The introduction of REPORT LUNS caused virtio-scsi to send two SCSI
commands (INQUIRY and REPORT LUNS) to each target rather than just
one.
However, INQUIRY will only fail if _no_ LUN is present on that target.
If all devices have LUN>0 is present, INQUIRY will report a dummy
device. Therefore,
The first patch fixes a regression where many disks on a single SCSI
target could prevent LUN0 from booting on another target.
The second patch then exploits the new scanning algorithm to speed up
the bus scan on virtio-scsi, which has a higher limit for the target
than the other adaptors and
The introduction of REPORT LUNS caused a potential regression when
many disks are present on the same SCSI target. If LUN0 is processed
too late, SeaBIOS might run out of memory and fail to register it.
There are two kinds of problems:
- QEMU presents the results of REPORT LUNS in the opposite
On Mon, Jun 19, 2017 at 06:21:09PM +0200, Paolo Bonzini wrote:
> The introduction of REPORT LUNS caused a potential regression when
> many disks are present on the same SCSI target. If LUN0 is processed
> too late, SeaBIOS might run out of memory and fail to register it.
>
> There are two kinds