by Hi,
I have some new issues on machine with AIC7880 that has been stable for
almost 5 years, and also after the last updates half a year ago. These
troubles have started recently when I removed a Permedia2 video card
(PCI), but the card has been put back, yet the troubles persist (I don't
think it's related, but I mention it anyways), so I think this is a
coincidence.
After some resets, the kernel discards /dev/sdb and /dev/sg1 and returns
EIO for all accesses to the partitions mounted from that drive, but
remains otherwise stable.
I'm pasting the kernel log below. I wonder if this is:
a) a hint to a driver bug that by sheer luck hasn't hit me so far
(I'd rebuild a more recent kernel in that case -- but this is a
production machine, so I can't play try this try that games. "Try
this, it should have been fixed" games are fine though.)
b) hardware deterioration (but sformat -verify hasn't found any issues
on any of the drives attached) -- the SCSI bus configuration has not
changed in years. "smartctl -a" 5.1 reports S.M.A.R.T. "GOOD" condition.
c) firmware bug in hard disk drives (unchanged for years).
Any insights appreciated. (I'm stripping date and host and kernel: tag, all
messages were logged in the same second.)
scsi0:A:2: no active SCB for reconnecting target - issuing BUS DEVICE RESET
SAVED_SCSIID == 0x27, SAVED_LUN == 0x0, ARG_1 == 0x3e ACCUM = 0x0
SEQ_FLAGS == 0x0, SCBPTR == 0xf, BTT == 0xff, SINDEX == 0x31
SCSIID == 0x27, SCB_SCSIID == 0x27, SCB_LUN == 0x0, SCB_TAG == 0xff, SCB_CONTROL == 0x68
SCSIBUSL == 0x3e, SCSISIGI == 0xe6
SXFRCTL0 == 0x88
SEQCTL == 0x10
scsi0: Dumping Card State in Message-in phase, at SEQADDR 0x1b6
ACCUM = 0x0, SINDEX = 0x31, DINDEX = 0xc0, ARG_2 = 0xf
HCNT = 0x0 SCBPTR = 0xf
SCSISEQ = 0x12, SBLKCTL = 0x2
DFCNTRL = 0x4, DFSTATUS = 0x6d
LASTPHASE = 0xe0, SCSISIGI = 0xe6, SXFRCTL0 = 0x88
SSTAT0 = 0x7, SSTAT1 = 0x3
STACK == 0x190, 0x160, 0xe4, 0x13b
SCB count = 68
Kernel NEXTQSCB = 62
Card NEXTQSCB = 62
QINFIFO entries:
Waiting Queue entries:
Disconnected Queue entries: 12:1
QOUTFIFO entries:
Sequencer Free SCB List: 15 0 8 5 10 3 6 2 13 9 11 7 1 14 4
Sequencer SCB Info: 0(c 0x68, s 0x27, l 0, t 0xff) 1(c 0x68, s 0x7, l 0, t 0xff) 2(c 0x68, s 0x7, l 0, t 0xff) 3(c 0x68, s 0x7, l 0, t
0xff) 4(c 0x68, s 0x7, l 0, t 0xff) 5(c 0x68, s 0x27, l 0, t 0xff) 6(c 0x68, s 0x7, l 0, t 0xff) 7(c 0x68, s 0x7, l 0, t 0xff) 8(c 0x68, s 0x27, l 0, t 0xff) 9(c 0x
68, s 0x7, l 0, t 0xff) 10(c 0x68, s 0x7, l 0, t 0xff) 11(c 0x68, s 0x7, l 0, t 0xff) 12(c 0x6c, s 0x27, l 0, t 0x1) 13(c 0x68, s 0x7, l 0, t 0xff) 14(c 0x68, s 0x7,
l 0, t 0xff) 15(c 0x68, s 0x27, l 0, t 0xff)
Pending list: 1(c 0x68, s 0x27, l 0)
Kernel Free SCB list: 48 30 57 31 13 15 24 11 58 26 49 22 60 59 43 23 38 51 36 19 41 46 61 21 63 3 17 14 34 0 20 55 53 42 2 10 12 8 28
6 9 56 5 54 44 47 27 16 32 52 50 45 67 35 18 7 39 25 4 40 29 37 33 66 65 64
DevQ(0:0:0): 0 waiting
DevQ(0:2:0): 0 waiting
DevQ(0:4:0): 0 waiting
scsi0: Bus Device Reset on A:2. 1 SCBs aborted
--
Matthias Andree
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/