commit 0462c56c290a99a7f03e817ae5b843116dfb575c upstream.

The commit 80dd33cf72d1 ("drivers: base: Fix device link removal")
introduces a workqueue to release the consumer and supplier devices used
in the devlink.
In the job queued, devices are release and in turn, when all the
references to these devices are dropped, the release function of the
device itself is called.

Nothing is present to provide some synchronisation with this workqueue
in order to ensure that all ongoing releasing operations are done and
so, some other operations can be started safely.

For instance, in the following sequence:
  1) of_platform_depopulate()
  2) of_overlay_remove()

During the step 1, devices are released and related devlinks are removed
(jobs pushed in the workqueue).
During the step 2, OF nodes are destroyed but, without any
synchronisation with devlink removal jobs, of_overlay_remove() can raise
warnings related to missing of_node_put():
  ERROR: memory leak, expected refcount 1 instead of 2

Indeed, the missing of_node_put() call is going to be done, too late,
from the workqueue job execution.

Introduce device_link_wait_removal() to offer a way to synchronize
operations waiting for the end of devlink removals (i.e. end of
workqueue jobs).
Also, as a flushing operation is done on the workqueue, the workqueue
used is moved from a system-wide workqueue to a local one.

Cc: stable@vger.kernel.org
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Tested-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Nuno Sa <nuno.sa@analog.com>
Reviewed-by: Saravana Kannan <saravanak@google.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20240325152140.198219-2-herve.codina@bootlin.com
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

[ Upstream commit 6442d79d880cf7a2fff18779265d657fef0cce4c ]

fw_devlink can detect most overlapping/intersecting cycles. However it was
missing a few corner cases because of an incorrect optimization logic that
tries to avoid repeating cycle detection for devices that are already
marked as part of a cycle.

Here's an example provided by Xu Yang (edited for clarity):

                    usb
                  +-----+
   tcpc           |     |
  +-----+         |  +--|
  |     |----------->|EP|
  |--+  |         |  +--|
  |EP|<-----------|     |
  |--+  |         |  B  |
  |     |         +-----+
  |  A  |            |
  +-----+            |
     ^     +-----+   |
     |     |     |   |
     +-----|  C  |<--+
           |     |
           +-----+
           usb-phy

Node A (tcpc) will be populated as device 1-0050.
Node B (usb) will be populated as device 38100000.usb.
Node C (usb-phy) will be populated as device 381f0040.usb-phy.

The description below uses the notation:
consumer --> supplier
child ==> parent

1. Node C is populated as device C. No cycles detected because cycle
   detection is only run when a fwnode link is converted to a device link.

2. Node B is populated as device B. As we convert B --> C into a device
   link we run cycle detection and find and mark the device link/fwnode
   link cycle:
   C--> A --> B.EP ==> B --> C

3. Node A is populated as device A. As we convert C --> A into a device
   link, we see it's already part of a cycle (from step 2) and don't run
   cycle detection. Thus we miss detecting the cycle:
   A --> B.EP ==> B --> A.EP ==> A

Looking at it another way, A depends on B in one way:
A --> B.EP ==> B

But B depends on A in two ways and we only detect the first:
B --> C --> A
B --> A.EP ==> A

To detect both of these, we remove the incorrect optimization attempt in
step 3 and run cycle detection even if the fwnode link from which the
device link is being created has already been marked as part of a cycle.

Reported-by: Xu Yang <xu.yang_2@nxp.com>
Closes: https://lore.kernel.org/lkml/DU2PR04MB8822693748725F85DC0CB86C8C792@DU2PR04MB8822.eurprd04.prod.outlook.com/
Fixes: 3fb16866b51d ("driver core: fw_devlink: Make cycle detection more robust")
Signed-off-by: Saravana Kannan <saravanak@google.com>
Tested-by: Xu Yang <xu.yang_2@nxp.com>
Link: https://lore.kernel.org/r/20240202095636.868578-3-saravanak@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

commit 7fddac12c38237252431d5b8af7b6d5771b6d125 upstream.

device_link_flag_is_sync_state_only() correctly returns true on the flags
of an existing device link that only implements sync_state() functionality.
However, it incorrectly and confusingly returns false if it's called with
DL_FLAG_SYNC_STATE_ONLY.

This bug doesn't manifest in any of the existing calls to this function,
but fix this confusing behavior to avoid future bugs.

Fixes: 67cad5c67019 ("driver core: fw_devlink: Add DL_FLAG_CYCLE support to device links")
Signed-off-by: Saravana Kannan <saravanak@google.com>
Tested-by: Xu Yang <xu.yang_2@nxp.com>
Link: https://lore.kernel.org/r/20240202095636.868578-2-saravanak@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The commit d21fdd07cea4 ("driver core: Return proper error code when
dev_set_name() fails") rewrote the logic of handling the dev_set_name()
error codes, but missed the point that initially set error value to
-EINVAL might be rewritten and hence the error path can't be triggered
at some circumstances. To fix this, make sure that error variable is
set to -EINVAL when other conditionals are false.

Reported-by: syzbot+bdfb03b1ec8b342c12cb@syzkaller.appspotmail.com
Fixes: d21fdd07cea4 ("driver core: Return proper error code when dev_set_name() fails")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230828145824.3895288-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Pull driver core updates from Greg KH:
 "Here is a small set of driver core updates and additions for 6.6-rc1.

  Included in here are:

   - stable kernel documentation updates

   - class structure const work from Ivan on various subsystems

   - kernfs tweaks

   - driver core tests!

   - kobject sanity cleanups

   - kobject structure reordering to save space

   - driver core error code handling fixups

   - other minor driver core cleanups

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

* tag 'driver-core-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (32 commits)
  driver core: Call in reversed order in device_platform_notify_remove()
  driver core: Return proper error code when dev_set_name() fails
  kobject: Remove redundant checks for whether ktype is NULL
  kobject: Add sanity check for kset->kobj.ktype in kset_register()
  drivers: base: test: Add missing MODULE_* macros to root device tests
  drivers: base: test: Add missing MODULE_* macros for platform devices tests
  drivers: base: Free devm resources when unregistering a device
  drivers: base: Add basic devm tests for platform devices
  drivers: base: Add basic devm tests for root devices
  kernfs: fix missing kernfs_iattr_rwsem locking
  docs: stable-kernel-rules: mention that regressions must be prevented
  docs: stable-kernel-rules: fine-tune various details
  docs: stable-kernel-rules: make the examples for option 1 a proper list
  docs: stable-kernel-rules: move text around to improve flow
  docs: stable-kernel-rules: improve structure by changing headlines
  base/node: Remove duplicated include
  kernfs: attach uuid for every kernfs and report it in fsid
  kernfs: add stub helper for kernfs_generic_poll()
  x86/resctrl: make pseudo_lock_class a static const structure
  x86/MSR: make msr_class a static const structure
  ...

Pull clk subsystem updates from Stephen Boyd:
 "This pull request is full of clk driver changes. In fact, there aren't
  any changes to the clk framework this time around. That's probably
  because everyone was on vacation (yours truly included). We did lose a
  couple clk drivers this time around because nobody was using those
  devices. That skews the diffstat a bit, but either way, nothing looks
  out of the ordinary here. The usual suspects are chugging along adding
  support for more SoCs and fixing bugs.

  If I had to choose, I'd say the theme for the past few months has been
  "polish". There's quite a few patches that migrate to
  devm_platform_ioremap_resource() in here. And there's more than a
  handful of patches that move the NR_CLKS define from the DT binding
  header to the driver. There's even patches that migrate drivers to use
  clk_parent_data and clk_hw to describe clk tree topology. It seems
  that the spring (summer?) cleaning bug got some folks, or the
  semiconductor shortage finally hit the software side.

  New Drivers:
   - StarFive JH7110 SoC clock drivers
   - Qualcomm IPQ5018 Global Clock Controller driver
   - Versa3 clk generator to support 48KHz playback/record with audio
     codec on RZ/G2L SMARC EVK

  Removed Drivers:
   - Remove non-OF mmp clk drivers
   - Remove OXNAS clk driver

  Updates:
   - Add __counted_by to struct clk_hw_onecell_data and struct
     spmi_pmic_div_clk_cc
   - Move defines for numbers of clks (NR_CLKS) from DT headers to
     drivers
   - Introduce kstrdup_and_replace() and use it
   - Add PLL rates for Rockchip rk3568
   - Add the display clock tree for Rockchip rv1126
   - Add Audio Clock Generator (ADG) clocks on Renesas R-Car Gen3 and
     RZ/G2 SoCs
   - Convert sun9i-mmc clock to use
     devm_platform_get_and_ioremap_resource()
   - Fix function name in a comment in ccu_mmc_timing.c
   - Parameter name correction for ccu_nkm_round_rate()
   - Implement CLK_SET_RATE_PARENT for Allwinner NKM clocks, i.e.
     consider alternative parent rates when determining clock rates
   - Set CLK_SET_RATE_PARENT for Allwinner A64 pll-mipi
   - Support finding closest (as opposed to closest but not higher)
     clock rate for NM, NKM, mux and div type clocks, as use it for
     Allwinner A64 pll-video0
   - Prefer current parent rate if able to generate ideal clock rate for
     Allwinner NKM clocks
   - Clean up Qualcomm SMD RPM driver, with interconnect bus clocks
     moved out to the interconnect drivers
   - Fix various PM runtime bugs across many Qualcomm clk drivers
   - Migrate Qualcomm MDM9615 is to parent_hw and parent_data
   - Add network related resets on Qualcomm IPQ4019
   - Add a couple missing USB related clocks to Qualcomm IPQ9574
   - Add missing gpll0_sleep_clk_src to Qualcomm MSM8917 global clock
     controller
   - In the Qualcomm QDU1000 global clock controller, GDSCs, clkrefs,
     and GPLL1 are added, while PCIe pipe clock, SDCC rcg ops are
     corrected
   - Add missing GDSCs to and correct GDSCs for the SC8280XP global
     clock controller driver
   - Support retention for the Qualcomm SC8280XP display clock
     controller GDSCs.
   - Qualcommm's SDCC apps_clk_src is marked with CLK_OPS_PARENT_ENABLE
     to fix issues with missing parent clocks across sc7180, sm7150,
     sm6350 and sm8250, while sm8450 is corrected to use floor ops
   - Correct Qualcomm SM6350 GPU clock controller's clock supplies
   - Drop unwanted clocks from the Qualcomm IPQ5332 GCC driver
   - Add missing OXILICX GDSC to Qualcomm MSM8226 GCC
   - Change the delay in the Qualcomm reset controller to fsleep() for
     correctness
   - Extend the Qualcomm SM83550 Video clock controller to support
     SC8280XP
   - Add graphics clock support on Renesas RZ/G2M, RZ/G2N, RZ/G2E, and
     R-Car H3, M3-W, and M3-N SoCs
   - Add Clocked Serial Interface (CSI) clocks on Renesas RZ/V2M
   - Add PWM (MTU3) clock and reset on Renesas RZ/G2UL and RZ/Five
   - Add the PDM IPC clock for i.MX93
   - Add 519.75MHz frequency support for i.MX9 PLL
   - Simplify the .determine_rate() implementation for i.MX GPR mux
   - Make the i.MX8QXP LPCG clock use devm_platform_ioremap_resource()
   - Add the audio mux clock to i.MX8
   - Fix the SPLL2 MULT range for PLLv4
   - Update the SPLL2 type in i.MX8ULP
   - Fix the SAI4 clock on i.MX8MP
   - Add silicon revision print for i.MX25 on clocks init
   - Drop the return value from __mx25_clocks_init()
   - Fix the clock pauses on no-op set_rate for i.MX8M composite clock
   - Drop restrictions for i.MX PLL14xx and fix its max prediv value
   - Drop the 393216000 and 361267200 from i.MX PLL14xx rate table to
     allow glitch free switching"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (207 commits)
  clk: qcom: Fix SM_GPUCC_8450 dependencies
  clk: lmk04832: Support using PLL1_LD as SPI readback pin
  clk: lmk04832: Don't disable vco clock on probe fail
  clk: lmk04832: Set missing parent_names for output clocks
  clk: mvebu: Convert to devm_platform_ioremap_resource()
  clk: nuvoton: Convert to devm_platform_ioremap_resource()
  clk: socfpga: agilex: Convert to devm_platform_ioremap_resource()
  clk: ti: Use devm_platform_get_and_ioremap_resource()
  clk: mediatek: Convert to devm_platform_ioremap_resource()
  clk: hsdk-pll: Convert to devm_platform_ioremap_resource()
  clk: gemini: Convert to devm_platform_ioremap_resource()
  clk: fsl-sai: Convert to devm_platform_ioremap_resource()
  clk: bm1880: Convert to devm_platform_ioremap_resource()
  clk: axm5516: Convert to devm_platform_ioremap_resource()
  clk: actions: Convert to devm_platform_ioremap_resource()
  clk: cdce925: Remove redundant of_match_ptr()
  clk: pxa910: Move number of clocks to driver source
  clk: pxa1928: Move number of clocks to driver source
  clk: pxa168: Move number of clocks to driver source
  clk: mmp2: Move number of clocks to driver source
  ...

It's logically correct to call the removal notifiers in the reversed order
as it might be dependent to each other. Luckily, platform_notify_remove()
currently is not used and the others have no dependency use, but theoretically
it's still possible.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230818133654.767986-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Whe device_add() tries to assign a device name with help of
dev_set_name() the error path explicitly uses -EINVAL, while
the function may return something different.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230817091221.463721-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In the current code, devres_release_all() only gets called if the device
has a bus and has been probed.

This leads to issues when using bus-less or driver-less devices where
the device might never get freed if a managed resource holds a reference
to the device. This is happening in the DRM framework for example.

We should thus call devres_release_all() in the device_del() function to
make sure that the device-managed actions are properly executed when the
device is unregistered, even if it has neither a bus nor a driver.

This is effectively the same change than commit 2f8d16a996da ("devres:
release resources on device_del()") that got reverted by commit
a525a3ddeaca ("driver core: free devres in device_release") over
memory leaks concerns.

This patch effectively combines the two commits mentioned above to
release the resources both on device_del() and device_release() and get
the best of both worlds.

Fixes: a525a3ddeaca ("driver core: free devres in device_release")
Signed-off-by: David Gow <davidgow@google.com>
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20230720-kunit-devm-inconsistencies-test-v3-3-6aa7e074f373@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Replace open coded functionality of kstrdup_and_replace() with a call.

Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230804143910.15504-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Stephen Boyd <sboyd@kernel.org>