oops and lockups with aic79xx 2.6.15+

I don't know enough about this problem to create a bugzilla report yet.

I didn't notice any problems with my Toshiba scsi DVD-ROM SD-M1401 until upgrading to 2.6.15.1 from 2.6.13.4. (this system, a Tyan Thunder K7X-Pro (S2469UNG) dual 2400mp, has been stable since 2.4.18). Now, most cd/dvd's I try cause a long listing of scsi errors in syslog. It could be that the drive is going bad, but the system no longer gracefully recovers from scsi errors. The other scsi drive, a YAMAHA CRW2100S cd-rw does not generate any read errors on all the cd's I have tried. Both 2.6.15.1 and 2.6.16.12 are causing system lockups. I have rebooted back into 2.6.13.4 and my system behaves normally, ie, CD read errors are simply reported to the shell, and processing continues.

In 2.6.15.1 I started to seeing these:

Apr 30 22:13:12 quickest automount[29812]: mount(generic): mounted /dev/dvdrom type iso9660 on /mnt/cdrom
Apr 30 22:13:58 quickest kernel: sr 1:0:2:0: Attempting to queue an ABORT message:CDB: 0x28 0x0 0x0 0x0 0x0 0x1e 0x0 0x0 0x1 0x0
Apr 30 22:13:58 quickest kernel: scsi1: At time of recovery, card was not paused
Apr 30 22:13:58 quickest kernel: >>>>>>>>>>>>>>>>>> Dump Card State Begins <<<<<<<<<<<<<<<<<
Apr 30 22:13:58 quickest kernel: scsi1: Dumping Card State at program address 0x94 Mode 0x11
Apr 30 22:13:58 quickest kernel: Card was paused
Apr 30 22:13:58 quickest kernel: HS_MAILBOX[0x0] INTCTL[0x80]:(SWTMINTMASK) SEQINTSTAT[0x0]
Apr 30 22:13:58 quickest kernel: SAVED_MODE[0x11] DFFSTAT[0x31]:(CURRFIFO_1|FIFO0FREE|FIFO1FREE)
Apr 30 22:13:58 quickest kernel: SCSISIGI[0x0]:(P_DATAOUT) SCSIPHASE[0x0] SCSIBUS[0x0]
Apr 30 22:13:58 quickest kernel: LASTPHASE[0x1]:(P_DATAOUT|P_BUSFREE) SCSISEQ0[0x0]
Apr 30 22:13:58 quickest kernel: SCSISEQ1[0x12]:(ENAUTOATNP|ENRSELI) SEQCTL0[0x10]:(FASTMODE)
Apr 30 22:13:58 quickest kernel: SEQINTCTL[0x80]:(INTVEC1DSL) SEQ_FLAGS[0xc0]:(NO_CDB_SENT|NOT_IDENTIFIED)
Apr 30 22:13:58 quickest kernel: SEQ_FLAGS2[0x0] SSTAT0[0x0] SSTAT1[0x0] SSTAT2[0x0]
Apr 30 22:13:58 quickest kernel: SSTAT3[0x0] PERRDIAG[0xc0]:(HIPERR|HIZERO) SIMODE1[0xa4]:(ENSCSIPERR|ENSCSIRST|ENSELTIMO)
Apr 30 22:13:58 quickest kernel: LQISTAT0[0x0] LQISTAT1[0x0] LQISTAT2[0x0] LQOSTAT0[0x0]
Apr 30 22:13:58 quickest kernel: LQOSTAT1[0x0] LQOSTAT2[0x0]
Apr 30 22:13:58 quickest kernel:
Apr 30 22:13:58 quickest kernel: SCB Count = 4 CMDS_PENDING = 1 LASTSCB 0x2 CURRSCB 0x2 NEXTSCB 0x0
Apr 30 22:13:58 quickest kernel: qinstart = 30690 qinfifonext = 30690
Apr 30 22:13:58 quickest kernel: QINFIFO:
Apr 30 22:13:58 quickest kernel: WAITING_TID_QUEUES:
Apr 30 22:13:58 quickest kernel: Pending list:
Apr 30 22:13:58 quickest kernel:   3 FIFO_USE[0x0] SCB_CONTROL[0x44]:(DISCONNECTED|DISCENB)
Apr 30 22:13:58 quickest kernel: SCB_SCSIID[0x27]
Apr 30 22:13:58 quickest kernel: Total 1
Apr 30 22:13:58 quickest kernel: Kernel Free SCB list: 2 1 0
Apr 30 22:13:58 quickest kernel: Sequencer Complete DMA-inprog list:
Apr 30 22:13:58 quickest kernel: Sequencer Complete list:
Apr 30 22:13:58 quickest kernel: Sequencer DMA-Up and Complete list:
Apr 30 22:13:58 quickest kernel:
Apr 30 22:13:58 quickest kernel: scsi1: FIFO0 Free, LONGJMP == 0x80ff, SCB 0x0
Apr 30 22:13:58 quickest kernel: SEQIMODE[0x3f]:(ENCFG4TCMD|ENCFG4ICMD|ENCFG4TSTAT|ENCFG4ISTAT|ENCFG4DATA|ENSAVEPTRS) 

Apr 30 22:13:58 quickest kernel: SEQINTSRC[0x0] DFCNTRL[0x0] DFSTATUS[0x89]:(FIFOEMP|HDONE|PRELOAD_AVAIL)
Apr 30 22:13:58 quickest kernel: SG_CACHE_SHADOW[0x2]:(LAST_SEG) SG_STATE[0x0] DFFSXFRCTL[0x0]
Apr 30 22:13:58 quickest kernel: SOFFCNT[0x0] MDFFSTAT[0x5]:(FIFOFREE|DLZERO) SHADDR = 0x00, SHCNT = 0x0
Apr 30 22:13:58 quickest kernel: HADDR = 0x00, HCNT = 0x0 CCSGCTL[0x10]:(SG_CACHE_AVAIL)
Apr 30 22:13:58 quickest kernel: scsi1: FIFO1 Free, LONGJMP == 0x81ec, SCB 0x2
Apr 30 22:13:59 quickest kernel: SEQIMODE[0x3f]:(ENCFG4TCMD|ENCFG4ICMD|ENCFG4TSTAT|ENCFG4ISTAT|ENCFG4DATA|ENSAVEPTRS) 

Apr 30 22:13:59 quickest kernel: SEQINTSRC[0x0] DFCNTRL[0x0] DFSTATUS[0x89]:(FIFOEMP|HDONE|PRELOAD_AVAIL)
Apr 30 22:13:59 quickest kernel: SG_CACHE_SHADOW[0x2]:(LAST_SEG) SG_STATE[0x0] DFFSXFRCTL[0x0]
Apr 30 22:13:59 quickest kernel: SOFFCNT[0x0] MDFFSTAT[0x5]:(FIFOFREE|DLZERO) SHADDR = 0x00, SHCNT = 0x0
Apr 30 22:13:59 quickest kernel: HADDR = 0x00, HCNT = 0x0 CCSGCTL[0x10]:(SG_CACHE_AVAIL)
Apr 30 22:13:59 quickest kernel: LQIN: 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x0
Apr 30 22:13:59 quickest kernel: scsi1: LQISTATE = 0x0, LQOSTATE = 0x0, OPTIONMODE = 0x42
Apr 30 22:13:59 quickest kernel: scsi1: OS_SPACE_CNT = 0x20 MAXCMDCNT = 0x0
Apr 30 22:13:59 quickest kernel:
Apr 30 22:13:59 quickest kernel: SIMODE0[0xc]:(ENOVERRUN|ENIOERR)
Apr 30 22:13:59 quickest kernel: CCSCBCTL[0x0]
Apr 30 22:13:59 quickest kernel: scsi1: REG0 == 0x3, SINDEX = 0x122, DINDEX = 0xe1
Apr 30 22:13:59 quickest kernel: scsi1: SCBPTR == 0x2, SCB_NEXT == 0xff00, SCB_NEXT2 == 0xff44
Apr 30 22:13:59 quickest kernel: CDB 43 2 0 80 9 f8
Apr 30 22:13:59 quickest kernel: STACK: 0x29 0x15 0x0 0x0 0x0 0x0 0x0 0x0
Apr 30 22:13:59 quickest kernel: <<<<<<<<<<<<<<<<< Dump Card State Ends >>>>>>>>>>>>>>>>>>
Apr 30 22:13:59 quickest kernel: (scsi1:A:2:0): Device is disconnected, re-queuing SCB
Apr 30 22:13:59 quickest kernel: Recovery code sleeping
Apr 30 22:13:59 quickest kernel: (scsi1:A:2:0): Abort Message Sent
Apr 30 22:13:59 quickest kernel: (scsi1:A:2:0): SCB 3 - Abort Tag Completed.
Apr 30 22:13:59 quickest kernel: Recovery SCB completes
Apr 30 22:13:59 quickest kernel: found == 0x1
Apr 30 22:13:59 quickest kernel: Recovery code awake

Here is the oops from 2.6.16.12:

quickest kernel: Oops: 0000 [#1]
quickest kernel: PREEMPT SMP
quickest kernel: CPU:    0
quickest kernel: EIP is at ahd_linux_run_command+0x201/0x648
quickest kernel: eax: f45ac380   ebx: 00000001   ecx: 00000000   edx: f45ac380
quickest kernel: esi: f45ac3d0   edi: 00000000   ebp: f7c93e60   esp: e8fa5b28
quickest kernel: ds: 007b   es: 007b   ss: 0068
quickest kernel: Process xmms (pid: 17463, threadinfo=e8fa4000 task=c9966560)
quickest kernel: Stack: <0>00000000 c1708e00 00000000 00200202 f747a000 f45ac380 f7c9b368 00200002
quickest kernel:        c02a7fb7 f747a000 f7c9b368 f45ac380 00001055 00200202 f7c9b800 f45ac380
quickest kernel:        00000000 c027d129 f45ac380 c027d324 00000001 f7c37794 f7c9b000 f7ca004c
quickest kernel: Call Trace:
quickest kernel:  [<c02a7fb7>] ahd_linux_queue+0x7b/0x84
quickest kernel:  [<c027d129>] scsi_dispatch_cmd+0x135/0x1ea
quickest kernel:  [<c027d324>] scsi_done+0x0/0x1c
quickest kernel:  [<c0281deb>] scsi_request_fn+0x1ad/0x2c8
quickest kernel:  [<c02073df>] elv_insert+0x147/0x14a
quickest kernel:  [<c02074a9>] elv_add_request+0x31/0x46
quickest kernel:  [<c020a1da>] blk_execute_rq_nowait+0x36/0x44
quickest kernel:  [<c020a270>] blk_execute_rq+0x88/0xb0
quickest kernel:  [<c020a43c>] blk_end_sync_rq+0x0/0x22
quickest kernel:  [<c01555db>] bio_phys_segments+0x1b/0x20
quickest kernel:  [<c020b3b8>] blk_rq_bio_prep+0x2a/0x86
quickest kernel:  [<c02b31ec>] cdrom_read_cdda_bpc+0x15e/0x1bc
quickest kernel:  [<c02b326d>] cdrom_read_cdda+0x23/0x90
quickest kernel:  [<c02b4285>] mmc_ioctl+0x457/0x8c6
quickest kernel:  [<f962cb3c>] _nv000419rm+0x24/0x28 [nvidia]
quickest kernel:  [<f9525f21>] _nv002668rm+0x1d/0x2c [nvidia]
quickest kernel:  [<f9525f2a>] _nv002668rm+0x26/0x2c [nvidia]
quickest kernel:  [<c020d9f7>] scsi_cmd_ioctl+0x7f/0x3a0
quickest kernel:  [<f969bd57>] _nv005436rm+0x23/0x28 [nvidia]
quickest kernel:  [<f9525f21>] _nv002668rm+0x1d/0x2c [nvidia]
quickest kernel:  [<f9525f21>] _nv002668rm+0x1d/0x2c [nvidia]
quickest kernel:  [<f9525f2a>] _nv002668rm+0x26/0x2c [nvidia]
quickest kernel:  [<f969bd57>] _nv005436rm+0x23/0x28 [nvidia]
quickest kernel:  [<f969bd57>] _nv005436rm+0x23/0x28 [nvidia]
quickest kernel:  [<f9525f21>] _nv002668rm+0x1d/0x2c [nvidia]
quickest kernel:  [<f9525f2a>] _nv002668rm+0x26/0x2c [nvidia]
quickest kernel:  [<c02b3d26>] cdrom_ioctl+0xa4c/0xad2
quickest kernel:  [<f969b4ca>] _nv005201rm+0x8a/0x94 [nvidia]
quickest kernel:  [<f953d576>] rm_set_interrupts+0x12e/0x164 [nvidia]
quickest kernel:  [<c0218136>] fast_clear_page+0xa/0x5e
quickest kernel:  [<c01395cb>] buffered_rmqueue+0x14d/0x210
quickest kernel:  [<c02af2b7>] sr_block_ioctl+0x3f/0x5a
quickest kernel:  [<c020bf67>] blkdev_driver_ioctl+0x3f/0x58
quickest kernel:  [<c020c001>] blkdev_ioctl+0x81/0x1c2
quickest kernel:  [<c015778c>] block_ioctl+0x1a/0x1e
quickest kernel:  [<c0160141>] do_ioctl+0x55/0x68
quickest kernel:  [<c01602a4>] vfs_ioctl+0x58/0x190
quickest kernel:  [<c0160439>] sys_ioctl+0x5d/0x64
quickest kernel:  [<c01029ed>] syscall_call+0x7/0xb
quickest kernel: Code: b7 83 86 00 00 00 66 85 c0 0f 84 43 03 00 00 8b 7b 70 0f b7 d8 8b 44 24 2c 83 78 3c 03 0f 84 22 03 00 00 85 db 0f 84 f7 02 00 00 <83> 7f 0c 00 0f 84 ed 02 00 00 85 db 74 29 31 c9 89 da 8b 04 0f

Please advise,

Kyle Davenport - unix sys admin consultant - Dallas TX
_____________________________
One OS to fool them all
One browser to find them
One email client to bring them all
And through security holes, blind them...
- Paul Johnson

Please accept my apologies if I get some of this wrong - I'm still a kernelnewbie.

Do you get lockups with 2.6.14? There's been some changes to the scsi code around this time such as using scsi_execute_req() and scsi_execute(). It's possible something broke when the new API was introduced - I've got a lockup with kernels 2.6.14 and above (though I don't know the exact cause).

-snip-

The call trace shows calls to functions from the 'nvidia' driver (if I've understood that correctly). Is this the nVidia graphics driver? That driver might be buggy, and causing the lockup. You might want to try reading the CD when the 'nvidia' driver is unloaded.

Hope this helps,

Srdjan Todorovic