Re: [Qemu-devel] [RFC PATCH v1 0/5] Enable virtio-scsi boot from /dev/sgX

[Eric Farman]

On 05/05/2017 03:41 AM, Fam Zheng wrote:
> On Wed, 04/26 16:46, Eric Farman wrote:
> > The short version of what happens is the host device driver rejects our
> > requests because the transfer lengths are too long for it to satisfy.
> > A virtio-scsi disk connected via scsi-generic is fine as a non-boot device
> > because the guest kernel is able to break up the requests for us.  So we just
> > need to handle this situation for the boot process.
> >
> > Patches 2-N in this series do that, but rely on us to specify the max_sectors
> > parameter for the virtio-scsi-ccw device:
> >
> >   /usr/bin/qemu-system-s390x ...
> >     -device virtio-scsi-ccw,id=scsi0,devno=fe.0.0001,max_sectors=2048
>
> Can you instead do an INQUIRY from the bios code to check the Block Limits page?
> The response is intercepted by hw/scsi/scsi-generic.c to merge in the host LUN's
> limits. That's how Linux kernel finds the granularity for request splitting.

It's a good idea (Thanks Christian, Paolo, and Fam :), but this leads to 
other difficulties.

We get a value of x3fffff when sending that to a scsi-disk from bios 
code.  That's fully emulated though, in scsi_disk_emulate_inquiry.  And 
that's the scenario that already works.

While there is indeed code in hw/scsi/scsi-generic.c to wire that in, 
that only happens after the I/O goes to the device itself.  The Block 
Limits page isn't supported [1] and thus it gets rejected with "invalid 
field in cdb".  We never get to that fixup code you reference, since the 
returned len is zero.

Should I be refactoring this code to always patch in that block limit 
regardless of a response from the host/device?  (That is, when page xb0 
isn't supported by the hw.)

 - Eric

[1] If I issue an EVPD page x00 from the QEMU bios code, I only see 
pages xb1, xc0, and c1 are supported.  If I look at the supported pages 
from the host, I see a few more but still not xb0:

$ sg_inq --page=0 /dev/sda
VPD INQUIRY: Supported VPD pages page
   Supported VPD pages:
     0x0        Supported VPD pages
     0x80       Unit serial number
     0x83       Device identification
     0x86       Extended INQUIRY data
     0xb1       Block device characteristics (sbc3)
     0xc0       vendor: Firmware numbers (seagate); Unit path report (EMC)
     0xc1       vendor: Date code (seagate)
$ sg_inq --page=0 /dev/sg0
VPD INQUIRY: Supported VPD pages page
   Supported VPD pages:
     0x0        Supported VPD pages
     0x80       Unit serial number
     0x83       Device identification
     0x86       Extended INQUIRY data
     0xb1       Block device characteristics (sbc3)
     0xc0       vendor: Firmware numbers (seagate); Unit path report (EMC)
     0xc1       vendor: Date code (seagate)


> That way, patch 1 is not necessary too. I don't like it because it doesn't
> always work considering LUN hotplug.
>
> Fam