Bug 11619 - Hotplug doesn't work with SAS1064
Hotplug doesn't work with SAS1064
Status: CLOSED OBSOLETE
Product: IO/Storage
Classification: Unclassified
Component: SCSI
All Linux
: P1 high
Assigned To: Eric Moore
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2008-09-22 02:49 UTC by Vlad Seliverstov
Modified: 2012-10-31 12:17 UTC (History)
4 users (show)

See Also:
Kernel Version: 2.6.31
Tree: Mainline
Regression: Yes


Attachments

Description Vlad Seliverstov 2008-09-22 02:49:02 UTC
Latest working kernel version: 2.6.20
Earliest failing kernel version: 2.6.22
Hardware Environment: Any Intel-powered server with SAS1064 PCI-X SAS controller
Problem Description: I have linux-box with SAS1064 PCI-X SAS Controller, and I'm experiencing problems with hotswapping. When I detach disk from system, replace it with another one, and trying to attach it again, nothing happens. After reboot, it's OK - I have this disk attached and functioning. Interesting, that if I try to attach old disk just after detaching it's OK too. 
I've tried following Vanilla kernel builds from linux git tree: 2.6.20, 2.6.22, 2.6.24, 2.6.26, 2.6.27-rc6 and it's worked only with 2.6.20. So, I tried to bisect buggy commit between 2.6.20 and 2.6.22 versions.
Here is bad commit: df9e062ad994c4db683377b108c0dbed4690e4b0 

Steps to reproduce: 
- detach disk with:
# echo "scsi remove-single-device x y z k" > /proc/scsi/scsi
- physically remove it and insert another one;
- try to attach disk with:
# echo "scsi add-single-device x y z k" > /proc/scsi/scsi
or with
echo "- - -" > /sys/class/scsi_host/hostN/scan
Comment 1 Anonymous Emailer 2008-09-22 16:11:21 UTC
Reply-To: akpm@linux-foundation.org


(switched to email.  Please respond via emailed reply-to-all, not via the
bugzilla web interface).

On Mon, 22 Sep 2008 02:49:03 -0700 (PDT)
bugme-daemon@bugzilla.kernel.org wrote:

> http://bugzilla.kernel.org/show_bug.cgi?id=11619
> 
>            Summary: Hotplug doesn't work with SAS1064
>            Product: IO/Storage
>            Version: 2.5
>      KernelVersion: 2.6.27-rc6
>           Platform: All
>         OS/Version: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: high
>           Priority: P1
>          Component: SCSI
>         AssignedTo: linux-scsi@vger.kernel.org
>         ReportedBy: vseliverstov@yandex-team.ru
> 
> 
> Latest working kernel version: 2.6.20
> Earliest failing kernel version: 2.6.22
> Hardware Environment: Any Intel-powered server with SAS1064 PCI-X SAS
> controller
> Problem Description: I have linux-box with SAS1064 PCI-X SAS Controller, and
> I'm experiencing problems with hotswapping. When I detach disk from system,
> replace it with another one, and trying to attach it again, nothing happens.
> After reboot, it's OK - I have this disk attached and functioning. Interesting,
> that if I try to attach old disk just after detaching it's OK too. 
> I've tried following Vanilla kernel builds from linux git tree: 2.6.20, 2.6.22,
> 2.6.24, 2.6.26, 2.6.27-rc6 and it's worked only with 2.6.20. So, I tried to
> bisect buggy commit between 2.6.20 and 2.6.22 versions.
> Here is bad commit: df9e062ad994c4db683377b108c0dbed4690e4b0 
> 
> Steps to reproduce: 
> - detach disk with:
> # echo "scsi remove-single-device x y z k" > /proc/scsi/scsi
> - physically remove it and insert another one;
> - try to attach disk with:
> # echo "scsi add-single-device x y z k" > /proc/scsi/scsi
> or with
> echo "- - -" > /sys/class/scsi_host/hostN/scan
> 
> 

Eric, df9e062ad994c4db683377b108c0dbed4690e4b0 was your

commit df9e062ad994c4db683377b108c0dbed4690e4b0
Author: Eric Moore <eric.moore@lsi.com>
Date:   Mon Jan 29 09:46:21 2007 -0700

    [SCSI] fusion - serialize target resets in mptsas.c

Comment 2 Vlad Seliverstov 2008-10-14 06:11:41 UTC
Can I request progress status of this bug? I have more than 100 boxes with this controller, and I can't change faulty hard drives without reboot. And I can't use 2.6.20 on them too. :(

Andrew Morton wrote:
> (switched to email.  Please respond via emailed reply-to-all, not via the
> bugzilla web interface).
> 
> On Mon, 22 Sep 2008 02:49:03 -0700 (PDT)
> bugme-daemon@bugzilla.kernel.org wrote:
> 
>> http://bugzilla.kernel.org/show_bug.cgi?id=11619
>>
>>            Summary: Hotplug doesn't work with SAS1064
>>            Product: IO/Storage
>>            Version: 2.5
>>      KernelVersion: 2.6.27-rc6
>>           Platform: All
>>         OS/Version: Linux
>>               Tree: Mainline
>>             Status: NEW
>>           Severity: high
>>           Priority: P1
>>          Component: SCSI
>>         AssignedTo: linux-scsi@vger.kernel.org
>>         ReportedBy: vseliverstov@yandex-team.ru
>>
>>
>> Latest working kernel version: 2.6.20
>> Earliest failing kernel version: 2.6.22
>> Hardware Environment: Any Intel-powered server with SAS1064 PCI-X SAS
>> controller
>> Problem Description: I have linux-box with SAS1064 PCI-X SAS Controller, and
>> I'm experiencing problems with hotswapping. When I detach disk from system,
>> replace it with another one, and trying to attach it again, nothing happens.
>> After reboot, it's OK - I have this disk attached and functioning. Interesting,
>> that if I try to attach old disk just after detaching it's OK too. 
>> I've tried following Vanilla kernel builds from linux git tree: 2.6.20, 2.6.22,
>> 2.6.24, 2.6.26, 2.6.27-rc6 and it's worked only with 2.6.20. So, I tried to
>> bisect buggy commit between 2.6.20 and 2.6.22 versions.
>> Here is bad commit: df9e062ad994c4db683377b108c0dbed4690e4b0 
>>
>> Steps to reproduce: 
>> - detach disk with:
>> # echo "scsi remove-single-device x y z k" > /proc/scsi/scsi
>> - physically remove it and insert another one;
>> - try to attach disk with:
>> # echo "scsi add-single-device x y z k" > /proc/scsi/scsi
>> or with
>> echo "- - -" > /sys/class/scsi_host/hostN/scan
>>
>>
> 
> Eric, df9e062ad994c4db683377b108c0dbed4690e4b0 was your
> 
> commit df9e062ad994c4db683377b108c0dbed4690e4b0
> Author: Eric Moore <eric.moore@lsi.com>
> Date:   Mon Jan 29 09:46:21 2007 -0700
> 
>     [SCSI] fusion - serialize target resets in mptsas.c
> 



Comment 3 Konstantin A. Lepikhov 2009-02-16 07:23:16 UTC
ping
Comment 4 Konstantin A. Lepikhov 2009-04-06 09:13:55 UTC
I'm think it's definitely hw problem. I can reproduce this error with  LSISAS1064 A3, FwRev=01060000h but not with LSISAS1064 A4 and 01180100h firmware.
Comment 5 Jordi Ferrer 2011-01-24 08:47:09 UTC
I have the same problem with an LSISAS1064E and a newer firmware:

# lspci -v
[...]
01:00.0 SCSI storage controller: LSI Logic / Symbios Logic SAS1064ET PCI-Express Fusion-MPT SAS (rev 08)
        Subsystem: IBM Device 03bb
        Flags: bus master, fast devsel, latency 0, IRQ 16
        I/O ports at 3000 [size=256]
        Memory at 91b10000 (64-bit, non-prefetchable) [size=16K]
        Memory at 91b00000 (64-bit, non-prefetchable) [size=64K]
        Expansion ROM at 91e00000 [disabled] [size=2M]
        Capabilities: [50] Power Management version 2
        Capabilities: [68] Express Endpoint, MSI 00
        Capabilities: [98] MSI: Enable- Count=1/1 Maskable- 64bit+
        Capabilities: [b0] MSI-X: Enable- Count=1 Masked-
        Capabilities: [100] Advanced Error Reporting
        Kernel driver in use: mptsas
[...]

Amb using kernel 2.6.31:

# uname -a
Linux server 2.6.31-gentoo-r6 #3 SMP Tue Mar 9 13:18:20 CET 2010 x86_64 Intel(R) Xeon(R) CPU X3430 @ 2.40GHz GenuineIntel GNU/Linux

Firmware seems much newer:

# lsiutil

LSI Logic MPT Configuration Utility, Version 1.60, July 11, 2008

1 MPT Port found

     Port Name         Chip Vendor/Type/Rev    MPT Rev  Firmware Rev  IOC
 1.  /proc/mpt/ioc0    LSI Logic SAS1064E B3     105      011b5600     0


And I can reproduce the same error explained in the first post, but, plugging back the old hard drive, this one is detected correctly.

Any Ideas?
Comment 6 Alan 2012-10-30 15:11:20 UTC
If this is still seen on modern kernels then please re-open/update
Comment 7 Konstantin Lepikhov 2012-10-31 11:54:35 UTC
(In reply to comment #6)
> If this is still seen on modern kernels then please re-open/update

According our last check this problem doesn't occur with kernels >= 2.6.37.
Comment 8 Alan 2012-10-31 12:17:31 UTC
thanks

Note You need to log in before you can comment on or make changes to this bug.