Add several defines and structs to allow IPA kernel tests running.
Change-Id: I9c8f114b28aa617bdf1bfe56b44b73852a178ece
Acked-by: Dmitry Kogan <dmitryk@qti.qualcomm.com>
Signed-off-by: Amir Levy <alevy@codeaurora.org>
Configure the sleep state for the wlan bootstrap pin for
msm8996. This configuration helps in power saving during
device sleep state.
CRs-Fixed: 1040537
Change-Id: Ia35f8bba901826b55722cb140c953f32b18610de
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
The QCA wlan chipset uses tlmm, msm and pmic gpio for the WLAN_EN
signal pin. The tlmm or msm gpio uses pinctrl or gpio library for
configuring the sleep/active state runtime.
The pmic gpio runtime configuration is not feasible using gpio
library or pinctrl framework. Convert the pmic gpio to fixed
regulator to runtime control(enable/disable) via regulator framework.
CRs-Fixed: 1040537
Change-Id: I4963c2753b463d5852a847a20fe280107abca337
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
The QCA wlan chipset uses tlmm, msm and pmic gpio for the WLAN_EN
signal pin. The tlmm or msm gpio uses pinctrl or gpio library for
configuring the sleep/active state runtime.
The pmic gpio runtime configuration is not feasible using gpio
library or pinctrl framework. Convert the pmic gpio to fixed
regulator to runtime control(enable/disable) via regulator framework.
CRs-Fixed: 1040537
Change-Id: Ie74a659f309f248d335e03ca7a0a00244e9715b4
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
Currently when ptrain_done interrupt is handled, the driver invariably
issues a ESR measurement. This is not required if qnovo is disabled.
Change-Id: I8c81dadc38da69ba7248905af2d772f7ce918239
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
In some cases, a skew needs to be applied on the learnt capacity
to counter the error. Add support for it.
Change-Id: I7f80088e7db4e3414d86983722c92e4fc14323e0
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Currently, when the charging status transitions to a state other
than charging, cycle counter algorithm stores the cycle count for
all the buckets for which the counting had started. This is fine
with respect to the algorithm.
However with qnovo enabled charging, this can cause issues when
charging status can go to not charging intermittently causing the
cycle count to be stored multiple times for a bucket. Modify the
logic to check for the charge termination or the presence of
input to go through all the buckets for storing the count. Also,
increment and store the counter only if battery SOC had increased
more than 2 LSBs for that SOC bucket.
While at it, run cycle_counter algorithm as a function instead
of a work. Also, keep the usage of cycle counter feature enable
flag inside the cycle counter APIs.
Change-Id: I62a92964ccbc6b965af09696deddc6fa8366a841
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Currently if there is a spike in system load or a thermal event which
causes the battery current to change dramatically then the TTF can jump.
While a battery is charging the TTF should be monotonically decreasing.
Track the TTF starting with the first estimate and set hard bounds of -2
and -0.1 on the slope. The negative slope ensures the TTF is
monotonically decreasing and the hard bounds on the slope smooth out
significant changes in the TTF.
Change-Id: I68a934599ff25bc5a9eb67b372b28a723532a540
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
QNOVO provides a table of expected charge current settings across the
charge cycle. Use this table to calculate the time it will take to
charge in each step. The sum of these steps is the total time it will
take to fully charge a battery.
Change-Id: I0ed48a2a63886531e761e7ce1d175a600060eaf8
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Micro resolution and 64-bit division is unnecessary; use milli
resolution instead. Measuring the IBATT and VBATT periodically during
sleep is necessary for accuracy. Don't clear the IBATT and VBATT buffers
on suspend to account for the higher charge current during suspend.
Prime the IBATT and VBATT buffers with 10 samples to get a more accurate
first estimate. Introduce a ttf mode to separate the differences in the
QNOVO version of TTF.
Change-Id: Ibc591dd5d38d4bbb712d8906755040d59181f008
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Update tx power of lower 5GHz band of US country according
to master sheet.
CRs-Fixed: 2084867
Change-Id: I18c500d5d125ef73b47c6cda120541df98881537
Signed-off-by: Rajeev Kumar Sirasanagandla <rsirasan@codeaurora.org>
* refs/heads/tmp-59ff2e1
Linux 4.4.78
kvm: vmx: allow host to access guest MSR_IA32_BNDCFGS
kvm: vmx: Check value written to IA32_BNDCFGS
kvm: x86: Guest BNDCFGS requires guest MPX support
kvm: vmx: Do not disable intercepts for BNDCFGS
KVM: x86: disable MPX if host did not enable MPX XSAVE features
tracing: Use SOFTIRQ_OFFSET for softirq dectection for more accurate results
PM / QoS: return -EINVAL for bogus strings
PM / wakeirq: Convert to SRCU
sched/topology: Optimize build_group_mask()
sched/topology: Fix overlapping sched_group_mask
crypto: caam - fix signals handling
crypto: sha1-ssse3 - Disable avx2
crypto: atmel - only treat EBUSY as transient if backlog
crypto: talitos - Extend max key length for SHA384/512-HMAC and AEAD
mm: fix overflow check in expand_upwards()
tpm: Issue a TPM2_Shutdown for TPM2 devices.
Add "shutdown" to "struct class".
tpm: Provide strong locking for device removal
tpm: Get rid of chip->pdev
selftests/capabilities: Fix the test_execve test
mnt: Make propagate_umount less slow for overlapping mount propagation trees
mnt: In propgate_umount handle visiting mounts in any order
mnt: In umount propagation reparent in a separate pass
vt: fix unchecked __put_user() in tioclinux ioctls
exec: Limit arg stack to at most 75% of _STK_LIM
s390: reduce ELF_ET_DYN_BASE
powerpc: move ELF_ET_DYN_BASE to 4GB / 4MB
arm64: move ELF_ET_DYN_BASE to 4GB / 4MB
arm: move ELF_ET_DYN_BASE to 4MB
binfmt_elf: use ELF_ET_DYN_BASE only for PIE
checkpatch: silence perl 5.26.0 unescaped left brace warnings
fs/dcache.c: fix spin lockup issue on nlru->lock
mm/list_lru.c: fix list_lru_count_node() to be race free
kernel/extable.c: mark core_kernel_text notrace
tools/lib/lockdep: Reduce MAX_LOCK_DEPTH to avoid overflowing lock_chain/: Depth
parisc/mm: Ensure IRQs are off in switch_mm()
parisc: DMA API: return error instead of BUG_ON for dma ops on non dma devs
parisc: use compat_sys_keyctl()
parisc: Report SIGSEGV instead of SIGBUS when running out of stack
irqchip/gic-v3: Fix out-of-bound access in gic_set_affinity
cfg80211: Check if PMKID attribute is of expected size
cfg80211: Validate frequencies nested in NL80211_ATTR_SCAN_FREQUENCIES
cfg80211: Define nla_policy for NL80211_ATTR_LOCAL_MESH_POWER_MODE
brcmfmac: fix possible buffer overflow in brcmf_cfg80211_mgmt_tx()
rds: tcp: use sock_create_lite() to create the accept socket
vrf: fix bug_on triggered by rx when destroying a vrf
net: ipv6: Compare lwstate in detecting duplicate nexthops
ipv6: dad: don't remove dynamic addresses if link is down
net: handle NAPI_GRO_FREE_STOLEN_HEAD case also in napi_frags_finish()
bpf: prevent leaking pointer via xadd on unpriviledged
net: prevent sign extension in dev_get_stats()
tcp: reset sk_rx_dst in tcp_disconnect()
net: dp83640: Avoid NULL pointer dereference.
ipv6: avoid unregistering inet6_dev for loopback
net/phy: micrel: configure intterupts after autoneg workaround
net: sched: Fix one possible panic when no destroy callback
net_sched: fix error recovery at qdisc creation
ANDROID: android-verity: mark dev as rw for linear target
ANDROID: sdcardfs: Remove unnecessary lock
ANDROID: binder: don't check prio permissions on restore.
Add BINDER_GET_NODE_DEBUG_INFO ioctl
UPSTREAM: cpufreq: schedutil: Trace frequency only if it has changed
UPSTREAM: cpufreq: schedutil: Avoid reducing frequency of busy CPUs prematurely
UPSTREAM: cpufreq: schedutil: Refactor sugov_next_freq_shared()
UPSTREAM: cpufreq: schedutil: Fix per-CPU structure initialization in sugov_start()
UPSTREAM: cpufreq: schedutil: Pass sg_policy to get_next_freq()
UPSTREAM: cpufreq: schedutil: move cached_raw_freq to struct sugov_policy
UPSTREAM: cpufreq: schedutil: Rectify comment in sugov_irq_work() function
UPSTREAM: cpufreq: schedutil: irq-work and mutex are only used in slow path
UPSTREAM: cpufreq: schedutil: enable fast switch earlier
UPSTREAM: cpufreq: schedutil: Avoid indented labels
Linux 4.4.77
saa7134: fix warm Medion 7134 EEPROM read
x86/mm/pat: Don't report PAT on CPUs that don't support it
ext4: check return value of kstrtoull correctly in reserved_clusters_store
staging: comedi: fix clean-up of comedi_class in comedi_init()
staging: vt6556: vnt_start Fix missing call to vnt_key_init_table.
tcp: fix tcp_mark_head_lost to check skb len before fragmenting
md: fix super_offset endianness in super_1_rdev_size_change
md: fix incorrect use of lexx_to_cpu in does_sb_need_changing
perf tools: Use readdir() instead of deprecated readdir_r() again
perf tests: Remove wrong semicolon in while loop in CQM test
perf trace: Do not process PERF_RECORD_LOST twice
perf dwarf: Guard !x86_64 definitions under #ifdef else clause
perf pmu: Fix misleadingly indented assignment (whitespace)
perf annotate browser: Fix behaviour of Shift-Tab with nothing focussed
perf tools: Remove duplicate const qualifier
perf script: Use readdir() instead of deprecated readdir_r()
perf thread_map: Use readdir() instead of deprecated readdir_r()
perf tools: Use readdir() instead of deprecated readdir_r()
perf bench numa: Avoid possible truncation when using snprintf()
perf tests: Avoid possible truncation with dirent->d_name + snprintf
perf scripting perl: Fix compile error with some perl5 versions
perf thread_map: Correctly size buffer used with dirent->dt_name
perf intel-pt: Use __fallthrough
perf top: Use __fallthrough
tools strfilter: Use __fallthrough
tools string: Use __fallthrough in perf_atoll()
tools include: Add a __fallthrough statement
mqueue: fix a use-after-free in sys_mq_notify()
RDMA/uverbs: Check port number supplied by user verbs cmds
KEYS: Fix an error code in request_master_key()
ath10k: override CE5 config for QCA9377
x86/uaccess: Optimize copy_user_enhanced_fast_string() for short strings
x86/tools: Fix gcc-7 warning in relocs.c
gfs2: Fix glock rhashtable rcu bug
USB: serial: qcserial: new Sierra Wireless EM7305 device ID
USB: serial: option: add two Longcheer device ids
pinctrl: sh-pfc: Update info pointer after SoC-specific init
pinctrl: mxs: atomically switch mux and drive strength config
pinctrl: sunxi: Fix SPDIF function name for A83T
pinctrl: meson: meson8b: fix the NAND DQS pins
pinctrl: sh-pfc: r8a7791: Fix SCIF2 pinmux data
sysctl: report EINVAL if value is larger than UINT_MAX for proc_douintvec
sysctl: don't print negative flag for proc_douintvec
mac80211_hwsim: Replace bogus hrtimer clockid
usb: Fix typo in the definition of Endpoint[out]Request
usb: usbip: set buffer pointers to NULL after free
Add USB quirk for HVR-950q to avoid intermittent device resets
USB: serial: cp210x: add ID for CEL EM3588 USB ZigBee stick
usb: dwc3: replace %p with %pK
drm/virtio: don't leak bo on drm_gem_object_init failure
tracing/kprobes: Allow to create probe with a module name starting with a digit
mm: fix classzone_idx underflow in shrink_zones()
bgmac: reset & enable Ethernet core before using it
driver core: platform: fix race condition with driver_override
fs: completely ignore unknown open flags
fs: add a VALID_OPEN_FLAGS
ANDROID: binder: add RT inheritance flag to node.
ANDROID: binder: improve priority inheritance.
ANDROID: binder: add min sched_policy to node.
ANDROID: binder: add support for RT prio inheritance.
ANDROID: binder: push new transactions to waiting threads.
ANDROID: binder: remove proc waitqueue
FROMLIST: binder: remove global binder lock
FROMLIST: binder: fix death race conditions
FROMLIST: binder: protect against stale pointers in print_binder_transaction
FROMLIST: binder: protect binder_ref with outer lock
FROMLIST: binder: use inner lock to protect thread accounting
FROMLIST: binder: protect transaction_stack with inner lock.
FROMLIST: binder: protect proc->threads with inner_lock
FROMLIST: binder: protect proc->nodes with inner lock
FROMLIST: binder: add spinlock to protect binder_node
FROMLIST: binder: add spinlocks to protect todo lists
FROMLIST: binder: use inner lock to sync work dq and node counts
FROMLIST: binder: introduce locking helper functions
FROMLIST: binder: use node->tmp_refs to ensure node safety
FROMLIST: binder: refactor binder ref inc/dec for thread safety
FROMLIST: binder: make sure accesses to proc/thread are safe
FROMLIST: binder: make sure target_node has strong ref
FROMLIST: binder: guarantee txn complete / errors delivered in-order
FROMLIST: binder: refactor binder_pop_transaction
FROMLIST: binder: use atomic for transaction_log index
FROMLIST: binder: add more debug info when allocation fails.
FROMLIST: binder: protect against two threads freeing buffer
FROMLIST: binder: remove dead code in binder_get_ref_for_node
FROMLIST: binder: don't modify thread->looper from other threads
FROMLIST: binder: avoid race conditions when enqueuing txn
FROMLIST: binder: refactor queue management in binder_thread_read
FROMLIST: binder: add log information for binder transaction failures
FROMLIST: binder: make binder_last_id an atomic
FROMLIST: binder: change binder_stats to atomics
FROMLIST: binder: add protection for non-perf cases
FROMLIST: binder: remove binder_debug_no_lock mechanism
FROMLIST: binder: move binder_alloc to separate file
FROMLIST: binder: separate out binder_alloc functions
FROMLIST: binder: remove unneeded cleanup code
FROMLIST: binder: separate binder allocator structure from binder proc
FROMLIST: binder: Use wake up hint for synchronous transactions.
Revert "android: binder: move global binder state into context struct."
sched: walt: fix window misalignment when HZ=300
ANDROID: android-base.cfg: remove CONFIG_CGROUP_DEBUG
ANDROID: sdcardfs: use mount_nodev and fix a issue in sdcardfs_kill_sb
Conflicts:
drivers/android/binder.c
drivers/net/wireless/ath/ath10k/pci.c
Change-Id: Ic6f82c2ec9929733a16a03bb3b745187e002f4f6
Signed-off-by: Blagovest Kolenichev <bkolenichev@codeaurora.org>
This is a preparation change for merging binder from
android-4.4@59ff2e1 (v4.4.78).
Code matching changes below will be re-added properly with
upcoming change:
0cebb40 FROMLIST: binder: Use wake up hint for synchronous transactions.
Reverted changes:
=================
e3ba00c android: binder: Don't use sched_preempt_enable_no_resched.
The correct function is prempt_enable_no_resched(). The other
function is reserved for the scheduler core.
Change-Id: Ib36697de003f6a59a608a0024d5351dc15ff8715
Signed-off-by: Todd Kjos <tkjos@google.com>
Git-commit: 776e5bca6446b3aac03b4685b4f4f72446ddcba0
Git-repo: https://android.googlesource.com/kernel/msm
[odhyade@codeaurora.org: resolve trivial merge conflicts]
Signed-off-by: Omprakash Dhyade <odhyade@codeaurora.org>
3956eab android: binder: Use wake up hint for synchronous transactions.
Use wake_up_interruptible_sync() to hint to the scheduler binder
transactions are synchronous wakeups. Disable premption while waking
to avoid ping-ponging on the binder lock.
Change-Id: Ic406a232d0873662f80148e37acefe5243d912a0
Signed-off-by: Todd Kjos <tkjos@google.com>
Git-commit: 443c026e90820170aa3db2c21d2933ae5922f900
Git-repo: https://android.googlesource.com/kernel/msm
Signed-off-by: Omprakash Dhyade <odhyade@codeaurora.org>
Change-Id: I51a11f4325ce22aaa9666efedec1ce51d1d7036a
Signed-off-by: Blagovest Kolenichev <bkolenichev@codeaurora.org>
These were hacks and no longer needed as next binder from
android-4.4@59ff2e1 (v4.4.78) has been reworked to use
fine-grained locking.
This is a preparation for clean merge of binder changes from
android-4.4@59ff2e1 (v4.4.78), which follows right after this
commit.
Reverted changes:
=================
773fc2f android: binder: use copy_from_user_preempt_disabled
To keep the driver consistent, and until we have
fine-grained locking in place.
Change-Id: I90fe02cdedb8a5677b900a68528fb443b9204322
Signed-off-by: Riley Andrews <riandrews@google.com>
Git-repo: https://source.codeaurora.org/quic/la/kernel/msm-4.4
Git-commit: 5c9ce54ca3a66a57e4ebfe3ae71c5733b6bcc579
Signed-off-by: Vikram Mulukutla <markivx@codeaurora.org>
821e02f CHROMIUM: android: binder: Fix potential scheduling-while-atomica
Commit f1e7f0a724f6 ("android: binder: Disable preemption while holding
the global binder lock.") re-enabled preemption around most of the sites
where calls to potentially sleeping functions were made, but missed
__alloc_fd(), which can sleep if the fdtable needs to be resized.
Re-enable preemption around __alloc_fd() as well as __fd_install() which
can now sleep in upstream kernels as of commit 8a81252b77 ("fs/file.c:
don't acquire files->file_lock in fd_install()").
BUG=chrome-os-partner:44012
TEST=Build and boot on Smaug.
Change-Id: I9819c4b95876f697e75b1b84810b6c520d9c33ec
Signed-off-by: Andrew Bresticker <abrestic@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/308582
Reviewed-by: Stephen Barber <smbarber@chromium.org>
Reviewed-by: Riley Andrews <riandrews@google.com>
Git-repo: https://source.codeaurora.org/quic/la/kernel/msm-4.4
Git-commit: c267ff1d548ed1bdad6a08f1c70776c5e60d569e
Signed-off-by: Vikram Mulukutla <markivx@codeaurora.org>
e4045d6 android: binder: Disable preemption while holding the global binder lock.
Change-Id: I90fe02cdedb8a5677b900a68528fb443b9204322
Signed-off-by: Riley Andrews <riandrews@google.com>
Git-repo: https://source.codeaurora.org/quic/la/kernel/msm-4.4
Git-commit: 5c9ce54ca3a66a57e4ebfe3ae71c5733b6bcc579
Signed-off-by: Vikram Mulukutla <markivx@codeaurora.org>
Change-Id: Id97e04f3694d76c2b6a1e7c0dc39da15d83c2867
Signed-off-by: Blagovest Kolenichev <bkolenichev@codeaurora.org>
QMI client is not disconnected if driver is failed to load
due to mode response timeout.
Disconnect QMI client if driver is failed to load after
starting QMI services.
Change-Id: I36b059fbe3254990ad6daa13533122497805e649
Signed-off-by: Govind Singh <govinds@codeaurora.org>
There is a possibility of integer overflow in the arithmetic
calculation for cmd_size. Fix this by adding checks for such
arithmetic.
Change-Id: I2298a32f8ba3411decb29f55bb7b55e2214de35a
Signed-off-by: Abhilash Kumar <krabhi@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.
Debugging Note : &pK prints only Zeros as address. If you need actual
address information, write 0 to kptr_restrict.
echo 0 > /proc/sys/kernel/kptr_restrict
Change-Id: If3e3a7d5c737b8b7e2c2f78e2814747db4a770cd
Signed-off-by: Liangliang Lu <luliang@codeaurora.org>
Whenever userspace calls read system call for ctrl_dev, ccid driver
waits till it receives control request from host. Once it receives
control data, it unblocks read and passes data to userspace. But if
cable is disconnected, then ccid driver is not unblocking read due
to which next open of ctrl_dev failing. This results in ccid functionality
not working after cable reconnection as userspace not able to open
ctrl_dev. Hence fix the issue by checking for online status and if
online flag is cleared, unblock read and return -ENODEV.
Also align request buffer length to OUT endpoint's max packet size,
when userspace tries to queue buffer length that is not multiple of
OUT endpoint max packet as USB controller expects length multiple of
maxpacket.
Change-Id: I97d1819c21919b6c66b7a1e5fbf1c452ec691639
Signed-off-by: Vijayavardhan Vennapusa <vvreddy@codeaurora.org>
Vote 1.8V IO and XTAL regulators before 3.3V VREG for WLAN hardware
to strictly follow the hardware requirement. Make sure 3.3V VREG is
voted at least 100us after 1.8V IO.
Change-Id: Ifc52c2062349a9913e6c998573b62d111faa5886
CRs-fixed: 1009287
Signed-off-by: Yue Ma <yuem@codeaurora.org>
During preemption microcode does save restore for all perf
counters. If we read the power counters at preemption boundary
we might get abnormal value from the perf counter. This will
result in showing incorrect GPU busy percentage. Fix this by
setting the abnormal power perf counter value with zero.
Change-Id: I96ba367ceeeb92d6adb507d0d917113297b4b58d
Signed-off-by: Abhilash Kumar <krabhi@codeaurora.org>
This enum is used to specify what packet type (SOP, SOP', SOP'')
to the PD PHY when sending and receiving messages. To avoid
confusion with the message type field of a message header (e.g.
MSG_REQUEST, MSG_ACCEPT, etc), rename the enum to pd_sop_type
and change the parameter name from 'type' to 'sop'. Also change
the enum pd_sig_type parameter name to 'sig' for similar clarity.
This patch is cosmetic, no functional changes.
Change-Id: Idf74f7635ec5afee3538815f2efd9bc032ee6d89
Signed-off-by: Jack Pham <jackp@codeaurora.org>
