3b3cecea215895b2235f2eb70e9e6859b379630a: "mm: convert totalram_pages and totalhigh_pages variables to atomic"
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
This hack sucks, was never necessary and instead just
writing to the subsystem's sysfs node is enough.
This reverts commits 6a16747eb1c3e0c53fd06e710913f14273f831d4 and 011f53380857e4729806931f2a0a4949a9503903.
* Our camera HAL uses boot time for buffer timestamp, rather than
system monotonic time. This leads to issues as framework uses system
monotonic time as reference start time for timestamp adjustment.
* This patch is taken from stock kernel source.
Change-Id: Ia4fac1d48e2206a7befd0030585776371dd8c3a9
Signed-off-by: Subhajeet Muhuri <subhajeet.muhuri@aosip.dev>
This feature implemented for power optimization
with OMX but not well compliant with Codec 2.0.
Disable it to unblock VT frame drop issue.
Bug: 149071324
Test: VT frame rate test
Change-Id: I187958f4da10d5936c0f0fbd5060301e55ac7f29
Signed-off-by: Wen Chang Liu <wenchangliu@google.com>
* this makes it easier to identify kernel builds
and track it down in history, since we already
do this in localversion
* example: QuicksilveR-surya-20220219-0000-638c37bb.zip
Replace 'schedule(); try_to_freeze();' with a call to freezable_schedule().
Tasks calling freezable_schedule() set the PF_FREEZER_SKIP flag
before calling schedule(). Unlike tasks calling schedule();
try_to_freeze() tasks calling freezable_schedule() are not awaken by
try_to_freeze_tasks(). Instead they call try_to_freeze() when they
wake up if the freeze is still underway.
It is not a problem since sleeping tasks can't do anything which isn't
allowed for a frozen task while sleeping.
The result is a potential performance gain during freeze, since less
tasks have to be awaken.
For instance on a bare Debian vm running a 4.19 stable kernel, the
number of tasks skipped in freeze_task() went up from 12 without the
patch to 32 with the patch (out of 448), an increase of > x2.5.
Signed-off-by: Hugo Lefeuvre <hle@owl.eu.com>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20190207200352.GA27859@behemoth.owl.eu.com.local
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Although data dependencies and one-way, semi-permeable barriers provided by
spin locks satisfy most ordering needs here, it is still possible for some
I/O writes to be reordered with respect to one another in a dangerous way.
One such example is that the interrupt status bit could be cleared *after*
the interrupt is unmasked when enabling the IRQ, potentially leading to a
spurious interrupt if there's an interrupt pending from when the IRQ was
disabled.
To prevent dangerous I/O write reordering, restore the minimum amount of
barriers needed to ensure writes are ordered as intended.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Sending synchronous IPIs to other CPUs involves spinning with preemption
disabled in order to wait for each IPI to finish. Keeping preemption off
for long periods of time like this is bad for system jitter, not to mention
the perf event IPIs are sent and flushed one at a time for each event for
each CPU rather than all at once for all the CPUs.
Since the way perf events are currently read is quite naive, rewrite it to
make it exploit parallelism and go much faster. IPIs for reading each perf
event are now sent to all CPUs asynchronously so that each CPU can work on
reading the events in parallel, and the dispatching CPU now sleeps rather
than spins when waiting for the IPIs to finish. Before the dispatching CPU
starts waiting though, it works on reading events for itself and then
reading events which can be read from any CPU in order to derive further
parallelism, and then waits for the IPIs to finish afterwards if they
haven't already.
Furthermore, there's now only one IPI sent to read all of a CPU's events
rather than an IPI sent for reading each event, which significantly speeds
up the event reads and reduces the number of IPIs sent.
This also checks for active SCM calls on a per-CPU basis rather than a
global basis so that unrelated CPUs don't get their counter reads skipped
and so that some CPUs can still receive fresh counter readings.
Overall, this makes the memlat driver much faster and more efficient, and
eliminates significant system jitter previously caused by IPI abuse.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
This reverts commit 1d647604a0baf169928deebffe0450132c27f24b.
There are a number of problems with this optimization attempt, such as how
a perf error code could be returned to one of the counter values, a CPU
could enter an SCM call for just one event and result in garbage
computations or a divide-by-zero when calculating the stall percentage,
and IRQs are disabled as a heavy-handed way of preventing CPU migration.
Toss this commit out so that it can be replaced by something better.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
There's no reason to constantly use snprintf() to generate pretty debug
strings from hot paths. We don't need them, so remove them.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
Constantly allocating and freeing all of the data structures associated
with atomic commits adds up and incurs a lot of latency not only when
allocating, but also when freeing. Since we know what the maximum number
of CRTCs, planes, and connectors is, we can skip the constant allocation-
and-free for the same structures and instead just recycle them via a lock-
less list. This also moves the commit cleanup so that it comes after CRTC
waiters are woken up, allowing the ioctl to proceed without waiting around
for some housekeeping to finish.
Since it's difficult to audit which parameters, if any, could exceed the
defined maximums in the msm_kms driver, dynamic allocations are retained as
a fallback so that userspace can't craft a malicious ioctl that results in
buffer overflows.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
This reverts commit e3cb7f446cdf36b5e64852c18cd05dde34ecf55f.
This creates a large TTWU burden, which isn't worth it for such a small
amount of cleanup. The cleanup will become inconsequential anyway since the
memory allocations will be replaced with buffer recycling.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Android sets the value to 50ms via vold's IdleMaint service. Since
500ms is too long for GC to colllect all invalid segments in time
which results in performance degradation.
On un-encrypted device, vold fails to set this value to 50ms thus
degrades the performance over time.
Based on [1].
[1] https://github.com/topjohnwu/Magisk/pull/5462
Signed-off-by: Panchajanya1999 <rsk52959@gmail.com>
Change-Id: I80f2c29558393d726d5e696aaf285096c8108b23
Signed-off-by: Panchajanya1999 <rsk52959@gmail.com>
Import the latest version of the Arm Optimized Routines strncmp function based
on the upstream code of string/aarch64/strncmp.S at commit 189dfefe37d5 from:
https://github.com/ARM-software/optimized-routines
This latest version includes MTE support.
Note that for simplicity Arm have chosen to contribute this code to Linux under
GPLv2 rather than the original MIT OR Apache-2.0 WITH LLVM-exception license.
Arm is the sole copyright holder for this code.
Co-authored-by: Joey Gouly <joey.gouly@arm.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20220301101435.19327-3-joey.gouly@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Divyanshu-Modi <divyan.m05@gmail.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
Import the latest version of the Arm Optimized Routines strcmp function based
on the upstream code of string/aarch64/strcmp.S at commit 189dfefe37d5 from:
https://github.com/ARM-software/optimized-routines
This latest version includes MTE support.
Note that for simplicity Arm have chosen to contribute this code to Linux under
GPLv2 rather than the original MIT OR Apache-2.0 WITH LLVM-exception license.
Arm is the sole copyright holder for this code.
Signed-off-by: Joey Gouly <joey.gouly@arm.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20220301101435.19327-2-joey.gouly@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Divyanshu-Modi <divyan.m05@gmail.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
Sending synchronous IPIs to other CPUs involves spinning with preemption
disabled in order to wait for each IPI to finish. Keeping preemption off
for long periods of time like this is bad for system jitter, not to mention
the watchdog's IPIs are sent and flushed one at a time for each CPU rather
than all at once for all the CPUs to be pinged.
Since the existing IPI ping machinery is quite lacking, rewrite it entirely
to address all of its performance shortcomings. This not only replaces the
synchronous IPIs with asynchronous ones, but also allows the IPIs to run in
parallel. The IPI ping and wait mechanisms are now much more efficient via
the use of generic_exec_single() (since smp_call_function_single_async()
disables preemption when all it really needs is migration disabled), and
by sleeping rather than spinning while waiting for the IPIs to finish.
This also does away with the ping_start and ping_end arrays as they don't
make sense with the parallel, asynchronous execution of the IPIs anymore.
They are instead replaced by a mask indicating which CPUs were pinged so
that a watchdog bark can still print out which CPU(s) stopped responding.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
commit 1139aeb1c521eb4a050920ce6c64c36c4f2a3ab7 upstream.
As of commit 966a967116e6 ("smp: Avoid using two cache lines for struct
call_single_data"), the smp code prefers 32-byte aligned call_single_data
objects for performance reasons, but the block layer includes an instance
of this structure in the main 'struct request' that is more senstive
to size than to performance here, see 4ccafe032005 ("block: unalign
call_single_data in struct request").
The result is a violation of the calling conventions that clang correctly
points out:
block/blk-mq.c:630:39: warning: passing 8-byte aligned argument to 32-byte aligned parameter 2 of 'smp_call_function_single_async' may result in an unaligned pointer access [-Walign-mismatch]
smp_call_function_single_async(cpu, &rq->csd);
It does seem that the usage of the call_single_data without cache line
alignment should still be allowed by the smp code, so just change the
function prototype so it accepts both, but leave the default alignment
unchanged for the other users. This seems better to me than adding
a local hack to shut up an otherwise correct warning in the caller.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Jens Axboe <axboe@kernel.dk>
Link: https://lkml.kernel.org/r/20210505211300.3174456-1-arnd@kernel.org
[nc: Fix conflicts]
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To make sure detach and attach netif instead of stop
and wakeup and also to update transfer timer update.
Change-Id: I9d589b7f9f6fe98f778df509d3c16f339dfdeea1
Signed-off-by: Michael Adisumarta <madisuma@codeaurora.org>
Signed-off-by: Andrzej Perczak <linux@andrzejperczak.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
After allocating and zeroing pages from system memory or the pool we need
to ensure that the cache is synchronized so that it doesn't cause problems
down the road. Use dma_sync_sg_for_device to make sure the allocated pages
are clean. This isn't the best way to handle this but we haven't yet come
up with a better way and this does the job.
xNombre: Backport to 4.14, should help with expensive zero-memsetting page vs
using dedicated clear_page.
Inspired by: 2b085f4448
Change-Id: Ic0dedbade48b700015bec172cf9b64e436364b4a
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Andrzej Perczak <linux@andrzejperczak.com>
Signed-off-by: Adithya R <gh0strider.2k18.reborn@gmail.com>
Increase the total number of camera ID's supported.
Change-Id: Icf9cf2e48757d437a2b7dbbe618e886dc0203f6d
Signed-off-by: Om Parkash <quic_oparkash@quicinc.com>
Add support for ToF camera sensor on trinket.
Change-Id: I184862d787ecace17337054cbda3d32681f86c5d
Signed-off-by: Om Parkash <quic_oparkash@quicinc.com>
Query for unsigned pd support on cDSP domain and update
the unsigned_support flag during fastrpc_init process.
Change-Id: I61f4c748ad08155f418422183acc8473a7b0e0a8
Signed-off-by: Vamsi Krishna Gattupalli <quic_vgattupa@quicinc.com>
Reject the session when third party applications
try to spawn signed PD and channel configured as secure.
Change-Id: Ic450a8c7dad430dfcdc4ae7354e29e63d9fae4a3
Acked-by: Krishnaiah Tadakamalla <ktadakam@qti.qualcomm.com>
Signed-off-by: Jeya R <jeyr@codeaurora.org>
To check if DSP is supported or not, we make
fs call for subsystem device node. This node
is not accessible to untrusted applications.
Use subsystem status flag instead to avoid
permission issues and return proper error
in case subsystem is not up.
Change-Id: Ia19e31b899600e5d765c0a3582bdf9132c9b67bf
Acked-by: Ekansh Gupta <ekangupt@qti.qualcomm.com>
Signed-off-by: Vamsi krishna Gattupalli <vgattupa@codeaurora.org>
Shared memory is initialized by CRM and used by
other drivers; with CRM not active other drivers
would fail to access the shared memory if
memory manager is deinit. Reader Writer locks can
prevent the open/close/ioctl calls from other drivers
if CRM open/close is already being processed.
Issue observed with the below sequence if drivers
are opened from UMD directly without this change.
CRM Open successful,ICP open successful,
CRM close in progress, ICP open successful,
mem mgr deinit and CRM close successful,
ICP tries to access HFI memory and results in crash.
This change helps to serialze the calls and prevents
issue.
CRs-Fixed: 3019488
Change-Id: If260a948918ffc1e7d8e20564fe7434731c8da9e
Signed-off-by: Tejas Prajapati <quic_tpraja@quicinc.com>
PMI632 does not have DCIN input. Add changes to
detect DCIN or uUSB insertion via GPIO and enable
charging for the same.
Change-Id: I4f5600041048d2fe1bfd8ded030fb718e55beaca
Signed-off-by: Umang Chheda <quic_uchheda@quicinc.com>
