Date: Thu, 27 Mar 1997 08:10:43 +0100 From: Andreas Klemm <andreas@klemm.gtn.com> To: "Justin T. Gibbs" <gibbs@plutotech.com> Cc: Kachun Lee <kachun@zippo.com>, freebsd-current@freebsd.org Subject: Re: ahc crashes Message-ID: <19970327081043.13955@klemm.gtn.com> In-Reply-To: <199703261954.MAA29414@pluto.plutotech.com>; from Justin T. Gibbs on Wed, Mar 26, 1997 at 12:55:06PM -0700 References: <19970326201659.02877@klemm.gtn.com> <199703261954.MAA29414@pluto.plutotech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Mar 26, 1997 at 12:55:06PM -0700, Justin T. Gibbs wrote:
> >First of all congradulations, you are making big progress !
>
> What model is the drive you're dumping off of? It certainly seems
> to have problems with tagged queuing and it may be that what you really
> want to do is reduce the number of openings from 8 down to 3 in i386/scsi/
> aic7xxx.c. It may also be that you need a firmware update on that drive
> in order to properly do tagged queueing.
>
> Se my other post to current about how tagged queueing can still fail if a
> device returns QUEUE FULL status.
You seem to be right ... reducing the tags to 3 brings more stability, but
on the long run the machine also hangs ...
Although I reduced the tags to 3 the following message appears at the beginning
of the dump:
sd0: Tagged openings reduced to 3
Then this after about 30 Minutes:
ahc0: WARNING no command for scb 4 (cmdcmplt)
QOUTCNT == 1
And this seems to be a hardware error (defective sector) or ?
sd1: data overrun of 16715782 bytes detected in Data-In phase. Tag == 0x0. Forcing a retry.
sd1: Have seen Data Phase. Length = 12288. NumSGs = 3.
sg[0] - Addr 0x2cbc000 : Length 4096
sg[1] - Addr 0x2c75000 : Length 4096
sg[2] - Addr 0x2cfc000 : Length 4096
swap_pager: I/O error - pagein failed; blkno 25824, size 12288, error 5
vm_fault: pager input (probably hardware) error, PID 687 failure
ahc0: WARNING no command for scb 0 (cmdcmplt)
QOUTCNT == 1
Mar 27 05:22:04 klemm /kernel: pid 687 (dump), uid 0: exited on signal 11
Mar 27 05:22:04 klemm /kernel: pid 685 (dump), uid 0: exited on signal 11
Mar 27 05:22:04 klemm /kernel: pid 688 (dump), uid 0: exited on signal 11
Mar 27 05:22:04 klemm /kernel: pid 686 (dump), uid 0: exited on signal 11
Mar 27 05:22:04 klemm /kernel: pid 684 (dump), uid 0: exited on signal 11
Mar 27 05:22:04 klemm /kernel: pid 630 (sh), uid 0: exited on signal 11 (core dumped)
This was my patch:
RCS file: /local/CVS/src/sys/i386/scsi/aic7xxx.c,v
retrieving revision 1.109
diff -u -r1.109 aic7xxx.c
--- aic7xxx.c 1997/03/24 17:42:25 1.109
+++ aic7xxx.c 1997/03/27 04:08:21
@@ -1892,7 +1892,8 @@
if (ahc->scb_data->maxhscbs >= 16
|| (ahc->flags & AHC_PAGESCBS)) {
/* Default to 8 tags */
- xs->sc_link->opennings += 13;
+ /* XXX xs->sc_link->opennings += 13; */
+ xs->sc_link->opennings = 3;
} else {
/*
* Default to 4 tags on whimpy
@@ -1902,7 +1903,8 @@
* slots. We should really have a better
* way of providing fairness.
*/
- xs->sc_link->opennings += 2;
+ /* XXX xs->sc_link->opennings += 2; */
+ xs->sc_link->opennings = 3;
}
}
}
And I asked IBM via e-mail if I can get a firmware update ...
--
andreas@klemm.gtn.com /\/\___ Wiechers & Partner Datentechnik GmbH
Andreas Klemm ___/\/\/ Support Unix -- andreas.klemm@wup.de
pgp p-key http://www-swiss.ai.mit.edu/~bal/pks-toplev.html >>> powered by <<<
ftp://sunsite.unc.edu/pub/Linux/system/Printing/aps-491.tgz >>> FreeBSD <<<
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970327081043.13955>