The commit <a0a9abc9fc> (pil: Fix minidump collection during
SSR) failed to set the value of 'ret' in pil_do_minidump() upon
the return failure from prepare_minidump_segments(). As a result,
the 'goto seg_mapping_fail' will end up returning uninitialized
value. Hence, assign an error code explicitly to avoid this situation.
Change-Id: I20f0d2c4fd006483c73541937025ed50b265e2b4
Signed-off-by: Raghavendra Rao Ananta <rananta@codeaurora.org>
Restore the generic system heap implementation such that it does not have
any MSM specific additions as of android11-5.4
commit a9a13eeea9 ("Merge 5.4.45 into android-5.4-stable").
Change-Id: Iedc33377d5fa64513d43459d305ff98ce2091a61
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Add support to configure bus dcvs sampling interval through the
sysfs entry. Also expose cur_ab vote value to sysfs.
$ echo 25000 > /sys/class/devfreq/soc:qcom,kgsl-busmon/sampling_interval
$ cat /sys/class/devfreq/soc:qcom,kgsl-busmon/cur_ab .
Change-Id: I77df55de2766d3fd818c27f717cfc4ae598bb0bd
Signed-off-by: Hareesh Gundu <hareeshg@codeaurora.org>
Remove the support for sending event notifications via
event_notify as there are no users of it. Update the
ABI snapshot accordingly.
Change-Id: I22f28339ed1e834ae0dbd762562a52a158c12406
Signed-off-by: Raghavendra Rao Ananta <rananta@codeaurora.org>
In preparation for restoring the GKI system heap, move the existing
system heap into the vendor heap space, leaving the GKI system heap
ID slot empty. For now, alias allocations targeted at ION_HEAP_SYSTEM to
ION_SYSTEM_HEAP_ID until all clients have migrated. Once that
happens, we can remove the aliasing.
Change-Id: Ib1bd74c2a54ebfdcb14da2b04f237ae1b88db189
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
The MSM implementation of the system heap currently lives
in the same source files as where the GKI system heap source
lives. This is problematic, as the MSM system heap needs to
co-exist with the GKI system heap. While they share a lot of
code, it is not possible for them to use the same source files,
as the same symbols would have to be available for GKI for the
system heap, as well as for the MSM system heap. Additionally,
the MSM system heap changes some of the signatures of the shared
functions, which makes it even more complicated for the two heaps
to share source files. Thus, in preparation to bring back the
GKI system heap, move the MSM system heap to its own files. This
change does not introduce any functional impact. Also, namespace
the page pool interface, as it has common global symbols with the
GKI ION page pool interface, which will lead to linking issues
with duplicate symbols.
Change-Id: I7cadb752add96d2f3a392240946c643c673711a1
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Lots of symbols have now been removed from our tracking, so update the
.xml file to notice this so that if we need to track them in the future,
we will see it.
This is primarily due to the building in more Bluetooth protocols.
Bug: 157965270
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Ic3a12af18fb8583e6d7bf21b28758d0d6e194f98
This change removes vbatt feature code from 5.4
kernel to avoid compilation issue realted to adc tm
api's. This feature will be enabled once conflicts
between icnss2 and adc_tm configs is resolved.
Change-Id: If875e93edfea3bbcccd3194f95f534933131e42d
Signed-off-by: Sandeep Singh <sandsing@codeaurora.org>
GPUCC registers are valid for all A6xx GPUs. Hence add
these register to A6xx snapshot dumping.
Change-Id: I6b8d1a856ed096afff7d22b4382ec6ac09b7b1fb
Signed-off-by: Hareesh Gundu <hareeshg@codeaurora.org>
Multiple platforms need these, so build them in so we don't need to
support an api we don't really need to.
Bug: 157965270
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Icc99311bd349e226cf61deba98414d5e48750c54
memblock_dbg is not up because of parse_early_param get called after
setup_machine_fd. To capture fdt reserved info below pr_info is added.
Change-Id: Id39ce71c3de63778e987d7ab09a6d41c6aaddc4f
Signed-off-by: Faiyaz Mohammed <faiyazm@codeaurora.org>
[vinmenon@codeaurora.org: move the location of print for 5.4]
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
Signed-off-by: Prakash Gupta <guptap@codeaurora.org>
For INITRD and initmem memory is reserved through "memblock_reserve"
during boot up but it is free via "free_reserved_area" instead
of "memblock_free".
For example:
[ 0.294848] Freeing initrd memory: 12K.
[ 0.696688] Freeing unused kernel memory: 4096K.
To get the start and end address of the above freed memory and to account
proper memblock added memblock_dbg log in "free_reserved_area".
After adding log:
[ 0.294837] memblock_free: [0x00000083600000-0x00000083603000] free_initrd_mem+0x20/0x28
[ 0.294848] Freeing initrd memory: 12K.
[ 0.695246] memblock_free: [0x00000081600000-0x00000081a00000] free_initmem+0x70/0xc8
[ 0.696688] Freeing unused kernel memory: 4096K.
Change-Id: I5ebbb9c1ced481e0ce2acd839e85526aa8dadf75
Signed-off-by: Faiyaz Mohammed <faiyazm@codeaurora.org>
Lots of symbols have now been removed from our tracking, so update the
.xml file to notice this so that if we need to track them in the future,
we will see it.
This is primarily due to the building in of one Bluetooth protocol
driver, which was pulling in a lot of things.
Bug: 157965270
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I1681acd4840dccc3c5f04a0356819b1ab3aae3ad
Might as well build it in so we don't have to worry about ABI issues
over time.
Bug: 157965270
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I285bf3e3bf4bb9c649c3ac667834c6d12c7e1f4a
It was enabled for arm64, but for some reason disabled for x86, which
led to lots of confusion when trying to do some build testing...
Bug: 157965270
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Ic8e295c4d1f1cb368d386f2ba5ee1a5edbb83a96
Notify WLAN host drive FW down as soon as SMMU fault happens so
that it can do necessary cleanup as earlier as possible.
Change-Id: Ie34243de296d35ced3f2c1a1e4d50a6c551ae88d
Signed-off-by: Yue Ma <yuem@codeaurora.org>
SMMU fault is observed due to spurious AXI transaction even after
gcc video_axi clocks are disabled. HW underlying may not actually turn off
clocks since HW enable signals can still be asserted/waiting on certain
scenarios like when CVP and MVP both are power collapsed.
Use clk_branch2_force_off_ops to guarantee that clocks are actually
turned off when SW disables them.
Change-Id: Iaa9250a5962f3e74e971b4fc859d246960f792b0
Signed-off-by: Jagadeesh Kona <jkona@codeaurora.org>
This reverts commit ec877350d0.
This is a preparation change for merging android-5.4-stable.50
into msm-5.4 branch. Newer version of the imported change is
committed already as:
4c9af3eb8e f2fs: fix use-after-free when accessing bio->bi_crypt_context
Change-Id: Ie0ff68e50ad1cfe8a3a257b2d24e55a000bbe5c2
Signed-off-by: Blagovest Kolenichev <bkolenichev@codeaurora.org>
To minimize wake up latency for binder transactions,
set low_latency flag.
Change-Id: If103199fc96e35d90b0eaaf32da38ae135976658
Signed-off-by: Kishore Sri venkata Ganesh Bolisetty <bsrivenk@codeaurora.org>
This reverts commit 20235bca1c.
Revert this for GKI configuration.
Change-Id: I9e2b54ff875987c7bb7f9a6a42f62132d673ae0f
Signed-off-by: Asutosh Das <asutoshd@codeaurora.org>
An extra vote is held and not released during cases of multiple
or back to back time synchronization requests where the response
is still pending and the doorbell is not rung. As only one
response is expected, we can safely remove the latest vote
placed.
Change-Id: I551ac2721af79162959ca301a61657cdf17cf135
Signed-off-by: Bhaumik Bhatt <bbhatt@codeaurora.org>
Remove the flush_cache_all usage from power off driver.
For the crash reset flow, this api is not needed on
latest SoCs and secure world can flush cpu caches.
Change-Id: If50364810e63c72c09ada026f7f295112414c413
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
ABI changes as a result of SPF padding additions for GKI 1.0.
Leaf changes summary: 3072 artifacts changed (34 filtered out)
Changed leaf types summary: 3 leaf types changed
Removed/Changed/Added functions summary: 0 Removed, 3026 Changed (34 filtered out), 0 Added function
Removed/Changed/Added variables summary: 0 Removed, 43 Changed, 0 Added variable
3026 functions with some sub-type change:
...
'struct mm_struct at mm_types.h:383:1' changed:
type size changed from 7040 to 7104 (in bits)
there are data member changes:
...
2847 impacted interfaces
'struct vm_area_struct at mm_types.h:293:1' changed:
type size changed from 1728 to 1792 (in bits)
1 data member insertion:
'u64 vm_area_struct::android_vendor_data1', at offset 1728 (in bits) at mm_types.h:370:1
2847 impacted interfaces
'struct vm_fault at mm.h:418:1' changed:
type size changed from 960 to 1088 (in bits)
2 data member insertions:
'u64 vm_fault::android_vendor_data1', at offset 960 (in bits) at mm.h:457:1
'u64 vm_fault::android_vendor_data2', at offset 1024 (in bits) at mm.h:458:1
2847 impacted interfaces
Bug: 155842923
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Change-Id: Ieb63c9897da176c75215c60dcbe72d2022203caa
Fields are added on mm_struct, vm_area_struct and vm_fault
to support vendor feature, speculative page fault.
Bug: 155842923
Change-Id: Iedda086d6f83f1353da5a2ed1e24f7ba7dde43f2
Merged-In: Iedda086d6f83f1353da5a2ed1e24f7ba7dde43f2
Signed-off-by: Hyesoo Yu <hyesoo.yu@samsung.com>
Doorbell resources may not be ready when the qrtr haven driver probes.
If this happens on the secondary vm, then communication will stall
because both peers will have missed the incoming notification.
Add a retry to the secondary vm that calls the doorbell with the
blocking option if the first send fails.
Change-Id: I4f1f47383708a1fc108870317b350f8db5d76f4a
Signed-off-by: Chris Lew <clew@codeaurora.org>
Currently, with LED symmetry enabled, turning on a switch without
configuring current in a channel first results in errors being seen.
Without LED symmetry enabled, the same action results only in
qti_flash_led_strobe() being called with the value argument set to zero.
Bring about the same behaviour for LED symmetry as well.
Change-Id: I2d6301ec3cc19d89acbd0545509485866ffa9085
Signed-off-by: Guru Das Srinagesh <gurus@codeaurora.org>
This enables both CONFIG_USB_STORAGE=y and CONFIG_USB_UAS=y so that
systems that want to support usb storage devices will work with the GKI
base kernel image.
Bug: 161091250
Bug: 161190032
Cc: zhang sanshan <pete.zhang@nxp.com>
Cc: JaeHun Jung <jh0801.jung@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Ibda9ad03aea3f192c057fd3e479742d52d060f7b
Change trip threshold, trip clear, passive polling delay and algorithm
type parameter name in the thermal zone configuration output to
set_temp, clr_temp, passive_delay and algo_type from trip_threshold,
trip_threshold_clr, passive_polling_delay and policy.
Change-Id: I5d531eaf0990202b7a0537e1d3a2ed3b145543ab
Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org>
IV_INO_LBLK_32 policies introduced the possibility that logically
contiguous data blocks might not have contiguous DUNs (because of
potential DUN wraparound). As such, whenever a page is merged into a
bio, fscrypt_mergeable_bio() must be called to check DUN contiguity.
Further, fscrypt inline encryption does not handle the case when the DUN
wraps around within a page (which can happen when the data unit size !=
PAGE_SIZE). For now, we handle that by disallowing inline encryption
with IV_INO_LBLK_32 policies when the data unit size != PAGE_SIZE (and
dropping the now redundant check for this in fscrypt_dio_supported()).
Bug: 144046242
Change-Id: I9cb414fcc284b197b9d3d1b9643029c6b875df5a
Signed-off-by: Satya Tangirala <satyat@google.com>
The path will be released when fail to enable any sink or link. The
path must be deleted from active paths before release it. Otherwise
it will get invalid address of path when enable or disable the paths.
Change-Id: Id37fdbc3854bae414f043bd049bc6af2e90af5fa
Signed-off-by: Mao Jinlong <jinlmao@codeaurora.org>
Remove mass storage as module. Now the driver is set default to "y" in
gki_defconfig.
Change-Id: Iacf1b5f363ff8ad8c8bdd0079c39a11e5f1f5ee4
Signed-off-by: Murali Nalajala <mnalajal@codeaurora.org>
