When a device model's I/O operation fails, we execute the error action. This lets layers above QEMU implement thin provisioning, or attempt to correct errors before they reach the guest. But when the I/O operation fails because its invalid, reporting the error to the guest is the only sensible action.
SCSI does that, but virtio-blk and IDE don't. Fix them. No other device model supports error actions. v4: * PATCH 4: Fix the DMA new error path, and update the commit message accordingly [Kevin] v3: * PATCH 2: Resolve semantic conflict with commit 671ec3f "virtio-blk: Convert VirtIOBlockReq.elem to pointer" [Kevin] v2: * PATCH 2: Commit message spelling fix [Fam] * PATCH 4: New, covering IDE Markus Armbruster (4): virtio-blk: Factor common checks out of virtio_blk_handle_read/write() virtio-blk: Bypass error action and I/O accounting on invalid r/w virtio-blk: Treat read/write beyond end as invalid ide: Treat read/write beyond end as invalid hw/block/virtio-blk.c | 45 +++++++++++++++++++++++++++++---------------- hw/ide/core.c | 28 ++++++++++++++++++++++++++++ 2 files changed, 57 insertions(+), 16 deletions(-) -- 1.9.3
