As the best clocksource is not selected till core boot completion,
only periodic tick timer works and it increases jiffies by one at
every tick updates. If interrupt is disabled more than one tick(10ms),
timer interrupts are missed and jiffies can't be updated at every
10ms and it can be behind the real time. So make it possible to select
the best clocksource right after arm arch timer initialization, so that
jiffies can be increased by multiple counts since then.
Change-Id: Id8c4e3ce9b9e44061fef7ad7e678ca1c27d84bb1
Signed-off-by: Se Wang (Patrick) Oh <sewango@codeaurora.org>
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
As the best clocksource is not selected till core boot completion,
only periodic tick timer works and it increases jiffies by one at
every tick updates. If interrupt is disabled more than one tick(10ms),
timer interrupts are missed and jiffies can't be updated at every
10ms and it can be behind the real time. So add API to force re-
selection of the best clocksource among registered clocksources so
that the best clocksource can be selected whenever it is available.
Change-Id: I481de3cdf1df8f0e35ed10aee7ab3882bf7a35b3
Signed-off-by: Se Wang (Patrick) Oh <sewango@codeaurora.org>
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
Use mmc_try_claim_host with a timeout instead of mmc_claim_host in
mmc_sd_detect. This is to ensure that mmc rescan work item is doesn't
get blocked on claim_host for longer period.
In the pm_suspend path, we cancel the mmc_rescan work item.
If this work item is already scheduled, suspend would be blocked till
mmc_rescan gets finished. In case, mmc_rescan is blocked on claim_host
lock, pm_suspend could get blocked for longer period. This can result
in momentary UI freeze since pm_suspend is blocked for longer duration.
This change is to prevent this scenario.
Change-Id: Ib93bae6745a153bad3579ae42f46c3c3a7c1b95a
Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Add support for completing failed cmdq requests with error.
If any cmdq request is failed, we enter into error recovery path and
re-queue the same request. But in some cases (e.g. requests failed due
to invalid crypto configuration) same request shouldn't be re-queued
but should complete the request with error so that the a new requested
would be issued.
In mmc_blk_cmdq_complete_rq(), for a request if error information is
present and flag which indicates to skip error handling is set
then complete that request with error info.
Change-Id: I9c4a446bb27b4d82a0847d0bfb4481b314df479c
Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Currently driver is calling composite_setup_complete() when request
queuing to control endpoint fails. During disconnect or composition switch,
ep_queue() fails with -ESHUTDOWN return value. In this case also, driver is
calling composite_setup_complete(), which leads to invalid pointer
dereference. Fix it by not calling composite_setup_complete() in case of
return value of -ESHUTDOWN as anyhow composite_unbind() will take care of
clearing pending flags before freeing request buffers.
Change-Id: I87ea6ecb1e925c6b36dede59486e49ba3a4e90c7
Signed-off-by: Vijayavardhan Vennapusa <vvreddy@codeaurora.org>
In msm_copy_camera_private_ioctl_args function in
msm_buf_mgr camera driver arg is pointing to an address
in userspace and not kernel.
Done changes to use copy_from_user to fix it.
Change-Id: Ia9b747dcf86b448656a5d3676455ccb4eccd4e5a
Signed-off-by: Manish Poddar <mpoddar@codeaurora.org>
Calibration data configuration is done from user space
for WCN3990 target. Skip cal configuration from debug
routine.
Change-Id: I349575a04029e80ef4efc2ac486f5716561c37d8
Signed-off-by: Govind Singh <govinds@codeaurora.org>
With ICE HCI addition in SDHC, couple of errors conditions
got introduced. These errors would be detected by ICE HCI and
triggers interrupt.
The two errors are:
1. Generic Crypto Error (GCE): This error is detected during the
transfer. Need to reset SDHC to recover.
2. Invalid Crypto Configuration Error (ICCE): The HW detects it
before the beginning of the transfer and terminates the transfer
before actual execution started.
Driver has to handle these error for recovering ICE-HCI from
these error conditions. This patch is for handling there errors
in cmdq mode.
Change-Id: Ia10075aca429cbc49ccccdae5c0499c7a176b8a2
Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Enable CONFIG_VT in msmcortex mediabox defonfig. This is needed for
Weston compositor functionality.
CRs-Fixed: 2002381
Change-Id: Id7d57d4b1862252af40fe1d8b296a4e5a8ee0fba
Signed-off-by: Manoj Kumar AVM <manojavm@codeaurora.org>
Improve touch will be used on sdm630 QRD so add support for it.
Change-Id: I4e7868f81dbd86a425375e98f4fea4a40f53887f
Signed-off-by: zhaoyuan <yzhao@codeaurora.org>
If there is an OVP event happened in the bootloader and the
interrupt is not handled before, then as soon as the WLED
module is enabled, fault interrupt can fire. This is seen
sometimes during bootup. This is harmless as such and cannot
be stopped completely. However, it would be good to keep the
OVP fault interrupt disabled at the startup. When WLED module
is enabled again, it will get enabled anyways after ~10 ms.
While at it, print the error message in OVP fault interrupt
handler only when the fault status is set.
Change-Id: I0759c87137a52a8872b37d126e71c13f76d10e05
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Currently, WLED over-voltage and short-circuit fault interrupts
are configured for both rising and falling edge. Configure them
to be on rising edge alone as that should be sufficient.
Change-Id: I9e19cf08e77432a4db5f18894ee445d38f1ad802
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Currently, WLED over-voltage and short-circuit fault interrupts
are configured for both rising and falling edge. Configure them
to be on rising edge alone as that should be sufficient.
Change-Id: I7cdaf126c07daa9afb2c0db68ee760a9a4a2956a
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
snd_card_free is a blocking function which waits for ALSA device to
be closed before freeing it. This could be blocked forever if userspace
has device node opened when midi function is unbound.
Replace it with snd_card_free_when_closed which can be safely used if
userspce has the device node opened.
It is not a blocking function and defers freeing of the card until
userspace closes the ALSA device.
Change-Id: I6becc44dc0f0c1131f9b7dcd1e4a151d830dde2b
CRs-fixed: 1116790
Signed-off-by: Manu Gautam <mgautam@codeaurora.org>
Add support for cable connect/disconnect acknowledgment
between HDMI/DisplayPort audio codec driver and display
driver to replace sysfs node mechanism.
CRs-Fixed: 2013097
Change-Id: I3b385a5e9933f7e99f1dbb9848a2664f687bd7d7
Signed-off-by: Kuirong Wang <kuirongw@codeaurora.org>
For the 16-bit FG peripheral register address check the lower-byte
for secure_access instead of the upper-byte.
CRs-Fixed: 2013922
Change-Id: I647aa4eadccb007e6e974eae65cef162c5a40a0d
Signed-off-by: Anirudh Ghayal <aghayal@codeaurora.org>