From: Mark Hounschell on
On 05/26/2010 02:34 PM, Mark Hounschell wrote:
> My application uses SCSI disks of 256, 512, 768, and 1024 sector sizes. It
> uses the sg_io interface to talk to these devices. As of 2.6.27 if any SCSI
> disk is attached that has been formatted with a 256 byte sector size, the
> boot process hangs when it gets to the point of discovering USB devices.
> 512, 768, and 1024 sector sizes do not seem to trigger this.
>
> I have bisected this problem to the following commit:
>
> # git bisect good
> 427e59f09fdba387547106de7bab980b7fff77be is first bad commit
> commit 427e59f09fdba387547106de7bab980b7fff77be
> Author: James Bottomley <James.Bottomley(a)HansenPartnership.com>
> Date: Sat Mar 8 18:24:17 2008 -0600
>
> [SCSI] make use of the residue value
>
> USB sometimes doesn't return an error but instead returns a residue
> value indicating part (or all) of the command wasn't completed. So if
> the driver _done() error processing indicates the command was fully
> processed, subtract off the residue so that this USB error gets
> propagated.
>
> Cc: Alan Stern <stern(a)rowland.harvard.edu>
> Signed-off-by: James Bottomley <James.Bottomley(a)HansenPartnership.com>
>
> :040000 040000 d3bad84ebe1bc231e8e7d6267907ca62fd4d0dcd
> c85f8cb8bd4910724f0101e41054555980727e16 M drivers
>
> Now, what the hell USB has to do with my SCSI disks is beyond me. I have a
> feeling that this commit is just uncovering another problem. I've attached
> a dmesg from a working kernel and pointed out where the boot hangs and my
> SCSI disk light is on solid. Once it turns on solid the only recourse is to
> recycle power to the disk. The reset button does nothing to it. In fact a
> second will not even see the device because it is out to lunch. Not even a
> SCSI bus reset clears what ever has happened to the disk.
>
> This does the same thing on a 2.6.34 kernel also. Anything I can do to
> help, I'm available.
>
> Thanks and regards
> Mark
>

I've attached the output from a serial console while trying to boot the
kernel with the above commit in. These disks have no partition table BTW.

Thanks
Mark