lkml.org 
[lkml]   [2008]   [Jan]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Problem with ata layer in 2.6.24
Gene Heskett wrote:
> Greeting;
>
> I had to reboot early this morning due to a freezeup, and I had a
> bunch of these in the messages log:
> ==============
> Jan 27 19:42:11 coyote kernel: [42461.915961] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
> Jan 27 19:42:11 coyote kernel: [42461.915973] ata1.00: cmd ca/00:08:b1:66:46/00:00:00:00:00/e8 tag 0 dma 4096 out
> Jan 27 19:42:11 coyote kernel: [42461.915974] res 40/00:01:00:4f:c2/00:00:00:00:00/00 Emask 0x4 (timeout)
> Jan 27 19:42:11 coyote kernel: [42461.915978] ata1.00: status: { DRDY }
> Jan 27 19:42:11 coyote kernel: [42461.916005] ata1: soft resetting link
> Jan 27 19:42:12 coyote kernel: [42462.078216] ata1.00: configured for UDMA/100
> Jan 27 19:42:12 coyote kernel: [42462.078232] ata1: EH complete
> Jan 27 19:42:12 coyote kernel: [42462.090700] sd 0:0:0:0: [sda] 390721968 512-byte hardware sectors (200050 MB)
> Jan 27 19:42:12 coyote kernel: [42462.114230] sd 0:0:0:0: [sda] Write Protect is off
> Jan 27 19:42:12 coyote kernel: [42462.115079] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't
> support DPO or FUA
> ===============
> That one showed up about 2 hours ago, so I expect I'll be locked
> up again before I've managed a 24 hour uptime. This drive passed
> a 'smartctl -t long /dev/sda' with flying colors after the reboot
> this morning.
>
> Two instances were logged after I had rebooted to 2.6.24 from 2.6.24-rc8:
>
> Jan 24 20:46:33 coyote kernel: [ 0.000000] Linux version 2.6.24 (root@coyote.coyote.den) (gcc version 4.1.2 20070925
> (Red Hat 4.1.2-33)) #1 SMP Thu Jan 24 20:17:55 EST 2008
> ----
> Jan 27 02:28:29 coyote kernel: [193207.445158] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
> Jan 27 02:28:29 coyote kernel: [193207.445170] ata1.00: cmd 35/00:08:f9:24:0a/00:00:17:00:00/e0 tag 0 dma 4096 out
> Jan 27 02:28:29 coyote kernel: [193207.445172] res 40/00:01:00:4f:c2/00:00:00:00:00/00 Emask 0x4 (timeout)
> Jan 27 02:28:29 coyote kernel: [193207.445175] ata1.00: status: { DRDY }
> Jan 27 02:28:29 coyote kernel: [193207.445202] ata1: soft resetting link
> Jan 27 02:28:29 coyote kernel: [193207.607384] ata1.00: configured for UDMA/100
> Jan 27 02:28:29 coyote kernel: [193207.607399] ata1: EH complete
> Jan 27 02:28:29 coyote kernel: [193207.609681] sd 0:0:0:0: [sda] 390721968 512-byte hardware sectors (200050 MB)
> Jan 27 02:28:29 coyote kernel: [193207.619277] sd 0:0:0:0: [sda] Write Protect is off
> Jan 27 02:28:29 coyote kernel: [193207.649041] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't
> support DPO or FUA
> Jan 27 02:30:06 coyote kernel: [193304.336929] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
> Jan 27 02:30:06 coyote kernel: [193304.336940] ata1.00: cmd ca/00:20:69:22:a6/00:00:00:00:00/e7 tag 0 dma 16384 out
> Jan 27 02:30:06 coyote kernel: [193304.336942] res 40/00:01:00:4f:c2/00:00:00:00:00/00 Emask 0x4 (timeout)
> Jan 27 02:30:06 coyote kernel: [193304.336945] ata1.00: status: { DRDY }
> Jan 27 02:30:06 coyote kernel: [193304.336972] ata1: soft resetting link
> Jan 27 02:30:06 coyote kernel: [193304.499210] ata1.00: configured for UDMA/100
> Jan 27 02:30:06 coyote kernel: [193304.499226] ata1: EH complete
> Jan 27 02:30:06 coyote kernel: [193304.499714] sd 0:0:0:0: [sda] 390721968 512-byte hardware sectors (200050 MB)
> Jan 27 02:30:06 coyote kernel: [193304.499857] sd 0:0:0:0: [sda] Write Protect is off
> Jan 27 02:30:06 coyote kernel: [193304.502315] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't
> support DPO or FUA
>
> None were logged during the time I was running an -rc7 or -rc8.
>
> The previous hits on this resulted in the udma speed being downgraded
> till it was actually running in pio just before the freeze that
> required the hardware reset button.

Unfortunately there are 1001 different causes for timeouts, so we need
to drill down into the hardware, libata version, and ACPI version (most
notably).


> I'll reboot to -rc8 right now and resume. If its the drive, I should see it.
> If not, then 2.6.24 is where I'll point the finger.

There was also an ACPI update, which always affects interrupt handling
(whose symptom can sometimes be a timeout).

Definitely interesting in test results from what you describe.

Jeff




\
 
 \ /
  Last update: 2008-01-28 20:11    [W:0.117 / U:0.552 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site