The pagevec batching causes lru_add_drain_all which is too expensive
sometimes. This patch adds a new vendor hook to drain the pagevec
immediately depending on the page's type.
Bug: 251881967
Signed-off-by: Minchan Kim <minchan@google.com>
Change-Id: Id17e14e69197993ddad511a40c96e51674c02834
Signed-off-by: Richard Chang <richardycc@google.com>
(cherry picked from commit 2f8253b7e6e563cc19cffa120c72f6f528664103)
Add vendor hook for compaction begin/end. The first use would be
to measure compaction durations.
Bug: 229927848
Test: local kernel build test
Signed-off-by: Robin Hsu <robinhsu@google.com>
Change-Id: I3d95434bf49b37199056dc9ddfc36a59a7de17b7
Signed-off-by: Richard Chang <richardycc@google.com>
(cherry picked from commit 13b6bd38bb1f43bfffdb08c8f3a4a20d36ccd670)
CMA first allocation policy for movable makes CMA(Upstream doesn't)
area always full. It's good for memory efficiency since it could use
up CMA available memory most of time. However, it could cause
cma_alloc slow since it causes a lot page migration all the time.
Let's add vendor hook for someone who want to restore CMA allocation
policy to upstream so they will see less page migration in cma_alloc.
If the vendor_hook returns false, the rmqueue_bulk return 0 without
filling pcp->lists so get_populated_pcp_list will return NULL.
Once get_populated_pcp_list returns NULL, __rmqueue_pcplist will retry
the page allocation with original migratetype(currently, original
migratetype couldn't be MIGRATE_CMA) so the retrial will find
available pages from !MIGRATE_CMA free list.
Bug: 231978523
Signed-off-by: Minchan Kim <minchan@google.com>
Change-Id: Ia031d9bc6f34085b892a8d9923bf5b9b1794f94a
Signed-off-by: Richard Chang <richardycc@google.com>
(cherry picked from commit 0ca85e35bf5b4a7ff08f00a060c83e4a82380b64)
This reverts commit 816e633095. This is
not needed in this branch. Upstream changes will be picked instead.
Bug: 255532233
Change-Id: I011e2377b0444d59d122f35f4d5527de81bf9510
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Add vendor enable_thermal_genl_check logic.
Filter on-die tz genl event.
To avoid thermal-hal being woken up all the time by thermal genl events,
only the selected thermal_zone and cooling_device can send events from
kernel.
Bug: 170682696
Test: boot and thermal-hal can receive thermal genl events from kernel
Change-Id: Idb3f4b07a2a2740c01d8785910878bfe6edc832d
Signed-off-by: davidchao <davidchao@google.com>
Some hardware has a way to restore all keyslots at once that is
significantly faster than restoring each keyslot individually, as is
done by blk_ksm_reprogram_all_keys(). Add a hook
"android_rvh_ufs_reprogram_all_keys" that allows overriding the
restoration of all keyslots after UFS reset. This may sleep, so this
must be a "restricted" Android vendor hook rather than a regular one.
Note that currently this functionality can't be upstreamed, as support
for the hardware that needs it would need to be upstreamed first.
(cherry picked from commit e2e063f507)
Bug: 162257402
Bug: 181905172
Bug: 241106918
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Kenny Root <kroot@google.com>
Change-Id: I0b25393a5131941f085892560e08a64e63cd1369
Currently the core UFS driver does not have a vops to notify when the
device is operational. This commit introduces a hook, which serves to
notify device completing initialization and is ready to accept I/O.
This is required by the FIPS140-2 [1] self integrity test of inline
encryption engine, which must run whenever the host controller is reset.
The code requires sleeping while waiting for I/O to complete and allocating
some memory dynamically, which requires the vendor hook to be restricted.
[1] https://csrc.nist.gov/publications/detail/fips/140/2/final
Bug: 185809932
Signed-off-by: Konstantin Vyshetsky <vkon@google.com>
Change-Id: I6f476f9c2e2b50574d2898c3f1ef6b648d92df24
(cherry picked from commit e774e4eca6)
Change how vendor hooks are instantiated to promote
more complete structure definition in the ABI XML
description without complicating hook definition for
partners.
We don't want to force partners to include all
headers as part of the hook definition in include/trace/hooks/
since that causes extra headers to be included in
source files that runs the risk of changing visibility
resulting in CRC changes to KMI symbols. Instead continue
the practice of using forward declarations in the hook
header files.
Instead of instantiating all hook tracepoints globally in
drivers/android/vendor_hooks.c, use subsystem-specific
vendor_hooks.c if inclusion of subsystem-specific
header files is required. This avoids namespace collisions
between internal header files and limits the exposure
to the internal headers to the instantiation, not the
call sites.
In this patch, all of the scheduler related hooks are
instantiated in kernel/sched/vendor_hooks.c which can
cleanly include scheduler-related header files to
provide full type visibility.
Bug: 233047575
Signed-off-by: Todd Kjos <tkjos@google.com>
Change-Id: Ife5a66c2968de73e3f6d05840411310611e2e175
Vendor may have need to track rt util.
Bug: 201261299
Signed-off-by: Rick Yiu <rickyiu@google.com>
Change-Id: I2f4e5142c6bc8574ee3558042e1fb0dae13b702d
Vendor may have the need to implement their own util tracking.
Bug: 201260585
Signed-off-by: Rick Yiu <rickyiu@google.com>
Change-Id: I973902e6ff82a85ecd029ac5a78692d629df1ebe
Reserve sufficient nl80211 attributes in enum nl80211_commands and enum
nl80211_attr to backport upstream UAPI changes.
Reserve data fields for upstream backport changes and vendor hooks.
Also, add vendor hooks to set/get additional info from driver based on
the cfg80211 current context. These APIs can be used to configure/fetch
data of the new params added in upstream cfg80211_op calls.
Ex:
upstream kernel added a new param to existing cfg80211_op "add_key()".
The addtional param info added in the upstream kernel can be configured
using "android_vh_cfg80211_set_context" before calling "add_key".
Bug: 233387627
Change-Id: I99f696ad291ae5965fc6d91bf565cc0c9d423dd5
Signed-off-by: Veerendranath Jakkam <quic_vjakkam@quicinc.com>
Some shrinker add lock in count_objects() may cause lock contention
issues and lead all task stall on slab shrink. Add vendor hook in
do_shrink_slab() for shrinker->count_objects() latency measuring.
Add 3 oem data in shrink_control struct. Two is for
shrinker->count_objects() and shrinker->scan_objects() latency
measuring, other one to store priority, some shrinker know the reclaimer
priority can control the memory reclaim more better.
Bug: 188684131
Change-Id: I80e9d90179bb52a99c54d9a067c6fcee835bb2ad
Signed-off-by: rongqianfeng <rongqianfeng@vivo.com>
Some vendors want to add a field when a 'sruct sock' is added so give a
hook to handle this. Any memory allocated when
trace_android_rvh_sk_alloc() is called needs to be freed when
trace_android_rvh_sk_free() is called.
Note, if trace_android_rvh_sk_alloc() fails, be sure to be able to
handle this in trace_android_rvh_sk_free(), but that should not be an
issue as that needs to be addressed in vendor code that runs for 'struct
sock' objects that have been created before the vendor code is loaded no
matter what.
Bug: 171013716
Signed-off-by: Vignesh Saravanaperumal <vignesh1.s@samsung.com>
Change-Id: I108a2f31d2dcc228f46159816deee6235afafbbd
Some vendors want to add a field when a 'sruct nf_conn' is added so give a
hook to handle this. Any memory allocated when
trace_android_rvh_nf_conn_alloc() is called needs to be freed when
trace_android_rvh_nf_conn_free() is called.
Note, if trace_android_rvh_nf_conn_alloc() fails, be sure to be able to
handle this in trace_android_rvh_nf_conn_free(), but that should not be
an issue as that needs to be addressed in vendor code that runs for
'struct nf_conn' objects that have been created before the vendor code
is loaded no matter what.
Bug: 171013716
Signed-off-by: Vignesh Saravanaperumal <vignesh1.s@samsung.com>
Change-Id: I4d2b025196a3df7ba4adec313c90483811cac728
The associated vendor hooks/data are used for implementing dynamic memory.low protection based on memcgv2.
Bug: 232723420
Test: build pass
Change-Id: I2e92bdc2840af1eaaa08ee6427d2a82d78390005
Signed-off-by: zhaoyang.huang <zhaoyang.huang@unisoc.com>
Add vendor hook to get signal for vendor-specific tuning.
Conflicts:
drivers/android/vendor_hooks.c
Bug: 184898838
Signed-off-by: Zhuguangqing <zhuguangqing@xiaomi.com>
Change-Id: I83a28b0a6eb413976f4c57f2314d008ad792fa0d
(cherry picked from commit d623f1ff72)
When watchdog or anr occur, we need to read
dev/binderfs/binder_logs/proc/pid or dev/binderfs/binder_logs/state
node to know the time-consuming information of the binder call.
We need to add the time-consuming information of binder transaction.
Bug: 190413570
Signed-off-by: zhang chuang <zhangchuang3@xiaomi.com>
Change-Id: I0423d4e821d5cd725a848584133dc7245cbc233a
(cherry picked from commit eabe9707f2)
Signed-off-by: Carlos Llamas <cmllamas@google.com>
[cmllamas: fix trivial merge conflict in vendor_hooks.c]
Need this vendor hook to support patial init functionality. This hook needs to be restricted as the additional vendor handling need to perform operations (such as wait_for_completion) in non-atomic context.
One of the main functionality of Partial init feature is to reduce latency during resume.
Bug: 231654364
Link: https://lore.kernel.org/lkml/ea318551ec0b43429ce47bfac9ebf285@quicinc.com/#r
Change-Id: Ia1cb0b7173db3318d2d6686f4bd4719f3cfefd8b
Signed-off-by: kamasali <quic_kamasali@quicinc.com>
Add hooks to init oem data of mutex.
Bug: 231527236
Signed-off-by: Dezhi Huang <huangdezhi@hihonor.com>
Change-Id: Id0aeac168e81bd3d88051657c32ba709c329dbdd
Add hook to tcp/udp to collect network info and do performance tuning.
Bug: 190523684
Change-Id: Id790a381d5ce6c35a747697510f14678ccf3ff2f
Signed-off-by: Liangliang Li <liliangliang@vivo.com>
(cherry picked from commit 81c8161bed)
We want to get binder free space information in the binder
transaction, but this needs to hold the mutex lock. So we add
this restrict hook to do this.
Bug: 205648032
Change-Id: Ie1f377018da686bd62f5ac2d1e5421899741e6d5
Signed-off-by: Liangliang Li <liliangliang@vivo.com>
(cherry picked from commit 2f3f5731de)
Add a hook in account_process_tick, which help us to get information
about the high load task and the cpu they running on.
Bug: 183260319
Change-Id: I54162ce3c65bd69e08d2d4747e4d4883efe4c442
Signed-off-by: Liujie Xie <xieliujie@oppo.com>
(cherry picked from commit 63399b4e2e)
Add vendor hook for passing events to client drivers.
We have some limitation to achieve the goal of input performance
without modification of kernel. So I would like you to add vendor
hook and please review the reason why I want to add it.
As the users trigger input actions, we need to support the input
operation and the what the users want to do so that there's no delay.
So we've developed and used "input booster" controling some resources
properly during input operations according to input type.
Android already has the booster in framework to support input actions
however this way has a long latency to control resources.
To reduce latency, we've placed the booster in kernel side for high
responsiveness and it guarantees faster than framework booster.
In addtion, I applied Samsung's own concept to the booster.
And it has been operated stably on all Galaxy models for about 10 years.
So we need
- add function to notify input events delivered to client in evdev.
Bug: 228927595
Signed-off-by: HongCheol Shim <hongc.shim@samsung.corp-partner.google.com>
Change-Id: Ibe199c148974b6788e02c192f950cb6802745f68
Add hook to dup_task_struct for vendor data fields initialisation.
Bug: 188004638
Change-Id: I4b58604ee822fb8d1e0cc37bec72e820e7318427
Signed-off-by: Liangliang Li <liliangliang@vivo.com>
(cherry picked from commit f66d96b14a)
Need this verndor hook to support wakeup functionality
for sdio cards.
Bug: 231668342
Change-Id: I1d1a34d6d02f4558edf56eccdd09df57d6232c05
Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Signed-off-by: Kamasali Satyanarayan <quic_kamasali@quicinc.com>
Add a hook in irqtime_account_process_tick, which helps to get
information about the high load task.
Bug: 187904818
Change-Id: I644f7d66b09d047ca6b0a0fbd2915a6387c8c007
Signed-off-by: Liangliang Li <liliangliang@vivo.com>
(cherry picked from commit fe580539f6)
Add vendor hook to thermal to allow vendor to selectively disable
thermal cooling device stats feature based on requirement. It helps
vendor to optimize memory footprint due to this feature especially
for low memory devices.
Bug: 218825214
Change-Id: I2ec72505f03575e09229c54765584614b16a3904
Signed-off-by: Manaf Meethalavalappu Pallikunhi <quic_manafm@quicinc.com>
Add hook in try_to_unmap_one() to trace this function for debug memory
swap bugs.
Bug: 198385827
Change-Id: I1fdbe60e09bb491b949e06a07133710453ecca03
Signed-off-by: Jiewen Wang <jiewen.wang@vivo.com>
(cherry picked from commit 955f917251)
Add hook in mmap_region() to record the vma and address information
of monitored processes.
Bug: 198385827
Change-Id: I0bde29113b47ca7f4a9f5d42a54188e791ca3b7e
Signed-off-by: Jiewen Wang <jiewen.wang@vivo.com>
(cherry picked from commit 878e0caa77)
This vendor hook let us initialize payload of the request.
Bug: 188749221
Change-Id: I51d6a3010ac0ab36066dbe1368158592832112b7
Signed-off-by: Yang Yang <yang.yang@vivo.com>
(cherry picked from commit 2faed77792)
This vendor hook let us attach oem data as payload to the request.
The payload is used by oem driver for debugging purpose.
Bug: 188749221
Change-Id: Iac598bd9cce836dac0efe9198a3e7752928f351a
Signed-off-by: Yang Yang <yang.yang@vivo.com>
(cherry picked from commit eecc725a8e)
Provide a vendor hook to skip cma-pages to add in pcplist when
free_unref_page_commit.
The patch is revelant to skip drain_all_pages in alloc_contig_range,
the revelant hooks is android_vh_cma_drain_all_pages_bypass
which is to avoid to delay in drain pcppages when drain_all_pages.
In most case, pcp->high is small so that free-pages with other mt_types
can also fill with pcplist full.
Bug: 224732340
Bug: 234405962
Signed-off-by: Peifeng Li <lipeifeng@oppo.com>
Change-Id: Ifdeeed9f8934d87671ec3fa6787a02675b993082
Provide a vendor hook to allow drain_all_pages to be skipped
during alloc_contig_range in some cases to avoid delays caused by
it in cases when the benefits of draining pcp lists are known
to be small.
Bug: 224732340
Bug: 234405962
Signed-off-by: Peifeng Li <lipeifeng@oppo.com>
Change-Id: I0a82f668cf985ad5344d666c0c6372a7e61c3798
Provide a vendor hook to allow drain_all_pages to be skipped
during direct reclaim in some cases to avoid delays caused by
it in cases when the benefits of draining pcp lists are known
to be small.
Bug: 220811627
Bug: 234405962
Signed-off-by: Liujie Xie <xieliujie@oppo.com>
Change-Id: I0805241f81e0a94afcf62c98e97cff125d4061e2
Provide a vendor hook to allow page_referenced to be skipped
during shrink_active_list to avoid heavy cpuloading caused by
it.
Bug: 220878851
Bug: 234405962
Signed-off-by: Liujie Xie <xieliujie@oppo.com>
Signed-off-by: Peifeng Li <lipeifeng@oppo.com>
Change-Id: Ie0e369f8f8739fea59a95470af20ab0e976869d1
Some of the irq migration paths call chip set affinity, after
current CPU is marked offline in cpu_online_mask. These
chip set affinity calls do not invoke vendor trace hooks.
So, convert gic_v3_set_affinity() vendor hook to a restricted
hook, to allow trace hook to be called from these irq migration
paths.
Bug: 187161770
Change-Id: I8f45536deb1ba1dc6be861ca4fc2b32306a5c50a
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
(cherry picked from commit 3bd9ad7eb4)
Currently the code has single hook for tracking irqs. However
modules need to deduce start and end of the irq.
Create separate hooks for irq start and end since the
cputime has already figured it out.
Bug: 231341763
Change-Id: Ie0dd503b283d83f69d01171ebd1cd6127c3bafd0
Signed-off-by: Shaleen Agrawal <quic_shalagra@quicinc.com>
Signed-off-by: Ashay Jaiswal <quic_ashayj@quicinc.com>
UNISOC's solution allows more requests to be issued to improve read and
write performance, so vendor hooks are added to support this function.
Bug: 229920089
Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com>
Change-Id: Ic9423b651884111d286646d0d66a033fce3b44d7
While TOS is running alongside with linux, cpu power off operation by linux
may need be denied by TOS in some scenarios.
This patch added two hooks in psci_tos_resident_on and psci_cpu_suspend
to hook cpu off operation.
The function psci_tos_resident_on originally is used to check if TOS is resident on
a specific cpu and that cpu is dedicated for running TOS exclusively. If so, that
cpu can not be power off. Actually if TOS supports SMP, TOS may need deny any
other cpu to power down in some cases, i.e. there are no-expired timers in TOS.
Thus the first hook for psci_tos_resident_on is used to determine if
the specific cpu is allowed to power off in the cpu hotplug path.
Besides cpu hotplug, a cpu also can power off by cpu_suspend.
The second hook for psci_cpu_suspend determines if cpu suspend should go through
or not. When the same conditions described above meets, cpu suspend will break up.
The hook is same as commit 88d88955ae0b8b1f1a555d7810beb6c8ca4ca0f1 from android12-5.4
Bug: 229067021
Change-Id: Ib329beeff20f0cfef263f6a7813280d33f6a5eaa
Signed-off-by: Jian Gong <Jian.Gong@unisoc.com>
