Date: Sun, 18 Aug 1996 22:44:33 -0400 (EDT) From: "R.D. Thrush" <rd@thrush.com> To: gibbs@freefall.freebsd.org (Justin T. Gibbs) Cc: freebsd-scsi@freebsd.org Subject: Re: scsi-diffs.960815.gz uploaded Message-ID: <199608190244.WAA10451@tarpit.thrush.com> In-Reply-To: gibbs@freefall.freebsd.org's message of 18 Aug 96 16:36:01 GMT References: Your message of "Sun, 18 Aug 1996 10:54:52 EDT." <199608181454.KAA25570@tarpit.thrush.com> <199608181636.JAA01036@freefall.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
>>>>> "gibbs" == Justin T Gibbs <gibbs@freefall.freebsd.org> writes:
>> I tried the diffs and had a system lockup. The following 2 messages
>> were on the console:
>>
>> ahc0: WARNING no command for scb2 (cmdcmplt)
>> QOUTCNT == 0
>> ahc0: WARNING no command for scb19 (cmdcmplt)
>> QOUTCNT == 0
gibbs> You're using SCB_PAGING right? Try turning it off and see if it happens
gibbs> again. I think there is a bug in the SCB_PAGING code and since the new
gibbs> code is faster, it probably exacerbated the problem. You should also
gibbs> trim down your config file. All of those unattached devices are just
gibbs> useless kernel bloat.
Thanks, I removed AHC_SCBPAGING_ENABLE from the kernel config and
have now processed several GBs of news without the forementioned
problem. I think this lends credence to your suspicion of a bug in
the SCB_PAGING code...
As mentioned in a separate email, the diffs seem to have resulted in
lossage of the "sps tps msps" stats as shown by iostat as well vmstat
disk stats.
comet3:home/rd 516>iostat 2
tty sd0 sd1 sd2 sd3 cpu
tin tout sps tps msps sps tps msps sps tps msps sps tps msps us ni sy in id
0 737 0 0 0.0 0 0 0.0 0 0 0.0 0 0 0.0 4 2 9 2 82
0 294 0 0 0.0 0 0 0.0 0 0 0.0 0 0 0.0 1 0 2 0 97
Thanks again for the prompt reply to the previous report.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199608190244.WAA10451>