Trond Myklebust <trond.myklebust@fys.uio.no> wrote on Wed, 14 Jun 2000
21:56:07 +0200 (CEST):
> >>>>> "A " == Alexei I Adamovich <lexa@adam.botik.ru> writes:
>
> TM> There are also a couple of other minor SMP race fixes that
> TM> should go in to the final 2.4.0. I haven't sent them to Alan,
> TM> but perhaps I should...
>
> A > Should we try them?
>
> I've sent them in to Alan. They are present in 2.4.0-test1-ac17 and
> upwards. Christian Iseli, who's had problems with lockups and hangs
> under plain 2.4.0-test1 reported that 2.4.0-test1-ac18 seemed to have
> fixed his problem.
>
> In addition, there are some small knfsd issues which Russell King
> helped me track down. They relate to the way in which we organize the
> duplicate request cache. Related to this is an issue wrt. the xid
> allocation in the client on machines without an RTC. However neither
> of these should cause you too Oops.
>
2.4.0-test1-ac18 in my case was helpless: Stress.sh returned a lot of
checksum errors but kernel even didn't notice anything. But with ac19
I had next message from syslog:
> Jun 15 10:33:38 adam kernel: iput: Aieee, semaphore in use inode 03:06/496185, count=-1012063976
Seems, it's knfsd issue (03:06 is /dev/hda6). Am I right?
This time (with ac19) I had 1 "cp" process stalled after interrupting
the scipt with <ctrl-C>.
> adam:~/benchmarks/stress # ps T -lwww
> F S UID PID PPID C PRI NI ADDR SZ WCHAN TTY TIME CMD
> 100 S 0 270 1 0 60 0 - 451 wait4 tty4 0:00 login -- root
> 100 S 0 275 270 0 71 0 - 545 wait4 tty4 0:00 -bash
! 000 D 0 1337 1 0 60 0 - 263 down tty4 0:00 cp -ax Documentation/00-INDEX
! Documentation/BUG-HUNTING Documentation/CVS Documentation/Changes
<snip> ... <snip>
! /testfs/stress/7
> 100 R 0 14934 275 0 73 0 - 650 - tty4 0:00 ps T -lwww
> adam:~/benchmarks/stress # cat /proc/1337/status
> Name: cp
> State: D (disk sleep)
> Pid: 1337
> PPid: 1
> TracerPid: 0
> Uid: 0 0 0 0
> Gid: 0 0 0 0
> FDSize: 32
> Groups: 0 1 14 15 16 17 65534
> VmSize: 1052 kB
> VmLck: 0 kB
> VmRSS: 336 kB
> VmData: 36 kB
> VmStk: 12 kB
> VmExe: 28 kB
> VmLib: 948 kB
> SigPnd: 0000000000000002
> SigBlk: 0000000000000000
> SigIgn: 8000000000000000
> SigCgt: 0000000000000000
> CapInh: 0000000000000000
> CapPrm: 00000000fffffeff
> CapEff: 00000000fffffeff
But this time (with ac18 and ac19 both) "/usr/sbin/kexportfs -au" didn't
showed sigsegv/oops.
Thanks,
Alexei I. Adamovich
----------------------------------------------------------------
Res. Centre for Multiprocessor Systems, | e-mail:
PSI RAS, Pereslavl-Zalessky 152140 Russia | lexa@adam.botik.ru
----------------------------------------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:16 EST