Re: [PATCH] storvsc: get rid of homegrown copy_{to,from}_bounce_buffer()

2015-09-22 Thread Christoph Hellwig
On Tue, Sep 22, 2015 at 06:27:50PM +0200, Vitaly Kuznetsov wrote:
> Storvsc driver needs to ensure there are no 'holes' in the presented
> sg list (all segments in the middle of the list need to be of PAGE_SIZE).

I think it should instead set a virt_boundary.  That's what we added for
the NVMe driver which has the same requirements, and Sagi recently also
switched iSER to it after we ensured that flag is handled correctly by
the SG_IO ioctl.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/


Re: [PATCH] storvsc: get rid of homegrown copy_{to,from}_bounce_buffer()

2015-09-23 Thread Vitaly Kuznetsov
Christoph Hellwig  writes:

> On Tue, Sep 22, 2015 at 06:27:50PM +0200, Vitaly Kuznetsov wrote:
>> Storvsc driver needs to ensure there are no 'holes' in the presented
>> sg list (all segments in the middle of the list need to be of PAGE_SIZE).
>
> I think it should instead set a virt_boundary.  That's what we added for
> the NVMe driver which has the same requirements, and Sagi recently also
> switched iSER to it after we ensured that flag is handled correctly by
> the SG_IO ioctl.

Wow,

I checked and blk_queue_virt_boundary(sdevice->request_queue, PAGE_SIZE
- 1) seems to be solving the issue completely, no bounce buffer
required. I'll test more and send v2 with removing the rest.

-- 
  Vitaly
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/