New: TRIM support is broken for Crucial M4 drives

https://bugzilla.kernel.org/show_bug.cgi?id=36722

Summary: TRIM support is broken for Crucial M4 drives
Product: File System
Version: 2.5
Kernel Version: 2.6.38.8
Platform: All
OS/Version: Linux
Tree: Mainline
Status: NEW
Severity: normal
Priority: P1
Component: ext4
AssignedTo: [email protected]
ReportedBy: [email protected]
Regression: No

Created an attachment (id=60842)
--> (https://bugzilla.kernel.org/attachment.cgi?id=60842)
dmesg output

The TRIM is disabled as per the following messages:

# dmesg | grep "discard not"
[   18.614000] EXT4-fs warning (device sdi8): release_blocks_on_commit:2674:
discard not supported, disabling
[42486.763000] EXT4-fs warning (device sdi11): release_blocks_on_commit:2674:
discard not supported, disabling

but TRIM support is there:

# hdparm -I /dev/sdi | grep TRIM
           *    Data Set Management TRIM supported (limit 8 blocks)
           *    Deterministic read data after TRIM

The drive is on this controller and not part of any RAID:

# lspci | grep LSI

04:00.0 Serial Attached SCSI controller: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)

The dmesg is attached.

Can someone also say what this message means:

[42087.056000] EXT4-fs (sdi11): Unaligned AIO/DIO on inode 654340 by

VirtualBox; performance will be poor.

Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=36722

--- Comment #1 from Lukas Czerner <[email protected]> 2011-06-06 09:04:51 --- Created an attachment (id=60892)
--> (https://bugzilla.kernel.org/attachment.cgi?id=60892)
BLKDISCARD testing program

Hi,

thanks for reporting! Could you please provide the output of this simple test program ? Also the output of the following command.

grep $ /sys/block/sda/queue/discard_*

What kernel version are you using ?

Thanks!
-Lukas

Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=36722

--- Comment #2 from devsk <[email protected]> 2011-06-06 14:41:55 --- I am on 2.6.38.8

$ grep $ /sys/block/sdk/queue/discard_*
/sys/block/sdk/queue/discard_granularity:0
/sys/block/sdk/queue/discard_max_bytes:0
/sys/block/sdk/queue/discard_zeroes_data:0

# /tmp/discard /dev/sdk5
discard: Operation not supported

Is this because of the LSI card not passing the TRIM onto the drive? Why does hdparm report that TRIM is supported then?

Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=36722

Martin K. Petersen <[email protected]> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]

--- Comment #3 from Martin K. Petersen <[email protected]> 2011-06-06 17:43:15 --- hdparm asks the drive directly whereas the LSI card must be able to translate from SCSI to ATA commands. None the LSI cards I have support that functionality yet.

You can try sending us the output of:

# sg_readcap -l /dev/sdk

# sg_vpd -p thpv /dev/sdk

Martin K. Petersen      Oracle Linux Engineering

-- 
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=36722

Alan <[email protected]> changed:

What |Removed |Added
----------------------------------------------------------------------------

             Status|NEW                         |RESOLVED
                 CC|                            |[email protected]
         Resolution|                            |DOCUMENTED

Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching the assignee of the bug.