While searching for first hole (get_first_hole) and next free hole
(get_next_hole) from rbtree, rb_first() and rb_next() functions may
return NULL pointer. A NULL pointer check can avoid the NULL pointer
dereference.
Change-Id: I21c918fff3ae995a9d8b2201030401d80ed1fd9e
Signed-off-by: Venkateswara Rao Tadikonda <vtadik@codeaurora.org>
Add documentation and TDM interface changes in msm8996
machine driver to support TDM on msm8996 MTP.
Change-Id: I5ae3c491a57c400deb6a2574ef0e1be90f7f0000
Signed-off-by: Satya Krishna Pindiproli <satyak@codeaurora.org>
This driver is enabled on hypervisor builds for the purpose of
creating a DRM device that DRM Master can open and listen on for
VBlank and Page Flip events.
Change-Id: Ib9b2df5ae0e2c33145663b3f515595e8f433a850
Signed-off-by: Camus Wong <camusw@codeaurora.org>
In case of an exception, there could be a deadlock:
-000|__switch_to()
-001|context_switch(inline)
-001|__schedule()
-002|schedule()
-003|schedule_timeout()
-004|do_wait_for_common(inline)
-004|__wait_for_common(inline)
-004|wait_for_common()
-005|wait_for_completion()
-006|destroy_work_on_stack(inline)
-006|flush_work()
-007|ufshcd_suspend()
-008|ufshcd_runtime_suspend()
-009|ufshcd_pltfrm_runtime_suspend()
-010|pm_generic_runtime_suspend()
-011|__rpm_callback(inline)
-011|rpm_callback()
-012|rpm_suspend()
-013|rpm_idle()
-014|__pm_runtime_idle()
-015|ufshcd_exception_event_handler()
-016|__read_once_size(inline)
-016|static_key_count(inline)
-016|static_key_false(inline)
-016|trace_workqueue_execute_end(inline)
-016|process_one_work()
-017|__read_once_size(inline)
-017|list_empty(inline)
-017|worker_thread()
-018|kthread()
-019|ret_from_fork(asm)
This is caused because in exception handler
pm_runtime_put_sync is invoked which calls ufshcd_suspend.
And ufshcd_suspend flushes the exception handler work
& deadlocks.
Hence, schedule the suspend work in a different context.
Change-Id: I02524cee5ab2b64e0ad5de873d5318aef7b6ac95
Signed-off-by: Asutosh Das <asutoshd@codeaurora.org>
Signed-off-by: Sayali Lokhande <sayalil@codeaurora.org>
Override the property fsmgr_flags by adding the options verify and
slotselect for the vendor node in the fstab so that A/B is enabled and the
partition is verified by the device mapper target dm-verity. Also delete
the system node from the fstab as it is already mounted by the kernel.
Change-Id: Ib65f1d261c95b995f3999ba24517a69488346ca9
Signed-off-by: Binoy Jayan <bjayan@codeaurora.org>
GEM address space range provided to drm mm allocator is not correct.
Start address and size of the address space should be passed to drm mm
initializer. Instead of size, end address of address space range is
passed to drm mm initializer.
msm_gem_get_iova() returns 0, even if memory mapping to virtual
address space is failed. Return proper error value to user space
driver incase of error.
Change-Id: I356a63b873094cb465909437dcdab9513e27b7ff
Signed-off-by: Venkateswara Rao Tadikonda <vtadik@codeaurora.org>
Support usb, lpass and modem clocks in msm8996 virtual clock
frontend driver.
Change-Id: I5ee4f2d951f133a8403442707e633563efdc5c65
Signed-off-by: Zhiqiang Tu <ztu@codeaurora.org>
Enable virtual clock for uart, spi, i2c and sdhci.
Change-Id: Id0dd32c520eeb956390fb1e0e1a4f762c6ae47f9
Signed-off-by: Zhiqiang Tu <ztu@codeaurora.org>
Currently our driver is notifying ETIMEDOUT to PMIC
incase of floating charger. Consider if we are inserting
an SDP charger slowly that time also we will detect
it as floating charger then also we are notifying
ETIMEDOUT irrespective of current that is notified from
gadget driver. Hence notify charging current properly
if enumeration happens even in case of
floating charger so that PMIC can update real charger
type to SDP
Signed-off-by: Sai Krishna Juturi <jsaikrishna@codeaurora.org>
Change-Id: Ibeeeafac0834c9df7099a3f0bad274ddb393de47
* refs/heads/tmp-aed4c54
Linux 4.4.96
Revert "drm: bridge: add DT bindings for TI ths8135"
ecryptfs: fix dereference of NULL user_key_payload
x86/microcode/intel: Disable late loading on model 79
regulator: fan53555: fix I2C device ids
can: kvaser_usb: Ignore CMD_FLUSH_QUEUE_REPLY messages
can: kvaser_usb: Correct return value in printout
can: sun4i: fix loopback mode
scsi: sg: Re-fix off by one in sg_fill_request_table()
scsi: zfcp: fix erp_action use-before-initialize in REC action trace
assoc_array: Fix a buggy node-splitting case
Input: gtco - fix potential out-of-bound access
Input: elan_i2c - add ELAN0611 to the ACPI table
xen/gntdev: avoid out of bounds access in case of partial gntdev_mmap()
fuse: fix READDIRPLUS skipping an entry
spi: uapi: spidev: add missing ioctl header
usb: xhci: Handle error condition in xhci_stop_device()
ceph: unlock dangling spinlock in try_flush_caps()
ALSA: hda - fix headset mic problem for Dell machines with alc236
ALSA: hda/realtek - Add support for ALC236/ALC3204
workqueue: replace pool->manager_arb mutex with a flag
sched: EAS: upmigrate misfit current task
sched: avoid pushing tasks to an offline CPU
sched: Extend active balance to accept 'push_task' argument
Revert "sched/core: Warn if ENERGY_AWARE is enabled but data is missing"
Revert "sched/core: fix have_sched_energy_data build warning"
FROMLIST: kbuild: clang: fix build failures with sparse check
Revert "Revert "UPSTREAM: efi/libstub/arm64: Set -fpie when building the EFI stub""
BACKPORT: efi/libstub: Unify command line param parsing
Conflicts:
drivers/usb/host/xhci-hub.c
kernel/sched/core.c
kernel/sched/fair.c
Change-Id: Ie36ce5de516f02b2d553043009d9afee64e7ff24
Signed-off-by: Blagovest Kolenichev <bkolenichev@codeaurora.org>
Check the ASM buffer index boundary before using it
to access the buffer to avoid out of array bounds error.
CRs-Fixed: 2145996
Change-Id: I97ad482dec0803debd01f19a9aeb531090fc1644
Signed-off-by: vigneshk <vigneshk@codeaurora.org>
As the IMEM approach to store reset reason is deprecated, so use
the spare register in PMIC power-on peripheral to store the oem
restart reason.
Bit-2 to bit-7 of SOFT_RB_SPARE for hard reset reason, so divide
existing range into regions, value 0 to 31 for common defined features
and 32 to 63 values are for OEMs/ODMs specific features.
Change-Id: Ib8e64d6dd5a335325d0469280e41e9d10ec9b893
Signed-off-by: Lijuan Gao <lijuang@codeaurora.org>
Fix OTG enable error path and drop stale votes
on USB_ICL votable.
Change-Id: Ia0fce0ef75bb83ece72cc6a5a8f294b2500a0166
Signed-off-by: Ashay Jaiswal <ashayj@codeaurora.org>
In existing implementation if vmap fails device panics for
NULL pointer access while doing pil_memset_io.
Instead check for the NULL pointer.
Change-Id: I1b9fd084a8c168b70f6e01134e4122445e4873b4
Signed-off-by: Swetha Chikkaboraiah <schikk@codeaurora.org>
Currently when the maximum available current per flash switch is
calculated, it is returned along with the error code. This will
be confusing when a real error occurs. Fix this by returning
error code separately and obtain the current by passing it by
reference.
Change-Id: I9e2a3436e19cc37319316603e6dfa45241df8a49
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
For improved performance, move RX handling from a thread to
a hard IRQ handler. This helps to decrease the chance of missing
incoming back-to-back messages (e.g. Accept -> PS_Ready following
a successful power request) due to the pending RX token in case
the IRQ thread is delayed due to high system load. policy_engine
needs to perform the buffer allocation with the GFP_ATOMIC flag.
Change-Id: Iacb2623280f1ac4de9f483a8abbf5dc055041be7
Signed-off-by: Jack Pham <jackp@codeaurora.org>
It is virtual clock frontend driver which communicates
with host backend clock service/driver over HABMM. It
only supports msm8996 virtual platform currently.
Change-Id: Icfee28c301fa4a583b45e5d364432535643eb9b7
Signed-off-by: Zhiqiang Tu <ztu@codeaurora.org>
Having this flag was causing PD TX messages to incorrectly
get blocked in the case where the policy engine state machine
worker got scheduled to run in the RX IRQ, but before the
handler exits and clears the flag. This caused false errors
that an RX message is pending. Remove usage of this flag as
it was doing more harm than good.
Change-Id: Iffccfc9ffafac6017192d86d8d97d5534890c7b5
Signed-off-by: Jack Pham <jackp@codeaurora.org>
When sending/receiving a hard reset in source mode, the
PE_SRC_TRANSITION_TO_DEFAULT state turns off VBUS/VCONN and
sleeps for 750ms. During this fairly long period any Type-C
event such as cable disconnect or reconnect won't get
processed as the workqueue is blocked by this wait condition.
This can lead to incorrect state handling, for example when
a sink is removed and a source is attached; the state machine
will think a sink is still attached. Fix this by removing
the msleep() and set a timer to restart the state machine
to PE_UNKNOWN which will turn VBUS/VCONN back on and start
again in PE_SRC_STARTUP. This frees up the workqueue to
process any potential incoming events in the meantime.
Change-Id: I282551a58252a672f319d1559eac9f4ad4c3ca8a
Signed-off-by: Jack Pham <jackp@codeaurora.org>
Entering SNK_Ready or SRC_Ready state from usbpd_set_state()
is somewhat terminal as it is the end state of a contract
negotiation. If there was a previous pending VDM message
waiting to be sent, for example the start of a Discovery which
was queued prior to a PR Swap, it would not get sent. Check for
this and enqueue usbpd_sm once again so that it could be sent
out. Otherwise, start the Discovery process in case we are
in DFP mode, and do that for SNK_Ready as well.
Change-Id: I4516ccb0deec8cdabdb62d539d7d2b66dc23f1ee
Signed-off-by: Jack Pham <jackp@codeaurora.org>
Add android-verity support so that the system image is verified by
the kernel during the boot time. Also set the x509 key file to be
used for verification.
Change-Id: I17ffc2944ef9af556d9fdc05d62734de384dc27c
Signed-off-by: Binoy Jayan <bjayan@codeaurora.org>
As per the hardware recommendation, update ESR tight and broad
filter coefficients for low temperature (< 10 C) to 3 %.
CRs-Fixed: 2155610
Change-Id: I7e5a77bcd14ad64459f6bd13f0be0551dcbefdba
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
If all cores for a cluster are offline or not brought up during
lmh dcvsh probe, it fails to register cooling device and it leads
to probe failure. If later these cores come online, lmh dcvsh doesn't
re-probe it and eventually that cluster lmh dcvsh is broken.
Register for cpu online notification and register cpu cooling device
in cpu online notification callback context. It makes sure that
both lmh dcvsh probing and cooling device register are proper.
Change-Id: Id5345a578aab920d12bd9d9e2ccf1d5059582de9
Signed-off-by: Manaf Meethalavalappu Pallikunhi <manafm@codeaurora.org>
rh850_2_apq interrupt is connected to gpio_122 on auto cdp boards
CRs-Fixed: 2009305
Change-Id: Id60ec07a7bc4757050ad2a1c39123ec561581082
Signed-off-by: Alex Yakavenka <ayakav@codeaurora.org>
WLAN module supports LRO, so enable CONFIG_INET_LRO for SDM660.
Remove it from defconfig to enable as Kconfig has default set to enable.
Change-Id: I5681c47caa18931ab1a3ea2b09c3a1f9abc38777
CRs-Fixed: 2144533
Signed-off-by: Poddar Siddarth <siddpodd@codeaurora.org>
NTN_CLK_SYNC line should be set low before bringing Neutrino out
of reset for msm8996 agave board and auto CDP board.
Change-Id: I7b213a7dad2dbdedd3fe42a88972dc36cbf2e5b5
Signed-off-by: Zou Shunxiang <shunxian@codeaurora.org>
