Date: Wed, 20 Aug 1997 00:55:57 +0200 From: Lindgren Fredrik <fli@solnet.sollentuna.se> To: aic7xxx@FreeBSD.ORG Subject: Problems with 2.0.31-pre5 Message-ID: <51F537775B63D0119D8D00805FBEB256471415@venture.edu.sollentuna.se>
next in thread | raw e-mail | index | archive | help
Hi! I've been having some problems with my News server running 2.0.30 so I though I'd give the pre-2.0.31 versions a try since they have a newer version of the AIC7xxx drivers. The News machine is a Compaq Deskpro 6000 (PPro200) with 128Mb ram. It's got a Adaptec 3980 controller with the boot/root/overview/ history disk ( a 4 gig Barracuda) on channel A (scsi0) and a SCSI-SCSI RAID controller on channel B (scsi1) ID 6. The array controller is configured for RAID0 over seven 2Gb 7200rpm Conner disks. Lately, I've been getting alot o these (with 2.0.30): Jul 19 08:59:44 borg kernel: Couldn't get a free page..... Jul 19 08:59:45 borg kernel: eth0: Couldn't allocate a sk_buff of size 96. The machine has uptimes in order of weeks though. Well, I installed 3.0.31-pre5. In the morning I found the machine endlessly printing these on the console: Aug 16 04:13:26 borg kernel: scsi : aborting command due to timeout : pid 4705553, scsi1, channel 0, id 6, lun 0 Read (10) 00 00 2b 21 f5 00 00 04 00 Aug 16 04:13:26 borg kernel: aic7xxx: (abort) Aborting scb 6, TCL 6/0/0 Aug 16 04:13:26 borg kernel: SCSI host 1 abort (pid 4705553) timed out - resetting Aug 16 04:13:26 borg kernel: SCSI bus is being reset for host 1 channel 0. Aug 16 04:13:26 borg kernel: aic7xxx: (reset) target/channel 6/0 Aug 16 04:13:26 borg kernel: aic7xxx: (abort_reset) scb state 0x200, Data-Out phase, SCSISIGI = 0x0 Aug 16 04:13:27 borg kernel: aic7xxs being reset for host 1 channel 0. Aug 16 04:13:27 borg kernel: aic7xxx: (reset) target/channel 6/0 Aug 16 04:13:27 borg kernel: aic7xxx: (abort_reset) scb state 0x200, Data-Out phase, SCSISIGI = 0x0 Aug 16 04:13:27 borg kernel: aic7xxx: (reset_device) target/channel -1/A, active_scb 5 Aug 16 04:13:27 borg kernel: aic7xxx: (reset_channel) Resetting current channel A Aug 16 04:13:27 borg kernel: aic7xxx: (reset_channel) Channel reset, sequencer restarted Also, I still had the "Couldn't get a free page....." messages. I rebooted the machine, and it went up ok. But, when I wanted to try pre7 and ran LILO, I got the following: geo_comp_addr: Cylinder number is too big (1031 > 1023) Obviously it's been messing on the boot disk too. pre6 won't boot at all on this machine. And I haven't had the time to try and fix the boot disk yet. Anyway, any ideas? Should I try the patches posted on the list recently, if so, which pre-2.0.31kernel should I use as a base? Oh, yes. As a side note; I get these line in the syslog when doing cat /proc/scsi/aic7xxx/0 with 2.0.31-pre5: Aug 16 18:17:07 borg kernel: aic7xxx: possible overflow at loop 6:8 Are these dangerous?! :) Regards, Fredrik Lindgren System Admin, SolNET, Sweden
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51F537775B63D0119D8D00805FBEB256471415>