Robust context attempts to perform a rendering that takes too long
whether due to an infinite loop in a shader or even just a rendering
operation that takes too long on the given hardware. This type of
attempts can result into GPU faults. Robust context expect driver
to replay IB instead skip IB and if it fails on replay context has
to be invalidated.
KGSL_CONTEXT_INVALIDATE_ON_FAULT flag allows draw context to execute
only replay policy on GPU fault recovery instead of going to default
recovery policy. User space has to set this flag during the context
creation.
Change-Id: If42dc5afc7d5ed1226b73ae5abfa2648d7acf2c3
Signed-off-by: Hareesh Gundu <hareeshg@codeaurora.org>
If the user provides a profile buffer identified with a buffer type
MSM_SUBMIT_CMD_PROFILE_BUF, then the driver records the kernel clock
time and gpu ticks at the time of cmdstream submission, and the GPU
records the ticks just before the start of the cmdstream execution and
right after the end of the cmdstream execution.
Change-Id: Ic6298ec5919b18e976ae089ffb0860b8165ce4f3
Signed-off-by: Sharat Masetty <smasetty@codeaurora.org>
Only the channel owner app can lock/unlock a channel ion buf.
However, if the app share the ION buf FD with its child tasks,
they might wrongly free the same ion buf twice.
The ion driver panic if a non-valid ION buff handle is provided.
Change-Id: Ia8166df5ea314949090f7e94e90eff3a3ed78b19
Signed-off-by: Amir Samuelov <amirs@codeaurora.org>
Wait for glink LOCAL_DISCONNECT event, even if the channel is not
connected.
Change-Id: I03f20b6dbb2f46d7abd6996e8d298e991cf3f82b
Signed-off-by: Amir Samuelov <amirs@codeaurora.org>
Add a check for cpu unbound deferrable timer expiry and raise
softirq for handling the expired timers so that the CPU can
process the cpu unbound deferrable times as early as possible
when a cpu tries to enter/exit idle loop.
Change-Id: Ieffa74fa22a4d25493f5590b5ac1e0d784fcbbad
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
(https://lkml.org/lkml/2016/12/13/579)
posix_acl_update_mode() could possibly clear 'acl', if so
we leak the memory pointed by 'acl'. Save this pointer
before calling posix_acl_update_mode() and release the memory
if 'acl' really gets cleared.
Reported-by: Mark Salyzyn <salyzyn@android.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Greg Kurz <groug@kaod.org>
Cc: Eric Van Hensbergen <ericvh@gmail.com>
Cc: Ron Minnich <rminnich@sandia.gov>
Cc: Latchesar Ionkov <lucho@ionkov.net>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Bug: 32458736
Change-Id: Ia78da401e6fd1bfd569653bd2cd0ebd3f9c737a0
Git-repo: https://android.googlesource.com/kernel/msm/
Git-commit: 7c780387c931f63fe67830c817f65c1272512bb1
Signed-off-by: Dennis Cagle <d-cagle@codeaurora.org>
The background scan completion takes more time when the station is
having heavy uplink traffic. The scan state machine decides to fall
back to home channel on every off-channel visit when there are pending
frames in tx queue. bgscan completion took ~30sec on dual band US
regulatory card.
scan period = (20 active channels * probe timeout) +
(12 passive channels * passive probe timeout) +
(32 * timeout on home channel) +
(32 * flush timeout)
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Git-commit: f9616e0f88
Git-repo: git://git.kernel.org/pub/scm/linux/kernel/git/
jberg/mac80211-next.git
CRs-Fixed: 2036907
Change-Id: I8b57fba12f0d42fa7d01243210206d432fbf9757
Signed-off-by: Padma, Santhosh Kumar <skpadma@codeaurora.org>
Deferring all the PP features programming to post pingpong done
in command mode panels causing performance issues in certain
use cases. To fix this, defer only the programming of features
with single buffered registers and program features with double
buffered registers before wait for pingpong.
Change-Id: I6a1e8114b50c558f667bde4db5c0ba57009d6f50
Signed-off-by: Sravan Kumar D.V.N <sravank1@codeaurora.org>
Fix RX sensitivity issue for SSPHY data lines for SDM660 by
writing zero value to PERIPH_SS_USB3PHY_QSERDES_RXA/B_RX_MODE_00
registers. Otherwise it could cause superspeed enumeration failure
with some Host machines.
Change-Id: I065c2a37f962e21809f9d810e1884e0eca08943e
Signed-off-by: Vijayavardhan Vennapusa <vvreddy@codeaurora.org>
On 32bit systems, bitmap storing the iova information is stored in per
page and each page represents the 128MB of IOVA. So, for the
alloc_iova() to return the requested range, iova size should be aligned
to 128MB.
Change-Id: I66eb15cdbb3d6730cc4d84cad9f0c652f91278d6
Signed-off-by: Charan Teja Reddy <charante@codeaurora.org>
The wcnss platform driver update the wlan calibration data
by the user space wlan daemon. The wlan user space daemon store
the updated wlan calibration data reported by wlan firmware in
user space and write it back to the wcnss platform calibration
data buffer for the calibration data download and update.
During the wlan calibration data store and retrieve operation
there are some potential race condition which leads to memory leak
and buffer overflow during the context switch.
Fix the above issue by adding protection code and avoid usage of
global pointer during the device file read and write operation.
CRs-Fixed: 2015858
Change-Id: Ib5b57eb86dcb4e6ed799b5222d06396eaabfaad3
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
Isolated cpu can enter to wfi if sleep is disabled and stay in
wfi until it wakes up and selects deeper lpm again.
Select cpu lpm for isolated cpu when sleep is disabled.
Change-Id: Id936ebaac0118d48d07e74aff7569f93b8b98b8b
Signed-off-by: Maulik Shah <mkshah@codeaurora.org>
Dump sync logs only when a fence is not signalled
for SYNC_DUMP_TIME_LIMIT or more. This is necessary
to reduce CPU hogging because of excessive sync logs
dumped due to fence timeouts which are less than
SYNC_DUMP_TIME_LIMIT and are harmless.
Change-Id: Id21a02925dd45a01361d096048a25d1879ebde9e
Signed-off-by: Deepak Kumar <dkumar@codeaurora.org>
Add mixer control to swap the speaker channels.
Use MFC module to update the channel mapping.
When playback is started with speaker device and
device orientation changes, swapping L/R channel
mixer control command is sent to DSP with the MFC
module.
CRs-Fixed: 1026248
Change-Id: I1d80ecc7f1e3cfb8f8cdf71c384e92a07023be5c
Signed-off-by: kunleiz <kunleiz@codeaurora.org>
Set the size of bitmap area properly to ensure the proper iova address
to be returned by the bitmap_() API's through alloc_iova(). The current
implementation causes the bitmap index search out of bounds of a page
which can lead to the randomn corruption issues.
Change-Id: Ie57ca0b7465c91d318145ed270b7d47d822c8eeb
Fixes:I88ddd98a76b ("arm/arm64: dma-mapping: Fix iova region size")
Signed-off-by: Charan Teja Reddy <charante@codeaurora.org>
With fast mapping, the iova range is zero in the absence of
mapping object from the client. In such case, set the iova
space using the DOMAIN_ATTR_GEOMETRY dommain attribute.
CRs-Fixed: 2020868
Change-Id: I28790728c02ecb2d0a921bf5a218fb8b16662619
Signed-off-by: Sudarshan Rajagopalan <sudaraja@codeaurora.org>
This separates the kref for ion handles into two components.
Userspace requests through the ioctl will hold at most one
reference to the internally used kref. All additional requests
will increment a separate counter, and the original reference is
only put once that counter hits 0. This protects the kernel from
a poorly behaving userspace.
Bug: 34276203
Change-Id: Ibc36bc4405788ed0fea7337b541cad3be2b934c0
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Git-repo: https://android.googlesource.com/kernel/msm/
Git-commit: 20abfcc16884a5af973a5e91dd013ddd789c44f4
[d-cagle@codeaurora.org: Resolve style issues]
Signed-off-by: Dennis Cagle <d-cagle@codeaurora.org>
