Re: [BUG] Regression introduced with "block: split bios to max possible length"

[Keith Busch]

On Thu, Jan 21, 2016 at 02:34:28PM -0700, Jens Axboe wrote:
> On 01/21/2016 07:57 AM, Stefan Haberland wrote:
> >Hi,
> >
> >unfortunately commit e36f62042880 "block: split bios to maxpossible length"
> >breaks the DASD driver on s390.  We expect the block requests to be
> >multiple
> >of 4k in size. With the patch applied I see the requests split up in
> >multiple
> >of 512 byte and therefore the requests get rejected and lots of I/Os fail.
> 
> Sigh, that's definitely a bug. I'll take a look at it and see if I
> can get a tested fix in for 4.5-rc1. If not, we'll revert it, again.

My apologies for the trouble. I trust it really is broken, but I don't
quite see how. The patch supposedly splits the transfer to the max size
the request queue says it allows. How does the max allowed size end up
an invalid multiple?