Upon allocation failure iommu_domain_alloc() returns NULL.
Driver is not checking for NULL value, which leads to NULL
pointer dereference later.
Change-Id: I9788412af3a34aca8fa8519e7ae6f3349f263a1e
Signed-off-by: Hemant Kumar <hemantk@codeaurora.org>
Updating the GPU SMMU settings to match the settings
used during pre-sil testing.
Change-Id: I70e0b87c3575f5778726ae2855d68bf0e560206a
Signed-off-by: Jonathan Wicks <jwicks@codeaurora.org>
When the regulators are not managed by SDHCI host driver (i.e.,
when host->vmmc and host->vmmcq are absent), get the regulator
current capabilities through a new host op get_current_limit().
Change-Id: I097349ad4d2e8ffbafbeb5fe65910894fe909bd6
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Signed-off-by: Sayali Lokhande <sayalil@codeaurora.org>
Add support to allow setting various flags on the branch clock
pertaining to PERIPH, RETAIN_PERIPH, RETAIN_MEM set and clear by clients
which require this support.
Change-Id: I59ddc1b3b677bd0d7fa838afc9a6cbfc10f98409
Signed-off-by: Taniya Das <tdas@codeaurora.org>
The consumers of the dummy clock provider would require support for dummy
resets, which would make a dummy assert/deassert reset signals. The dummy
reset controller registers for 0-100 reset signals.
The consumers node would have the reset phandle and reset specific pairs
and the reset signal names.
test_node {
....
resets = <&clock_gcc GCC_PHY_BCR>;
reset-names = "phy_reset";
};
Change-Id: I242d2f36a6b4af53c658dbaa7a0f32e46cd3af33
Signed-off-by: Taniya Das <tdas@codeaurora.org>
With zcache using zbud, strange locking scenarios are
observed. The first problem seen is:
Core 2 waiting on mapping->tree_lock which is taken by core 6
do_raw_spin_lock
raw_spin_lock_irq
atomic_cmpxchg
page_freeze_refs
__remove_mapping
shrink_page_list
Core 6 after taking mapping->tree_lock is waiting on zbud pool lock
which is held by core 5
zbud_alloc
zcache_store_page
__cleancache_put_page
cleancache_put_page
__delete_from_page_cache
spin_unlock_irq
__remove_mapping
shrink_page_list
shrink_inactive_list
Core 5 after taking zbud pool lock from zbud_free received an IRQ, and
after IRQ exit, softirqs were scheduled and end_page_writeback tried to
lock on mapping->tree_lock which is already held by Core 6. Deadlock.
do_raw_spin_lock
raw_spin_lock_irqsave
test_clear_page_writeba
end_page_writeback
ext4_finish_bio
ext4_end_bio
bio_endio
blk_update_request
end_clone_bio
bio_endio
blk_update_request
blk_update_bidi_request
blk_end_bidi_request
blk_end_request
mmc_blk_cmdq_complete_r
mmc_cmdq_softirq_done
blk_done_softirq
static_key_count
static_key_false
trace_softirq_exit
__do_softirq()
tick_irq_exit
irq_exit()
set_irq_regs
__handle_domain_irq
gic_handle_irq
el1_irq
exception
__list_del_entry
list_del
zbud_free
zcache_load_page
__cleancache_get_page(?
This shows that allowing softirqs while holding zbud pool lock
can result in deadlocks. To fix this, 'commit 6a1fdaa36272
("mm: zbud: prevent softirq during zbud alloc, free and reclaim")'
decided to take spin_lock_bh during zbud_free, zbud_alloc and
zbud_reclaim. But this resulted in another deadlock.
spin_bug()
do_raw_spin_lock()
_raw_spin_lock_irqsave()
test_clear_page_writeback()
end_page_writeback()
ext4_finish_bio()
ext4_end_bio()
bio_endio()
blk_update_request()
end_clone_bio()
bio_endio()
blk_update_request()
blk_update_bidi_request()
blk_end_request()
mmc_blk_cmdq_complete_rq()
mmc_cmdq_softirq_done()
blk_done_softirq()
__do_softirq()
do_softirq()
__local_bh_enable_ip()
_raw_spin_unlock_bh()
zbud_alloc()
zcache_store_page()
__cleancache_put_page()
__delete_from_page_cache()
__remove_mapping()
shrink_page_list()
Here, the spin_unlock_bh resulted in explicit invocation of do_sofirq,
which resulted in the acquisition of mapping->tree_lock which was already
taken by __remove_mapping.
The new fix considers the following facts.
1) zcache_store_page is always be called from __delete_from_page_cache
with mapping->tree_lock held and interrupts disabled. Thus zbud_alloc
which is called only from zcache_store_page is always called with
interrupts disabled.
2) zbud_free and zbud_reclaim_page can be called from zcache with or
without interrupts disabled. So an interrupt while holding zbud pool
lock can result in do_softirq and acquisition of mapping->tree_lock.
(1) implies zbud_alloc need not explicitly disable bh. But disable
interrupts to make sure zbud_alloc is safe with zcache, irrespective
of future changes. This will fix the second scenario.
(2) implies zbud_free and zbud_reclaim_page should use spin_lock_irqsave,
so that interrupts will not be triggered and inturn softirqs.
spin_lock_bh can't be used because a spin_unlock_bh can triger a softirq
even in interrupt context. This will fix the first scenario.
Change-Id: Ibc810525dddf97614db41643642fec7472bd6a2c
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
Set correct permissions for sysfs nodes.
1. Show method of write only sysfs attributes is set to NULL
since the attributes are write-only.
2. Store method of read only sysfs attributes is set to NULL
since the attributes are read-only.
Change-Id: I93b14e22da57a8a47a02080a05a9b8c666970272
Signed-off-by: Abinaya P <abinayap@codeaurora.org>
This deliberately changes the behavior of the per-cpuset
cpus file to not be effected by hotplug. When a cpu is offlined,
it will be removed from the cpuset/cpus file. When a cpu is onlined,
if the cpuset originally requested that that cpu was part of the cpuset, that
cpu will be restored to the cpuset. The cpus files still
have to be hierachical, but the ranges no longer have to be out of
the currently online cpus, just the physically present cpus.
Change-Id: I3efbae24a1f6384be1e603fb56f0d3baef61d924
[ohaugan@codeaurora.org: Port to 4.4]
Git-commit: f180bcac788464a0baf3d79d76dd86d6972ea413
Git-repo: https://android.googlesource.com/kernel/common/msm.git
Signed-off-by: Olav Haugan <ohaugan@codeaurora.org>
The location of some fuel gauge SRAM registers have changed in
PMICOBALT v2. Add a new SRAM register map for v2 to handle these changes.
Change-Id: I1fcfce8e56b1d4e8b8f54457193cd547fb5e3de7
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
IPA-driver should request for
aggre2_noc_clk via bus driver.
Change-Id: I2f44b44a27ae93c6dd46cc6829d68ea1117ba237
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
Wlan need to pass the VA to let the IPA-driver
decide when to suspend the WDI pipe in SMMU
enabled scenario.
Change-Id: I00ae7f088c7411c8a6420f5e7ab12276d65fd269
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
This change is to fix issues reported when OUTPUT and OUTPUT2
ports enabled in Venus. In this case following are the rules
- s_fmt on OUTPUT port updates OUTPUT port dimensions.
- s_fmt on CAPTURE port updates CAPTURE port dimensions.
- reconfig dimensions are always updated on OUTPUT port.
- All clocks scaling, bus scaling uses max of both ports.
- DCVS is based on reconfig, CAPTURE and OUTPUT port dims.
CRs-Fixed: 1044772
Change-Id: Ib45424c70f995d856ad27107ae3ea24a59ca4e74
Signed-off-by: Praneeth Paladugu <ppaladug@codeaurora.org>
Signed-off-by: Abdulla Anam <abdullahanam@codeaurora.org>
Set the status property of the qcom,memlat-mon-cpu4 device. This is
required for partial binned devices.
Change-Id: I246a5f5149e492aba4f7c3920296d4bd3d873de0
Signed-off-by: David Keitel <dkeitel@codeaurora.org>
Using first fit algorithm to select the remote rx intent from the
list is not optimal way.
Optimize the selection of intent from list using the best fit algorithm.
CRs-Fixed: 1058750
Change-Id: I7b2a70188975b75a0fbcd2a6cb26f28cc0258532
Signed-off-by: Dhoat Harpal <hdhoat@codeaurora.org>
Antenna Sharing GPIO need to be toggled between QCA PCIe and SDIO
devices based on the requirement. If SDIO device is to be operational,
this GPIO should be turned ON high.
Expose API to CLD Driver to control this GPIO for dual-wifi platforms.
CRs-Fixed: 1013494
Change-Id: I271ba273a4c70717f066d219ff02e8f0f56e2ac3
Signed-off-by: Komal Kumar <kseelam@codeaurora.org>
Dynamic Mode Change Feature Request, needs to power ON
QCA Chip during system bootup to get Firmware Capabilities.
Today the implementation is only during driver load, platform driver
power up the chip and call the device driver probe and call device driver
remove and cut the chip power during driver unload.
As part of this feature request, driver needs API to power down and
power up chip after ifconfig down and before ifconfig up.
CRs-Fixed: 1009901
Change-Id: I8e964c6bb8d6d8fb6f4bfba17a6b24b97509953c
Signed-off-by: Komal Kumar <kseelam@codeaurora.org>
Interrupt Storm is observed on the APPS side, when PCIe link down is
simulated, as FW keeps generating interrupts to host, as it's not aware
of the PCIe link down.
When PCIe link down is identified, wlan platform driver schedules work
to start recovery. During Recovery we cut down power to QCA6174, which
stops the interrupt storm. During the window where platform driver
schedule work and the work actually gets schedules, interrupt storm
happened and the APPS triggered watchdog.
Change-Id: Ica3c32007bc9b8acce79c5bb2175790ff27180d0
CRs-Fixed: 967956
Signed-off-by: Komal Seelam <kseelam@codeaurora.org>
The WIFI PCIe Clock is controlled by WLAN_EN gpio.
On Dual-WiFi Platforms, QCA WIFI SDIO chip clock sourced by PCIe Clock.
Turning off WLAN_EN gpio would result in SDIO WIFI Functional Failures.
Control the gpio, based on the platform.
CRs-Fixed: 990796
Change-Id: I1399cd8aa53d1c6fd6b8ff0c5cce66b978770dbd
Signed-off-by: Komal Seelam <kseelam@codeaurora.org>
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
Fix issues raised by static analyzer:
1. initialize "retval" before being returned from the driver function.
2. Check return value of the function create_singlethread_workqueue()
and return -ENOMEM if it failed. If creation of the workqueue
failed, then both queue_work() and queue_delayed_work() calls will
later crash, as they expect the workqueue pointer to be a non-NULL
value. Also add clean-up code for this.
3. check return value of the snprintf() call as array 'buf' of size
16 may use index value(s) 16...20.
CRs-Fixed: 995687
Change-Id: I89d9f7cacbcf23de43a7e96556d1ac65911126d6
Signed-off-by: Shantanu Jain <shjain@codeaurora.org>
vdd-wlan-xtal-aon-supply(LD0-4) need to be voted by cnss platform
driver to keep wlan XO(LDO-5) ON during APPS suspend on mdmcalifornium
platforms when wlan is turned ON as per Power UP sequence. Currently
LDO-4 is only voted by USB driver when USB is connected. On removing USB,
LDO-4 is turned off resulting LDO-5 turned off.
When WLAN is turned on removing LDO-5 vote resulted in beaconing failures
on mdmcalifornium platforms. Hence add a dummy vote to LDO-4 when WLAN is
turned ON.
Change-Id: I0c6e7a6c2dc6a51fb19a9d31ceb6823fe9a05ae2
CRs-Fixed: 996537
Signed-off-by: Komal Seelam <kseelam@codeaurora.org>
Synaptics touch controller does not go to sleep in suspend because
L6 is still enabled. Disable L6 in suspend to ensure CTP go into
sleep mode.
Change-Id: I40f3f380c1a85c292029d4e22e49bebf25bd198c
Signed-off-by: Mao Li <maol@codeaurora.org>
During device resume, the touch resume function is called after
display resume. Touch resume function will take about >200ms.
Defer the touch resume function to a workqueue to reduce the total
device resume time. An optional DT property is added to enable this
on targets that need this feature.
This change has been merged. Now, touch driver was updated to V2.6,
so, merge this change to V2.6 driver.
Change-Id: Ica477f35bd4fda59eb49c9b4f5e88b460366c761
Signed-off-by: zhaoyuan <yzhao@codeaurora.org>
Because of a change in the underlying bus driver, the secure touch
layer in the input drivers is now required to control directly the
clocks which are needed during the touch sessions. The clocks which
are associated with the underlying I2C bus are clk_gcc_blsp1_ahb_clk
and clk_gcc_blsp1_qup3_i2c_apps_clk. These clocks are turned on/off
when the secure touch session is started/ended.
CRs-Fixed: 990820
Change-Id: Ie2f6f5b736b27792e6e4dc2c39064e0ebdc03d7b
Signed-off-by: Shantanu Jain <shjain@codeaurora.org>
Signed-off-by: Abinaya P <abinayap@codeaurora.org>
Update F12 touch report function with adding input_sync to
correct the pointer reporting status.
Change-Id: Iefe24525a9551736a6044697b447a25e2c5c65ca
Signed-off-by: Mao Li <maol@codeaurora.org>
This patch creates two sysfs files for secure touch - secure_touch
and secure_touch_enable, which will be accessed by secure UI app.
We also define the store and show function for these sysfs files.
Change-Id: Id43118120d4a1f0682904f48b3584d3ba62ef1dd
Signed-off-by: Shantanu Jain <shjain@codeaurora.org>
8953 QRD supports Synaptics touch controller. For synaptics_dsx_2.6 driver
to work with the TrustZone secure touch, all the touch interrupts must
be forwarded between the Linux Kernel and the TrustZone input driver.
Add APIs and configuration to support this.
Change-Id: I47520f9ba91b7645dff2c11b4c58a0b7aeed2765
Signed-off-by: Shantanu Jain <shjain@codeaurora.org>
Add support for getting additional mappings
from dtsi in order to create a mapping on
IPA wlan SMMU context bank.
Change-Id: I92922e494c745ea4c6bf4f473515dea4eccbd62e
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
Add diag over glink support to communicate to WDSP.
Change-Id: I60cce19bd82030990a1267b81ed28c25ac8601ec
Signed-off-by: Sreelakshmi Gownipalli <sgownipa@codeaurora.org>
IPA-driver should request for aggre2_noc_clk
via bus driver, add entry in dtsi.
Change-Id: I98726deff99b373344fe75f8bdcb11fcad83e621
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
The lineptr interrupt is expected to be triggered and have the
ability to be updated instantly whenever the panel is fetching
data. Previously, enabling lineptr was tied to kickoff and disabling
to pp_done work. Add capability to change the lineptr value instantly,
when auto-refresh is enabled and avoid disabling lineptr during that time.
Once enabled, lineptr can only be disabled when 0 is written to the node
while auto-refresh is enabled.
Change-Id: I9a1a478c857efd988984a0efb0a2b6475030c7ec
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
IPA might have Information leak and device crash due to
kernel heap overread in IPA driver when processing
WAN_IOC_ADD_FLT_RULE_INDEX ioctl. The fix is to add
check on max number of filter rules send to modem.
Change-Id: I454e04d05cfcb7af8fc4bd2b4a1bade55c4684d0
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
Compile Video driver as a Loadable Kernel Module
on msmcortex.
CRs-Fixed: 1058390
Change-Id: I166e3e3854439c8ae15fd1e5a10e671da76ef044
Signed-off-by: Chinmay Sawarkar <chinmays@codeaurora.org>
qseecom_send_modfd_cmd converts ION buffer's virtual address to
scatter gather(SG) list and then sends them to TA by populating
SG list into message buffer. As the physical memory address in
SG list is used directly by TA, this allows a malicious TA to
access/corrupt arbitrary physical memory and may lead to the
process gaining kernel/root privileges. Thus, make changes to
have the QSEEComm driver passing a list of whitelist buffers
that is allowed to be mapped by TA, and the QSEE kernel, in turn,
should add checks to the register_shared_buffer syscall to make
sure the shared buffers an application is mapping falls within
one of these whitelist buffers.
CRs-fixed: 1021945
Change-Id: I776ead0030cad167afcf41ab985db7151a42d126
Signed-off-by: Zhen Kong <zkong@codeaurora.org>
Adjust phy-qcom-ufs-qmp-v3 settings based on post silicon validation.
Change-Id: Ide0ca4714679bb576a8069bdd7720507074ccc47
Signed-off-by: Venkat Gopalakrishnan <venkatg@codeaurora.org>
Buck damage was seen to go away once the dead time were increased
in addition to increasing the OOB comparator threshold.
CRs-Fixed: 1055113
Change-Id: I8979d203fc7c35ec23e23874c8dd8f980a1cc619
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Commit 85635da429 ("qcom-charger: smb138x: fix buck damage on
v1.1 hardware") introduced wa_flags as the first member
in the struct. However there are implict assumptions in the driver that
the first member is the library structure. This causes memory corruption
and an eventual watch dog bite reset.
Fix it by ensuring that the library struct is explicitly initialized
from the parent struct.
CRs-Fixed: 1055113
Change-Id: I38474a228ec7b245dc330b4a37847c9f1bf29c2a
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
'commit 5b42bd749e8c ("usb: gadget: gsi: Optimize TRB's buffer allocation
functionality")' removed handling of EVT_HOST_READY event with state
STATE_INITIALIZED to optimized USB GSI state machine. This handling is
required for specifically when RNDIS interface is being enabled through
device manager, otherwise RNDIS interface doesn't get private IP address
back. Fix this issue by adding back handling of EVT_HOST_READY event with
state STATE_INITIALIZED.
CRs-Fixed: 1050838
Change-Id: I7010005b37b4709f54de58d52a56cb7f5e88bd08
Signed-off-by: Mayank Rana <mrana@codeaurora.org>
header change for CNE to pass gateway
info to IPA
Change-Id: I0f0a84a12c3ba514f7309b16c73f5b8ed631626d
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
Fix a static analysis issue at IPA driver,
IPA unit-tests I/S. Do not dereference a pointer even
if it can potentially be NULL.
CRs-Fixed: 1055710
Change-Id: Ie7eff49f136fe429ef10c6067ce7d8aae2694d3a
Signed-off-by: Ghanim Fodi <gfodi@codeaurora.org>
Focaltech's CTP FT6436 is able to behave like a proximity sensor.
Enable the driver support this new feature.
Also cleared the chekpatch warning on 3.18 kernel.
Change-Id: I7a6ec3a387536c512637b0bd8dab95e7cceca212
Signed-off-by: Mao Li <maol@codeaurora.org>
Signed-off-by: Sudhakar Manapati <smanap@codeaurora.org>
Signed-off-by: Abinaya P <abinayap@codeaurora.org>
Update the Display-Port PHY and PLL configuration
with the recommended settings. Remove the
support for 9.72Ghz VCO frequency. Update the divider
settings to support the new frequency plan.
Update the Phy Aux settings and voltage/pre-emphasis
settings according to recommended configuration.
Change-Id: Ic4d206da3dc6b45214e7601e7556cfb0bef81a7d
Signed-off-by: Chandan Uddaraju <chandanu@codeaurora.org>
Add compatibility checks for the new LM configs sent by dest
scalar with the DSC slice width/height during validation. Without
this check, the DSC config check would fail at a later point during
ctl start.
Change-Id: I55e483be5629cfe26cca3b740d206845a1caa903
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
msmcobalt has a different command mode MISR CRC register offset. Correct
the offset for the MISR control and signature registers, and correct the
block_id by lookup the INTF number from CTL.
CRs-Fixed: 1049910
Change-Id: If4fb44d217ed84067a0732334523ddb1e435e4e7
Signed-off-by: Benjamin Chan <bkchan@codeaurora.org>
sysctl_sched_wake_to_idle is a means to allow or disallow a global
task placement preference for idle CPUs. It has been unused thus
far since we've preferred to use a per-task flag instead to control
placement for individual tasks. Using this global flag, however, does
allow greater flexibility for testing and system evaluation.
Incorporate sysctl_sched_wake_to_idle in the placement policy.
Change-Id: I7e830bc914eb9c159ae18f165bc8b0278ec9af40
Signed-off-by: Syed Rameez Mustafa <rameezmustafa@codeaurora.org>
Do the aggregation for frequency only when the total group busy time
is above sched_freq_aggregate_threshold. This filtering is especially
needed for the cases where groups are created by including all threads
of an application process. This knob can be tuned to apply aggregation
only for the heavy workload applications.
When this knob is enabled and load is aggregated, the load is not
clipped to 100% @ current frequency to ramp up the frequency faster.
Change-Id: Icfd91c85938def101a989af3597d3dcaa8026d16
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
The load reporting during frequency alert notifications is broken under
load aggregation. When aggregation is enabled, the total group busy
time is accounted towards the maximum busy CPU of a frequency domain.
If this CPU has a notification pending, it's group busy time alone is
accounted and other CPU's group busy time is completely ignored.
Similarly if any CPU other than maximum busy CPU has a pending
notification, its group busy time is accounted twice.
Maintain the frequency alert notification flag per frequency domain.
When the notification is pending, don't clip the load to 100% @ fur
for any of the CPUs in the frequency domain.
Change-Id: Iebc7d74d6fafa20430fa1c7d80f34a6ab198832d
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
When sysctl_sched_enable_thread_grouping is set to 1, any new tasks
created are put in the same group as their group leader.
Change-Id: If1837dd7c8120c8b097cfffa1dc52eb4781f1641
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
