In current code, kgsl_close_device holds device mutex and calls
last_close in case open_count goes to zero. But as part of
last_close call, device mutex is released intermittently to wait
for active count to go to zero and also for dispatcher idle. This
intermittent release of device mutex opens up a window where a
kgsl_open_device call can go ahead and call first_open. As this
first_open is triggered before last_close could finish it messes
up the complete state machine. To fix this, move decrement of
open_count after last_close to allow any open/close happening when
last_close is in progress to just increment/decrement open_count
without triggering first_open or last_close.
Change-Id: Ibb2567e3ef0dc1864f2bfca346b0682792d0b675
Signed-off-by: Deepak Kumar <[email protected]>
Move IPC_LOG inside mutex lock and bfefore deleting
cb_txn inside process_tzcb_req to make sure that
we are not accessing cb_req value that has gone bad.
Change-Id: Icc04bab52d1364d883b0a4bb07ff4cf2cc8ce6f8
Signed-off-by: Anmolpreet Kaur <[email protected]>
Few functions can be handled directly without having
the gpuid checks. So cleanup these functions to avoid
unnecessary gpuid checks.
Change-Id: If5b88adb4d21d327a39dc92d657d1bf5066f6667
Signed-off-by: Hareesh Gundu <[email protected]>
Cleanup debugfs, dev nodes and gpios on driver removal.
Change-Id: Icd18103ef0d968edb9f2ab254321a14323dd1c5a
Signed-off-by: Sneh Shah <[email protected]>
Report the QC3+ adapter power-rating via "power_now" in the usb
power-supply class.
Change-Id: I0d0e9a3e9d2f5d57f5e8696dea8f4e17a8f2df5f
Signed-off-by: Ashish Chavan <[email protected]>
Added process exit checks when adding
file init memory to ramdump collection
to fix use after free for file init memory.
Change-Id: I8991d27e4c83f572a6baacc478a09eb42ea76d8f
Acked-by: Krishnaiah Tadakamalla <[email protected]>
Signed-off-by: Jeya R <[email protected]>
With Port Trigger feature, the DL traffic forward ports get opened
once the UL traffic is sent on respective trigger ports.
Change-Id: I7fc67a48ab4d079aa3e7c1b5fa81b4611f54493f
Acked-by: Vinisha Varre <[email protected]>
Signed-off-by: Paras Jain <[email protected]>
Enable thermal framework features and other thermal drivers like QMI
sensor and mitigation driver for sdxlemur.
Change-Id: I1e4a45ada65b3bff83cfc902cc059547787fbce2
Signed-off-by: Ram Chandrasekar <[email protected]>
Add support for new QMI sensors. Here is the list of sensors that are
newly added,
qtm_therm
modem_bcl_warn
sdr0_pa0
sdr0_pa1
sdr0_pa2
sdr0_pa3
sdr0_pa4
sdr0_pa5
sdr0
sdr1_pa0
sdr1_pa1
sdr1_pa2
sdr1_pa3
sdr1_pa4
sdr1_pa5
sdr1
mmw0
mmw1
mmw2
mmw3
mmw_ific0
This will enable these sensors to be registered with thermal framework.
Change-Id: I5ea49478afee82f7058247ee7e0e505955423e72
Signed-off-by: Ram Chandrasekar <[email protected]>
Channel release function is not making intent wait condition true
before calling wake_up, due to this thread waiting on intent is
stuck and does not wake up until timeout happens. By this time
channel context is freed and this casues finish wait to fail due
to access of wait queue in freed channel context.
Mark intent request complete before calling wake_up.
CRs-Fixed: 2914979
Change-Id: Ia4162c9e62f0f5a7bd04ca0e09c2cd58319ef341
Signed-off-by: Deepak Kumar Singh <[email protected]>
Some platforms do not support DRV suspend and resume.
On these platforms the access to MSI config is not
prevented as part of suspend leading to NOC faults.
Fix this by preventing the access to the same as part
of suspend, and allowing it as part of resume.
Change-Id: I435d9f1bad3843154f5a4ebff21cd8e358802817
Signed-off-by: Sriharsha Allenki <[email protected]>
SCMI handle will remain NULL in case of SCMI failures during
initialization.
Update all store functions to check for NULL before accessing
vendor ops structure.
Change-Id: Ied56f62785579078fb06fb526e29aaef9d45e00d
Signed-off-by: Santosh Mardi <[email protected]>
Add an interface to enable/disable boost concurrency. When
enabled the charger in buck-mode can act as a boost over a
separate path. The feature is supported only on PM5100 and
can be enabled using "qcom,concurrency-mode-supported" flag.
Change-Id: I3ab78af2123add2c4d5cba96500e1576b3c7deb3
Signed-off-by: Ashish Chavan <[email protected]>
Propagate the pcie aspm settings from original auto (both metal
and gvm) defconfigs to make sure they're aligned with the
settings on top of the newly adopted kernel branch.
Change-Id: I260cd9b236e715183d11bbc3e625641b57c93ee8
Signed-off-by: wadesong <[email protected]>
Signed-off-by: Hans Chang <[email protected]>
If PCIe enumeration fails with primary PCIe RC port, retry with
other PCIe RC ports if any in the list from device tree using
the same logic.
Change-Id: I4e803f74dfeda914a9944ab2d14a4055fc223301
Signed-off-by: Yue Ma <[email protected]>
