When audio_aio_open() is triggered in blocking fashion, it frees audio
resouces and returns error code. In the current implementation, callers
did not check the error code which can lead to dereference to freed
resources and cause system crash. Update the callers to check return
code and then return error back to application to avoid the issue. In
the meanwhile, update the audio_aio_open not to free memory for caller
and leave the caller to free the memory in error condition.
CRs-fixed: 436653
Change-Id: If4dda158cb988c5837517817b1afa0572c30696e
Signed-off-by: Kuirong Wang <kuirongw@codeaurora.org>
mmc_stop_bkops calls cancel_delayed_work_sync while the host is taken.
This can cause a deadlock since the delayed work is blocked on claiming
the host. In order to prevent that, mmc_start_bkops will not
be blocked waiting on mmc_claim_host and if the host is already taken
BKOPS will not be performed.
CRs-Fixed: 446090
Change-Id: I5cda5489e254764b3231b8a12da9e3c241c969c3
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit 8571386b83bea1cbadd81c75bf0614da30b6b7c0)
If there is a race condition between the delayed BKOPS work and
the runtime suspend, and the suspend occurs before the delayed work,
the delayed work should be canceled.
Since mmc_claim_host can fail in case the host is suspended, we should
check if the delayed work is canceled before trying to claim the host.
CRs-Fixed: 443236
Change-Id: I9df506de4da6f9277a89ed608724f5b6e8585bfd
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit 5a3d216c005725baad96b6858dbcd370056e470c)
In order to keep card endurance, it is best not to check the need for
BKOPS every time the MMC is idle.
However, the decision when to check the card need for BKOPS according
to number of changed sectors doesn't fit all cards sizes. Therefore,
the check for BKOPS need is based on percentages of changed sectors
of the card size.
In order to assure that BKOPS will be triggered again in case it was
interrupted, the accumulated number of changed sectors is cleared only
in case BKOPS was completed without interruption.
Change-Id: I598f64b652a524a431b87d103fd5890b808f11b1
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit e1eae988976536834759a45cf7f62e1cb2b82838)
BKOPS operations can take long time and cause bad user experience and
failure of time critical operations as EFS sync.
In order to prevent the above failures all the levels will be handled only
in idle time BKOPS handling and will allow interrupting the BKOPS when
needed.
In case the card raised an exception with a need for BKOPS, a flag will
be set to indicate MMC to start the BKOPS activity when it becomes idle.
Fixed-CRs: 432027
Change-Id: I5f7b43569c0242f0fea83355f76f286b1ad037bc
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit 5f8ac3b955e44def61c2238192000ffe5f126714)
The timeout for the BKOPS polling is increased in order to allow the
card to complete the BKOPS before going to suspend.
In case we exit due to timeout we need to stop the BKOPS activity
performed by the card.
Change-Id: I39686de0c06a57242f11c8b1c9ae50b33417c088
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit ea1385e6d7542d4e3ad510de94ead00bece7823e)
mmc_cache_ctrl was called in runtime suspend before MMC interrupted
BKOPS in case it is still running on the card. This caused the cache
disable to timeout.
Therefore, mmc_cache_ctrl has to move to mmc_suspend where we are sure
that the card can go into suspend and there is no pending activity.
CRs-Fixed: 436805
Change-Id: I76281a766f48ccacc24685a64067df4e60cfe1d8
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit 67a51386fd99be62e5e68f4bc76ee00c19099cd6)
In order to allow the card to perform BKOPS as early as possible without
causing degradation to a flow of requests, we reduce the time of the
BKOPS delayed work to 200ms.
Based on experience, 200ms will allow requests to be performed without
a delay of checking the BKOPS.
This patch also allows the host controller to set any other time if
required.
Change-Id: I1e56bfd0c9471abf4da50befe6d14fe0199129e7
Signed-off-by: Maya Erez <merez@codeaurora.org>
(cherry picked from commit 70b00220e7564b00e3b0747260fe2c2dbe62f45d)
Remove the delay in HDMI resume sequence by not waiting for HPD
to be switched on. HDMI panel will not be turned on by default
when the device resumes and would instead be turned on by HPD
sense if it is still connected.
CRs-Fixed: 440559
Change-Id: Id276dddbcebf9226e9688aa1b5dad4b281545b21
Signed-off-by: Ajay Singh Parmar <aparmar@codeaurora.org>
This change lets the user modify the fps level dynamically using
sysfs interface.
Change-Id: Ie99bbca5ec1a0b5ff5fa3570874b254e3fb7235c
Signed-off-by: Vishnuvardhan Prodduturi <vproddut@codeaurora.org>
Stop parsing an indirect buffer if an error is encountered (such as
a missing buffer). This is a pretty good indication that the buffers
are not reliable and the further the parser goes with a unreliable
buffer the more likely it is to get confused.
Change-Id: Ic0dedbadf28ef374c9afe70613048d3c31078ec6
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
The snapshot timestamp is currently based on wall time which makes
it very difficult to debug against a kernel log based on boot time.
It doesn't help matters that the snapshot timestamp is output in
human unfriendly hex. Use boot time instead of wall time for the
snapshot value and print it out in decimal.
Change-Id: Ic0dedbadf9b7d799a3cb4d5917787bccd350c90a
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
When taking a snapshot from a ramdump it is usually important that
we get the last IB2 that was executed so if we find it during
parsing put it in the static snapshot region as opposed to the
dynamic one.
Change-Id: Ic0dedbad878b961ebe29afe9145443e9bfd0bc6a
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
If there are a lot of sequential commands in the ringbuffer for the
same context there is a high probablity that some of the memory
for the older commands has been reused and so the contents may not
be reliable. Increase robustness by making sure we don't parse the
same IB over and over again and be more tolerant of errors and
strange sizes.
Change-Id: Ic0dedbade8504ea258ceea40593cb75e7d6211c8
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
When debugfs entity "ft_user_control" is set, it enables userspace to
control GPU fault tolerance policy. By default it is disabled.
Change-Id: Ia93ccf6d9595f6e7cda71863f9067819aaa0a68d
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
GPU pagefault policy can be set to below options:
a) enable/disable GPU HALT on pagefaults
b) log only one pagefault per page.
GPU pagefault policy can be controled using panel file
and debugfs.
Change-Id: I0613e0d087fc45eb5e519e1c834a33f32376672f
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
Log only one pagefault error per page. If we log all the
pagefault errors it causes lot of messages in kernel logs
and sometimes leads to watchdog timeouts.
Change-Id: If8ee4ee28f6289277eb98d4d8ba73dfda6fc1133
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
If userspace requests no fault tolerance using context flag
KGSL_CONTEXT_NO_FAULT_TOLERANCE, do not attempt denial of
service detection and do not attempt fault tolerance when
a hang is detected.
Change-Id: I5de0e4ae6e8407365725c30fcb0e00a601460ebc
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
Allow userspace to control fault tolerance policy,
this allows fault tolerance policy to be controlled
using panel file.
Change-Id: I991edf6f082384bc69454058fe5df3b5f535aa4c
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
Full postmortem logs take 600ms to 1sec and this affects
fault tolerance time. Reduce postmortem logs and print
only necessary information. Full postmortem logs can
be enabled using debugfs.
Change-Id: I9039746bbdcf67ceebc02f21e84ab873531035a5
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
GPU fault tolerance should take minimal time,
set the GPU clock to max freqency to acheive this.
When we set GPU clock to max freqency we vote for
high CPU clock and bus bandwidth.
Change-Id: I98a3f7c24f0939c08f9b12257817d17414e3f128
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
When expired global timestamp is not progressing,
check if an IB is talking more than 2 seconds
with no update in GPU read pointer, IB1 base
IB1 size, IB2 base, IB2 size and detect it as a
long running IB. Invalidate the context of the IB
and do not attempt fault tolerance on this context.
Change-Id: I7e9780f5aed73a5c8807069aa506bec16a1576b4
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
Store process id and name in context structure, this
information is used to print the process name when
when we detect a error in the driver, example
GPU fault tolerance, commands from bad context
in issueibcmds.
Change-Id: I6acd63933754fdec80d8274d3e1380ff2c23770e
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
Renaming recovery to fault tolerance and modifying
the functions and log messages accordingly.
Change-Id: I5f249806026ac514c4aff7da45c3a4e8cc2f8c34
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
Recovery algorithm is changed to
step 1: retry the same commands that hung the GPU
step 2: if step 1 fails, nop just the IB that hung the GPU
and retry
step 3: if step 2 fails, skip commands in current context
till end of frame and retry
step 4: if step 3 fails mark context as bad and execute
the remaining commands from good contexts.
Previously we used to return -EDEADLK when recovery succeeds,
this is the same error code if the context is not recoverable.
With new policy if recovery succeeds we return -EAGAIN so that
userspace treats recovered context differently from the ones
that are not recoverable. If recovery fails we mark the context
as bad and return -EDEADLK.
Change-Id: I9fa3c40801964186866b6002e62f19cf6aa41361
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
When a hang happens locate the index in ringbuffer from where replay
of recovery needs to be re-tried during initialization of recevery.
If the hang detection is false then hang reported by postmortem and
snapshot may be different, but recovery will always start from the
index obtained during initialization. Also, this allows us to capture
data into the snapshot that is in line with the point from where
recovery occurred.
Change-Id: Ie3877f0c81dad2ede888db7cb7d3ff13e2761763
Signed-off-by: Shubhraprakash Das <sadas@codeaurora.org>
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
In current design waittimestamp can trigger hang detection
in less than 50ms. Check if hang detection was called less than
50ms ago, if true do no attempt hang detection.
Change-Id: Iaeb2501316c4194957fbe211b6023eeb869bd0d5
Signed-off-by: Tarun Karra <tkarra@codeaurora.org>
For the playback use of compressed pass through on HDMI device,
post processing features are not supported and should not be sent to
DSP. Hence the change is now to send the volume only for the playback
of decode and render in DSP and skip for compressed pass through
CRs-Fixed: 442140
Change-Id: I2125f08baf1676f6bb1d13463a1f01dfe280963f
Signed-off-by: Subhash Chandra Bose Naripeddy <snariped@codeaurora.org>
Add wcnss reset interface, this interface generates an FIQ to Riva.
Upon receiving this interrupt, Riva will execute the ERR_FATAL
routine and invoke SSR.
Conflicts:
drivers/net/wireless/wcnss/wcnss_wlan.c
include/linux/wcnss_wlan.h
Change-Id: I74c293121cb69b312ed980902a12019adb700792
Signed-off-by: Sameer Thalappil <sameert@codeaurora.org>
Sanitize command should never be issued with timeout argument of 0,
cause then it invokes timeout.
Moreover, sanitize command should not be called from routines that
does erase/trim/secure-discard commands, but only from a routine that
issues the sanitize spesifically.(mmc_blk_issue_sanitize_rq)
Change-Id: Ia3e3737f4bad149edc2248b8a30d74a9bd812be1
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
(cherry picked from commit 9a87427db940047907aee0dc86965645656f1ca9)
Signed-off-by: Neha Pandey <nehap@codeaurora.org>
There is maximum number of COPPs ASM session can be routed
to even though there are far more number of AFE ports.
Add the check to make sure caller is not routing to more
than ADM can actually support.
Change-Id: Iee97dd7e5d21820c87244a422fa252c2228185f9
Signed-off-by: Patrick Lai <plai@codeaurora.org>
Remove unnecessary check as it should be reasonable
to assume valid parameter would be issued through
debugfs interface
Change-Id: I1a23d8c5324b813bf9501c5462684234926b5dcc
Signed-off-by: Patrick Lai <plai@codeaurora.org>
AUDIO_AAC_DUAL_MONO_PL_PR is 0 anyway and dual mono
setting is defined as unsigned value. Hence,
there is no need if new setting is less than
AUDIO_AAC_DUAL_MONO_PL_PR.
Change-Id: I2e18a875e211d05d100d827fcd5b67545de168ac
Signed-off-by: Patrick Lai <plai@codeaurora.org>
PID ID 9079 is a clone of 9048 intended for use over HSUSB
connection. The net device name for its interfaces will be
prefixed "rmnet2_usbX". The number and names of its control
interfaces will need to be initialized from userspace.
Change-Id: I98d25f23246c93f96a934dcfc2411c4b6d970b15
Signed-off-by: Jack Pham <jackp@codeaurora.org>
Currently rmnet_usb supports a maximum of 4 net interfaces on a
single device at a time. This change allows the possibility of
multiple concurrent USB devices with RmNet interfaces to be
supported. The number and names of devices supported are now
specified by the following new module parameters:
* no_rmnet_devs - default: 1
* no_rmnet_insts_per_dev - default: 4
* rmnet_dev_names - default: "hsicctl"
* rmnet_data_init - initializes device nodes with above params
Userspace must now write to "rmnet_data_init" in order for the
control device nodes to be created.
Change-Id: I1e36d85b7fab4290ef64ef635162eb6e4de84e50
Signed-off-by: Jack Pham <jackp@codeaurora.org>
New characterization data shows that vdd_dig voltage for L2@384MHz
can be lowered to the low level. Update the data in this patch.
Change-Id: Ia80ee9397b3b433b96f2bce49a86d5ef7b024fc5
Signed-off-by: Tianyi Gou <tgou@codeaurora.org>
(cherry picked from commit f684b386cbe45229e12a0eb778b933c69bceb0d4)
Signed-off-by: Neha Pandey <nehap@codeaurora.org>
Allow software to selectively read/write MDP color calibration registers.
This is needed to allow calibration tool to individually configure
registers.
Change-Id: I5db2dcbca0b3433c7fedf9033ceff0cf6a5e0e1f
Signed-off-by: Carl Vanderlip <carlv@codeaurora.org>
(cherry picked from commit afc6fbf054d079761d5c2ee85242e56d640ee409)
Signed-off-by: Pravin Tamkhane <pravint@codeaurora.org>
A new SMD control channel is opened with WCNSS, currently only the
version information is exchanged thru this channel. WCNSS version is
exposed thru /sys interface.
Conflicts:
drivers/net/wireless/wcnss/wcnss_wlan.c
Change-Id: I45638bdd992846fa6087b6131180684d30805002
Signed-off-by: Sameer Thalappil <sameert@codeaurora.org>
The shutdown ioctl now gets the system monitor id for the modem to shutdown
from the platform data structure rather than using a hardcoded value.
Change-Id: Ibe7666075903258932014bcf4c79aba0f71184fd
Signed-off-by: Ameya Thakur <ameyat@codeaurora.org>
Some platforms have the ap2mdm_errfatal line shared between
two external modems. Current implementation is to assert that
gpio during the SSR shutdown callback but that does not allow
for independent restart of each modem. On most platforms it is
actually not necessary to assert the ap2mdm_errfatal during
restart of the mdm. It is necessary during kernel panic of the
apps processor but that is already done in the panic handler.
The use case where it is necessary to assert the errfatal gpio
is where the mdm needs to be restarted by the SSR module, for
example on platforms where the internal and external modems need
to be restarted together. This patch adds a platform data parameter
to identify those platforms.
Change-Id: Ie4f5cbed752a04a80f1e4bfb50da03f52e010045
Signed-off-by: Joel King <joelking@codeaurora.org>
For the DSDA2 target with two external modems, independent
subsystem restart of each modem is required. This flag is
used by the mdm driver to prevent it from asserting the
ap2mdm_errfatal during subsystem shutdown for the DSDA2 target,
otherwise both modems would reset because they share the same
gpio.
Change-Id: I27aedaa2b1b8756079710e41d8a4a9572b8ff83a
Signed-off-by: Joel King <joelking@codeaurora.org>
Add support for a secondary instance of an external modem over
the USB/HSIC transport.
Change-Id: I1891ccc14279dfcbcf5e3555a6d42f42049e208a
Signed-off-by: Jack Pham <jackp@codeaurora.org>
Enhance the driver to support (currently up to 2) multiple sysmon
interface instances. This is done by classifying the supported PIDs
as either the first or second allowable instance. Add PID 9079 and
allow it to be instantiated as the second instance. Also make use
of the new USB_DEVICE_INTERFACE_NUMBER macro.
Change-Id: I6f9fbf37d151dcce1d0bcf5204460ce0732730f1
Signed-off-by: Jack Pham <jackp@codeaurora.org>
The SSR reference count now gets decremented only once during a SSR cycle.
Ramdump timeout is now set correctly for each mdm.
Change-Id: I3f1d577805d4a6cc5171a69107f8529a9d165b04
Signed-off-by: Ameya Thakur <ameyat@codeaurora.org>
Use the USB_DEVICE_INTERFACE_NUMBER macro to add the supported
devices and their corresponding interface numbers. This replaces
the cumbersome per-PID bitfield used to map supported interfaces.
Change-Id: I738d79cda929a450086df8fb27247882df8bc696
Signed-off-by: Jack Pham <jackp@codeaurora.org>
Metadata shared input is used to communicate between the
video core and video driver. Allocate it from firmware
heap for specific targets as MFC command heap size is
limited.
CRs-Fixed: 450118
Change-Id: I3d04cf6dd498ae67adcec7a8cba513bd7932d158
Signed-off-by: Shobhit Pandey <cshopan@codeaurora.org>
Signed-off-by: Deepak Verma <dverma@codeaurora.org>
In the use case of audio playback in non tunnel mode, in audio release
unmap is called twice, before and after audio disable. This is
resulting to request DSP to unmap the region twice. Remove the unmap of
the ion region before audio disable.
CRs-Fixed: 443535
Change-Id: Iba44647063c85ab48091888ab5a583b6208dbb3e
Signed-off-by: Subhash Chandra Bose Naripeddy <snariped@codeaurora.org>
