As snapshot can be recollected and snapshot size can vary on
each recollection if we add snapshot region in minidump during
each recollection then we need to remove it and add it again
to ensure proper snapshot size is captured in minidump. To
avoid this add snapshot region to minidump once and with complete
snapshot memory size. Also, take care of removing it during close
to ensure minidump doesn't try to access already freed memory region.
Change-Id: I5eded015c0544a03eec835da57b6aaf64defc139
Signed-off-by: Deepak Kumar <[email protected]>
With rapid hotplug test case there could be a corner case when the SD card
can go bad and removed quickly even before we try to reset the card as
part of recovery.
This will lead to reset failure with ENOMEDIUM and since recovery failed
so we set the corrupted_card flag and try to remove the card.
Since corrupted_card will be set so as part of rescan we'll remove the
card and return before setting the bus_refs properly leading to card not
getting detected after this.
Only return after setting host->bus_refs properly when card is set as
corrupted.
Change-Id: Id14790f9b28d8fa58d31b034be48db5486ccfcb7
Signed-off-by: Sarthak Garg <[email protected]>
Having a global PCI device table share between all PCIe ports
is not necessary. There is potential race conditions if required
locks are present. remove the global PCI device table completely
and utilize the bus list for each PCIe port instead.
Change-Id: I48ebc086fcf56e4b9a9496615e4185c3893ae278
Signed-off-by: Tony Truong <[email protected]>
During enumeration, the host can send a clear feature EP_HALT. When
the UDC receives this command, it will cancel any pending requests and
return -ECONNRESET to the function driver. Since the clear EP_HALT
occurs after the cser driver has already submitted the OUT USB reqs,
there is no other method to restart these transfers. Resubmit OUT
reqs specifically when ECONNRESET is seen.
Change-Id: I925170d389048f84579c9b53128effdbef5c9d6a
Signed-off-by: Wesley Cheng <[email protected]>
If none of the "if" conditions are met, an uninitialized
parb_phase will be used. Initialize parb_phase so that it can
be handled properly.
Change-Id: I682ece7330cd257566ea740c4498bc786ae95588
Signed-off-by: Tony Truong <[email protected]>
If a client defines the qcom,iommu-group property, and the iommu group
does not yet exist, of_get_device_group will create a new group with a
refcount of one. The following call to iommu_group_ref_get() increments
the refcount unnecessarily and causes a resource leak.
If a client defines the qcom,iommu-group property, and the iommu group
already exists, modify __bus_lookup_iommu_group() to return with
an additional refcount.
Change-Id: I5c9e62e50737ee4128a591ea2b14a0d47a1be162
Signed-off-by: Patrick Daly <[email protected]>
Fix the compilation issue caused by QCOM_SCM and QCOM_SCM_QCPE
config flags by removing the redundancy of having a separate
qcom_scm-smc_qcpe.c file for adding QCPE support.
Change-Id: Idaa998565464ebeef16990afdc2de2c1a3094a5e
Signed-off-by: katamoni sai abhilash <[email protected]>
Enable Devfreq monitors, devices, and governors on
sdmsteppeauto target.
Change-Id: I8b206eb3f6efa8e52807352f4edb2f32542c90df
Signed-off-by: Shreyas K K <[email protected]>
Enable core hang detect support on sdmsteppeauto targets.
Change-Id: I992252cda7a5d95cc559fdf3635dec49cca2da7f
Signed-off-by: Shreyas K K <[email protected]>
Enable service locator and service notifier support on sdmsteppeauto
target.
Change-Id: I8cd97abd03b9f91178bdc6efce08d18e743d5c08
Signed-off-by: Shreyas K K <[email protected]>
Enable Uncompressed kernel for gvm to get kernel parsed
from host.
Change-Id: I659409112aab8ea78de81e94969053aa2e57fe23
Signed-off-by: Venkata Rao Kakani <[email protected]>
Enable DM_INIT to support dm-mod.create command line parameter
to mount dm-verity enabled system partition.
Change-Id: Ia9afb608d216f0b3295fa2ebb4d6aebc0a497ef0
Signed-off-by: Venkata Rao Kakani <[email protected]>
When OVP fault is detected during WLED driver probe, SW first needs to
confirm that OVP event is sustained over a minimum time duration
before fixing it. After reading fault status and seeing OVP is set,
add time delay equal to OVP_INT_TIMER value, and ensure OVP fault
is still seen, before running auto calibration algorithm.
Change-Id: I4fbc75fa1f63af53fe2e1b5298e0e277cc8a74b8
Signed-off-by: Jishnu Prakash <[email protected]>
Enable Pagescope read feature on sdxlemur also remove the
software workaround that made for pagescope read feature
on qpic 2.0 controllers because this got fixed on sdxlemur.
Change-Id: I61a2e16d308f80096628024030148a2d6e822ea8
Signed-off-by: Pradeep P V K <[email protected]>
Update the priority based on the hw recommendation.
Change-Id: I6a53adf7f6c8b4584d146637fd7a4a094fa2dea8
Signed-off-by: Veera Vegivada <[email protected]>
Enable boot stats driver and boot time marker config
options for gen3auto targets.
Change-Id: If9b24937d2fe1df9f8a3eb5f0e8ee1d0b4c4d28f
Signed-off-by: Shreyas K K <[email protected]>
The 'commit 5adc2a6e19 ("slim-msm-ngd: Disable core irq
in SSR/PDR Down notification")' added irq_enable and disable from
ngd_slim_enable function. irq will be enabled only if msm_slim_qmi_init
success where as disabled by default in else block. If QMI init is
failed we will disable the IRQ twice during SSR testing.
As the interrupt line is not enabled, capability exchange is timing out and
leading to freed memory access inside ngd_xfer_msg.
Move irq enable in ndg_slim_enable without depending on QMI init.
Change-Id: Id32b0d283b1653bb4085231694031b0df126cf4b
Signed-off-by: Chandana Kishori Chiluveru <[email protected]>
Commit 8799bc1e517f ("usb: dwc3: Disable parkmode for Gen1 controllers")
was added to disable the park mode which is a debug feature. But, doing
this indirectly affect the host mode throughput for targets using Gen1
controllers in HS mode.
Therefore enabling park mode only for HS to improve the throughput
numbers in host mode.
Change-Id: Iea9e44d0ed0f9e2f9bba6731cae29229d9b46bb1
Signed-off-by: Udipto Goswami <[email protected]>
In kgsl_mem_entry_create, access map_count only if entry is allocated
successfully to avoid invalid access.
Change-Id: I57bce1aec2da6a27b6d13dbee96ed86a45c9660c
Signed-off-by: Deepak Kumar <[email protected]>
Touch uses gpio as irq in LA. As per latest change from pinctrl,
wakeup bit is disabled/enabled when disable_irq/enable_irq is called.
During transtion from LA to LE, disable_irq is called leading to
wakeup bit getting disabled and in LE Touch interrupts does not come.
So, use disable_irq_wake to disable Touch interrupts while transition
from LA to LE and use enable_irq_wake to enable back Touch interrupts
while transition from LE to LA.
Change-Id: Iad3391af40535d51daeef881c3d62b3ec893fff2
Signed-off-by: Ritesh Kumar <[email protected]>
Spi slave will support only mode1, so set the CPHA parameter
for spi slave.
wrongly added condition by merge conflict, hence removing it.
Change-Id: I37f9b159b01c9253de1a9e8b46e41160fa940660
Signed-off-by: Aniket Randive <[email protected]>
In the current deign, it is expected for the charger driver
to detect OTG and enable it without the USB driver to
explicitly enable it. As this support exists in the driver,
remove the OTG regulator configuration.
Change-Id: I165b144a2a8be88b493142de3a4a645e07f06253
Signed-off-by: Kavya Nunna <[email protected]>
Currently we are accessing snapshot structure after signaling
dump_gate. In corner cases this worker can get preempted
before driver reads this variable, and eventually client
dumps the snapshot and snapshot object gets released.
When worker gets back in execution it will read freed memory
and will case kernel panic.
To fix this race, do completion for dump_gate after finishing
work in worker kgsl_snapshot_save_frozen_objs.
Change-Id: Iaf3c9f9e75a6e3e19653d2d166e2da624796ec98
Signed-off-by: Pankaj Gupta <[email protected]>
Signed-off-by: Deepak Kumar <[email protected]>
