Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 28 Feb 2006 12:38:58 -0800
From:      othermark <atkin901@yahoo.com>
To:        freebsd-stable@freebsd.org
Subject:   Re: "sio12: 178 more interrupt-level buffer overflows" on 6.1-PRERELEASE
Message-ID:  <du2cd2$g4v$1@sea.gmane.org>
References:  <20060227203625.GA19758@intserv.int1.b.intern> <20060227221844.GA21696@intserv.int1.b.intern>

next in thread | previous in thread | raw e-mail | index | archive | help
Holger Kipp wrote:
> On Mon, Feb 27, 2006 at 09:36:25PM +0100, Holger Kipp wrote:
>> Hello,
>> 
>> This looks like PR 51982
> 
> Yes - I just recompiled my kernel with the suggested
> change to sio.c, and the problem goes away... hmm.
> 
> 
>> Feb 27 21:03:17 dialout kernel: sio12: 24 more interrupt-level buffer
>> overflows (total 433) Feb 27 21:03:56 dialout kernel: sio12: 178 more
>> interrupt-level buffer overflows (total 611) Feb 27 21:04:13 dialout
>> kernel: sio12: 71 more interrupt-level buffer overflows (total 682) Feb
>> 27 21:05:56 dialout kernel: sio12: 172 more interrupt-level buffer
>> overflows (total 854) Feb 27 21:06:06 dialout kernel: sio12: 79 more
>> interrupt-level buffer overflows (total 933) Feb 27 21:06:07 dialout
>> kernel: sio12: 4 more interrupt-level buffer overflows (total 937) Feb 27
>> 21:07:56 dialout kernel: sio12: 23 more interrupt-level buffer overflows
>> (total 960) Feb 27 21:08:06 dialout kernel: sio12: 26 more
>> interrupt-level buffer overflows (total 986)

Yep, I get that on three different machines.  The faster the machine, the
less likely to overflow.  And as you saw, increasing the multiplier fixes
the problem.  That "patch" would make a good tunable candidate either for
the multiplier or the actual buffer value.

-- 
othermark
atkin901 at nospam dot yahoo dot com
(!wired)?(coffee++):(wired);




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?du2cd2$g4v$1>