Fix ordering of link creation between node->prev and prev->next in
osq_lock(). A case in which the status of optimistic spin queue is
CPU6->CPU2 in which CPU6 has acquired the lock. At this point if CPU0
comes in to acquire osq_lock, it will update the tail count. After tail
count update if CPU2 starts to unqueue itself from optimistic spin queue,
it will find updated tail count with CPU0 and update CPU2 node->next to
NULL in osq_wait_next(). If reordering of following stores happen then
prev->next where prev being CPU2 would be updated to point to CPU0 node:
node->prev = prev;
WRITE_ONCE(prev->next, node);
At this point if next instruction
WRITE_ONCE(next->prev, prev);
in CPU2 path is committed before the update of CPU0 node->prev = prev then
CPU0 node->prev will point to CPU6 node. At this point if CPU0 path's
node->prev = prev is committed resulting in change of CPU0 prev back to
CPU2 node. CPU2 node->next is NULL currently, so if CPU0 gets into unqueue
path of osq_lock it will keep spinning in infinite loop as condition
prev->next == node will never be true.
Change-Id: I48d847096daf3c228de90ae1cd2a6415b7bde65a
Signed-off-by: Prateek Sood <prsood@codeaurora.org>
The legacy detection workaround change restricts DCP to 1.5A draw,
regardless of legacy/non-legacy type. This was to prevent a legacy
(non-compliant) cable, incorrectly detected as non-legacy cable from
drawing more than 1.5A.
The issue with legacy bit is that hardware could end up reporting
non-legacy for a legacy cable (not the other way round). We want to
ensure that for non-compliant legacy cable SDP/CDP current limits
are honoured and that for a Rp = 10k legacy cable VBUS shouldn't go
higher than 5V.
This cap of 1.5A on DCP is not necessary. Moreover AICL will limit the
current if needed.
To realize this force the Rp based current whenever DCP or float
is seen via the LEGACY_UNKNOWN_VOTER. LEGACY_UNKNOWN_VOTER is not
removed unless a confirmed hvdcp or pd is seen.
Change-Id: I89505e9db4f045aaf71ab0ee534de783ea4d2df3
Signed-off-by: Ashay Jaiswal <ashayj@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Excessive logging due to several successive memory allocation failure
may cause a watchdog bite. Hence, this change adds ratelimit to logging
on memory allocation failure.
Change-Id: I8e5d78918a32c48ef7fa587f3dc63cbd1f065d5f
Signed-off-by: Deepak Kumar <dkumar@codeaurora.org>
prepare_to_wait() will enqueue the thread on the given queue
and put it into the given execution state,
which is TASK_INTERRUPTIBLE.
Further processing in function, calls mutex_lock(),
will go into a new version of the going-to-sleep code,
changing the task state.
That, of course, may well interfere with the outer
sleeping code.
So, nesting of sleeping primitives in this way is discouraged.
And new warning was added to point out this kind of nesting.
Fix the nesting of sleeping primitives with the new solution
provide in linux kernel.
Change-Id: Id1a5f64472cd2d63e679706c6482db98f89ec765
Signed-off-by: Mohammed Javid <mjavid@codeaurora.org>
Resolve memory out of bound access by correcting the
length of the buffer to be copied.
Change-Id: I2cc74a664399913acf67464a5f6827b100522676
Signed-off-by: Revathi Uddaraju <revathiu@codeaurora.org>
Currently, there is possibility of out-of-bound accesses during
handling of data in non-hdlc path. The patch adds proper protection
when processing non-hdlc packet information to fix the issue.
CRs-Fixed: 2029216
Change-Id: I07c466f85bd8ac08226948fea86b1d8567e68431
Signed-off-by: Hardik Arya <harya@codeaurora.org>
The pointer req_frm is coming from userspace, it may overflow stream_info.
Adding a bound check to prevent the same.
CRs-fixed: 2008683
Change-Id: I8682e09ff2ab7ba490bbbd9e20db978493c5f3e4
Signed-off-by: Senthil Kumar Rajagopal <skrajago@codeaurora.org>
Signed-off-by: Andy Sun <bins@codeaurora.org>
Live Lock due to task spinning while unqueue of CPU osq_node
from optimistic_spin_queue. Task T1 had decremented mutex count to
acquire the lock on CPU0. Before setting owner it got preempted. On
CPU1 task T2 acquired osq_lock and started spinning on owner of mutex
with preemption disabled. CPU1 runq has one task, so need_resched will
not be set. On CPU0 task T3 tried to acquire osq_lock to spin on the
same mutex. At this time following scenario causes soft lockup:
After preemption of task T1, RT task T3 tried to acquire the same
mutex. It will start spinning on the osq_lock until the lock is available
or need_resched is set. For RT task, need_resched will not be set. Task T3
will not be able to bail out of the infinite loop.
Change-Id: Ifd7506047119a22e14b15459ac6b04b410ba1c84
Signed-off-by: Prateek Sood <prsood@codeaurora.org>
Current upstream parser only handles RGB deep color
modes.
Add support in the SDE EDID parser module to parse
HDMI VSDB block and indicate support for YUV 420
deep color modes in the sink.
Change-Id: If6c007263094e7716a29cae503d3e3471ae04306
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
Add default value support for plane enum property which caller is
able to pass in different default value when initialize the enum
property list.
Change-Id: I57595bf7c42b0e528a18ab0951655a169b00d611
Signed-off-by: Jin Li <jinl@codeaurora.org>
Signed-off-by: Yunyun Cao <yunyunc@codeaurora.org>
Conntrack, if enabled, verifies complete checksum on all tcp
packets. If a packet is corrupted, it sets ip_summed field
to CHECKSUM_COMPLETE and checksum valid field to false.
With these changes such packet will be dropped.
Packets that are corrupted can go into userspace even when
the checksum is wrong. The io_vec library pushes the data into
pipe before checking the checksum. If the checksum is wrong,
the copied data is not reverted. Users observe corrupted data
in the application memory due to such corrupted packets.
This fix is to plug one such hole.
This change might be redundant on kernels after v4.9.
Change-Id: Iffface598d0fa2b25fb9c20c7aa6443aab9d8aea
Signed-off-by: Tejaswi Tanikella <tejaswit@codeaurora.org>
Neighbour index of step table possibly have same position,
so i2c operation reported invalid size parameters.
we add protection condition to return success value.
Change-Id: I7dab8f44a99c7c3c7d6996c8decb8bcd09c246c9
Signed-off-by: penliu <pengfeiliu@codeaurora.org>
Sysfs attribute "inactivity_timer" not being used anymore so remove that
entry from sysfs.
CRs-Fixed: 2048725
Change-Id: I30d2d4a3cb9899b7975a6395eab90337fc006488
Signed-off-by: Arumuga Durai A <cadurai@codeaurora.org>
Userspace can set the default IRQ affinity setting by writing into
/proc/irq/default_smp_affinity file. When an IRQ affinity is
broken during isolation/hotplug,override the affinity to online and
un-isolated CPUs from the default affinity CPUs. If no such CPU
is available, then only override with cpu_online_mask.
Change-Id: I7578728ed0d7c17c5890d9916cfd6451d1968568
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
With commit bfc60d474137 ("genirq: Use irq_set_affinity_locked to change
irq affinity"), affinity listeners receive the notification when the irq
affinity is changed during migration. If there is no online and
un-isolated CPU available from the user specified affinity, the affinity
is overridden with all online and un-isolated CPUs. The same cpumask is
notified to PM QOS affinity listener which applies PM_QOS_CPU_DMA_LATENCY
vote to all those CPUs. As the low level irqchip driver sets affinity to
only one CPU, do the same while overriding the affinity during migration.
Change-Id: I0bcb75dd356658da100fbeeefd33ef8b121f4d6d
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
Add support of cnss_utils for WLAN. Export APIs to WLAN
host driver which are used to set/get WLAN related
information. These information could be retrived later
when WLAN driver is reloaded.
CRs-Fixed: 2060693
Change-Id: I2c8c36602ec8af133946ff00c41ce648a2628041
Signed-off-by: Yuanyuan Liu <yuanliu@codeaurora.org>
When write masters are reloaded pingpong status regi-
-ster will not be reset. Instead, it would be holding
a stale data, until new axi_done irq is interrupted.
So, place a check to validate the pingpong register
value based on the reloaded status of write masters.
Change-Id: Id14b886154f8a8ef8d5c05338023d8172d6925d0
Signed-off-by: Lokesh Kumar Aakulu <lkumar@codeaurora.org>
debug offset comes from the user and can hold any value which can
cause unaligned access. This change fixes the unaligned access
problem on debug offset by properly aligning it.
Change-Id: Ie4de9a12433f6ffd568c6c86928b71a5537b0dff
Signed-off-by: Harsh Sahu <hsahu@codeaurora.org>
The PM_QOS_CPU_DMA_LATENCY vote attached to an IRQ is discarded,
if it is affined to an isolated CPU. So we need atleast 1 CPU
in online and un-isolate state. The scheduler rejects isolating
a CPU if it is the only online and un-isolated CPU in the system.
Add the same check for CPU hotplug.
Change-Id: I5bdfe6e3bb0352ed3ae5a2de90097b73d248f3fc
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
IPA needs to be kept disabled in APQ 8098 platform.
Change-Id: If340d72d1dc18f2861635528ea1a37a4d81421bd
Signed-off-by: Mohammed Javid <mjavid@codeaurora.org>
when set_cs() is called by framework, first check the device suspend
status and if not suspended then proceed for register access. Let system
suspend call SPI runtime suspend but make sure set_cs doesn't race with
it by holding the mutex.
Change-Id: Id6d05ac37227bf3358438eeedf279ddd1c9aab8b
Signed-off-by: Mukesh Kumar Savaliya <msavaliy@codeaurora.org>
Format specifier %p can leak kernel addresses while not valuing the
kptr_restrict system settings. When kptr_restrict is set to (1), kernel
pointers printed using the %pK format specifier will be replaced with 0's
Change-Id: Iff8d82b12e958b938fc767bf3e8c3a3c8fc65c2a
Signed-off-by: mohamed sunfeer <msunfeer@codeaurora.org>
ACS fails as channel details are null in survey info.
Add the missing fields in survey_info structure and
populate the channel details in survey results.
CRs-Fixed: 2027495
Change-Id: I1ca933ded4ca039845203ccd51d77cf7cef1c000
Signed-off-by: Ashutosh Kumar <askuma@codeaurora.org>
In the current implementation, if panel is configured in split
mode and supports multiple resolutions, then the pixel clock
calculation for the supported modes does not take split mode
into consideration. This causes issues when recovery or charger
application try to configure display. So fix the same by
recalculating the pixel clock with taking care of proper width
in case of split mode panel.
Change-Id: Ie6b50bcd67d3e283610f8b04ac0a974b3527e552
Signed-off-by: Sandeep Panda <spanda@codeaurora.org>
Defines the non-removable property for UFS device node.
This basically lets the driver know that UFS is the
boot device.
Change-Id: I1c8aa298af4afc7e42d0c8a3690f7d3360e37647
Signed-off-by: Sayali Lokhande <sayalil@codeaurora.org>