PD#SWPL-155395
Problem:
When modules send mailbox message to aocpu, sometimes
aocpu can not response because of stuck or interrupt
signal lost, at this time driver modules need to know
whether aocpu is alive or not, this is helpful to debug.
Solution:
Add aocpu alive check mechanism, when aocpu did not
response, system will check and print aocpu status.
Verify:
T5M-AY309
Change-Id: Idd1c47fd99505f3a159c86390c28303f5a5b60e9
Signed-off-by: Yao Jie <jie.yao@amlogic.com>
PD#SWPL-163844
Problem:
switch to 59.94hz, not set 60hz mode when qms enable
because frac mode
Solution:
set frac 0 when brr mode
Verify:
s7d
Test:
DRM-TX-75
Change-Id: I69fc38ffee456d2a917f586e7ed04dad3aa8a22d
Signed-off-by: Wenjie Qiao <wenjie.qiao@amlogic.com>
PD#SWPL-158042
Problem:
drm module automic-check and automic-enable use differnet hpd status
information when hdmitx hotplug
Solution:
[hpd] add system timein hdmitx state
Verify:
S5/SC2
Test:
DRM-TX-103
Change-Id: Id0f37cf01c5266a720f2881ee8cd9fab274b79b9
Signed-off-by: yahui.liu <yahui.liu@amlogic.com>
PD#SWPL-143076
Problem:
mmu copy cannot access address over 4G.
Solution:
set alloc from dma32 zone when config_alloc_flags bit0 is set.
Verify:
S5
Change-Id: I142f182dff1509e591d20ed0e440a6851f8d7029
Signed-off-by: gan.zhang <gan.zhang@amlogic.com>
PD#SWPL-161016
Problem:
When s7 is in standby and wakes up, it will turn off and on the vpu
power domain. When it is turned off, the reg of the relevant modules
will be reset, the output no signal when it wakes up.
Solution:
hdmitx driver needs to reinitialize the required top register when
it wake up
Verify:
s7
Test:
DRM-TX-78, DRM-TX-79, DRM-TX-80, DRM-TX-81
Change-Id: Ie9d458087fcc9ea672ec477d46b88a6d7ddcec81
Signed-off-by: xiang.wu1 <xiang.wu1@amlogic.com>
PD#SWPL-161276
Problem:
set arb and urgent default val
Solution:
add arb and urgent default val
Verify:
s7
Change-Id: Ic0dd2f35e65e9b1580970d82096d067e9a15da5c
Signed-off-by: hai.cao <hai.cao@amlogic.com>
PD#SWPL-157220
Problem:
When the edid only supports 576p 16x9 and hwc set to 576p50hz,
576p 4x3 is selected as the output mode, resulting in output failure.
Solution:
When mode match timing name, 16x9 timing is preferred.
EDID only support 4x3, match 4x3 timing
EDID only support 16x9, match 16x9 timing
EDID support 4x3 and 16x9, match 16x9 timing
Verify:
bluebell
Test:
DRM-TX-107
Change-Id: I49fb7114edbde884c14724dbc403df5384968059
Signed-off-by: zhou.han <zhou.han@amlogic.com>
PD#SWPL-158260
Problem:
amlvideo2 want to get hdcp status
Solution:
set hdcp flag in vf if HDCP input
Verify:
t7c
Change-Id: I1d5ddcb5d3e5d1f467e9de7f20dc76a96fdd94b4
Signed-off-by: haitao.liu <haitao.liu@amlogic.com>
PD#GH-3737
PD#SWPL-155123
Problem:
play HDR video under DV_STD, output as SDR
Solution:
auto change the color depth to support HDR under DV_STD
Verify:
SC2
S5
Test:
DRM-TX-90
Change-Id: I255011995b80d2cf01499ea8c3584cbda5e54c5f
Signed-off-by: ruofei.zhao <ruofei.zhao@amlogic.com>
PD#SWPL-152701
Problem:
Currently,we found that the Sanitize/Erase command may take longer
than that on the first factory reset after flashing.
This will cause the Sanitize/Erase command to timeout before completing.
Solution:
Enable max timeout in the eMMC controller only for Sanitize.
This will allow the factory reset command to always complete
successfully with the drawback that on eMMC failures, the device will
be stuck unless rebooted.
Verify:
T5d
Change-Id: Ifea2edfe758b2f723f2c02d369468e8ba9629fe3
Signed-off-by: Long <long.yu@amlogic.com>
PD#SWPL-158107
Problem:
tracer is not written in the edid parsing process
Solution:
edid tracer post-processing
Verify:
SC2
S5
Test:
DRM-TX-99
Change-Id: Ie802d13da665ae0846a6829ecca0618583d54166
Signed-off-by: ruofei.zhao <ruofei.zhao@amlogic.com>
PD#SWPL-96632
Problem:
hdmitx_common_validate_format_para has no mutex lock protect.
when hdmitx plug event happen in encoder atomic check stage,
it will lead to info mismatch
Solution:
add hdmitx_common_validate_mode_locked api which has mutex lock
to protect the encoder check flow.
Will revert it when we move edid parering to drm.
Verify:
ah212
Test:
DRM-OSD-5
Change-Id: I7b4c98fac8378133166100f9d22ddcfe0af66407
Signed-off-by: Ao Xu <ao.xu@amlogic.com>
PD#SWPL-147355
Problem:
due to lockup when dmc irq ratio is too high
Solution:
use dmc irq thread when irq ratio is too high
Verify:
local
Change-Id: Ie13d584f81563e3c2e111fb5d6033f0e2aa0d337
Signed-off-by: qinglin.li <qinglin.li@amlogic.com>
PD#SWPL-153877
Problem:
conflict between drm and dv when write afbc top ctrl.
Solution:
dv add graphic interface to drm.
drm get dv status and write corresponding reg bit.
Verify:
t7c
Change-Id: I98f27d82f77af8aad35fba65b66b66f482d0802a
Signed-off-by: jialong.jiang <jialong.jiang@amlogic.com>
PD#SWPL-146509
Problem:
add new interface for di back
Solution:
1.add api for release buffer when unreg/reg
2.add api for post/pre num
Verify:
sc2
Change-Id: Ie1eadc81751b3ca6da2d1f337e94d54dc1455c6c
Signed-off-by: qianqian.cai <qianqian.cai@amlogic.com>
PD#SWPL-113283
Problem:
when use dma to read a small amount of data to a memory by kzalloc,
kzalloc memory len less than dma minimum handling length,
e.g. kzalloc one byte memory, but dma moves data with a width 64 bits,
so it would destroy the 7 bytes memory space on the right side of the
requested memory.
BUG: KFENCE: memory corruption in meson_nfc_exec_op+0x3b4/0x4c8
Corrupted memory at 0x00000000527f2fd0 [ ! ! ! ! ! ! ! . . . . . . . ]
meson_nfc_exec_op+0x3b4/0x4c8
nand_status_op+0x1ac/0x294
nand_erase_op+0x19c/0x328
nand_erase_nand+0x1d0/0x3cc
nand_erase+0x14/0x20
mtd_erase+0x174/0x1f0
mtdchar_ioctl+0x1c1c/0x2704
mtdchar_unlocked_ioctl+0x50/0x74
__arm64_sys_ioctl+0xa4/0x114
invoke_syscall+0x5c/0x140
el0_svc_common.llvm.4567374417812983254+0xb4/0xf0
do_el0_svc+0x24/0x84
el0_svc+0x2c/0xa4
el0t_64_sync_handler+0x68/0xb4
el0t_64_sync+0x1b4/0x1b8
kfence-#52: 0x00000000d4c8f37e-0x00000000d4c8f37e, size=1, cache=kmalloc-64
Solution:
align dma minimum handling length to kzalloc
Verify:
AT301_T962D4-K35E(1.5G) #256
Change-Id: I541b6acecaa9dd69ec6466bd7040af94b45ae596
Signed-off-by: zhikui.cui <zhikui.cui@amlogic.com>
PD#SWPL-154473
Problem:
other module can not get video real axis
Solution:
provide get video real axis function
Verify:
all chips
Change-Id: I0dae2a30c2841396cda7ee4bfefd57bfff186d59
Signed-off-by: hai.cao <hai.cao@amlogic.com>
PD#SWPL-141096
Problem:
add vdec unsupport format status use by media_hal
Solution:
add vdec unsupport format status use by media_hal
Verify:
txhd2
Change-Id: I2ecf78ac8e91ca912f331dc8a8ed8a6494610dd5
Signed-off-by: zehong.luo <zehong.luo@amlogic.com>
PD#SWPL-155145
Problem:
Tracer prints abnormally when playing hdr10 plus video
Solution:
fix the abnormal printing problem of tracer when
playing hdr10 plus video
Verify:
SC2
S5
Test:
DRM-TX-91
Change-Id: Ia081b5667bfb01d540c497faa32479c38b084da9
Signed-off-by: ruofei.zhao <ruofei.zhao@amlogic.com>
PD#SWPL-135758
Problem:
there is no s7 pwm tee driver
Solution:
add pwm tee driver
Verify:
s7 pxp
Change-Id: If1ec9b943340687173827e114df1207001ec2daa
Signed-off-by: junyi.zhao <junyi.zhao@amlogic.com>
PD#SWPL-151874
Problem:
need do aipq for T5M.
Solution:
modify code to support.
Verify:
T5M
Change-Id: Ief28052e1df8f67fff4e291319511d48ca6e2075
Signed-off-by: chen.xu <chen.xu@amlogic.com>
PD#SWPL-151864
Problem:
dv not support different size for two slice
Solution:
vpp config same size for vd1 two slice, it means that
dv input size is same for two slice and overlap is
fixed 96. dv output size maybe different, it depends
on vpp slice[0].hsize and slice[1].hsize
Verify:
t3x
Change-Id: I9986a84968e63939843996ae59d42d82c7707c67
Signed-off-by: Pengcheng Chen <pengcheng.chen@amlogic.com>
PD#SWPL-151748
Problem:
update hdmitx version for release.
sabrina u / kernel 5.15: hdmitx driver refactoring
sabrina release base from trunk11.20
audio mat question modified by cl:380286
Solution:
update hdmitx version
Verify:
verify on ohm
Test:
DRM-TX-40
Change-Id: Ie5ccf28423bbcecfdaddaf97a64fbfd692c3bf23
Signed-off-by: yahui.liu <yahui.liu@amlogic.com>
Yao Jie