diff options
| -rw-r--r-- | block/quorum.c | 177 |
1 files changed, 129 insertions, 48 deletions
diff --git a/block/quorum.c b/block/quorum.c index 0de07bb036..0160fe3ae9 100644 --- a/block/quorum.c +++ b/block/quorum.c @@ -29,6 +29,7 @@ #define QUORUM_OPT_VOTE_THRESHOLD "vote-threshold" #define QUORUM_OPT_BLKVERIFY "blkverify" #define QUORUM_OPT_REWRITE "rewrite-corrupted" +#define QUORUM_OPT_READ_PATTERN "read-pattern" /* This union holds a vote hash value */ typedef union QuorumVoteValue { @@ -79,6 +80,8 @@ typedef struct BDRVQuorumState { bool rewrite_corrupted;/* true if the driver must rewrite-on-read corrupted * block if Quorum is reached. */ + + QuorumReadPattern read_pattern; } BDRVQuorumState; typedef struct QuorumAIOCB QuorumAIOCB; @@ -122,6 +125,7 @@ struct QuorumAIOCB { bool is_read; int vote_ret; + int child_iter; /* which child to read in fifo pattern */ }; static bool quorum_vote(QuorumAIOCB *acb); @@ -148,7 +152,6 @@ static AIOCBInfo quorum_aiocb_info = { static void quorum_aio_finalize(QuorumAIOCB *acb) { - BDRVQuorumState *s = acb->common.bs->opaque; int i, ret = 0; if (acb->vote_ret) { @@ -158,7 +161,8 @@ static void quorum_aio_finalize(QuorumAIOCB *acb) acb->common.cb(acb->common.opaque, ret); if (acb->is_read) { - for (i = 0; i < s->num_children; i++) { + /* on the quorum case acb->child_iter == s->num_children - 1 */ + for (i = 0; i <= acb->child_iter; i++) { qemu_vfree(acb->qcrs[i].buf); qemu_iovec_destroy(&acb->qcrs[i].qiov); } @@ -261,6 +265,21 @@ static void quorum_rewrite_aio_cb(void *opaque, int ret) quorum_aio_finalize(acb); } +static BlockDriverAIOCB *read_fifo_child(QuorumAIOCB *acb); + +static void quorum_copy_qiov(QEMUIOVector *dest, QEMUIOVector *source) +{ + int i; + assert(dest->niov == source->niov); + assert(dest->size == source->size); + for (i = 0; i < source->niov; i++) { + assert(dest->iov[i].iov_len == source->iov[i].iov_len); + memcpy(dest->iov[i].iov_base, + source->iov[i].iov_base, + source->iov[i].iov_len); + } +} + static void quorum_aio_cb(void *opaque, int ret) { QuorumChildRequest *sacb = opaque; @@ -268,6 +287,21 @@ static void quorum_aio_cb(void *opaque, int ret) BDRVQuorumState *s = acb->common.bs->opaque; bool rewrite = false; + if (acb->is_read && s->read_pattern == QUORUM_READ_PATTERN_FIFO) { + /* We try to read next child in FIFO order if we fail to read */ + if (ret < 0 && ++acb->child_iter < s->num_children) { + read_fifo_child(acb); + return; + } + + if (ret == 0) { + quorum_copy_qiov(acb->qiov, &acb->qcrs[acb->child_iter].qiov); + } + acb->vote_ret = ret; + quorum_aio_finalize(acb); + return; + } + sacb->ret = ret; acb->count++; if (ret == 0) { @@ -348,19 +382,6 @@ static bool quorum_rewrite_bad_versions(BDRVQuorumState *s, QuorumAIOCB *acb, return count; } -static void quorum_copy_qiov(QEMUIOVector *dest, QEMUIOVector *source) -{ - int i; - assert(dest->niov == source->niov); - assert(dest->size == source->size); - for (i = 0; i < source->niov; i++) { - assert(dest->iov[i].iov_len == source->iov[i].iov_len); - memcpy(dest->iov[i].iov_base, - source->iov[i].iov_base, - source->iov[i].iov_len); - } -} - static void quorum_count_vote(QuorumVotes *votes, QuorumVoteValue *value, int index) @@ -620,34 +641,62 @@ free_exit: return rewrite; } -static BlockDriverAIOCB *quorum_aio_readv(BlockDriverState *bs, - int64_t sector_num, - QEMUIOVector *qiov, - int nb_sectors, - BlockDriverCompletionFunc *cb, - void *opaque) +static BlockDriverAIOCB *read_quorum_children(QuorumAIOCB *acb) { - BDRVQuorumState *s = bs->opaque; - QuorumAIOCB *acb = quorum_aio_get(s, bs, qiov, sector_num, - nb_sectors, cb, opaque); + BDRVQuorumState *s = acb->common.bs->opaque; int i; - acb->is_read = true; - for (i = 0; i < s->num_children; i++) { - acb->qcrs[i].buf = qemu_blockalign(s->bs[i], qiov->size); - qemu_iovec_init(&acb->qcrs[i].qiov, qiov->niov); - qemu_iovec_clone(&acb->qcrs[i].qiov, qiov, acb->qcrs[i].buf); + acb->qcrs[i].buf = qemu_blockalign(s->bs[i], acb->qiov->size); + qemu_iovec_init(&acb->qcrs[i].qiov, acb->qiov->niov); + qemu_iovec_clone(&acb->qcrs[i].qiov, acb->qiov, acb->qcrs[i].buf); } for (i = 0; i < s->num_children; i++) { - bdrv_aio_readv(s->bs[i], sector_num, &acb->qcrs[i].qiov, nb_sectors, - quorum_aio_cb, &acb->qcrs[i]); + bdrv_aio_readv(s->bs[i], acb->sector_num, &acb->qcrs[i].qiov, + acb->nb_sectors, quorum_aio_cb, &acb->qcrs[i]); } return &acb->common; } +static BlockDriverAIOCB *read_fifo_child(QuorumAIOCB *acb) +{ + BDRVQuorumState *s = acb->common.bs->opaque; + + acb->qcrs[acb->child_iter].buf = qemu_blockalign(s->bs[acb->child_iter], + acb->qiov->size); + qemu_iovec_init(&acb->qcrs[acb->child_iter].qiov, acb->qiov->niov); + qemu_iovec_clone(&acb->qcrs[acb->child_iter].qiov, acb->qiov, + acb->qcrs[acb->child_iter].buf); + bdrv_aio_readv(s->bs[acb->child_iter], acb->sector_num, + &acb->qcrs[acb->child_iter].qiov, acb->nb_sectors, + quorum_aio_cb, &acb->qcrs[acb->child_iter]); + + return &acb->common; +} + +static BlockDriverAIOCB *quorum_aio_readv(BlockDriverState *bs, + int64_t sector_num, + QEMUIOVector *qiov, + int nb_sectors, + BlockDriverCompletionFunc *cb, + void *opaque) +{ + BDRVQuorumState *s = bs->opaque; + QuorumAIOCB *acb = quorum_aio_get(s, bs, qiov, sector_num, + nb_sectors, cb, opaque); + acb->is_read = true; + + if (s->read_pattern == QUORUM_READ_PATTERN_QUORUM) { + acb->child_iter = s->num_children - 1; + return read_quorum_children(acb); + } + + acb->child_iter = 0; + return read_fifo_child(acb); +} + static BlockDriverAIOCB *quorum_aio_writev(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov, @@ -787,10 +836,33 @@ static QemuOptsList quorum_runtime_opts = { .type = QEMU_OPT_BOOL, .help = "Rewrite corrupted block on read quorum", }, + { + .name = QUORUM_OPT_READ_PATTERN, + .type = QEMU_OPT_STRING, + .help = "Allowed pattern: quorum, fifo. Quorum is default", + }, { /* end of list */ } }, }; +static int parse_read_pattern(const char *opt) +{ + int i; + + if (!opt) { + /* Set quorum as default */ + return QUORUM_READ_PATTERN_QUORUM; + } + + for (i = 0; i < QUORUM_READ_PATTERN_MAX; i++) { + if (!strcmp(opt, QuorumReadPattern_lookup[i])) { + return i; + } + } + + return -EINVAL; +} + static int quorum_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { @@ -832,28 +904,37 @@ static int quorum_open(BlockDriverState *bs, QDict *options, int flags, } s->threshold = qemu_opt_get_number(opts, QUORUM_OPT_VOTE_THRESHOLD, 0); - - /* and validate it against s->num_children */ - ret = quorum_valid_threshold(s->threshold, s->num_children, &local_err); + ret = parse_read_pattern(qemu_opt_get(opts, QUORUM_OPT_READ_PATTERN)); if (ret < 0) { + error_setg(&local_err, "Please set read-pattern as fifo or quorum"); goto exit; } + s->read_pattern = ret; - /* is the driver in blkverify mode */ - if (qemu_opt_get_bool(opts, QUORUM_OPT_BLKVERIFY, false) && - s->num_children == 2 && s->threshold == 2) { - s->is_blkverify = true; - } else if (qemu_opt_get_bool(opts, QUORUM_OPT_BLKVERIFY, false)) { - fprintf(stderr, "blkverify mode is set by setting blkverify=on " - "and using two files with vote_threshold=2\n"); - } + if (s->read_pattern == QUORUM_READ_PATTERN_QUORUM) { + /* and validate it against s->num_children */ + ret = quorum_valid_threshold(s->threshold, s->num_children, &local_err); + if (ret < 0) { + goto exit; + } - s->rewrite_corrupted = qemu_opt_get_bool(opts, QUORUM_OPT_REWRITE, false); - if (s->rewrite_corrupted && s->is_blkverify) { - error_setg(&local_err, - "rewrite-corrupted=on cannot be used with blkverify=on"); - ret = -EINVAL; - goto exit; + /* is the driver in blkverify mode */ + if (qemu_opt_get_bool(opts, QUORUM_OPT_BLKVERIFY, false) && + s->num_children == 2 && s->threshold == 2) { + s->is_blkverify = true; + } else if (qemu_opt_get_bool(opts, QUORUM_OPT_BLKVERIFY, false)) { + fprintf(stderr, "blkverify mode is set by setting blkverify=on " + "and using two files with vote_threshold=2\n"); + } + + s->rewrite_corrupted = qemu_opt_get_bool(opts, QUORUM_OPT_REWRITE, + false); + if (s->rewrite_corrupted && s->is_blkverify) { + error_setg(&local_err, + "rewrite-corrupted=on cannot be used with blkverify=on"); + ret = -EINVAL; + goto exit; + } } /* allocate the children BlockDriverState array */ |