lkml.org 
[lkml]   [2012]   [Jul]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] virtio-blk: allow toggling host cache between writeback and writethrough
Il 04/07/2012 18:26, Sasha Levin ha scritto:
> On Tue, 2012-07-03 at 15:19 +0200, Paolo Bonzini wrote:
>> diff --git a/include/linux/virtio_blk.h b/include/linux/virtio_blk.h
>> index e0edb40..18a1027 100644
>> --- a/include/linux/virtio_blk.h
>> +++ b/include/linux/virtio_blk.h
>> @@ -37,8 +37,9 @@
>> #define VIRTIO_BLK_F_RO 5 /* Disk is read-only */
>> #define VIRTIO_BLK_F_BLK_SIZE 6 /* Block size of disk is available*/
>> #define VIRTIO_BLK_F_SCSI 7 /* Supports scsi command passthru */
>> -#define VIRTIO_BLK_F_FLUSH 9 /* Cache flush command support */
>> +#define VIRTIO_BLK_F_WCE 9 /* Writeback mode enabled after reset */
>> #define VIRTIO_BLK_F_TOPOLOGY 10 /* Topology information is available */
>> +#define VIRTIO_BLK_F_CONFIG_WCE 11 /* Writeback mode available in config */
>
> Wouldn't this change break any usermode code that implements virtio-blk?

No, the change is really just clarifying the existing spec, and
mandating that virtio-blk implementations follow certain assumptions of
the Linux driver.

In particular, the Linux driver is already assuming that the host
exposes VIRTIO_BLK_F_FLUSH if and only if it exposes a volatile write
cache. This works because if you have a writeback cache, but provide no
way to flush it, the guest driver really cannot do anything about it
anyway. Might as well treat it as writethrough, i.e. blk_queue_flush(q, 0).

QEMU in fact has already behaved like that, and even called the flag
VIRTIO_BLK_F_WCACHE instead of VIRRTIO_BLK_F_FLUSH.

Paolo


\
 
 \ /
  Last update: 2012-07-04 19:01    [W:0.124 / U:0.528 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site