[Qemu-devel] Re: [PATCH] block: Make BSG detection more sane in hdev_open()

[Nicholas A. Bellinger]

On Sun, 2010-08-22 at 10:31 +0200, Christoph Hellwig wrote:
> On Sat, Aug 21, 2010 at 04:01:15PM -0700, Nicholas A. Bellinger wrote:
> > From: Nicholas Bellinger <address@hidden>
> > 
> > Greetings hch, tomo and Co,
> 
> What tree is this against?  I can't see any specificc BSG support in qemu.
> 

The Megasas / SGL passthrough friendly qemu-kvm.git/scsi-bsg updated to
v0.12.5 is here:

http://git.kernel.org/?p=virt/kvm/nab/qemu-kvm.git;a=shortlog;h=refs/heads/scsi-bsg

and the BSG driver is here:

http://git.kernel.org/?p=virt/kvm/nab/qemu-kvm.git;a=blob;f=hw/scsi-bsg.c;hb=refs/heads/scsi-bsg


> Even more I think all this in the wrong place.  The only reason
> SG_IO support was shoe-horned into raw-posix is that we only had -drive
> and no way to specify other devices.  We now have -device and a scsi
> generic device should be just that without an attached driver, so that
> we can get rid of all the special casing for sd devices.
> 

Hmmm, Ok, I think I see what you mean here..

So dropping the explict checking in raw-posix.c to rensure a '-drive
file=/dev/bsg/H:C:T:L' used with '-device scsi-bsg' is really a BSG
major is OK.  And this should instead do similar checks to my rev1 patch
in hw/scsi-bsg.c:bsg_generic_initfn(), and just drop the bdrv_is_bsg()
(and brdv_is_sg() for hw/scsi-generic.c) stuff altogether, yes..?

Thanks!

--nab