xc0078ff8
(null)() at 0xc05df924
(null)() at 0xc05dfa70
(null)() at 0xc031ac98
(null)() at 0xc02b994c
(null)() at 0xc0554f84
(null)() at 0xc02b4c68
(null)() at 0xc02b866c
(null)() at 0xc02b86c0
(null)() at 0xc0281c8c
(null)() at 0xc0078030
db>

Could you please fix this?

Marius