platform/kernel/linux-rpi.git
4 years agoMerge branch 'nvme-5.7-rc1' of git://git.infradead.org/nvme into for-5.7/drivers
Jens Axboe [Wed, 25 Mar 2020 22:20:51 +0000 (16:20 -0600)]
Merge branch 'nvme-5.7-rc1' of git://git.infradead.org/nvme into for-5.7/drivers

Pull 5.7 NVMe updates from Keith.

* 'nvme-5.7-rc1' of git://git.infradead.org/nvme: (42 commits)
  nvme: cleanup namespace identifier reporting in nvme_init_ns_head
  nvme: rename __nvme_find_ns_head to nvme_find_ns_head
  nvme: refactor nvme_identify_ns_descs error handling
  nvme-tcp: Add warning on state change failure at nvme_tcp_setup_ctrl
  nvme-rdma: Add warning on state change failure at nvme_rdma_setup_ctrl
  nvme: Fix controller creation races with teardown flow
  nvme: Make nvme_uninit_ctrl symmetric to nvme_init_ctrl
  nvme: Fix ctrl use-after-free during sysfs deletion
  nvme-pci: Re-order nvme_pci_free_ctrl
  nvme: Remove unused return code from nvme_delete_ctrl_sync
  nvme: Use nvme_state_terminal helper
  nvme: release ida resources
  nvme: Add compat_ioctl handler for NVME_IOCTL_SUBMIT_IO
  nvmet-tcp: optimize tcp stack TX when data digest is used
  nvme-fabrics: Use scnprintf() for avoiding potential buffer overflow
  nvme-multipath: do not reset on unknown status
  nvmet-rdma: allocate RW ctxs according to mdts
  nvmet-rdma: Implement get_mdts controller op
  nvmet: Add get_mdts op for controllers
  nvme-pci: properly print controller address
  ...

4 years agonvme: cleanup namespace identifier reporting in nvme_init_ns_head
Christoph Hellwig [Wed, 25 Mar 2020 13:19:37 +0000 (14:19 +0100)]
nvme: cleanup namespace identifier reporting in nvme_init_ns_head

Lift the common namespace identifier reporting between the shared
namespace and new nshead cases into common code.  This also means
one less lock is held while doing I/O.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: rename __nvme_find_ns_head to nvme_find_ns_head
Christoph Hellwig [Wed, 25 Mar 2020 13:19:36 +0000 (14:19 +0100)]
nvme: rename __nvme_find_ns_head to nvme_find_ns_head

There is no non __-prefixed version, so make the name a little more
readable.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: refactor nvme_identify_ns_descs error handling
Christoph Hellwig [Wed, 25 Mar 2020 13:19:35 +0000 (14:19 +0100)]
nvme: refactor nvme_identify_ns_descs error handling

Move the handling of an error into the function from the caller, and
only do it for an actual error on the admin command itself, not the
command parsing, as that should be enough to deal with devices claiming
a bogus version compliance.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-tcp: Add warning on state change failure at nvme_tcp_setup_ctrl
Israel Rukshin [Tue, 24 Mar 2020 15:29:45 +0000 (17:29 +0200)]
nvme-tcp: Add warning on state change failure at nvme_tcp_setup_ctrl

The transition to LIVE state should not fail in case of a new controller.
Moving to DELETING state before nvme_tcp_create_ctrl() allocates all the
resources may leads to NULL dereference at teardown flow (e.g., IO tagset,
admin_q, connect_q).

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-rdma: Add warning on state change failure at nvme_rdma_setup_ctrl
Israel Rukshin [Tue, 24 Mar 2020 15:29:44 +0000 (17:29 +0200)]
nvme-rdma: Add warning on state change failure at nvme_rdma_setup_ctrl

The transition to LIVE state should not fail in case of a new controller.
Moving to DELETING state before nvme_tcp_create_ctrl() allocates all the
resources may leads to NULL dereference at teardown flow (e.g., IO tagset,
admin_q, connect_q).

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Fix controller creation races with teardown flow
Israel Rukshin [Tue, 24 Mar 2020 15:29:43 +0000 (17:29 +0200)]
nvme: Fix controller creation races with teardown flow

Calling nvme_sysfs_delete() when the controller is in the middle of
creation may cause several bugs. If the controller is in NEW state we
remove delete_controller file and don't delete the controller. The user
will not be able to use nvme disconnect command on that controller again,
although the controller may be active. Other bugs may happen if the
controller is in the middle of create_ctrl callback and
nvme_do_delete_ctrl() starts. For example, freeing I/O tagset at
nvme_do_delete_ctrl() before it was allocated at create_ctrl callback.

To fix all those races don't allow the user to delete the controller
before it was fully created.

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Make nvme_uninit_ctrl symmetric to nvme_init_ctrl
Israel Rukshin [Tue, 24 Mar 2020 15:29:42 +0000 (17:29 +0200)]
nvme: Make nvme_uninit_ctrl symmetric to nvme_init_ctrl

Put the ctrl reference count at nvme_uninit_ctrl as opposed to
nvme_init_ctrl which takes it. This decrease the reference count at the
core layer instead of decreasing it on each transport separately.
Also move the call of nvme_uninit_ctrl at PCI driver after calling to
nvme_release_prp_pools and nvme_dev_unmap, in order to put the reference
count after using the dev. This is safe because those functions use
nvme_dev which is freed only later at nvme_pci_free_ctrl.

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Fix ctrl use-after-free during sysfs deletion
Israel Rukshin [Tue, 24 Mar 2020 15:29:41 +0000 (17:29 +0200)]
nvme: Fix ctrl use-after-free during sysfs deletion

In case nvme_sysfs_delete() is called by the user before taking the ctrl
reference count, the ctrl may be freed during the creation and cause the
bug. Take the reference as soon as the controller is externally visible,
which is done by cdev_device_add() in nvme_init_ctrl(). Also take the
reference count at the core layer instead of taking it on each transport
separately.

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-pci: Re-order nvme_pci_free_ctrl
Israel Rukshin [Tue, 24 Mar 2020 15:29:40 +0000 (17:29 +0200)]
nvme-pci: Re-order nvme_pci_free_ctrl

Destroy the resources in the same order like in nvme_probe error flow to
improve code readability.

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Remove unused return code from nvme_delete_ctrl_sync
Israel Rukshin [Tue, 24 Mar 2020 15:29:39 +0000 (17:29 +0200)]
nvme: Remove unused return code from nvme_delete_ctrl_sync

The return code of nvme_delete_ctrl_sync is never used, so change it to
void.

Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Use nvme_state_terminal helper
Israel Rukshin [Tue, 10 Mar 2020 14:39:10 +0000 (16:39 +0200)]
nvme: Use nvme_state_terminal helper

Improve code readability.

Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: release ida resources
Max Gurtovoy [Wed, 18 Mar 2020 15:27:59 +0000 (17:27 +0200)]
nvme: release ida resources

ida instances allocate some internal memory in addition to the base
'struct ida'. Use ida_destroy() to release that memory at module_exit().

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Add compat_ioctl handler for NVME_IOCTL_SUBMIT_IO
masahiro31.yamada@kioxia.com [Thu, 5 Mar 2020 11:13:29 +0000 (11:13 +0000)]
nvme: Add compat_ioctl handler for NVME_IOCTL_SUBMIT_IO

Currently 32 bit application gets ENOTTY when it calls
compat_ioctl with NVME_IOCTL_SUBMIT_IO in 64 bit kernel.

The cause is that the results of sizeof(struct nvme_user_io),
which is used to define NVME_IOCTL_SUBMIT_IO,
are not same between 32 bit compiler and 64 bit compiler.

* 32 bit: the result of sizeof nvme_user_io is 44.
* 64 bit: the result of sizeof nvme_user_io is 48.

64 bit compiler seems to add 32 bit padding for multiple of 8 bytes.

This patch adds a compat_ioctl handler.
The handler replaces NVME_IOCTL_SUBMIT_IO32 with NVME_IOCTL_SUBMIT_IO
in case 32 bit application calls compat_ioctl for submit in 64 bit kernel.
Then, it calls nvme_ioctl as usual.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Masahiro Yamada (KIOXIA) <masahiro31.yamada@kioxia.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvmet-tcp: optimize tcp stack TX when data digest is used
Sagi Grimberg [Thu, 12 Mar 2020 23:06:39 +0000 (16:06 -0700)]
nvmet-tcp: optimize tcp stack TX when data digest is used

If we have a 4-byte data digest to send to the wire, but we
have more data to send, set MSG_MORE to tell the stack
that more is coming.

Reviewed-by: Mark Wunderlich <mark.wunderlich@intel.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-fabrics: Use scnprintf() for avoiding potential buffer overflow
Takashi Iwai [Wed, 11 Mar 2020 08:50:37 +0000 (09:50 +0100)]
nvme-fabrics: Use scnprintf() for avoiding potential buffer overflow

Since snprintf() returns the would-be-output size instead of the
actual output size, the succeeding calls may go beyond the given
buffer limit.  Fix it by replacing with scnprintf().

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-multipath: do not reset on unknown status
John Meneghini [Thu, 20 Feb 2020 01:05:38 +0000 (10:05 +0900)]
nvme-multipath: do not reset on unknown status

The nvme multipath error handling defaults to controller reset if the
error is unknown. There are, however, no existing nvme status codes that
indicate a reset should be used, and resetting causes unnecessary
disruption to the rest of IO.

Change nvme's error handling to first check if failover should happen.
If not, let the normal error handling take over rather than reset the
controller.

Based-on-a-patch-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: John Meneghini <johnm@netapp.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvmet-rdma: allocate RW ctxs according to mdts
Max Gurtovoy [Sun, 8 Mar 2020 10:55:05 +0000 (12:55 +0200)]
nvmet-rdma: allocate RW ctxs according to mdts

Current nvmet-rdma code allocates MR pool budget based on queue size,
assuming both host and target use the same "max_pages_per_mr" count.
After limiting the mdts value for RDMA controllers, we know the factor
of maximum MR's per IO operation. Thus, make sure MR pool will be
sufficient for the required IO depth and IO size.

That is, say host's SQ size is 100, then the MR pool budget allocated
currently at target will also be 100 MRs. But 100 IO WRITE Requests
with 256 sg_count(IO size above 1MB) require 200 MRs when target's
"max_pages_per_mr" is 128.

Reported-by: Krishnamraju Eraparaju <krishna2@chelsio.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
4 years agonvmet-rdma: Implement get_mdts controller op
Max Gurtovoy [Sun, 8 Mar 2020 10:55:04 +0000 (12:55 +0200)]
nvmet-rdma: Implement get_mdts controller op

Set the maximal data transfer size to be 1MB (currently mdts is
unlimited). This will allow calculating the amount of MR's that
one ctrl should allocate to fulfill it's capabilities.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
4 years agonvmet: Add get_mdts op for controllers
Max Gurtovoy [Sun, 8 Mar 2020 10:55:03 +0000 (12:55 +0200)]
nvmet: Add get_mdts op for controllers

Some transports, such as RDMA, would like to set the Maximum Data
Transfer Size (MDTS) according to device/port/ctrl characteristics.
This will enable the transport to set the optimal MDTS according to
controller needs and device capabilities. Add a new nvmet transport
op that is called during ctrl identification. This will not effect
transports that don't implement this option. The return value of the new
op is according to the NVMe spec definition for MDTS.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Israel Rukshin <israelr@mellanox.com>
4 years agonvme-pci: properly print controller address
Max Gurtovoy [Mon, 9 Mar 2020 15:04:12 +0000 (17:04 +0200)]
nvme-pci: properly print controller address

Align PCI address print with fabrics address that is printed with
newline character.

Before:
[root@server40 linux]# cat /sys/class/nvme/nvme2/address
0000:0b:00.0[root@server40 linux]#

After:
[root@server40 linux]# cat /sys/class/nvme/nvme2/address
0000:0b:00.0
[root@server40 linux]#

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
4 years agonvme-tcp: break from io_work loop if recv failed
Sagi Grimberg [Wed, 26 Feb 2020 00:43:24 +0000 (16:43 -0800)]
nvme-tcp: break from io_work loop if recv failed

If we failed to receive data from the socket, don't try
to further process it, we will for sure be handling a queue
error at this point. While no issue was seen with the
current behavior thus far, its safer to cease socket processing
if we detected an error.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-tcp: move send failure to nvme_tcp_try_send
Sagi Grimberg [Wed, 26 Feb 2020 00:43:23 +0000 (16:43 -0800)]
nvme-tcp: move send failure to nvme_tcp_try_send

Consolidate the request failure handling code to where
it is being fetched (nvme_tcp_try_send).

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvmet-tcp: fix maxh2cdata icresp parameter
Sagi Grimberg [Wed, 26 Feb 2020 00:42:27 +0000 (16:42 -0800)]
nvmet-tcp: fix maxh2cdata icresp parameter

MAXH2CDATA is not zero based. Also no reason to limit ourselves to
1M transfers as we can do more easily. Make this an arbitrary limit
of 16M.

Reported-by: Wenhua Liu <liuw@vmware.com>
Cc: stable@vger.kernel.org # v5.0+
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-tcp: optimize queue io_cpu assignment for multiple queue maps
Sagi Grimberg [Tue, 25 Feb 2020 23:53:09 +0000 (15:53 -0800)]
nvme-tcp: optimize queue io_cpu assignment for multiple queue maps

Currently, queue io_cpu assignment is done sequentially for default,
read and poll queues based on queue id. This causes miss-alignment between
context of CPU initiating I/O and the I/O worker thread processing
queued requests or completions.

Change to modify queue io_cpu assignment to take into account queue
maps offset. Each queue io_cpu will start at zero for each queue map.
This essentially aligns read/poll queues to start over the same range as
default queues.

Testing performed by Mark with:
- ram device (nvmet)
- single CPU core (pinned)
- 100% 4k reads
- engine io_uring (not using sq_thread option)
- hipri flag set

Micro-benchmark results show a net gain of:
- increase of 18%-29% in IOPs
- reduction of 16%-22% in average latency
- reduction of 7%-23% in 99.99% latency

Baseline:
========
QDepth/Batch | IOPs [k] | Avg. Lat [us] | 99.99% Lat [us]
-----------------------------------------------------------------
1/1  | 32.4 | 30.11 | 50.94
32/8 | 179 | 168.20 | 371

CPU alignment:
=============
QDepth/Batch | IOPs [k] | Avg. Lat [us] | 99.99% Lat [us]
-----------------------------------------------------------------
1/1  | 38.5 |   25.18 | 39.16
32/8 | 231 |   130.75 | 343

Reported-by: Mark Wunderlich <mark.wunderlich@intel.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-pci: Simplify nvme_poll_irqdisable
Keith Busch [Wed, 4 Mar 2020 17:17:01 +0000 (09:17 -0800)]
nvme-pci: Simplify nvme_poll_irqdisable

The timeout handler can use the existing nvme_poll() if it needs to
check a polled queue, allowing nvme_poll_irqdisable() to handle only
irq driven queues for the remaining callers.

Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-pci: Remove two-pass completions
Keith Busch [Mon, 2 Mar 2020 16:56:53 +0000 (08:56 -0800)]
nvme-pci: Remove two-pass completions

Completion handling had been done in two steps: find all new completions
under a lock, then handle those completions outside the lock. This was
done to make the locked section as short as possible so that other
threads using the same lock wait less time.

The driver no longer shares locks during completion, and is in fact
lockless for interrupt driven queues, so the optimization no longer
serves its original purpose. Replace the two-pass completion queue
handler with a single pass that completes entries immediately.

Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-pci: Remove tag from process cq
Keith Busch [Mon, 2 Mar 2020 16:45:04 +0000 (08:45 -0800)]
nvme-pci: Remove tag from process cq

The only user for tagged completion was for timeout handling. That user,
though, really only cares if the timed out command is completed, which
we can safely check within the timeout handler.

Remove the tag check to simplify completion handling.

Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme-pci: slimmer CQ head update
Alexey Dobriyan [Fri, 28 Feb 2020 18:45:19 +0000 (21:45 +0300)]
nvme-pci: slimmer CQ head update

Update CQ head with pre-increment operator. This saves subtraction of 1
and a few registers.

Also update phase with "^= 1". This generates only one RMW instruction.

ffffffff815ba150 <nvme_update_cq_head>:
ffffffff815ba150:       0f b7 47 70             movzx  eax,WORD PTR [rdi+0x70]
ffffffff815ba154:       83 c0 01                add    eax,0x1
ffffffff815ba157:       66 89 47 70             mov    WORD PTR [rdi+0x70],ax
ffffffff815ba15b:       66 3b 47 68             cmp    ax,WORD PTR [rdi+0x68]
ffffffff815ba15f:       74 01                   je     ffffffff815ba162 <nvme_update_cq_head+0x12>
ffffffff815ba161:       c3                      ret
ffffffff815ba162:       31 c0                   xor    eax,eax
ffffffff815ba164:       80 77 74 01      ===>   xor    BYTE PTR [rdi+0x74],0x1
ffffffff815ba168:       66 89 47 70             mov    WORD PTR [rdi+0x70],ax
ffffffff815ba16c:       c3                      ret

add/remove: 0/0 grow/shrink: 0/3 up/down: 0/-119 (-119)
Function                                     old     new   delta
nvme_poll                                    690     678     -12
nvme_dev_disable                            1230    1177     -53
nvme_irq                                     613     559     -54

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
4 years agonvmet: check ncqr & nsqr for set-features cmd
Amit Engel [Sun, 1 Mar 2020 00:28:41 +0000 (16:28 -0800)]
nvmet: check ncqr & nsqr for set-features cmd

For set feature command when setting up NVME_FEAT_NUM_QUEUES, check
Number of I/O Completion Queues Requested (NCQR) and Number of I/O
Submission Queues Requested (NSQR) before we proceed, for invalid values
(i.e. 65535) return an appropriate NVMe invalid field status.

Signed-off-by: Amit Engel <Amit.Engel@dell.com>
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Check for readiness more quickly, to speed up boot time
Josh Triplett [Sat, 29 Feb 2020 02:52:28 +0000 (18:52 -0800)]
nvme: Check for readiness more quickly, to speed up boot time

After initialization, nvme_wait_ready checks for readiness every 100ms,
even though the drive may be ready far sooner than that. This delays
system boot by hundreds of milliseconds. Reduce the delay, checking for
readiness every millisecond instead.

Boot-time tests on an AWS c5.12xlarge:

Before:
[    0.546936] initcall nvme_init+0x0/0x5b returned 0 after 37 usecs
...
[    0.764178] nvme nvme0: 2/0/0 default/read/poll queues
[    0.768424]  nvme0n1: p1
[    0.774132] EXT4-fs (nvme0n1p1): mounted filesystem with ordered data mode. Opts: (null)
[    0.774146] VFS: Mounted root (ext4 filesystem) on device 259:1.
...
[    0.788141] Run /sbin/init as init process

After:
[    0.537088] initcall nvme_init+0x0/0x5b returned 0 after 37 usecs
...
[    0.543457] nvme nvme0: 2/0/0 default/read/poll queues
[    0.548473]  nvme0n1: p1
[    0.554339] EXT4-fs (nvme0n1p1): mounted filesystem with ordered data mode. Opts: (null)
[    0.554344] VFS: Mounted root (ext4 filesystem) on device 259:1.
...
[    0.567931] Run /sbin/init as init process

Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: log additional message for controller status
Rupesh Girase [Thu, 27 Feb 2020 16:45:26 +0000 (22:15 +0530)]
nvme: log additional message for controller status

Log the controller status to know more about issue if it
lies within kernel nvme subsytem or controller is unhealthy.

Signed-off-by: Rupesh Girase <rgirase@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulakrni@wdc.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: code cleanup nvme_identify_ns_desc()
Chaitanya Kulkarni [Wed, 19 Feb 2020 16:14:31 +0000 (08:14 -0800)]
nvme: code cleanup nvme_identify_ns_desc()

The function nvme_identify_ns_desc() has 3 levels of nesting which make
error message to exceeded > 80 char per line which is not aligned with
the kernel code standards and rest of the NVMe subsystem code.

Add a helper function to move the processing of the log when the
command is successful by reducing the nesting and keeping the
code < 80 char per line.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: Don't deter users from enabling hwmon support
Jean Delvare [Tue, 11 Feb 2020 12:41:36 +0000 (13:41 +0100)]
nvme: Don't deter users from enabling hwmon support

I see no good reason for the "If unsure, say N" advice in the description
of the NVME_HWMON configuration option. It is not dangerous, it does
not select any other option, and has a fairly low overhead.

As the option is already not enabled by default, further suggesting
hesitant users to not enable it is not useful anyway. Unlike some other
options where the description alone may not be sufficient for users to
make a decision, NVME_HWMON is pretty simple to grasp in my opinion,
so just let the user do what they want.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: expose hostid via sysfs for fabrics controllers
Sagi Grimberg [Sat, 8 Feb 2020 01:13:54 +0000 (17:13 -0800)]
nvme: expose hostid via sysfs for fabrics controllers

We allow userspace to connect with a custom hostid which is useful for
certain use-cases. However there is is no way to tell what is the hostid
used to connect to a given controller.

Expose this so userspace can correlate controllers based on hostid.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agonvme: expose hostnqn via sysfs for fabrics controllers
Sagi Grimberg [Sat, 8 Feb 2020 01:13:53 +0000 (17:13 -0800)]
nvme: expose hostnqn via sysfs for fabrics controllers

We allow userspace to connect with a custom hostnqn which is useful for
certain use-cases. However there is no way to tell what is the hostnqn
used to connect to a given controller.

Expose this so userspace can correlate controllers based on hostnqn.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
4 years agobcache: remove dupplicated declaration from btree.h
Coly Li [Wed, 25 Mar 2020 01:30:57 +0000 (09:30 +0800)]
bcache: remove dupplicated declaration from btree.h

Commit 253a99d95d5b ("bcache: move macro btree() and btree_root()
into btree.h") makes two duplicated declaration into btree.h,
typedef int (btree_map_keys_fn)();
int bch_btree_map_keys();

The kbuild test robot <lkp@intel.com> detects and reports this
problem and this patch fixes it by removing the duplicated ones.

Fixes: 253a99d95d5b ("bcache: move macro btree() and btree_root() into btree.h")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: optimize barrier usage for atomic operations
Coly Li [Sun, 22 Mar 2020 06:03:05 +0000 (14:03 +0800)]
bcache: optimize barrier usage for atomic operations

The idea of this patch is from Davidlohr Bueso, he posts a patch
for bcache to optimize barrier usage for read-modify-write atomic
bitops. Indeed such optimization can also apply on other locations
where smp_mb() is used before or after an atomic operation.

This patch replaces smp_mb() with smp_mb__before_atomic() or
smp_mb__after_atomic() in btree.c and writeback.c,  where it is used
to synchronize memory cache just earlier on other cores. Although
the locations are not on hot code path, it is always not bad to mkae
things a little better.

Signed-off-by: Coly Li <colyli@suse.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: optimize barrier usage for Rmw atomic bitops
Davidlohr Bueso [Sun, 22 Mar 2020 06:03:04 +0000 (14:03 +0800)]
bcache: optimize barrier usage for Rmw atomic bitops

We can avoid the unnecessary barrier on non LL/SC architectures,
such as x86. Instead, use the smp_mb__after_atomic().

Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: Use scnprintf() for avoiding potential buffer overflow
Takashi Iwai [Sun, 22 Mar 2020 06:03:03 +0000 (14:03 +0800)]
bcache: Use scnprintf() for avoiding potential buffer overflow

Since snprintf() returns the would-be-output size instead of the
actual output size, the succeeding calls may go beyond the given
buffer limit.  Fix it by replacing with scnprintf().

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: make bch_sectors_dirty_init() to be multithreaded
Coly Li [Sun, 22 Mar 2020 06:03:02 +0000 (14:03 +0800)]
bcache: make bch_sectors_dirty_init() to be multithreaded

When attaching a cached device (a.k.a backing device) to a cache
device, bch_sectors_dirty_init() is called to count dirty sectors
and stripes (see what bcache_dev_sectors_dirty_add() does) on the
cache device.

The counting is done by a single thread recursive function
bch_btree_map_keys() to iterate all the bcache btree nodes.
If the btree has huge number of nodes, bch_sectors_dirty_init() will
take quite long time. In my testing, if the registering cache set has
a existed UUID which matches a already registered cached device, the
automatical attachment during the registration may take more than
55 minutes. This is too long for waiting the bcache to work in real
deployment.

Fortunately when bch_sectors_dirty_init() is called, no other thread
will access the btree yet, it is safe to do a read-only parallelized
dirty sectors counting by multiple threads.

This patch tries to create multiple threads, and each thread tries to
one-by-one count dirty sectors from the sub-tree indexed by a root
node key which the thread fetched. After the sub-tree is counted, the
counting thread will continue to fetch another root node key, until
the fetched key is NULL. How many threads in parallel depends on
the number of keys from the btree root node, and the number of online
CPU core. The thread number will be the less number but no more than
BCH_DIRTY_INIT_THRD_MAX. If there are only 2 keys in root node, it
can only be 2x times faster by this patch. But if there are 10 keys
in the root node, with this patch it can be 10x times faster.

Signed-off-by: Coly Li <colyli@suse.de>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: make bch_btree_check() to be multithreaded
Coly Li [Sun, 22 Mar 2020 06:03:01 +0000 (14:03 +0800)]
bcache: make bch_btree_check() to be multithreaded

When registering a cache device, bch_btree_check() is called to check
all btree nodes, to make sure the btree is consistent and not
corrupted.

bch_btree_check() is recursively executed in a single thread, when there
are a lot of data cached and the btree is huge, it may take very long
time to check all the btree nodes. In my testing, I observed it took
around 50 minutes to finish bch_btree_check().

When checking the bcache btree nodes, the cache set is not running yet,
and indeed the whole tree is in read-only state, it is safe to create
multiple threads to check the btree in parallel.

This patch tries to create multiple threads, and each thread tries to
one-by-one check the sub-tree indexed by a key from the btree root node.
The parallel thread number depends on how many keys in the btree root
node. At most BCH_BTR_CHKTHREAD_MAX (64) threads can be created, but in
practice is should be min(cpu-number/2, root-node-keys-number).

Signed-off-by: Coly Li <colyli@suse.de>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: add bcache_ prefix to btree_root() and btree() macros
Coly Li [Sun, 22 Mar 2020 06:03:00 +0000 (14:03 +0800)]
bcache: add bcache_ prefix to btree_root() and btree() macros

This patch changes macro btree_root() and btree() to bcache_btree_root()
and bcache_btree(), to avoid potential generic name clash in future.

NOTE: for product kernel maintainers, this patch can be skipped if
you feel the rename stuffs introduce inconvenince to patch backport.

Suggested-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agobcache: move macro btree() and btree_root() into btree.h
Coly Li [Sun, 22 Mar 2020 06:02:59 +0000 (14:02 +0800)]
bcache: move macro btree() and btree_root() into btree.h

In order to accelerate bcache registration speed, the macro btree()
and btree_root() will be referenced out of btree.c. This patch moves
them from btree.c into btree.h with other relative function declaration
in btree.h, for the following changes.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agorsxx: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Wed, 12 Feb 2020 19:46:02 +0000 (13:46 -0600)]
rsxx: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertenly introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoMerge branch 'md-next' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/song...
Jens Axboe [Tue, 17 Mar 2020 21:15:34 +0000 (15:15 -0600)]
Merge branch 'md-next' of ssh://gitolite./linux/kernel/git/song/md into for-5.7/drivers

Pull MD fixes from Song.

* 'md-next' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/song/md:
  block: keep bdi->io_pages in sync with max_sectors_kb for stacked devices
  md: check arrays is suspended in mddev_detach before call quiesce operations

4 years agoblock: keep bdi->io_pages in sync with max_sectors_kb for stacked devices
Konstantin Khlebnikov [Fri, 28 Feb 2020 14:51:48 +0000 (17:51 +0300)]
block: keep bdi->io_pages in sync with max_sectors_kb for stacked devices

Field bdi->io_pages added in commit 9491ae4aade6 ("mm: don't cap request
size based on read-ahead setting") removes unneeded split of read requests.

Stacked drivers do not call blk_queue_max_hw_sectors(). Instead they set
limits of their devices by blk_set_stacking_limits() + disk_stack_limits().
Field bio->io_pages stays zero until user set max_sectors_kb via sysfs.

This patch updates io_pages after merging limits in disk_stack_limits().

Commit c6d6e9b0f6b4 ("dm: do not allow readahead to limit IO size") fixed
the same problem for device-mapper devices, this one fixes MD RAIDs.

Fixes: 9491ae4aade6 ("mm: don't cap request size based on read-ahead setting")
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Bob Liu <bob.liu@oracle.com>
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Signed-off-by: Song Liu <songliubraving@fb.com>
4 years agomd: check arrays is suspended in mddev_detach before call quiesce operations
Guoqing Jiang [Tue, 11 Feb 2020 10:10:04 +0000 (11:10 +0100)]
md: check arrays is suspended in mddev_detach before call quiesce operations

Don't call quiesce(1) and quiesce(0) if array is already suspended,
otherwise in level_store, the array is writable after mddev_detach
in below part though the intention is to make array writable after
resume.

mddev_suspend(mddev);
mddev_detach(mddev);
...
mddev_resume(mddev);

And it also causes calltrace as follows in [1].

[48005.653834] WARNING: CPU: 1 PID: 45380 at kernel/kthread.c:510 kthread_park+0x77/0x90
[...]
[48005.653976] CPU: 1 PID: 45380 Comm: mdadm Tainted: G           OE     5.4.10-arch1-1 #1
[48005.653979] Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./J4105-ITX, BIOS P1.40 08/06/2018
[48005.653984] RIP: 0010:kthread_park+0x77/0x90
[48005.654015] Call Trace:
[48005.654039]  r5l_quiesce+0x3c/0x70 [raid456]
[48005.654052]  raid5_quiesce+0x228/0x2e0 [raid456]
[48005.654073]  mddev_detach+0x30/0x70 [md_mod]
[48005.654090]  level_store+0x202/0x670 [md_mod]
[48005.654099]  ? security_capable+0x40/0x60
[48005.654114]  md_attr_store+0x7b/0xc0 [md_mod]
[48005.654123]  kernfs_fop_write+0xce/0x1b0
[48005.654132]  vfs_write+0xb6/0x1a0
[48005.654138]  ksys_write+0x67/0xe0
[48005.654146]  do_syscall_64+0x4e/0x140
[48005.654155]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[48005.654161] RIP: 0033:0x7fa0c8737497

[1]: https://bugzilla.kernel.org/show_bug.cgi?id=206161

Signed-off-by: Guoqing Jiang <guoqing.jiang@cloud.ionos.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
4 years agofloppy: rename the global "fdc" variable to "current_fdc"
Willy Tarreau [Sun, 1 Mar 2020 19:55:55 +0000 (20:55 +0100)]
floppy: rename the global "fdc" variable to "current_fdc"

This is done in order to remove the confusion that arises at some places
in the code where local variables or arguments shadow the global variable.
It is already visible that some places are a bit awkward and iterate over
the global variable, for the sole reason that they used to rely on it being
named "fdc" in order to get the correct address when using FD_DOR. These
ones are easy to spot by searching for "for (current_fdc...".

Some more cleanup is definitely possible. For example
"fdc_state[current_fdc].somefield" is used all over the code and would
probably be better with "fdc_state->somefield" with fdc_state being set
when current_fdc is assigned. This would require to pass the pointer to
the current state instead of the current_fdc to the I/O functions.

Link: https://lore.kernel.org/r/20200301195555.11154-7-w@1wt.eu
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: separate the FDC's base address from its registers
Willy Tarreau [Sun, 1 Mar 2020 19:55:54 +0000 (20:55 +0100)]
floppy: separate the FDC's base address from its registers

FDC registers FD_STATUS, FD_DATA, FD_DOR, FD_DIR and FD_DCR used to be
defined relative to FD_IOPORT, which is the FDC's base address, itself
a macro depending on the "fdc" local or global variable.

This patch changes this so that the register macros above now only
reference the address offset, and that the FDC's address is explicitly
passed in each call to fd_inb() and fd_outb(), thus removing the macro.
With this change there is no more implicit usage of the local/global
"fdc" variable.

One place in the ARM code used to check if the port was equal to FD_DOR,
this was changed to testing the register by applying a mask to the port,
as was already done in the sparc code.

There are still occurrences of fd_inb() and fd_outb() in the PARISC
code and these ones remain unaffected since they already used to work
with a base address and a register offset.

The sparc, m68k and parisc code could now be slightly cleaned up to
benefit from the macro definitions above instead of the equivalent
hard-coded values.

Link: https://lore.kernel.org/r/20200301195555.11154-6-w@1wt.eu
Cc: Ian Molton <spyro@f2s.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: introduce new functions fdc_inb() and fdc_outb()
Willy Tarreau [Sun, 1 Mar 2020 19:55:53 +0000 (20:55 +0100)]
floppy: introduce new functions fdc_inb() and fdc_outb()

These two functions replace fd_inb() and fd_outb() in that they take
the FDC in argument. This will ease the separation of the base address
and the port everywhere the code is used.

Link: https://lore.kernel.org/r/20200301195555.11154-5-w@1wt.eu
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: prepare ARM code to simplify base address separation
Willy Tarreau [Sun, 1 Mar 2020 19:55:52 +0000 (20:55 +0100)]
floppy: prepare ARM code to simplify base address separation

The fd_outb() macro on ARM relies on a special fd_setdor() macro when
the register is FD_DOR and both will need to be changed to accept a
separate base address. Let's just remerge them to simplify the change
and make this code more easily reviewable.

Link: https://lore.kernel.org/r/20200301195555.11154-4-w@1wt.eu
Cc: Ian Molton <spyro@f2s.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: remove incomplete support for second FDC from ARM code
Willy Tarreau [Sun, 1 Mar 2020 19:55:51 +0000 (20:55 +0100)]
floppy: remove incomplete support for second FDC from ARM code

The ARM code was written with the apparent hope to one day support
a second FDC except that the code was incomplete and only touches
the first one, which is also reflected by N_FDC==1. However this
made its fd_outb() macro artificially depend on the global or local
"fdc" variable.

Let's get rid of this and make it explicit it doesn't rely on this
variable anymore.

Link: https://lore.kernel.org/r/20200301195555.11154-3-w@1wt.eu
Cc: Ian Molton <spyro@f2s.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: remove dead code for drives scanning on ARM
Willy Tarreau [Sun, 1 Mar 2020 19:55:50 +0000 (20:55 +0100)]
floppy: remove dead code for drives scanning on ARM

On ARM, function fd_scandrives pre-dates Git era, is #ifed 0 out, not
used, and cannot even compile since it references an fdc variable that's
not declared anywhere (supposed to be the global one that we're turning
to current_fdc apparently).

There was also an ifdefde out include of mach/floppy.h that does not
exist anymore either. Let's get rid of them since they complicate the
fixing of the driver.

Link: https://lore.kernel.org/r/20200301195555.11154-2-w@1wt.eu
Cc: Ian Molton <spyro@f2s.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand the reply_buffer macros
Willy Tarreau [Mon, 24 Feb 2020 21:23:52 +0000 (22:23 +0100)]
floppy: cleanup: expand the reply_buffer macros

Several macros were used to access reply_buffer[] at discrete positions
without making it obvious they were relying on this. These ones have
been replaced by their offset in the reply buffer to make these accesses
more obvious.

Link: https://lore.kernel.org/r/20200224212352.8640-11-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand the R/W / format command macros
Willy Tarreau [Mon, 24 Feb 2020 21:23:51 +0000 (22:23 +0100)]
floppy: cleanup: expand the R/W / format command macros

Various macros were used to access raw_cmd for R/W or format commands
without making it obvious that raw_cmd->cmd[] was used. Let's expand
the macros to make this more obvious.

Link: https://lore.kernel.org/r/20200224212352.8640-10-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro DRWE
Willy Tarreau [Mon, 24 Feb 2020 21:23:50 +0000 (22:23 +0100)]
floppy: cleanup: expand macro DRWE

This macro doesn't bring much value and only slightly obfuscates the
code by silently using global variable "current_drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-9-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro DRS
Willy Tarreau [Mon, 24 Feb 2020 21:23:49 +0000 (22:23 +0100)]
floppy: cleanup: expand macro DRS

This macro doesn't bring much value and only slightly obfuscates the
code by silently using global variable "current_drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-8-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro DP
Willy Tarreau [Mon, 24 Feb 2020 21:23:48 +0000 (22:23 +0100)]
floppy: cleanup: expand macro DP

This macro doesn't bring much value and only slightly obfuscates the
code by silently using global variable "current_drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-7-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro UDRWE
Willy Tarreau [Mon, 24 Feb 2020 21:23:47 +0000 (22:23 +0100)]
floppy: cleanup: expand macro UDRWE

This macro doesn't bring much value and only slightly obfuscates the
code by silently using local variable "drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-6-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro UDRS
Willy Tarreau [Mon, 24 Feb 2020 21:23:46 +0000 (22:23 +0100)]
floppy: cleanup: expand macro UDRS

This macro doesn't bring much value and only slightly obfuscates the
code by silently using local variable "drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-5-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro UDP
Willy Tarreau [Mon, 24 Feb 2020 21:23:45 +0000 (22:23 +0100)]
floppy: cleanup: expand macro UDP

This macro doesn't bring much value and only slightly obfuscates the
code by silently using local variable "drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-4-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro UFDCS
Willy Tarreau [Mon, 24 Feb 2020 21:23:44 +0000 (22:23 +0100)]
floppy: cleanup: expand macro UFDCS

This macro doesn't bring much value and only slightly obfuscates the
code by silently using local variable "drive", let's expand it.

Link: https://lore.kernel.org/r/20200224212352.8640-3-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agofloppy: cleanup: expand macro FDCS
Willy Tarreau [Mon, 24 Feb 2020 21:23:43 +0000 (22:23 +0100)]
floppy: cleanup: expand macro FDCS

Macro FDCS silently uses identifier "fdc" which may be either the
global one or a local one. Let's expand the macro to make this more
obvious.

Link: https://lore.kernel.org/r/20200224212352.8640-2-w@1wt.eu
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: describe the usage of fault injection param
Dongli Zhang [Thu, 12 Mar 2020 22:01:40 +0000 (15:01 -0700)]
null_blk: describe the usage of fault injection param

As null_blk is a very good start point to test block layer, this patch
adds description and comments to 'timeout', 'requeue' and 'init_hctx' to
explain how to use fault injection with null_blk.

The nvme has similar with nvme_core.fail_request in the form of comment.

Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: fix spurious IO errors after failed past-wp access
Alexey Dobriyan [Wed, 12 Feb 2020 20:23:20 +0000 (23:23 +0300)]
null_blk: fix spurious IO errors after failed past-wp access

Steps to reproduce:

BLKRESETZONE zone 0

// force EIO
pwrite(fd, buf, 4096, 4096);

[issue more IO including zone ioctls]

It will start failing randomly including IO to unrelated zones because of
->error "reuse". Trigger can be partition detection as well if test is not
run immediately which is even more entertaining.

The fix is of course to clear ->error where necessary.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alexey Dobriyan (SK hynix) <adobriyan@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonbd: requeue command if the soecket is changed
Hou Pu [Fri, 28 Feb 2020 06:40:30 +0000 (01:40 -0500)]
nbd: requeue command if the soecket is changed

In commit 2da22da5734 (nbd: fix zero cmd timeout handling v2),
it is allowed to reset timer when it fires if tag_set.timeout
is set to zero. If the server is shutdown and a new socket
is reconfigured, the request should be requeued to be processed by
new server instead of waiting for response from the old one.

Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Hou Pu <houpu@bytedance.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonbd: enable replace socket if only one connection is configured
Hou Pu [Fri, 28 Feb 2020 06:40:29 +0000 (01:40 -0500)]
nbd: enable replace socket if only one connection is configured

Nbd server with multiple connections could be upgraded since
560bc4b (nbd: handle dead connections). But if only one conncection
is configured, after we take down nbd server, all inflight IO
would finally timeout and return error. We could requeue them
like what we do with multiple connections and wait for new socket
in submit path.

Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Hou Pu <houpu@bytedance.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoblock/drbd: delete invalid function drbd_md_mark_dirty_
Jackie Liu [Sat, 7 Mar 2020 02:39:25 +0000 (10:39 +0800)]
block/drbd: delete invalid function drbd_md_mark_dirty_

We deleted last_md_mark_dirty long ago, this function no longer needs to
exist, delete it, otherwise a compilation error will occur when DEBUG is
opened.

Fixes: ac0acb9e39ac ("drbd: use drbd_device_post_work() in more place")
Signed-off-by: Jackie Liu <liuyun01@kylinos.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agolightnvm: pblk: Use scnprintf() for avoiding potential buffer overflow
Takashi Iwai [Wed, 11 Mar 2020 07:44:39 +0000 (08:44 +0100)]
lightnvm: pblk: Use scnprintf() for avoiding potential buffer overflow

Since snprintf() returns the would-be-output size instead of the
actual output size, the succeeding calls may go beyond the given
buffer limit.  Fix it by replacing with scnprintf().

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoblock: aoe: Use scnprintf() for avoiding potential buffer overflow
Takashi Iwai [Wed, 11 Mar 2020 07:12:58 +0000 (08:12 +0100)]
block: aoe: Use scnprintf() for avoiding potential buffer overflow

Since snprintf() returns the would-be-output size instead of the
actual output size, the succeeding calls may go beyond the given
buffer limit.  Fix it by replacing with scnprintf().

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoloop: Only freeze block queue when needed.
Martijn Coenen [Tue, 10 Mar 2020 13:06:54 +0000 (14:06 +0100)]
loop: Only freeze block queue when needed.

__loop_update_dio() can be called as a part of loop_set_fd(), when the
block queue is not yet up and running; avoid freezing the block queue in
that case, since that is an expensive operation.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Martijn Coenen <maco@android.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoloop: Only change blocksize when needed.
Martijn Coenen [Tue, 10 Mar 2020 13:12:30 +0000 (14:12 +0100)]
loop: Only change blocksize when needed.

Return early in loop_set_block_size() if the requested block size is
identical to the one we already have; this avoids expensive calls to
freeze the block queue.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martijn Coenen <maco@android.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: Add support for init_hctx() fault injection
Bart Van Assche [Tue, 10 Mar 2020 04:26:23 +0000 (21:26 -0700)]
null_blk: Add support for init_hctx() fault injection

This makes it possible to test the error path in blk_mq_realloc_hw_ctxs()
and also several error paths in null_blk.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: Handle null_add_dev() failures properly
Bart Van Assche [Tue, 10 Mar 2020 04:26:22 +0000 (21:26 -0700)]
null_blk: Handle null_add_dev() failures properly

If null_add_dev() fails then null_del_dev() is called with a NULL argument.
Make null_del_dev() handle this scenario correctly. This patch fixes the
following KASAN complaint:

null-ptr-deref in null_del_dev+0x28/0x280 [null_blk]
Read of size 8 at addr 0000000000000000 by task find/1062

Call Trace:
 dump_stack+0xa5/0xe6
 __kasan_report.cold+0x65/0x99
 kasan_report+0x16/0x20
 __asan_load8+0x58/0x90
 null_del_dev+0x28/0x280 [null_blk]
 nullb_group_drop_item+0x7e/0xa0 [null_blk]
 client_drop_item+0x53/0x80 [configfs]
 configfs_rmdir+0x395/0x4e0 [configfs]
 vfs_rmdir+0xb6/0x220
 do_rmdir+0x238/0x2c0
 __x64_sys_unlinkat+0x75/0x90
 do_syscall_64+0x6f/0x2f0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: Fix the null_add_dev() error path
Bart Van Assche [Tue, 10 Mar 2020 04:26:21 +0000 (21:26 -0700)]
null_blk: Fix the null_add_dev() error path

If null_add_dev() fails, clear dev->nullb.

This patch fixes the following KASAN complaint:

BUG: KASAN: use-after-free in nullb_device_submit_queues_store+0xcf/0x160 [null_blk]
Read of size 8 at addr ffff88803280fc30 by task check/8409

Call Trace:
 dump_stack+0xa5/0xe6
 print_address_description.constprop.0+0x26/0x260
 __kasan_report.cold+0x7b/0x99
 kasan_report+0x16/0x20
 __asan_load8+0x58/0x90
 nullb_device_submit_queues_store+0xcf/0x160 [null_blk]
 configfs_write_file+0x1c4/0x250 [configfs]
 __vfs_write+0x4c/0x90
 vfs_write+0x145/0x2c0
 ksys_write+0xd7/0x180
 __x64_sys_write+0x47/0x50
 do_syscall_64+0x6f/0x2f0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x7ff370926317
Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 24
RSP: 002b:00007fff2dd2da48 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 00007ff370926317
RDX: 0000000000000002 RSI: 0000559437ef23f0 RDI: 0000000000000001
RBP: 0000559437ef23f0 R08: 000000000000000a R09: 0000000000000001
R10: 0000559436703471 R11: 0000000000000246 R12: 0000000000000002
R13: 00007ff370a006a0 R14: 00007ff370a014a0 R15: 00007ff370a008a0

Allocated by task 8409:
 save_stack+0x23/0x90
 __kasan_kmalloc.constprop.0+0xcf/0xe0
 kasan_kmalloc+0xd/0x10
 kmem_cache_alloc_node_trace+0x129/0x4c0
 null_add_dev+0x24a/0xe90 [null_blk]
 nullb_device_power_store+0x1b6/0x270 [null_blk]
 configfs_write_file+0x1c4/0x250 [configfs]
 __vfs_write+0x4c/0x90
 vfs_write+0x145/0x2c0
 ksys_write+0xd7/0x180
 __x64_sys_write+0x47/0x50
 do_syscall_64+0x6f/0x2f0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Freed by task 8409:
 save_stack+0x23/0x90
 __kasan_slab_free+0x112/0x160
 kasan_slab_free+0x12/0x20
 kfree+0xdf/0x250
 null_add_dev+0xaf3/0xe90 [null_blk]
 nullb_device_power_store+0x1b6/0x270 [null_blk]
 configfs_write_file+0x1c4/0x250 [configfs]
 __vfs_write+0x4c/0x90
 vfs_write+0x145/0x2c0
 ksys_write+0xd7/0x180
 __x64_sys_write+0x47/0x50
 do_syscall_64+0x6f/0x2f0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Fixes: 2984c8684f96 ("nullb: factor disk parameters")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: Fix changing the number of hardware queues
Bart Van Assche [Tue, 10 Mar 2020 04:26:20 +0000 (21:26 -0700)]
null_blk: Fix changing the number of hardware queues

Instead of initializing null_blk hardware queues explicitly after the
request queue has been created, provide .init_hctx() and .exit_hctx()
callback functions. The latter functions are not only called during
request queue allocation but also when the number of hardware queues
changes. Allocate nr_cpu_ids queues during initialization to support
increasing the number of hardware queues above the initial hardware
queue count.

This change fixes increasing the number of hardware queues above the
initial number of hardware queues and also keeps nullb->nr_queues in
sync with the number of hardware queues.

Fixes: 45919fbfe1c4 ("null_blk: Enable modifying 'submit_queues' after an instance has been configured")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agonull_blk: Suppress an UBSAN complaint triggered when setting 'memory_backed'
Bart Van Assche [Tue, 10 Mar 2020 04:26:19 +0000 (21:26 -0700)]
null_blk: Suppress an UBSAN complaint triggered when setting 'memory_backed'

Although it is not clear to me why UBSAN complains when 'memory_backed'
is set, this patch suppresses the UBSAN complaint that is triggered when
setting that configfs attribute.

UBSAN: Undefined behaviour in drivers/block/null_blk_main.c:327:1
load of value 16 is not a valid value for type '_Bool'
CPU: 2 PID: 8396 Comm: check Not tainted 5.6.0-rc1-dbg+ #14
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
Call Trace:
 dump_stack+0xa5/0xe6
 ubsan_epilogue+0x9/0x26
 __ubsan_handle_load_invalid_value+0x6d/0x76
 nullb_device_memory_backed_store.cold+0x2c/0x38 [null_blk]
 configfs_write_file+0x1c4/0x250 [configfs]
 __vfs_write+0x4c/0x90
 vfs_write+0x145/0x2c0
 ksys_write+0xd7/0x180
 __x64_sys_write+0x47/0x50
 do_syscall_64+0x6f/0x2f0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoblk-mq: Fix a recently introduced regression in blk_mq_realloc_hw_ctxs()
Bart Van Assche [Tue, 10 Mar 2020 04:26:18 +0000 (21:26 -0700)]
blk-mq: Fix a recently introduced regression in blk_mq_realloc_hw_ctxs()

q->nr_hw_queues must only be updated once it is known that
blk_mq_realloc_hw_ctxs() has succeeded. Otherwise it can happen that
reallocation fails and that q->nr_hw_queues is larger than the number of
allocated hardware queues. This patch fixes the following crash if
increasing the number of hardware queues fails:

BUG: KASAN: null-ptr-deref in blk_mq_map_swqueue+0x775/0x810
Write of size 8 at addr 0000000000000118 by task check/977

CPU: 3 PID: 977 Comm: check Not tainted 5.6.0-rc1-dbg+ #8
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
Call Trace:
 dump_stack+0xa5/0xe6
 __kasan_report.cold+0x65/0x99
 kasan_report+0x16/0x20
 check_memory_region+0x140/0x1b0
 memset+0x28/0x40
 blk_mq_map_swqueue+0x775/0x810
 blk_mq_update_nr_hw_queues+0x468/0x710
 nullb_device_submit_queues_store+0xf7/0x1a0 [null_blk]
 configfs_write_file+0x1c4/0x250 [configfs]
 __vfs_write+0x4c/0x90
 vfs_write+0x145/0x2c0
 ksys_write+0xd7/0x180
 __x64_sys_write+0x47/0x50
 do_syscall_64+0x6f/0x2f0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Fixes: ac0d6b926e74 ("block: Reduce the amount of memory required per request queue")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Cc: Keith Busch <kbusch@kernel.org>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoblk-mq: Keep set->nr_hw_queues and set->map[].nr_queues in sync
Bart Van Assche [Tue, 10 Mar 2020 04:26:17 +0000 (21:26 -0700)]
blk-mq: Keep set->nr_hw_queues and set->map[].nr_queues in sync

blk_mq_map_queues() and multiple .map_queues() implementations expect that
set->map[HCTX_TYPE_DEFAULT].nr_queues is set to the number of hardware
queues. Hence set .nr_queues before calling these functions. This patch
fixes the following kernel warning:

WARNING: CPU: 0 PID: 2501 at include/linux/cpumask.h:137
Call Trace:
 blk_mq_run_hw_queue+0x19d/0x350 block/blk-mq.c:1508
 blk_mq_run_hw_queues+0x112/0x1a0 block/blk-mq.c:1525
 blk_mq_requeue_work+0x502/0x780 block/blk-mq.c:775
 process_one_work+0x9af/0x1740 kernel/workqueue.c:2269
 worker_thread+0x98/0xe40 kernel/workqueue.c:2415
 kthread+0x361/0x430 kernel/kthread.c:255

Fixes: ed76e329d74a ("blk-mq: abstract out queue map") # v5.0
Reported-by: syzbot+d44e1b26ce5c3e77458d@syzkaller.appspotmail.com
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Cc: Johannes Thumshirn <jth@kernel.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoblk-mq: Fix a comment in include/linux/blk-mq.h
Bart Van Assche [Tue, 10 Mar 2020 04:26:16 +0000 (21:26 -0700)]
blk-mq: Fix a comment in include/linux/blk-mq.h

The 'hctx_list' member of struct blk_mq_hw_ctx is not a list head but
instead an entry in q->unused_hctx_list. Fix the comment above this
struct member.

Fixes: d386732bc142 ("blk-mq: fill header with kernel-doc")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Cc: André Almeida <andrealmeid@collabora.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoLinux 5.6-rc5
Linus Torvalds [Mon, 9 Mar 2020 00:44:44 +0000 (17:44 -0700)]
Linux 5.6-rc5

4 years agoMerge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Linus Torvalds [Mon, 9 Mar 2020 00:36:22 +0000 (17:36 -0700)]
Merge tag 'armsoc-fixes' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Olof Johansson:
 "We've been accruing these for a couple of weeks, so the batch is a bit
  bigger than usual.

  Largest delta is due to a led-bl driver that is added -- there was a
  miscommunication before the merge window and the driver didn't make it
  in. Due to this, the platforms needing it regressed. At this point, it
  seemed easier to add the new driver than unwind the changes.

  Besides that, there are a handful of various fixes:

   - AMD tee memory leak fix

   - A handful of fixlets for i.MX SCU communication

   - A few maintainers woke up and realized DEBUG_FS had been missing
     for a while, so a few updates of that.

  ... and the usual collection of smaller fixes to various platforms"

* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (37 commits)
  ARM: socfpga_defconfig: Add back DEBUG_FS
  arm64: dts: socfpga: agilex: Fix gmac compatible
  ARM: bcm2835_defconfig: Explicitly restore CONFIG_DEBUG_FS
  arm64: dts: meson: fix gxm-khadas-vim2 wifi
  arm64: dts: meson-sm1-sei610: add missing interrupt-names
  ARM: meson: Drop unneeded select of COMMON_CLK
  ARM: dts: bcm2711: Add pcie0 alias
  ARM: dts: bcm283x: Add missing properties to the PWR LED
  tee: amdtee: fix memory leak in amdtee_open_session()
  ARM: OMAP2+: Fix compile if CONFIG_HAVE_ARM_SMCCC is not set
  arm: dts: dra76x: Fix mmc3 max-frequency
  ARM: dts: dra7: Add "dma-ranges" property to PCIe RC DT nodes
  bus: ti-sysc: Fix 1-wire reset quirk
  ARM: dts: r8a7779: Remove deprecated "renesas, rcar-sata" compatible value
  soc: imx-scu: Align imx sc msg structs to 4
  firmware: imx: Align imx_sc_msg_req_cpu_start to 4
  firmware: imx: scu-pd: Align imx sc msg structs to 4
  firmware: imx: misc: Align imx sc msg structs to 4
  firmware: imx: scu: Ensure sequential TX
  ARM: dts: imx7-colibri: Fix frequency for sd/mmc
  ...

4 years agoMerge tag 'edac_urgent-2020-03-08' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 9 Mar 2020 00:33:52 +0000 (17:33 -0700)]
Merge tag 'edac_urgent-2020-03-08' of git://git./linux/kernel/git/ras/ras

Pull EDAC fix from Borislav Petkov:
 "Error reporting fix for synopsys_edac: do not overwrite partial
  decoded error message (Sherry Sun)"

* tag 'edac_urgent-2020-03-08' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras:
  EDAC/synopsys: Do not print an error with back-to-back snprintf() calls

4 years agoMerge tag 'char-misc-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sun, 8 Mar 2020 15:49:44 +0000 (10:49 -0500)]
Merge tag 'char-misc-5.6-rc5' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc fixes from Greg KH:
 "Here are four small char/misc driver fixes for reported issues for
  5.6-rc5.

  These fixes are:

   - binder fix for a potential use-after-free problem found (took two
     tries to get it right)

   - interconnect core fix

   - altera-stapl driver fix

  All four of these have been in linux-next for a while with no reported
  issues"

* tag 'char-misc-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  binder: prevent UAF for binderfs devices II
  interconnect: Handle memory allocation errors
  altera-stapl: altera_get_note: prevent write beyond end of 'key'
  binder: prevent UAF for binderfs devices

4 years agoMerge tag 'driver-core-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 8 Mar 2020 15:39:40 +0000 (10:39 -0500)]
Merge tag 'driver-core-5.6-rc5' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core and debugfs fixes from Greg KH:
 "Here are four small driver core / debugfs patches for 5.6-rc3:

   - debugfs api cleanup now that all debugfs_create_regset32() callers
     have been fixed up. This was waiting until after the -rc1 merge as
     these fixes came in through different trees

   - driver core sync state fixes based on reports of minor issues found
     in the feature

  All of these have been in linux-next with no reported issues"

* tag 'driver-core-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  driver core: Skip unnecessary work when device doesn't have sync_state()
  driver core: Add dev_has_sync_state()
  driver core: Call sync_state() even if supplier has no consumers
  debugfs: remove return value of debugfs_create_regset32()

4 years agoMerge tag 'tty-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sun, 8 Mar 2020 15:35:04 +0000 (10:35 -0500)]
Merge tag 'tty-5.6-rc5' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial fixes from Greg KH:
 "Here are some small tty/serial fixes for 5.6-rc5

  Just some small serial driver fixes, and a vt core fixup, full details
  are:

   - vt fixes for issues found by syzbot

   - serdev fix for Apple boxes

   - fsl_lpuart serial driver fixes

   - MAINTAINER update for incorrect serial files

   - new device ids for 8250_exar driver

   - mvebu-uart fix

  All of these have been in linux-next with no reported issues"

* tag 'tty-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  tty: serial: fsl_lpuart: free IDs allocated by IDA
  Revert "tty: serial: fsl_lpuart: drop EARLYCON_DECLARE"
  serdev: Fix detection of UART devices on Apple machines.
  MAINTAINERS: Add missed files related to Synopsys DesignWare UART
  serial: 8250_exar: add support for ACCES cards
  tty:serial:mvebu-uart:fix a wrong return
  vt: selection, push sel_lock up
  vt: selection, push console lock down

4 years agoMerge tag 'usb-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sun, 8 Mar 2020 15:32:23 +0000 (10:32 -0500)]
Merge tag 'usb-5.6-rc5' of git://git./linux/kernel/git/gregkh/usb

Pull USB/PHY fixes from Greg KH:
 "Here are some small USB and PHY driver fixes for reported issues for
  5.6-rc5.

  Included in here are:

   - phy driver fixes

   - new USB quirks

   - USB cdns3 gadget driver fixes

   - USB hub core fixes

  All of these have been in linux-next with no reported issues"

* tag 'usb-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  usb: dwc3: gadget: Update chain bit correctly when using sg list
  usb: core: port: do error out if usb_autopm_get_interface() fails
  usb: core: hub: do error out if usb_autopm_get_interface() fails
  usb: core: hub: fix unhandled return by employing a void function
  usb: storage: Add quirk for Samsung Fit flash
  usb: quirks: add NO_LPM quirk for Logitech Screen Share
  usb: usb251xb: fix regulator probe and error handling
  phy: allwinner: Fix GENMASK misuse
  usb: cdns3: gadget: toggle cycle bit before reset endpoint
  usb: cdns3: gadget: link trb should point to next request
  phy: mapphone-mdm6600: Fix timeouts by adding wake-up handling
  phy: brcm-sata: Correct MDIO operations for 40nm platforms
  phy: ti: gmii-sel: do not fail in case of gmii
  phy: ti: gmii-sel: fix set of copy-paste errors
  phy: core: Fix phy_get() to not return error on link creation failure
  phy: mapphone-mdm6600: Fix write timeouts with shorter GPIO toggle interval

4 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Linus Torvalds [Sun, 8 Mar 2020 01:52:55 +0000 (19:52 -0600)]
Merge tag 'for-linus' of git://git./linux/kernel/git/rdma/rdma

Pull rdma fixes from Jason Gunthorpe:
 "Nothing particularly exciting, some small ODP regressions from the mmu
  notifier rework, another bunch of syzkaller fixes, and a bug fix for a
  botched syzkaller fix in the first rc pull request.

   - Fix busted syzkaller fix in 'get_new_pps' - this turned out to
     crash on certain HW configurations

   - Bug fixes for various missed things in error unwinds

   - Add a missing rcu_read_lock annotation in hfi/qib

   - Fix two ODP related regressions from the recent mmu notifier
     changes

   - Several more syzkaller bugs in siw, RDMA netlink, verbs and iwcm

   - Revert an old patch in CMA as it is now shown to not be allocating
     port numbers properly"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  RDMA/iwcm: Fix iwcm work deallocation
  RDMA/siw: Fix failure handling during device creation
  RDMA/nldev: Fix crash when set a QP to a new counter but QPN is missing
  RDMA/odp: Ensure the mm is still alive before creating an implicit child
  RDMA/core: Fix protection fault in ib_mr_pool_destroy
  IB/mlx5: Fix implicit ODP race
  IB/hfi1, qib: Ensure RCU is locked when accessing list
  RDMA/core: Fix pkey and port assignment in get_new_pps
  RMDA/cm: Fix missing ib_cm_destroy_id() in ib_cm_insert_listen()
  RDMA/rw: Fix error flow during RDMA context initialization
  RDMA/core: Fix use of logical OR in get_new_pps
  Revert "RDMA/cma: Simplify rdma_resolve_addr() error flow"

4 years agoMerge tag 'io_uring-5.6-2020-03-07' of git://git.kernel.dk/linux-block
Linus Torvalds [Sat, 7 Mar 2020 20:20:29 +0000 (14:20 -0600)]
Merge tag 'io_uring-5.6-2020-03-07' of git://git.kernel.dk/linux-block

Pull io_uring fixes from Jens Axboe:
 "Here are a few io_uring fixes that should go into this release. This
  contains:

   - Removal of (now) unused io_wq_flush() and associated flag (Pavel)

   - Fix cancelation lockup with linked timeouts (Pavel)

   - Fix for potential use-after-free when freeing percpu ref for fixed
     file sets

   - io-wq cancelation fixups (Pavel)"

* tag 'io_uring-5.6-2020-03-07' of git://git.kernel.dk/linux-block:
  io_uring: fix lockup with timeouts
  io_uring: free fixed_file_data after RCU grace period
  io-wq: remove io_wq_flush and IO_WQ_WORK_INTERNAL
  io-wq: fix IO_WQ_WORK_NO_CANCEL cancellation

4 years agoMerge tag 'block-5.6-2020-03-07' of git://git.kernel.dk/linux-block
Linus Torvalds [Sat, 7 Mar 2020 20:14:38 +0000 (14:14 -0600)]
Merge tag 'block-5.6-2020-03-07' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "Here are a few fixes that should go into this release. This contains:

   - Revert of a bad bcache patch from this merge window

   - Removed unused function (Daniel)

   - Fixup for the blktrace fix from Jan from this release (Cengiz)

   - Fix of deeper level bfqq overwrite in BFQ (Carlo)"

* tag 'block-5.6-2020-03-07' of git://git.kernel.dk/linux-block:
  block, bfq: fix overwrite of bfq_group pointer in bfq_find_set_group()
  blktrace: fix dereference after null check
  Revert "bcache: ignore pending signals when creating gc and allocator thread"
  block: Remove used kblockd_schedule_work_on()

4 years agoMerge tag 'media/v5.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Sat, 7 Mar 2020 18:00:13 +0000 (12:00 -0600)]
Merge tag 'media/v5.6-2' of git://git./linux/kernel/git/mchehab/linux-media

Pull media fixes from Mauro Carvalho Chehab:

 - a fix for the media controller links in both hantro driver and in
   v4l2-mem2mem core

 - some fixes for the pulse8-cec driver

 - vicodec: handle alpha channel for RGB32 formats, as it may be used

 - mc-entity.c: fix handling of pad flags

* tag 'media/v5.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  media: hantro: Fix broken media controller links
  media: mc-entity.c: use & to check pad flags, not ==
  media: v4l2-mem2mem.c: fix broken links
  media: vicodec: process all 4 components for RGB32 formats
  media: pulse8-cec: close serio in disconnect, not adap_free
  media: pulse8-cec: INIT_DELAYED_WORK was called too late

4 years agoio_uring: fix lockup with timeouts
Pavel Begunkov [Fri, 6 Mar 2020 22:15:22 +0000 (01:15 +0300)]
io_uring: fix lockup with timeouts

There is a recipe to deadlock the kernel: submit a timeout sqe with a
linked_timeout (e.g.  test_single_link_timeout_ception() from liburing),
and SIGKILL the process.

Then, io_kill_timeouts() takes @ctx->completion_lock, but the timeout
isn't flagged with REQ_F_COMP_LOCKED, and will try to double grab it
during io_put_free() to cancel the linked timeout. Probably, the same
can happen with another io_kill_timeout() call site, that is
io_commit_cqring().

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoMerge tag 's390-5.6-5' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Sat, 7 Mar 2020 14:12:47 +0000 (08:12 -0600)]
Merge tag 's390-5.6-5' of git://git./linux/kernel/git/s390/linux

Pull s390 fixes from Vasily Gorbik:

 - Fix panic in gup_fast on large pud by providing an implementation of
   pud_write. This has been overlooked during migration to common gup
   code.

 - Fix unexpected write combining on PCI stores.

* tag 's390-5.6-5' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/pci: Fix unexpected write combine on resource
  s390/mm: fix panic in gup_fast on large pud

4 years agoMerge tag 'powerpc-5.6-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sat, 7 Mar 2020 14:10:34 +0000 (08:10 -0600)]
Merge tag 'powerpc-5.6-4' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:
 "Some more powerpc fixes for 5.6:

   - One fix for a recent regression to our breakpoint/watchpoint code.

   - Another fix for our KUAP support, this time a missing annotation in
     a rarely used path in signal handling.

   - A fix for our handling of a CPU feature that effects the PMU, when
     booting guests in some configurations.

   - A minor fix to our linker script to explicitly include the .BTF
     section.

  Thanks to: Christophe Leroy, Desnes A. Nunes do Rosario, Leonardo
  Bras, Naveen N. Rao, Ravi Bangoria, Stefan Berger"

* tag 'powerpc-5.6-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/mm: Fix missing KUAP disable in flush_coherent_icache()
  powerpc: fix hardware PMU exception bug on PowerVM compatibility mode systems
  powerpc: Include .BTF section
  powerpc/watchpoint: Don't call dar_within_range() for Book3S

4 years agoMerge tag 'for-linus-5.6b-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 7 Mar 2020 14:04:54 +0000 (08:04 -0600)]
Merge tag 'for-linus-5.6b-rc5-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fixes from Juergen Gross:
 "Four fixes and a small cleanup patch:

   - two fixes by Dongli Zhang fixing races in the xenbus driver

   - two fixes by me fixing issues introduced in 5.6

   - a small cleanup by Gustavo Silva replacing a zero-length array with
     a flexible-array"

* tag 'for-linus-5.6b-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/blkfront: fix ring info addressing
  xen/xenbus: fix locking
  xenbus: req->err should be updated before req->state
  xenbus: req->body should be updated before req->state
  xen: Replace zero-length array with flexible-array member

4 years agoMerge tag 'for-linus-2020-03-07' of gitolite.kernel.org:pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 7 Mar 2020 14:01:43 +0000 (08:01 -0600)]
Merge tag 'for-linus-2020-03-07' of gitolite.pub/scm/linux/kernel/git/brauner/linux

Pull thread fixes from Christian Brauner:
 "Here are a few hopefully uncontroversial fixes:

   - Use RCU_INIT_POINTER() when initializing rcu protected members in
     task_struct to fix sparse warnings.

   - Add pidfd_fdinfo_test binary to .gitignore file"

* tag 'for-linus-2020-03-07' of gitolite.kernel.org:pub/scm/linux/kernel/git/brauner/linux:
  selftests: pidfd: Add pidfd_fdinfo_test in .gitignore
  exit: Fix Sparse errors and warnings
  fork: Use RCU_INIT_POINTER() instead of rcu_access_pointer()

4 years agoMerge tag 'sound-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Sat, 7 Mar 2020 13:59:30 +0000 (07:59 -0600)]
Merge tag 'sound-5.6-rc5' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "The regular "bump-in-the-middle" updates, containing mostly ASoC-
  related fixes at this time. All changes are reasonably small.

  A few entries are for ASoC and ALSA core parts (DAPM, PCM, topology)
  for followups of the recent changes and potential buffer overflow by
  snprintf(), while the rest are (both new and old) device-specific
  fixes for Intel, meson, tas2562, rt1015, as well as the usual HD-audio
  quirks"

* tag 'sound-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (25 commits)
  ALSA: sgio2audio: Remove usage of dropped hw_params/hw_free functions
  ALSA: hda/realtek - Enable the headset of ASUS B9450FA with ALC294
  ALSA: hda/realtek - Fix silent output on Gigabyte X570 Aorus Master
  ALSA: hda/realtek - Add Headset Button supported for ThinkPad X1
  ALSA: hda/realtek - Add Headset Mic supported
  ASoC: wm8741: Fix typo in Kconfig prompt
  ASoC: stm32: sai: manage rebind issue
  ASoC: SOF: Fix snd_sof_ipc_stream_posn()
  ASoC: rt1015: modify pre-divider for sysclk
  ASoC: rt1015: add operation callback function for rt1015_dai[]
  ASoC: soc-component: tidyup snd_soc_pcm_component_sync_stop()
  ASoC: dapm: Correct DAPM handling of active widgets during shutdown
  ASoC: tas2562: Fix sample rate error message
  ASoC: Intel: Skylake: Fix available clock counter incrementation
  ASoC: soc-pcm/soc-compress: don't use snd_soc_dapm_stream_stop()
  ASoC: meson: g12a: add tohdmitx reset
  ASoC: pcm512x: Fix unbalanced regulator enable call in probe error path
  ASoC: soc-core: fix for_rtd_codec_dai_rollback() macro
  ASoC: topology: Fix memleak in soc_tplg_manifest_load()
  ASoC: topology: Fix memleak in soc_tplg_link_elems_load()
  ...

4 years agoMerge tag 'asoc-fix-v5.6-rc4' of https://git.kernel.org/pub/scm/linux/kernel/git...
Takashi Iwai [Sat, 7 Mar 2020 06:24:36 +0000 (07:24 +0100)]
Merge tag 'asoc-fix-v5.6-rc4' of https://git./linux/kernel/git/broonie/sound into for-linus

ASoC: Fixes for v5.6

More fixes that have arrived since the merge window, spread out all
over.  There's a few things like the operation callback addition for
rt1015 and the meson reset addition which add small new bits of
functionality to fix non-working systems, they're all very small and for
parts of newly added functionality.

4 years agoMerge tag 'linux-kselftest-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 6 Mar 2020 23:03:37 +0000 (17:03 -0600)]
Merge tag 'linux-kselftest-5.6-rc5' of git://git./linux/kernel/git/shuah/linux-kselftest

Pull kselftest update from Shuah Khan:
 "This consists of a cleanup patch to undo changes to global .gitignore
  that added selftests/lkdtm objects and add them to a local
  selftests/lkdtm/.gitignore.

  Summary of Linus's comments on local vs. global gitignore scope:

   - Keep local gitignore patterns in local files.

   - Put only global gitignore patterns in the top-level gitignore file.

  Local scope keeps things much better separated. It also incidentally
  means that if a directory gets renamed, the gitignore file continues
  to work unless in the case of renaming the actual files themselves
  that are named in the gitignore"

* tag 'linux-kselftest-5.6-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
  selftest/lkdtm: Use local .gitignore