linux-stable.git/drivers/rpmsg/rpmsg_core.c, branch v4.14.331

rpmsg: Fix possible refcount leak in rpmsg_register_device_override()

2023-11-08T10:21:08+00:00

commit d7bd416d35121c95fe47330e09a5c04adbc5f928 upstream.

rpmsg_register_device_override need to call put_device to free vch when
driver_set_override fails.

Fix this by adding a put_device() to the error path.

Fixes: bb17d110cbf2 ("rpmsg: Fix calling device_lock() on non-initialized device")
Reviewed-by: Krzysztof Kozlowski 
Signed-off-by: Hangyu Hua 
Link: https://lore.kernel.org/r/20220624024120.11576-1-hbh25y@gmail.com
Signed-off-by: Mathieu Poirier 
(cherry picked from commit d7bd416d35121c95fe47330e09a5c04adbc5f928)
Signed-off-by: Lee Jones 
Signed-off-by: Greg Kroah-Hartman

rpmsg: Fix calling device_lock() on non-initialized device

2023-11-08T10:21:08+00:00

commit bb17d110cbf270d5247a6e261c5ad50e362d1675 upstream.

driver_set_override() helper uses device_lock() so it should not be
called before rpmsg_register_device() (which calls device_register()).
Effect can be seen with CONFIG_DEBUG_MUTEXES:

  DEBUG_LOCKS_WARN_ON(lock->magic != lock)
  WARNING: CPU: 3 PID: 57 at kernel/locking/mutex.c:582 __mutex_lock+0x1ec/0x430
  ...
  Call trace:
   __mutex_lock+0x1ec/0x430
   mutex_lock_nested+0x44/0x50
   driver_set_override+0x124/0x150
   qcom_glink_native_probe+0x30c/0x3b0
   glink_rpm_probe+0x274/0x350
   platform_probe+0x6c/0xe0
   really_probe+0x17c/0x3d0
   __driver_probe_device+0x114/0x190
   driver_probe_device+0x3c/0xf0
   ...

Refactor the rpmsg_register_device() function to use two-step device
registering (initialization + add) and call driver_set_override() in
proper moment.

This moves the code around, so while at it also NULL-ify the
rpdev->driver_override in error path to be sure it won't be kfree()
second time.

Fixes: 42cd402b8fd4 ("rpmsg: Fix kfree() of static memory on setting driver_override")
Reported-by: Marek Szyprowski 
Signed-off-by: Krzysztof Kozlowski 
Tested-by: Marek Szyprowski 
Link: https://lore.kernel.org/r/20220429195946.1061725-2-krzysztof.kozlowski@linaro.org
Signed-off-by: Greg Kroah-Hartman 
(cherry picked from commit bb17d110cbf270d5247a6e261c5ad50e362d1675)
Signed-off-by: Lee Jones 
Signed-off-by: Greg Kroah-Hartman

Revert "rpmsg: core: add support to power domains for devices"

2018-09-29T10:06:04+00:00

This reverts commit 1ed3a93072307265d6385031b72929a904b50f87 which is
commit fe782affd0f440a4e60e2cc81b8f2eccb2923113 upstream

Rafael reports that this patch causes problems:
	> -rc2 looks good. There is a problem on dragonboard during boot that was
	> introduced in v4.14.71 that I didn't notice last week. We'll bisect it
	> and report back later this week. dragonboard on the other branches (4.9,
	> 4.18, mainline) looks fine.

	As Dan pointed out, during validation, we have bisected this issue on
	a dragonboard 410c (can't find root device) to the following commit
	for v4.14:

	[1ed3a9307230] rpmsg: core: add support to power domains for devices

	There is an on-going discussion on "[PATCH] rpmsg: core: add support
	to power domains for devices" about this patch having other
	dependencies and breaking something else on v4.14 as well.

so drop it.

Reported-by: Rafael Tinoco 
Cc: Srinivas Kandagatla 
Cc: Bjorn Andersson 
Cc: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

rpmsg: core: add support to power domains for devices

2018-09-19T20:43:41+00:00

[ Upstream commit fe782affd0f440a4e60e2cc81b8f2eccb2923113 ]

Some of the rpmsg devices need to switch on power domains to communicate
with remote processor. For example on Qualcomm DB820c platform LPASS
power domain needs to switched on for any kind of audio services.
This patch adds the missing power domain support in rpmsg core.

Without this patch attempting to play audio via QDSP on DB820c would
reboot the system.

Signed-off-by: Srinivas Kandagatla 
Signed-off-by: Bjorn Andersson 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

Merge tag 'rpmsg-v4.13' of git://github.com/andersson/remoteproc

2017-07-06T22:38:31+00:00

Pull rpmsg updates from Bjorn Andersson:
 "This introduces the Qualcomm GLINK protocol driver and
  DeviceTree-based modalias support, as well as a number of smaller
  fixes"

* tag 'rpmsg-v4.13' of git://github.com/andersson/remoteproc:
  rpmsg: Make modalias work for DeviceTree based devices
  rpmsg: Drop VIRTUALIZATION dependency from RPMSG_VIRTIO
  rpmsg: Don't overwrite release op of rpdev
  rpmsg: virtio_rpmsg_bus: cleanup multiple assignment to ops
  rpmsg: virtio_rpmsg_bus: fix nameservice address
  rpmsg: cleanup incorrect function in dev_err message
  rpmsg: virtio_rpmsg_bus: fix announce for devices without endpoint
  rpmsg: Introduce Qualcomm RPM glink driver
  soc: qcom: Add device tree binding for GLINK RPM
  rpmsg: Release rpmsg devices in backends

rpmsg: Make modalias work for DeviceTree based devices

2017-06-29T04:49:24+00:00

When rpmsg devices are expected to be matched based on their compatible
the modalias should reflect this, so that module autoloading has a
chance to match and load the appropriate module.

Tested-by: Rob Clark 
Reported-by: Rob Clark 
Signed-off-by: Bjorn Andersson

rpmsg: Don't overwrite release op of rpdev

2017-06-27T05:39:56+00:00

b0b03b811963 ("rpmsg: Release rpmsg devices in backends") attempted to
correct the ownership of freeing rpmsg device memory. But the patch
is not complete, in that the rpmsg core will overwrite the release op as
the device is being registered.

Fixes: b0b03b811963 ("rpmsg: Release rpmsg devices in backends")
Reported-by: Henri Roosen 
Signed-off-by: Bjorn Andersson

rpmsg: use dev_groups and not dev_attrs for bus_type

2017-06-09T09:00:46+00:00

The dev_attrs field has long been "depreciated" and is finally being
removed, so move the driver to use the "correct" dev_groups field
instead for struct bus_type.

Cc: Ohad Ben-Cohen 
Acked-by: Bjorn Andersson 
Cc: 
Signed-off-by: Greg Kroah-Hartman

rpmsg: rpmsg_create_ept() returns NULL on error

2017-02-13T08:52:54+00:00

The parameter validation incorrectly returned an ERR_PTR(), which is not
handled by the callers to rpmsg_create_ept(), per the definition NULL
should be returned.

Fixes: 93e9324431c9 ("rpmsg: Handle invalid parameters in public API")
Reported-by: Dan Carpenter 
Signed-off-by: Bjorn Andersson

rpmsg: Introduce "poll" to endpoint ops

2017-01-18T18:43:15+00:00

This allows rpmsg backends to implement polling of the outgoing buffer,
which provides poll support to user space when using the rpmsg character
device.

Signed-off-by: Bjorn Andersson