| Message ID | 20180710063124.2263-1-famz@redhat.com |
|---|---|
| Headers | show |
| Series | block: Fix dst reading after tail copy offloading | expand |
Am 10.07.2018 um 08:31 hat Fam Zheng geschrieben: > Based-on: 20180709163719.87107-1-vsementsov@virtuozzo.com > > v3: - Rebase onto Vladmir's: > [PATCH v5 0/4] fix image fleecing > on top of master, which has blklogwrites to be converted to BdrvChild for > the bdrv_co_pdiscard() parameter. [Kevin] > - Add file_ prefix to file protocol trace points. [Kevin] > - Assert that BdrvTrackedRequest bytes and offset don't overflow INT64_MAX. > [Kevin] > - Don't misuse req->offset/req->bytes in bdrv_co_write_req_prepare/finish. > [Kevin] > - Fix stat64_max. [Kevin] > - Keep lines within 80 columns. [Kevin] > - Grammar fix in commit message and rev-by. [Eric] > > Qcow2 allocates new clusters after the end of the file. If it is the destinaton > of copy offloading, we must adjust dst->bs->total_sectors. Otherwise, further > reads will drop to the "beyond EOF" code path and return zeroes, which problem > is caught by iotests 222. > > Follow the logic in the normal write code and update bs->total_sectors after > I/O is done. > > While at it, add a few convenient trace points to aid future debug experiences > in the topic. Thanks, applied to the block branch. Kevin
Based-on: 20180709163719.87107-1-vsementsov@virtuozzo.com v3: - Rebase onto Vladmir's: [PATCH v5 0/4] fix image fleecing on top of master, which has blklogwrites to be converted to BdrvChild for the bdrv_co_pdiscard() parameter. [Kevin] - Add file_ prefix to file protocol trace points. [Kevin] - Assert that BdrvTrackedRequest bytes and offset don't overflow INT64_MAX. [Kevin] - Don't misuse req->offset/req->bytes in bdrv_co_write_req_prepare/finish. [Kevin] - Fix stat64_max. [Kevin] - Keep lines within 80 columns. [Kevin] - Grammar fix in commit message and rev-by. [Eric] Qcow2 allocates new clusters after the end of the file. If it is the destinaton of copy offloading, we must adjust dst->bs->total_sectors. Otherwise, further reads will drop to the "beyond EOF" code path and return zeroes, which problem is caught by iotests 222. Follow the logic in the normal write code and update bs->total_sectors after I/O is done. While at it, add a few convenient trace points to aid future debug experiences in the topic. Fam Zheng (10): block: Prefix file driver trace points with "file_" block: Add copy offloading trace points block: Use BdrvChild to discard block: Use uint64_t for BdrvTrackedRequest byte fields block: Extract common write req handling block: Fix handling of image enlarging write block: Use common req handling for discard block: Use common req handling in copy offloading block: Fix bdrv_co_truncate overlap check block: Use common write req handling in truncate block/blkdebug.c | 2 +- block/blklogwrites.c | 2 +- block/blkreplay.c | 2 +- block/block-backend.c | 2 +- block/copy-on-read.c | 2 +- block/file-posix.c | 4 +- block/file-win32.c | 2 +- block/io.c | 213 ++++++++++++++++++++++++-------------- block/iscsi.c | 3 + block/mirror.c | 2 +- block/qcow2-refcount.c | 2 +- block/raw-format.c | 2 +- block/throttle.c | 2 +- block/trace-events | 10 +- include/block/block.h | 4 +- include/block/block_int.h | 4 +- 16 files changed, 163 insertions(+), 95 deletions(-)