Re: SCSI fix (a621bac3)missing in stable

[James Bottomley <jejb@linux.vnet.ibm.com>]

On Mon, 2016-06-20 at 12:23 +0200, Jinpu Wang wrote:
> On Fri, Jun 17, 2016 at 5:51 PM, James Bottomley
> <jejb@linux.vnet.ibm.com> wrote:
> > On Fri, 2016-06-17 at 17:29 +0200, Jinpu Wang wrote:
> > > On Wed, Jun 15, 2016 at 10:04 AM, Jinpu Wang
> > > <jinpu.wang@profitbricks.com> wrote:
> > > > On Tue, Jun 14, 2016 at 10:03 PM, Greg KH <
> > > > gregkh@linuxfoundation.org> wrote:
> > > > > On Wed, Jun 08, 2016 at 06:47:40PM +0200, Jinpu Wang wrote:
> > > > > > Hi all,
> > > > > > 
> > > > > > We found the SCSI fix is missing in any stable only in
> > > > > > linux
> > > > > > 4.7-rc2
> > > > > > [PATCH] scsi_lib: correctly retry failed zero length
> > > > > > REQ_TYPE_FS
> > > > > >  commands
> > > > > > 
> > > > > > commit a621bac3044ed6f7ec5fa0326491b2d4838bfa93 upstream.
> > > > > > 
> > > > > > So we backport & tested on linux-3.12 (as we're using this
> > > > > > kernel).
> > > > > > Could you review if we did it right?
> > > > > 
> > > > > This "backport" includes a bunch of checks and comments that
> > > > > are
> > > > > not in
> > > > > the upstream kernel, so I don't really want to include it in
> > > > > the
> > > > > 3.14-stable queue unless you get the scsi maintainers to ack
> > > > > it.
> > > > > 
> > > > > thanks,
> > > > > 
> > > > > greg k-h
> > > > 
> > > > Thanks Greg!
> > > > 
> > > > Hi James, hi Martin
> > > > 
> > > > Could you give your Ack on the patch?
> > > > 
> > > > --
> > > 
> > > Ping?
> > 
> > How can I review this?  What you're proposing certainly isn't a
> > backport.  What you're actually trying to do is to port around this
> > commit:
> > 
> > ommit bc85dc500f9df9b2eec15077e5046672c46adeaa
> > Author: Christoph Hellwig <hch@lst.de>
> > Date:   Thu May 1 16:51:03 2014 +0200
> > 
> >     scsi: remove scsi_end_request
> > 
> > Have you tried simply backporting that to your tree and then
> > applying
> > to work around?
> > 
> > James
> > 
> 
> Thanks James for suggestions.
> I did backport bc85dc500, and backport a621bac3, it passed my local
> tests.
> 
> Could you review patches attached?

I'm afraid 3.12 is way too old for me to remember the current state of
play.  However, if those two patches backport unmodified and you've
tested it, then you're good to go.

James