From: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com> We are generally moving to int64_t for both offset and bytes parameters on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for fast zeroing large disk chunks, up to the whole disk. We chose signed type, to be consistent with off_t (which is signed) and with possibility for signed return type (where negative value means error). All requests in block/io must not overflow BDRV_MAX_LENGTH, all external users of BdrvTrackedRequest already have corresponding assertions, so we are safe. Add some assertions still. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com> Message-Id: <20201211183934.169161-9-vsement...@virtuozzo.com> Reviewed-by: Eric Blake <ebl...@redhat.com> Signed-off-by: Eric Blake <ebl...@redhat.com> --- include/block/block_int.h | 4 ++-- block/io.c | 14 +++++++---- block/io.c.rej | 50 ++++++++++----------------------------- 3 files changed, 24 insertions(+), 44 deletions(-) diff --git a/include/block/block_int.h b/include/block/block_int.h index 5bbbf9ee0af9..7f41f0990cc0 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -79,12 +79,12 @@ enum BdrvTrackedRequestType { typedef struct BdrvTrackedRequest { BlockDriverState *bs; int64_t offset; - uint64_t bytes; + int64_t bytes; enum BdrvTrackedRequestType type; bool serialising; int64_t overlap_offset; - uint64_t overlap_bytes; + int64_t overlap_bytes; QLIST_ENTRY(BdrvTrackedRequest) list; Coroutine *co; /* owner, used for deadlock detection */ diff --git a/block/io.c b/block/io.c index b56db913da30..1c23587d18c6 100644 --- a/block/io.c +++ b/block/io.c @@ -717,10 +717,10 @@ static void tracked_request_end(BdrvTrackedRequest *req) static void tracked_request_begin(BdrvTrackedRequest *req, BlockDriverState *bs, int64_t offset, - uint64_t bytes, + int64_t bytes, enum BdrvTrackedRequestType type) { - assert(bytes <= INT64_MAX && offset <= INT64_MAX - bytes); + bdrv_check_request(offset, bytes, &error_abort); *req = (BdrvTrackedRequest){ .bs = bs, @@ -741,8 +741,10 @@ static void tracked_request_begin(BdrvTrackedRequest *req, } static bool tracked_request_overlaps(BdrvTrackedRequest *req, - int64_t offset, uint64_t bytes) + int64_t offset, int64_t bytes) { + bdrv_check_request(offset, bytes, &error_abort); + /* aaaa bbbb */ if (offset >= req->overlap_offset + req->overlap_bytes) { return false; @@ -810,8 +812,10 @@ static void tracked_request_set_serialising(BdrvTrackedRequest *req, uint64_t align) { int64_t overlap_offset = req->offset & ~(align - 1); - uint64_t overlap_bytes = ROUND_UP(req->offset + req->bytes, align) - - overlap_offset; + int64_t overlap_bytes = + ROUND_UP(req->offset + req->bytes, align) - overlap_offset; + + bdrv_check_request(req->offset, req->bytes, &error_abort); if (!req->serialising) { qatomic_inc(&req->bs->serialising_in_flight); diff --git a/block/io.c.rej b/block/io.c.rej index f52df016263b..ae2f972d1af9 100644 --- a/block/io.c.rej +++ b/block/io.c.rej @@ -1,40 +1,16 @@ diff a/block/io.c b/block/io.c (rejected hunks) -@@ -2138,20 +2139,32 @@ int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, - return 0; - } +@@ -798,10 +800,12 @@ bool bdrv_mark_request_serialising(BdrvTrackedRequest *req, uint64_t align) + { + BlockDriverState *bs = req->bs; + int64_t overlap_offset = req->offset & ~(align - 1); +- uint64_t overlap_bytes = ROUND_UP(req->offset + req->bytes, align) +- - overlap_offset; ++ int64_t overlap_bytes = ++ ROUND_UP(req->offset + req->bytes, align) - overlap_offset; + bool waited; -+ if (!(flags & BDRV_REQ_ZERO_WRITE)) { -+ /* -+ * Pad request for following read-modify-write cycle. -+ * bdrv_co_do_zero_pwritev() does aligning by itself, so, we do -+ * alignment only if there is no ZERO flag. -+ */ -+ padded = bdrv_pad_request(bs, &qiov, &qiov_offset, &offset, &bytes, -+ &pad); -+ } ++ bdrv_check_request(req->offset, req->bytes, &error_abort); + - bdrv_inc_in_flight(bs); -- /* -- * Align write if necessary by performing a read-modify-write cycle. -- * Pad qiov with the read parts and be sure to have a tracked request not -- * only for bdrv_aligned_pwritev, but also for the reads of the RMW cycle. -- */ - tracked_request_begin(&req, bs, offset, bytes, BDRV_TRACKED_WRITE); - - if (flags & BDRV_REQ_ZERO_WRITE) { -+ assert(!padded); - ret = bdrv_co_do_zero_pwritev(child, offset, bytes, flags, &req); - goto out; - } - -- if (bdrv_pad_request(bs, &qiov, &qiov_offset, &offset, &bytes, &pad)) { -+ if (padded) { -+ /* -+ * Request was unaligned to request_alignment and therefore padded. -+ * We are going to do read-modify-write. User is not prepared to widened -+ * request intersections with other requests, so we serialize the -+ * request. -+ */ - bdrv_mark_request_serialising(&req, align); - bdrv_padding_rmw_read(child, &req, &pad, false); - } + qemu_co_mutex_lock(&bs->reqs_lock); + if (!req->serialising) { + qatomic_inc(&req->bs->serialising_in_flight); -- 2.30.0