GitHub/LineageOS/android_kernel_motorola_exynos9610.git
6 months agoRevert "[HACK] Make kernel panic when verity error is detected for debugging"
Hyeyeon Chung [Wed, 29 Aug 2018 08:33:36 +0000 (17:33 +0900)]
Revert "[HACK] Make kernel panic when verity error is detected for debugging"

This reverts commit c674aa0e66e07baf5d3e73ee10a9114c05e30bc6.

Change-Id: I5559860846bfb8508ec8957d9edcf0d423b5484a

6 months ago[COMMON] media: smfc: change variable for HW runtime
Seungchul Kim [Fri, 18 Aug 2017 06:31:59 +0000 (15:31 +0900)]
[COMMON] media: smfc: change variable for HW runtime

Variable that stores the HW runtime of smfc is changed to
print a HW runtime on platform log.

Change-Id: Ia89bd96fe4b38fc4baa978b7a380b93600b1b046
Signed-off-by: Seungchul Kim <sc377.kim@samsung.com>
6 months ago[COMMON] media: smfc: add devfreq setting
Seungchul Kim [Wed, 2 Aug 2017 01:50:38 +0000 (10:50 +0900)]
[COMMON] media: smfc: add devfreq setting

devfreq setting for smfc is added to callback of runtime PM.
If dtsi defines minlock of smfc, 'runtime resume' sets a minlock
that is defined in dtsi and 'runtime suspend' sets a minlock to 0.

Change-Id: Ic1cb9fc24db0aa773add277a49ece9d709c92174
Signed-off-by: Seungchul Kim <sc377.kim@samsung.com>
6 months ago[COMMON] media: smfc: measure HW runtime
Seungchul Kim [Thu, 1 Jun 2017 01:27:27 +0000 (10:27 +0900)]
[COMMON] media: smfc: measure HW runtime

In 4.9 kernel, HW runtime is stored in timestamp
that moves from vb2_v4l2_buffer to vb2_buffer.

Change-Id: I805da6e155f683d10e96824b61d83dd229893635
Signed-off-by: Seungchul Kim <sc377.kim@samsung.com>
6 months ago[9610] wlbt: fix prevent issue CID:240470,240473,240480
Youngsoo Kim [Wed, 29 Aug 2018 14:03:42 +0000 (23:03 +0900)]
[9610] wlbt: fix prevent issue CID:240470,240473,240480

drivers/net/wireless/scsc/hip4_smapper.c

CID 240470 (#1 of 1): Out-of-bounds write (OVERRUN)
6. overrun-local: Overrunning array control->lookuptable of 10 bytes at byte offset 251 using index bank->bank (which evaluates to 251).

CID 240473 (#1 of 1): Out-of-bounds read (OVERRUN)
4. overrun-local: Overrunning array control->lookuptable of 10 bytes at byte offset 10 using index bank_num (which evaluates to 10).

drivers/misc/samsung/scsc/mxlogger.h

CID 240480 (#1 of 1): Uninitialized scalar variable (UNINIT)
2. uninit_use_in_call: Using uninitialized value sync_r. Field sync_r.fw_time is uninitialized when calling memcpy. [Note: The source code implementation of the function has been overridden by a builtin model.]

Change-Id: I5fd7e08d84d0d2e21c9416229f827043d55a2e79

6 months agodma-buf: fix wrong variable of task to check condition
hyesoo.yu [Thu, 23 Aug 2018 04:49:28 +0000 (13:49 +0900)]
dma-buf: fix wrong variable of task to check condition

dmabuf trace should release all references between task
and buffer when the task exit.

To know when the task exit, the dmabuf trace registers
anon file descriptor and that should be closed only by
do_exit. Thus, the task need to have 'PF_EXITING'
when closing the file system. If not, it is possible
for user to manage file descriptor incorrectly, so we
check the flags of task.
However, release function could be called by kernel thread
though delayed work, we have to check the task from
file inode's private data instead of 'current'

Change-Id: Id2750fc0f4231817537e183df7bc75d010b348b0
Signed-off-by: hyesoo.yu <hyesoo.yu@samsung.com>
6 months agodma-buf: fix uninitialized variable on error state
hyesoo.yu [Wed, 22 Aug 2018 05:42:58 +0000 (14:42 +0900)]
dma-buf: fix uninitialized variable on error state

Change-Id: Ifdb8a7e013575a0cc7885c0522d86b2dafb0f854
Signed-off-by: hyesoo.yu <hyesoo.yu@samsung.com>
6 months agodma-buf: introduce dma-buf trace
hyesoo.yu [Fri, 29 Jun 2018 00:56:53 +0000 (09:56 +0900)]
dma-buf: introduce dma-buf trace

We introduce the concept to get information about
what process has allocated and used the dmabuf.

The reference is registered and released by the user.
This information is supported by debugfs with pss.

The temporary reference is always registered
by allocation to indicate the user to request allocation,
and this reference is deleted when user requests
a reference explicitly.

The reference ensure to be released when the buffer
is removed or the user process of handle is exited.

Change-Id: I4c147ad91058cf04b81f51753cbefcc420ec8389
Signed-off-by: hyesoo.yu <hyesoo.yu@samsung.com>
6 months agoandroid: ion: show heaptype in the buffer list
Cho KyongHo [Wed, 11 Jul 2018 07:37:26 +0000 (16:37 +0900)]
android: ion: show heaptype in the buffer list

Heap name tells the purpose of heaps and heap type shows how heaps
really work. Showing heap types in the list of buffers helps
identifying the property of buffers more accurately.

Change-Id: I77c97a589e199797ba1898dc696b4d369ea3538a
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agodma-buf: include exp_name to anon file name
Cho KyongHo [Tue, 3 Apr 2018 11:33:37 +0000 (20:33 +0900)]
dma-buf: include exp_name to anon file name

exp_name is introduced for debugging purpose. And we often list the
open file descriptors under /proc/<pid>/fd to find the list of anon
files of dma-buf optained by the process of <pid>. However it is not
possible to associate a specific open file descriptor listed in
/proc/<pid>/fd with the buffer list scrapped from
debugfs/dma-buf/bufinfo.

Attaching exp_name to the anon file name of dma-buf helps identifying
the exported module of the the buffers listed in /proc/<pid>/fd.

dma_buf.exp_name also has the copy of the string instead of pointing
where the string given by the exporter. This gives the exporter chances
to set different exp_name to the different buffers. Therefore we can
identify exact buffers from /proc/<pid>/fd and find the properties of
the buffers. The properties are defined by the exporter and never
exported elsewhere.

Change-Id: Ia3222d1c90b40739cf2e19805a11f366a3b73089
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agoandroid: ion: add buffer id
Cho KyongHo [Fri, 23 Mar 2018 08:33:16 +0000 (17:33 +0900)]
android: ion: add buffer id

Now buffer has its unique id within 0 and 2048. It will be used to
identify a specific ION buffer exported to dma-buf.

Change-Id: I69bc91207bef14a4711cb9c1b02b3c20caf5cd91
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agoandroid: ion: attach buffer id to dmabuf.exp_name
Cho KyongHo [Tue, 3 Apr 2018 11:55:27 +0000 (20:55 +0900)]
android: ion: attach buffer id to dmabuf.exp_name

dmabuf.exp_name is introduced for debugging purpose but it is not
capable of distinguishing different buffers from the same exporter.
Attaching unique buffer id after the module name makes the exp_name
unique.

This commit is only applicable after 'commit e6c7e8059376 ("dma-buf:
include exp_name to anon file name")' that copies exp_name provided by
the exporter instead of pointing it.

Change-Id: Idab0249e1e51007d62c15a8b0dbbb8bec03eaf75
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[9610] soc: samsung: cal-if: Corrected id_tbl bit field.
Jeonghoon Jang [Mon, 20 Aug 2018 00:38:25 +0000 (09:38 +0900)]
[9610] soc: samsung: cal-if: Corrected id_tbl bit field.

Change-Id: I4f5ee21594cff56939ec0a2bdc01f14b2819afbf
Signed-off-by: Jeonghoon Jang <jnghn.jang@samsung.com>
6 months ago[COMMON][HACK] fimc-is2: start sfr dump when hardware timeout at ISP
dan huh [Wed, 22 Aug 2018 10:18:33 +0000 (19:18 +0900)]
[COMMON][HACK] fimc-is2: start sfr dump when hardware timeout at ISP

PR JIRA ID: CPR-206

Change-Id: I2e3d229548af6303d0c627a5bba2c9715b3b5251
Signed-off-by: dan huh <dan.huh@samsung.com>
6 months agomm/gup: fix condition to check need migration
Hyesoo Yu [Tue, 28 Aug 2018 10:59:18 +0000 (19:59 +0900)]
mm/gup: fix condition to check need migration

If POLL_CMA is not set or POLL_GET is not set,
get_user_pages need not to migrate CMA pages.
But current conditional expression, get_user_pages
need not to migration if both POLL_GET and POLL_CMA
is not set.

Change-Id: I243c8b03ae8099027559567231f09c7fb8fede51
Signed-off-by: Hyesoo Yu <hyesoo.yu@samsung.com>
6 months ago[9610] soc: samsung: epx: support dynamic loading framework
Sangkyu Kim [Tue, 27 Mar 2018 08:26:37 +0000 (17:26 +0900)]
[9610] soc: samsung: epx: support dynamic loading framework

Change-Id: Ibc7ed92b8b668f0ade0460cf73c70397a9deff95
Signed-off-by: Sangkyu Kim <skwith.kim@samsung.com>
6 months ago[9610] arm64: dts: insert epx attribute
Sangkyu Kim [Fri, 27 Jul 2018 10:25:54 +0000 (19:25 +0900)]
[9610] arm64: dts: insert epx attribute

Change-Id: Ie3387faa72604f72bf46abfb715f238ebd4aeef1
Signed-off-by: Sangkyu Kim <skwith.kim@samsung.com>
6 months ago[9610] arm64: dts: add arm pmu event
Sangkyu Kim [Fri, 24 Aug 2018 12:29:10 +0000 (21:29 +0900)]
[9610] arm64: dts: add arm pmu event

Change-Id: I1da12d0e55b196a14687c4795b6f71b73a3c9caa
Signed-off-by: Sangkyu Kim <skwith.kim@samsung.com>
6 months ago[9610] arm64: modifed cpu-mif table for performance optimization
Sangkyu Kim [Thu, 23 Aug 2018 12:15:39 +0000 (21:15 +0900)]
[9610] arm64: modifed cpu-mif table for performance optimization

Change-Id: I1ad678fcf9a50fb56db70313e7a96d4ef1702b7b
Signed-off-by: Sangkyu Kim <skwith.kim@samsung.com>
6 months ago[RAMEN9610-4063]GPU_MAX_CLOCK_LIMITE is set from ECT.
Yoojin Park [Mon, 27 Aug 2018 12:16:52 +0000 (21:16 +0900)]
[RAMEN9610-4063]GPU_MAX_CLOCK_LIMITE is set from ECT.

GPU_MAX_CLOCK_LIMITE is fixed value originally.
it makes mis-matching issue to ECT. So, Codes are changed to set after reading ECT.

Change-Id: I6b0593da4f7949e506851c8449a26b035d9fb9b3
Signed-off-by: Yoojin Park <yoojin1.park@samsung.com>
6 months ago[9610] wlbt: Fix for tcp option error
dujeong.lee [Mon, 27 Aug 2018 05:28:18 +0000 (14:28 +0900)]
[9610] wlbt: Fix for tcp option error

Change-Id: Ibef06ff802017d22462c8a49f66fa93e189c26d1
Signed-off-by: dujeong.lee <dujeong.lee@samsung.com>
6 months ago[9610] wlbt: Fix skb pointer exception
Yuseok Kim [Mon, 27 Aug 2018 07:33:25 +0000 (16:33 +0900)]
[9610] wlbt: Fix skb pointer exception

In order to prevent access to skb->len after mem allocation has been released,
tx_bytes data is stored before slsi_mlme_send_frame_data()i is called.

Change-Id: Ib7e32959c921b843180c7c6563ef57d7105f48b6

6 months ago[9610] : ccic : add rid initiation
Junhan Bae [Tue, 31 Jul 2018 07:29:08 +0000 (16:29 +0900)]
[9610] : ccic : add rid initiation

Change-Id: Ie55ee71d5a87003afe9227f42ce06bb1e7f18849
Signed-off-by: Junhan Bae <junhan84.bae@samsung.com>
6 months ago[9610] muic: fix noti
Kim Taejeong [Tue, 24 Jul 2018 10:58:10 +0000 (19:58 +0900)]
[9610] muic: fix noti

Change-Id: Id4aecd1f466b351405ff52927ab90eb0249e72fc
Signed-off-by: Kim Taejeong <tj.kim@samsung.com>
6 months ago[9610] arm64: dts: Add UART switch for CP
Jiyoung Jeong [Tue, 24 Jul 2018 12:39:04 +0000 (21:39 +0900)]
[9610] arm64: dts: Add UART switch for CP

Add UART switch for CP in modem-ss360ap-sit-pdata.dtsi

Change-Id: Id20d6e4e29fce004e383509969e6447fed58f96c
Signed-off-by: Jiyoung Jeong <ji_0.jeong@samsung.com>
6 months ago[9610] drivers: modem_if: Add UART switch for CP
Jiyoung Jeong [Wed, 25 Jul 2018 01:46:36 +0000 (10:46 +0900)]
[9610] drivers: modem_if: Add UART switch for CP

Add UART switch file in modem_if drivers

Change-Id: I28c4f5041fa6b6f38f1368571a6cfcb54fee1bb0
Signed-off-by: Jiyoung Jeong <ji_0.jeong@samsung.com>
6 months ago[9610] drivers: muic: Support CP UART
Jiyoung Jeong [Wed, 25 Jul 2018 01:51:01 +0000 (10:51 +0900)]
[9610] drivers: muic: Support CP UART

Change-Id: I3ecd6d0cfd6d52ed6db78f4f5d04323304b4c95d
Signed-off-by: Jiyoung Jeong <ji_0.jeong@samsung.com>
6 months agoRevert "[HACK][COMMON] scsi: assume device cache off"
JaeHun Jung [Thu, 23 Aug 2018 08:08:09 +0000 (17:08 +0900)]
Revert "[HACK][COMMON] scsi: assume device cache off"

This reverts commit 451871a9ea3d65d360d93ecdaf24eb71dfa9b4f2.

Change-Id: I1c25c7b679b7bbaf9d716379a0ed0d2572c0a56f
Signed-off-by: JaeHun Jung <jh0801.jung@samsung.com>
6 months ago[9610] wlbt: fix prevent issue
Youngsoo Kim [Mon, 27 Aug 2018 09:29:16 +0000 (18:29 +0900)]
[9610] wlbt: fix prevent issue

CID 237209 Uninitialized pointer read (UNINIT)
35. uninit_use_in_call: Using uninitialized value ht_operation_ie when calling slsi_modify_ies_on_channel_switch.

Change-Id: Id1913c5a9b92e4ec77c7ecac4ea21e1d82a0633c

6 months ago[9610] arm64: configs: add USB audio configuration
Kisang Lee [Thu, 23 Aug 2018 08:22:13 +0000 (17:22 +0900)]
[9610] arm64: configs: add USB audio configuration

Change-Id: I590cc0805be8bd562ed7c94565dfb869985a4eae
Signed-off-by: Kisang Lee <kisang80.lee@samsung.com>
6 months agogud: Enable FOLL_CMA
DongHyun Cha [Tue, 21 Aug 2018 04:00:02 +0000 (13:00 +0900)]
gud: Enable FOLL_CMA

Change-Id: I5d2b49d09d0e6cf7964951554f31dd7dd1bb8ae9
Signed-off-by: DongHyun Cha <dhyun.cha@samsung.com>
6 months agomm: hpa: fix killing low oom_adj tasks
Cho KyongHo [Fri, 6 Apr 2018 14:07:38 +0000 (23:07 +0900)]
mm: hpa: fix killing low oom_adj tasks

oom_skip_task() returns true if the give task should not be killed.
But hpa_killer() skips the given task if oom_skip_task() returns false.

Change-Id: If7d829e7056021fe63bae99f735377b685578db6
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: hpa: fix reference to NULL task_struct
Cho KyongHo [Fri, 6 Apr 2018 12:05:25 +0000 (21:05 +0900)]
mm: hpa: fix reference to NULL task_struct

hpa_killer() finds the task with the largest oom_adj_score. It
repeatedly holds and releases references to the tasks during traversing
the list of tasks.
But it forgot the fact that selected that holds the pointer to the
task with the largest oom_adj_score for now is initialized by NULL.

Change-Id: Id374a8d5e2b1bbefc797a314f4eb36783258bbb1
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm, compaction: remove ALLOC_CMA from __compaction_suitable()
Cho KyongHo [Mon, 17 Apr 2017 05:59:30 +0000 (14:59 +0900)]
mm, compaction: remove ALLOC_CMA from __compaction_suitable()

The following patch forces ALLOC_CMA to __zone_watermark_ok() in
__compaction_suitable() because the migration target free page can
be allocated from the free list of MIGRATE_CMA:
    984fdba6a32e mm, compaction: use proper alloc_flags in __compaction_suitable()

However, it is not true from the following patch:
    bf72522dabeb mm: compaction: exclude cma pages from compaction

Therefore, force 0 to __zone_watermark_ok() and revert the alloc_flag
in __isolate_free_page().

Change-Id: I95b3ac6e705f5546b93a8b5c4ceabeaa73d5642d
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: compaction: exclude cma pages from compaction
Cho KyongHo [Mon, 20 Feb 2017 08:26:19 +0000 (17:26 +0900)]
mm: compaction: exclude cma pages from compaction

The following commits avoid allocating CMA pages for allocation
requests without GFP_HIGHUSER_MOVABLE.

commit 8e980f429 ("mm: restrict cma pages to GFP_HIGHUSER_MOVABLE")
commit 61019deeb ("mm: allocate CMA pages first for GFP_HIGHUSER_MOVABLE")
commit 9db6a5edc ("mm: do not alloc cma page for shmem")

However one exception is found that a CMA page is allocated without
GFP_HIGHUSER_MOVABLE: compaction. The freepage scanner of compaction
gathers free pages to copy the contents of migration target pages.
It gathers CMA pages to migrate movable pages which might be allocated
without GFP_HIGHUSER_MOVABLE.

To avoid pinning CMA pages unexpectedly, CMA pages should be excluded
from the freescanner of the compaction. Likewise CMA pages are also
excluded from the migration target page scanner because CMA pages are
not helpful for the purpose of the compaction. CMA page blocks are not
allocated for either of THP or unmovable pages.

Change-Id: Ic4524f650eea19cfa8e2f197653c465dd5f0db97
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: restrict cma pages to GFP_HIGHUSER_MOVABLE
Cho KyongHo [Thu, 5 Jan 2017 12:25:29 +0000 (21:25 +0900)]
mm: restrict cma pages to GFP_HIGHUSER_MOVABLE

pcp list of MIGRATE_MOVABLE can have CMA pages and the free pages in
the list are ready for allocation with __GFP_MOVABLE. Since CMA pages
should be allocated only if GFP_HIGHUSER_MOVABLE is specified, it is
required to verify if the selected page in the pcp list is proper for
allocation according to gfp mask.

Change-Id: I5c74d74edc6adce293a3db59abdbb1acf4a01c14
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: allocate CMA pages first for GFP_HIGHUSER_MOVABLE
Cho KyongHo [Thu, 5 Jan 2017 01:23:07 +0000 (10:23 +0900)]
mm: allocate CMA pages first for GFP_HIGHUSER_MOVABLE

CMA pages are not actively used and tend to be remained unused because
the free list of CMA pages are only lookuped as the fallback of
movable page allocations. Large amount of free CMA pages causes
problems about unmovable page allocation failure as well as memory
unefficiency.
If free CMA pages are more than the low wmark, page reclamation does
not work for unmovable page allocations. This results in depletion of
free pages available for unmovable page allocations.
To prevent this impact by free CMA pages, we should always try to
consume all free CMA pages.

Change-Id: Idc51f0327f6706d27dafbbb69475bf53eb0153eb
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: do not alloc cma page for shmem
Cho KyongHo [Wed, 21 Dec 2016 02:17:20 +0000 (11:17 +0900)]
mm: do not alloc cma page for shmem

Some page caches of shmem tends to be pinned for a long time.
Therefore shmem should not allocate page caches from CMA because cma
pages should be always ready to be reclaimed promptly.

Change-Id: I024dad765feec07b0028851b0b98b7db7f3c85a5
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: show page state if isolation is failed
Cho KyongHo [Fri, 9 Jun 2017 13:32:44 +0000 (22:32 +0900)]
mm: show page state if isolation is failed

alloc_contig_range() shows the following message to the kernel log if
isolation of the given page ranges has been failed:
"__alloc_contig_range: [8fcd00, 900000) PFNs busy"

But the message provides no information but isolation of some pages in
that range is failed.
If isolation of specific pages is failed repeatedly, we should find
the reason. Identifying the state of pages that are not isolated is
very important to start solving the failure.

Change-Id: I83fa34f9c49989076e1ba20f5b92c570bdd180c0
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: show page state only for cma
Cho KyongHo [Tue, 31 Oct 2017 12:03:37 +0000 (21:03 +0900)]
mm: show page state only for cma

migrate_pages() is not only called by cma but also by compaction and
memory hotplug-out.
"fbca0939fc96 [COMMON] mm: show page state on migration failure"
inserts printing simple states of the pages that are failed in
migration to another pages. Migration for cma and memory hotplug-out
seldom fails but that of compaction does not since compaction tries
to migrate all anonymous pages. Therefore the state of the unmigrated
pages is not necessary for compaction. Rather it may suffer from too
many unnecessary mirgration failre log messages.

Change-Id: Ica55bbcacc94a7b16dc76410b404852a98bc482a
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm: show page state on migration failure
Cho KyongHo [Wed, 20 Sep 2017 07:54:55 +0000 (16:54 +0900)]
mm: show page state on migration failure

It is very useful to find the reason of the migration failure if we
know what the page is not migrated and the state of the page.

Change-Id: If6def51cbe8ac5b48a49c860a925633a36b498a2
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months agomm/gup: check FOLL_CMA eariler than others
Cho KyongHo [Mon, 20 Aug 2018 12:43:19 +0000 (21:43 +0900)]
mm/gup: check FOLL_CMA eariler than others

__need_migrate_cma_page() checks a lot of conditions if the current
get_user_pages() request need to migrate CMA pages to some other pages.
But FOLL_CMA is given by very few users. We don't need to spend time
for condition check for the users without FOLL_CMA.

Change-Id: I973b3c99140e71fbf78cb6280e94ef5e97c9824a
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[COMMON] mm: gup: add local pagevec flush
Cho KyongHo [Tue, 18 Jul 2017 01:42:28 +0000 (10:42 +0900)]
[COMMON] mm: gup: add local pagevec flush

Change-Id: Ic6a899001fcea0f0df64f3c07bb67f998144eb31
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[COMMON] mm: gup: drain pagevec if FOLL_CMA is set
Cho KyongHo [Fri, 9 Jun 2017 13:39:22 +0000 (22:39 +0900)]
[COMMON] mm: gup: drain pagevec if FOLL_CMA is set

Some users of get_user_pages() give FOLL_CMA not to pin CMA pages.
However a CMA page may not be migrated even though FOLL_CMA is set
because it is not in LRU but in per-cpu pagevec.
The previous approach to solve this problem is the invocation of
migrate_prep_local() that drains the pagevec of the local CPU. It was
the only choice in if it is called in __need_migrate_cma_page(). It is
because __need_migrate_cma_page() is called under a spinlock of page
table is held. This approach unfortunately does not drain pagevecs of
other CPUs and makes CMA fail to allocate physically contiguous memory
at last.
We need a different approach to get over the problem: clearing pagevec
of all CPUs if FOLL_CMA is set unconditionally. It degrades the
performance of get_user_pages() with FOLL_CMA if no CMA page is found
in the given range. But it is not a problem because there is very few
user that gives FOLL_CMA to get_user_pages().

Change-Id: Idcd651d5e2bc7ec66d17b3296e742b7e714f4b48
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[COMMON] mm: gup: fix invalid comparison for CMA
Cho KyongHo [Tue, 27 Dec 2016 11:33:02 +0000 (20:33 +0900)]
[COMMON] mm: gup: fix invalid comparison for CMA

CMA pages should not be pinned for a long time because they should be
always ready for reclamation by the user's request. If a user of gup
wants to pin a series of pages for a long time he/she should give
FOLL_CMA to let gup migrate the CMA pages that are in pinning to
non-CMA pages.

In order to determine if a pinning page is in a CMA page, gpu should
compare if the migrate type of the pageblock of the page is
MIGRATE_CMA. But the comparison is not correct and the migration
never happens.

SVACE #82317

Change-Id: Id4ef40957365bdbe586f29c4ba0e6318e232719a
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[COMMON] mm: fix cma page migration race condition
Sunyoung Kang [Tue, 14 Jun 2016 06:09:20 +0000 (15:09 +0900)]
[COMMON] mm: fix cma page migration race condition

The below callstack makes VM_BUG_ON in race condition.
So this fixes it by moving the page isolation into before
pte_unmap_unlock().

----------------------------------------------------------
CPU0 CPU1
__need_migrate_cma_page
   PageLRU(page)
pte_unmap_unlock
isolate_migrate_pages_range
migrate_pages # clear pageLRU
# _count == 0
__migrate_cma_pinpage
   __isolate_lru_page # failure on !PageLRU
get_page_foll # VM_BUG on page->_count == 0
----------------------------------------------------------

This patch effects that get_user_pages() may stop on a page that is
under a racing but it guarantees that the state of the page is
consistent. Therefore, the caller of get_user_pages() should always
check if the pinning all required pages is successful.

Change-Id: I7947db746e8aaed23ea7344ee0483073bbd2d51c
Signed-off-by: Sunyoung Kang <sy0816.kang@samsung.com>
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[COMMON] mm: cma page migration for pinned pages
Jinsung Yang [Tue, 3 Mar 2015 05:24:46 +0000 (14:24 +0900)]
[COMMON] mm: cma page migration for pinned pages

If the changes in the patch 'cma: redirect page allocation to CMA' were
applied, many user pages could be allocated from cma page block. However,
this can cause migration failures due to pinned pages by __get_user_pages().
This patch adds to support cma page migration at page pinning time if gup_flags
has FOLL_CMA.

Change-Id: I70389f1ddee9697af653c4a50eab161925113979
Signed-off-by: Jinsung Yang <jsgood.yang@samsung.com>
Signed-off-by: Cho KyongHo <pullip.cho@samsung.com>
6 months ago[9610] dtsi: Modfiy a dvfs level for capture
dan huh [Tue, 21 Aug 2018 06:46:15 +0000 (15:46 +0900)]
[9610] dtsi: Modfiy a dvfs level for capture

PR JIRA ID: CPR-193

Change-Id: If819dc7f8df17a474a75978ee5cec80a97c4b1cb
Signed-off-by: dan huh <dan.huh@samsung.com>
6 months ago[9610] wlbt : disable log_collection
Youngsoo Kim [Wed, 22 Aug 2018 16:24:28 +0000 (01:24 +0900)]
[9610] wlbt : disable log_collection

temporary disable log_collection cause it make delay.

Change-Id: Icd410801dcd58ef80c8ea1212352696d5770f62f

6 months ago[COMMON] cpufreq: remove condition performance governor
Sangkyu Kim [Wed, 22 Aug 2018 07:32:17 +0000 (16:32 +0900)]
[COMMON] cpufreq: remove condition performance governor

Change-Id: I993d99194c8459f48ff286f842d1aa82826bf283
Signed-off-by: Sangkyu Kim <skwith.kim@samsung.com>
6 months ago[9610][ERD] dtsi: remove sensor pin interrupt
Boojin Kim [Tue, 21 Aug 2018 12:20:41 +0000 (21:20 +0900)]
[9610][ERD] dtsi: remove sensor pin interrupt

Change-Id: Ic013fa415f56f8bc06471c56d7c7cc13a5f1f9c7
Signed-off-by: Boojin Kim <boojin.kim@samsung.com>
6 months agoRevert "[9610] arm64: dts: add gpio pins"
Boojin Kim [Tue, 21 Aug 2018 12:19:37 +0000 (21:19 +0900)]
Revert "[9610] arm64: dts: add gpio pins"

This reverts commit b91347f9ed63cf8ce6609a4cca8d47a14eebdd5c.

Change-Id: Iade8c85023b36e47502757dec4329c75ab98f5d4
Signed-off-by: Boojin Kim <boojin.kim@samsung.com>
6 months ago[9610] arm64: configs: modify configs for user & userdebug
Youngwan Kim [Thu, 16 Aug 2018 08:23:52 +0000 (17:23 +0900)]
[9610] arm64: configs: modify configs for user & userdebug

for fixing warning pop-up in platform and VTS fail

TODO: add same config to the eng build

Change-Id: I68acfe1b57a9f85183bf82ac5669d2b6f3598504
Signed-off-by: Youngwan Kim <y103.kim@samsung.com>
6 months ago[9610] fimc-is2: Fix a SCAR error at DVFS file
dan huh [Thu, 16 Aug 2018 23:33:58 +0000 (08:33 +0900)]
[9610] fimc-is2: Fix a SCAR error at DVFS file

PR JIRA ID: CPR-171

Change-Id: Id094c678808b63771d8d46d61a3269a4498f4f5c
Signed-off-by: dan huh <dan.huh@samsung.com>
6 months ago[9610] fimc-is2: Modify a csis-3aa mux setting for dual, pip
dan huh [Thu, 16 Aug 2018 05:16:40 +0000 (14:16 +0900)]
[9610] fimc-is2: Modify a csis-3aa mux setting for dual, pip

PR JIRA ID: CPR-171

Change-Id: I0fe6389c11cf9664e0e4fb2b9319eeae8f83923c
Signed-off-by: dan huh <dan.huh@samsung.com>
6 months ago[9610] fimc-is2: Add a PIP scenario for detect
dan huh [Thu, 16 Aug 2018 05:15:17 +0000 (14:15 +0900)]
[9610] fimc-is2: Add a PIP scenario for detect

PR JIRA ID: CPR-173

Change-Id: I19b3d246d045258af8b94b5a2cb75a67c04a6533
Signed-off-by: dan huh <dan.huh@samsung.com>
6 months ago[9610] dtsi: Add a PIP scenario DVFS level
dan huh [Thu, 16 Aug 2018 05:13:22 +0000 (14:13 +0900)]
[9610] dtsi: Add a PIP scenario DVFS level

PR JIRA ID: CPR-173

Change-Id: Ibb1848f4b322faf85e60dc2795fa5fa6d4e94015
Signed-off-by: dan huh <dan.huh@samsung.com>
6 months ago[COMMON] fimc-is: fix a bug TDNR mode setting at each other instance
Wooyeon Kim [Tue, 14 Aug 2018 07:50:22 +0000 (16:50 +0900)]
[COMMON] fimc-is: fix a bug TDNR mode setting at each other instance

 - if device is used a multi-instance, TDNR mode should set 2DNR mode
   as other instance is operated

 PR JIRA ID: CPR-166

Change-Id: I25254bcf4a2579cdbd4b16609a3257d42fa6392a
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[9610][ERD] arm64: config: enable CONFIG_ANDROID_LOW_MEMORY_KILLER
Youngmin Nam [Thu, 16 Aug 2018 11:57:13 +0000 (20:57 +0900)]
[9610][ERD] arm64: config: enable CONFIG_ANDROID_LOW_MEMORY_KILLER

Change-Id: I69782fc86326bafdcf422b33730c24bb671cdd02
Signed-off-by: Youngmin Nam <youngmin.nam@samsung.com>
6 months agoRevert "staging, android: remove lowmemory killer from the tree"
Donghoon Yu [Wed, 4 Jul 2018 03:39:34 +0000 (12:39 +0900)]
Revert "staging, android: remove lowmemory killer from the tree"

This reverts commit 915e70f9263d56fbf103742265025f7a492aa625.

Change-Id: I358f22b6f059cbfbfca64726a0da6979202fb8fe

6 months ago[9610] wlbt: increment scsc_release to 6.44.0
scsc_gerrit [Thu, 16 Aug 2018 12:29:02 +0000 (13:29 +0100)]
[9610] wlbt: increment scsc_release to 6.44.0

Import FW for iteration 6.44.0

Change-Id: I3e5d95c15835ad98af4c22225767776b1a110bc9
SCSC-Bug-Id: scsc_iteration_6.44.0_p_ww
Signed-off-by: scsc_gerrit <scsc_gerrit@samsung.com>
6 months ago[9610] FAPI update for TEST SAP
Pragya Gupta [Tue, 7 Aug 2018 09:55:22 +0000 (15:25 +0530)]
[9610] FAPI update for TEST SAP

FAPI update for TEST SAP

Change-Id: I17079aa3c473f2411f2031f21794f026c44281d6
SCSC-Bug-Id:SSB-42581
Signed-off-by: Pragya Gupta <pragya.gupta@samsung.com>
6 months ago[9610][7885][7872] wlbt: P2P Scan cleanup on intf down
Jayaprakash Sangaru [Thu, 2 Aug 2018 12:20:04 +0000 (17:50 +0530)]
[9610][7885][7872] wlbt: P2P Scan cleanup on intf down

Send mlme-del-scan.req when interface goes down for p2p-group interface

SCSC-Bug-Id: SSB-42480

Change-Id: I496c00b5872aaebfe860a3fa432cb5cad78c304a
Signed-off-by: Jayaprakash Sangaru <j.sangaru@samsung.com>
6 months ago[9610] wlbt: SMAPPER: add defence to access SKB
Debabrata Purohit [Fri, 3 Aug 2018 15:04:54 +0000 (16:04 +0100)]
[9610] wlbt: SMAPPER: add defence to access SKB

hip4_smapper_get_skb_data() dereferences pointer
without checking it's valid or not.
Add defence for it.

Change-Id: I4933202ac72fd1ae0d5476b43cd183f5394488e3
SCSC-Bug-Id: SSB-42553
Signed-off-by: Debabrata Purohit <d.purohit@samsung.com>
6 months ago[9610] wlbt: fw panic copy files to /sdcard/log
Purnendu Kapadia [Tue, 14 Aug 2018 14:48:27 +0000 (15:48 +0100)]
[9610] wlbt: fw panic copy files to /sdcard/log

on fw panic we want to generate tar file via wlbtd. We were calling
EVENT_SCSC to create moredump in /sdcard/log. Now wlbtd takes care of this
by copying moredump and other logs in /sdcard/log (temporarily) until
test systems can cope with new sable and tar file.
Removing call to EVENT_SCSC as it is redundant.

Change-Id: I9f12379a077a3e5d751200504c8eeb2b1984f4e9
SCSC-Bug-Id: SSB-42639
Signed-off-by: Purnendu Kapadia <p.kapadia@samsung.com>
6 months ago[9610] wlbt: default path where tar is generated
Purnendu Kapadia [Mon, 13 Aug 2018 17:10:01 +0000 (18:10 +0100)]
[9610] wlbt: default path where tar is generated

Android P expects things in /data/vendor so use /data/vendor/log/wifi as
default path where sable and tar containing sable will be generated

Change-Id: Iadcd02df5d5c71731a0bb92f04b254e9489b9dc1
SCSC-Bug-Id: SSB-42639
Signed-off-by: Purnendu Kapadia <p.kapadia@samsung.com>
6 months ago[9610] wlbt: Add support for DRAM sbl dump and mmap
Albert Cano [Fri, 3 Aug 2018 10:46:56 +0000 (11:46 +0100)]
[9610] wlbt: Add support for DRAM sbl dump and mmap

Since Android P does not allow writting files from the Kernel, the patch
provides a mechanism to buffer the sbl file in DRAM (vmalloc) and allow mmap ops
so the wlbt could dump the file from Userland

Change-Id: I46348bf45b65a851e9878e8c65678ded0f420611
SCSC-Bug-Id: SSB-42521
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
(cherry picked from commit e29a986dac1a22849026aeb0e88b89ba07f3270e)

6 months ago[7570][7872][7885][9610]wlbt: SGI40 should not be enabled in 2.4G beacon
Srishti Piplani [Fri, 27 Jul 2018 12:42:15 +0000 (18:12 +0530)]
[7570][7872][7885][9610]wlbt: SGI40 should not be enabled in 2.4G beacon

If the DUT is only 20MHz capable, SGI40 should not be enabled.
Changes done to remove the SGI40 check if the width is 20 MHz.

Change-Id: I38f1b2d5eb9ed850fb8ccec3c460633fbcbedccd
SCSC-Bug-Id: SSB-42343
Signed-off-by: Srishti Piplani <srishti.p@samsung.com>
6 months ago[9610] wlbt: Added modparams to Sable and MXLOGGER
Cristian Marussi [Wed, 25 Jul 2018 19:03:12 +0000 (20:03 +0100)]
[9610] wlbt: Added modparams to Sable and MXLOGGER

Sable could be now disabled as a whole using modparams:
file generation is inhibited and MXLOGGER underlying mechanism
is halted, refraining from sending the needed config to FW
MXLOGGER module. All FW logs will suddendly flow to the Host
and NO Sable file could be generated.

Change-Id: I76b15756ed9ec589bce3b7fc17434719fc9ded4c
SCSC-Bug-ID: SSB-42258
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
6 months ago[9610] wlbt: WLBTD collect log files and tar them
Purnendu Kapadia [Wed, 18 Jul 2018 14:46:58 +0000 (15:46 +0100)]
[9610] wlbt: WLBTD collect log files and tar them

For SBL support, we need wlbtd to collect various logs into a single tarball.
This patch adds the driver side changes to support this. A new EVENT_SABLE
and supporting infrastructure is introduced to handle trigger from log collection
(with reason for generating .sbl file) and EVENT_SABLE sent to wlbtd to generate
other logs (logcat, dmesg, mx.dump etc) and create tar containing .sbl file alongwith other logs.

for generating a tar this change dependens on change to wlbtd with EVENT_SABLE support.

Change-Id: I7a356327ef8cb74d5ff5cb7ce3fd12b880127601
SCSC-Bug-Id: SSB-41551
Signed-off-by: Purnendu Kapadia <p.kapadia@samsung.com>
6 months ago[9610] wlbt: Fix MXLOGGER DRAM allocation
Cristian Marussi [Mon, 23 Jul 2018 08:53:18 +0000 (09:53 +0100)]
[9610] wlbt: Fix MXLOGGER DRAM allocation

MXLOGGER must use new DRAM dedicated area and related API.

Change-Id: I9540667b9c854e2d51b61d778e8167b41b6367a9
SCSC-Bug-Id: SSB-42105
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
6 months ago[9610] wlbt: Shared DRAM memory partitioning
Roman Geller [Tue, 17 Jul 2018 12:18:29 +0000 (13:18 +0100)]
[9610] wlbt: Shared DRAM memory partitioning

Added a second dram partition section for the new debug mechanism
which will store debug info in DRAM buffers

Change-Id: I6a3dad523bff21ec1c242bc8c66a3d775cd66c8f
SCSC-Bug-Id: SSB-41224
Signed-off-by: Roman Geller <r.geller@samsung.com>
6 months ago[9610] wlbt: fix force_disable_mxlogger modparam
Cristian Marussi [Fri, 20 Jul 2018 12:34:42 +0000 (13:34 +0100)]
[9610] wlbt: fix force_disable_mxlogger modparam

Disabling mxlogger using 'force_disable_mxlogger' modparam
is now effective also at system boot-time.

Change-Id: Iccef147570a05b2b92c70c15c7c4a92c067a3ba1
SCSC-Bug-Id: SSB-42080
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
6 months ago[9610] wlbt: Fixed deadlock in mxlogger
Albert Cano [Tue, 17 Jul 2018 12:40:36 +0000 (13:40 +0100)]
[9610] wlbt: Fixed deadlock in mxlogger

The following sequence leads to a deadlock when mxlogger is
deinitialzing and a log collection is triggered:

mxlogger_deinit
---> gets mutex_lock(&mxlogger->lock);
***** preemption *****
__scsc_log_collector_collect_to_file
---> gets mutex_lock(&log_status.log_mutex);
//calls
mxlogger_collect:
---> sitting in already claimed mutex_lock(&mxlogger->lock);
***** preemption *****
mxlogger_deinit
//calls
scsc_log_collector_unregister_client
---> sitting in alreaday claimed mutex_lock(&log_status.log_mutex)
DEADLOCK

To avoid the deadlock condition, the scsc_log_collector_unregister_client
calls are moved outside the mxlogger_deinit lock.

Hip4.c has also been addressed

Client registration / unregistration serialization has also been
addressed

Change-Id: I5a85de46bbd0564c4409ed6049cb9f3a8c987ae1
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months ago[9610] wlbt: Enable log collection by default
Albert Cano [Mon, 16 Jul 2018 20:08:09 +0000 (21:08 +0100)]
[9610] wlbt: Enable log collection by default

Enable log collection by default

Change-Id: I6447d7b7cc562c577437ca986761c82937081cac
SCSC-Bug-Id: SSB-41954
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
6 months ago[9610] wlbt: Increase MMAP max area properly
Cristian Marussi [Tue, 7 Aug 2018 17:47:04 +0000 (18:47 +0100)]
[9610] wlbt: Increase MMAP max area properly

Since we use DRAM partitioned allocation we should make
available for mmapp'ing a wider area.

Change-Id: Ic13d73195adadb9cbe1d193d870ba4b414cb9e01
SCSC-Bug-Id: SSB-42658
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
6 months ago[9610] wlbt: Create and register Wlan sable file clients
Albert Cano [Thu, 12 Jul 2018 18:29:26 +0000 (19:29 +0100)]
[9610] wlbt: Create and register Wlan sable file clients

Create and register wlan sable file client to collect hip4 samples and
hcf a sbl file.

Add UDI as a mxlogger observer

Change-Id: I69d14126920ffc01d07b976847318ebeed2789fb
SCSC-Bug-Id: SSB-41876
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months ago[9610] wlbt: Create and register logring sable file client
Albert Cano [Thu, 12 Jul 2018 18:22:08 +0000 (19:22 +0100)]
[9610] wlbt: Create and register logring sable file client

Create and register logring sable file client to collect logring in the
sbl file.

Change-Id: I98cc68afb49d26644879193b45a2735fce681fca
SCSC-Bug-Id: SSB-41875
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months ago[9610] wlbt: Enable SMAPPER by default
Albert Cano [Wed, 8 Aug 2018 08:36:25 +0000 (09:36 +0100)]
[9610] wlbt: Enable SMAPPER by default

Enable SMAPPER by default.

Fixed dependency by addding dma-mapping header.

Change-Id: I67a3e8304c79243252c77b5953c1cfff23c0774e
SCSC-Bug-Id: SSB-42637
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months agowlbt: Implement mxlogger module
Albert Cano [Thu, 12 Jul 2018 17:20:05 +0000 (18:20 +0100)]
wlbt: Implement mxlogger module

Implement mxlogger module, responsible of:
- Defining and providing logging buffers in Shared
DRAM
- Providing a communication channel from/to FW to
send receive low level logging commands (START,
STOP, SWITCH, FW_TRIGGER...)
- Creating an interface to register/unregister
local or global observers
- Implementing the control mechanism to instruct the
FW to store logs in Shared DRAM or stream logs to
Host

Keep logs streamed to host by default

Change-Id: Ia831573630a07c63fb3420aa28bcddd932296548
SCSC-Bug-Id: SSB-41855
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months agowlbt: Implement log collection Framework
Albert Cano [Thu, 12 Jul 2018 12:26:32 +0000 (13:26 +0100)]
wlbt: Implement log collection Framework

Implemented log collection Framework.

- Interested clients/writers register with the framework
- On event trigger, framework will execute clients callback
- Framework will generate sbl file as described in
http://confluence/display/PLAT/Sable+-+File+Format
- Add entry in proc interface to trigger log collection from
Userland (User manual trigger and dumpsys)

Change-Id: I8a37af7c8aac983df2bb9963e1832d769893d530
SCSC-Bug-Id: SSB-41841
Signed-off-by: Cristian Marussi <c.marussi@samsung.com>
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months ago[9610] wlbt: SMAPPER: fix access to SKB data
Debabrata Purohit [Fri, 3 Aug 2018 09:59:24 +0000 (10:59 +0100)]
[9610] wlbt: SMAPPER: fix access to SKB data

Fix incorrect access to Ethernet header in S-mapper
mode.

Change-Id: I5893340536a1836fa7cbb04ec1e0b284ff8ff210
SCSC-Bug-Id: SSB-42532
Signed-off-by: Debabrata Purohit <d.purohit@samsung.com>
6 months ago[7570][7872][7885][9610] wlbt: Dwell time in mlme-send-frame.req
Srishti Piplani [Fri, 6 Jul 2018 12:22:52 +0000 (17:52 +0530)]
[7570][7872][7885][9610] wlbt: Dwell time in mlme-send-frame.req

For all frames sent via mlme-send-frame_req and that expect an
answer, such as EAPOL (except M4), ARP, DHCP..etc. the host shall
set a dwell-time.The default dwelltime may be equal to
unifiForcedScheduleDuration.

Change-Id: I9601bc780b3d84d14d72bf3f5eba24315c83be0c
SCSC-Bug-Id: SSB-41631
Signed-off-by: Srishti Piplani <srishti.p@samsung.com>
6 months ago[9610][7885][7872] wlbt: Passive scan support
JayaPrakash Sangaru [Tue, 31 Jul 2018 10:12:30 +0000 (15:42 +0530)]
[9610][7885][7872] wlbt: Passive scan support

In some cases supplicant requests passive scan. in these case
scan_request has zero ssids. Support passive scan when ssids
in scan request is zero.

SCSC-Bug-Id: SSB-42405

Change-Id: I18aa04abe56192d5501247032690aaba6c1d20e5
Signed-off-by: JayaPrakash Sangaru <j.sangaru@samsung.com>
6 months ago[9610] [7885] wlbt : Don't set the host_state in wlanlite mode
Pragya Gupta [Fri, 27 Jul 2018 11:26:27 +0000 (16:56 +0530)]
[9610] [7885] wlbt : Don't set the host_state in wlanlite mode

Don't set the host_state for wlanlite mode

SCSC-Bug-Id:SSB-42261

Change-Id: I080dc5ca5bac3266a0760c380216a29ad193824e
Signed-off-by: Pragya Gupta <pragya.gupta@samsung.com>
6 months ago[9610] wlbt: monitor mode - type: Radio tap
Debabrata Purohit [Wed, 27 Jun 2018 13:35:21 +0000 (14:35 +0100)]
[9610] wlbt: monitor mode - type: Radio tap

Change the interface encapsulation type to
Radio tap for monitor interface.

SCSC-Bug-Id: SSB-41477
Change-Id: Iffbb0a404cc99a77c3b116705db29687141008b8
Signed-off-by: Debabrata Purohit <d.purohit@samsung.com>
6 months ago[9610] wlbt: remove tcp_get_info call
Debabrata Purohit [Tue, 19 Jun 2018 15:31:30 +0000 (16:31 +0100)]
[9610] wlbt: remove tcp_get_info call

Remove call to tcp_get_info(). It is called in
software IRQ context and as it may sleep
in kernels later than 4.9 it can cause deadlock.

Change-Id: I7ab1bcae50e98a374caae743bc90a5993028ae25
SCSC-Bug-Id: SSB-41049
Signed-off-by: Debabrata Purohit <d.purohit@samsung.com>
6 months ago[9610] wlbt: Fixed WLAN banks to PHYs banks mapping
Albert Cano [Mon, 30 Jul 2018 16:26:13 +0000 (17:26 +0100)]
[9610] wlbt: Fixed WLAN banks to PHYs banks mapping

Incorrect mapping occurs when WLAN is allocating banks in the WLAN array.

WLAN will use the phy/logical map to configure the SMAPPER FW however
when FW returns entries mapped to phy banks, WLAN is not able to map to
their own logical WLAN banks.

This patch adds a lookup table to map PHY->WLAN banks

Change-Id: Iffaa3484cfe5298dfd0f1a5e0619364e8f27cca6
SCSC-Bug-Id: SSB-42392
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months ago[9610] wlbt: Fixed number of entries computed in small banks
Albert Cano [Mon, 30 Jul 2018 14:34:40 +0000 (15:34 +0100)]
[9610] wlbt: Fixed number of entries computed in small banks

Incorrect indexing when computing the number of
entries in small banks.

Also fixed misleading logging message

Change-Id: I9ebd7e63cc184a361e7fb74f1693173a0471f342
SCSC-Bug-Id: SSB-42388
Signed-off-by: Albert Cano <a.canocamps@samsung.com>
6 months ago[9610][7885][7872] wlbt: Rename wifi-hal return values enum
Jayaprakash Sangaru [Wed, 25 Jul 2018 09:33:19 +0000 (15:03 +0530)]
[9610][7885][7872] wlbt: Rename wifi-hal return values enum

wifi-hal return vaalues enum is defined in enchanced logging and
in NAN which is creating some issues. Rename these enum params in
NAN.

SCSC-Bug-Id: SSB-42223

Change-Id: I1afe03e412a03d1e39db8df96a94f510f8ac9e2c
Signed-off-by: Jayaprakash Sangaru <j.sangaru@samsung.com>
6 months ago[HACK] Make kernel panic when verity error is detected for debugging
Hyeyeon Chung [Thu, 16 Aug 2018 10:16:26 +0000 (19:16 +0900)]
[HACK] Make kernel panic when verity error is detected for debugging

Change-Id: If35e3c21cdbe8fe314dc4f1df878f31a3e0e3924
Signed-off-by: Hyeyeon Chung <hyeon.chung@samsung.com>
6 months ago[IMS-WFC] Enable Kernel Config for ePDG Client(Charon)
jihun.no [Wed, 1 Aug 2018 03:39:09 +0000 (12:39 +0900)]
[IMS-WFC] Enable Kernel Config for ePDG Client(Charon)

Change-Id: I9a95183035dd450f7bc63f585f6de28a9de9c9fe
Signed-off-by: jihun.no <jihun.no@samsung.com>
6 months ago[COMMON] fimc-is2: add code related to initial AE for 6B2
Wooyeon Kim [Tue, 7 Aug 2018 08:58:25 +0000 (17:58 +0900)]
[COMMON] fimc-is2: add code related to initial AE for 6B2

 PR JIRA ID: CPR-134

Change-Id: I425d87e072fc3488d8ff123d63dd568c6febe6f9
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[COMMON] fimc-is2: add code related to initial AE for 2T7SX
Wooyeon Kim [Tue, 7 Aug 2018 08:54:31 +0000 (17:54 +0900)]
[COMMON] fimc-is2: add code related to initial AE for 2T7SX

 PR JIRA ID: CPR-134

Change-Id: Id77783ad9d32d165f3b613bfa07fc8e6aab34d30
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[COMMON] fimc-is2: add code related to initial AE for 2P7SQ
Wooyeon Kim [Tue, 7 Aug 2018 08:59:20 +0000 (17:59 +0900)]
[COMMON] fimc-is2: add code related to initial AE for 2P7SQ

 PR JIRA ID: CPR-134

Change-Id: Ib4c2a616f31a547ba1ade96e95c0041e32fcb6ac
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[9610][ERD] dts: camera: enable initial AE feature
Wooyeon Kim [Tue, 7 Aug 2018 11:37:34 +0000 (20:37 +0900)]
[9610][ERD] dts: camera: enable initial AE feature

 PR JIRA ID: CPR-134

Change-Id: Iaae228abc2ddca6365a8fa5b7f5dc12ef97dc409
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[COMMON] fimc-is2: modify initial AE code for commonize
mourui [Wed, 7 Mar 2018 09:26:04 +0000 (17:26 +0800)]
[COMMON] fimc-is2: modify initial AE code for commonize

 -  apply last AE value to next initial value

 PR JIRA ID: CPR-134

Change-Id: I86edcd79044df658e7b1c28b036224628919fb65
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[COMMON] fimc-is2: fix a build error
Wooyeon Kim [Mon, 13 Aug 2018 00:46:07 +0000 (09:46 +0900)]
[COMMON] fimc-is2: fix a build error

 PR JIRA ID: CPR-153

Change-Id: I131e32750c14f1d5f9dad8708ca7743d9a9d8261
Signed-off-by: Wooyeon Kim <wooy88.kim@samsung.com>
6 months ago[9610] fimc-is2: move structure for capture node address info
Wooki Min [Tue, 24 Jul 2018 10:56:42 +0000 (19:56 +0900)]
[9610] fimc-is2: move structure for capture node address info

The structure for capture node address info
is moved from meta data to framemgr.

 PR JIRA ID: CPR-153

Change-Id: I293a1a258224388f33e192b52025352f5458be6d
Signed-off-by: Wooki Min <wooki.min@samsung.com>
6 months ago[COMMON] fimc-is2: move structure for capture node address info
Wooki Min [Tue, 24 Jul 2018 08:33:04 +0000 (17:33 +0900)]
[COMMON] fimc-is2: move structure for capture node address info

The structure for capture node address info
is moved from meta data to framemgr.

 PR JIRA ID: CPR-153

Change-Id: I15353f17ff267263792db36e737226533838d761
Signed-off-by: Wooki Min <wooki.min@samsung.com>