Skip to content

Various block related fixes

Jens Axboe linux-block linux-nvme rust-for-linux stable 1 version Applied

Status from this message: Applied, thanks!

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest98 patches 104 repliesApplied

Patches in v1 Applied

#Subject and filesSizeRepliesApplied
1/98blk-cgroup: keep the old blkg while reassociating a bioblock/blk-cgroup.c+5 -21 reply
2/98blk-iocost: free struct ioc after a grace periodblock/blk-iocost.c+2 -11 reply
3/98blk-throttle: don't truncate the cgroup v1 iops limitsblock/blk-throttle.c+2 -11 reply
4/98block: fix use-after-free of the disk in sync_bdevs()block/bdev.c+10 -01 reply
5/98block: don't lose the status of a failed flush requestblock/blk-flush.c+7 -11 reply
6/98blk-mq: don't batch allocate tags from an inactive hctxblock/blk-mq-tag.c+12 -01 reply
7/98block: keep zone write plugs with plugged BIOs aliveblock/blk-zoned.c+8 -11 reply
8/98block: keep the zone write plug of a failed writeblock/blk-zoned.c+6 -21 reply
9/98blk-mq: end the request if getting a keyslot failsblock/blk-mq.c+3 -23 replies
10/98block: don't unplug a zone under a failed writeblock/blk-zoned.c+12 -10 replies
11/98block: check dma alignment of bvec iters in async dioblock/bio.c block/blk.h block/fops.c+41 -352 replies
12/98drbd: stop the start_resync timer of a dying devicedrivers/block/drbd/drbd_main.c+1 -00 replies
13/98blk-cgroup: fix root io.stat byte counts on 32-bitblock/blk-cgroup.c+3 -31 reply
14/98drbd: take req_lock to skip the initial sync as peerdrivers/block/drbd/drbd_receiver.c+2 -00 replies
15/98zloop: refuse zone files without read/write_iterdrivers/block/zloop.c+10 -03 replies
16/98block/rnbd-srv: don't free the device on a sysfs errordrivers/block/rnbd/rnbd-srv-sysfs.c drivers/block/rnbd/rnbd-srv.c+13 -60 replies
17/98block/rnbd-srv: drop the sess_dev kobject unlockeddrivers/block/rnbd/rnbd-srv-sysfs.c drivers/block/rnbd/rnbd-srv.c+13 -70 replies
18/98aoe: don't free a device with its work queueddrivers/block/aoe/aoecmd.c drivers/block/aoe/aoedev.c+5 -15 replies
19/98blk-cgroup: don't free a blkg with its punt work queuedblock/blk-cgroup.c block/blk-cgroup.h+14 -51 reply
20/98mtip32xx: don't mark the disk dead in the irq handlerdrivers/block/mtip32xx/mtip32xx.c drivers/block/mtip32xx/mtip32xx.h+32 -63 replies
21/98mtip32xx: abort timed out commands of a removed devicedrivers/block/mtip32xx/mtip32xx.c+1 -10 replies
22/98mtip32xx: abort commands in flight on surprise removaldrivers/block/mtip32xx/mtip32xx.c+23 -52 replies
23/98blk-iolatency: flush enable_work after going offlineblock/blk-iolatency.c+2 -11 reply
24/98zram: accept lz4 dictionaries of more than 64 KiBdrivers/block/zram/backend_lz4.c drivers/block/zram/backend_lz4hc.c+6 -24 replies
25/98blk-throttle: pass the I/O counts of a dead group onblock/blk-throttle.c+12 -11 reply
26/98blk-throttle: make blkio.reset_stats reset I/O countsblock/blk-throttle.c+9 -01 reply
27/98blk-mq: don't trust the cookie in blk_mq_poll()block/blk-mq.c+14 -11 reply
28/98blk-iocost: check again that the disk is aliveblock/blk-iocost.c+5 -02 replies
29/98blk-mq: keep sharing tags fairly after a queue updateblock/blk-mq.c+3 -01 reply
30/98blk-mq: don't bring back flags of a completed requestblock/blk-mq.c block/blk-timeout.c block/blk.h+17 -21 reply
31/98blk-mq: don't forget a timer the driver has restartedblock/blk-mq.c+4 -11 reply
32/98blk-mq-dma: initialize the dma state for P2P, tooblock/blk-mq-dma.c+2 -01 reply
33/98blk-mq-dma: map the metadata a passthrough request hasblock/blk-mq-dma.c+8 -01 reply
34/98block: don't enter reclaim under rq_qos_mutexblock/blk-cgroup.c block/blk-iocost.c block/blk-iolatency.c +1 more+16 -81 reply
35/98blktrace: remove the debugfs files before the channelkernel/trace/blktrace.c+8 -90 replies
36/98blk-mq: quiesce a queue that joins a quiesced tag setblock/blk-mq.c include/linux/blk-mq.h+28 -03 replies
37/98blk-mq: exit the hctxs of a queue under tag_list_lockblock/blk-mq.c+8 -22 replies
38/98blk-mq: set up a queue under tag_list_lockblock/blk-mq.c+18 -31 reply
39/98rnbd-srv: fix use-after-free on two CLOSE messagesdrivers/block/rnbd/rnbd-srv.c+8 -71 reply
40/98rnbd-srv: publish the sess_dev at the end of OPENdrivers/block/rnbd/rnbd-srv.c+8 -21 reply
41/98zram: check a bio's range against the device sizedrivers/block/zram/zram_drv.c+22 -02 replies
42/98block: partitions/mac: compute the sectors in 64 bitsblock/partitions/mac.c+2 -22 replies
43/98blk-mq: free the bitmaps of tags in the SRCU callbackblock/blk-mq-tag.c+4 -42 replies
44/98block: apply the segment size to the integrity bufferinclude/linux/blkdev.h+5 -41 reply
45/98nvmet: don't truncate the length of big discardsdrivers/nvme/target/io-cmd-bdev.c+2 -11 reply
46/98nvme-loop: delete the controller if a reset failsdrivers/nvme/target/loop.c+3 -20 replies
47/98nvme: refuse metadata sizes above 255 bytesdrivers/nvme/host/core.c+4 -01 reply
48/98nvmet-tcp: rewind the iterator after non-data recordsdrivers/nvme/target/tcp.c+7 -53 replies
49/98nvmet-tcp: ask for the TLS record type on every receivedrivers/nvme/target/tcp.c+5 -01 reply
50/98nvme-tcp: wait for an in-flight TLS handshake callbackdrivers/nvme/host/tcp.c+6 -11 reply
51/98nvmet-tcp: look again after a short read with TLSdrivers/nvme/target/tcp.c+43 -110 replies
52/98nvme-tcp: keep the socket while the queue is polleddrivers/nvme/host/tcp.c+22 -70 replies
53/98nvme-tcp: don't get stuck on a TLS control recorddrivers/nvme/host/tcp.c+86 -11 reply
54/98nvme-tcp: start error recovery after a TLS KeyUpdatedrivers/nvme/host/tcp.c+8 -01 reply
55/98nvme-tcp: don't act on PDUs for a command being sentdrivers/nvme/host/tcp.c+73 -140 replies
56/98nvme-tcp: don't queue the AER while it is being sentdrivers/nvme/host/tcp.c+11 -40 replies
57/98nvme-tcp: don't send SIGPIPE to whoever submits I/Odrivers/nvme/host/tcp.c+10 -50 replies
58/98nvme-fabrics: reconnect if the association is gonedrivers/nvme/host/fabrics.c+13 -20 replies
59/98nvmet-auth: serialize the commands of a controllerdrivers/nvme/target/core.c drivers/nvme/target/fabrics-cmd-auth.c drivers/nvme/target/nvmet.h+11 -01 reply
60/98nvmet-auth: keep the negotiated hash per queuedrivers/nvme/target/auth.c drivers/nvme/target/fabrics-cmd-auth.c drivers/nvme/target/nvmet.h+25 -231 reply
61/98nvmet-auth: authenticate the I/O queues againdrivers/nvme/target/auth.c drivers/nvme/target/fabrics-cmd.c+7 -110 replies
62/98nvmet: don't mix queues with and without TLSdrivers/nvme/target/core.c drivers/nvme/target/fabrics-cmd.c drivers/nvme/target/nvmet.h+11 -01 reply
63/98nvmet-auth: refuse a reply without a DH valuedrivers/nvme/target/auth.c drivers/nvme/target/fabrics-cmd-auth.c+4 -01 reply
64/98nvmet-auth: keep the old keys if the new ones faildrivers/nvme/target/auth.c+28 -411 reply
65/98nvmet-auth: take only AUTH_Negotiate for a negotiationdrivers/nvme/target/fabrics-cmd-auth.c+2 -11 reply
66/98nvmet-auth: free the old data on a new AUTH_Negotiatedrivers/nvme/target/fabrics-cmd-auth.c+1 -01 reply
67/98nvmet-auth: fix the check of the offered DH groupsdrivers/nvme/target/fabrics-cmd-auth.c+5 -221 reply
68/98nvmet: accept I/O queues on the controller's port onlydrivers/nvme/target/core.c+5 -01 reply
69/98nvmet: confine fatal errors of unauthenticated queuesdrivers/nvme/target/auth.c drivers/nvme/target/core.c drivers/nvme/target/fabrics-cmd-auth.c +3 more+63 -91 reply
70/98nvmet-tcp: don't wait for the host in the accept workdrivers/nvme/target/tcp.c+59 -80 replies
71/98nvmet: check the host NQN against the TLS PSK identitydrivers/nvme/target/fabrics-cmd.c+33 -00 replies
72/98nvmet: fix NULL dereference of PCI endpoint targetsdrivers/nvme/target/nvmet.h+2 -10 replies
73/98nvmet-tcp: stop listening when a port is removeddrivers/nvme/target/tcp.c+2 -01 reply
74/98nvme-tcp: keep the page frag cache with the controllerdrivers/nvme/host/tcp.c+13 -70 replies
75/98nvmet-tcp: hold the queue while the TLS timer is armeddrivers/nvme/target/tcp.c+10 -20 replies
76/98nvmet: fix NULL dereference in debugfs host_traddrdrivers/nvme/target/core.c+3 -02 replies
77/98nvmet-tcp: let the accept work come to an enddrivers/nvme/target/tcp.c+6 -21 reply
78/98nvmet-tcp: don't read a removed port at Connect timedrivers/nvme/target/tcp.c+4 -10 replies
79/98nvme-tcp: start error recovery when a send failsdrivers/nvme/host/tcp.c+2 -01 reply
80/98nvme-loop: don't BUG() if a scan runs beside a resetdrivers/nvme/target/loop.c+0 -21 reply
81/98nvmet: fix pc_ref use-after-free in preempt and abortdrivers/nvme/target/nvmet.h drivers/nvme/target/pr.c+69 -280 replies
82/98nvmet: move preempt and abort work off nvmet-wqdrivers/nvme/target/core.c drivers/nvme/target/nvmet.h drivers/nvme/target/pr.c+16 -20 replies
83/98nvmet-rdma: treat a failed connect as a disconnectdrivers/nvme/target/rdma.c+7 -80 replies
84/98nvmet-rdma: disconnect only established queuesdrivers/nvme/target/rdma.c+13 -20 replies
85/98nvmet-rdma: take a device off the list lastdrivers/nvme/target/rdma.c+7 -30 replies
86/98nvmet-rdma: queue the release work under the mutexdrivers/nvme/target/rdma.c+6 -60 replies
87/98nvmet-rdma: disconnect a queue on device removaldrivers/nvme/target/rdma.c+8 -90 replies
88/98nvmet-rdma: flush nvmet-wq on module exitdrivers/nvme/target/rdma.c+6 -00 replies
89/98nvmet-rdma: reject queues larger than max_cqedrivers/nvme/target/rdma.c+11 -20 replies
90/98nvmet-rdma: disable interrupts under rsp_wr_wait_lockdrivers/nvme/target/rdma.c+9 -62 replies
91/98nvmet-rdma: don't run the wait list recursivelydrivers/nvme/target/rdma.c+25 -30 replies
92/98nvmet-rdma: fix a lost run of the wait listdrivers/nvme/target/rdma.c+17 -10 replies
93/98nvmet-fc: only send read data for successful commandsdrivers/nvme/target/fc.c+4 -41 reply
94/98nvmet: wait only for executing I/O in preempt and abortdrivers/nvme/target/core.c drivers/nvme/target/nvmet.h drivers/nvme/target/pr.c+45 -30 replies
95/98nvmet-rdma: fix double puts after a failed RDMA WRITEdrivers/nvme/target/rdma.c+4 -10 replies
96/98nvmet: take sq->ref in nvmet_sq_fatal_error()drivers/nvme/target/core.c drivers/nvme/target/fabrics-cmd-auth.c drivers/nvme/target/nvmet.h+23 -21 reply
97/98nvmet-rdma: hold sq->ref to post a T10-PI Read's SENDdrivers/nvme/target/rdma.c+13 -02 replies
98/98nvmet: don't clear sq->ctrl for a Connect with sqsize 0drivers/nvme/target/fabrics-cmd.c+1 -21 reply

Reviews trailers per patch

37 of 98 patches have a Reviewed-by or Acked-by.

PatchTejun HeoJohannes ThumshirnHannes ReineckeChristoph HellwigNilay ShroffJack Wang
1/98A-b
2/98A-b
3/98A-b
4/98R-b
5/98R-b
6/98
7/98
8/98
9/98
10/98
11/98
12/98
13/98A-b
14/98
15/98R-b
16/98
17/98
18/98
19/98A-b
20/98
21/98
22/98
23/98A-b
24/98
25/98A-b
26/98A-b
27/98R-b
28/98
29/98R-b
30/98R-b
31/98
32/98R-b
33/98
34/98A-b
35/98
36/98R-b
37/98R-b
38/98
39/98A-b
40/98A-b
41/98
42/98
43/98R-b
44/98R-b
45/98R-b
46/98
47/98
48/98
49/98R-b
50/98
51/98
52/98
53/98
54/98
55/98
56/98
57/98
58/98
59/98
60/98
61/98
62/98
63/98R-b
64/98R-b
65/98R-b
66/98R-b
67/98R-b
68/98R-b
69/98
70/98
71/98
72/98
73/98R-b
74/98
75/98
76/98
77/98R-b
78/98
79/98R-b
80/98R-b
81/98
82/98
83/98
84/98
85/98
86/98
87/98
88/98
89/98
90/98
91/98
92/98
93/98R-b
94/98
95/98
96/98R-b
97/98
98/98R-b

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes. Scroll the table sideways for all 6 reviewers.

Read the v1 thread

Cover letter [PATCHSET 0/98] Various block related fixes

Hi,

Got a giant pile of stuff dumped on me, and went through and did some vetting and, particularly, commit message rewriting to get it to a somewhat comprehensible shape. If you're CC'ed on a patch, please review it. Feel free to queue it up, or reply back why the particular patch is garbage, if it is garbage.

Thanks!

Signature · 1 line
Jens Axboe