Re: [Xen-devel] [PATCH 1/3] qemu-xen-traditional: use O_DIRECT to open disk images for IDE

[Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>]

On Thu, 29 Mar 2012, Ian Jackson wrote:

> Ian Campbell writes ("Re: [Xen-devel] [PATCH 1/3] qemu-xen-traditional: use 
> O_DIRECT to open disk images for IDE"):
> > commit 82787c6f689d869ad349df83ec3f58702afe00fe
> > Author: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>
> > Date:   Mon Mar 2 11:21:51 2009 +0000
> > 
> >     Override default cache mode for disk images to write-back
> >     
> >     Upstream qemu changed the default cache mode to write-through (ie,
> >     O_DSYNC) which is much slower.  We do not need this as we have
> >     explicit control of cacheing with the IDE cache control commands.
> ...
> 
> OK, so the reason for this is as follows:
> 
> There are IDE commands for controlling cacheing in the disk.  These
> are the commands 0x02 and 0x82 (near l.2666 of hw/ide.c in
> qemu-xen-unstable).  If the host has enabled the write cache, the disk
> is allowed to buffer written data in its own ram.  The host is
> expected to either disable the cache, or flush the cache
> (WIN_FLUSH_CACHE, near l.2717) as and when appropriate.
> 
> We implement this in qemu-xen-unstable as follows:
> 
> The bdrv_write calls are expected to be write-back cached.  The cache
> enable/disable commands set a variable write_cache in the ide code.
> If write_cache is 0, we run bdrv_flush after each write.  In any case
> we run bdrv_flush when we get a cache flush command.  bdrv_flush
> calls, effectively, fsync().
> 
> So I think qemu-xen-unstable the use of write-back cacheing for
> bdrv_write is correct.  (Apart from the wrinkle that the emulated
> drive cache, write_cache, is enabled by default - but I wouldn't be
> surprised if that accurately emulated the behaviour of real disks.)

You have missed my other emails on this thread about other disk formats
and related bug reports.



> How is this done in qemu-upstream ?
 
WIN_FLUSH_CACHE calls ide_flush_cache that calls bdrv_aio_flush.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel