linux-toradex.git/drivers/md, branch v5.11-rc5

Merge tag 'block-5.11-2021-01-24' of git://git.kernel.dk/linux-block

2021-01-24T20:24:35+00:00

Pull block fixes from Jens Axboe:

 - NVMe pull request from Christoph:
      - fix a status code in nvmet (Chaitanya Kulkarni)
      - avoid double completions in nvme-rdma/nvme-tcp (Chao Leng)
      - fix the CMB support to cope with NVMe 1.4 controllers (Klaus Jensen)
      - fix PRINFO handling in the passthrough ioctl (Revanth Rajashekar)
      - fix a double DMA unmap in nvme-pci

 - lightnvm error path leak fix (Pan)

 - MD pull request from Song:
      - Flush request fix (Xiao)

* tag 'block-5.11-2021-01-24' of git://git.kernel.dk/linux-block:
  lightnvm: fix memory leak when submit fails
  nvme-pci: fix error unwind in nvme_map_data
  nvme-pci: refactor nvme_unmap_data
  md: Set prev_flush_start and flush_bio in an atomic way
  nvmet: set right status on error in id-ns handler
  nvme-pci: allow use of cmb on v1.4 controllers
  nvme-tcp: avoid request double completion for concurrent nvme_tcp_timeout
  nvme-rdma: avoid request double completion for concurrent nvme_rdma_timeout
  nvme: check the PRINFO bit before deciding the host buffer length

dm: avoid filesystem lookup in dm_get_dev_t()

2021-01-21T20:06:45+00:00

This reverts commit
644bda6f3460 ("dm table: fall back to getting device using name_to_dev_t()")

dm_get_dev_t() is just used to convert an arbitrary 'path' string
into a dev_t. It doesn't presume that the device is present; that
check will be done later, as the only caller is dm_get_device(),
which does a dm_get_table_device() later on, which will properly
open the device.

So if the path string already _is_ in major:minor representation
we can convert it directly, avoiding a recursion into the filesystem
to lookup the block device.

This avoids a hang in multipath_message() when the filesystem is
inaccessible.

Fixes: 644bda6f3460 ("dm table: fall back to getting device using name_to_dev_t()")
Cc: stable@vger.kernel.org
Signed-off-by: Hannes Reinecke 
Signed-off-by: Martin Wilck 
Reviewed-by: Christoph Hellwig 
Signed-off-by: Mike Snitzer

dm crypt: fix copy and paste bug in crypt_alloc_req_aead

2021-01-21T20:06:44+00:00

In commit d68b29584c25 ("dm crypt: use GFP_ATOMIC when allocating
crypto requests from softirq") code was incorrectly copy and pasted
from crypt_alloc_req_skcipher()'s crypto request allocation code to
crypt_alloc_req_aead(). It is OK from runtime perspective as both
simple encryption request pointer and AEAD request pointer are part of
a union, but may confuse code reviewers.

Fixes: d68b29584c25 ("dm crypt: use GFP_ATOMIC when allocating crypto requests from softirq")
Cc: stable@vger.kernel.org # v5.9+
Reported-by: Pavel Machek 
Signed-off-by: Ignat Korchagin 
Signed-off-by: Mike Snitzer

dm integrity: conditionally disable "recalculate" feature

2021-01-21T20:05:21+00:00

Otherwise a malicious user could (ab)use the "recalculate" feature
that makes dm-integrity calculate the checksums in the background
while the device is already usable. When the system restarts before all
checksums have been calculated, the calculation continues where it was
interrupted even if the recalculate feature is not requested the next
time the dm device is set up.

Disable recalculating if we use internal_hash or journal_hash with a
key (e.g. HMAC) and we don't have the "legacy_recalculate" flag.

This may break activation of a volume, created by an older kernel,
that is not yet fully recalculated -- if this happens, the user should
add the "legacy_recalculate" flag to constructor parameters.

Cc: stable@vger.kernel.org
Signed-off-by: Mikulas Patocka 
Reported-by: Daniel Glockner 
Signed-off-by: Mike Snitzer

dm integrity: fix a crash if "recalculate" used without "internal_hash"

2021-01-21T19:41:43+00:00

Recalculate can only be specified with internal_hash.

Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Mikulas Patocka 
Signed-off-by: Mike Snitzer

md: Set prev_flush_start and flush_bio in an atomic way

2021-01-20T16:18:10+00:00

One customer reports a crash problem which causes by flush request. It
triggers a warning before crash.

        /* new request after previous flush is completed */
        if (ktime_after(req_start, mddev->prev_flush_start)) {
                WARN_ON(mddev->flush_bio);
                mddev->flush_bio = bio;
                bio = NULL;
        }

The WARN_ON is triggered. We use spin lock to protect prev_flush_start and
flush_bio in md_flush_request. But there is no lock protection in
md_submit_flush_data. It can set flush_bio to NULL first because of
compiler reordering write instructions.

For example, flush bio1 sets flush bio to NULL first in
md_submit_flush_data. An interrupt or vmware causing an extended stall
happen between updating flush_bio and prev_flush_start. Because flush_bio
is NULL, flush bio2 can get the lock and submit to underlayer disks. Then
flush bio1 updates prev_flush_start after the interrupt or extended stall.

Then flush bio3 enters in md_flush_request. The start time req_start is
behind prev_flush_start. The flush_bio is not NULL(flush bio2 hasn't
finished). So it can trigger the WARN_ON now. Then it calls INIT_WORK
again. INIT_WORK() will re-initialize the list pointers in the
work_struct, which then can result in a corrupted work list and the
work_struct queued a second time. With the work list corrupted, it can
lead in invalid work items being used and cause a crash in
process_one_work.

We need to make sure only one flush bio can be handled at one same time.
So add spin lock in md_submit_flush_data to protect prev_flush_start and
flush_bio in an atomic way.

Reviewed-by: David Jeffery 
Signed-off-by: Xiao Ni 
Signed-off-by: Song Liu

Merge tag 'for-5.11/dm-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

2021-01-16T02:01:17+00:00

Pull device mapper fixes from Mike Snitzer:

 - Fix DM-raid's raid1 discard limits so discards work.

 - Select missing Kconfig dependencies for DM integrity and zoned
   targets.

 - Four fixes for DM crypt target's support to optionally bypass kcryptd
   workqueues.

 - Fix DM snapshot merge supports missing data flushes before committing
   metadata.

 - Fix DM integrity data device flushing when external metadata is used.

 - Fix DM integrity's maximum number of supported constructor arguments
   that user can request when creating an integrity device.

 - Eliminate DM core ioctl logging noise when an ioctl is issued without
   required CAP_SYS_RAWIO permission.

* tag 'for-5.11/dm-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm crypt: defer decryption to a tasklet if interrupts disabled
  dm integrity: fix the maximum number of arguments
  dm crypt: do not call bio_endio() from the dm-crypt tasklet
  dm integrity: fix flush with external metadata device
  dm: eliminate potential source of excessive kernel log noise
  dm snapshot: flush merged data before committing metadata
  dm crypt: use GFP_ATOMIC when allocating crypto requests from softirq
  dm crypt: do not wait for backlogged crypto request completion in softirq
  dm zoned: select CONFIG_CRC32
  dm integrity: select CRYPTO_SKCIPHER
  dm raid: fix discard limits for raid1

dm crypt: defer decryption to a tasklet if interrupts disabled

2021-01-14T14:54:37+00:00

On some specific hardware on early boot we occasionally get:

[ 1193.920255][    T0] BUG: sleeping function called from invalid context at mm/mempool.c:381
[ 1193.936616][    T0] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/69
[ 1193.953233][    T0] no locks held by swapper/69/0.
[ 1193.965871][    T0] irq event stamp: 575062
[ 1193.977724][    T0] hardirqs last  enabled at (575061): [] tick_nohz_idle_exit+0xe2/0x3e0
[ 1194.002762][    T0] hardirqs last disabled at (575062): [] flush_smp_call_function_from_idle+0x4f/0x80
[ 1194.029035][    T0] softirqs last  enabled at (575050): [] asm_call_irq_on_stack+0x12/0x20
[ 1194.054227][    T0] softirqs last disabled at (575043): [] asm_call_irq_on_stack+0x12/0x20
[ 1194.079389][    T0] CPU: 69 PID: 0 Comm: swapper/69 Not tainted 5.10.6-cloudflare-kasan-2021.1.4-dev #1
[ 1194.104103][    T0] Hardware name: NULL R162-Z12-CD/MZ12-HD4-CD, BIOS R10 06/04/2020
[ 1194.119591][    T0] Call Trace:
[ 1194.130233][    T0]  dump_stack+0x9a/0xcc
[ 1194.141617][    T0]  ___might_sleep.cold+0x180/0x1b0
[ 1194.153825][    T0]  mempool_alloc+0x16b/0x300
[ 1194.165313][    T0]  ? remove_element+0x160/0x160
[ 1194.176961][    T0]  ? blk_mq_end_request+0x4b/0x490
[ 1194.188778][    T0]  crypt_convert+0x27f6/0x45f0 [dm_crypt]
[ 1194.201024][    T0]  ? rcu_read_lock_sched_held+0x3f/0x70
[ 1194.212906][    T0]  ? module_assert_mutex_or_preempt+0x3e/0x70
[ 1194.225318][    T0]  ? __module_address.part.0+0x1b/0x3a0
[ 1194.237212][    T0]  ? is_kernel_percpu_address+0x5b/0x190
[ 1194.249238][    T0]  ? crypt_iv_tcw_ctr+0x4a0/0x4a0 [dm_crypt]
[ 1194.261593][    T0]  ? is_module_address+0x25/0x40
[ 1194.272905][    T0]  ? static_obj+0x8a/0xc0
[ 1194.283582][    T0]  ? lockdep_init_map_waits+0x26a/0x700
[ 1194.295570][    T0]  ? __raw_spin_lock_init+0x39/0x110
[ 1194.307330][    T0]  kcryptd_crypt_read_convert+0x31c/0x560 [dm_crypt]
[ 1194.320496][    T0]  ? kcryptd_queue_crypt+0x1be/0x380 [dm_crypt]
[ 1194.333203][    T0]  blk_update_request+0x6d7/0x1500
[ 1194.344841][    T0]  ? blk_mq_trigger_softirq+0x190/0x190
[ 1194.356831][    T0]  blk_mq_end_request+0x4b/0x490
[ 1194.367994][    T0]  ? blk_mq_trigger_softirq+0x190/0x190
[ 1194.379693][    T0]  flush_smp_call_function_queue+0x24b/0x560
[ 1194.391847][    T0]  flush_smp_call_function_from_idle+0x59/0x80
[ 1194.403969][    T0]  do_idle+0x287/0x450
[ 1194.413891][    T0]  ? arch_cpu_idle_exit+0x40/0x40
[ 1194.424716][    T0]  ? lockdep_hardirqs_on_prepare+0x286/0x3f0
[ 1194.436399][    T0]  ? _raw_spin_unlock_irqrestore+0x39/0x40
[ 1194.447759][    T0]  cpu_startup_entry+0x19/0x20
[ 1194.458038][    T0]  secondary_startup_64_no_verify+0xb0/0xbb

IO completion can be queued to a different CPU by the block subsystem as a "call
single function/data". The CPU may run these routines from the idle task, but it
does so with interrupts disabled.

It is not a good idea to do decryption with irqs disabled even in an idle task
context, so just defer it to a tasklet (as is done with requests from hard irqs).

Fixes: 39d42fa96ba1 ("dm crypt: add flags to optionally bypass kcryptd workqueues")
Cc: stable@vger.kernel.org # v5.9+
Signed-off-by: Ignat Korchagin 
Signed-off-by: Mike Snitzer

dm integrity: fix the maximum number of arguments

2021-01-12T20:55:06+00:00

Advance the maximum number of arguments from 9 to 15 to account for
all potential feature flags that may be supplied.

Linux 4.19 added "meta_device"
(356d9d52e1221ba0c9f10b8b38652f78a5298329) and "recalculate"
(a3fcf7253139609bf9ff901fbf955fba047e75dd) flags.

Commit 468dfca38b1a6fbdccd195d875599cb7c8875cd9 added
"sectors_per_bit" and "bitmap_flush_interval".

Commit 84597a44a9d86ac949900441cea7da0af0f2f473 added
"allow_discards".

And the commit d537858ac8aaf4311b51240893add2fc62003b97 added
"fix_padding".

Signed-off-by: Mikulas Patocka 
Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Mike Snitzer

dm crypt: do not call bio_endio() from the dm-crypt tasklet

2021-01-12T18:51:17+00:00

Sometimes, when dm-crypt executes decryption in a tasklet, we may get
"BUG: KASAN: use-after-free in tasklet_action_common.constprop..."
with a kasan-enabled kernel.

When the decryption fully completes in the tasklet, dm-crypt will call
bio_endio(), which in turn will call clone_endio() from dm.c core code. That
function frees the resources associated with the bio, including per bio private
structures. For dm-crypt it will free the current struct dm_crypt_io, which
contains our tasklet object, causing use-after-free, when the tasklet is being
dequeued by the kernel.

To avoid this, do not call bio_endio() from the current tasklet context, but
delay its execution to the dm-crypt IO workqueue.

Fixes: 39d42fa96ba1 ("dm crypt: add flags to optionally bypass kcryptd workqueues")
Cc:  # v5.9+
Signed-off-by: Ignat Korchagin 
Signed-off-by: Mike Snitzer