platform/kernel/linux-rpi.git
2 years agomedia: subdev: add active state to struct v4l2_subdev
Tomi Valkeinen [Tue, 12 Apr 2022 09:42:43 +0000 (10:42 +0100)]
media: subdev: add active state to struct v4l2_subdev

Add a new 'active_state' field to struct v4l2_subdev to which we can
store the active state of a subdev. This will place the subdev
configuration into a known place, allowing us to use the state directly
from the v4l2 framework, thus simplifying the drivers.

Also add functions v4l2_subdev_init_finalize() and
v4l2_subdev_cleanup(), which will allocate and free the active state.
The functions are named in a generic way so that they can be also used
for other subdev initialization work.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: subdev: rename subdev-state alloc & free
Tomi Valkeinen [Tue, 12 Apr 2022 09:42:42 +0000 (10:42 +0100)]
media: subdev: rename subdev-state alloc & free

v4l2_subdev_alloc_state() and v4l2_subdev_free_state() are not supposed
to be used by the drivers. However, we do have a few drivers that use
those at the moment, so we need to expose these functions for the time
being.

Prefix the functions with __ to mark the functions as internal.

At the same time, rename them to v4l2_subdev_state_alloc and
v4l2_subdev_state_free to match the style used for other functions like
video_device_alloc() and media_request_alloc().

Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: v4l2-subdev: drop extra #ifdef
Tomi Valkeinen [Tue, 12 Apr 2022 09:42:41 +0000 (10:42 +0100)]
media: v4l2-subdev: drop extra #ifdef

subdev_open() is inside #ifdef CONFIG_VIDEO_V4L2_SUBDEV_API, which
depends on CONFIG_MEDIA_CONTROLLER, so there's no need for an extra

Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: v4l2-subdev: fix #endif comments
Tomi Valkeinen [Tue, 12 Apr 2022 09:42:40 +0000 (10:42 +0100)]
media: v4l2-subdev: fix #endif comments

Add comments after #endifs to clarify their scope.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: i2c: ov5645: Fix media bus format
Jacopo Mondi [Mon, 28 Mar 2022 16:32:20 +0000 (17:32 +0100)]
media: i2c: ov5645: Fix media bus format

The ov5645 driver reports as its unique supported format
MEDIA_BUS_FMT_UYVY8_2X8, which is not correct as the sensor
uses the MIPI CSI-2 serial bus.

Fix that by using MEDIA_BUS_FMT_UYVY8_1X16 instead.

Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: mc: delete redundant code in __media_device_unregister_entity
Hangyu Hua [Thu, 24 Mar 2022 10:27:52 +0000 (10:27 +0000)]
media: mc: delete redundant code in __media_device_unregister_entity

media_gobj_destroy has already set graph_obj.mdev to NULL. There is no
need to set it again.

[Sakari Ailus: Remove extra newline, rewrap commit msg]

Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: staging/intel-ipu3: Cleanup dummy buffers via helper
Umang Jain [Fri, 18 Mar 2022 12:18:26 +0000 (12:18 +0000)]
media: staging/intel-ipu3: Cleanup dummy buffers via helper

Use an existing helper imgu_video_nodes_exit() on imgu-video-node
initialization failure path to cleanup dummy buffers.

Signed-off-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: staging/intel-ipu3: Reset imgu_video_device sequence
Umang Jain [Fri, 18 Mar 2022 12:18:25 +0000 (12:18 +0000)]
media: staging/intel-ipu3: Reset imgu_video_device sequence

Reset the sequence number of imgu_video_device on stream start.
Failing to do so results in sequence number getting incremented
for consecutive stream on/off cycles.

Signed-off-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: v4l2-async: Use endpoints in __v4l2_async_nf_add_fwnode_remote()
Laurent Pinchart [Mon, 21 Mar 2022 14:51:34 +0000 (14:51 +0000)]
media: v4l2-async: Use endpoints in __v4l2_async_nf_add_fwnode_remote()

Matching on device fwnode handles is deprecated in favour of endpoint
fwnode handles. Switch the __v4l2_async_nf_add_fwnode_remote() function
to use the latter. The match code handles backward compatibility by
falling by to the device fwnode handle, so this shouldn't introduce any
regression.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: staging: media: ipu3-imgu: Request specific firmware binary
Sakari Ailus [Fri, 18 Mar 2022 16:08:12 +0000 (16:08 +0000)]
media: staging: media: ipu3-imgu: Request specific firmware binary

Primarily request a specific revision of the IPU3 firmware that the driver
is known to work with,
irci_irci_ecr-master_20161208_0213_20170112_1500.bin. Some distros only
ship this while others provide a symlink called ipu3-fw.bin, which the
driver only requested previously.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: staging: media: ipu3: Fix AF x_start position when rightmost stripe is used
Kate Hsuan [Thu, 17 Mar 2022 07:57:13 +0000 (07:57 +0000)]
media: staging: media: ipu3: Fix AF x_start position when rightmost stripe is used

For the AF configuration, if the rightmost stripe is used, the AF scene
will be at the incorrect location of the sensor.

The AF coordinate may be set to the right part of the sensor. This
configuration would lead to x_start being greater than the
down_scaled_stripes offset and the leftmost stripe would be disabled
and only the rightmost stripe is used to control the AF coordinate. If
the x_start doesn't perform any adjustments, the AF coordinate will be
at the wrong place of the sensor since down_scaled_stripes offset
would be the new zero of the coordinate system.

In this patch, if only the rightmost stripe is used, x_start should
minus down_scaled_stripes offset to maintain its correctness of AF
scene coordinate.

Signed-off-by: Kate Hsuan <hpa@redhat.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: i2c: max9286: fix kernel oops when removing module
Laurentiu Palcu [Mon, 7 Mar 2022 16:46:07 +0000 (16:46 +0000)]
media: i2c: max9286: fix kernel oops when removing module

When removing the max9286 module we get a kernel oops:

Unable to handle kernel paging request at virtual address 000000aa00000094
Mem abort info:
  ESR = 0x96000004
  EC = 0x25: DABT (current EL), IL = 32 bits
  SET = 0, FnV = 0
  EA = 0, S1PTW = 0
  FSC = 0x04: level 0 translation fault
Data abort info:
  ISV = 0, ISS = 0x00000004
  CM = 0, WnR = 0
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000880d85000
[000000aa00000094] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 96000004 [#1] PREEMPT SMP
Modules linked in: fsl_jr_uio caam_jr rng_core libdes caamkeyblob_desc caamhash_desc caamalg_desc crypto_engine max9271 authenc crct10dif_ce mxc_jpeg_encdec
CPU: 2 PID: 713 Comm: rmmod Tainted: G         C        5.15.5-00057-gaebcd29c8ed7-dirty #5
Hardware name: Freescale i.MX8QXP MEK (DT)
pstate: 80000005 (Nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : i2c_mux_del_adapters+0x24/0xf0
lr : max9286_remove+0x28/0xd0 [max9286]
sp : ffff800013a9bbf0
x29: ffff800013a9bbf0 x28: ffff00080b6da940 x27: 0000000000000000
x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000
x23: ffff000801a5b970 x22: ffff0008048b0890 x21: ffff800009297000
x20: ffff0008048b0f70 x19: 000000aa00000064 x18: 0000000000000000
x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
x14: 0000000000000014 x13: 0000000000000000 x12: ffff000802da49e8
x11: ffff000802051918 x10: ffff000802da4920 x9 : ffff000800030098
x8 : 0101010101010101 x7 : 7f7f7f7f7f7f7f7f x6 : fefefeff6364626d
x5 : 8080808000000000 x4 : 0000000000000000 x3 : 0000000000000000
x2 : ffffffffffffffff x1 : ffff00080b6da940 x0 : 0000000000000000
Call trace:
 i2c_mux_del_adapters+0x24/0xf0
 max9286_remove+0x28/0xd0 [max9286]
 i2c_device_remove+0x40/0x110
 __device_release_driver+0x188/0x234
 driver_detach+0xc4/0x150
 bus_remove_driver+0x60/0xe0
 driver_unregister+0x34/0x64
 i2c_del_driver+0x58/0xa0
 max9286_i2c_driver_exit+0x1c/0x490 [max9286]
 __arm64_sys_delete_module+0x194/0x260
 invoke_syscall+0x48/0x114
 el0_svc_common.constprop.0+0xd4/0xfc
 do_el0_svc+0x2c/0x94
 el0_svc+0x28/0x80
 el0t_64_sync_handler+0xa8/0x130
 el0t_64_sync+0x1a0/0x1a4

The Oops happens because the I2C client data does not point to
max9286_priv anymore but to v4l2_subdev. The change happened in
max9286_init() which calls v4l2_i2c_subdev_init() later on...

Besides fixing the max9286_remove() function, remove the call to
i2c_set_clientdata() in max9286_probe(), to avoid confusion, and make
the necessary changes to max9286_init() so that it doesn't have to use
i2c_get_clientdata() in order to fetch the pointer to priv.

Fixes: 66d8c9d2422d ("media: i2c: Add MAX9286 driver")
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@oss.nxp.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: v4l: ioctl: Set bus_info in v4l_querycap()
Sakari Ailus [Sat, 22 Jan 2022 13:23:18 +0000 (13:23 +0000)]
media: v4l: ioctl: Set bus_info in v4l_querycap()

The bus_info field is set by most drivers based on the type of the device
bus as well as the name of the device. Do this in v4l_querycap() so
drivers don't need to. This keeps compatibility with non-default and silly
bus_info.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: mc: Set bus_info in media_device_init()
Sakari Ailus [Sat, 22 Jan 2022 11:31:37 +0000 (11:31 +0000)]
media: mc: Set bus_info in media_device_init()

Set bus_info field based on struct device in media_device_init() and
remove corresponding code from drivers.

Also update media_device_init() documentation: the dev field must be now
initialised before calling it.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: mc: Provide a helper for setting bus_info field
Sakari Ailus [Sat, 22 Jan 2022 11:25:57 +0000 (11:25 +0000)]
media: mc: Provide a helper for setting bus_info field

The bus_info or a similar field exists in a lot of structs, yet drivers
tend to set the value of that field by themselves in a determinable way.
Thus provide a helper for doing this. To be used in subsequent patches.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: mc: media_device_init() initialises a media_device, not media_entity
Sakari Ailus [Wed, 9 Mar 2022 09:12:01 +0000 (09:12 +0000)]
media: mc: media_device_init() initialises a media_device, not media_entity

The documentation for media_device_init() had several references to
(struct) media_entity where it should have referred to struct media_device
instead. Fix this.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: mc: Remove redundant documentation
Sakari Ailus [Thu, 27 Jan 2022 09:25:48 +0000 (09:25 +0000)]
media: mc: Remove redundant documentation

Remove redundant kerneldoc documentation in mc-device.c. The functions are
already documented in media-device.h, where non-redundant documentation is
also moved.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: ccs: Use unsigned int as index to an array
Sakari Ailus [Wed, 16 Mar 2022 07:53:56 +0000 (07:53 +0000)]
media: ccs: Use unsigned int as index to an array

Use an unsigned int to index an array instead of a signed one.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: ccs: Use %u for printing unsigned values, remove extra debug print
Sakari Ailus [Wed, 16 Mar 2022 07:51:26 +0000 (07:51 +0000)]
media: ccs: Use %u for printing unsigned values, remove extra debug print

Use %u for printing unsigned integer or u32 values. In a lot of cases %d
was being used instead.

Also remove an extra debug print --- the number of lanes is already
printed by V4L2 when parsing fwnode endpoints when dynamic debug is
enabled.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: v4l2-async: Create links during v4l2_async_match_notify()
Daniel Scally [Wed, 2 Mar 2022 22:03:04 +0000 (22:03 +0000)]
media: v4l2-async: Create links during v4l2_async_match_notify()

Upon an async fwnode match, there's some typical behaviour that the
notifier and matching subdev will want to do. For example, a notifier
representing a sensor matching to an async subdev representing its
VCM will want to create an ancillary link to expose that relationship
to userspace.

To avoid lots of code in individual drivers, try to build these links
within v4l2 core.

Signed-off-by: Daniel Scally <djrscally@gmail.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: entity: Add support for ancillary links
Daniel Scally [Wed, 2 Mar 2022 22:03:03 +0000 (22:03 +0000)]
media: entity: Add support for ancillary links

Add functions to create ancillary links, so that they don't need to
be manually created by users.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <djrscally@gmail.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: entity: Add link_type_name() helper
Daniel Scally [Wed, 2 Mar 2022 22:03:02 +0000 (22:03 +0000)]
media: entity: Add link_type_name() helper

Now we have three types of media link, printing the right name during
debug output is slightly more complicated. Add a helper function to
make it easier.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <djrscally@gmail.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: media.h: Add new media link type
Daniel Scally [Wed, 2 Mar 2022 22:03:01 +0000 (22:03 +0000)]
media: media.h: Add new media link type

To describe in the kernel the connection between devices and their
supporting peripherals (for example, a camera sensor and the vcm
driving the focusing lens for it), add a new type of media link
to introduce the concept of these ancillary links.

Add some elements to the uAPI documentation to explain the new link
type, their purpose and some aspects of their current implementation.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <djrscally@gmail.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: entity: Skip non-data links in graph iteration
Daniel Scally [Wed, 2 Mar 2022 22:03:00 +0000 (22:03 +0000)]
media: entity: Skip non-data links in graph iteration

When iterating over the media graph, don't follow links that are not
data links.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <djrscally@gmail.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: entity: skip non-data link when removing reverse links
Yunke Cao [Tue, 12 Apr 2022 06:23:13 +0000 (07:23 +0100)]
media: entity: skip non-data link when removing reverse links

The original implementation removes reverse links for any input link and
assumes the presense of sink/source.
It fails when the link is a not a data link.
media_entity_remove_links when there's an ancillary link can also fail.

We only need to remove reverse links for a data link.

Signed-off-by: Yunke Cao <yunkec@google.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: imx: imx-mipi-csis: Add support for JPEG_1X8
Jacopo Mondi [Wed, 9 Mar 2022 10:04:48 +0000 (10:04 +0000)]
media: imx: imx-mipi-csis: Add support for JPEG_1X8

Add support for MEDIA_BUS_FMT_JPEG_1X8 media bus code to the
CSIS driver.

The MEDIA_BUS_FMT_JPEG_1X8 code is mapped to the RAW8 CSI-2 Data Type,
while the CSI-2 specification suggests to use User Defined Data Type 1.

As reported in the comment, the CSIS interface captures arbitrary Data
Types by using a pixel sampling mode not supported by the IP core
connected to it on i.MX SoCs.

As some sensors, such as OV5640, support sending JPEG data on the RAW8
Data Type and capture operations work correcty with such configuration,
map MEDIA_BUS_FMT_JPEG_1X8 to Data Type 0x2a.

Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec: add optional adap_configured callback
Hans Verkuil [Thu, 3 Feb 2022 12:11:15 +0000 (12:11 +0000)]
media: cec: add optional adap_configured callback

This new optional callback is called when the adapter is fully configured
or fully unconfigured. Some drivers may have to take action when this
happens.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec: add xfer_timeout_ms field
Hans Verkuil [Sat, 13 Nov 2021 11:05:24 +0000 (11:05 +0000)]
media: cec: add xfer_timeout_ms field

Allow drivers to change the transmit timeout value, i.e. after how
long should a transmit be considered 'lost', i.e. the corresponding
cec_transmit_done_ts was never called.

Some CEC devices have their own timeout, and so this timeout value must be
longer than that hardware timeout value. If it is shorter then the
framework would consider the transmit lost, even though it is effectively
still in progress at the hardware level.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec: use call_op and check for !unregistered
Hans Verkuil [Thu, 17 Mar 2022 08:51:20 +0000 (08:51 +0000)]
media: cec: use call_op and check for !unregistered

Use call_(void_)op consistently in the CEC core framework. Ditto
for the cec pin ops. And check if !adap->devnode.unregistered before
calling each op. This avoids calls to ops when the device has been
unregistered and the underlying hardware may be gone.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec.h: add cec_msg_recv_is_rx/tx_result helpers
Hans Verkuil [Wed, 9 Mar 2022 10:55:43 +0000 (10:55 +0000)]
media: cec.h: add cec_msg_recv_is_rx/tx_result helpers

These two helper functions return true if the received message
contains the result of a previous non-blocking transmit. Either
the tx_status result (cec_msg_recv_is_tx_result) of the transmit,
or the rx_status result (cec_msg_recv_is_rx_result) of the reply
to the original transmit.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec: correctly pass on reply results
Hans Verkuil [Sat, 13 Nov 2021 11:02:36 +0000 (11:02 +0000)]
media: cec: correctly pass on reply results

The results of non-blocking transmits were not correctly communicated
to userspace.

Specifically:

1) if a non-blocking transmit was canceled, then rx_status wasn't set to 0
   as it should.
2) if the non-blocking transmit succeeded, but the corresponding reply
   never arrived (aborted or timed out), then tx_status wasn't set to 0
   as it should, and rx_status was hardcoded to ABORTED instead of the
   actual reason, such as TIMEOUT. In addition, adap->ops->received() was
   never called, so drivers that want to do message processing themselves
   would not be informed of the failed reply.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec: abort if the current transmit was canceled
Hans Verkuil [Thu, 3 Mar 2022 14:01:44 +0000 (14:01 +0000)]
media: cec: abort if the current transmit was canceled

If a transmit-in-progress was canceled, then, once the transmit
is done, mark it as aborted and refrain from retrying the transmit.

To signal this situation the new transmit_in_progress_aborted field is
set to true.

The old implementation would just set adap->transmitting to NULL and
set adap->transmit_in_progress to false, but on the hardware level
the transmit was still ongoing. However, the framework would think
the transmit was aborted, and if a new transmit was issued, then
it could overwrite the HW buffer containing the old transmit with the
new transmit, leading to garbled data on the CEC bus.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cec: call enable_adap on s_log_addrs
Hans Verkuil [Thu, 3 Mar 2022 13:55:08 +0000 (13:55 +0000)]
media: cec: call enable_adap on s_log_addrs

Don't enable/disable the adapter if the first fh is opened or the
last fh is closed, instead do this when the adapter is configured
or unconfigured, and also when we enter Monitor All or Monitor Pin
mode for the first time or we exit the Monitor All/Pin mode for the
last time.

However, if needs_hpd is true, then do this when the physical
address is set or cleared: in that case the adapter typically is
powered by the HPD, so it really is disabled when the HPD is low.
This case (needs_hpd is true) was already handled in this way, so
this wasn't changed.

The problem with the old behavior was that if the HPD goes low when
no fh is open, and a transmit was in progress, then the adapter would
be disabled, typically stopping the transmit immediately which
leaves a partial message on the bus, which isn't nice and can confuse
some adapters.

It makes much more sense to disable it only when the adapter is
unconfigured and we're not monitoring the bus, since then you really
won't be using it anymore.

To keep track of this store a CEC activation count and call adap_enable
only when it goes from 0 to 1 or back to 0.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: meson-ir-tx: remove superfluous dev_err()
Yihao Han [Thu, 10 Mar 2022 07:40:59 +0000 (07:40 +0000)]
media: meson-ir-tx: remove superfluous dev_err()

Remove dev_err() messages after platform_get_irq*() failures.
platform_get_irq() already prints an error.

Generated by: scripts/coccinelle/api/platform_get_irq.cocci

Signed-off-by: Yihao Han <hanyihao@vivo.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: dib0700_devices: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:38 +0000 (22:41 +0200)]
media: dvb-usb: dib0700_devices: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/028699ec71158dbc49d710a4259eb8cdb7f673cb.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: dibusb-mc: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:37 +0000 (22:41 +0200)]
media: dvb-usb: dibusb-mc: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/f1d4fa4960b709152ae693800c830e19a4bc1f48.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: dibusb-mb: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:36 +0000 (22:41 +0200)]
media: dvb-usb: dibusb-mb: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/a300c26ad4e9bb913e86eeaf0ec7d72b9e7d5d3e.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: vp7045: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:35 +0000 (22:41 +0200)]
media: dvb-usb: vp7045: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/0241bf842bf592dfa01b0ef4916afda396194f98.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: vp702x: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:34 +0000 (22:41 +0200)]
media: dvb-usb: vp702x: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/13bf6aab3909fae5da4c9a24c114b15e76abd146.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: umt-010: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:33 +0000 (22:41 +0200)]
media: dvb-usb: umt-010: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/5fd3c469faa115856f48037019e607edcb41d458.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: ttusb2: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:32 +0000 (22:41 +0200)]
media: dvb-usb: ttusb2: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/b447d9fd3832da5eff6267e8fe742c431f1133f2.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: technisat-usb2: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:31 +0000 (22:41 +0200)]
media: technisat-usb2: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro.

Link: https://lore.kernel.org/linux-media/c125c28aeb3d4344b632e1f99d81c433917f2a4c.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: pctv452e: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:30 +0000 (22:41 +0200)]
media: dvb-usb: pctv452e: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/283a8c6bdf9778f832b4f6acc104c06688281668.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: opera1: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:29 +0000 (22:41 +0200)]
media: dvb-usb: opera1: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/08861d80b6706ac1ed04a68959ebb78f27cb028d.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: nova-t-usb2: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:28 +0000 (22:41 +0200)]
media: dvb-usb: nova-t-usb2: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/58f1a356b7b75bbefef3aa07cd99896c446df32f.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: m920x: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:27 +0000 (22:41 +0200)]
media: dvb-usb: m920x: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/c6f6bab97c39561add54f69a75980f4d453f7c17.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: gp8psk: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:26 +0000 (22:41 +0200)]
media: dvb-usb: gp8psk: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/0d32148747df677f0c930605389c12b190c09bdf.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dw2102: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:25 +0000 (22:41 +0200)]
media: dw2102: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro, and rename some PIDs.

Link: https://lore.kernel.org/linux-media/828998ef3f0843bab4e84780e42f8f0802f57be7.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dtv5100: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:24 +0000 (22:41 +0200)]
media: dtv5100: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro.

Link: https://lore.kernel.org/linux-media/6e7183735aacf33fff86bc709a38aafb6b858dff.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: dtt200u: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:23 +0000 (22:41 +0200)]
media: dvb-usb: dtt200u: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/9fa3d1add4c58e1320dcc18578fda2d0106becda.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: digitv: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:22 +0000 (22:41 +0200)]
media: digitv: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro.

Link: https://lore.kernel.org/linux-media/68de8820a361e61c25bf7402acac71b3770ff906.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cxusb: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:21 +0000 (22:41 +0200)]
media: cxusb: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro.

Link: https://lore.kernel.org/linux-media/bd1e61664e234252de3dfac16aab8bfc35b7bcd7.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: cinergyT2-core: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:20 +0000 (22:41 +0200)]
media: cinergyT2-core: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro and an enum with the USB model supported by
this driver.

Link: https://lore.kernel.org/linux-media/4b8212adab277a2bf84ab04480eb6fd37edda74f.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: az6027: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:19 +0000 (22:41 +0200)]
media: dvb-usb: az6027: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/65b9775c39dcd21e5cb75a86e1e7b99b7d6eefcd.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: af9005: use the newer dvb-usb macros for USB device
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:18 +0000 (22:41 +0200)]
media: af9005: use the newer dvb-usb macros for USB device

In order to make the drivers under dvb-usb more homogeneous,
use the new macro.

Link: https://lore.kernel.org/linux-media/9b1749763465815af92f0a4d8f210fe170c549d5.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: a800: use an enum for the device number
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:17 +0000 (22:41 +0200)]
media: dvb-usb: a800: use an enum for the device number

The device number is currently a value that needs to be the same
on two separate tables, but the code doesn't actually enforce it,
leading to errors as boards get added or removed.

Fix it by using an enum.

Link: https://lore.kernel.org/linux-media/dc8f9ec6cc8f2e16967a61752a292c46622c01dc.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: Add helper macros for using USB VID/PID
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:16 +0000 (22:41 +0200)]
media: dvb-usb: Add helper macros for using USB VID/PID

In order to use designated initializers and to avoid avoid big lines
at the USB ID tables, define some helper macros.

Link: https://lore.kernel.org/linux-media/f82e376dea2e9b922f51a03d1e7730b03e49cc7d.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: vp702x: reference to usb ID table
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:15 +0000 (22:41 +0200)]
media: dvb-usb: vp702x: reference to usb ID table

There are two commented entries that are pointing to the wrong
places. Fix them.

Link: https://lore.kernel.org/linux-media/fe9ee24510431e6baad5244d8a27e56ce167fc36.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb: move USB IDs to dvb-usb-ids.h
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:14 +0000 (22:41 +0200)]
media: dvb-usb: move USB IDs to dvb-usb-ids.h

Almost all drivers based on dvb-usb place their USB IDs
at dvb-usb-ids.h. In order to make it more standard, place
the remaining ones also there.

Link: https://lore.kernel.org/linux-media/7b32d5383169d23082758a7b69edef2f099202f3.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agomedia: dvb-usb-ids.h: sort entries
Mauro Carvalho Chehab [Mon, 28 Mar 2022 20:41:13 +0000 (22:41 +0200)]
media: dvb-usb-ids.h: sort entries

The entries there are alphabetically sorted, but some are at the
wrong place. Re-sort them.

While here, replace spaces by tabs where needed.

Link: https://lore.kernel.org/linux-media/0208dbba189b754b999759f06c2584242c879f4d.1648499509.git.mchehab@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
2 years agoLinux 5.18-rc3
Linus Torvalds [Sun, 17 Apr 2022 20:57:31 +0000 (13:57 -0700)]
Linux 5.18-rc3

2 years agoMerge tag 'for-linus-5.18-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 17 Apr 2022 17:29:10 +0000 (10:29 -0700)]
Merge tag 'for-linus-5.18-rc3-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fixlet from Juergen Gross:
 "A single cleanup patch for the Xen balloon driver"

* tag 'for-linus-5.18-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/balloon: don't use PV mode extra memory for zone device allocations

2 years agoMerge tag 'x86-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 17 Apr 2022 16:55:59 +0000 (09:55 -0700)]
Merge tag 'x86-urgent-2022-04-17' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Thomas Gleixner:
 "Two x86 fixes related to TSX:

   - Use either MSR_TSX_FORCE_ABORT or MSR_IA32_TSX_CTRL to disable TSX
     to cover all CPUs which allow to disable it.

   - Disable TSX development mode at boot so that a microcode update
     which provides TSX development mode does not suddenly make the
     system vulnerable to TSX Asynchronous Abort"

* tag 'x86-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/tsx: Disable TSX development mode at boot
  x86/tsx: Use MSR_TSX_CTRL to clear CPUID bits

2 years agoMerge tag 'timers-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 17 Apr 2022 16:53:01 +0000 (09:53 -0700)]
Merge tag 'timers-urgent-2022-04-17' of git://git./linux/kernel/git/tip/tip

Pull timer fixes from Thomas Gleixner:
 "A small set of fixes for the timers core:

   - Fix the warning condition in __run_timers() which does not take
     into account that a CPU base (especially the deferrable base) never
     has a timer armed on it and therefore the next_expiry value can
     become stale.

   - Replace a WARN_ON() in the NOHZ code with a WARN_ON_ONCE() to
     prevent endless spam in dmesg.

   - Remove the double star from a comment which is not meant to be in
     kernel-doc format"

* tag 'timers-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tick/sched: Fix non-kernel-doc comment
  tick/nohz: Use WARN_ON_ONCE() to prevent console saturation
  timers: Fix warning condition in __run_timers()

2 years agoMerge tag 'smp-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 17 Apr 2022 16:46:15 +0000 (09:46 -0700)]
Merge tag 'smp-urgent-2022-04-17' of git://git./linux/kernel/git/tip/tip

Pull SMP fixes from Thomas Gleixner:
 "Two fixes for the SMP core:

   - Make the warning condition in flush_smp_call_function_queue()
     correct, which checked a just emptied list head for being empty
     instead of validating that there was no pending entry on the
     offlined CPU at all.

   - The @cpu member of struct cpuhp_cpu_state is initialized when the
     CPU hotplug thread for the upcoming CPU is created. That's too late
     because the creation of the thread can fail and then the following
     rollback operates on CPU0. Get rid of the CPU member and hand the
     CPU number to the involved functions directly"

* tag 'smp-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  cpu/hotplug: Remove the 'cpu' member of cpuhp_cpu_state
  smp: Fix offline cpu check in flush_smp_call_function_queue()

2 years agoMerge tag 'irq-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 17 Apr 2022 16:42:03 +0000 (09:42 -0700)]
Merge tag 'irq-urgent-2022-04-17' of git://git./linux/kernel/git/tip/tip

Pull irq fix from Thomas Gleixner:
 "A single fix for the interrupt affinity spreading logic to take into
  account that there can be an imbalance between present and possible
  CPUs, which causes already assigned bits to be overwritten"

* tag 'irq-urgent-2022-04-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq/affinity: Consider that CPUs on nodes can be unbalanced

2 years agoMerge tag 'for-v5.18-rc' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux...
Linus Torvalds [Sun, 17 Apr 2022 16:36:27 +0000 (09:36 -0700)]
Merge tag 'for-v5.18-rc' of git://git./linux/kernel/git/sre/linux-power-supply

Pull power supply fixes from Sebastian Reichel:

 - Fix a regression with battery data failing to load from DT

* tag 'for-v5.18-rc' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply:
  power: supply: Reset err after not finding static battery
  power: supply: samsung-sdi-battery: Add missing charge restart voltages

2 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sun, 17 Apr 2022 16:31:27 +0000 (09:31 -0700)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Regular set of fixes for drivers and the dev-interface"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: ismt: Fix undefined behavior due to shift overflowing the constant
  i2c: dev: Force case user pointers in compat_i2cdev_ioctl()
  i2c: dev: check return value when calling dev_set_name()
  i2c: qcom-geni: Use dev_err_probe() for GPI DMA error
  i2c: imx: Implement errata ERR007805 or e7805 bus frequency limit
  i2c: pasemi: Wait for write xfers to finish

2 years agoMerge tag 'devicetree-fixes-for-5.18-2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 17 Apr 2022 00:07:50 +0000 (17:07 -0700)]
Merge tag 'devicetree-fixes-for-5.18-2' of git://git./linux/kernel/git/robh/linux

Pull devicetree fixes from Rob Herring:

 - Fix scalar property schemas with array constraints

 - Fix 'enum' lists with duplicate entries

 - Fix incomplete if/then/else schemas

 - Add Renesas RZ/V2L SoC support to Mali Bifrost binding

 - Maintainers update for Marvell irqchip

* tag 'devicetree-fixes-for-5.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: display: panel-timing: Define a single type for properties
  dt-bindings: Fix array constraints on scalar properties
  dt-bindings: gpu: mali-bifrost: Document RZ/V2L SoC
  dt-bindings: net: snps: remove duplicate name
  dt-bindings: Fix 'enum' lists with duplicate entries
  dt-bindings: irqchip: mrvl,intc: refresh maintainers
  dt-bindings: Fix incomplete if/then/else schemas
  dt-bindings: power: renesas,apmu: Fix cpus property limits
  dt-bindings: extcon: maxim,max77843: fix ports type

2 years agoMerge tag 'gpio-fixes-for-v5.18-rc3' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 17 Apr 2022 00:01:43 +0000 (17:01 -0700)]
Merge tag 'gpio-fixes-for-v5.18-rc3' of git://git./linux/kernel/git/brgl/linux

Pull gpio fixes from Bartosz Golaszewski:
 "A single fix for gpio-sim and two patches for GPIO ACPI pulled from
  Andy:

   - fix the set/get_multiple() callbacks in gpio-sim

   - use correct format characters in gpiolib-acpi

   - use an unsigned type for pins in gpiolib-acpi"

* tag 'gpio-fixes-for-v5.18-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpio: sim: fix setting and getting multiple lines
  gpiolib: acpi: Convert type for pin to be unsigned
  gpiolib: acpi: use correct format characters

2 years agoMerge tag 'soc-fixes-5.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Linus Torvalds [Sat, 16 Apr 2022 23:51:39 +0000 (16:51 -0700)]
Merge tag 'soc-fixes-5.18-2' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Arnd Bergmann:
 "There are a number of SoC bugfixes that came in since the merge
  window, and more of them are already pending.

  This batch includes:

   - A boot time regression fix for davinci that triggered on
     multi_v5_defconfig when booting any platform

   - Defconfig updates to address removed features, changed symbol names
     or dependencies, for gemini, ux500, and pxa

   - Email address changes for Krzysztof Kozlowski

   - Build warning fixes for ep93xx and iop32x

   - Devicetree warning fixes across many platforms

   - Minor bugfixes for the reset controller, memory controller and SCMI
     firmware subsystems plus the versatile-express board"

* tag 'soc-fixes-5.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (34 commits)
  ARM: config: Update Gemini defconfig
  arm64: dts: qcom/sdm845-shift-axolotl: Fix boolean properties with values
  ARM: dts: align SPI NOR node name with dtschema
  ARM: dts: Fix more boolean properties with values
  arm/arm64: dts: qcom: Fix boolean properties with values
  arm64: dts: imx: Fix imx8*-var-som touchscreen property sizes
  arm: dts: imx: Fix boolean properties with values
  arm64: dts: tegra: Fix boolean properties with values
  arm: dts: at91: Fix boolean properties with values
  arm: configs: imote2: Drop defconfig as board support dropped.
  ep93xx: clock: Don't use plain integer as NULL pointer
  ep93xx: clock: Fix UAF in ep93xx_clk_register_gate()
  ARM: vexpress/spc: Fix all the kernel-doc build warnings
  ARM: vexpress/spc: Fix kernel-doc build warning for ve_spc_cpu_in_wfi
  ARM: config: u8500: Re-enable AB8500 battery charging
  ARM: config: u8500: Add some common hardware
  memory: fsl_ifc: populate child nodes of buses and mfd devices
  ARM: config: Refresh U8500 defconfig
  firmware: arm_scmi: Fix sparse warnings in OPTEE transport driver
  firmware: arm_scmi: Replace zero-length array with flexible-array member
  ...

2 years agoMerge tag 'random-5.18-rc3-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 16 Apr 2022 23:42:53 +0000 (16:42 -0700)]
Merge tag 'random-5.18-rc3-for-linus' of git://git./linux/kernel/git/crng/random

Pull random number generator fixes from Jason Donenfeld:

 - Per your suggestion, random reads now won't fail if there's a page
   fault after some non-zero amount of data has been read, which makes
   the behavior consistent with all other reads in the kernel.

 - Rather than an inconsistent mix of random_get_entropy() returning an
   unsigned long or a cycles_t, now it just returns an unsigned long.

 - A memcpy() was replaced with an memmove(), because the addresses are
   sometimes overlapping. In practice the destination is always before
   the source, so not really an issue, but better to be correct than
   not.

* tag 'random-5.18-rc3-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random:
  random: use memmove instead of memcpy for remaining 32 bytes
  random: make random_get_entropy() return an unsigned long
  random: allow partial reads if later user copies fail

2 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sat, 16 Apr 2022 20:38:26 +0000 (13:38 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "13 fixes, all in drivers.

  The most extensive changes are in the iscsi series (affecting drivers
  qedi, cxgbi and bnx2i), the next most is scsi_debug, but that's just a
  simple revert and then minor updates to pm80xx"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: iscsi: MAINTAINERS: Add Mike Christie as co-maintainer
  scsi: qedi: Fix failed disconnect handling
  scsi: iscsi: Fix NOP handling during conn recovery
  scsi: iscsi: Merge suspend fields
  scsi: iscsi: Fix unbound endpoint error handling
  scsi: iscsi: Fix conn cleanup and stop race during iscsid restart
  scsi: iscsi: Fix endpoint reuse regression
  scsi: iscsi: Release endpoint ID when its freed
  scsi: iscsi: Fix offload conn cleanup when iscsid restarts
  scsi: iscsi: Move iscsi_ep_disconnect()
  scsi: pm80xx: Enable upper inbound, outbound queues
  scsi: pm80xx: Mask and unmask upper interrupt vectors 32-63
  Revert "scsi: scsi_debug: Address races following module load"

2 years agoMerge tag 'intel-gpio-v5.18-2' of gitolite.kernel.org:pub/scm/linux/kernel/git/andy...
Bartosz Golaszewski [Sat, 16 Apr 2022 19:57:00 +0000 (21:57 +0200)]
Merge tag 'intel-gpio-v5.18-2' of gitolite.pub/scm/linux/kernel/git/andy/linux-gpio-intel into gpio/for-current

intel-gpio for v5.18-2

* Couple of fixes related to handling unsigned value of the pin from ACPI

gpiolib:
 -  acpi: Convert type for pin to be unsigned
 -  acpi: use correct format characters

2 years agoMerge tag 'dma-mapping-5.18-2' of git://git.infradead.org/users/hch/dma-mapping
Linus Torvalds [Sat, 16 Apr 2022 18:20:21 +0000 (11:20 -0700)]
Merge tag 'dma-mapping-5.18-2' of git://git.infradead.org/users/hch/dma-mapping

Pull dma-mapping fix from Christoph Hellwig:

 - avoid a double memory copy for swiotlb (Chao Gao)

* tag 'dma-mapping-5.18-2' of git://git.infradead.org/users/hch/dma-mapping:
  dma-direct: avoid redundant memory sync for swiotlb

2 years agorandom: use memmove instead of memcpy for remaining 32 bytes
Jason A. Donenfeld [Wed, 13 Apr 2022 23:50:38 +0000 (01:50 +0200)]
random: use memmove instead of memcpy for remaining 32 bytes

In order to immediately overwrite the old key on the stack, before
servicing a userspace request for bytes, we use the remaining 32 bytes
of block 0 as the key. This means moving indices 8,9,a,b,c,d,e,f ->
4,5,6,7,8,9,a,b. Since 4 < 8, for the kernel implementations of
memcpy(), this doesn't actually appear to be a problem in practice. But
relying on that characteristic seems a bit brittle. So let's change that
to a proper memmove(), which is the by-the-books way of handling
overlapping memory copies.

Reviewed-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
2 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Fri, 15 Apr 2022 22:57:18 +0000 (15:57 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge misc fixes from Andrew Morton:
 "14 patches.

  Subsystems affected by this patch series: MAINTAINERS, binfmt, and
  mm (tmpfs, secretmem, kasan, kfence, pagealloc, zram, compaction,
  hugetlb, vmalloc, and kmemleak)"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  mm: kmemleak: take a full lowmem check in kmemleak_*_phys()
  mm/vmalloc: fix spinning drain_vmap_work after reading from /proc/vmcore
  revert "fs/binfmt_elf: use PT_LOAD p_align values for static PIE"
  revert "fs/binfmt_elf: fix PT_LOAD p_align values for loaders"
  hugetlb: do not demote poisoned hugetlb pages
  mm: compaction: fix compiler warning when CONFIG_COMPACTION=n
  mm: fix unexpected zeroed page mapping with zram swap
  mm, page_alloc: fix build_zonerefs_node()
  mm, kfence: support kmem_dump_obj() for KFENCE objects
  kasan: fix hw tags enablement when KUNIT tests are disabled
  irq_work: use kasan_record_aux_stack_noalloc() record callstack
  mm/secretmem: fix panic when growing a memfd_secret
  tmpfs: fix regressions from wider use of ZERO_PAGE
  MAINTAINERS: Broadcom internal lists aren't maintainers

2 years agoMerge tag 'for-5.18/dm-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 15 Apr 2022 22:20:59 +0000 (15:20 -0700)]
Merge tag 'for-5.18/dm-fixes-2' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull device mapper fixes from Mike Snitzer:

 - Fix memory corruption in DM integrity target when tag_size is less
   than digest size.

 - Fix DM multipath's historical-service-time path selector to not use
   sched_clock() and ktime_get_ns(); only use ktime_get_ns().

 - Fix dm_io->orig_bio NULL pointer dereference in dm_zone_map_bio() due
   to 5.18 changes that overlooked DM zone's use of ->orig_bio

 - Fix for regression that broke the use of dm_accept_partial_bio() for
   "abnormal" IO (e.g. WRITE ZEROES) that does not need duplicate bios

 - Fix DM's issuing of empty flush bio so that it's size is 0.

* tag 'for-5.18/dm-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm: fix bio length of empty flush
  dm: allow dm_accept_partial_bio() for dm_io without duplicate bios
  dm zone: fix NULL pointer dereference in dm_zone_map_bio
  dm mpath: only use ktime_get_ns() in historical selector
  dm integrity: fix memory corruption when tag_size is less than digest size

2 years agomm: kmemleak: take a full lowmem check in kmemleak_*_phys()
Patrick Wang [Fri, 15 Apr 2022 02:14:04 +0000 (19:14 -0700)]
mm: kmemleak: take a full lowmem check in kmemleak_*_phys()

The kmemleak_*_phys() apis do not check the address for lowmem's min
boundary, while the caller may pass an address below lowmem, which will
trigger an oops:

  # echo scan > /sys/kernel/debug/kmemleak
  Unable to handle kernel paging request at virtual address ff5fffffffe00000
  Oops [#1]
  Modules linked in:
  CPU: 2 PID: 134 Comm: bash Not tainted 5.18.0-rc1-next-20220407 #33
  Hardware name: riscv-virtio,qemu (DT)
  epc : scan_block+0x74/0x15c
   ra : scan_block+0x72/0x15c
  epc : ffffffff801e5806 ra : ffffffff801e5804 sp : ff200000104abc30
   gp : ffffffff815cd4e8 tp : ff60000004cfa340 t0 : 0000000000000200
   t1 : 00aaaaaac23954cc t2 : 00000000000003ff s0 : ff200000104abc90
   s1 : ffffffff81b0ff28 a0 : 0000000000000000 a1 : ff5fffffffe01000
   a2 : ffffffff81b0ff28 a3 : 0000000000000002 a4 : 0000000000000001
   a5 : 0000000000000000 a6 : ff200000104abd7c a7 : 0000000000000005
   s2 : ff5fffffffe00ff9 s3 : ffffffff815cd998 s4 : ffffffff815d0e90
   s5 : ffffffff81b0ff28 s6 : 0000000000000020 s7 : ffffffff815d0eb0
   s8 : ffffffffffffffff s9 : ff5fffffffe00000 s10: ff5fffffffe01000
   s11: 0000000000000022 t3 : 00ffffffaa17db4c t4 : 000000000000000f
   t5 : 0000000000000001 t6 : 0000000000000000
  status: 0000000000000100 badaddr: ff5fffffffe00000 cause: 000000000000000d
    scan_gray_list+0x12e/0x1a6
    kmemleak_scan+0x2aa/0x57e
    kmemleak_write+0x32a/0x40c
    full_proxy_write+0x56/0x82
    vfs_write+0xa6/0x2a6
    ksys_write+0x6c/0xe2
    sys_write+0x22/0x2a
    ret_from_syscall+0x0/0x2

The callers may not quite know the actual address they pass(e.g. from
devicetree).  So the kmemleak_*_phys() apis should guarantee the address
they finally use is in lowmem range, so check the address for lowmem's
min boundary.

Link: https://lkml.kernel.org/r/20220413122925.33856-1-patrick.wang.shcn@gmail.com
Signed-off-by: Patrick Wang <patrick.wang.shcn@gmail.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm/vmalloc: fix spinning drain_vmap_work after reading from /proc/vmcore
Omar Sandoval [Fri, 15 Apr 2022 02:14:01 +0000 (19:14 -0700)]
mm/vmalloc: fix spinning drain_vmap_work after reading from /proc/vmcore

Commit 3ee48b6af49c ("mm, x86: Saving vmcore with non-lazy freeing of
vmas") introduced set_iounmap_nonlazy(), which sets vmap_lazy_nr to
lazy_max_pages() + 1, ensuring that any future vunmaps() immediately
purge the vmap areas instead of doing it lazily.

Commit 690467c81b1a ("mm/vmalloc: Move draining areas out of caller
context") moved the purging from the vunmap() caller to a worker thread.
Unfortunately, set_iounmap_nonlazy() can cause the worker thread to spin
(possibly forever).  For example, consider the following scenario:

 1. Thread reads from /proc/vmcore. This eventually calls
    __copy_oldmem_page() -> set_iounmap_nonlazy(), which sets
    vmap_lazy_nr to lazy_max_pages() + 1.

 2. Then it calls free_vmap_area_noflush() (via iounmap()), which adds 2
    pages (one page plus the guard page) to the purge list and
    vmap_lazy_nr. vmap_lazy_nr is now lazy_max_pages() + 3, so the
    drain_vmap_work is scheduled.

 3. Thread returns from the kernel and is scheduled out.

 4. Worker thread is scheduled in and calls drain_vmap_area_work(). It
    frees the 2 pages on the purge list. vmap_lazy_nr is now
    lazy_max_pages() + 1.

 5. This is still over the threshold, so it tries to purge areas again,
    but doesn't find anything.

 6. Repeat 5.

If the system is running with only one CPU (which is typicial for kdump)
and preemption is disabled, then this will never make forward progress:
there aren't any more pages to purge, so it hangs.  If there is more
than one CPU or preemption is enabled, then the worker thread will spin
forever in the background.  (Note that if there were already pages to be
purged at the time that set_iounmap_nonlazy() was called, this bug is
avoided.)

This can be reproduced with anything that reads from /proc/vmcore
multiple times.  E.g., vmcore-dmesg /proc/vmcore.

It turns out that improvements to vmap() over the years have obsoleted
the need for this "optimization".  I benchmarked `dd if=/proc/vmcore
of=/dev/null` with 4k and 1M read sizes on a system with a 32GB vmcore.
The test was run on 5.17, 5.18-rc1 with a fix that avoided the hang, and
5.18-rc1 with set_iounmap_nonlazy() removed entirely:

    |5.17  |5.18+fix|5.18+removal
  4k|40.86s|  40.09s|      26.73s
  1M|24.47s|  23.98s|      21.84s

The removal was the fastest (by a wide margin with 4k reads).  This
patch removes set_iounmap_nonlazy().

Link: https://lkml.kernel.org/r/52f819991051f9b865e9ce25605509bfdbacadcd.1649277321.git.osandov@fb.com
Fixes: 690467c81b1a  ("mm/vmalloc: Move draining areas out of caller context")
Signed-off-by: Omar Sandoval <osandov@fb.com>
Acked-by: Chris Down <chris@chrisdown.name>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorevert "fs/binfmt_elf: use PT_LOAD p_align values for static PIE"
Andrew Morton [Fri, 15 Apr 2022 02:13:58 +0000 (19:13 -0700)]
revert "fs/binfmt_elf: use PT_LOAD p_align values for static PIE"

Despite Mike's attempted fix (925346c129da117122), regressions reports
continue:

  https://lore.kernel.org/lkml/cb5b81bd-9882-e5dc-cd22-54bdbaaefbbc@leemhuis.info/
  https://bugzilla.kernel.org/show_bug.cgi?id=215720
  https://lkml.kernel.org/r/b685f3d0-da34-531d-1aa9-479accd3e21b@leemhuis.info

So revert this patch.

Fixes: 9630f0d60fec ("fs/binfmt_elf: use PT_LOAD p_align values for static PIE")
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Chris Kennelly <ckennelly@google.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Fangrui Song <maskray@google.com>
Cc: H.J. Lu <hjl.tools@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Cc: Sandeep Patil <sspatil@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Song Liu <songliubraving@fb.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorevert "fs/binfmt_elf: fix PT_LOAD p_align values for loaders"
Andrew Morton [Fri, 15 Apr 2022 02:13:55 +0000 (19:13 -0700)]
revert "fs/binfmt_elf: fix PT_LOAD p_align values for loaders"

Commit 925346c129da11 ("fs/binfmt_elf: fix PT_LOAD p_align values for
loaders") was an attempt to fix regressions due to 9630f0d60fec5f
("fs/binfmt_elf: use PT_LOAD p_align values for static PIE").

But regressionss continue to be reported:

  https://lore.kernel.org/lkml/cb5b81bd-9882-e5dc-cd22-54bdbaaefbbc@leemhuis.info/
  https://bugzilla.kernel.org/show_bug.cgi?id=215720
  https://lkml.kernel.org/r/b685f3d0-da34-531d-1aa9-479accd3e21b@leemhuis.info

This patch reverts the fix, so the original can also be reverted.

Fixes: 925346c129da11 ("fs/binfmt_elf: fix PT_LOAD p_align values for loaders")
Cc: H.J. Lu <hjl.tools@gmail.com>
Cc: Chris Kennelly <ckennelly@google.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Song Liu <songliubraving@fb.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Sandeep Patil <sspatil@google.com>
Cc: Fangrui Song <maskray@google.com>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agohugetlb: do not demote poisoned hugetlb pages
Mike Kravetz [Fri, 15 Apr 2022 02:13:52 +0000 (19:13 -0700)]
hugetlb: do not demote poisoned hugetlb pages

It is possible for poisoned hugetlb pages to reside on the free lists.
The huge page allocation routines which dequeue entries from the free
lists make a point of avoiding poisoned pages.  There is no such check
and avoidance in the demote code path.

If a hugetlb page on the is on a free list, poison will only be set in
the head page rather then the page with the actual error.  If such a
page is demoted, then the poison flag may follow the wrong page.  A page
without error could have poison set, and a page with poison could not
have the flag set.

Check for poison before attempting to demote a hugetlb page.  Also,
return -EBUSY to the caller if only poisoned pages are on the free list.

Link: https://lkml.kernel.org/r/20220307215707.50916-1-mike.kravetz@oracle.com
Fixes: 8531fc6f52f5 ("hugetlb: add hugetlb demote page support")
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Reviewed-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm: compaction: fix compiler warning when CONFIG_COMPACTION=n
Charan Teja Kalla [Fri, 15 Apr 2022 02:13:49 +0000 (19:13 -0700)]
mm: compaction: fix compiler warning when CONFIG_COMPACTION=n

The below warning is reported when CONFIG_COMPACTION=n:

   mm/compaction.c:56:27: warning: 'HPAGE_FRAG_CHECK_INTERVAL_MSEC' defined but not used [-Wunused-const-variable=]
      56 | static const unsigned int HPAGE_FRAG_CHECK_INTERVAL_MSEC = 500;
         |                           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Fix it by moving 'HPAGE_FRAG_CHECK_INTERVAL_MSEC' under
CONFIG_COMPACTION defconfig.

Also since this is just a 'static const int' type, use #define for it.

Link: https://lkml.kernel.org/r/1647608518-20924-1-git-send-email-quic_charante@quicinc.com
Signed-off-by: Charan Teja Kalla <quic_charante@quicinc.com>
Reported-by: kernel test robot <lkp@intel.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Nitin Gupta <nigupta@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm: fix unexpected zeroed page mapping with zram swap
Minchan Kim [Fri, 15 Apr 2022 02:13:46 +0000 (19:13 -0700)]
mm: fix unexpected zeroed page mapping with zram swap

Two processes under CLONE_VM cloning, user process can be corrupted by
seeing zeroed page unexpectedly.

      CPU A                        CPU B

  do_swap_page                do_swap_page
  SWP_SYNCHRONOUS_IO path     SWP_SYNCHRONOUS_IO path
  swap_readpage valid data
    swap_slot_free_notify
      delete zram entry
                              swap_readpage zeroed(invalid) data
                              pte_lock
                              map the *zero data* to userspace
                              pte_unlock
  pte_lock
  if (!pte_same)
    goto out_nomap;
  pte_unlock
  return and next refault will
  read zeroed data

The swap_slot_free_notify is bogus for CLONE_VM case since it doesn't
increase the refcount of swap slot at copy_mm so it couldn't catch up
whether it's safe or not to discard data from backing device.  In the
case, only the lock it could rely on to synchronize swap slot freeing is
page table lock.  Thus, this patch gets rid of the swap_slot_free_notify
function.  With this patch, CPU A will see correct data.

      CPU A                        CPU B

  do_swap_page                do_swap_page
  SWP_SYNCHRONOUS_IO path     SWP_SYNCHRONOUS_IO path
                              swap_readpage original data
                              pte_lock
                              map the original data
                              swap_free
                                swap_range_free
                                  bd_disk->fops->swap_slot_free_notify
  swap_readpage read zeroed data
                              pte_unlock
  pte_lock
  if (!pte_same)
    goto out_nomap;
  pte_unlock
  return
  on next refault will see mapped data by CPU B

The concern of the patch would increase memory consumption since it
could keep wasted memory with compressed form in zram as well as
uncompressed form in address space.  However, most of cases of zram uses
no readahead and do_swap_page is followed by swap_free so it will free
the compressed form from in zram quickly.

Link: https://lkml.kernel.org/r/YjTVVxIAsnKAXjTd@google.com
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Reported-by: Ivan Babrou <ivan@cloudflare.com>
Tested-by: Ivan Babrou <ivan@cloudflare.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
Cc: Nitin Gupta <ngupta@vflare.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: David Hildenbrand <david@redhat.com>
Cc: <stable@vger.kernel.org> [4.14+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm, page_alloc: fix build_zonerefs_node()
Juergen Gross [Fri, 15 Apr 2022 02:13:43 +0000 (19:13 -0700)]
mm, page_alloc: fix build_zonerefs_node()

Since commit 6aa303defb74 ("mm, vmscan: only allocate and reclaim from
zones with pages managed by the buddy allocator") only zones with free
memory are included in a built zonelist.  This is problematic when e.g.
all memory of a zone has been ballooned out when zonelists are being
rebuilt.

The decision whether to rebuild the zonelists when onlining new memory
is done based on populated_zone() returning 0 for the zone the memory
will be added to.  The new zone is added to the zonelists only, if it
has free memory pages (managed_zone() returns a non-zero value) after
the memory has been onlined.  This implies, that onlining memory will
always free the added pages to the allocator immediately, but this is
not true in all cases: when e.g. running as a Xen guest the onlined new
memory will be added only to the ballooned memory list, it will be freed
only when the guest is being ballooned up afterwards.

Another problem with using managed_zone() for the decision whether a
zone is being added to the zonelists is, that a zone with all memory
used will in fact be removed from all zonelists in case the zonelists
happen to be rebuilt.

Use populated_zone() when building a zonelist as it has been done before
that commit.

There was a report that QubesOS (based on Xen) is hitting this problem.
Xen has switched to use the zone device functionality in kernel 5.9 and
QubesOS wants to use memory hotplugging for guests in order to be able
to start a guest with minimal memory and expand it as needed.  This was
the report leading to the patch.

Link: https://lkml.kernel.org/r/20220407120637.9035-1-jgross@suse.com
Fixes: 6aa303defb74 ("mm, vmscan: only allocate and reclaim from zones with pages managed by the buddy allocator")
Signed-off-by: Juergen Gross <jgross@suse.com>
Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Reviewed-by: Wei Yang <richard.weiyang@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm, kfence: support kmem_dump_obj() for KFENCE objects
Marco Elver [Fri, 15 Apr 2022 02:13:40 +0000 (19:13 -0700)]
mm, kfence: support kmem_dump_obj() for KFENCE objects

Calling kmem_obj_info() via kmem_dump_obj() on KFENCE objects has been
producing garbage data due to the object not actually being maintained
by SLAB or SLUB.

Fix this by implementing __kfence_obj_info() that copies relevant
information to struct kmem_obj_info when the object was allocated by
KFENCE; this is called by a common kmem_obj_info(), which also calls the
slab/slub/slob specific variant now called __kmem_obj_info().

For completeness, kmem_dump_obj() now displays if the object was
allocated by KFENCE.

Link: https://lore.kernel.org/all/20220323090520.GG16885@xsang-OptiPlex-9020/
Link: https://lkml.kernel.org/r/20220406131558.3558585-1-elver@google.com
Fixes: b89fb5ef0ce6 ("mm, kfence: insert KFENCE hooks for SLUB")
Fixes: d3fb45f370d9 ("mm, kfence: insert KFENCE hooks for SLAB")
Signed-off-by: Marco Elver <elver@google.com>
Reviewed-by: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz> [slab]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokasan: fix hw tags enablement when KUNIT tests are disabled
Vincenzo Frascino [Fri, 15 Apr 2022 02:13:37 +0000 (19:13 -0700)]
kasan: fix hw tags enablement when KUNIT tests are disabled

Kasan enables hw tags via kasan_enable_tagging() which based on the mode
passed via kernel command line selects the correct hw backend.
kasan_enable_tagging() is meant to be invoked indirectly via the cpu
features framework of the architectures that support these backends.
Currently the invocation of this function is guarded by
CONFIG_KASAN_KUNIT_TEST which allows the enablement of the correct backend
only when KUNIT tests are enabled in the kernel.

This inconsistency was introduced in commit:

  ed6d74446cbf ("kasan: test: support async (again) and asymm modes for HW_TAGS")

... and prevents to enable MTE on arm64 when KUNIT tests for kasan hw_tags are
disabled.

Fix the issue making sure that the CONFIG_KASAN_KUNIT_TEST guard does not
prevent the correct invocation of kasan_enable_tagging().

Link: https://lkml.kernel.org/r/20220408124323.10028-1-vincenzo.frascino@arm.com
Fixes: ed6d74446cbf ("kasan: test: support async (again) and asymm modes for HW_TAGS")
Signed-off-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoirq_work: use kasan_record_aux_stack_noalloc() record callstack
Zqiang [Fri, 15 Apr 2022 02:13:34 +0000 (19:13 -0700)]
irq_work: use kasan_record_aux_stack_noalloc() record callstack

On PREEMPT_RT kernel and KASAN is enabled.  the kasan_record_aux_stack()
may call alloc_pages(), and the rt-spinlock will be acquired, if currently
in atomic context, will trigger warning:

  BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:46
  in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 239, name: bootlogd
  Preemption disabled at:
  [<ffffffffbab1a531>] rt_mutex_slowunlock+0xa1/0x4e0
  CPU: 3 PID: 239 Comm: bootlogd Tainted: G        W 5.17.1-rt17-yocto-preempt-rt+ #105
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014
  Call Trace:
     __might_resched.cold+0x13b/0x173
     rt_spin_lock+0x5b/0xf0
     get_page_from_freelist+0x20c/0x1610
     __alloc_pages+0x25e/0x5e0
     __stack_depot_save+0x3c0/0x4a0
     kasan_save_stack+0x3a/0x50
     __kasan_record_aux_stack+0xb6/0xc0
     kasan_record_aux_stack+0xe/0x10
     irq_work_queue_on+0x6a/0x1c0
     pull_rt_task+0x631/0x6b0
     do_balance_callbacks+0x56/0x80
     __balance_callbacks+0x63/0x90
     rt_mutex_setprio+0x349/0x880
     rt_mutex_slowunlock+0x22a/0x4e0
     rt_spin_unlock+0x49/0x80
     uart_write+0x186/0x2b0
     do_output_char+0x2e9/0x3a0
     n_tty_write+0x306/0x800
     file_tty_write.isra.0+0x2af/0x450
     tty_write+0x22/0x30
     new_sync_write+0x27c/0x3a0
     vfs_write+0x3f7/0x5d0
     ksys_write+0xd9/0x180
     __x64_sys_write+0x43/0x50
     do_syscall_64+0x44/0x90
     entry_SYSCALL_64_after_hwframe+0x44/0xae

Fix it by using kasan_record_aux_stack_noalloc() to avoid the call to
alloc_pages().

Link: https://lkml.kernel.org/r/20220402142555.2699582-1-qiang1.zhang@intel.com
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm/secretmem: fix panic when growing a memfd_secret
Axel Rasmussen [Fri, 15 Apr 2022 02:13:31 +0000 (19:13 -0700)]
mm/secretmem: fix panic when growing a memfd_secret

When one tries to grow an existing memfd_secret with ftruncate, one gets
a panic [1].  For example, doing the following reliably induces the
panic:

    fd = memfd_secret();

    ftruncate(fd, 10);
    ptr = mmap(NULL, 10, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0);
    strcpy(ptr, "123456789");

    munmap(ptr, 10);
    ftruncate(fd, 20);

The basic reason for this is, when we grow with ftruncate, we call down
into simple_setattr, and then truncate_inode_pages_range, and eventually
we try to zero part of the memory.  The normal truncation code does this
via the direct map (i.e., it calls page_address() and hands that to
memset()).

For memfd_secret though, we specifically don't map our pages via the
direct map (i.e.  we call set_direct_map_invalid_noflush() on every
fault).  So the address returned by page_address() isn't useful, and
when we try to memset() with it we panic.

This patch avoids the panic by implementing a custom setattr for
memfd_secret, which detects resizes specifically (setting the size for
the first time works just fine, since there are no existing pages to try
to zero), and rejects them with EINVAL.

One could argue growing should be supported, but I think that will
require a significantly more lengthy change.  So, I propose a minimal
fix for the benefit of stable kernels, and then perhaps to extend
memfd_secret to support growing in a separate patch.

[1]:

  BUG: unable to handle page fault for address: ffffa0a889277028
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD afa01067 P4D afa01067 PUD 83f909067 PMD 83f8bf067 PTE 800ffffef6d88060
  Oops: 0002 [#1] PREEMPT SMP DEBUG_PAGEALLOC PTI
  CPU: 0 PID: 281 Comm: repro Not tainted 5.17.0-dbg-DEV #1
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
  RIP: 0010:memset_erms+0x9/0x10
  Code: c1 e9 03 40 0f b6 f6 48 b8 01 01 01 01 01 01 01 01 48 0f af c6 f3 48 ab 89 d1 f3 aa 4c 89 c8 c3 90 49 89 f9 40 88 f0 48 89 d1 <f3> aa 4c 89 c8 c3 90 49 89 fa 40 0f b6 ce 48 b8 01 01 01 01 01 01
  RSP: 0018:ffffb932c09afbf0 EFLAGS: 00010246
  RAX: 0000000000000000 RBX: ffffda63c4249dc0 RCX: 0000000000000fd8
  RDX: 0000000000000fd8 RSI: 0000000000000000 RDI: ffffa0a889277028
  RBP: ffffb932c09afc00 R08: 0000000000001000 R09: ffffa0a889277028
  R10: 0000000000020023 R11: 0000000000000000 R12: ffffda63c4249dc0
  R13: ffffa0a890d70d98 R14: 0000000000000028 R15: 0000000000000fd8
  FS:  00007f7294899580(0000) GS:ffffa0af9bc00000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: ffffa0a889277028 CR3: 0000000107ef6006 CR4: 0000000000370ef0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  Call Trace:
   ? zero_user_segments+0x82/0x190
   truncate_inode_partial_folio+0xd4/0x2a0
   truncate_inode_pages_range+0x380/0x830
   truncate_setsize+0x63/0x80
   simple_setattr+0x37/0x60
   notify_change+0x3d8/0x4d0
   do_sys_ftruncate+0x162/0x1d0
   __x64_sys_ftruncate+0x1c/0x20
   do_syscall_64+0x44/0xa0
   entry_SYSCALL_64_after_hwframe+0x44/0xae
  Modules linked in: xhci_pci xhci_hcd virtio_net net_failover failover virtio_blk virtio_balloon uhci_hcd ohci_pci ohci_hcd evdev ehci_pci ehci_hcd 9pnet_virtio 9p netfs 9pnet
  CR2: ffffa0a889277028

[lkp@intel.com: secretmem_iops can be static]
Signed-off-by: kernel test robot <lkp@intel.com>
[axelrasmussen@google.com: return EINVAL]

Link: https://lkml.kernel.org/r/20220324210909.1843814-1-axelrasmussen@google.com
Link: https://lkml.kernel.org/r/20220412193023.279320-1-axelrasmussen@google.com
Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: <stable@vger.kernel.org>
Cc: kernel test robot <lkp@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agotmpfs: fix regressions from wider use of ZERO_PAGE
Hugh Dickins [Fri, 15 Apr 2022 02:13:27 +0000 (19:13 -0700)]
tmpfs: fix regressions from wider use of ZERO_PAGE

Chuck Lever reported fsx-based xfstests generic 075 091 112 127 failing
when 5.18-rc1 NFS server exports tmpfs: bisected to recent tmpfs change.

Whilst nfsd_splice_action() does contain some questionable handling of
repeated pages, and Chuck was able to work around there, history from
Mark Hemment makes clear that there might be similar dangers elsewhere:
it was not a good idea for me to pass ZERO_PAGE down to unknown actors.

Revert shmem_file_read_iter() to using ZERO_PAGE for holes only when
iter_is_iovec(); in other cases, use the more natural iov_iter_zero()
instead of copy_page_to_iter().

We would use iov_iter_zero() throughout, but the x86 clear_user() is not
nearly so well optimized as copy to user (dd of 1T sparse tmpfs file
takes 57 seconds rather than 44 seconds).

And now pagecache_init() does not need to SetPageUptodate(ZERO_PAGE(0)):
which had caused boot failure on arm noMMU STM32F7 and STM32H7 boards

Link: https://lkml.kernel.org/r/9a978571-8648-e830-5735-1f4748ce2e30@google.com
Fixes: 56a8c8eb1eaf ("tmpfs: do not allocate pages on read")
Signed-off-by: Hugh Dickins <hughd@google.com>
Reported-by: Patrice CHOTARD <patrice.chotard@foss.st.com>
Reported-by: Chuck Lever III <chuck.lever@oracle.com>
Tested-by: Chuck Lever III <chuck.lever@oracle.com>
Cc: Mark Hemment <markhemm@googlemail.com>
Cc: Patrice CHOTARD <patrice.chotard@foss.st.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Lukas Czerner <lczerner@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: Broadcom internal lists aren't maintainers
Joe Perches [Fri, 15 Apr 2022 02:13:24 +0000 (19:13 -0700)]
MAINTAINERS: Broadcom internal lists aren't maintainers

Convert the broadcom internal list M: and L: entries to R: as exploder
email addresses are neither maintainers nor mailing lists.

Reorder the entries as necessary.

Link: https://lkml.kernel.org/r/04eb301f5b3adbefdd78e76657eff0acb3e3d87f.camel@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoi2c: ismt: Fix undefined behavior due to shift overflowing the constant
Borislav Petkov [Tue, 5 Apr 2022 15:15:11 +0000 (17:15 +0200)]
i2c: ismt: Fix undefined behavior due to shift overflowing the constant

Fix:

  drivers/i2c/busses/i2c-ismt.c: In function ‘ismt_hw_init’:
  drivers/i2c/busses/i2c-ismt.c:770:2: error: case label does not reduce to an integer constant
    case ISMT_SPGT_SPD_400K:
    ^~~~
  drivers/i2c/busses/i2c-ismt.c:773:2: error: case label does not reduce to an integer constant
    case ISMT_SPGT_SPD_1M:
    ^~~~

See https://lore.kernel.org/r/YkwQ6%2BtIH8GQpuct@zn.tnic for the gory
details as to why it triggers with older gccs only.

Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
2 years agoi2c: dev: Force case user pointers in compat_i2cdev_ioctl()
Andy Shevchenko [Mon, 11 Apr 2022 18:07:52 +0000 (21:07 +0300)]
i2c: dev: Force case user pointers in compat_i2cdev_ioctl()

Sparse has warned us about wrong address space for user pointers:

  i2c-dev.c:561:50: warning: incorrect type in initializer (different address spaces)
  i2c-dev.c:561:50:    expected unsigned char [usertype] *buf
  i2c-dev.c:561:50:    got void [noderef] __user *

Force cast the pointer to (__u8 *) that is used by I²C core code.

Note, this is an additional fix to the previously addressed similar issue
in the I2C_RDWR case in the same function.

Fixes: 3265a7e6b41b ("i2c: dev: Add __user annotation")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
2 years agoi2c: dev: check return value when calling dev_set_name()
Andy Shevchenko [Mon, 11 Apr 2022 18:07:51 +0000 (21:07 +0300)]
i2c: dev: check return value when calling dev_set_name()

If dev_set_name() fails, the dev_name() is null, check the return
value of dev_set_name() to avoid the null-ptr-deref.

Fixes: 1413ef638aba ("i2c: dev: Fix the race between the release of i2c_dev and cdev")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
2 years agoi2c: qcom-geni: Use dev_err_probe() for GPI DMA error
Bjorn Andersson [Tue, 12 Apr 2022 21:26:01 +0000 (14:26 -0700)]
i2c: qcom-geni: Use dev_err_probe() for GPI DMA error

The GPI DMA engine driver can be compiled as a module, in which case the
likely probe deferral "error" shows up in the kernel log. Switch to
using dev_err_probe() to silence this warning and to ensure that
"devices_deferred" in debugfs carries this information.

Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
2 years agoi2c: imx: Implement errata ERR007805 or e7805 bus frequency limit
Marek Vasut [Fri, 8 Apr 2022 17:15:24 +0000 (19:15 +0200)]
i2c: imx: Implement errata ERR007805 or e7805 bus frequency limit

The i.MX8MP Mask Set Errata for Mask 1P33A, Rev. 2.0 has description of
errata ERR007805 as below. This errata is found on all MX8M{M,N,P,Q},
MX7{S,D}, MX6{UL{,L,Z},S{,LL,X},S,D,DL,Q,DP,QP} . MX7ULP, MX8Q, MX8X
are not affected. MX53 and older status is unknown, as the errata
first appears in MX6 errata sheets from 2016 and the latest errata
sheet for MX53 is from 2015. Older SoC errata sheets predate the
MX53 errata sheet. MX8ULP and MX9 status is unknown as the errata
sheet is not available yet.

"
ERR007805 I2C: When the I2C clock speed is configured for 400 kHz,
the SCL low period violates the I2C spec of 1.3 uS min

Description: When the I2C module is programmed to operate at the
maximum clock speed of 400 kHz (as defined by the I2C spec), the SCL
clock low period violates the I2C spec of 1.3 uS min. The user must
reduce the clock speed to obtain the SCL low time to meet the 1.3us
I2C minimum required. This behavior means the SoC is not compliant
to the I2C spec at 400kHz.

Workaround: To meet the clock low period requirement in fast speed
mode, SCL must be configured to 384KHz or less.
"

Implement the workaround by matching on the affected SoC specific
compatible strings and by limiting the maximum bus frequency in case
the SoC is affected.

Signed-off-by: Marek Vasut <marex@denx.de>
To: linux-i2c@vger.kernel.org
Acked-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
2 years agoi2c: pasemi: Wait for write xfers to finish
Martin Povišer [Tue, 29 Mar 2022 18:38:17 +0000 (20:38 +0200)]
i2c: pasemi: Wait for write xfers to finish

Wait for completion of write transfers before returning from the driver.
At first sight it may seem advantageous to leave write transfers queued
for the controller to carry out on its own time, but there's a couple of
issues with it:

 * Driver doesn't check for FIFO space.

 * The queued writes can complete while the driver is in its I2C read
   transfer path which means it will get confused by the raising of
   XEN (the 'transaction ended' signal). This can cause a spurious
   ENODATA error due to premature reading of the MRXFIFO register.

Adding the wait fixes some unreliability issues with the driver. There's
some efficiency cost to it (especially with pasemi_smb_waitready doing
its polling), but that will be alleviated once the driver receives
interrupt support.

Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
Reviewed-by: Sven Peter <sven@svenpeter.dev>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
2 years agodm: fix bio length of empty flush
Shin'ichiro Kawasaki [Fri, 15 Apr 2022 08:45:13 +0000 (17:45 +0900)]
dm: fix bio length of empty flush

The commit 92986f6b4c8a ("dm: use bio_clone_fast in alloc_io/alloc_tio")
removed bio_clone_fast() call from alloc_tio() when ci->io->tio is
available. In this case, ci->bio is not copied to ci->io->tio.clone.
This is fine since init_clone_info() sets same values to ci->bio and
ci->io->tio.clone.

However, when incoming bios have REQ_PREFLUSH flag, __send_empty_flush()
prepares a zero length bio on stack and set it to ci->bio. At this time,
ci->io->tio.clone still keeps non-zero length. When alloc_tio() chooses
this ci->io->tio.clone as the bio to map, it is passed to targets as
non-empty flush bio. It causes bio length check failure in dm-zoned and
unexpected operation such as dm_accept_partial_bio() call.

To avoid the non-empty flush bio, set zero length to ci->io->tio.clone
in __send_empty_flush().

Fixes: 92986f6b4c8a ("dm: use bio_clone_fast in alloc_io/alloc_tio")
Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2 years agoMerge tag 'block-5.18-2022-04-15' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 15 Apr 2022 18:38:55 +0000 (11:38 -0700)]
Merge tag 'block-5.18-2022-04-15' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:

 - Moving of lower_48_bits() to the block layer and a fix for the
   unaligned_be48 added with that originally (Alexander, Keith)

 - Fix a bad WARN_ON() for trim size checking (Ming)

 - A polled IO timeout fix for null_blk (Ming)

 - Silence IO error printing for dead disks (Christoph)

 - Compat mode range fix (Khazhismel)

 - NVMe pull request via Christoph:
     - Tone down the error logging added this merge window a bit
       (Chaitanya Kulkarni)
     - Quirk devices with non-unique unique identifiers (Christoph)

* tag 'block-5.18-2022-04-15' of git://git.kernel.dk/linux-block:
  block: don't print I/O error warning for dead disks
  block/compat_ioctl: fix range check in BLKGETSIZE
  nvme-pci: disable namespace identifiers for Qemu controllers
  nvme-pci: disable namespace identifiers for the MAXIO MAP1002/1202
  nvme: add a quirk to disable namespace identifiers
  nvme: don't print verbose errors for internal passthrough requests
  block: null_blk: end timed out poll request
  block: fix offset/size check in bio_trim()
  asm-generic: fix __get_unaligned_be48() on 32 bit platforms
  block: move lower_48_bits() to block