Re: Sound recording over-run
In <[🔎] 19981111122427.A16755@wau.mis.ah.nl> Paul Slootman <paul@wau.mis.ah.nl> wrote:
>>In article <[🔎] Pine.LNX.3.96.981111120402.3229A-100000@solzhenitsyn.ens.gu.edu.au> you write:
>>What does the message
>>kernel: Sound: Recording overrun
>>mean?
>
>It means that your application can't take the data out of the kernel
>buffers quickly enough, which means that the kernel buffers overflow
>(i.e. are overrun). This can be caused by ide disks and/or too much
>memory... if you have too much memory, the periodic flush of the
>buffer cache takes too long. There's a utility you can use to tune
>those parameters (it's not hdparm, at least not the debian version;
>I'm not sure what I mean _is_ in debian...).
Thanks. The problem application is probably synaesthesia, similar to that
reported by "Alexander N. Benner" <Nikodemus@innocent.com>
I'll try to find the utility Paul Slootman refers to. Regarding the kernel
error - there is one, but not before the overrrun. /var/log/syslog shows:
Nov 10 17:08:03 solzhenitsyn kernel: Sound: Recording overrun
Nov 10 17:08:34 solzhenitsyn last message repeated 2975 times
Nov 10 17:08:39 solzhenitsyn last message repeated 718 times
Nov 10 17:08:39 solzhenitsyn kernel: Unable to handle kernel paging request at v
irtual address 00001000
Nov 10 17:08:39 solzhenitsyn kernel: current->tss.cr3 = 009b5000, %cr3 = 009b500
0
Nov 10 17:08:39 solzhenitsyn kernel: *pde = 00000000
Nov 10 17:08:39 solzhenitsyn kernel: Oops: 0002
Nov 10 17:08:39 solzhenitsyn kernel: CPU: 0
Nov 10 17:08:39 solzhenitsyn kernel: EIP: 0010:[<c01a8bb4>]
Nov 10 17:08:39 solzhenitsyn kernel: EFLAGS: 00010206
Nov 10 17:08:39 solzhenitsyn kernel: eax: 00010080 ebx: c80020d4 ecx: 000010
00 edx: 00001000
Nov 10 17:08:39 solzhenitsyn kernel: esi: 00001000 edi: 00001000 ebp: 000000
01 esp: c13bdf40
Nov 10 17:08:39 solzhenitsyn kernel: ds: 0018 es: 0018 ss: 0018 u
Nov 10 17:08:39 solzhenitsyn kernel: Process xsynaesthesia (pid: 28300,
process
nr: 64, stackpage=c13bd000)
Nov 10 17:08:39 solzhenitsyn kernel: Stack: c1eff6e0 bffffb50 c7c556c0
00000000
00001000 bffffb50 00000001 c01a8c0f
Nov 10 17:08:39 solzhenitsyn kernel: 00000000 c7c556c0 c55f2430
c01a752b
00000003 c7c556c0 c7c556c0 c0121e97
Nov 10 17:08:39 solzhenitsyn kernel: c55f2430 c7c556c0 c7c556c0
00000000
fffffff7 c0121f23 c7c556c0 c7c556c0
Nov 10 17:08:39 solzhenitsyn kernel: Call Trace: [<c01a8c0f>] [<c01a752b>]
[<c01
21e97>] [<c0121f23>] [<c0121fa4>] [<c0107a38>] [<c010002b>]
Nov 10 17:08:39 solzhenitsyn kernel: Code: f3 aa ff 44 24 20 8b 73 2c 39 74 24
2
0 7c a5 80 4b 14 80 5b
Nov 10 17:08:39 solzhenitsyn kernel: Sound: Recording overrun
Nov 10 17:09:10 solzhenitsyn last message repeated 5332 times
This raises some further questions.
1) The offending process is dead & buried. Is there some way, short of
a re-boot (which I'd like to avoid at the moment) for the kernel to
"recover" (by which I guess I mean stop producing these error messages)
2) Should I report this to the kernel list?
----
Robert King, Australian Environmental Studies, Griffith University, Australia
3875 6677 Robert.King@ens.gu.edu.au http://www.ens.gu.edu.au/robertk/
The Society for the Preservation of Tithesis commends your
and scrutable use of delible and defatigable, which are gainly,
sipid and couth. We are gruntled and consolate that you have the
ertia and eptitude to choose such putably pensible tithesis, which we parage.
Reply to: