Re: [PATCH v2] Separate target visibility from reaped state information

[James Bottomley <James.Bottomley@HansenPartnership.com>]

On Mon, 2015-11-16 at 16:18 -0800, Bart Van Assche wrote:
> Instead of representing the states "visible in sysfs" and
> "has been removed from the target list" by a single state
> variable, use two variables to represent this information.
> 
> This patch avoids that SCSI device removal can trigger a
> soft lockup.
> 
> See also:
> * "scsi: restart list search after unlock in scsi_remove_target"
>   (commit 40998193560d).
> * "scsi_remove_target: fix softlockup regression on hot remove"
>   (commit bc3f02a795d3).

OK, could you justify this, please ... like with traces and things.

The theory on which

commit 40998193560dab6c3ce8d25f4fa58a23e252ef38
Author: Christoph Hellwig <hch@lst.de>
Date:   Mon Oct 19 16:35:46 2015 +0200

    scsi: restart list search after unlock in scsi_remove_target

Was based is that the race you're claiming to be fixing no longer exists
because it was fixed by

commit f2495e228fce9f9cec84367547813cbb0d6db15a
Author: James Bottomley <JBottomley@Parallels.com>
Date:   Tue Jan 21 07:01:41 2014 -0800

    [SCSI] dual scan thread bug fix

If that isn't the case, we can fix it, but I'd like to see the evidence.

Thanks,

James