Pull vfs file updates from Christian Brauner:
"This is the work to cleanup and shrink struct file significantly.
Right now, (focusing on x86) struct file is 232 bytes. After this
series struct file will be 184 bytes aka 3 cacheline and a spare 8
bytes for future extensions at the end of the struct.
With struct file being as ubiquitous as it is this should make a
difference for file heavy workloads and allow further optimizations in
the future.
- struct fown_struct was embedded into struct file letting it take up
32 bytes in total when really it shouldn't even be embedded in
struct file in the first place. Instead, actual users of struct
fown_struct now allocate the struct on demand. This frees up 24
bytes.
- Move struct file_ra_state into the union containg the cleanup hooks
and move f_iocb_flags out of the union. This closes a 4 byte hole
we created earlier and brings struct file to 192 bytes. Which means
struct file is 3 cachelines and we managed to shrink it by 40
bytes.
- Reorder struct file so that nothing crosses a cacheline.
I suspect that in the future we will end up reordering some members
to mitigate false sharing issues or just because someone does
actually provide really good perf data.
- Shrinking struct file to 192 bytes is only part of the work.
Files use a slab that is SLAB_TYPESAFE_BY_RCU and when a kmem cache
is created with SLAB_TYPESAFE_BY_RCU the free pointer must be
located outside of the object because the cache doesn't know what
part of the memory can safely be overwritten as it may be needed to
prevent object recycling.
That has the consequence that SLAB_TYPESAFE_BY_RCU may end up
adding a new cacheline.
So this also contains work to add a new kmem_cache_create_rcu()
function that allows the caller to specify an offset where the
freelist pointer is supposed to be placed. Thus avoiding the
implicit addition of a fourth cacheline.
- And finally this removes the f_version member in struct file.
The f_version member isn't particularly well-defined. It is mainly
used as a cookie to detect concurrent seeks when iterating
directories. But it is also abused by some subsystems for
completely unrelated things.
It is mostly a directory and filesystem specific thing that doesn't
really need to live in struct file and with its wonky semantics it
really lacks a specific function.
For pipes, f_version is (ab)used to defer poll notifications until
a write has happened. And struct pipe_inode_info is used by
multiple struct files in their ->private_data so there's no chance
of pushing that down into file->private_data without introducing
another pointer indirection.
But pipes don't rely on f_pos_lock so this adds a union into struct
file encompassing f_pos_lock and a pipe specific f_pipe member that
pipes can use. This union of course can be extended to other file
types and is similar to what we do in struct inode already"
* tag 'vfs-6.12.file' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (26 commits)
fs: remove f_version
pipe: use f_pipe
fs: add f_pipe
ubifs: store cookie in private data
ufs: store cookie in private data
udf: store cookie in private data
proc: store cookie in private data
ocfs2: store cookie in private data
input: remove f_version abuse
ext4: store cookie in private data
ext2: store cookie in private data
affs: store cookie in private data
fs: add generic_llseek_cookie()
fs: use must_set_pos()
fs: add must_set_pos()
fs: add vfs_setpos_cookie()
s390: remove unused f_version
ceph: remove unused f_version
adi: remove unused f_version
mm: Removed @freeptr_offset to prevent doc warning
...
The Gen6 devices have the same problem and the same Solution as the Gen5
ones.
Some TongFang barebones have touchpad and/or keyboard issues after
suspend, fixable with nomux + reset + noloop + nopnp. Luckily, none of
them have an external PS/2 port so this can safely be set for all of
them.
I'm not entirely sure if every device listed really needs all four quirks,
but after testing and production use, no negative effects could be
observed when setting all four.
Signed-off-by: Werner Sembach <wse@tuxedocomputers.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20240910094008.1601230-3-wse@tuxedocomputers.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
There's no need to get the length of an DT array property before
parsing the array. of_property_read_variable_u32_array() takes a
minimum and maximum length and returns the actual length (or error
code).
This is part of a larger effort to remove callers of of_get_property()
and similar functions. of_get_property() leaks the DT property data
pointer which is a problem for dynamically allocated nodes which may
be freed.
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://lore.kernel.org/r/20240913200827.546649-1-robh@kernel.org
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
- drop flags, they were not used anyway
- add OF ID match table
- process "autorepeat", "debounce-delay-ms", prescale from device tree
- drop platform data usage and it's header
- keymap goes from device tree now on
Signed-off-by: Nikita Shubin <nikita.shubin@maquefel.me>
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Acked-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
This regulator is not optional from the controller point of view,
so stop treating it as such. For hard-wired designs that omit the
regulator from their device trees regulator subsystem will create
a dummy instance.
This may introduce unnecessary delay of 100us in case of dummy
regulator, but if it is important the driver should be marked as
using asynchronous probing to avoid even longer delays waiting for
the command completions.
Also use usleep_range() instead of udelay() to avoid spinning.
Tested-by: Andreas Kemnade <andreas@kemnade.info> # Tolino Shine2HD
Link: https://lore.kernel.org/r/20240824055047.1706392-15-dmitry.torokhov@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Zforce touch data packet consists of a byte representing number of
contacts followed by several chunks with length of 9 bytes representing
each contact. Instead of accounting for the leading byte by increasing
offset of each field in contacts by one introduce a pointer to contact
data and point it appropriately. This avoids awkward constructs like:
point.prblty = payload[9 * i + 9];
which makes it seem like there is off-by-one error, in favor of more
straightforward:
point.prblty = p[8];
Tested-by: Andreas Kemnade <andreas@kemnade.info> # Tolino Shine2HD
Link: https://lore.kernel.org/r/20240824055047.1706392-11-dmitry.torokhov@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
There is a small chance that ts->suspending flag may change while the
interrupt handler is running. To make sure call to pm_relax() is not
skipped on accident use a temporary to hold the original value at the
beginning of interrupt. Use READ_ONCE() so that the value is actually
fetched at the right time.
Tested-by: Andreas Kemnade <andreas@kemnade.info> # Tolino Shine2HD
Link: https://lore.kernel.org/r/20240824055047.1706392-8-dmitry.torokhov@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Some TongFang barebones have touchpad and/or keyboard issues after
suspend, fixable with nomux + reset + noloop + nopnp. Luckily, none of
them have an external PS/2 port so this can safely be set for all of
them.
I'm not entirely sure if every device listed really needs all four quirks,
but after testing and production use, no negative effects could be
observed when setting all four.
Signed-off-by: Werner Sembach <wse@tuxedocomputers.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20240905164851.771578-1-wse@tuxedocomputers.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Arm SCMI updates for v6.12
Few main features include:
1. SCMI transport as stand-alone drivers
Currently the SCMI transport layer is being built embedded into in
the core SCMI stack. Some of these transports, despite being currently
part of the main SCMI module, are indeed also registered with different
subsystems like optee or virtio, and actively probed also by those.
This leads to a few awkward and convoluted tricks to properly handle
such interactions at boot time in the SCMI stack.
This change adds the new logic to the core SCMI stack so that each
existing transport is transitioned to be a standi-alone driver. With
that all the probe deferral and awkward retries between the SCMI
core stack and the transports has been removed, since no more needed.
2. Support for obtaining transport descriptors from the devicetree
SCMI platform firmwares might have different designs depending on
the platform. Some of the transport descriptors rely on such design.
E.g. the maximum receive channel timeout value might vary depending
on the specific underlying hardware and firmware design choices.
This change adds support for max-rx-timeout-ms property to describe
the transport needs of a specific platform design. It will be extended
in the future to obtain other such hardware/firmware dependent
transport related descriptors.
3. NXP i.MX95 specific SCMI vendor protocol extensions
SCMI specification allows vendor or platform-specific extensions to
the interface. NXP i.MX95 System Manager(SM) that implements SCMI
extends the interface to implement couple of vendor/platform specific
protocol, namely:
a. Battery Backed Module(BBM) Protocol
This protocol is intended provide access to the battery-backed
module. This contains persistent storage (GPR), an RTC, and the
ON/OFF button. The protocol can also provide access to similar
functions implemented via external board components.
b. MISC Protocol for misc settings
This includes controls that are misc settings/actions that must
be exposed from the SM to agents. They are device specific and
are usually define to access bit fields in various mix block
control modules, IOMUX_GPR, and other GPR/CSR owned by the SM.
4. SCMI debug/tracking metrics
Since SCMI involves interaction with the entity(software, firmware
and/or hardware) providing services or features, it is quite useful
to track certain metrics(for pure debugging purposes) like how many
messages were sent or received, were there any failures, what kind
of failures, ..etc. This feature adds support for the same via debugfs.
Apart from these main features, there are some miscellaneous updates, fixes
and cleanups.
* tag 'scmi-updates-6.12' of https://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux: (31 commits)
rtc: support i.MX95 BBM RTC
input: keyboard: support i.MX95 BBM module
firmware: imx: Add i.MX95 MISC driver
firmware: arm_scmi: Add initial support for i.MX MISC protocol
firmware: arm_scmi: Add initial support for i.MX BBM protocol
firmware: arm_scmi: Add NXP i.MX95 SCMI documentation
dt-bindings: firmware: Add i.MX95 SCMI Extension protocol
firmware: arm_scmi: Replace comma with the semicolon
firmware: arm_scmi: Replace the use of of_node_put() to __free(device_node)
firmware: arm_scmi: Fix trivial whitespace/coding style issues
firmware: arm_scmi: Use max-rx-timeout-ms from devicetree
dt-bindings: firmware: arm,scmi: Introduce property max-rx-timeout-ms
firmware: arm_scmi: Remove const from transport descriptors
firmware: arm_scmi: Simplify with scoped for each OF child loop
firmware: arm_scmi: Update various protocols versions
firmware: arm_scmi: Remove legacy transport-layer code
firmware: arm_scmi: Make VirtIO transport a standalone driver
firmware: arm_scmi: Make OPTEE transport a standalone driver
firmware: arm_scmi: Make SMC transport a standalone driver
firmware: arm_scmi: Make MBOX transport a standalone driver
...
Link: https://lore.kernel.org/r/20240830135918.2383664-1-sudeep.holla@arm.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Pull input fix from Dmitry Torokhov:
- a fix for Cypress PS/2 touchpad for regression introduced in 6.11
merge window where a timeout condition is incorrectly reported for
all extended Cypress commands
* tag 'input-for-v6.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: cypress_ps2 - fix waiting for command response
Commit 8bccf667f6 ("Input: cypress_ps2 - report timeouts when reading
command status") uncovered an existing problem with cypress_ps2 driver:
it tries waiting on a PS/2 device waitqueue without using the rest of
libps2. Unfortunately without it nobody signals wakeup for the
waiting process, and each "extended" command was timing out. But the
rest of the code simply did not notice it.
Fix this by switching from homegrown way of sending request to get
command response and reading it to standard ps2_command() which does
the right thing.
Reported-by: Woody Suwalski <terraluna977@gmail.com>
Tested-by: Woody Suwalski <terraluna977@gmail.com>
Fixes: 8bccf667f6 ("Input: cypress_ps2 - report timeouts when reading command status")
Link: https://lore.kernel.org/r/a8252e0f-dab4-ef5e-2aa1-407a6f4c7204@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
The BBM module provides BUTTON feature. To i.MX95, this module
is managed by System Manager and exported using System Management
Control Interface(SCMI). Linux could use i.MX SCMI BBM Extension
protocol to use BUTTON feature.
This driver is to use SCMI interface to enable pwrkey.
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Message-Id: <20240823-imx95-bbm-misc-v2-v8-7-e600ed9e9271@nxp.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Linus Torvalds