| Message ID | 20161123103912.26699-1-olaf@aepfle.de |
|---|---|
| State | New |
| Headers | show |
On 11/23/2016 04:39 AM, Olaf Hering wrote: > The guest sends discard requests as u64 sector/count pairs, but the > block layer operates internally with s64/s32 pairs. The conversion > leads to IO errors in the guest, the discard request is not processed. > > domU.cfg: > 'vdev=xvda, format=qcow2, backendtype=qdisk, target=/x.qcow2' > domU: > mkfs.ext4 -F /dev/xvda > Discarding device blocks: failed - Input/output error > > Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS. > Add input range checking to avoid overflow. > > Fixes f313520 ("xen_disk: add discard support") > > Signed-off-by: Olaf Hering <olaf@aepfle.de> > --- Qualifies as a bug fix, so requesting 2.8 inclusion. Reviewed-by: Eric Blake <eblake@redhat.com>
Am 23.11.2016 um 12:40 hat Eric Blake geschrieben: > On 11/23/2016 04:39 AM, Olaf Hering wrote: > > The guest sends discard requests as u64 sector/count pairs, but the > > block layer operates internally with s64/s32 pairs. The conversion > > leads to IO errors in the guest, the discard request is not processed. > > > > domU.cfg: > > 'vdev=xvda, format=qcow2, backendtype=qdisk, target=/x.qcow2' > > domU: > > mkfs.ext4 -F /dev/xvda > > Discarding device blocks: failed - Input/output error > > > > Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS. > > Add input range checking to avoid overflow. > > > > Fixes f313520 ("xen_disk: add discard support") > > > > Signed-off-by: Olaf Hering <olaf@aepfle.de> > > --- > > Qualifies as a bug fix, so requesting 2.8 inclusion. > Reviewed-by: Eric Blake <eblake@redhat.com> Stefano, are you going to merge this or should I take a look? Kevin
On Wed, Nov 23, 2016 at 10:39:12AM +0000, Olaf Hering wrote: > The guest sends discard requests as u64 sector/count pairs, but the > block layer operates internally with s64/s32 pairs. The conversion > leads to IO errors in the guest, the discard request is not processed. > > domU.cfg: > 'vdev=xvda, format=qcow2, backendtype=qdisk, target=/x.qcow2' > domU: > mkfs.ext4 -F /dev/xvda > Discarding device blocks: failed - Input/output error > > Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS. > Add input range checking to avoid overflow. > > Fixes f313520 ("xen_disk: add discard support") > > Signed-off-by: Olaf Hering <olaf@aepfle.de> Acked-by: Anthony PERARD <anthony.perard@citrix.com>
On Wed, 23 Nov 2016, Olaf Hering wrote: > The guest sends discard requests as u64 sector/count pairs, but the > block layer operates internally with s64/s32 pairs. The conversion > leads to IO errors in the guest, the discard request is not processed. > > domU.cfg: > 'vdev=xvda, format=qcow2, backendtype=qdisk, target=/x.qcow2' > domU: > mkfs.ext4 -F /dev/xvda > Discarding device blocks: failed - Input/output error > > Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS. > Add input range checking to avoid overflow. > > Fixes f313520 ("xen_disk: add discard support") > > Signed-off-by: Olaf Hering <olaf@aepfle.de> Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> > v3: > turn tab into spaces to fix checkpatch warning > v2: > adjust overflow check > add Fixes revspec because the initial commit also failed to convert u64 to s32 > adjust summary > > hw/block/xen_disk.c | 42 ++++++++++++++++++++++++++++++++++++------ > 1 file changed, 36 insertions(+), 6 deletions(-) > > diff --git a/hw/block/xen_disk.c b/hw/block/xen_disk.c > index 3a7dc19..456a2d5 100644 > --- a/hw/block/xen_disk.c > +++ b/hw/block/xen_disk.c > @@ -660,6 +660,38 @@ static void qemu_aio_complete(void *opaque, int ret) > qemu_bh_schedule(ioreq->blkdev->bh); > } > > +static bool blk_split_discard(struct ioreq *ioreq, blkif_sector_t sector_number, > + uint64_t nr_sectors) > +{ > + struct XenBlkDev *blkdev = ioreq->blkdev; > + int64_t byte_offset; > + int byte_chunk; > + uint64_t byte_remaining, limit; > + uint64_t sec_start = sector_number; > + uint64_t sec_count = nr_sectors; > + > + /* Wrap around, or overflowing byte limit? */ > + if (sec_start + sec_count < sec_count || > + sec_start + sec_count > INT64_MAX >> BDRV_SECTOR_BITS) { > + return false; > + } > + > + limit = BDRV_REQUEST_MAX_SECTORS << BDRV_SECTOR_BITS; > + byte_offset = sec_start << BDRV_SECTOR_BITS; > + byte_remaining = sec_count << BDRV_SECTOR_BITS; > + > + do { > + byte_chunk = byte_remaining > limit ? limit : byte_remaining; > + ioreq->aio_inflight++; > + blk_aio_pdiscard(blkdev->blk, byte_offset, byte_chunk, > + qemu_aio_complete, ioreq); > + byte_remaining -= byte_chunk; > + byte_offset += byte_chunk; > + } while (byte_remaining > 0); > + > + return true; > +} > + > static int ioreq_runio_qemu_aio(struct ioreq *ioreq) > { > struct XenBlkDev *blkdev = ioreq->blkdev; > @@ -708,12 +740,10 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq) > break; > case BLKIF_OP_DISCARD: > { > - struct blkif_request_discard *discard_req = (void *)&ioreq->req; > - ioreq->aio_inflight++; > - blk_aio_pdiscard(blkdev->blk, > - discard_req->sector_number << BDRV_SECTOR_BITS, > - discard_req->nr_sectors << BDRV_SECTOR_BITS, > - qemu_aio_complete, ioreq); > + struct blkif_request_discard *req = (void *)&ioreq->req; > + if (!blk_split_discard(ioreq, req->sector_number, req->nr_sectors)) { > + goto err; > + } > break; > } > default: >
On Wed, 23 Nov 2016, Kevin Wolf wrote: > Am 23.11.2016 um 12:40 hat Eric Blake geschrieben: > > On 11/23/2016 04:39 AM, Olaf Hering wrote: > > > The guest sends discard requests as u64 sector/count pairs, but the > > > block layer operates internally with s64/s32 pairs. The conversion > > > leads to IO errors in the guest, the discard request is not processed. > > > > > > domU.cfg: > > > 'vdev=xvda, format=qcow2, backendtype=qdisk, target=/x.qcow2' > > > domU: > > > mkfs.ext4 -F /dev/xvda > > > Discarding device blocks: failed - Input/output error > > > > > > Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS. > > > Add input range checking to avoid overflow. > > > > > > Fixes f313520 ("xen_disk: add discard support") > > > > > > Signed-off-by: Olaf Hering <olaf@aepfle.de> > > > --- > > > > Qualifies as a bug fix, so requesting 2.8 inclusion. > > Reviewed-by: Eric Blake <eblake@redhat.com> > > Stefano, are you going to merge this or should I take a look? I can merge it. Cheers, Stefano
Am 23. November 2016 13:27:13 MEZ, schrieb Kevin Wolf <kwolf@redhat.com>: >Am 23.11.2016 um 12:40 hat Eric Blake geschrieben: >> Qualifies as a bug fix, so requesting 2.8 inclusion. >> Reviewed-by: Eric Blake <eblake@redhat.com> Is this a can for 2.x? Olaf
Am 23. November 2016 21:44:50 MEZ, schrieb Olaf Hering <olaf@aepfle.de>:
>Is this a can for 2.x?
candidate
Olaf
diff --git a/hw/block/xen_disk.c b/hw/block/xen_disk.c index 3a7dc19..456a2d5 100644 --- a/hw/block/xen_disk.c +++ b/hw/block/xen_disk.c @@ -660,6 +660,38 @@ static void qemu_aio_complete(void *opaque, int ret) qemu_bh_schedule(ioreq->blkdev->bh); } +static bool blk_split_discard(struct ioreq *ioreq, blkif_sector_t sector_number, + uint64_t nr_sectors) +{ + struct XenBlkDev *blkdev = ioreq->blkdev; + int64_t byte_offset; + int byte_chunk; + uint64_t byte_remaining, limit; + uint64_t sec_start = sector_number; + uint64_t sec_count = nr_sectors; + + /* Wrap around, or overflowing byte limit? */ + if (sec_start + sec_count < sec_count || + sec_start + sec_count > INT64_MAX >> BDRV_SECTOR_BITS) { + return false; + } + + limit = BDRV_REQUEST_MAX_SECTORS << BDRV_SECTOR_BITS; + byte_offset = sec_start << BDRV_SECTOR_BITS; + byte_remaining = sec_count << BDRV_SECTOR_BITS; + + do { + byte_chunk = byte_remaining > limit ? limit : byte_remaining; + ioreq->aio_inflight++; + blk_aio_pdiscard(blkdev->blk, byte_offset, byte_chunk, + qemu_aio_complete, ioreq); + byte_remaining -= byte_chunk; + byte_offset += byte_chunk; + } while (byte_remaining > 0); + + return true; +} + static int ioreq_runio_qemu_aio(struct ioreq *ioreq) { struct XenBlkDev *blkdev = ioreq->blkdev; @@ -708,12 +740,10 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq) break; case BLKIF_OP_DISCARD: { - struct blkif_request_discard *discard_req = (void *)&ioreq->req; - ioreq->aio_inflight++; - blk_aio_pdiscard(blkdev->blk, - discard_req->sector_number << BDRV_SECTOR_BITS, - discard_req->nr_sectors << BDRV_SECTOR_BITS, - qemu_aio_complete, ioreq); + struct blkif_request_discard *req = (void *)&ioreq->req; + if (!blk_split_discard(ioreq, req->sector_number, req->nr_sectors)) { + goto err; + } break; } default:
The guest sends discard requests as u64 sector/count pairs, but the block layer operates internally with s64/s32 pairs. The conversion leads to IO errors in the guest, the discard request is not processed. domU.cfg: 'vdev=xvda, format=qcow2, backendtype=qdisk, target=/x.qcow2' domU: mkfs.ext4 -F /dev/xvda Discarding device blocks: failed - Input/output error Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS. Add input range checking to avoid overflow. Fixes f313520 ("xen_disk: add discard support") Signed-off-by: Olaf Hering <olaf@aepfle.de> --- v3: turn tab into spaces to fix checkpatch warning v2: adjust overflow check add Fixes revspec because the initial commit also failed to convert u64 to s32 adjust summary hw/block/xen_disk.c | 42 ++++++++++++++++++++++++++++++++++++------ 1 file changed, 36 insertions(+), 6 deletions(-)