Re: [PATCH] block: recalculate segment count for multi-segment discard requests correctly

Hi Jens, when you get a moment, could you take a quick look at this one for ack?

On Thu, Feb 4, 2021 at 11:49 AM Laurence Oberman [off-list ref] wrote:

On Thu, 2021-02-04 at 10:27 +0800, Ming Lei wrote:

quoted

On Mon, Feb 01, 2021 at 11:48:50AM -0500, David Jeffery wrote:

quoted

When a stacked block device inserts a request into another block
device
using blk_insert_cloned_request, the request's nr_phys_segments
field gets
recalculated by a call to blk_recalc_rq_segments in
blk_cloned_rq_check_limits. But blk_recalc_rq_segments does not
know how to
handle multi-segment discards. For disk types which can handle
multi-segment discards like nvme, this results in discard requests
which
claim a single segment when it should report several, triggering a
warning
in nvme and causing nvme to fail the discard from the invalid
state.

 WARNING: CPU: 5 PID: 191 at drivers/nvme/host/core.c:700
nvme_setup_discard+0x170/0x1e0 [nvme_core]
 ...
 nvme_setup_cmd+0x217/0x270 [nvme_core]
 nvme_loop_queue_rq+0x51/0x1b0 [nvme_loop]
 __blk_mq_try_issue_directly+0xe7/0x1b0
 blk_mq_request_issue_directly+0x41/0x70
 ? blk_account_io_start+0x40/0x50
 dm_mq_queue_rq+0x200/0x3e0
 blk_mq_dispatch_rq_list+0x10a/0x7d0
 ? __sbitmap_queue_get+0x25/0x90
 ? elv_rb_del+0x1f/0x30
 ? deadline_remove_request+0x55/0xb0
 ? dd_dispatch_request+0x181/0x210
 __blk_mq_do_dispatch_sched+0x144/0x290
 ? bio_attempt_discard_merge+0x134/0x1f0
 __blk_mq_sched_dispatch_requests+0x129/0x180
 blk_mq_sched_dispatch_requests+0x30/0x60
 __blk_mq_run_hw_queue+0x47/0xe0
 __blk_mq_delay_run_hw_queue+0x15b/0x170
 blk_mq_sched_insert_requests+0x68/0xe0
 blk_mq_flush_plug_list+0xf0/0x170
 blk_finish_plug+0x36/0x50
 xlog_cil_committed+0x19f/0x290 [xfs]
 xlog_cil_process_committed+0x57/0x80 [xfs]
 xlog_state_do_callback+0x1e0/0x2a0 [xfs]
 xlog_ioend_work+0x2f/0x80 [xfs]
 process_one_work+0x1b6/0x350
 worker_thread+0x53/0x3e0
 ? process_one_work+0x350/0x350
 kthread+0x11b/0x140
 ? __kthread_bind_mask+0x60/0x60
 ret_from_fork+0x22/0x30

This patch fixes blk_recalc_rq_segments to be aware of devices
which can
have multi-segment discards. It calculates the correct discard
segment
count by counting the number of bio as each discard bio is
considered its
own segment.

Signed-off-by: David Jeffery <redacted>
Tested-by: Laurence Oberman <redacted>
---
 block/blk-merge.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/block/blk-merge.c b/block/blk-merge.c
index 808768f6b174..fe7358bd5d09 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c

@@ -382,6 +382,13 @@ unsigned int blk_recalc_rq_segments(struct

request *rq)

    switch (bio_op(rq->bio)) {
    case REQ_OP_DISCARD:
+           if (queue_max_discard_segments(rq->q) > 1) {
+                   struct bio *bio = rq->bio;
+                   for_each_bio(bio)
+                           nr_phys_segs++;
+                   return nr_phys_segs;
+           }
+           /* fall through */
    case REQ_OP_SECURE_ERASE:
    case REQ_OP_WRITE_ZEROES:
            return 0;

blk_rq_nr_discard_segments() always returns >=1 segments, so no
similar
issue in case of single range discard.

Reviewed-by: Ming Lei <redacted>

And it can be thought as:

Fixes: 1e739730c5b9 ("block: optionally merge discontiguous discard
bios into a single request")

Great, can we get enough acks and push this through its urgent for me
Reviewed-by: Laurence Oberman <redacted>