sd or block: regression in 2.6.28-rc WRT manage_start_stop

Hi,

I only recently switched from 2.6.27.y to 2.6.28-rc6 and discovered the following regression with some FireWire HDDs (based on the OXUF 922, OXFW 921, OXUF 924DSB SBP-2 to IDE or SATA bridge chips).

The background is that the SBP-2 drivers switch sdev->manage_start_stop unconditionally on for all HDDs since Linux 2.6.27. For example, when I unload firewire-sbp2, the disk's spindle is stopped. When I then reload firewire-sbp2, the SCSI stack does not succeed to restart the HDD under Linux 2.6.28-rc6. It worked under 2.6.27.y.

This does not happen with some other disks (tested: HDDs based on OXFW 911, PL-3507, TSB42AA9). The difference between the failing disks and the unaffected disks is explained below.

=== example of a failing session, 2.6.26-rc8 ===

# modprobe -r firewire-sbp2

Nov 26 23:25:34 mini sd 2:0:0:0: [sdb] Synchronizing SCSI cache
Nov 26 23:25:34 mini sd 2:0:0:0: [sdb] Stopping disk
Nov 26 23:25:34 mini firewire_sbp2: released fw1.0, target 2:0:0

# modprobe firewire-sbp2

Nov 26 23:25:44 mini scsi3 : SBP-2 IEEE-1394
Nov 26 23:25:44 mini firewire_sbp2: fw1.0: logged in to LUN 0000 (0 retries)
Nov 26 23:25:44 mini scsi 3:0:0:0: Direct-Access-RBC ST340083 2A
       3.03 PQ: 0 ANSI: 4
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] 781422768 512-byte hardware
sectors: (400 GB/372 GiB)
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Write Protect is off
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Mode Sense: 11 00 00 00
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] 781422768 512-byte hardware
sectors: (400 GB/372 GiB)
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Write Protect is off
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Mode Sense: 11 00 00 00
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
Nov 26 23:25:44 mini sdb:<5>firewire_sbp2: fw1.0: sbp2_scsi_abort
Nov 26 23:25:44 mini sd 3:0:0:0: Device offlined - not ready after error
recovery
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Result: hostbyte=DID_OK
driverbyte=DRIVER_OK,SUGGEST_OK
Nov 26 23:25:44 mini end_request: I/O error, dev sdb, sector 0
Nov 26 23:25:44 mini Buffer I/O error on device sdb, logical block 0
Nov 26 23:25:44 mini sd 3:0:0:0: rejecting I/O to offline device
Nov 26 23:25:44 mini Buffer I/O error on device sdb, logical block 0
Nov 26 23:25:44 mini sd 3:0:0:0: rejecting I/O to offline device
Nov 26 23:25:44 mini Buffer I/O error on device sdb, logical block 0
Nov 26 23:25:44 mini ldm_validate_partition_table(): Disk read failed.
Nov 26 23:25:44 mini sd 3:0:0:0: rejecting I/O to offline device
Nov 26 23:25:44 mini Buffer I/O error on device sdb, logical block 0
Nov 26 23:25:44 mini sd 3:0:0:0: rejecting I/O to offline device
Nov 26 23:25:44 mini Buffer I/O error on device sdb, logical block 0
Nov 26 23:25:44 mini unable to read partition table
Nov 26 23:25:44 mini sd 3:0:0:0: [sdb] Attached SCSI disk
Nov 26 23:25:44 mini sd 3:0:0:0: Attached scsi generic sg1 type 14
Nov 26 23:25:46 mini firewire_core: phy config: card 0, new root=ffc2,
gap_count=7
Nov 26 23:25:46 mini firewire_sbp2: fw1.0: reconnected to LUN 0000 (0
retries)
Nov 26 23:25:53 mini firewire_core: phy config: card 0, new root=ffc2,
gap_count=7
Nov 26 23:25:53 mini firewire_sbp2: fw1.0: reconnected to LUN 0000 (0

retries)

=== example of a working session, 2.6.27.4, same disk ===

# modprobe -r firewire-sbp2

Nov 26 23:28:30 mini sd 2:0:0:0: [sdb] Synchronizing SCSI cache
Nov 26 23:28:30 mini sd 2:0:0:0: [sdb] Stopping disk
Nov 26 23:28:30 mini firewire_sbp2: released fw1.0, target 2:0:0

# modprobe firewire-sbp2

Nov 26 23:29:50 mini scsi4 : SBP-2 IEEE-1394
Nov 26 23:29:50 mini firewire_sbp2: fw1.0: logged in to LUN 0000 (0 retries)
Nov 26 23:29:50 mini scsi 4:0:0:0: Direct-Access-RBC ST340083 2A
       3.03 PQ: 0 ANSI: 4
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] 781422768 512-byte hardware
sectors (400088 MB)
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] Write Protect is off
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] Mode Sense: 11 00 00 00
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] 781422768 512-byte hardware
sectors (400088 MB)
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] Write Protect is off
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] Mode Sense: 11 00 00 00
Nov 26 23:29:50 mini sd 4:0:0:0: [sdb] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
Nov 26 23:29:58 mini sdb: sdb1
Nov 26 23:29:58 mini sd 4:0:0:0: [sdb] Attached SCSI disk
Nov 26 23:29:58 mini sd 4:0:0:0: Attached scsi generic sg1 type 14

Note how there are 8 seconds pause in the working session between when the cache type is determined and when the partition table is read.

Compare this with the failing session where there is a command abortion immediately after the cache type was read.

The failing disks spin up late at the moment when the "reconnected" messages are logged under 2.6.28-rc, and when they received the first READ command under 2.6.27.y. The still working disks spin up earlier or quicker.

Stefan Richter
-=====-==--- =-== ==-=-
http://arcgraph.de/sr/