New: 4.9.-rc4 produces hundreds of unusable scsi devices.

https://bugzilla.kernel.org/show_bug.cgi?id=187381

            Bug ID: 187381
           Summary: 4.9.-rc4 produces hundreds of unusable scsi devices.
           Product: SCSI Drivers
           Version: 2.5
    Kernel Version: 4.9-rc4
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: Other
          Assignee: [email protected]
          Reporter: [email protected]
        Regression: No

Created attachment 244091
--> https://bugzilla.kernel.org/attachment.cgi?id=244091&action=edit
Debian's kernel log file. Note that this contains functional boots with older kernels as well earlier in the file.

With both my hand compiled 4.9-rc4 from kernel.org sources and Ubuntu's 4.9.0-040900rc4-generic the kernel creates hundred of useless /dev/sd?? devices. This system is fine with various 4.4, and 4.8 kernels.

The logs are spammed with messages like these

2016-11-09T12:17:00.021765-08:00 Node001 kernel: [75590.503155] Dev sdjm:
unable to read RDB block 1
2016-11-09T12:17:00.021767-08:00 Node001 kernel: [75590.503188]  sdjm: unable
to read partition table
2016-11-09T12:17:00.021769-08:00 Node001 kernel: [75590.503191] sdjm: partition
table beyond EOD, enabling native capacity
2016-11-09T12:17:00.021776-08:00 Node001 kernel: [75590.504463] sd 1:3:126:0:
[sdjl] Sector size 0 reported, assuming 512.
2016-11-09T12:17:00.021778-08:00 Node001 kernel: [75590.504486] sd 1:3:127:0:
[sdjm] Sector size 0 reported, assuming 512.
2016-11-09T12:17:00.025711-08:00 Node001 kernel: [75590.504624] Dev sdjl:
unable to read RDB block 1
2016-11-09T12:17:00.025720-08:00 Node001 kernel: [75590.504625] Dev sdjm:
unable to read RDB block 1
2016-11-09T12:17:00.025722-08:00 Node001 kernel: [75590.504635]  sdjm: unable
to read partition table
2016-11-09T12:17:00.025724-08:00 Node001 kernel: [75590.504687] sdjm: partition
table beyond EOD, truncated
2016-11-09T12:17:00.025725-08:00 Node001 kernel: [75590.504697]  sdjl: unable
to read partition table
2016-11-09T12:17:00.025727-08:00 Node001 kernel: [75590.504701] sdjl: partition
table beyond EOD, truncated
2016-11-09T12:17:00.025728-08:00 Node001 kernel: [75590.505084] sd 1:3:127:0:
[sdjm] Sector size 0 reported, assuming 512.
2016-11-09T12:17:00.025730-08:00 Node001 kernel: [75590.505088] sd 1:3:126:0:
[sdjl] Sector size 0 reported, assuming 512.
2016-11-09T12:17:00.025732-08:00 Node001 kernel: [75590.505115] sd 1:3:126:0:
[sdjl] Attached SCSI disk
2016-11-09T12:17:00.025733-08:00 Node001 kernel: [75590.505151] sd 1:3:127:0:

[sdjm] Attached SCSI disk

You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=187381

--- Comment #1 from Samuel Flory Silbory <[email protected]> --- Created attachment 244101
--> https://bugzilla.kernel.org/attachment.cgi?id=244101&action=edit
Storcli output from system.

Note that this occurs with really old firmware and the current megaraid firmware. The megaraid configuration is 12 one drive RAID zero arrays.

You are receiving this mail because:
You are watching the assignee of the bug.

This was reported to the mailing list and should be fixed by this:

http://marc.info/?l=linux-scsi&m=147868920429684

We'll fast track this, but can you verify it fixes your issue.

Thanks,

James

Certainly I'll give it a try.
________________________________________
From: James Bottomley [[email protected]]
Sent: Wednesday, November 09, 2016 12:49 PM
To: [email protected]; [email protected]
Cc: Samuel Silbory
Subject: Re: [Bug 187381] New: 4.9.-rc4 produces hundreds of unusable scsi devices.

This was reported to the mailing list and should be fixed by this:

https://urldefense.proofpoint.com/v2/url?u=http-3A__marc.info_-3Fl-3Dlinux-2Dscsi-26m-3D147868920429684&d=CwICaQ&c=DZ-EF4pZfxGSU6MfABwx0g&r=4FOU_7YnNtQNKWSdicj67DITRyD0V0NPxX2P7qBxsr8&m=qqYheGXoM4mBPxyZZ7HM1rugSFukVAVfEjl1ifhaoqU&s=KCePcDLsNpEqob1iEoGGCb5Qf-smnMYk4XsK62Sp0h4&e=

We'll fast track this, but can you verify it fixes your issue.

Thanks,

James

Seems sane now.

root@pcsnode:~# uname -a
Linux pcsnode 4.9.0-rc4-1-default #1 SMP Wed Nov 9 13:59:28 PST 2016 x86_64 GNU/Linux
root@pcsnode:~# ls /dev/sd*
/dev/sda  /dev/sdc  /dev/sde  /dev/sdg	/dev/sdi  /dev/sdk   /dev/sdk2	/dev/sdk4  /dev/sdk6  /dev/sdl
/dev/sdb  /dev/sdd  /dev/sdf  /dev/sdh	/dev/sdj  /dev/sdk1  /dev/sdk3	/dev/sdk5  /dev/sdk7  /dev/sdm
root@pcsnode:~# 

The only difference compared to 4.8/4/4 is the boot drive on the SATA controller ended up sdk instead of sdm (the last drive). It's not an issue for me as I mount things by UUID

________________________________________
From: Samuel Silbory
Sent: Wednesday, November 09, 2016 1:19 PM
To: James Bottomley; [email protected]; [email protected]
Subject: RE: [Bug 187381] New: 4.9.-rc4 produces hundreds of unusable scsi devices.

Certainly I'll give it a try.
________________________________________
From: James Bottomley [[email protected]]
Sent: Wednesday, November 09, 2016 12:49 PM
To: [email protected]; [email protected]
Cc: Samuel Silbory
Subject: Re: [Bug 187381] New: 4.9.-rc4 produces hundreds of unusable scsi devices.

This was reported to the mailing list and should be fixed by this:

https://urldefense.proofpoint.com/v2/url?u=http-3A__marc.info_-3Fl-3Dlinux-2Dscsi-26m-3D147868920429684&d=CwICaQ&c=DZ-EF4pZfxGSU6MfABwx0g&r=4FOU_7YnNtQNKWSdicj67DITRyD0V0NPxX2P7qBxsr8&m=qqYheGXoM4mBPxyZZ7HM1rugSFukVAVfEjl1ifhaoqU&s=KCePcDLsNpEqob1iEoGGCb5Qf-smnMYk4XsK62Sp0h4&e=

We'll fast track this, but can you verify it fixes your issue.

Thanks,

James

https://bugzilla.kernel.org/show_bug.cgi?id=187381

--- Comment #2 from Samuel Flory Silbory <[email protected]> --- Certainly I'll give it a try.
________________________________________
From: James Bottomley [[email protected]]
Sent: Wednesday, November 09, 2016 12:49 PM
To: [email protected]; [email protected]
Cc: Samuel Silbory
Subject: Re: [Bug 187381] New: 4.9.-rc4 produces hundreds of unusable scsi devices.

This was reported to the mailing list and should be fixed by this:

https://urldefense.proofpoint.com/v2/url?u=http-3A__marc.info_-3Fl-3Dlinux-2Dscsi-26m-3D147868920429684&d=CwICaQ&c=DZ-EF4pZfxGSU6MfABwx0g&r=4FOU_7YnNtQNKWSdicj67DITRyD0V0NPxX2P7qBxsr8&m=qqYheGXoM4mBPxyZZ7HM1rugSFukVAVfEjl1ifhaoqU&s=KCePcDLsNpEqob1iEoGGCb5Qf-smnMYk4XsK62Sp0h4&e=

We'll fast track this, but can you verify it fixes your issue.

Thanks,

James

You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=187381

--- Comment #3 from Samuel Flory Silbory <[email protected]> --- Seems sane now.

root@pcsnode:~# uname -a
Linux pcsnode 4.9.0-rc4-1-default #1 SMP Wed Nov 9 13:59:28 PST 2016 x86_64
GNU/Linux
root@pcsnode:~# ls /dev/sd*
/dev/sda  /dev/sdc  /dev/sde  /dev/sdg    /dev/sdi  /dev/sdk   /dev/sdk2   
/dev/sdk4  /dev/sdk6  /dev/sdl
/dev/sdb  /dev/sdd  /dev/sdf  /dev/sdh    /dev/sdj  /dev/sdk1  /dev/sdk3   
/dev/sdk5  /dev/sdk7  /dev/sdm
root@pcsnode:~# 

The only difference compared to 4.8/4/4 is the boot drive on the SATA controller ended up sdk instead of sdm (the last drive). It's not an issue for me as I mount things by UUID

________________________________________
From: Samuel Silbory
Sent: Wednesday, November 09, 2016 1:19 PM
To: James Bottomley; [email protected]; [email protected]
Subject: RE: [Bug 187381] New: 4.9.-rc4 produces hundreds of unusable scsi devices.

Certainly I'll give it a try.
________________________________________
From: James Bottomley [[email protected]]
Sent: Wednesday, November 09, 2016 12:49 PM
To: [email protected]; [email protected]
Cc: Samuel Silbory
Subject: Re: [Bug 187381] New: 4.9.-rc4 produces hundreds of unusable scsi devices.

This was reported to the mailing list and should be fixed by this:

https://urldefense.proofpoint.com/v2/url?u=http-3A__marc.info_-3Fl-3Dlinux-2Dscsi-26m-3D147868920429684&d=CwICaQ&c=DZ-EF4pZfxGSU6MfABwx0g&r=4FOU_7YnNtQNKWSdicj67DITRyD0V0NPxX2P7qBxsr8&m=qqYheGXoM4mBPxyZZ7HM1rugSFukVAVfEjl1ifhaoqU&s=KCePcDLsNpEqob1iEoGGCb5Qf-smnMYk4XsK62Sp0h4&e=

We'll fast track this, but can you verify it fixes your issue.

Thanks,

James

You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.kernel.org/show_bug.cgi?id=187381

Tommy Wu <[email protected]> changed:

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

--- Comment #4 from Tommy Wu <[email protected]> --- I got same issue today.
After apply the patch, everything work fine now.

You are receiving this mail because:
You are watching the assignee of the bug.