18 Mar, 2021
1 commit
-
CL335550 [KERNEL SPACE] dev64x_234062: fix a cache refresh bug when import user logical(pfn map).
Change-Id: I98b54b94b87251f9e3a20bf446d8fbb1130b27eb
Signed-off-by: Richard Liu
09 Mar, 2021
1 commit
-
- M865SW-692: VSI V4L2 Engineer release package 20210307
M865SW-543: [VPU/V4L2] decode: some MJPEG decoded frames have mosaic
M865SW-597: [VPU/V4L2] Easy to met hang issue when handle errors or exceptions
M865SW-650: [VPU/V4L2] Calculate h264 level before encoding since ctrl sw will check it
M865SW-675: [VPU/V4L2] Need to implement different ctrls based on encoder or decoder
M865SW-677: [VPU/V4L2] DTRC: need to report rfc luma/chroma offset of output buffer to app
M865SW-685: [VPU/V4L2] decoder: 8MQ: adding new format for DTRC & RFC
M865SW-419: [VPU/V4L2] Pass stress test for resolution change stream
M865SW-523: [VPU/V4L2] decoder: correctly handle stream file that contained codec data only
M865SW-571: [VPU/V4L2] decoder: input has 240 frames, output only get 239 pictures
M865SW-658: [VPU/V4L2] encoder: support resolution change stream transcoding.
M865SW-662: [VPU/V4l2] H264EncSetRateCtrl return error if GOP length is larger than 300Signed-off-by: Zhou Peng
Change-Id: Ifd78246208056de2d72a7a82d39bb62887828e1e
01 Mar, 2021
1 commit
-
- M865SW-681: VSI V4L2 Engineer release package 20210226
M865SW-595: [VPU/V4L2] report hdr10 metadata to userspace
M865SW-608: [VPU/V4L2] decoder: enumerate CAPTURE formats according to the currently established coded format
M865SW-655: [VPU/V4l2] Refine assert reference for Android build
M865SW-663: [VPU/V4L2] decoder: app can't receive codec error event
M865SW-668: [VPU/V4L2] daemon: update vsidaemon ts manager to fix Android GTS issue
M865SW-671: [VPU/V4L2] decoder: seek a resolution changed vp9 stream meet decoding error
M865SW-676: [VPU/V4L2] decoder: please add new formats into verfiyPlanesize()
M865SW-678: [VPU/V4L2] decoder: please add 10bit depth into vsiv4l2_decidepixeldepth()
M865SW-572: [VPU/V4L2] decoder/encoder: one frame stream can't receive EOS
M865SW-674: [VPU/V4L2] the dtrc and dtrx output is incorrect when decoding 10bit streamSigned-off-by: Zhou Peng
Change-Id: I61a196742dade4255c806b0549da77d9d261d32e
07 Feb, 2021
1 commit
-
- M865SW-664: VSI V4L2 Engineer release package 20210205
M865SW-511: [VPU/V4L2] decoder: add extr v4l2 event to report 'PIC_CONSUMED' buffer
M865SW-554: [VPU/V4L2] decoder: decoder return wrong timestamp because of wrong consumed length
M865SW-595: [VPU/V4L2] report hdr10 metadata to userspace
M865SW-608: [VPU/V4L2] decoder: enumerate CAPTURE formats according to the currently established coded format
M865SW-648: [VPU/V4L2] encoder: streamoff failed in VSI_STATUS_INIT state
M865SW-654: [VPU/V4L2] Extend maximum h265 level to 5_2
M865SW-655: [VPU/V4l2] Refine assert reference for Android build
M865SW-659: [VPU/V4L2] Android meet dead waiting in open video device if rootfs does exists vsidaemon
M865SW-523: [VPU/V4L2] decoder: correctly handle stream file that contained codec data only
M865SW-572: [VPU/V4L2] decoder/encoder: one frame stream can't receive EOS
M865SW-644: [VPU/V4L2] decoder: can't recover to capture on state if stop cmd comes earlier than capture on cmdSigned-off-by: Zhou Peng
Change-Id: I3de0c8a409e2cd9f7334369442c8d037da2471b2
03 Feb, 2021
1 commit
-
camera hal will enum all /dev/video* node, it will hang at line 548 when visit
vpu enc/dec node(it will hang in case call_usermodehelper failed)
/vendor/bin/vsidaemon do not exist on android11-2.0.0 releaseChange-Id: I28efdfebbb73c7a4a2d809ba02529bc9778383cc
Signed-off-by: zhang sanshan
01 Feb, 2021
4 commits
-
- VSI v1.4.1 20210129 formal release
Change List (compare with v1.4):
Regenerating code to make modules' boundary clearer.
Adjust license declaration to meet NXP requirement. Passed NPX's license check.
Remove unused code from project or release.
Clean up code style to get conformant to kernel spec(driver) or google (daemon).
Better support of AVS and RV format in unit test.
Better support of 10-bit decoding.
Save bandwidth by reducing data transfer number.
Add extra event for communication between driver and application.
Resolve some potential sync problem.
Better support of crop feature in decoding.
Resolve several CTS cases.
Re-organize imx-vpu.h in conformance to standard kernel driver API's framework.
JIRA ticket 503, 523, 528, 546, 554, 557, 589, 597, 608, 626, 627, 628, 633, 640 are resolved in this releaseSigned-off-by: Zhou Peng
Change-Id: I4e2779b8e878f633f4e6f002e05d927cf750ccec -
add vpu_v4l2 device node for 8mm/8mq/8mp
Signed-off-by: Zhou Peng
Change-Id: If41af5aafa79764a3d208d6888f4b2a0eca7361b -
Define variables for daemon and logpath in android
Signed-off-by: xiahong.bao
Reviewed-by: Zhou Peng
Change-Id: Ide5102a03ea394f0d6c2f2fcd888770f10c55dbe -
Add vpu v4l2 driver 'drivers/mxc/hantro_v4l2/'
Source are based on VSI 20210121 releaseSigned-off-by: Zhou Peng
Change-Id: Ic879d0a19c75336349803c2e7115e64ce93b71bf
25 Jan, 2021
1 commit
-
HAVE_COMPAT_IOCTL define is removed from include/linux/fs.h
on 5.10 kernel by below kernel patch, driver should remove
HAVE_COMPAT_IOCTL check, other drivers don’t have such check.
Kernel patch:
https://lkml.org/lkml/2020/6/22/218Change-Id: I428f762f929d1689d558b18349dd4ce536ac51f6
Signed-off-by: Richard Liu
18 Jan, 2021
2 commits
-
GPU clock will switch to minClock/64, 32/64 64/64 of max GPU frequency
according to thermal event.Change-Id: Ibcfa98ec8976215119acecbd92f2e130f0a39a73
Signed-off-by: Zhang Bo -
change "ifdef CONFIG_DEVICE_THERMAL" to "if IS_ENABLED(CONFIG_DEVICE_THERMAL)"
to enable the module with IS_BUILTIN(option) or IS_MODULE(option)).Change-Id: I116ed60acae56629713163888cb4a7d5a2d6acb3
Signed-off-by: Dandan Sun
13 Jan, 2021
3 commits
-
driver will use the data to report the skipped frame.
but the vpu driver of 8q can't get the skipped frame id,
so set it to an invalid value.Signed-off-by: Ming Qian
Reviewed-by: Shijie Qin
(cherry picked from commit f9782fd6e516647ad73476bb9d39eb2dd892ae05) -
…_G and V4L2_PIX_FMT_VC1_ANNEX_L
Due to historical reasons,
we have made a mistake abort vc1_l and vc1_g,
they are reversed.
so we need to exchange them in driver,
and the userspace also needs to be modified accordinglySigned-off-by: Ming Qian <ming.qian@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
(cherry picked from commit 9e4f4f44257312c2b4df47d0005762e8068e1bd0) -
we have add a header file imx_vpu.h
to define the custom interface based on v4l2 driver.
to unify the interface,
we need to align the 8qm/qxp v4l2 driver to this header.Signed-off-by: Ming Qian
Reviewed-by: Shijie Qin
(cherry picked from commit 8cfb286daa653c5876c157d20bed6a44420bab27)
22 Dec, 2020
1 commit
-
include more bug-fixings for vulkan and opencl3.0
Signed-off-by: Xianzhong
18 Dec, 2020
6 commits
-
* mxc/next: (279 commits)
LF-2111-2: mxc: vpu_windsor: Avoid hide parameter
MLK-24960: [8QM_MEK/8QXP_MEK]mxc:vpu_malone: move timestamp manager to firmware
MLK-24959: [8QM_MEK/8QXP_MEK]mxc:vpu_windsor: move timestamp handler to firmware
MLK-24906: [8QM_MEK/8QXP_MEK]mxc:vpu_malone: report V4L2_COLORSPACE_DEFAULT before seq hdr found
MLK-24593:[8QM_MEK/8QXP_MEK]mxc:vpu_windsor: support change bitrate dynamically
... -
* origin/mxc/vpu: (123 commits)
LF-2111-2: mxc: vpu_windsor: Avoid hide parameter
MLK-24960: [8QM_MEK/8QXP_MEK]mxc:vpu_malone: move timestamp manager to firmware
MLK-24959: [8QM_MEK/8QXP_MEK]mxc:vpu_windsor: move timestamp handler to firmware
MLK-24906: [8QM_MEK/8QXP_MEK]mxc:vpu_malone: report V4L2_COLORSPACE_DEFAULT before seq hdr found
MLK-24593:[8QM_MEK/8QXP_MEK]mxc:vpu_windsor: support change bitrate dynamically
... -
* origin/mxc/sim: (33 commits)
LF-2932 mxc: sim/emvsim: fix competition when call pm_runtime_put
MLK-24363-2 mxc: emvsim: fix error report in probe stage
MLK-24363-1 mxc: emvsim: add error check when enable clk
MLK-24364 mxc: sim: add runtime pm support
LF-1762-7 mxc: sim: replace '---help---' in Kconfig files with 'help'
... -
* origin/mxc/mlb: (20 commits)
LF-2594-6 mxc: mlb: remove imx8qm/qxp support
LF-2225-7 mlb: drop the using of smp_read_barrier_depends
LF-665-12: mxc: mlb: fix build break due to un-exist pr_warning
MLK-23000-1 mxc: mlb: fix kernel dump when use gen_pool_dma_alloc
mxc mlb: Allow compilation on imx8
... -
* origin/mxc/ipu: (13 commits)
LF-2092 mxc: IPUv3: Correct di number when enabling DIRECT_ASYNC0 channel
LF-2091 media: platform: mxc_vout: Do not check 'crop->c.width/heightr.width/heightvidioc_s_selection
... -
* origin/mxc/hdmi:
LF-2739: mfd: mxc-hdmi-core: Complete cts value for 32kHz
mxc: hdmi-cec: Add imx6 hdmi cec driver
driver: mfd: hdmi: Add hdmi core driver
14 Dec, 2020
18 commits
-
Fix another declaration hides parameter 'attr'.
(coverity ID 5572378)Signed-off-by: Shijie Qin
Reviewed-by: ming_qian -
move timestamp to firmware, it will make our driver more easy to
upstreamSigned-off-by: Ming Qian
-
move timestamp to firmware, it will make our driver more easy to
upstreamSigned-off-by: Ming Qian
-
the new gst need driver report V4L2_COLORSPACE_DEFAULT
before seq hdr is found,
otherwise the gpaly may fail to playback some stream.
To be compatible with the new version of gst,
change the behavior of try_fmt.and the gst is adapted for kernel 5.10,
so this patch will only be pushed to 5.10,
won't be included in 5.4Signed-off-by: Ming Qian
Reviewed-by: Shijie Qin -
update the encoder parameters when bitrate is changed
Signed-off-by: Ming Qian
Reviewed-by: Shijie Qin -
use bframe to control low latency mode by default
deprecate the ctrl V4L2_CID_MPEG_VIDEO_H264_ASOSigned-off-by: Ming Qian
Reviewed-by: Shijie Qin -
gst will depend on the try_fmt api
if driver don't report the correct color space in try_fmt
gst may return failSigned-off-by: Ming Qian
Reviewed-by: Shijie Qin -
Fix declaration hides parameter 'attr'.
(coverity ID 5572378)Signed-off-by: Shijie Qin
Reviewed-by: ming_qian -
Bit field of 'valid_ts_received' is 1, change its typedef to
unsigned int. (coverity ID 9000734)Signed-off-by: Shijie Qin
Reviewed-by: ming_qian -
use the v4l2 api instead of custom api
Signed-off-by: Ming Qian
Reviewed-by: Shijie Qin -
1. enable color description in vui when encoding.
2. improve the get/set color description apiSigned-off-by: Ming Qian
Reviewed-by: Shijie Qin -
Update Kconfig to cope with upstream change
commit 84af7a6194e4 ("checkpatch: kconfig: prefer 'help' over
'---help---'").Signed-off-by: Dong Aisheng
-
Replace spin_lock() with spin_lock_irqsave() to avoid potential deadlock
Signed-off-by: Zhou Peng
(cherry picked from commit 85b5d239b0e2fb6c534064ff203f74a2e2b0e7e8) -
HW enable and disable are shared with the same function, driver need to
check the parameter to identify whether HW is enabled or notSigned-off-by: Zhou Peng
(cherry picked from commit 742c7adea8b19b047b6a5bce1c8d7da6895a3b71) -
If 'pm_runtime_get_sync()' fail, need to reset device power usage_count
to '0'. Otherwise, the usage_count always bigger than '0' and not do
'runtime_suspend()' actually.Signed-off-by: Shijie Qin
Reviewed-by: Anson Huang
Reviewed-by: ming_qian -
If 'pm_runtime_get_sync()' fail, need to reset device power usage_count
to '0'. Otherwise, the usage_count always bigger than '0' and not do
'runtime_suspend()' actually.Signed-off-by: Shijie Qin
Reviewed-by: Anson Huang
Reviewed-by: ming_qian -
there may be some redundant data before the encoded frame,
generally it won't bring any problem.
but on android, the redundant data may affect the mux,
can cause the mp4 can't be decoded.
so delete the redundant data in the driver.Signed-off-by: Ming Qian
Reviewed-by: Shijie Qin -
For more readable, use a handler table to insert scode for
each format.Signed-off-by: Shijie Qin
Reviewed-by: ming_qian