platform/kernel/linux-rpi.git
3 years agoMerge tag 'linux-kselftest-kunit-fixes-5.10-rc5' of git://git.kernel.org/pub/scm...
Linus Torvalds [Wed, 18 Nov 2020 19:57:55 +0000 (11:57 -0800)]
Merge tag 'linux-kselftest-kunit-fixes-5.10-rc5' of git://git./linux/kernel/git/shuah/linux-kselftest

Pull Kunit fixes from Shuah Khan:
 "Several fixes to Kunit documentation and tools, and to not pollute
  the source directory.

  Also remove the incorrect kunit .gitattributes file"

* tag 'linux-kselftest-kunit-fixes-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
  kunit: fix display of failed expectations for strings
  kunit: tool: fix extra trailing \n in raw + parsed test output
  kunit: tool: print out stderr from make (like build warnings)
  KUnit: Docs: usage: wording fixes
  KUnit: Docs: style: fix some Kconfig example issues
  KUnit: Docs: fix a wording typo
  kunit: Do not pollute source directory with generated files (test.log)
  kunit: Do not pollute source directory with generated files (.kunitconfig)
  kunit: tool: fix pre-existing python type annotation errors
  kunit: Fix kunit.py parse subcommand (use null build_dir)
  kunit: tool: unmark test_data as binary blobs

3 years agonet: dsa: mv88e6xxx: Wait for EEPROM done after HW reset
Andrew Lunn [Mon, 16 Nov 2020 16:43:01 +0000 (08:43 -0800)]
net: dsa: mv88e6xxx: Wait for EEPROM done after HW reset

When the switch is hardware reset, it reads the contents of the
EEPROM. This can contain instructions for programming values into
registers and to perform waits between such programming. Reading the
EEPROM can take longer than the 100ms mv88e6xxx_hardware_reset() waits
after deasserting the reset GPIO. So poll the EEPROM done bit to
ensure it is complete.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Ruslan Sushko <rus@sushko.dev>
Link: https://lore.kernel.org/r/20201116164301.977661-1-rus@sushko.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'mlxsw-couple-of-fixes'
Jakub Kicinski [Wed, 18 Nov 2020 19:07:03 +0000 (11:07 -0800)]
Merge branch 'mlxsw-couple-of-fixes'

Ido Schimmel says:

====================
mlxsw: Couple of fixes

Patch #1 fixes firmware flashing when CONFIG_MLXSW_CORE=y and
CONFIG_MLXFW=m.

Patch #2 prevents EMAD transactions from needlessly failing when the
system is under heavy load by using exponential backoff.

Please consider patch #2 for stable.
====================

Link: https://lore.kernel.org/r/20201117173352.288491-1-idosch@idosch.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agomlxsw: core: Use variable timeout for EMAD retries
Ido Schimmel [Tue, 17 Nov 2020 17:33:52 +0000 (19:33 +0200)]
mlxsw: core: Use variable timeout for EMAD retries

The driver sends Ethernet Management Datagram (EMAD) packets to the
device for configuration purposes and waits for up to 200ms for a reply.
A request is retried up to 5 times.

When the system is under heavy load, replies are not always processed in
time and EMAD transactions fail.

Make the process more robust to such delays by using exponential
backoff. First wait for up to 200ms, then retransmit and wait for up to
400ms and so on.

Fixes: caf7297e7ab5 ("mlxsw: core: Introduce support for asynchronous EMAD register access")
Reported-by: Denis Yulevich <denisyu@nvidia.com>
Tested-by: Denis Yulevich <denisyu@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agomlxsw: Fix firmware flashing
Ido Schimmel [Tue, 17 Nov 2020 17:33:51 +0000 (19:33 +0200)]
mlxsw: Fix firmware flashing

The commit cited below moved firmware flashing functionality from
mlxsw_spectrum to mlxsw_core, but did not adjust the Kconfig
dependencies. This makes it possible to have mlxsw_core as built-in and
mlxfw as a module. The mlxfw code is therefore not reachable from
mlxsw_core and firmware flashing fails:

# devlink dev flash pci/0000:01:00.0 file mellanox/mlxsw_spectrum-13.2008.1310.mfa2
devlink answers: Operation not supported

Fix by having mlxsw_core select mlxfw.

Fixes: b79cb787ac70 ("mlxsw: Move fw flashing code into core.c")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reported-by: Vadim Pasternak <vadimp@nvidia.com>
Tested-by: Vadim Pasternak <vadimp@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: Have netpoll bring-up DSA management interface
Florian Fainelli [Tue, 17 Nov 2020 03:52:34 +0000 (19:52 -0800)]
net: Have netpoll bring-up DSA management interface

DSA network devices rely on having their DSA management interface up and
running otherwise their ndo_open() will return -ENETDOWN. Without doing
this it would not be possible to use DSA devices as netconsole when
configured on the command line. These devices also do not utilize the
upper/lower linking so the check about the netpoll device having upper
is not going to be a problem.

The solution adopted here is identical to the one done for
net/ipv4/ipconfig.c with 728c02089a0e ("net: ipv4: handle DSA enabled
master network devices"), with the network namespace scope being
restricted to that of the process configuring netpoll.

Fixes: 04ff53f96a93 ("net: dsa: Add netconsole support")
Tested-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20201117035236.22658-1-f.fainelli@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoatl1e: fix error return code in atl1e_probe()
Zhang Changzhong [Tue, 17 Nov 2020 02:57:55 +0000 (10:57 +0800)]
atl1e: fix error return code in atl1e_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: a6a5325239c2 ("atl1e: Atheros L1E Gigabit Ethernet driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605581875-36281-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoatl1c: fix error return code in atl1c_probe()
Zhang Changzhong [Tue, 17 Nov 2020 02:55:21 +0000 (10:55 +0800)]
atl1c: fix error return code in atl1c_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 43250ddd75a3 ("atl1c: Atheros L1C Gigabit Ethernet driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605581721-36028-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoah6: fix error return code in ah6_input()
Zhang Changzhong [Tue, 17 Nov 2020 02:45:05 +0000 (10:45 +0800)]
ah6: fix error return code in ah6_input()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605581105-35295-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: usb: qmi_wwan: Set DTR quirk for MR400
Filip Moc [Tue, 17 Nov 2020 17:36:31 +0000 (18:36 +0100)]
net: usb: qmi_wwan: Set DTR quirk for MR400

LTE module MR400 embedded in TL-MR6400 v4 requires DTR to be set.

Signed-off-by: Filip Moc <dev@moc6.cz>
Acked-by: Bjørn Mork <bjorn@mork.no>
Link: https://lore.kernel.org/r/20201117173631.GA550981@moc6.cz
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoregulator: ti-abb: Fix array out of bound read access on the first transition
Nishanth Menon [Wed, 18 Nov 2020 14:50:09 +0000 (08:50 -0600)]
regulator: ti-abb: Fix array out of bound read access on the first transition

At the start of driver initialization, we do not know what bias
setting the bootloader has configured the system for and we only know
for certain the very first time we do a transition.

However, since the initial value of the comparison index is -EINVAL,
this negative value results in an array out of bound access on the
very first transition.

Since we don't know what the setting is, we just set the bias
configuration as there is nothing to compare against. This prevents
the array out of bound access.

NOTE: Even though we could use a more relaxed check of "< 0" the only
valid values(ignoring cosmic ray induced bitflips) are -EINVAL, 0+.

Fixes: 40b1936efebd ("regulator: Introduce TI Adaptive Body Bias(ABB) on-chip LDO driver")
Link: https://lore.kernel.org/linux-mm/CA+G9fYuk4imvhyCN7D7T6PMDH6oNp6HDCRiTUKMQ6QXXjBa4ag@mail.gmail.com/
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20201118145009.10492-1-nm@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agoASOC: Intel: kbl_rt5663_rt5514_max98927: Do not try to disable disabled clock
Guenter Roeck [Wed, 11 Nov 2020 20:54:34 +0000 (12:54 -0800)]
ASOC: Intel: kbl_rt5663_rt5514_max98927: Do not try to disable disabled clock

In kabylake_set_bias_level(), enabling mclk may fail if the clock has
already been enabled by the firmware. Attempts to disable that clock
later will fail with a warning backtrace.

mclk already disabled
WARNING: CPU: 2 PID: 108 at drivers/clk/clk.c:952 clk_core_disable+0x1b6/0x1cf
...
Call Trace:
 clk_disable+0x2d/0x3a
 kabylake_set_bias_level+0x72/0xfd [snd_soc_kbl_rt5663_rt5514_max98927]
 snd_soc_card_set_bias_level+0x2b/0x6f
 snd_soc_dapm_set_bias_level+0xe1/0x209
 dapm_pre_sequence_async+0x63/0x96
 async_run_entry_fn+0x3d/0xd1
 process_one_work+0x2a9/0x526
...

Only disable the clock if it has been enabled.

Fixes: 15747a802075 ("ASoC: eve: implement set_bias_level function for rt5514")
Cc: Brent Lu <brent.lu@intel.com>
Cc: Curtis Malainey <cujomalainey@chromium.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20201111205434.207610-1-linux@roeck-us.net
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agoxfs: return corresponding errcode if xfs_initialize_perag() fail
Yu Kuai [Wed, 18 Nov 2020 17:21:26 +0000 (09:21 -0800)]
xfs: return corresponding errcode if xfs_initialize_perag() fail

In xfs_initialize_perag(), if kmem_zalloc(), xfs_buf_hash_init(), or
radix_tree_preload() failed, the returned value 'error' is not set
accordingly.

Reported-as-fixing: 8b26c5825e02 ("xfs: handle ENOMEM correctly during initialisation of perag structures")
Fixes: 9b2471797942 ("xfs: cache unlinked pointers in an rhashtable")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
3 years agoxfs: ensure inobt record walks always make forward progress
Darrick J. Wong [Sat, 14 Nov 2020 17:59:22 +0000 (09:59 -0800)]
xfs: ensure inobt record walks always make forward progress

The aim of the inode btree record iterator function is to call a
callback on every record in the btree.  To avoid having to tear down and
recreate the inode btree cursor around every callback, it caches a
certain number of records in a memory buffer.  After each batch of
callback invocations, we have to perform a btree lookup to find the
next record after where we left off.

However, if the keys of the inode btree are corrupt, the lookup might
put us in the wrong part of the inode btree, causing the walk function
to loop forever.  Therefore, we add extra cursor tracking to make sure
that we never go backwards neither when performing the lookup nor when
jumping to the next inobt record.  This also fixes an off by one error
where upon resume the lookup should have been for the inode /after/ the
point at which we stopped.

Found by fuzzing xfs/460 with keys[2].startino = ones causing bulkstat
and quotacheck to hang.

Fixes: a211432c27ff ("xfs: create simplified inode walk function")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
3 years agoxfs: fix forkoff miscalculation related to XFS_LITINO(mp)
Gao Xiang [Sat, 14 Nov 2020 19:06:01 +0000 (11:06 -0800)]
xfs: fix forkoff miscalculation related to XFS_LITINO(mp)

Currently, commit e9e2eae89ddb dropped a (int) decoration from
XFS_LITINO(mp), and since sizeof() expression is also involved,
the result of XFS_LITINO(mp) is simply as the size_t type
(commonly unsigned long).

Considering the expression in xfs_attr_shortform_bytesfit():
  offset = (XFS_LITINO(mp) - bytes) >> 3;
let "bytes" be (int)340, and
    "XFS_LITINO(mp)" be (unsigned long)336.

on 64-bit platform, the expression is
  offset = ((unsigned long)336 - (int)340) >> 3 =
           (int)(0xfffffffffffffffcUL >> 3) = -1

but on 32-bit platform, the expression is
  offset = ((unsigned long)336 - (int)340) >> 3 =
           (int)(0xfffffffcUL >> 3) = 0x1fffffff
instead.

so offset becomes a large positive number on 32-bit platform, and
cause xfs_attr_shortform_bytesfit() returns maxforkoff rather than 0.

Therefore, one result is
  "ASSERT(new_size <= XFS_IFORK_SIZE(ip, whichfork));"

assertion failure in xfs_idata_realloc(), which was also the root
cause of the original bugreport from Dennis, see:
   https://bugzilla.redhat.com/show_bug.cgi?id=1894177

And it can also be manually triggered with the following commands:
  $ touch a;
  $ setfattr -n user.0 -v "`seq 0 80`" a;
  $ setfattr -n user.1 -v "`seq 0 80`" a

on 32-bit platform.

Fix the case in xfs_attr_shortform_bytesfit() by bailing out
"XFS_LITINO(mp) < bytes" in advance suggested by Eric and a misleading
comment together with this bugfix suggested by Darrick. It seems the
other users of XFS_LITINO(mp) are not impacted.

Fixes: e9e2eae89ddb ("xfs: only check the superblock version for dinode size calculation")
Cc: <stable@vger.kernel.org> # 5.7+
Reported-and-tested-by: Dennis Gilmore <dgilmore@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
3 years agoxfs: directory scrub should check the null bestfree entries too
Darrick J. Wong [Mon, 9 Nov 2020 00:32:42 +0000 (16:32 -0800)]
xfs: directory scrub should check the null bestfree entries too

Teach the directory scrubber to check all the bestfree entries,
including the null ones.  We want to be able to detect the case where
the entry is null but there actually /is/ a directory data block.

Found by fuzzing lbests[0] = ones in xfs/391.

Fixes: df481968f33b ("xfs: scrub directory freespace")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
3 years agoxfs: strengthen rmap record flags checking
Darrick J. Wong [Mon, 9 Nov 2020 00:32:41 +0000 (16:32 -0800)]
xfs: strengthen rmap record flags checking

We always know the correct state of the rmap record flags (attr, bmbt,
unwritten) so check them by direct comparison.

Fixes: d852657ccfc0 ("xfs: cross-reference reverse-mapping btree")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
3 years agoxfs: fix the minrecs logic when dealing with inode root child blocks
Darrick J. Wong [Mon, 9 Nov 2020 00:32:41 +0000 (16:32 -0800)]
xfs: fix the minrecs logic when dealing with inode root child blocks

The comment and logic in xchk_btree_check_minrecs for dealing with
inode-rooted btrees isn't quite correct.  While the direct children of
the inode root are allowed to have fewer records than what would
normally be allowed for a regular ondisk btree block, this is only true
if there is only one child block and the number of records don't fit in
the inode root.

Fixes: 08a3a692ef58 ("xfs: btree scrub should check minrecs")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
3 years agocan: m_can: process interrupt only when not runtime suspended
Jarkko Nikula [Tue, 15 Sep 2020 13:47:15 +0000 (16:47 +0300)]
can: m_can: process interrupt only when not runtime suspended

Avoid processing bogus interrupt statuses when the HW is runtime suspended and
the M_CAN_IR register read may get all bits 1's. Handler can be called if the
interrupt request is shared with other peripherals or at the end of free_irq().

Therefore check the runtime suspended status before processing.

Fixes: cdf8259d6573 ("can: m_can: Add PM Support")
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Link: https://lore.kernel.org/r/20200915134715.696303-1-jarkko.nikula@linux.intel.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agogfs2: Fix regression in freeze_go_sync
Bob Peterson [Wed, 18 Nov 2020 13:54:31 +0000 (08:54 -0500)]
gfs2: Fix regression in freeze_go_sync

Patch 541656d3a513 ("gfs2: freeze should work on read-only mounts") changed
the check for glock state in function freeze_go_sync() from "gl->gl_state
== LM_ST_SHARED" to "gl->gl_req == LM_ST_EXCLUSIVE".  That's wrong and it
regressed gfs2's freeze/thaw mechanism because it caused only the freezing
node (which requests the glock in EX) to queue freeze work.

All nodes go through this go_sync code path during the freeze to drop their
SHared hold on the freeze glock, allowing the freezing node to acquire it
in EXclusive mode. But all the nodes must freeze access to the file system
locally, so they ALL must queue freeze work. The freeze_work calls
freeze_func, which makes a request to reacquire the freeze glock in SH,
effectively blocking until the thaw from the EX holder. Once thawed, the
freezing node drops its EX hold on the freeze glock, then the (blocked)
freeze_func reacquires the freeze glock in SH again (on all nodes, including
the freezer) so all nodes go back to a thawed state.

This patch changes the check back to gl_state == LM_ST_SHARED like it was
prior to 541656d3a513.

Fixes: 541656d3a513 ("gfs2: freeze should work on read-only mounts")
Cc: stable@vger.kernel.org # v5.8+
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
3 years agocan: flexcan: flexcan_chip_start(): fix erroneous flexcan_transceiver_enable() during...
Marc Kleine-Budde [Wed, 18 Nov 2020 15:01:48 +0000 (16:01 +0100)]
can: flexcan: flexcan_chip_start(): fix erroneous flexcan_transceiver_enable() during bus-off recovery

If the CAN controller goes into bus off, the do_set_mode() callback with
CAN_MODE_START can be used to recover the controller, which then calls
flexcan_chip_start(). If configured, this is done automatically by the
framework or manually by the user.

In flexcan_chip_start() there is an explicit call to
flexcan_transceiver_enable(), which does a regulator_enable() on the
transceiver regulator. This results in a net usage counter increase, as there
is no corresponding flexcan_transceiver_disable() in the bus off code path.
This further leads to the transceiver stuck enabled, even if the CAN interface
is shut down.

To fix this problem the
flexcan_transceiver_enable()/flexcan_transceiver_disable() are moved out of
flexcan_chip_start()/flexcan_chip_stop() into flexcan_open()/flexcan_close().

Fixes: e955cead0311 ("CAN: Add Flexcan CAN controller driver")
Link: https://lore.kernel.org/r/20201118150148.2664024-1-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agoio_uring: order refnode recycling
Pavel Begunkov [Wed, 18 Nov 2020 14:56:26 +0000 (14:56 +0000)]
io_uring: order refnode recycling

Don't recycle a refnode until we're done with all requests of nodes
ejected before.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Cc: stable@vger.kernel.org # v5.7+
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 years agoio_uring: get an active ref_node from files_data
Pavel Begunkov [Wed, 18 Nov 2020 14:56:25 +0000 (14:56 +0000)]
io_uring: get an active ref_node from files_data

An active ref_node always can be found in ctx->files_data, it's much
safer to get it this way instead of poking into files_data->ref_list.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Cc: stable@vger.kernel.org # v5.7+
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 years agoiommu/vt-d: Avoid panic if iommu init fails in tboot system
Zhenzhong Duan [Tue, 10 Nov 2020 07:19:08 +0000 (15:19 +0800)]
iommu/vt-d: Avoid panic if iommu init fails in tboot system

"intel_iommu=off" command line is used to disable iommu but iommu is force
enabled in a tboot system for security reason.

However for better performance on high speed network device, a new option
"intel_iommu=tboot_noforce" is introduced to disable the force on.

By default kernel should panic if iommu init fail in tboot for security
reason, but it's unnecessory if we use "intel_iommu=tboot_noforce,off".

Fix the code setting force_on and move intel_iommu_tboot_noforce
from tboot code to intel iommu code.

Fixes: 7304e8f28bb2 ("iommu/vt-d: Correctly disable Intel IOMMU force on")
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@gmail.com>
Tested-by: Lukasz Hawrylko <lukasz.hawrylko@linux.intel.com>
Acked-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20201110071908.3133-1-zhenzhong.duan@gmail.com
Signed-off-by: Will Deacon <will@kernel.org>
3 years agodmaengine: fix error codes in channel_register()
Dan Carpenter [Fri, 13 Nov 2020 10:16:31 +0000 (13:16 +0300)]
dmaengine: fix error codes in channel_register()

The error codes were not set on some of these error paths.

Also the error handling was more confusing than it needed to be so I
cleaned it up and shuffled it around a bit.

Fixes: d2fb0a043838 ("dmaengine: break out channel registration")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20201113101631.GE168908@mwanda
Signed-off-by: Vinod Koul <vkoul@kernel.org>
3 years agox86/dumpstack: Do not try to access user space code of other tasks
Thomas Gleixner [Tue, 17 Nov 2020 20:23:34 +0000 (21:23 +0100)]
x86/dumpstack: Do not try to access user space code of other tasks

sysrq-t ends up invoking show_opcodes() for each task which tries to access
the user space code of other processes, which is obviously bogus.

It either manages to dump where the foreign task's regs->ip points to in a
valid mapping of the current task or triggers a pagefault and prints "Code:
Bad RIP value.". Both is just wrong.

Add a safeguard in copy_code() and check whether the @regs pointer matches
currents pt_regs. If not, do not even try to access it.

While at it, add commentary why using copy_from_user_nmi() is safe in
copy_code() even if the function name suggests otherwise.

Reported-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Tested-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20201117202753.667274723@linutronix.de
3 years agocan: kvaser_usb: kvaser_usb_hydra: Fix KCAN bittiming limits
Jimmy Assarsson [Sun, 15 Nov 2020 16:30:23 +0000 (17:30 +0100)]
can: kvaser_usb: kvaser_usb_hydra: Fix KCAN bittiming limits

Use correct bittiming limits for the KCAN CAN controller.

Fixes: aec5fb2268b7 ("can: kvaser_usb: Add support for Kvaser USB hydra family")
Signed-off-by: Jimmy Assarsson <extja@kvaser.com>
Link: https://lore.kernel.org/r/20201115163027.16851-2-jimmyassarsson@gmail.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agocan: kvaser_pciefd: Fix KCAN bittiming limits
Jimmy Assarsson [Sun, 15 Nov 2020 16:30:22 +0000 (17:30 +0100)]
can: kvaser_pciefd: Fix KCAN bittiming limits

Use correct bittiming limits for the KCAN CAN controller.

Fixes: 26ad340e582d ("can: kvaser_pciefd: Add driver for Kvaser PCIEcan devices")
Signed-off-by: Jimmy Assarsson <extja@kvaser.com>
Link: https://lore.kernel.org/r/20201115163027.16851-1-jimmyassarsson@gmail.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agodrm/sun4i: backend: Fix probe failure with multiple backends
Maxime Ripard [Fri, 6 Nov 2020 14:22:31 +0000 (15:22 +0100)]
drm/sun4i: backend: Fix probe failure with multiple backends

Commit e0d072782c73 ("dma-mapping: introduce DMA range map, supplanting
dma_pfn_offset") introduced a regression in our code since the second
backed to probe will now get -EINVAL back from dma_direct_set_offset and
will prevent the entire DRM device from probing.

Ignore -EINVAL as a temporary measure to get it back working, before
removing that call entirely.

Fixes: e0d072782c73 ("dma-mapping: introduce DMA range map, supplanting dma_pfn_offset")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
3 years agoipv4: use IS_ENABLED instead of ifdef
Florian Klink [Sun, 15 Nov 2020 22:45:09 +0000 (23:45 +0100)]
ipv4: use IS_ENABLED instead of ifdef

Checking for ifdef CONFIG_x fails if CONFIG_x=m.

Use IS_ENABLED instead, which is true for both built-ins and modules.

Otherwise, a
> ip -4 route add 1.2.3.4/32 via inet6 fe80::2 dev eth1
fails with the message "Error: IPv6 support not enabled in kernel." if
CONFIG_IPV6 is `m`.

In the spirit of b8127113d01e53adba15b41aefd37b90ed83d631.

Fixes: d15662682db2 ("ipv4: Allow ipv6 gateway with ipv4 routes")
Cc: Kim Phillips <kim.phillips@arm.com>
Signed-off-by: Florian Klink <flokli@flokli.de>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20201115224509.2020651-1-flokli@flokli.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoqed: fix ILT configuration of SRC block
Dmitry Bogdanov [Mon, 16 Nov 2020 13:29:44 +0000 (16:29 +0300)]
qed: fix ILT configuration of SRC block

The code refactoring of ILT configuration was not complete, the old
unused variables were used for the SRC block. That could lead to the memory
corruption by HW when rx filters are configured.
This patch completes that refactoring.

Fixes: 8a52bbab39c9 (qed: Debug feature: ilt and mdump)
Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: Dmitry Bogdanov <dbogdanov@marvell.com>
Link: https://lore.kernel.org/r/20201116132944.2055-1-dbogdanov@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoinet_diag: Fix error path to cancel the meseage in inet_req_diag_fill()
Wang Hai [Mon, 16 Nov 2020 08:20:18 +0000 (16:20 +0800)]
inet_diag: Fix error path to cancel the meseage in inet_req_diag_fill()

nlmsg_cancel() needs to be called in the error path of
inet_req_diag_fill to cancel the message.

Fixes: d545caca827b ("net: inet: diag: expose the socket mark to privileged processes.")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Link: https://lore.kernel.org/r/20201116082018.16496-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agotools/testing/scatterlist: Fix test to compile and run
Maor Gottlieb [Sun, 15 Nov 2020 12:06:23 +0000 (14:06 +0200)]
tools/testing/scatterlist: Fix test to compile and run

Add missing define of ALIGN_DOWN to make the test build and run.  In
addition, __sg_alloc_table_from_pages now support unaligned maximum
segment, so adapt the test result accordingly.

Fixes: 07da1223ec93 ("lib/scatterlist: Add support in dynamic allocation of SG table from pages")
Link: https://lore.kernel.org/r/20201115120623.139113-1-leon@kernel.org
Signed-off-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
3 years agobpf, sockmap: Avoid failures from skb_to_sgvec when skb has frag_list
John Fastabend [Mon, 16 Nov 2020 22:29:28 +0000 (14:29 -0800)]
bpf, sockmap: Avoid failures from skb_to_sgvec when skb has frag_list

When skb has a frag_list its possible for skb_to_sgvec() to fail. This
happens when the scatterlist has fewer elements to store pages than would
be needed for the initial skb plus any of its frags.

This case appears rare, but is possible when running an RX parser/verdict
programs exposed to the internet. Currently, when this happens we throw
an error, break the pipe, and kfree the msg. This effectively breaks the
application or forces it to do a retry.

Lets catch this case and handle it by doing an skb_linearize() on any
skb we receive with frags. At this point skb_to_sgvec should not fail
because the failing conditions would require frags to be in place.

Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556576837.73229.14800682790808797635.stgit@john-XPS-13-9370
3 years agobpf, sockmap: Handle memory acct if skb_verdict prog redirects to self
John Fastabend [Mon, 16 Nov 2020 22:29:08 +0000 (14:29 -0800)]
bpf, sockmap: Handle memory acct if skb_verdict prog redirects to self

If the skb_verdict_prog redirects an skb knowingly to itself, fix your
BPF program this is not optimal and an abuse of the API please use
SK_PASS. That said there may be cases, such as socket load balancing,
where picking the socket is hashed based or otherwise picks the same
socket it was received on in some rare cases. If this happens we don't
want to confuse userspace giving them an EAGAIN error if we can avoid
it.

To avoid double accounting in these cases. At the moment even if the
skb has already been charged against the sockets rcvbuf and forward
alloc we check it again and do set_owner_r() causing it to be orphaned
and recharged. For one this is useless work, but more importantly we
can have a case where the skb could be put on the ingress queue, but
because we are under memory pressure we return EAGAIN. The trouble
here is the skb has already been accounted for so any rcvbuf checks
include the memory associated with the packet already. This rolls
up and can result in unnecessary EAGAIN errors in userspace read()
calls.

Fix by doing an unlikely check and skipping checks if skb->sk == sk.

Fixes: 51199405f9672 ("bpf: skb_verdict, support SK_PASS on RX BPF path")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556574804.73229.11328201020039674147.stgit@john-XPS-13-9370
3 years agobpf, sockmap: Avoid returning unneeded EAGAIN when redirecting to self
John Fastabend [Mon, 16 Nov 2020 22:28:46 +0000 (14:28 -0800)]
bpf, sockmap: Avoid returning unneeded EAGAIN when redirecting to self

If a socket redirects to itself and it is under memory pressure it is
possible to get a socket stuck so that recv() returns EAGAIN and the
socket can not advance for some time. This happens because when
redirecting a skb to the same socket we received the skb on we first
check if it is OK to enqueue the skb on the receiving socket by checking
memory limits. But, if the skb is itself the object holding the memory
needed to enqueue the skb we will keep retrying from kernel side
and always fail with EAGAIN. Then userspace will get a recv() EAGAIN
error if there are no skbs in the psock ingress queue. This will continue
until either some skbs get kfree'd causing the memory pressure to
reduce far enough that we can enqueue the pending packet or the
socket is destroyed. In some cases its possible to get a socket
stuck for a noticeable amount of time if the socket is only receiving
skbs from sk_skb verdict programs. To reproduce I make the socket
memory limits ridiculously low so sockets are always under memory
pressure. More often though if under memory pressure it looks like
a spurious EAGAIN error on user space side causing userspace to retry
and typically enough has moved on the memory side that it works.

To fix skip memory checks and skb_orphan if receiving on the same
sock as already assigned.

For SK_PASS cases this is easy, its always the same socket so we
can just omit the orphan/set_owner pair.

For backlog cases we need to check skb->sk and decide if the orphan
and set_owner pair are needed.

Fixes: 51199405f9672 ("bpf: skb_verdict, support SK_PASS on RX BPF path")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556572660.73229.12566203819812939627.stgit@john-XPS-13-9370
3 years agobpf, sockmap: Use truesize with sk_rmem_schedule()
John Fastabend [Mon, 16 Nov 2020 22:28:26 +0000 (14:28 -0800)]
bpf, sockmap: Use truesize with sk_rmem_schedule()

We use skb->size with sk_rmem_scheduled() which is not correct. Instead
use truesize to align with socket and tcp stack usage of sk_rmem_schedule.

Suggested-by: Daniel Borkman <daniel@iogearbox.net>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556570616.73229.17003722112077507863.stgit@john-XPS-13-9370
3 years agobpf, sockmap: Ensure SO_RCVBUF memory is observed on ingress redirect
John Fastabend [Mon, 16 Nov 2020 22:28:06 +0000 (14:28 -0800)]
bpf, sockmap: Ensure SO_RCVBUF memory is observed on ingress redirect

Fix sockmap sk_skb programs so that they observe sk_rcvbuf limits. This
allows users to tune SO_RCVBUF and sockmap will honor them.

We can refactor the if(charge) case out in later patches. But, keep this
fix to the point.

Fixes: 51199405f9672 ("bpf: skb_verdict, support SK_PASS on RX BPF path")
Suggested-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556568657.73229.8404601585878439060.stgit@john-XPS-13-9370
3 years agobpf, sockmap: Fix partial copy_page_to_iter so progress can still be made
John Fastabend [Mon, 16 Nov 2020 22:27:46 +0000 (14:27 -0800)]
bpf, sockmap: Fix partial copy_page_to_iter so progress can still be made

If copy_page_to_iter() fails or even partially completes, but with fewer
bytes copied than expected we currently reset sg.start and return EFAULT.
This proves problematic if we already copied data into the user buffer
before we return an error. Because we leave the copied data in the user
buffer and fail to unwind the scatterlist so kernel side believes data
has been copied and user side believes data has _not_ been received.

Expected behavior should be to return number of bytes copied and then
on the next read we need to return the error assuming its still there. This
can happen if we have a copy length spanning multiple scatterlist elements
and one or more complete before the error is hit.

The error is rare enough though that my normal testing with server side
programs, such as nginx, httpd, envoy, etc., I have never seen this. The
only reliable way to reproduce that I've found is to stream movies over
my browser for a day or so and wait for it to hang. Not very scientific,
but with a few extra WARN_ON()s in the code the bug was obvious.

When we review the errors from copy_page_to_iter() it seems we are hitting
a page fault from copy_page_to_iter_iovec() where the code checks
fault_in_pages_writeable(buf, copy) where buf is the user buffer. It
also seems typical server applications don't hit this case.

The other way to try and reproduce this is run the sockmap selftest tool
test_sockmap with data verification enabled, but it doesn't reproduce the
fault. Perhaps we can trigger this case artificially somehow from the
test tools. I haven't sorted out a way to do that yet though.

Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556566659.73229.15694973114605301063.stgit@john-XPS-13-9370
3 years agonet/tls: Fix wrong record sn in async mode of device resync
Tariq Toukan [Sun, 15 Nov 2020 13:14:48 +0000 (15:14 +0200)]
net/tls: Fix wrong record sn in async mode of device resync

In async_resync mode, we log the TCP seq of records until the async request
is completed.  Later, in case one of the logged seqs matches the resync
request, we return it, together with its record serial number.  Before this
fix, we mistakenly returned the serial number of the current record
instead.

Fixes: ed9b7646b06a ("net/tls: Add asynchronous resync")
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Boris Pismenny <borisp@nvidia.com>
Link: https://lore.kernel.org/r/20201115131448.2702-1-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoio_uring: don't double complete failed reissue request
Jens Axboe [Tue, 17 Nov 2020 14:59:16 +0000 (07:59 -0700)]
io_uring: don't double complete failed reissue request

Zorro reports that an xfstest test case is failing, and it turns out that
for the reissue path we can potentially issue a double completion on the
request for the failure path. There's an issue around the retry as well,
but for now, at least just make sure that we handle the error path
correctly.

Cc: stable@vger.kernel.org
Fixes: b63534c41e20 ("io_uring: re-issue block requests that failed because of resources")
Reported-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 years agonetdevsim: set .owner to THIS_MODULE
Taehee Yoo [Sun, 15 Nov 2020 10:30:41 +0000 (10:30 +0000)]
netdevsim: set .owner to THIS_MODULE

If THIS_MODULE is not set, the module would be removed while debugfs is
being used.
It eventually makes kernel panic.

Fixes: 82c93a87bf8b ("netdevsim: implement couple of testing devlink health reporters")
Fixes: 424be63ad831 ("netdevsim: add UDP tunnel port offload support")
Fixes: 4418f862d675 ("netdevsim: implement support for devlink region and snapshots")
Fixes: d3cbb907ae57 ("netdevsim: add ACL trap reporting cookie as a metadata")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Link: https://lore.kernel.org/r/20201115103041.30701-1-ap420073@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoseccomp: Set PF_SUPERPRIV when checking capability
Mickaël Salaün [Fri, 30 Oct 2020 12:38:49 +0000 (13:38 +0100)]
seccomp: Set PF_SUPERPRIV when checking capability

Replace the use of security_capable(current_cred(), ...) with
ns_capable_noaudit() which set PF_SUPERPRIV.

Since commit 98f368e9e263 ("kernel: Add noaudit variant of
ns_capable()"), a new ns_capable_noaudit() helper is available.  Let's
use it!

Cc: Jann Horn <jannh@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Tyler Hicks <tyhicks@linux.microsoft.com>
Cc: Will Drewry <wad@chromium.org>
Cc: stable@vger.kernel.org
Fixes: e2cfabdfd075 ("seccomp: add system call filtering using BPF")
Signed-off-by: Mickaël Salaün <mic@linux.microsoft.com>
Reviewed-by: Jann Horn <jannh@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20201030123849.770769-3-mic@digikod.net
3 years agoptrace: Set PF_SUPERPRIV when checking capability
Mickaël Salaün [Fri, 30 Oct 2020 12:38:48 +0000 (13:38 +0100)]
ptrace: Set PF_SUPERPRIV when checking capability

Commit 69f594a38967 ("ptrace: do not audit capability check when outputing
/proc/pid/stat") replaced the use of ns_capable() with
has_ns_capability{,_noaudit}() which doesn't set PF_SUPERPRIV.

Commit 6b3ad6649a4c ("ptrace: reintroduce usage of subjective credentials in
ptrace_has_cap()") replaced has_ns_capability{,_noaudit}() with
security_capable(), which doesn't set PF_SUPERPRIV neither.

Since commit 98f368e9e263 ("kernel: Add noaudit variant of ns_capable()"), a
new ns_capable_noaudit() helper is available.  Let's use it!

As a result, the signature of ptrace_has_cap() is restored to its original one.

Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Eric Paris <eparis@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Serge E. Hallyn <serge@hallyn.com>
Cc: Tyler Hicks <tyhicks@linux.microsoft.com>
Cc: stable@vger.kernel.org
Fixes: 6b3ad6649a4c ("ptrace: reintroduce usage of subjective credentials in ptrace_has_cap()")
Fixes: 69f594a38967 ("ptrace: do not audit capability check when outputing /proc/pid/stat")
Signed-off-by: Mickaël Salaün <mic@linux.microsoft.com>
Reviewed-by: Jann Horn <jannh@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20201030123849.770769-2-mic@digikod.net
3 years agoenetc: Workaround for MDIO register access issue
Alex Marginean [Thu, 12 Nov 2020 18:26:08 +0000 (20:26 +0200)]
enetc: Workaround for MDIO register access issue

Due to a hardware issue, an access to MDIO registers
that is concurrent with other ENETC register accesses
may lead to the MDIO access being dropped or corrupted.
The workaround introduces locking for all register accesses
to the ENETC register space.  To reduce performance impact,
a readers-writers locking scheme has been implemented.
The writer in this case is the MDIO access code (irrelevant
whether that MDIO access is a register read or write), and
the reader is any access code to non-MDIO ENETC registers.
Also, the datapath functions acquire the read lock fewer times
and use _hot accessors.  All the rest of the code uses the _wa
accessors which lock every register access.
The commit introducing MDIO support is -
commit ebfcb23d62ab ("enetc: Add ENETC PF level external MDIO support")
but due to subsequent refactoring this patch is applicable on
top of a later commit.

Fixes: 6517798dd343 ("enetc: Make MDIO accessors more generic and export to include/linux/fsl")
Signed-off-by: Alex Marginean <alexandru.marginean@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Link: https://lore.kernel.org/r/20201112182608.26177-1-claudiu.manoil@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Tue, 17 Nov 2020 20:00:40 +0000 (12:00 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

Pull input fixes from Dmitry Torokhov:
 "A fix for use-after-free in the Sun keyboard driver, a fix to firmware
  updates on newer ICs in the Elan touchpad diver, and a couple misc
  driver fixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: elan_i2c - fix firmware update on newer ICs
  Input: resistive-adc-touch - fix kconfig dependency on IIO_BUFFER
  Input: sunkbd - avoid use-after-free in teardown paths
  Input: i8042 - allow insmod to succeed on devices without an i8042 controller
  Input: adxl34x - clean up a data type in adxl34x_probe()

3 years agonet/mlx5: fix error return code in mlx5e_tc_nic_init()
Wang Hai [Sat, 14 Nov 2020 11:52:23 +0000 (19:52 +0800)]
net/mlx5: fix error return code in mlx5e_tc_nic_init()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: aedd133d17bc ("net/mlx5e: Support CT offload for tc nic flows")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: E-Switch, Fail mlx5_esw_modify_vport_rate if qos disabled
Eli Cohen [Mon, 9 Nov 2020 09:35:52 +0000 (11:35 +0200)]
net/mlx5: E-Switch, Fail mlx5_esw_modify_vport_rate if qos disabled

Avoid calling mlx5_esw_modify_vport_rate() if qos is not enabled and
avoid unnecessary syndrome messages from firmware.

Fixes: fcb64c0f5640 ("net/mlx5: E-Switch, add ingress rate support")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Disable QoS when min_rates on all VFs are zero
Vladyslav Tarasiuk [Wed, 21 Oct 2020 08:05:41 +0000 (11:05 +0300)]
net/mlx5: Disable QoS when min_rates on all VFs are zero

Currently when QoS is enabled for VF and any min_rate is configured,
the driver sets bw_share value to at least 1 and doesn’t allow to set
it to 0 to make minimal rate unlimited. It means there is always a
minimal rate configured for every VF, even if user tries to remove it.

In order to make QoS disable possible, check whether all vports have
configured min_rate = 0. If this is true, set their bw_share to 0 to
disable min_rate limitations.

Fixes: c9497c98901c ("net/mlx5: Add support for setting VF min rate")
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Clear bw_share upon VF disable
Vladyslav Tarasiuk [Mon, 2 Nov 2020 11:45:24 +0000 (13:45 +0200)]
net/mlx5: Clear bw_share upon VF disable

Currently, if user disables VFs with some min and max rates configured,
they are cleared. But QoS data is not cleared and restored upon next VF
enable placing limits on minimal rate for given VF, when user expects
none.

To match cleared vport->info struct with QoS-related min and max rates
upon VF disable, clear vport->qos struct too.

Fixes: 556b9d16d3f5 ("net/mlx5: Clear VF's configuration on disabling SRIOV")
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Add handling of port type in rule deletion
Michael Guralnik [Mon, 2 Nov 2020 15:34:44 +0000 (17:34 +0200)]
net/mlx5: Add handling of port type in rule deletion

Handle destruction of rules with port destination type to enable
full destruction of flow.

Without this handling of TX rules the deletion of these rules fails.
Dmesg of flow destruction failure:

[  203.714146] mlx5_core 0000:00:0b.0: mlx5_cmd_check:753:(pid 342): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x144b7a)
[  210.547387] ------------[ cut here ]------------
[  210.548663] refcount_t: decrement hit 0; leaking memory.
[  210.550651] WARNING: CPU: 4 PID: 342 at lib/refcount.c:31 refcount_warn_saturate+0x5c/0x110
[  210.550654] Modules linked in: mlx5_ib mlx5_core ib_ipoib rdma_ucm rdma_cm iw_cm ib_cm ib_umad ib_uverbs ib_core
[  210.550675] CPU: 4 PID: 342 Comm: test Not tainted 5.8.0-rc2+ #116
[  210.550678] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
[  210.550680] RIP: 0010:refcount_warn_saturate+0x5c/0x110
[  210.550685] Code: c6 d1 1b 01 00 0f 84 ad 00 00 00 5b 5d c3 80 3d b5 d1 1b 01 00 75 f4 48 c7 c7 20 d1 15 82 c6 05 a5 d1 1b 01 01 e8 a7 eb af ff <0f> 0b eb dd 80 3d 99 d1 1b 01 00 75 d4 48 c7 c7 c0 cf 15 82 c6 05
[  210.550687] RSP: 0018:ffff8881642e77e8 EFLAGS: 00010282
[  210.550691] RAX: 0000000000000000 RBX: 0000000000000004 RCX: 0000000000000000
[  210.550694] RDX: 0000000000000027 RSI: 0000000000000004 RDI: ffffed102c85ceef
[  210.550696] RBP: ffff888161720428 R08: ffffffff8124c10e R09: ffffed103243beae
[  210.550698] R10: ffff8881921df56b R11: ffffed103243bead R12: ffff8881841b4180
[  210.550701] R13: ffff888161720428 R14: ffff8881616d0000 R15: ffff888161720380
[  210.550704] FS:  00007fc27f025740(0000) GS:ffff888192000000(0000) knlGS:0000000000000000
[  210.550706] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  210.550708] CR2: 0000557e4b41a6a0 CR3: 0000000002415004 CR4: 0000000000360ea0
[  210.550711] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  210.550713] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  210.550715] Call Trace:
[  210.550717]  mlx5_del_flow_rules+0x484/0x490 [mlx5_core]
[  210.550720]  ? mlx5_cmd_set_fte+0xa80/0xa80 [mlx5_core]
[  210.550722]  mlx5_ib_destroy_flow+0x17f/0x280 [mlx5_ib]
[  210.550724]  uverbs_free_flow+0x4c/0x90 [ib_uverbs]
[  210.550726]  destroy_hw_idr_uobject+0x41/0xb0 [ib_uverbs]
[  210.550728]  uverbs_destroy_uobject+0xaa/0x390 [ib_uverbs]
[  210.550731]  __uverbs_cleanup_ufile+0x129/0x1b0 [ib_uverbs]
[  210.550733]  ? uverbs_destroy_uobject+0x390/0x390 [ib_uverbs]
[  210.550735]  uverbs_destroy_ufile_hw+0x78/0x190 [ib_uverbs]
[  210.550737]  ib_uverbs_close+0x36/0x140 [ib_uverbs]
[  210.550739]  __fput+0x181/0x380
[  210.550741]  task_work_run+0x88/0xd0
[  210.550743]  do_exit+0x5f6/0x13b0
[  210.550745]  ? sched_clock_cpu+0x30/0x140
[  210.550747]  ? is_current_pgrp_orphaned+0x70/0x70
[  210.550750]  ? lock_downgrade+0x360/0x360
[  210.550752]  ? mark_held_locks+0x1d/0x90
[  210.550754]  do_group_exit+0x8a/0x140
[  210.550756]  get_signal+0x20a/0xf50
[  210.550758]  do_signal+0x8c/0xbe0
[  210.550760]  ? hrtimer_nanosleep+0x1d8/0x200
[  210.550762]  ? nanosleep_copyout+0x50/0x50
[  210.550764]  ? restore_sigcontext+0x320/0x320
[  210.550766]  ? __hrtimer_init+0xf0/0xf0
[  210.550768]  ? timespec64_add_safe+0x150/0x150
[  210.550770]  ? mark_held_locks+0x1d/0x90
[  210.550772]  ? lockdep_hardirqs_on_prepare+0x14c/0x240
[  210.550774]  __prepare_exit_to_usermode+0x119/0x170
[  210.550776]  do_syscall_64+0x65/0x300
[  210.550778]  ? trace_hardirqs_off+0x10/0x120
[  210.550781]  ? mark_held_locks+0x1d/0x90
[  210.550783]  ? asm_sysvec_apic_timer_interrupt+0xa/0x20
[  210.550785]  ? lockdep_hardirqs_on+0x112/0x190
[  210.550787]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  210.550789] RIP: 0033:0x7fc27f1cd157
[  210.550791] Code: Bad RIP value.
[  210.550793] RSP: 002b:00007ffd4db27ea8 EFLAGS: 00000246 ORIG_RAX: 0000000000000023
[  210.550798] RAX: fffffffffffffdfc RBX: ffffffffffffff80 RCX: 00007fc27f1cd157
[  210.550800] RDX: 00007fc27f025740 RSI: 00007ffd4db27eb0 RDI: 00007ffd4db27eb0
[  210.550803] RBP: 0000000000000016 R08: 0000000000000000 R09: 000000000000000e
[  210.550805] R10: 00007ffd4db27dc7 R11: 0000000000000246 R12: 0000000000400c00
[  210.550808] R13: 00007ffd4db285f0 R14: 0000000000000000 R15: 0000000000000000
[  210.550809] irq event stamp: 49399
[  210.550812] hardirqs last  enabled at (49399): [<ffffffff81172d36>] console_unlock+0x556/0x6f0
[  210.550815] hardirqs last disabled at (49398): [<ffffffff81172897>] console_unlock+0xb7/0x6f0
[  210.550818] softirqs last  enabled at (48706): [<ffffffff81e0037b>] __do_softirq+0x37b/0x60c
[  210.550820] softirqs last disabled at (48697): [<ffffffff81c00e2f>] asm_call_on_stack+0xf/0x20
[  210.550822] ---[ end trace ad18c0e6fa846454 ]---
[  210.581862] mlx5_core 0000:00:0c.0: mlx5_destroy_flow_table:2132:(pid 342): Flow table 262150 wasn't destroyed, refcount > 1

Fixes: a7ee18bdee83 ("RDMA/mlx5: Allow creating a matcher for a NIC TX flow table")
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix check if netdev is bond slave
Maor Dickman [Wed, 4 Nov 2020 12:10:30 +0000 (14:10 +0200)]
net/mlx5e: Fix check if netdev is bond slave

Bond events handler uses bond_slave_get_rtnl to check if net device
is bond slave. bond_slave_get_rtnl return the rcu rx_handler pointer
from the netdev which exists for bond slaves but also exists for
devices that are attached to linux bridge so using it as indication
for bond slave is wrong.

Fix by using netif_is_lag_port instead.

Fixes: 7e51891a237f ("net/mlx5e: Use netdev events to set/del egress acl forward-to-vport rule")
Signed-off-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix IPsec packet drop by mlx5e_tc_update_skb
Huy Nguyen [Tue, 3 Nov 2020 18:56:18 +0000 (12:56 -0600)]
net/mlx5e: Fix IPsec packet drop by mlx5e_tc_update_skb

Both TC and IPsec crypto offload use metadata_regB to store
private information. Since TC does not use bit 31 of regB, IPsec
will use bit 31 as the IPsec packet marker. The IPsec's regB usage
is changed to:
Bit31: IPsec marker
Bit30-24: IPsec syndrome
Bit23-0: IPsec obj id

Fixes: b2ac7541e377 ("net/mlx5e: IPsec: Add Connect-X IPsec Rx data path offload")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Set IPsec WAs only in IP's non checksum partial case.
Huy Nguyen [Thu, 22 Oct 2020 16:22:56 +0000 (11:22 -0500)]
net/mlx5e: Set IPsec WAs only in IP's non checksum partial case.

The IP's checksum partial still requires L4 csum flag on Ethernet WQE.
Make the IPsec WAs only for the IP's non checksum partial case
(for example icmd packet)

Fixes: 5be019040cb7 ("net/mlx5e: IPsec: Add Connect-X IPsec Tx data path offload")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Alaa Hleihel <alaa@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix refcount leak on kTLS RX resync
Maxim Mikityanskiy [Thu, 8 Oct 2020 09:34:10 +0000 (12:34 +0300)]
net/mlx5e: Fix refcount leak on kTLS RX resync

On resync, the driver calls inet_lookup_established
(__inet6_lookup_established) that increases sk_refcnt of the socket. To
decrease it, the driver set skb->destructor to sock_edemux. However, it
didn't work well, because the TCP stack also sets this destructor for
early demux, and the refcount gets decreased only once, while increased
two times (in mlx5e and in the TCP stack). It leads to a socket leak, a
TLS context leak, which in the end leads to calling tls_dev_del twice:
on socket close and on driver unload, which in turn leads to a crash.

This commit fixes the refcount leak by calling sock_gen_put right away
after using the socket, thus fixing all the subsequent issues.

Fixes: 0419d8c9d8f8 ("net/mlx5e: kTLS, Add kTLS RX resync support")
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agoMerge tag 's390-5.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Tue, 17 Nov 2020 19:22:03 +0000 (11:22 -0800)]
Merge tag 's390-5.10-4' of git://git./linux/kernel/git/s390/linux

Pull s390 fixes from Heiko Carstens:

 - fix system call exit path; avoid return to user space with any
   TIF/CIF/PIF set

 - fix file permission for cpum_sfb_size parameter

 - another small defconfig update

* tag 's390-5.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/cpum_sf.c: fix file permission for cpum_sfb_size
  s390: update defconfigs
  s390: fix system call exit path

3 years agoMerge tag 'mips_fixes_5.10_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips...
Linus Torvalds [Tue, 17 Nov 2020 19:15:08 +0000 (11:15 -0800)]
Merge tag 'mips_fixes_5.10_1' of git://git./linux/kernel/git/mips/linux

Pull MIPS fixes from Thomas Bogendoerfer:

 - fix bug preventing booting on several platforms

 - fix for build error, when modules need has_transparent_hugepage

 - fix for memleak in alchemy clk setup

* tag 'mips_fixes_5.10_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
  MIPS: Alchemy: Fix memleak in alchemy_clk_setup_cpu
  MIPS: kernel: Fix for_each_memblock conversion
  MIPS: export has_transparent_hugepage() for modules

3 years agotcp: only postpone PROBE_RTT if RTT is < current min_rtt estimate
Ryan Sharpelletti [Mon, 16 Nov 2020 17:44:13 +0000 (17:44 +0000)]
tcp: only postpone PROBE_RTT if RTT is < current min_rtt estimate

During loss recovery, retransmitted packets are forced to use TCP
timestamps to calculate the RTT samples, which have a millisecond
granularity. BBR is designed using a microsecond granularity. As a
result, multiple RTT samples could be truncated to the same RTT value
during loss recovery. This is problematic, as BBR will not enter
PROBE_RTT if the RTT sample is <= the current min_rtt sample, meaning
that if there are persistent losses, PROBE_RTT will constantly be
pushed off and potentially never re-entered. This patch makes sure
that BBR enters PROBE_RTT by checking if RTT sample is < the current
min_rtt sample, rather than <=.

The Netflix transport/TCP team discovered this bug in the Linux TCP
BBR code during lab tests.

Fixes: 0f8782ea1497 ("tcp_bbr: add BBR congestion control")
Signed-off-by: Ryan Sharpelletti <sharpelletti@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Link: https://lore.kernel.org/r/20201116174412.1433277-1-sharpelletti.kdev@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ftgmac100: Fix crash when removing driver
Joel Stanley [Tue, 17 Nov 2020 02:44:48 +0000 (13:14 +1030)]
net: ftgmac100: Fix crash when removing driver

When removing the driver we would hit BUG_ON(!list_empty(&dev->ptype_specific))
in net/core/dev.c due to still having the NC-SI packet handler
registered.

 # echo 1e660000.ethernet > /sys/bus/platform/drivers/ftgmac100/unbind
  ------------[ cut here ]------------
  kernel BUG at net/core/dev.c:10254!
  Internal error: Oops - BUG: 0 [#1] SMP ARM
  CPU: 0 PID: 115 Comm: sh Not tainted 5.10.0-rc3-next-20201111-00007-g02e0365710c4 #46
  Hardware name: Generic DT based system
  PC is at netdev_run_todo+0x314/0x394
  LR is at cpumask_next+0x20/0x24
  pc : [<806f5830>]    lr : [<80863cb0>]    psr: 80000153
  sp : 855bbd58  ip : 00000001  fp : 855bbdac
  r10: 80c03d00  r9 : 80c06228  r8 : 81158c54
  r7 : 00000000  r6 : 80c05dec  r5 : 80c05d18  r4 : 813b9280
  r3 : 813b9054  r2 : 8122c470  r1 : 00000002  r0 : 00000002
  Flags: Nzcv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment none
  Control: 00c5387d  Table: 85514008  DAC: 00000051
  Process sh (pid: 115, stack limit = 0x7cb5703d)
 ...
  Backtrace:
  [<806f551c>] (netdev_run_todo) from [<80707eec>] (rtnl_unlock+0x18/0x1c)
   r10:00000051 r9:854ed710 r8:81158c54 r7:80c76bb0 r6:81158c10 r5:8115b410
   r4:813b9000
  [<80707ed4>] (rtnl_unlock) from [<806f5db8>] (unregister_netdev+0x2c/0x30)
  [<806f5d8c>] (unregister_netdev) from [<805a8180>] (ftgmac100_remove+0x20/0xa8)
   r5:8115b410 r4:813b9000
  [<805a8160>] (ftgmac100_remove) from [<805355e4>] (platform_drv_remove+0x34/0x4c)

Fixes: bd466c3fb5a4 ("net/faraday: Support NCSI mode")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20201117024448.1170761-1-joel@jms.id.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: b44: fix error return code in b44_init_one()
Zhang Changzhong [Tue, 17 Nov 2020 03:02:11 +0000 (11:02 +0800)]
net: b44: fix error return code in b44_init_one()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 39a6f4bce6b4 ("b44: replace the ssb_dma API with the generic DMA API")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/1605582131-36735-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'perf-tools-fixes-for-v5.10-2020-11-17' of git://git.kernel.org/pub/scm...
Linus Torvalds [Tue, 17 Nov 2020 18:47:45 +0000 (10:47 -0800)]
Merge tag 'perf-tools-fixes-for-v5.10-2020-11-17' of git://git./linux/kernel/git/acme/linux

Pull perf tools fixes from Arnaldo Carvalho de Melo:

 - Fix file corruption due to event deletion in 'perf inject'.

 - Update arch/x86/lib/mem{cpy,set}_64.S copies used in 'perf bench mem
   memcpy', silencing perf build warning.

 - Avoid an msan warning in a copied stack in 'perf test'.

 - Correct tracepoint field name "flags" in ARM's CS-ETM hardware
   tracing 'perf test' entry.

 - Update branch sample pattern for cs-etm to cope with excluding guest
   in userspace counting.

 - Don't free "lock_seq_stat" if read_count isn't zero in 'perf lock'.

* tag 'perf-tools-fixes-for-v5.10-2020-11-17' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
  perf test: Avoid an msan warning in a copied stack.
  perf inject: Fix file corruption due to event deletion
  perf test: Update branch sample pattern for cs-etm
  perf test: Fix a typo in cs-etm testing
  tools arch: Update arch/x86/lib/mem{cpy,set}_64.S copies used in 'perf bench mem memcpy'
  perf lock: Don't free "lock_seq_stat" if read_count isn't zero
  perf lock: Correct field name "flags"

3 years agoqed: fix error return code in qed_iwarp_ll2_start()
Zhang Changzhong [Mon, 16 Nov 2020 13:07:13 +0000 (21:07 +0800)]
qed: fix error return code in qed_iwarp_ll2_start()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 469981b17a4f ("qed: Add unaligned and packed packet processing")
Fixes: fcb39f6c10b2 ("qed: Add mpa buffer descriptors for storing and processing mpa fpdus")
Fixes: 1e28eaad07ea ("qed: Add iWARP support for fpdu spanned over more than two tcp packets")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Acked-by: Michal Kalderon <michal.kalderon@marvell.com>
Link: https://lore.kernel.org/r/1605532033-27373-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'urgent-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck...
Linus Torvalds [Tue, 17 Nov 2020 18:31:56 +0000 (10:31 -0800)]
Merge branch 'urgent-fixes' of git://git./linux/kernel/git/paulmck/linux-rcu

Pull RCU fix from Paul McKenney:
 "A single commit that fixes a bug that was introduced a couple of merge
  windows ago, but which rather more recently converged to an
  agreed-upon fix. The bug is that interrupts can be incorrectly enabled
  while holding an irq-disabled spinlock. This can of course result in
  self-deadlocks.

  The bug is a bit difficult to trigger. It requires that a preempted
  task be blocking a preemptible-RCU grace period long enough to trigger
  an RCU CPU stall warning. In addition, an interrupt must occur at just
  the right time, and that interrupt's handler must acquire that same
  irq-disabled spinlock. Still, a deadlock is a deadlock.

  Furthermore, we do now have a fix, and that fix survives kernel test
  robot, -next, and rcutorture testing. It has also been verified by
  Sebastian as fixing the bug. Therefore..."

* 'urgent-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu:
  rcu: Don't invoke try_invoke_on_locked_down_task() with irqs disabled

3 years agodrm/sun4i: dw-hdmi: fix error return code in sun8i_dw_hdmi_bind()
Xiongfeng Wang [Mon, 16 Nov 2020 01:09:29 +0000 (09:09 +0800)]
drm/sun4i: dw-hdmi: fix error return code in sun8i_dw_hdmi_bind()

Fix to return a negative error code from the error handling case instead
of 0 in function sun8i_dw_hdmi_bind().

Fixes: b7c7436a5ff0 ("drm/sun4i: Implement A83T HDMI driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@siol.net>
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net>
Link: https://patchwork.freedesktop.org/patch/msgid/1605488969-5211-1-git-send-email-wangxiongfeng2@huawei.com
3 years agospi: npcm-fiu: Don't leak SPI master in probe error path
Lukas Wunner [Mon, 16 Nov 2020 08:23:10 +0000 (09:23 +0100)]
spi: npcm-fiu: Don't leak SPI master in probe error path

If the calls to of_match_device(), of_alias_get_id(),
devm_ioremap_resource(), devm_regmap_init_mmio() or devm_clk_get()
fail on probe of the NPCM FIU SPI driver, the spi_controller struct is
erroneously not freed.

Fix by switching over to the new devm_spi_alloc_master() helper.

Fixes: ace55c411b11 ("spi: npcm-fiu: add NPCM FIU controller driver")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: <stable@vger.kernel.org> # v5.4+: 5e844cc37a5c: spi: Introduce device-managed SPI controller allocation
Cc: <stable@vger.kernel.org> # v5.4+
Cc: Tomer Maimon <tmaimon77@gmail.com>
Link: https://lore.kernel.org/r/a420c23a363a3bc9aa684c6e790c32a8af106d17.1605512876.git.lukas@wunner.de
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agospi: dw: Set transfer handler before unmasking the IRQs
Serge Semin [Tue, 17 Nov 2020 09:40:54 +0000 (12:40 +0300)]
spi: dw: Set transfer handler before unmasking the IRQs

It turns out the IRQs most like can be unmasked before the controller is
enabled with no problematic consequences. The manual doesn't explicitly
state that, but the examples perform the controller initialization
procedure in that order. So the commit da8f58909e7e ("spi: dw: Unmask IRQs
after enabling the chip") hasn't been that required as I thought. But
anyway setting the IRQs up after the chip enabling still worth adding
since it has simplified the code a bit. The problem is that it has
introduced a potential bug. The transfer handler pointer is now
initialized after the IRQs are enabled. That may and eventually will cause
an invalid or uninitialized callback invocation. Fix that just by
performing the callback initialization before the IRQ unmask procedure.

Fixes: da8f58909e7e ("spi: dw: Unmask IRQs after enabling the chip")
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Link: https://lore.kernel.org/r/20201117094054.4696-1-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agoxtensa: uaccess: Add missing __user to strncpy_from_user() prototype
Laurent Pinchart [Mon, 31 Aug 2020 21:09:37 +0000 (00:09 +0300)]
xtensa: uaccess: Add missing __user to strncpy_from_user() prototype

When adding __user annotations in commit 2adf5352a34a, the
strncpy_from_user() function declaration for the
CONFIG_GENERIC_STRNCPY_FROM_USER case was missed. Fix it.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Message-Id: <20200831210937.17938-1-laurent.pinchart@ideasonboard.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
3 years agoALSA: usb-audio: Add delay quirk for all Logitech USB devices
Joakim Tjernlund [Tue, 17 Nov 2020 12:28:03 +0000 (13:28 +0100)]
ALSA: usb-audio: Add delay quirk for all Logitech USB devices

Found one more Logitech device, BCC950 ConferenceCam, which needs
the same delay here. This makes 3 out of 3 devices I have tried.

Therefore, add a delay for all Logitech devices as it does not hurt.

Signed-off-by: Joakim Tjernlund <joakim.tjernlund@infinera.com>
Cc: <stable@vger.kernel.org> # 4.19.y, 5.4.y
Link: https://lore.kernel.org/r/20201117122803.24310-1-joakim.tjernlund@infinera.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoMerge branch 'cpufreq/arm/fixes' of git://git.kernel.org/pub/scm/linux/kernel/git...
Rafael J. Wysocki [Tue, 17 Nov 2020 12:25:06 +0000 (13:25 +0100)]
Merge branch 'cpufreq/arm/fixes' of git://git./linux/kernel/git/vireshk/pm

Pull cpufreq-arm fixes for 5.10-rc5 from Viresh Kumar:

"- tegra186: Fix ->get() callback.
 - arm/scmi: Add dummy clock provider to fix failure."

* 'cpufreq/arm/fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vireshk/pm:
  cpufreq: scmi: Fix OPP addition failure with a dummy clock provider
  cpufreq: tegra186: Fix get frequency callback

3 years agoperf/x86: fix sysfs type mismatches
Sami Tolvanen [Fri, 13 Nov 2020 18:31:26 +0000 (10:31 -0800)]
perf/x86: fix sysfs type mismatches

This change switches rapl to use PMU_FORMAT_ATTR, and fixes two other
macros to use device_attribute instead of kobj_attribute to avoid
callback type mismatches that trip indirect call checking with Clang's
Control-Flow Integrity (CFI).

Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Link: https://lkml.kernel.org/r/20201113183126.1239404-1-samitolvanen@google.com
3 years agolockdep: Put graph lock/unlock under lock_recursion protection
Boqun Feng [Fri, 13 Nov 2020 11:05:03 +0000 (19:05 +0800)]
lockdep: Put graph lock/unlock under lock_recursion protection

A warning was hit when running xfstests/generic/068 in a Hyper-V guest:

[...] ------------[ cut here ]------------
[...] DEBUG_LOCKS_WARN_ON(lockdep_hardirqs_enabled())
[...] WARNING: CPU: 2 PID: 1350 at kernel/locking/lockdep.c:5280 check_flags.part.0+0x165/0x170
[...] ...
[...] Workqueue: events pwq_unbound_release_workfn
[...] RIP: 0010:check_flags.part.0+0x165/0x170
[...] ...
[...] Call Trace:
[...]  lock_is_held_type+0x72/0x150
[...]  ? lock_acquire+0x16e/0x4a0
[...]  rcu_read_lock_sched_held+0x3f/0x80
[...]  __send_ipi_one+0x14d/0x1b0
[...]  hv_send_ipi+0x12/0x30
[...]  __pv_queued_spin_unlock_slowpath+0xd1/0x110
[...]  __raw_callee_save___pv_queued_spin_unlock_slowpath+0x11/0x20
[...]  .slowpath+0x9/0xe
[...]  lockdep_unregister_key+0x128/0x180
[...]  pwq_unbound_release_workfn+0xbb/0xf0
[...]  process_one_work+0x227/0x5c0
[...]  worker_thread+0x55/0x3c0
[...]  ? process_one_work+0x5c0/0x5c0
[...]  kthread+0x153/0x170
[...]  ? __kthread_bind_mask+0x60/0x60
[...]  ret_from_fork+0x1f/0x30

The cause of the problem is we have call chain lockdep_unregister_key()
-> <irq disabled by raw_local_irq_save()> lockdep_unlock() ->
arch_spin_unlock() -> __pv_queued_spin_unlock_slowpath() -> pv_kick() ->
__send_ipi_one() -> trace_hyperv_send_ipi_one().

Although this particular warning is triggered because Hyper-V has a
trace point in ipi sending, but in general arch_spin_unlock() may call
another function having a trace point in it, so put the arch_spin_lock()
and arch_spin_unlock() after lock_recursion protection to fix this
problem and avoid similiar problems.

Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20201113110512.1056501-1-boqun.feng@gmail.com
3 years agosched/deadline: Fix priority inheritance with multiple scheduling classes
Juri Lelli [Tue, 17 Nov 2020 06:14:32 +0000 (07:14 +0100)]
sched/deadline: Fix priority inheritance with multiple scheduling classes

Glenn reported that "an application [he developed produces] a BUG in
deadline.c when a SCHED_DEADLINE task contends with CFS tasks on nested
PTHREAD_PRIO_INHERIT mutexes.  I believe the bug is triggered when a CFS
task that was boosted by a SCHED_DEADLINE task boosts another CFS task
(nested priority inheritance).

 ------------[ cut here ]------------
 kernel BUG at kernel/sched/deadline.c:1462!
 invalid opcode: 0000 [#1] PREEMPT SMP
 CPU: 12 PID: 19171 Comm: dl_boost_bug Tainted: ...
 Hardware name: ...
 RIP: 0010:enqueue_task_dl+0x335/0x910
 Code: ...
 RSP: 0018:ffffc9000c2bbc68 EFLAGS: 00010002
 RAX: 0000000000000009 RBX: ffff888c0af94c00 RCX: ffffffff81e12500
 RDX: 000000000000002e RSI: ffff888c0af94c00 RDI: ffff888c10b22600
 RBP: ffffc9000c2bbd08 R08: 0000000000000009 R09: 0000000000000078
 R10: ffffffff81e12440 R11: ffffffff81e1236c R12: ffff888bc8932600
 R13: ffff888c0af94eb8 R14: ffff888c10b22600 R15: ffff888bc8932600
 FS:  00007fa58ac55700(0000) GS:ffff888c10b00000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00007fa58b523230 CR3: 0000000bf44ab003 CR4: 00000000007606e0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 PKRU: 55555554
 Call Trace:
  ? intel_pstate_update_util_hwp+0x13/0x170
  rt_mutex_setprio+0x1cc/0x4b0
  task_blocks_on_rt_mutex+0x225/0x260
  rt_spin_lock_slowlock_locked+0xab/0x2d0
  rt_spin_lock_slowlock+0x50/0x80
  hrtimer_grab_expiry_lock+0x20/0x30
  hrtimer_cancel+0x13/0x30
  do_nanosleep+0xa0/0x150
  hrtimer_nanosleep+0xe1/0x230
  ? __hrtimer_init_sleeper+0x60/0x60
  __x64_sys_nanosleep+0x8d/0xa0
  do_syscall_64+0x4a/0x100
  entry_SYSCALL_64_after_hwframe+0x49/0xbe
 RIP: 0033:0x7fa58b52330d
 ...
 ---[ end trace 0000000000000002 ]—

He also provided a simple reproducer creating the situation below:

 So the execution order of locking steps are the following
 (N1 and N2 are non-deadline tasks. D1 is a deadline task. M1 and M2
 are mutexes that are enabled * with priority inheritance.)

 Time moves forward as this timeline goes down:

 N1              N2               D1
 |               |                |
 |               |                |
 Lock(M1)        |                |
 |               |                |
 |             Lock(M2)           |
 |               |                |
 |               |              Lock(M2)
 |               |                |
 |             Lock(M1)           |
 |             (!!bug triggered!) |

Daniel reported a similar situation as well, by just letting ksoftirqd
run with DEADLINE (and eventually block on a mutex).

Problem is that boosted entities (Priority Inheritance) use static
DEADLINE parameters of the top priority waiter. However, there might be
cases where top waiter could be a non-DEADLINE entity that is currently
boosted by a DEADLINE entity from a different lock chain (i.e., nested
priority chains involving entities of non-DEADLINE classes). In this
case, top waiter static DEADLINE parameters could be null (initialized
to 0 at fork()) and replenish_dl_entity() would hit a BUG().

Fix this by keeping track of the original donor and using its parameters
when a task is boosted.

Reported-by: Glenn Elliott <glenn@aurora.tech>
Reported-by: Daniel Bristot de Oliveira <bristot@redhat.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Daniel Bristot de Oliveira <bristot@redhat.com>
Link: https://lkml.kernel.org/r/20201117061432.517340-1-juri.lelli@redhat.com
3 years agosched: Fix rq->nr_iowait ordering
Peter Zijlstra [Thu, 24 Sep 2020 11:50:42 +0000 (13:50 +0200)]
sched: Fix rq->nr_iowait ordering

  schedule() ttwu()
    deactivate_task();   if (p->on_rq && ...) // false
    atomic_dec(&task_rq(p)->nr_iowait);
    if (prev->in_iowait)
      atomic_inc(&rq->nr_iowait);

Allows nr_iowait to be decremented before it gets incremented,
resulting in more dodgy IO-wait numbers than usual.

Note that because we can now do ttwu_queue_wakelist() before
p->on_cpu==0, we lose the natural ordering and have to further delay
the decrement.

Fixes: c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")
Reported-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: https://lkml.kernel.org/r/20201117093829.GD3121429@hirez.programming.kicks-ass.net
3 years agosched: Fix data-race in wakeup
Peter Zijlstra [Tue, 17 Nov 2020 08:08:41 +0000 (09:08 +0100)]
sched: Fix data-race in wakeup

Mel reported that on some ARM64 platforms loadavg goes bananas and
Will tracked it down to the following race:

  CPU0 CPU1

  schedule()
    prev->sched_contributes_to_load = X;
    deactivate_task(prev);

try_to_wake_up()
  if (p->on_rq &&) // false
  if (smp_load_acquire(&p->on_cpu) && // true
      ttwu_queue_wakelist())
        p->sched_remote_wakeup = Y;

    smp_store_release(prev->on_cpu, 0);

where both p->sched_contributes_to_load and p->sched_remote_wakeup are
in the same word, and thus the stores X and Y race (and can clobber
one another's data).

Whereas prior to commit c6e7bd7afaeb ("sched/core: Optimize ttwu()
spinning on p->on_cpu") the p->on_cpu handoff serialized access to
p->sched_remote_wakeup (just as it still does with
p->sched_contributes_to_load) that commit broke that by calling
ttwu_queue_wakelist() with p->on_cpu != 0.

However, due to

  p->XXX = X ttwu()
  schedule()   if (p->on_rq && ...) // false
    smp_mb__after_spinlock()   if (smp_load_acquire(&p->on_cpu) &&
    deactivate_task()       ttwu_queue_wakelist())
      p->on_rq = 0;         p->sched_remote_wakeup = Y;

We can be sure any 'current' store is complete and 'current' is
guaranteed asleep. Therefore we can move p->sched_remote_wakeup into
the current flags word.

Note: while the observed failure was loadavg accounting gone wrong due
to ttwu() cobbering p->sched_contributes_to_load, the reverse problem
is also possible where schedule() clobbers p->sched_remote_wakeup,
this could result in enqueue_entity() wrecking ->vruntime and causing
scheduling artifacts.

Fixes: c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")
Reported-by: Mel Gorman <mgorman@techsingularity.net>
Debugged-by: Will Deacon <will@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20201117083016.GK3121392@hirez.programming.kicks-ass.net
3 years agosched/fair: Fix overutilized update in enqueue_task_fair()
Quentin Perret [Thu, 12 Nov 2020 11:12:01 +0000 (11:12 +0000)]
sched/fair: Fix overutilized update in enqueue_task_fair()

enqueue_task_fair() attempts to skip the overutilized update for new
tasks as their util_avg is not accurate yet. However, the flag we check
to do so is overwritten earlier on in the function, which makes the
condition pretty much a nop.

Fix this by saving the flag early on.

Fixes: 2802bf3cd936 ("sched/fair: Add over-utilization/tipping point indicator")
Reported-by: Rick Yiu <rickyiu@google.com>
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20201112111201.2081902-1-qperret@google.com
3 years agoMIPS: Alchemy: Fix memleak in alchemy_clk_setup_cpu
Zhang Qilong [Fri, 13 Nov 2020 13:18:56 +0000 (21:18 +0800)]
MIPS: Alchemy: Fix memleak in alchemy_clk_setup_cpu

If the clk_register fails, we should free h before
function returns to prevent memleak.

Fixes: 474402291a0ad ("MIPS: Alchemy: clock framework integration of onchip clocks")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
3 years agommc: sdhci-of-arasan: Issue DLL reset explicitly
Manish Narani [Mon, 16 Nov 2020 08:32:45 +0000 (14:02 +0530)]
mmc: sdhci-of-arasan: Issue DLL reset explicitly

In the current implementation DLL reset will be issued for
each ITAP and OTAP setting inside ATF, this is creating issues
in some scenarios and this sequence is not inline with the TRM.
To fix the issue, DLL reset should be removed from the ATF and
host driver will request it explicitly.
This patch update host driver to explicitly request for DLL reset
before ITAP (assert DLL) and after OTAP (release DLL) settings.

Fixes: a5c8b2ae2e51 ("mmc: sdhci-of-arasan: Add support for ZynqMP Platform Tap Delays Setup")
Signed-off-by: Sai Krishna Potthuri <lakshmi.sai.krishna.potthuri@xilinx.com>
Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Acked-by: Michal Simek <michal.simek@xilinx.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/1605515565-117562-4-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
3 years agommc: sdhci-of-arasan: Use Mask writes for Tap delays
Manish Narani [Mon, 16 Nov 2020 08:32:44 +0000 (14:02 +0530)]
mmc: sdhci-of-arasan: Use Mask writes for Tap delays

Mask the ITAP and OTAP delay bits before updating with the new
tap value for Versal platform.

Fixes: 1a470721c8f5 ("sdhci: arasan: Add support for Versal Tap Delays")
Signed-off-by: Sai Krishna Potthuri <lakshmi.sai.krishna.potthuri@xilinx.com>
Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Acked-by: Michal Simek <michal.simek@xilinx.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/1605515565-117562-3-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
3 years agommc: sdhci-of-arasan: Allow configuring zero tap values
Manish Narani [Mon, 16 Nov 2020 08:32:43 +0000 (14:02 +0530)]
mmc: sdhci-of-arasan: Allow configuring zero tap values

Allow configuring the Output and Input tap values with zero to avoid
failures in some cases (one of them is SD boot mode) where the output
and input tap values may be already set to non-zero.

Fixes: a5c8b2ae2e51 ("mmc: sdhci-of-arasan: Add support for ZynqMP Platform Tap Delays Setup")
Signed-off-by: Sai Krishna Potthuri <lakshmi.sai.krishna.potthuri@xilinx.com>
Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Acked-by: Michal Simek <michal.simek@xilinx.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/1605515565-117562-2-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
3 years agommc: sdhci-pci: Prefer SDR25 timing for High Speed mode for BYT-based Intel controllers
Adrian Hunter [Thu, 12 Nov 2020 13:36:56 +0000 (15:36 +0200)]
mmc: sdhci-pci: Prefer SDR25 timing for High Speed mode for BYT-based Intel controllers

A UHS setting of SDR25 can give better results for High Speed mode.
This is because there is no setting corresponding to high speed.  Currently
SDHCI sets no value, which means zero which is also the setting for SDR12.
There was an attempt to change this in sdhci.c but it caused problems for
some drivers, so it was reverted and the change was made to sdhci-brcmstb
in commit 2fefc7c5f7d16e ("mmc: sdhci-brcmstb: Fix incorrect switch to HS
mode").  Several other drivers also do this.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: stable@vger.kernel.org # v5.4+
Link: https://lore.kernel.org/r/20201112133656.20317-1-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
3 years agoMerge tag 'iio-fixes-for-5.10a' of https://git.kernel.org/pub/scm/linux/kernel/git...
Greg Kroah-Hartman [Tue, 17 Nov 2020 09:53:00 +0000 (10:53 +0100)]
Merge tag 'iio-fixes-for-5.10a' of https://git./linux/kernel/git/jic23/iio into staging-linus

Jonathan writes:

First set of IIO and counter fixes for the 5.10 cycle.

IIO

cros_ec
 - Provide defauts for max and min frequency when older machines fail
   to return them correctly.
ingenic-adc
 - Fix wrong vref value for JZ4770 SoC
 - Fix AUX / VBAT readings when touchscreen in use by pausing touchscreen
   readings during a read of these channels.
kxcjk1013
 - Fix an issue with KIOX010A ACPI id using devices which need to run
   a ACPI device specific method to avoid leaving the keyboard disabled.
   Includes a minor precursor patch to make this fix easier to do.
mt6577-auxadc
 - Fix an issue with dev_comp not being set resulting in a null ptr deref.
st_lsm6dsx
 - Set a 10ms min shub slave timeout to handle fast snesors where more time
   is needed to set up the config than the cycles allowed.
stm32-adc
 - Fix an issue due to a clash between an ADC configured to use IRQs and
   a second configured to use DMA cause by some incorrect register masking.
vcnl4035
 - Kconfig missing dependency

Counter

ti-eqep
 - wrong value for max_register as one beyond the end instead of the end.

* tag 'iio-fixes-for-5.10a' of https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio:
  iio: accel: kxcjk1013: Add support for KIOX010A ACPI DSM for setting tablet-mode
  iio: accel: kxcjk1013: Replace is_smo8500_device with an acpi_type enum
  iio: light: fix kconfig dependency bug for VCNL4035
  iio/adc: ingenic: Fix AUX/VBAT readings when touchscreen is used
  iio/adc: ingenic: Fix battery VREF for JZ4770 SoC
  iio: imu: st_lsm6dsx: set 10ms as min shub slave timeout
  counter/ti-eqep: Fix regmap max_register
  iio: adc: stm32-adc: fix a regression when using dma and irq
  iio: adc: mediatek: fix unset field
  iio: cros_ec: Use default frequencies when EC returns invalid information

3 years agox86/microcode/intel: Check patch signature before saving microcode for early loading
Chen Yu [Fri, 13 Nov 2020 01:59:23 +0000 (09:59 +0800)]
x86/microcode/intel: Check patch signature before saving microcode for early loading

Currently, scan_microcode() leverages microcode_matches() to check
if the microcode matches the CPU by comparing the family and model.
However, the processor stepping and flags of the microcode signature
should also be considered when saving a microcode patch for early
update.

Use find_matching_signature() in scan_microcode() and get rid of the
now-unused microcode_matches() which is a good cleanup in itself.

Complete the verification of the patch being saved for early loading in
save_microcode_patch() directly. This needs to be done there too because
save_mc_for_early() will call save_microcode_patch() too.

The second reason why this needs to be done is because the loader still
tries to support, at least hypothetically, mixed-steppings systems and
thus adds all patches to the cache that belong to the same CPU model
albeit with different steppings.

For example:

  microcode: CPU: sig=0x906ec, pf=0x2, rev=0xd6
  microcode: mc_saved[0]: sig=0x906e9, pf=0x2a, rev=0xd6, total size=0x19400, date = 2020-04-23
  microcode: mc_saved[1]: sig=0x906ea, pf=0x22, rev=0xd6, total size=0x19000, date = 2020-04-27
  microcode: mc_saved[2]: sig=0x906eb, pf=0x2, rev=0xd6, total size=0x19400, date = 2020-04-23
  microcode: mc_saved[3]: sig=0x906ec, pf=0x22, rev=0xd6, total size=0x19000, date = 2020-04-27
  microcode: mc_saved[4]: sig=0x906ed, pf=0x22, rev=0xd6, total size=0x19400, date = 2020-04-23

The patch which is being saved for early loading, however, can only be
the one which fits the CPU this runs on so do the signature verification
before saving.

 [ bp: Do signature verification in save_microcode_patch()
       and rewrite commit message. ]

Fixes: ec400ddeff20 ("x86/microcode_intel_early.c: Early update ucode on Intel's CPU")
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: stable@vger.kernel.org
Link: https://bugzilla.kernel.org/show_bug.cgi?id=208535
Link: https://lkml.kernel.org/r/20201113015923.13960-1-yu.c.chen@intel.com
3 years agoMIPS: kernel: Fix for_each_memblock conversion
Thomas Bogendoerfer [Mon, 16 Nov 2020 17:45:15 +0000 (18:45 +0100)]
MIPS: kernel: Fix for_each_memblock conversion

The loop over all memblocks works with PFNs and not physical
addresses, so we need for_each_mem_pfn_range().

Fixes: b10d6bca8720 ("arch, drivers: replace for_each_membock() with for_each_mem_range()")
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Reviewed-by: Mike Rapoport <rppt@linux.ibm.com>
Reviewed-by: Serge Semin <fancer.lancer@gmail.com>
3 years agocpufreq: scmi: Fix OPP addition failure with a dummy clock provider
Sudeep Holla [Tue, 10 Nov 2020 11:10:40 +0000 (11:10 +0000)]
cpufreq: scmi: Fix OPP addition failure with a dummy clock provider

Commit dd461cd9183f ("opp: Allow dev_pm_opp_get_opp_table() to return
-EPROBE_DEFER") handles -EPROBE_DEFER for the clock/interconnects within
_allocate_opp_table() which is called from dev_pm_opp_add and it
now propagates the error back to the caller.

SCMI performance domain re-used clock bindings to keep it simple. However
with the above mentioned change, if clock property is present in a device
node, opps fails to get added with below errors until clk_get succeeds.

 cpu0: failed to add opp 450000000Hz
 cpu0: failed to add opps to the device
 ....(errors on cpu1-cpu4)
 cpu5: failed to add opp 450000000Hz
 cpu5: failed to add opps to the device

So, in order to fix the issue, we need to register dummy clock provider.
With the dummy clock provider, clk_get returns NULL(no errors!), then opp
core proceeds to add OPPs for the CPUs.

Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Fixes: dd461cd9183f ("opp: Allow dev_pm_opp_get_opp_table() to return -EPROBE_DEFER")
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
3 years agocpufreq: tegra186: Fix get frequency callback
Jon Hunter [Tue, 3 Nov 2020 11:55:14 +0000 (11:55 +0000)]
cpufreq: tegra186: Fix get frequency callback

Commit b89c01c96051 ("cpufreq: tegra186: Fix initial frequency")
implemented the CPUFREQ 'get' callback to determine the current
operating frequency for each CPU. This implementation used a simple
looked up to determine the current operating frequency. The problem
with this is that frequency table for different Tegra186 devices may
vary and so the default boot frequency for Tegra186 device may or may
not be present in the frequency table. If the default boot frequency is
not present in the frequency table, this causes the function
tegra186_cpufreq_get() to return 0 and in turn causes cpufreq_online()
to fail which prevents CPUFREQ from working.

Fix this by always calculating the CPU frequency based upon the current
'ndiv' setting for the CPU. Note that the CPU frequency for Tegra186 is
calculated by reading the current 'ndiv' setting, multiplying by the
CPU reference clock and dividing by a constant divisor.

Fixes: b89c01c96051 ("cpufreq: tegra186: Fix initial frequency")

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
3 years agoscsi: target: iscsi: Fix cmd abort fabric stop race
Mike Christie [Sat, 14 Nov 2020 01:46:18 +0000 (19:46 -0600)]
scsi: target: iscsi: Fix cmd abort fabric stop race

Maurizio found a race where the abort and cmd stop paths can race as
follows:

 1. thread1 runs iscsit_release_commands_from_conn and sets
    CMD_T_FABRIC_STOP.

 2. thread2 runs iscsit_aborted_task and then does __iscsit_free_cmd. It
    then returns from the aborted_task callout and we finish
    target_handle_abort and do:

    target_handle_abort -> transport_cmd_check_stop_to_fabric ->
lio_check_stop_free -> target_put_sess_cmd

    The cmd is now freed.

 3. thread1 now finishes iscsit_release_commands_from_conn and runs
    iscsit_free_cmd while accessing a command we just released.

In __target_check_io_state we check for CMD_T_FABRIC_STOP and set the
CMD_T_ABORTED if the driver is not cleaning up the cmd because of a session
shutdown. However, iscsit_release_commands_from_conn only sets the
CMD_T_FABRIC_STOP and does not check to see if the abort path has claimed
completion ownership of the command.

This adds a check in iscsit_release_commands_from_conn so only the abort or
fabric stop path cleanup the command.

Link: https://lore.kernel.org/r/1605318378-9269-1-git-send-email-michael.christie@oracle.com
Reported-by: Maurizio Lombardi <mlombard@redhat.com>
Reviewed-by: Maurizio Lombardi <mlombard@redhat.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
3 years agoscsi: libiscsi: Fix NOP race condition
Lee Duncan [Fri, 6 Nov 2020 19:33:17 +0000 (11:33 -0800)]
scsi: libiscsi: Fix NOP race condition

iSCSI NOPs are sometimes "lost", mistakenly sent to the user-land iscsid
daemon instead of handled in the kernel, as they should be, resulting in a
message from the daemon like:

  iscsid: Got nop in, but kernel supports nop handling.

This can occur because of the new forward- and back-locks, and the fact
that an iSCSI NOP response can occur before processing of the NOP send is
complete. This can result in "conn->ping_task" being NULL in
iscsi_nop_out_rsp(), when the pointer is actually in the process of being
set.

To work around this, we add a new state to the "ping_task" pointer. In
addition to NULL (not assigned) and a pointer (assigned), we add the state
"being set", which is signaled with an INVALID pointer (using "-1").

Link: https://lore.kernel.org/r/20201106193317.16993-1-leeman.duncan@gmail.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
3 years agoMerge branch 'bnxt_en-bug-fixes'
Jakub Kicinski [Tue, 17 Nov 2020 01:39:49 +0000 (17:39 -0800)]
Merge branch 'bnxt_en-bug-fixes'

Michael Chan says:

====================
bnxt_en: Bug fixes.

This first patch fixes a module eeprom A2h addressing issue.  The next
2 patches fix counter related issues.  The last one skips an
unsupported firmware call on the VF to avoid the error log.
====================

Link: https://lore.kernel.org/r/1605486472-28156-1-git-send-email-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Avoid unnecessary NVM_GET_DEV_INFO cmd error log on VFs.
Vasundhara Volam [Mon, 16 Nov 2020 00:27:52 +0000 (19:27 -0500)]
bnxt_en: Avoid unnecessary NVM_GET_DEV_INFO cmd error log on VFs.

VFs do not have access permissions to issue NVM_GET_DEV_INFO
firmware command.

Fixes: 4933f6753b50 ("bnxt_en: Add bnxt_hwrm_nvm_get_dev_info() to query NVM info.")
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Fix counter overflow logic.
Michael Chan [Mon, 16 Nov 2020 00:27:51 +0000 (19:27 -0500)]
bnxt_en: Fix counter overflow logic.

bnxt_add_one_ctr() adds a hardware counter to a software counter and
adjusts for the hardware counter wraparound against the mask.  The logic
assumes that the hardware counter is always smaller than or equal to
the mask.

This assumption is mostly correct.  But in some cases if the firmware
is older and does not provide the accurate mask, the driver can use
a mask that is smaller than the actual hardware mask.  This can cause
some extra carry bits to be added to the software counter, resulting in
counters that far exceed the actual value.  Fix it by masking the
hardware counter with the mask passed into bnxt_add_one_ctr().

Fixes: fea6b3335527 ("bnxt_en: Accumulate all counters.")
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Free port stats during firmware reset.
Michael Chan [Mon, 16 Nov 2020 00:27:50 +0000 (19:27 -0500)]
bnxt_en: Free port stats during firmware reset.

Firmware is unable to retain the port counters during any kind of
fatal or non-fatal resets, so we must clear the port counters to
avoid false detection of port counter overflow.

Fixes: fea6b3335527 ("bnxt_en: Accumulate all counters.")
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: read EEPROM A2h address using page 0
Edwin Peer [Mon, 16 Nov 2020 00:27:49 +0000 (19:27 -0500)]
bnxt_en: read EEPROM A2h address using page 0

The module eeprom address range returned by bnxt_get_module_eeprom()
should be 256 bytes of A0h address space, the lower half of the A2h
address space, and page 0 for the upper half of the A2h address space.

Fix the firmware call by passing page_number 0 for the A2h slave address
space.

Fixes: 42ee18fe4ca2 ("bnxt_en: Add Support for ETHTOOL_GMODULEINFO and ETHTOOL_GMODULEEEPRO")
Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ipa: lock when freeing transaction
Alex Elder [Sat, 14 Nov 2020 18:20:17 +0000 (12:20 -0600)]
net: ipa: lock when freeing transaction

Transactions sit on one of several lists, depending on their state
(allocated, pending, complete, or polled).  A spinlock protects
against concurrent access when transactions are moved between these
lists.

Transactions are also reference counted.  A newly-allocated
transaction has an initial count of 1; a transaction is released in
gsi_trans_free() only if its decremented reference count reaches 0.
Releasing a transaction includes removing it from the polled (or if
unused, allocated) list, so the spinlock is acquired when we release
a transaction.

The reference count is used to allow a caller to synchronously wait
for a committed transaction to complete.  In this case, the waiter
takes an extra reference to the transaction *before* committing it
(so it won't be freed), and releases its reference (calls
gsi_trans_free()) when it is done with it.

Similarly, gsi_channel_update() takes an extra reference to ensure a
transaction isn't released before the function is done operating on
it.  Until the transaction is moved to the completed list (by this
function) it won't be freed, so this reference is taken "safely."

But in the quiesce path, we want to wait for the "last" transaction,
which we find in the completed or polled list.  Transactions on
these lists can be freed at any time, so we (try to) prevent that
by taking the reference while holding the spinlock.

Currently gsi_trans_free() decrements a transaction's reference
count unconditionally, acquiring the lock to remove the transaction
from its list *only* when the count reaches 0.  This does not
protect the quiesce path, which depends on the lock to ensure its
extra reference prevents release of the transaction.

Fix this by only dropping the last reference to a transaction
in gsi_trans_free() while holding the spinlock.

Fixes: 9dd441e4ed575 ("soc: qcom: ipa: GSI transactions")
Reported-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/20201114182017.28270-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet/tls: fix corrupted data in recvmsg
Vadim Fedorenko [Sun, 15 Nov 2020 04:16:00 +0000 (07:16 +0300)]
net/tls: fix corrupted data in recvmsg

If tcp socket has more data than Encrypted Handshake Message then
tls_sw_recvmsg will try to decrypt next record instead of returning
full control message to userspace as mentioned in comment. The next
message - usually Application Data - gets corrupted because it uses
zero copy for decryption that's why the data is not stored in skb
for next iteration. Revert check to not decrypt next record if
current is not Application Data.

Fixes: 692d7b5d1f91 ("tls: Fix recvmsg() to be able to peek across multiple records")
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Link: https://lore.kernel.org/r/1605413760-21153-1-git-send-email-vfedorenko@novek.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: qualcomm: rmnet: Fix incorrect receive packet handling during cleanup
Subash Abhinov Kasiviswanathan [Fri, 13 Nov 2020 20:12:05 +0000 (13:12 -0700)]
net: qualcomm: rmnet: Fix incorrect receive packet handling during cleanup

During rmnet unregistration, the real device rx_handler is first cleared
followed by the removal of rx_handler_data after the rcu synchronization.

Any packets in the receive path may observe that the rx_handler is NULL.
However, there is no check when dereferencing this value to use the
rmnet_port information.

This fixes following splat by adding the NULL check.

Unable to handle kernel NULL pointer dereference at virtual
address 000000000000000d
pc : rmnet_rx_handler+0x124/0x284
lr : rmnet_rx_handler+0x124/0x284
 rmnet_rx_handler+0x124/0x284
 __netif_receive_skb_core+0x758/0xd74
 __netif_receive_skb+0x50/0x17c
 process_backlog+0x15c/0x1b8
 napi_poll+0x88/0x284
 net_rx_action+0xbc/0x23c
 __do_softirq+0x20c/0x48c

Fixes: ceed73a2cf4a ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation")
Signed-off-by: Sean Tranchetti <stranche@codeaurora.org>
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Link: https://lore.kernel.org/r/1605298325-3705-1-git-send-email-subashab@codeaurora.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: mvneta: fix possible memory leak in mvneta_swbm_add_rx_fragment
Lorenzo Bianconi [Fri, 13 Nov 2020 18:16:57 +0000 (19:16 +0100)]
net: mvneta: fix possible memory leak in mvneta_swbm_add_rx_fragment

Recycle the page running page_pool_put_full_page() in
mvneta_swbm_add_rx_fragment routine when the last descriptor
contains just the FCS or if the received packet contains more than
MAX_SKB_FRAGS fragments

Fixes: ca0e014609f0 ("net: mvneta: move skb build after descriptors processing")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/df6a2bad70323ee58d3901491ada31c1ca2a40b9.1605291228.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: stmmac: Use rtnl_lock/unlock on netif_set_real_num_rx_queues() call
Wong Vee Khee [Sun, 15 Nov 2020 07:42:10 +0000 (15:42 +0800)]
net: stmmac: Use rtnl_lock/unlock on netif_set_real_num_rx_queues() call

Fix an issue where dump stack is printed on suspend resume flow due to
netif_set_real_num_rx_queues() is not called with rtnl_lock held().

Fixes: 686cff3d7022 ("net: stmmac: Fix incorrect location to set real_num_rx|tx_queues")
Reported-by: Christophe ROULLIER <christophe.roullier@st.com>
Tested-by: Christophe ROULLIER <christophe.roullier@st.com>
Cc: Alexandre TORGUE <alexandre.torgue@st.com>
Reviewed-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: Wong Vee Khee <vee.khee.wong@intel.com>
Link: https://lore.kernel.org/r/20201115074210.23605-1-vee.khee.wong@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: bridge: add missing counters to ndo_get_stats64 callback
Heiner Kallweit [Fri, 13 Nov 2020 09:27:27 +0000 (10:27 +0100)]
net: bridge: add missing counters to ndo_get_stats64 callback

In br_forward.c and br_input.c fields dev->stats.tx_dropped and
dev->stats.multicast are populated, but they are ignored in
ndo_get_stats64.

Fixes: 28172739f0a2 ("net: fix 64 bit counters on 32 bit arches")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/58ea9963-77ad-a7cf-8dfd-fc95ab95f606@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ethernet: ti: cpsw: fix error return code in cpsw_probe()
Zhang Changzhong [Fri, 13 Nov 2020 06:49:33 +0000 (14:49 +0800)]
net: ethernet: ti: cpsw: fix error return code in cpsw_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 83a8471ba255 ("net: ethernet: ti: cpsw: refactor probe to group common hw initialization")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605250173-18438-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: stmmac: dwmac-intel-plat: fix error return code in intel_eth_plat_probe()
Zhang Changzhong [Fri, 13 Nov 2020 06:34:03 +0000 (14:34 +0800)]
net: stmmac: dwmac-intel-plat: fix error return code in intel_eth_plat_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 9efc9b2b04c7 ("net: stmmac: Add dwmac-intel-plat for GBE driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605249243-17262-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>