Since the SOF timestamp is only updated at SOF,
so we can't do self-correction at EOF. When the
frames already in sync, but the req is a EOF
triggered req, then the req can't be applied at
SOF, but it will fail to pass self-correction
checking at EOF. Then we will loop this progress,
the req can't be applied anymore.
CRs-Fixed: 2930497
Change-Id: I079cbd3cb9ac9541f049c5369baeee831f7444a2
Signed-off-by: Depeng Shao <depengs@codeaurora.org>
This change adds validity check for ife out res. If ife out res is
NULL we can run into crash or issues.
CRs-Fixed: 2915741
Change-Id: I8722b2a4e2634bda42f0080e00bf09050bbd6b91
Signed-off-by: Vikram Sharma <vikramsa@codeaurora.org>
This change is to reset csid_error_irq_count on flush
to prevent from recovery.
CRs-Fixed: 2926200
Change-Id: I8568fdc679317e41c31f4aeb9cf8c3ae579d8448
Signed-off-by: Ayush Kumar <ayushkr@codeaurora.org>
In Dual tfe scenario, tfe csid drop first frame in the ipp pix
path to sync both master and slave, but rdi frame will go to tfe.
this will cause issues in the camif and rdi sub sampling scenarios.
in sub sampling scenarios both camif and rdi should be subsampled
on same frames. To achieve this drop the rdi first frame from csid
in dual tfe scenarios.
Change-Id: I73e98a6d3d973af93bbe6afef7585a95578f2a2b
CRs-Fixed: 2923045
Signed-off-by: Ravikishore Pampana <rpampana@codeaurora.org>
Add support for YUV422 format 10 bit in CSID and bus drivers.
CRs-Fixed: 2923254
Change-Id: Icecc367d88509b9ed08f5d0cab8c4da788886a0a
Signed-off-by: Gaurav Jindal <gjindal@codeaurora.org>
This change is set the OT values for RT client of Camonc
to 8 for Yupik camera.
CRs-Fixed: 2924987
Change-Id: Icb7f30c23f5899558a41308fe313d46fd8d992a3
Signed-off-by: Ayush Kumar <ayushkr@codeaurora.org>
Pagefault_ops is protected by the mutex_lock to avoid the
list corruption by concurrent access by flush calls.
CRs-Fixed: 2904307
Change-Id: Ia6519a382ace7b89b6bd59a2dacfeca3c31e2705
Signed-off-by: Tejas Prajapati <tpraja@codeaurora.org>
when provider crash occurred, there was a CSID lane overflow
observed due to shutdown sequence from CRM(sensor, csid, csiphy).
To fix the issue need to change the shutdown sequence
(csid, sensor, csiphy).
This change will update the devices sequence in order while registering
and close the sequence accordingly.
CRs-Fixed: 2852076
Change-Id: Ia6d8022e995823bf031400c33528eb8544dc2e29
Signed-off-by: Alok Chauhan <alokc@codeaurora.org>
When camnoc fill level for RDI port goes beyond 25%, notify UMD
with BAD frames to take corrective actions.
CRs-Fixed: 2913934
Change-Id: I5427ce94cdd2a48d8647ff4593cb1605a2f90a37
Signed-off-by: Alok Pandey <akumarpa@codeaurora.org>
There is a possibility of composite groups in free dual
composite group list are not in sequential order. In case
of dual ife use case, if composite groups are not in
sequential manner then we landed into a scenario
where two different IFE with same resource have two
different composite group index, which is not expected.
This commit ensures if it is dual IFE usecase then at the time
of acquiring composite group, both IFEs have same composite
group ID for same resource.
CRs-Fixed: 2912721
Change-Id: I734d1d1660c1e70e0480e2dc0b8377a1c441c634
Signed-off-by: Shravya Samala <shravyas@codeaurora.org>
There is a chance that during acquire time while updating write master,
for few resources io buffers are not allocated. If Wm configurations
are enabled without receiving io buffers then there is a chance of
hitting page fault issues.
This commit takes care of enabling Wm configurations for each
client only if io buffers are allocated.
CRs-Fixed: 2887053
Change-Id: I868bd01dbd896dd86eeec4fc92a8a62e746f79cf
Signed-off-by: Shravya Samala <shravyas@codeaurora.org>
Validate active sequence id before clear bit. This avoid
invalid clear bit scenario if link is created but try to
deactivate with out link activate.
CRs-Fixed: 2909190
Change-Id: Ie12f6bb100e1e12c41b9addf516446f4088ae8a3
Signed-off-by: Ravikishore Pampana <rpampana@codeaurora.org>
Add data rate table in csiphy driver for cphy
lane for holi.
CRs-Fixed: 2913428
Change-Id: I57c6ab503ed6535ae80e2268e76fb29bfccad856
Signed-off-by: shiwgupt <shiwgupt@codeaurora.org>
For recovery of the bubble the request from active list
should be added back to pending request list.
CRs-Fixed: 2887188
Change-Id: I9bd14db82f96d2fda99bb542bd0df4c9c4a06745
Signed-off-by: Tejas Prajapati <tpraja@codeaurora.org>
Dump camnoc fill level registers for overflow and violations
along with the SOF, EPOCH and EOF timestamps.
CRs-Fixed: 2908055
Change-Id: I774140794eae6580f4332cd9cd39b8c65d7197a5
Signed-off-by: Tejas Prajapati <tpraja@codeaurora.org>
KASAN identified global-out-of-bounds issue on:
msm_cam_hw_cdm_dt_match
Due to no null-terminating struct in the array.
CRs-Fixed: 2883523
Change-Id: I36e4811f239993e1e6de158df959157217c28bfe
Signed-off-by: Karthik Jayakumar <kjayakum@codeaurora.org>
This change modifies regulator_get, regulator_put to devres entry.
Devres is linked list of arbitrarily sized memory areas associated
with a struct device. Each devres entry is associated with a release
function. All devres entries are released on driver detach. This
avoids potential leaking of resources on driver detach or having half
broken failure path implementation in probe call which would leak
resources when failure occurs.
CRs-Fixed: 2890861
Change-Id: Ibd46bc37bd6e31aeb2a5ea90ed75a1f596b57bf9
Signed-off-by: shiwgupt <shiwgupt@codeaurora.org>
If req#N in slot idx, and this req meets bubble issue,
then the req#(N+1) in slot (idx+1) may also be applied,
and its status is APPLIED. If req#(N+1) applies failed
after finishing bubble recovery due to some reasons,
but we still can move to next slot since the status of
slot (idx+1) is already APPLIED. This change resets
the status of slot (idx+1) to ADDED if it is APPLIED
during processing error.
CRs-Fixed: 2902809
Change-Id: I32f318569ad8f24fdf0f324603f4ea0036867e98
Signed-off-by: Depeng Shao <depengs@codeaurora.org>
Sometimes, UMD reports nonincrementing shutter request id
issue during monkey test on sHDR usecase, isp reports last
one req in active req list during bubble, so if the active
req count is 2, we will meet this issue.
This change gets the first active req during bubble in case
the active req cnt is 2.
CRs-Fixed: 2885087
Change-Id: I67d1ecfe9c88123e6d0887383961bfc676455a0d
Signed-off-by: Depeng Shao <depengs@codeaurora.org>
