- Apr 15, 2019
-
-
Cheng Gu authored
Adds MNH CPU related definitions to allow register access. Bug: 118765019 Change-Id: Ibcdf7485a81d3520461fb920442052c889a0e97a Signed-off-by:Cheng Gu <gucheng@google.com>
-
Jorge Troncoso authored
Add cpu_cg sysfs for halting MNH CPU. Bug: 110053899 Change-Id: I16cba1ba487628a40749a603d4ee986d28def3a1 Signed-off-by: -
Cheng Gu authored
This patch also prints BOOT_STAT on boot/resume failure, in addition to BOOT_TRACE. This helps collect data for stability issues. BOOT_STAT shows boot sequence in lower level firmware. BOOT_TRACE shows boot sequence in Easel kernel. Bug: 111652429 Bug: 118765019 Change-Id: I4d96b82e7ca9db1fb23560e97415e3ff064d03b9 Signed-off-by: -
Cheng Gu authored
On an emergency shutdown event, link may be in a bad state and may cause kernel BUG on our platform: Unhandled fault: synchronous external abort In such case, update PCIe link status before disabling PCIe irq, in order to avoid any read/write to PCIe configuration registers. In the normal case, the original sequence is not affected. Bug: 76434914 Change-Id: I0ef078c747660cbb31e57142cac12bfb6a2b3162 Signed-off-by: -
mandyshen authored
There would be no soc mitigation for a few minutes due to booting sequence issue when soc is less than or equal to 10. Add notification to bcl driver once maxfg init_work is done. // Before [ 0.769687] bcl_peripheral:bcl_evaluate_soc test: battery_percentage = 100 [ 193.033827] bcl_peripheral:bcl_evaluate_soc test: battery_percentage = 9 [ 193.034239] thermal_sys: thermal_cdev_update test: thermal-cpufreq-4 = 18 // After [ 0.772569] bcl_peripheral:bcl_evaluate_soc test: battery_percentage = 100 [ 2.538930] bcl_peripheral:bcl_evaluate_soc test: battery_percentage = 10 [ 2.539657] thermal_sys: thermal_cdev_update test: thermal-cpufreq-4 = 18 Bug: 125245731 Test: bcl read correct soc earlier in local test Change-Id: I45eff19491c335f637f56051a47cb99e1f919ee4 Signed-off-by:<mandyshen@google.com>
-
- Apr 12, 2019
-
-
TeYuan Wang authored
Bug: 129787335 test: verified thermal-engine throttling by emul temp Change-Id: I3b687c9205387df2365e41af51fbbaf107c47f56 Signed-off-by:TeYuan Wang <kamewang@google.com>
-
Chiawei Wang authored
Bug: 129024273 Test: run lmkd_unit_test and speed test Signed-off-by:Chiawei Wang <chiaweiwang@google.com> Change-Id: I983af15dad3b62c933bee0f4ba48feb01d50db53
-
- Apr 11, 2019
-
-
Wei Wang authored
Test: Build Change-Id: I303f9fdf287e37fd0848060ff0195aac354fa7dd Signed-off-by:
Wei Wang <wvw@google.com> Signed-off-by:
davidchao <davidchao@google.com>
-
Ram Chandrasekar authored
Use the new threshold notification API to include the temperature read by the tsens driver. This will make sure thermal framework will use the same temperature as tsens driver to evaluate. Bug: 129787335 Bug: 129129453 Change-Id: Id50986c407052717fddc0554065dc711a2641c14 Signed-off-by:
Ram Chandrasekar <rkumbako@codeaurora.org> Signed-off-by:
-
Ram Chandrasekar authored
With virtual sensors, temperature reads by different zones at different time can be different and hence can result in different set of thresholds. This can lead to a case where we can get interrupt for the same temperature back to back if the temperature is fluctuating. To avoid this, allow sensor drivers to read the temperature and provide that information along when notifying for trip notification. Thermal framework will now make sure all the thermal zones will use the same temperature in all thermal zones to evaluate. Bug: 129787335 Bug: 129129453 Change-Id: I10861be54f0948081a83a1e839d8f9401929c61b Signed-off-by:
-
Manaf Meethalavalappu Pallikunhi authored
Thermal core framework subscribes for suspend/resume notification. On resume notification it re-evaluates each thermal zone for temperature and cooling state update. Most of the qcom targets, large number of thermal zones are enabled for different mitigations. Re-evaluating each thermal zone during resume leads to multiple issues including delay in back to back suspend resume scenario, power penalty for frequent wake up due to re-setting trip threshold especially during cold temperature usecases. But almost all qcom target TSENS and ADC sensor interrupts are wakeup capable. More over all internal cooling devices are platform cooling devices and no impact for cooling states during suspend. With such sensor and cooling device support, thermal core resume re-evaluation for each thermal zone is redundant and it can be ignored to avoid issues mentioned above. Add wake-capable-sensor property to thermal zone devicetree node to denote that these sensors are wakeup capable. If a thermal zone has this property defined, thermal framework ignores resume re-evaluation and can service the threshold notification during the suspend/resume path. Bug: 129787335 Bug: 129129453 Change-Id: I07edf80ad29009378af4c70e750d01bde6f30806 Signed-off-by:
Manaf Meethalavalappu Pallikunhi <manafm@codeaurora.org> Signed-off-by:
-
Wei Wang authored
Test: build Change-Id: I2557039bbd23306c034b7ad8ee1883cb3559b9e6 Signed-off-by: -
Albert Wang authored
Enable support for USB-connected xplay, ps3, and other generic USB controllers. Cherrypick from Id3bfcd8a2b98041af732c41b1a3ede490845fb4e Bug: 129367726 Change-Id: I744f0876368653abe3162bc1e5101b58905b866e Signed-off-by:Albert Wang <albertccwang@google.com>
-
- Apr 10, 2019
-
-
Jack Wu authored
Increase start soc from default 15% to 50% to run the learning more frequently. cl-start-capacity: 50% Adjust the high/low boundary to improve the accuracy. cl-max-increment: 0.5% cl-max-decrement: 2% Bug: 120401569 Signed-off-by:Jack Wu <wjack@google.com> Change-Id: Ie0ff344be5937b6bc3c3e6562bc7c265f57c642f
-
- Apr 09, 2019
-
-
Kelly Rossmoyer authored
Multiple device families have hit issues recently in which recurring, undesired kernel wakeups occur that only get logged as "unknown" by BatteryStats. Investigation has revealed that this can occur for a few different reasons, including: * driver configuring IRQ with IRQF_NO_SUSPEND when it is actually wakeup capable and should use enable_irq_wake() instead * driver configuring IRQ with both IRQF_NO_SUSPEND and enable_irq_wake() even though the kernel documentation clearly states that that is not appropriate * assertion of unmapped HW IRQs * wakeups handled by TZ (e.g. watchdog, which leaves no evidence of the wake reason once TZ handling has completed) When these wakeups occur, the wakeup_reason logging in the kernel doesn't capture the precipitating event, which results in the sysfs node containing last wake reason remaining empty. When BatteryStats sees that, it logs the wakeup reason as "unknown". The only other two formats that it supports (when a reason is present) are: * "Abort: <reason>" - indicates that an attempt to suspend was aborted for the specified reason * "<IRQ> <name>" - indicates that an interrupt with the specified virtual (_not_ HW) IRQ# and action name woke the kernel This CL includes changes that address the second and third scenarios listed above (i.e. _not_ the case where a driver uses IRQF_NO_SUSPEND when it should use enable_irq_wake()). It adds a new log_bad_wake_reason() function to our existing wakeup_reason code, adds a call to that function from the GIC driver when an unmapped HW IRQ is detected (which should pretty much never occur on a healthy device), and adds a call to irq_may_run() - which executes for virtually _every_ interrupt - in cases where a non-wakeup-armed interrupt turns out to be misconfigured with IRQF_NO_SUSPEND and enable_irq_wake(). Bug: 127863679 Bug: 127862997 Test: STILL UNDER TEST, DO NOT MERGE Change-Id: I838c77a19767b66820cc9dc3715923533c0e2f43 Signed-off-by:
Kelly Rossmoyer <krossmo@google.com> Signed-off-by:
Thierry Strudel <tstrudel@google.com>
-
Robin Peng authored
Bug: 119540871 Change-Id: I3e3ee8de933592054e4ad96a2efb6d3015e27489
-
Nick Chung authored
This is factory request only for Bonito readout some info(b/112119789). It has the code change from userspace. But B1C1 didn't ask for this request. To play it safe, we need to use CONFIG_BOARD_BONITO flag to separate it only for Bonito. Bug: 129738690 Bug: 127891929 Test: CTS, GCA Change-Id: I363b593e0d17e251baf64f7308d029fd85e690d4 Signed-off-by:Nick Chung <nickchung@google.com>
-
Nick Chung authored
Due to the finetune is different. Bug: 127891929 Bug: 129738690 Test: CTS, GCA Change-Id: Iebd859d96b554cd9b2d4745e271270234c5bf680 Signed-off-by: -
Nick Chung authored
CCI_TIMEOUT msecs_to_jiffies(50) is B1C1 WAR Bug: 127891929 Bug: 129738690 Test: CTS, GCA Change-Id: Ica3c764e794f0741841e00cbecf91d20664b17ff Signed-off-by:
-
- Apr 03, 2019
-
-
Robin Peng authored
Bug: 119540871 Change-Id: I35a68cbc21747073dc452bad0bad52d5cdd7ba46 Signed-off-by:Robin Peng <robinpeng@google.com>
-
- Apr 02, 2019
-
-
Krishna Manikandan authored
The dma buffer associated with the gem object is already unmapped during put_iova. Avoid unmapping it again in put_pages. Bug: 128365207 Change-Id: Iac57e164dde6f3e5913070acbe74b42691049913 Signed-off-by:Krishna Manikandan <mkrishn@codeaurora.org>
-
- Apr 01, 2019
-
-
Robin Peng authored
1. Use BUILD_CONFIG to build for bluecross or bonito 2. Also add friendly build_*.sh to indicate build command for bluecross and bonito Bug: 119540871 Change-Id: Ia2fd5c08d2b95dad57a26a34cc0bb4a266f969a4 Signed-off-by: -
Mark Salyzyn authored
Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
Mark Salyzyn <salyzyn@google.com> Bug: 64114943 Bug: 129556081 Change-Id: Ie8a2b9977a21022c204a19f1a8d781ea5a23c656
-
Hridya Valsaraju authored
pick_prebuilt script no longer packages them. Test: make Bug: 129350739 Change-Id: I620fff2451df16c44bab3550bbb98beb864aabe5 Signed-off-by:Hridya Valsaraju <hridya@google.com>
-
Petri Gynther authored
Bug: 115649324 Test: Manual testing Change-Id: I21779913e404ab776de622e109436f4f89122127 Signed-off-by:Petri Gynther <pgynther@google.com>
-
Robin Peng authored
Bug: 119540871 Change-Id: Ic655747be2e2733844ffd77251c2e275b32794c1 Signed-off-by:
-
- Mar 29, 2019
-
-
Petri Gynther authored
commit 86a136d9 ("Revert "ANDROID: input: keychord: Add keychord driver"") is the base for android-4.9-q branch. * Revert "ANDROID: input: keychord: Add keychord driver" drivers/input/misc/Kconfig drivers/input/misc/Makefile Revert "ANDROID: input: misc: keychord: move header to uapi" Revert "ANDROID: input: misc: keychord: log when keychord triggered" Revert "ANDROID: keychord: Fix a slab out-of-bounds read." Revert "Use %zu to print resid (size_t)." Revert "ANDROID: keychord: Fix races in keychord_write." Revert "ANDROID: keychord: Fix for a memory leak in keychord." Revert "ANDROID: keychord: Check for write data size" ANDROID: drop CONFIG_INPUT_KEYCHORD from cuttlefish and ranchu * BACKPORT: mm/debug.c: provide useful debugging information for VM_BUG mm/debug.c * UPSTREAM: bug: use %pB in BUG and stack protector failure kernel/panic.c lib/bug.c UPSTREAM: x86/alternative: Print unadorned pointers UPSTREAM: trace_uprobe: Display correct offset in uprobe_events * UPSTREAM: usercopy: Remove pointer from overflow report mm/usercopy.c UPSTREAM: Do not hash userspace addresses in fault handlers UPSTREAM: mm/slab.c: do not hash pointers when debugging slab UPSTREAM: kasan: use %px to print addresses instead of %p * BACKPORT: vsprintf: add printk specifier %px lib/vsprintf.c * BACKPORT: printk: hash addresses printed with %p lib/vsprintf.c * ANDROID: Fix race in crng_reseed() drivers/char/random.c * UPSTREAM: siphash: implement HalfSipHash1-3 for hash tables include/linux/siphash.h lib/siphash.c * UPSTREAM: siphash: add cryptographically secure PRF include/linux/siphash.h lib/Kconfig.debug lib/Makefile lib/siphash.c * BACKPORT: vsprintf: refactor %pK code out of pointer() lib/vsprintf.c BACKPORT: docs: correct documentation for %pK Merge upstream-f2fs-stable-linux-4.9.y into android-4.9 * BACKPORT: filemap: add a comment about FAULT_FLAG_RETRY_NOWAIT behavior mm/filemap.c * BACKPORT: filemap: drop the mmap_sem for all blocking operations mm/filemap.c * BACKPORT: filemap: kill page_cache_read usage in filemap_fault include/linux/pagemap.h mm/filemap.c * ANDROID: binder: remove extra declaration left after backport drivers/android/binder.c * UPSTREAM: net: socket: set sock->sk to NULL after calling proto_ops::release() net/socket.c * FROMGIT: binder: fix BUG_ON found by selinux-testsuite drivers/android/binder.c ANDROID: dm-bow: Fix 32 bit compile errors * ANDROID: fix 4.9 backport of psi header refactoring kernel/sched/core.c * UPSTREAM: mm: proc: smaps_rollup: fix pss_locked calculation fs/proc/task_mmu.c * UPSTREAM: binder: fix handling of misaligned binder object drivers/android/binder.c UPSTREAM: binder: fix sparse issue in binder_alloc_selftest.c * BACKPORT: binder: use userspace pointer as base of buffer space drivers/android/binder.c drivers/android/binder_alloc.c drivers/android/binder_alloc.h drivers/android/binder_trace.h * UPSTREAM: binder: fix kerneldoc header for struct binder_buffer drivers/android/binder_alloc.h * BACKPORT: binder: remove user_buffer_offset drivers/android/binder.c drivers/android/binder_alloc.c drivers/android/binder_alloc.h * UPSTREAM: binder: remove kernel vm_area for buffer space drivers/android/binder_alloc.c * UPSTREAM: binder: avoid kernel vm_area for buffer fixups drivers/android/binder.c * BACKPORT: binder: add function to copy binder object from buffer drivers/android/binder.c * BACKPORT: binder: add functions to copy to/from binder buffers drivers/android/binder.c drivers/android/binder_alloc.c drivers/android/binder_alloc.h * UPSTREAM: binder: create userspace-to-binder-buffer copy function drivers/android/binder.c drivers/android/binder_alloc.c drivers/android/binder_alloc.h ANDROID: Add dm-bow to cuttlefish configuration ANDROID: dm-bow: Backport to 4.9 ANDROID: dm-bow: backport to 4.14 * ANDROID: dm-bow: Add dm-bow feature drivers/md/Kconfig drivers/md/Makefile * f2fs: set pin_file under CAP_SYS_ADMIN fs/f2fs/file.c * f2fs: fix to avoid deadlock in f2fs_read_inline_dir() fs/f2fs/inline.c * f2fs: fix to adapt small inline xattr space in __find_inline_xattr() fs/f2fs/xattr.c * f2fs: fix to do sanity check with inode.i_inline_xattr_size fs/f2fs/inode.c fs/f2fs/super.c fs/f2fs/xattr.h * f2fs: give some messages for inline_xattr_size fs/f2fs/super.c * f2fs: don't trigger read IO for beyond EOF page fs/f2fs/data.c * f2fs: fix to add refcount once page is tagged PG_private fs/f2fs/checkpoint.c fs/f2fs/data.c fs/f2fs/dir.c fs/f2fs/f2fs.h fs/f2fs/node.c fs/f2fs/segment.c * f2fs: remove wrong comment in f2fs_invalidate_page() fs/f2fs/data.c * f2fs: fix to use kvfree instead of kzfree fs/f2fs/xattr.c * f2fs: print more parameters in trace_f2fs_map_blocks include/trace/events/f2fs.h * f2fs: trace f2fs_ioc_shutdown fs/f2fs/file.c include/trace/events/f2fs.h * f2fs: fix to avoid deadlock of atomic file operations fs/f2fs/segment.c * f2fs: fix to dirty inode for i_mode recovery fs/f2fs/file.c * f2fs: give random value to i_generation fs/f2fs/f2fs.h fs/f2fs/namei.c fs/f2fs/super.c * f2fs: no need to take page lock in readdir fs/f2fs/dir.c * f2fs: fix to update iostat correctly in IPU path fs/f2fs/segment.c * f2fs: fix encrypted page memory leak fs/f2fs/data.c * f2fs: make fault injection covering __submit_flush_wait() fs/f2fs/segment.c * f2fs: fix to retry fill_super only if recovery failed fs/f2fs/super.c * f2fs: silence VM_WARN_ON_ONCE in mempool_alloc fs/f2fs/data.c * f2fs: correct spelling mistake include/linux/f2fs_fs.h * f2fs: fix wrong #endif fs/f2fs/f2fs.h * f2fs: don't clear CP_QUOTA_NEED_FSCK_FLAG fs/f2fs/checkpoint.c * f2fs: don't allow negative ->write_io_size_bits fs/f2fs/super.c * f2fs: fix to check inline_xattr_size boundary correctly fs/f2fs/f2fs.h fs/f2fs/super.c include/linux/f2fs_fs.h * Revert "f2fs: fix to avoid deadlock of atomic file operations" fs/f2fs/segment.c * Revert "f2fs: fix to check inline_xattr_size boundary correctly" fs/f2fs/f2fs.h fs/f2fs/super.c include/linux/f2fs_fs.h f2fs: do not use mutex lock in atomic context * f2fs: fix potential data inconsistence of checkpoint fs/f2fs/data.c fs/f2fs/f2fs.h fs/f2fs/file.c fs/f2fs/inline.c * f2fs: fix to avoid deadlock of atomic file operations fs/f2fs/segment.c * f2fs: fix to check inline_xattr_size boundary correctly fs/f2fs/f2fs.h fs/f2fs/super.c include/linux/f2fs_fs.h * f2fs: jump to label 'free_node_inode' when failing from d_make_root() fs/f2fs/super.c f2fs: fix to document inline_xattr_size option * f2fs: fix to data block override node segment by mistake fs/f2fs/super.c * f2fs: fix typos in code comments include/linux/f2fs_fs.h * f2fs: use xattr_prefix to wrap up fs/f2fs/xattr.c * f2fs: sync filesystem after roll-forward recovery fs/f2fs/checkpoint.c fs/f2fs/node.c fs/f2fs/super.c * fs: export evict_inodes fs/inode.c fs/internal.h include/linux/fs.h * f2fs: flush quota blocks after turnning it off fs/f2fs/super.c * f2fs: avoid null pointer exception in dcc_info fs/f2fs/f2fs.h * f2fs: don't wake up too frequently, if there is lots of IOs fs/f2fs/segment.h * f2fs: try to keep CP_TRIMMED_FLAG after successful umount fs/f2fs/segment.c * f2fs: add quick mode of checkpoint=disable for QA fs/f2fs/checkpoint.c fs/f2fs/f2fs.h fs/f2fs/file.c fs/f2fs/segment.c fs/f2fs/super.c include/linux/f2fs_fs.h * f2fs: run discard jobs when put_super fs/f2fs/f2fs.h fs/f2fs/segment.c fs/f2fs/super.c fs/f2fs/sysfs.c * f2fs: fix to set sbi dirty correctly fs/f2fs/f2fs.h * f2fs: UBSAN: set boolean value iostat_enable correctly fs/f2fs/sysfs.c * f2fs: add brackets for macros fs/f2fs/f2fs.h * f2fs: check if file namelen exceeds max value fs/f2fs/dir.c * f2fs: fix to trigger fsck if dirent.name_len is zero fs/f2fs/dir.c * f2fs: no need to check return value of debugfs_create functions fs/f2fs/debug.c fs/f2fs/f2fs.h fs/f2fs/super.c * f2fs: export FS_NOCOW_FL flag to user fs/f2fs/f2fs.h fs/f2fs/file.c * f2fs: check inject_rate validity during configuring fs/f2fs/sysfs.c * f2fs: remove set but not used variable 'err' fs/f2fs/data.c * f2fs: fix compile warnings: 'struct *' declared inside parameter list include/trace/events/f2fs.h * f2fs: change error code to -ENOMEM from -EINVAL fs/f2fs/super.c Change-Id: I199547b8a925245ea0fd613296fc5cedb762c5f6 Signed-off-by:
-
Mark Salyzyn authored
Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit 630a1e7f. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit f4d1cf12. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit c4be12ac. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit 913d980e. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit a1e4c795. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit 59584701. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit 72a8dae2. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
This reverts commit f6738522. Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
Mark Salyzyn authored
Remove keychord driver, replaced in user space by https://android-review.googlesource.com/c/677629 . Signed-off-by:
-
This CL fixes race conditions inside easelcomm_stop_local and also fixes a race condition in easelcomm command handler. Bug: 112309571 Bug: 112312381 Change-Id: I7a7e8188869e66b48af10ef03438e2f8068dcf41 Signed-off-by:Alexander Perez <alexperez@google.com>
-
Hridya Valsaraju authored
pick_prebuilt script does not look for it any longer. Test: make Bug: 129350739 Change-Id: Id0413901932ace63712d1414937c937e47bccc83 Signed-off-by: -
Petri Gynther authored
Switch to 1 MiB static log buffer in __log_buf[]: define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT) static char __log_buf[__LOG_BUF_LEN] __aligned(LOG_ALIGN); instead of having the log buffer reallocated at boot by: setup_log_buf() log_buf_add_cpu() log_buf_len_update() new_log_buf = memblock_virt_alloc_nopanic() There is no need to do this reallocation for the log buffer. Change-Id: I81b9f1e2a89fe9e16dd702bda968c88877041083 Signed-off-by:
-
