PD#SWPL-148670
Problem:
https://scgit.amlogic.com/#/c/375286/
the patch closes some trace events nodes.it will lead to cts fail.
Solution:
set allow_trace_enable to 0 for opening the trace event nodes for cts
test
Verify:
local.
Change-Id: I175a72c123e3bf71926bd08ac61427fdf7c0c3e6
Signed-off-by: Dezhen Wang <dezhen.wang@amlogic.com>
PD#SWPL-147985
Problem:
little-high-prio interactive task schedule latency maybe very big
when task group schedule enable
Solution:
increase sched_task_high_prio to optimize schedule latency
Verify:
sc2
Change-Id: I9c0448e48c391432fc7d31805970fcbaceb6bb80
Signed-off-by: Hanjie Lin <hanjie.lin@amlogic.com>
PD#SWPL-146932
Problem:
power led off when device is running
Solution:
set this gpio when device is running
Verify:
sc2 ah212
Change-Id: Ib31e3ff674c061e560f8bb1035250e57bb9ea332
Signed-off-by: Junyi Zhao <junyi.zhao@amlogic.com>
Signed-off-by: Luan Yuan <luan.yuan@amlogic.com>
PD#SWPL-146713
Problem:
data read from udisk on otg usb port is error
Solution:
set u2drd wr_outstanding as 0
Verify:
be311
Change-Id: I12a5e155f73526f31afd0f4ae48b9ee031fe2450
Signed-off-by: qi duan <qi.duan@amlogic.com>
PD#SWPL-148278
Problem:
CONFIG_AMLOGIC_ZSTD is to upgrate the zstd version in old kernel5.15.
Now we will not maintain the zstd function in this kernel version.
Solution:
set CONFIG_AMLOGIC_ZSTD to n
Verify:
local
Change-Id: I49fe3baa5e1b13d42a06949cb8711fde9ebcdd04
Signed-off-by: Dezhen Wang <dezhen.wang@amlogic.com>
PD#SWPL-148281
Problem:
with EDID Y420 cmdb which payload len = 1,
it will not store the vic and will cause
Y420 is treated as not supported
Solution:
store the vic which support Y420
Verify:
ah212
Test:
DRM-TX-72
Change-Id: I12ad2458458ee3ef2fc8cfbfafe6a842ff1bab37
Signed-off-by: hang cheng <hang.cheng@amlogic.com>
PD#SWPL-144323
Problem:
always hdr, when uboot enters the kernel, it will send the
sdr package first and then the hdr package, causing the TV
to flash black once or the color will darken.
Solution:
When the kernel starts up, vout takes priority over vpp.
the HDR/DV module will use vinfo information, it needs
to attach vinfo after the EDID is parsed and before the
HDR/DV module is enabled.
Verify:
ohm
Test:
DRM-TX-70
Change-Id: Ifb8e4700e33db96a2298d57d5413a8fceb232fc0
Signed-off-by: xiang.wu1 <xiang.wu1@amlogic.com>
PD#SWPL-146602
Problem:
when there are some modules which have similar name, the script will
not put the module with the short name in the modules.order
Solution:
fix the problem
Verify:
local
Change-Id: I2623b8b3ab1e44c515f5bca8217a41bb75f64cc4
Signed-off-by: Dezhen Wang <dezhen.wang@amlogic.com>
PD#SWPL-145586
Problem:
When the vic of VSDB is stored first, the vic of video data block
is inserted at the end, and an error occurs in 420 capability map
Solution:
add SVD_VIC, used to store SVD in VDB, mapped with 420 CMDB
Verify:
S5/AX201-S928X
Test:
DRM-TX-44
Change-Id: I18b42e9c304b01b6712d9640f4bc08020595f305
Signed-off-by: ruofei.zhao <ruofei.zhao@amlogic.com>
PD#SWPL-141797
Problem:
under suspend, plugin event will not be notified to
drm until resume
Solution:
sync hpd between drm and hdmitx side
Verify:
Ohm + RDK, suspend/resume
Test:
DRM-TX-38
Change-Id: I61f78dad4847d036148a4a8631279437b54eed46
Signed-off-by: hang cheng <hang.cheng@amlogic.com>
PD#SWPL-139469
Problem:
Dirty date in cacheline may pollute data in memory when
flush cacheline
Solution:
Use noncache mapping in afbc decompress lib
Verify:
T7C
Change-Id: I4f6ed2ad6473c0aaab80d175663434b38dfb2af5
Signed-off-by: zhentao.guo <zhentao.guo@amlogic.com>
PD#SWPL-145324
Problem:
kernel NULL pointer panic when cpu off line
Solution:
The root cause is that context for zram is configured as 2 part:
one for direct reclaim and one for kswapd. They are not for per-cpu
but when a core is off, zram will free per-cpu context so caused panic
on other cpus. So we should remove free code for crypto hooks.
Verify:
T5M
Change-Id: I2e4c9362bda0ddf74c91c3dec55d39dd532f46c2
Signed-off-by: Tao Zeng <tao.zeng@amlogic.com>
PD#SWPL-146339
Problem:
may overflow in some case.
Solution:
expand scs size.
Verify:
t5m.
Change-Id: I72910fc922e7339d5c0ac51f702b55b7e92b4dcc
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
PD#SWPL-146891
PD#TV-101060
Problem:
no timeout when dvbc signal is abnormal
Solution:
remove judgments of detecting Qam in the detection of SR
Verify:
TC8000
Change-Id: I0d772a0b1c3140d77415ea9bfec8658ae22175b3
Signed-off-by: caiyi.xu <caiyi.xu@amlogic.com>
PD#SWPL-139016
Problem:
demod becomes abnormal after the second
pm enable.
Solution:
1.remove before change.
2.add pm force suspend in shutdown interface.
Verify:
verified by T5M
Change-Id: I97535dfb3ad9cefa1304a6653ed4f85719640a7b
Signed-off-by: min.yang <min.yang@amlogic.com>
PD#SWPL-145324
Problem:
1, using contiguous conter for type identify is not uniform distribution
when compress for some tasks;
2, zram driver will alloc per-cpu contex for zstd, which will cause
about 3MB memory;
Solution:
1, using a fixed type array for uniform distribution of mix type
selection;
2, create 2 context: one dedicated for kswapd task; another one
dedicated for direct reclaim;
Verify:
t5m
Change-Id: I06683383e06d9601c9e6522bcd0fbcfa68a55987
Signed-off-by: Tao Zeng <tao.zeng@amlogic.com>
PD#SWPL-145324
Problem:
cma alloc too slow.
Solution:
let the idle cpu allocate more.
reduce drain lru in cma task.
Verify:
local.
Change-Id: I5f46112dcc4d5eb8fe7b21526d09a4078edc2752
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
PD#SWPL-145324
Problem:
kernel NULL pointer panic when cpu off line
Solution:
The root cause is that context for zram is configured as 2 part:
one for direct reclaim and one for kswapd. They are not for per-cpu
but when a core is off, zram will free per-cpu context so caused panic
on other cpus. So we should remove free code for crypto hooks.
Verify:
T5M
Change-Id: I2e4c9362bda0ddf74c91c3dec55d39dd532f46c2
Signed-off-by: Tao Zeng <tao.zeng@amlogic.com>
PD#SWPL-145324
Problem:
add early_param to switch in bootload.
Solution:
add early_param for trace events white list.
Verify:
local.
Change-Id: I4179e804433c61036c22868664e8c89552c5eb50
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
PD#SWPL-145324
Problem:
scs cost too many memory, 4kb per thread;
Solution:
reduce it to 1kb per thread
Verify:
local
Change-Id: I850a6c3fefa26d12527bd8e011a1de347079b865
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
PD#SWPL-145324
Problem:
There are too many trace event inodes
Solution:
Add white list for trace event, this can save about 10MB memory
Verify:
t5m
Change-Id: I7ecd89ba145e37c52d51842c1ab52f34a38ebdab
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
PD#SWPL-145324
Problem:
zstd is high compress ratio but low speed, lz4 is fast but low
compress ratio. Need balance them.
Solution:
Add a mixed crypto layer and can set ratio of lz4/zstd
Verify:
t5m
Change-Id: I4fbe930d05ef66336acacaae640fa47855d4b47b
Signed-off-by: Tao Zeng <tao.zeng@amlogic.com>
PD#SWPL-146459
Problem:
When the color depth is 444, 8bit, there is no signal output
after the resolution is switched from 720x480p to 640x480p
Solution:
Add clock configuration for 640x480p59.94hz
Verify:
S905X4/S905Y4
Test:
DRM-TX-45
Change-Id: I7acc0538fb133398bd7544e98a0f7449390d4eb9
Signed-off-by: ruofei.zhao <ruofei.zhao@amlogic.com>
PD#SWPL-140850
Problem:
5.15 ioc move to uapi.
Solution:
5.15 ioc move to uapi.
Verify:
t3.
Change-Id: Icc1bd8efb807a84d55fd3b380fd3a016b34f4849
Signed-off-by: jialong.jiang <jialong.jiang@amlogic.com>
PD#SWPL-146346
Problem:
UI response is not quickly on TC8000 projects
Solution:
Change kswapd filecache reclaim policy. Do not reclaim page cache
with mapcount larger than 1 (shared by 2 processes or above) to help
keep more mapped file caches in LRU list.
Verify:
T5M
Change-Id: I081a92debe94af608071468487b122529e6cdbda
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
PD#SWPL-145369
Problem:
try to avoid cma pool in low memory.
Solution:
add check the cma pool in low memory.
Verify:
local.
Change-Id: I5ec3a59dbd32dc4f94fbf368ce2736eb4d85ff68
Signed-off-by: Jianxiong Pan <jianxiong.pan@amlogic.com>
huiwen.yu