[Qemu-devel] [PATCH 0/4] scsi-block: receive the right SCSI status on reads and writes

[Paolo Bonzini]

scsi-block uses the block layer for reads and writes in order to avoid
allocating bounce buffers as big as the transferred data.  We know how
to split a large transfer to multiple reads and writes, and thus we can
use scsi-disk.c's existing code to do I/O in multiple chunks (for non-s/g
SCSI hosts) or through the DMA helpers (for s/g SCSI hosts).

Unfortunately, this has the side effect of eating the SCSI status except
in the very few cases where we can convert an errno code back to a SCSI
status.  It puts a big wrench in persistent reservations support in the
guest, for example.

Luckily, splitting a large transfer into multiple SBC commands is just as
easy, and this is what the last patch does.  It takes the original CDB,
patches in a modified starting sector and sector count, and executes the
SCSI command through blk_aio_ioctl.  It is also easy to pass a QEMUIOVector
to SG_IO, so that s/g SCSI hosts keep the performance.

The other patches are preparatory work.  Patch 2 modifies the DMA helpers
to make the API more flexible, because the SCSI subsystem would like to
have a SCSIRequest pointer passed to the I/O function instead of the
BlockBackend pointer.  Patches 1 and 3 introduce the two callbacks that
scsi-block can override.

This conflicts with Eric's work on the byte-based API; I plan to rebase it
once those patches are in.  In the meanwhile, I would especially appreciate
the review of patch 2.

Paolo Bonzini (4):
  scsi-disk: introduce a common base class
  dma-helpers: change BlockBackend to opaque value in DMAIOFunc
  scsi-disk: introduce dma_readv and dma_writev
  scsi-block: always use SG_IO

 dma-helpers.c        |  51 +++++++++---
 hw/ide/core.c        |  12 +--
 hw/ide/internal.h    |   4 +-
 hw/ide/macio.c       |   4 +-
 hw/scsi/scsi-disk.c  | 229 ++++++++++++++++++++++++++++++++++++++++++---------
 include/sysemu/dma.h |  11 +--
 6 files changed, 247 insertions(+), 64 deletions(-)

-- 
2.5.5