From 4f02fb7617ba12ac15d261c654b9759ea8f1f1ef Mon Sep 17 00:00:00 2001 From: Joseph Qi Date: Sat, 30 Sep 2017 14:38:49 +0800 Subject: blk-throttle: fix possible io stall when upgrade to max There is a case which will lead to io stall. The case is described as follows. /test1 |-subtest1 /test2 |-subtest2 And subtest1 and subtest2 each has 32 queued bios already. Now upgrade to max. In throtl_upgrade_state, it will try to dispatch bios as follows: 1) tg=subtest1, do nothing; 2) tg=test1, transfer 32 queued bios from subtest1 to test1; no pending left, no need to schedule next dispatch; 3) tg=subtest2, do nothing; 4) tg=test2, transfer 32 queued bios from subtest2 to test2; no pending left, no need to schedule next dispatch; 5) tg=/, transfer 8 queued bios from test1 to /, 8 queued bios from test2 to /, 8 queued bios from test1 to /, and 8 queued bios from test2 to /; note that test1 and test2 each still has 16 queued bios left; 6) tg=/, try to schedule next dispatch, but since disptime is now (update in tg_update_disptime, wait=0), pending timer is not scheduled in fact; 7) In throtl_upgrade_state it totally dispatches 32 queued bios and with 32 left. test1 and test2 each has 16 queued bios; 8) throtl_pending_timer_fn sees the left over bios, but could do nothing, because throtl_select_dispatch returns 0, and test1/test2 has no pending tg. The blktrace shows the following: 8,32 0 0 2.539007641 0 m N throtl upgrade to max 8,32 0 0 2.539072267 0 m N throtl /test2 dispatch nr_queued=16 read=0 write=16 8,32 7 0 2.539077142 0 m N throtl /test1 dispatch nr_queued=16 read=0 write=16 So force schedule dispatch if there are pending children. Reviewed-by: Shaohua Li Signed-off-by: Joseph Qi Signed-off-by: Jens Axboe --- block/blk-throttle.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'block') diff --git a/block/blk-throttle.c b/block/blk-throttle.c index 0fea76aa0f3f..17816a028dcb 100644 --- a/block/blk-throttle.c +++ b/block/blk-throttle.c @@ -1911,11 +1911,11 @@ static void throtl_upgrade_state(struct throtl_data *td) tg->disptime = jiffies - 1; throtl_select_dispatch(sq); - throtl_schedule_next_dispatch(sq, false); + throtl_schedule_next_dispatch(sq, true); } rcu_read_unlock(); throtl_select_dispatch(&td->service_queue); - throtl_schedule_next_dispatch(&td->service_queue, false); + throtl_schedule_next_dispatch(&td->service_queue, true); queue_work(kthrotld_workqueue, &td->dispatch_work); } -- cgit v1.2.3 From 70e62f4bacdf31ea8a59f241c9229120cd06d9d1 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Tue, 3 Oct 2017 14:57:16 -0700 Subject: blk-mq-debugfs: fix device sched directory for default scheduler In blk_mq_debugfs_register(), I remembered to set up the per-hctx sched directories if a default scheduler was already configured by blk_mq_sched_init() from blk_mq_init_allocated_queue(), but I didn't do the same for the device-wide sched directory. Fix it. Fixes: d332ce091813 ("blk-mq-debugfs: allow schedulers to register debugfs attributes") Signed-off-by: Omar Sandoval Signed-off-by: Jens Axboe --- block/blk-mq-debugfs.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'block') diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c index 980e73095643..de294d775acf 100644 --- a/block/blk-mq-debugfs.c +++ b/block/blk-mq-debugfs.c @@ -815,10 +815,14 @@ int blk_mq_debugfs_register(struct request_queue *q) goto err; /* - * blk_mq_init_hctx() attempted to do this already, but q->debugfs_dir + * blk_mq_init_sched() attempted to do this already, but q->debugfs_dir * didn't exist yet (because we don't know what to name the directory * until the queue is registered to a gendisk). */ + if (q->elevator && !q->sched_debugfs_dir) + blk_mq_debugfs_register_sched(q); + + /* Similarly, blk_mq_init_hctx() couldn't do this previously. */ queue_for_each_hw_ctx(q, hctx, i) { if (!hctx->debugfs_dir && blk_mq_debugfs_register_hctx(q, hctx)) goto err; -- cgit v1.2.3 From eab40cf336065e8d765e006b81ff48c5c114b365 Mon Sep 17 00:00:00 2001 From: Benjamin Block Date: Tue, 3 Oct 2017 12:48:37 +0200 Subject: bsg-lib: fix use-after-free under memory-pressure When under memory-pressure it is possible that the mempool which backs the 'struct request_queue' will make use of up to BLKDEV_MIN_RQ count emergency buffers - in case it can't get a regular allocation. These buffers are preallocated and once they are also used, they are re-supplied with old finished requests from the same request_queue (see mempool_free()). The bug is, when re-supplying the emergency pool, the old requests are not again ran through the callback mempool_t->alloc(), and thus also not through the callback bsg_init_rq(). Thus we skip initialization, and while the sense-buffer still should be good, scsi_request->cmd might have become to be an invalid pointer in the meantime. When the request is initialized in bsg.c, and the user's CDB is larger than BLK_MAX_CDB, bsg will replace it with a custom allocated buffer, which is freed when the user's command is finished, thus it dangles afterwards. When next a command is sent by the user that has a smaller/similar CDB as BLK_MAX_CDB, bsg will assume that scsi_request->cmd is backed by scsi_request->__cmd, will not make a custom allocation, and write into undefined memory. Fix this by splitting bsg_init_rq() into two functions: - bsg_init_rq() is changed to only do the allocation of the sense-buffer, which is used to back the bsg job's reply buffer. This pointer should never change during the lifetime of a scsi_request, so it doesn't need re-initialization. - bsg_initialize_rq() is a new function that makes use of 'struct request_queue's initialize_rq_fn callback (which was introduced in v4.12). This is always called before the request is given out via blk_get_request(). This function does the remaining initialization that was previously done in bsg_init_rq(), and will also do it when the request is taken from the emergency-pool of the backing mempool. Fixes: 50b4d485528d ("bsg-lib: fix kernel panic resulting from missing allocation of reply-buffer") Cc: # 4.11+ Reviewed-by: Hannes Reinecke Reviewed-by: Johannes Thumshirn Reviewed-by: Christoph Hellwig Signed-off-by: Benjamin Block Signed-off-by: Jens Axboe --- block/bsg-lib.c | 27 +++++++++++++++++++++------ 1 file changed, 21 insertions(+), 6 deletions(-) (limited to 'block') diff --git a/block/bsg-lib.c b/block/bsg-lib.c index dbddff8174e5..15d25ccd51a5 100644 --- a/block/bsg-lib.c +++ b/block/bsg-lib.c @@ -207,20 +207,34 @@ static int bsg_init_rq(struct request_queue *q, struct request *req, gfp_t gfp) struct bsg_job *job = blk_mq_rq_to_pdu(req); struct scsi_request *sreq = &job->sreq; + /* called right after the request is allocated for the request_queue */ + + sreq->sense = kzalloc(SCSI_SENSE_BUFFERSIZE, gfp); + if (!sreq->sense) + return -ENOMEM; + + return 0; +} + +static void bsg_initialize_rq(struct request *req) +{ + struct bsg_job *job = blk_mq_rq_to_pdu(req); + struct scsi_request *sreq = &job->sreq; + void *sense = sreq->sense; + + /* called right before the request is given to the request_queue user */ + memset(job, 0, sizeof(*job)); scsi_req_init(sreq); + + sreq->sense = sense; sreq->sense_len = SCSI_SENSE_BUFFERSIZE; - sreq->sense = kzalloc(sreq->sense_len, gfp); - if (!sreq->sense) - return -ENOMEM; job->req = req; - job->reply = sreq->sense; + job->reply = sense; job->reply_len = sreq->sense_len; job->dd_data = job + 1; - - return 0; } static void bsg_exit_rq(struct request_queue *q, struct request *req) @@ -251,6 +265,7 @@ struct request_queue *bsg_setup_queue(struct device *dev, const char *name, q->cmd_size = sizeof(struct bsg_job) + dd_job_size; q->init_rq_fn = bsg_init_rq; q->exit_rq_fn = bsg_exit_rq; + q->initialize_rq_fn = bsg_initialize_rq; q->request_fn = bsg_request_fn; ret = blk_init_allocated_queue(q); -- cgit v1.2.3 From 95d78c28b5a85bacbc29b8dba7c04babb9b0d467 Mon Sep 17 00:00:00 2001 From: Vitaly Mayatskikh Date: Fri, 22 Sep 2017 01:18:39 -0400 Subject: fix unbalanced page refcounting in bio_map_user_iov bio_map_user_iov and bio_unmap_user do unbalanced pages refcounting if IO vector has small consecutive buffers belonging to the same page. bio_add_pc_page merges them into one, but the page reference is never dropped. Cc: stable@vger.kernel.org Signed-off-by: Vitaly Mayatskikh Signed-off-by: Al Viro --- block/bio.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'block') diff --git a/block/bio.c b/block/bio.c index b38e962fa83e..0d6439e89acb 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1383,6 +1383,7 @@ struct bio *bio_map_user_iov(struct request_queue *q, offset = offset_in_page(uaddr); for (j = cur_page; j < page_limit; j++) { unsigned int bytes = PAGE_SIZE - offset; + unsigned short prev_bi_vcnt = bio->bi_vcnt; if (len <= 0) break; @@ -1397,6 +1398,13 @@ struct bio *bio_map_user_iov(struct request_queue *q, bytes) break; + /* + * check if vector was merged with previous + * drop page reference if needed + */ + if (bio->bi_vcnt == prev_bi_vcnt) + put_page(pages[j]); + len -= bytes; offset = 0; } -- cgit v1.2.3 From 2b04e8f6bbb196cab4b232af0f8d48ff2c7a8058 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 23 Sep 2017 15:51:23 -0400 Subject: more bio_map_user_iov() leak fixes we need to take care of failure exit as well - pages already in bio should be dropped by analogue of bio_unmap_pages(), since their refcounts had been bumped only once per reference in bio. Cc: stable@vger.kernel.org Signed-off-by: Al Viro --- block/bio.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'block') diff --git a/block/bio.c b/block/bio.c index 0d6439e89acb..9e9606d26cc6 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1331,6 +1331,7 @@ struct bio *bio_map_user_iov(struct request_queue *q, int ret, offset; struct iov_iter i; struct iovec iov; + struct bio_vec *bvec; iov_for_each(iov, i, *iter) { unsigned long uaddr = (unsigned long) iov.iov_base; @@ -1375,7 +1376,12 @@ struct bio *bio_map_user_iov(struct request_queue *q, ret = get_user_pages_fast(uaddr, local_nr_pages, (iter->type & WRITE) != WRITE, &pages[cur_page]); - if (ret < local_nr_pages) { + if (unlikely(ret < local_nr_pages)) { + for (j = cur_page; j < page_limit; j++) { + if (!pages[j]) + break; + put_page(pages[j]); + } ret = -EFAULT; goto out_unmap; } @@ -1431,10 +1437,8 @@ struct bio *bio_map_user_iov(struct request_queue *q, return bio; out_unmap: - for (j = 0; j < nr_pages; j++) { - if (!pages[j]) - break; - put_page(pages[j]); + bio_for_each_segment_all(bvec, bio, j) { + put_page(bvec->bv_page); } out: kfree(pages); -- cgit v1.2.3 From 1cfd0ddd82232804e03f3023f6a58b50dfef0574 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 24 Sep 2017 10:21:15 -0400 Subject: bio_copy_user_iov(): don't ignore ->iov_offset Since "block: support large requests in blk_rq_map_user_iov" we started to call it with partially drained iter; that works fine on the write side, but reads create a copy of iter for completion time. And that needs to take the possibility of ->iov_iter != 0 into account... Cc: stable@vger.kernel.org #v4.5+ Signed-off-by: Al Viro --- block/bio.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'block') diff --git a/block/bio.c b/block/bio.c index 9e9606d26cc6..101c2a9b5481 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1239,8 +1239,8 @@ struct bio *bio_copy_user_iov(struct request_queue *q, */ bmd->is_our_pages = map_data ? 0 : 1; memcpy(bmd->iov, iter->iov, sizeof(struct iovec) * iter->nr_segs); - iov_iter_init(&bmd->iter, iter->type, bmd->iov, - iter->nr_segs, iter->count); + bmd->iter = *iter; + bmd->iter.iov = bmd->iov; ret = -ENOMEM; bio = bio_kmalloc(gfp_mask, nr_pages); -- cgit v1.2.3