RE: [RESEND PATCH v1 0/7] Performance improvement of decoder

From: jackson . lee
Date: Sun Apr 13 2025 - 22:04:37 EST

Next message: yohan.joung: "[PATCH v2] f2fs: Improve large section GC by locating valid block segments"
Previous message: Shuai Xue: "Re: [PATCH v10 10/11] arm64: idle: export arch_cpu_idle()"
In reply to: Nicolas Dufresne: "Re: [RESEND PATCH v1 0/7] Performance improvement of decoder"
Next in thread: Nicolas Dufresne: "Re: [RESEND PATCH v1 0/7] Performance improvement of decoder"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi Nicolas

> -----Original Message-----
> From: Nicolas Dufresne <nicolas.dufresne@xxxxxxxxxxxxx>
> Sent: Saturday, April 12, 2025 1:58 AM
> To: jackson.lee <jackson.lee@xxxxxxxxxxxxxxx>; mchehab@xxxxxxxxxx;
> hverkuil-cisco@xxxxxxxxx; sebastian.fricke@xxxxxxxxxxxxx;
> bob.beckett@xxxxxxxxxxxxx; dafna.hirschfeld@xxxxxxxxxxxxx
> Cc: linux-media@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; lafley.kim
> <lafley.kim@xxxxxxxxxxxxxxx>; b-brnich@xxxxxx; hverkuil@xxxxxxxxx; Nas
> Chung <nas.chung@xxxxxxxxxxxxxxx>
> Subject: Re: [RESEND PATCH v1 0/7] Performance improvement of decoder
>
> Hi Jackson,
>
> Le jeudi 10 avril 2025 à 12:39 +0900, Jackson.lee a écrit :
> > From: Jackson Lee <jackson.lee@xxxxxxxxxxxxxxx>
> >
> > v4l2-compliance results:
> > ========================
> >
> > v4l2-compliance 1.28.1-5233, 64 bits, 64-bit time_t
> >
> > Buffer ioctls:
> >                 warn: v4l2-test-buffers.cpp(693): VIDIOC_CREATE_BUFS
> > not supported
> >                 warn: v4l2-test-buffers.cpp(693): VIDIOC_CREATE_BUFS
> > not supported
> >         test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
> >         test CREATE_BUFS maximum buffers: OK
> >         test VIDIOC_EXPBUF: OK
> >         test Requests: OK (Not Supported)
> >
> > Total for wave5-dec device /dev/video0: 46, Succeeded: 46, Failed: 0,
> > Warnings: 2
> > Total for wave5-enc device /dev/video1: 46, Succeeded: 46, Failed: 0,
> > Warnings: 0
> >
> > Fluster test results:
> > =====================
> >
> > Running test suite JCT-VC-HEVC_V1 with decoder GStreamer-H.265-V4L2-
> > Gst1.0
> > Using 3 parallel job(s)
> > Ran 133/147 tests successfully               in 41.629 secs
>
> Same results here.
>
> >
> > (1 test fails because of not supporting to parse multi frames, 1 test
> > fails because of a missing frame and slight corruption,
> > 2 tests fail because of sizes which are incompatible with the IP, 11
> > tests fail because of unsupported 10 bit format)
> >
> >
> > Running test suite JVT-AVC_V1 with decoder GStreamer-H.264-V4L2-
> > Gst1.0
> > Using 3 parallel job(s)
> > Ran 78/135 tests successfully               in 44.578 secs
> >
> > (57 fail because the hardware is unable to decode MBAFF / FMO / Field
> > / Extended profile streams.)
>
> Same results here. There is also JVT-FR-EXT test suite now that you should
> include. 23/69 here, without incident, but I did not analyze the inner
> results, I'll leave that to you. Though, a quick looks shows that
> YUV422 does not work anymore.
>
> >
> > Seek test
> > =====================
> > 1. gst-play-1.0 seek.264
> > 2. this will use waylandsink since gst-play-1.0 uses playbin.
> >    if you don't want to hook up display,
> >    you can run gst-play-1.0 seek.264 --videosink=fakevideosink instead
> > 3. Let pipeline run for 2-3 seconds 4. press SPACE key to pause 5.
> > press 0 to reset press SPACE to start again
> >
> > gst-play-1.0 seek.264 --videosink=fakevideosink Press 'k' to see a
> > list of keyboard shortcuts.
> > Now playing /root/seek.264
> > Redistribute latency...
> > Redistribute latency...
> > Redistribute latency...
> > Redistribute latency...
> > Redistribute latency...aused
> > 0:00:09.9 / 0:00:09.7
> > Reached end of play list.
>
> So, I managed to resurrect my device. Once side effect of this series is
> that the driver is no longer silent in normal cases. Pretty 'q'
> while playing, or seeking seems to fill the kernel logs with these two
> error.
>
> [ 5037.457307] vdec 4210000.video-codec: wave5_vpu_dec_finish_decode:
> could not get output info.
> [ 5037.457436] vdec 4210000.video-codec:
> wave5_vpu_firmware_command_queue_error_check: result not ready: 0x800
>
> This needs fixing for the next version. The condition that makes these non
> error needs to be tested so that we don't get spammed anymore. They also
> occur while running fluster.
>
> >
> > Sequence Change test
> > =====================
> > gst-launch-1.0 filesrc location=./switch_1080p_720p_240frames.h264 !
> > h264parse ! v4l2h264dec ! filesink location=./h264_output_420.yuv
> > Setting pipeline to PAUSED ...
> > Pipeline is PREROLLING ...
> > Redistribute latency...
> > Redistribute latency...
> > Pipeline is PREROLLED ...
> > Setting pipeline to PLAYING ...
> > Redistribute latency...
> > New clock: GstSystemClock
> > Got EOS from element "pipeline0".
> > Execution ended after 0:00:05.772414400 Setting pipeline to NULL ...
> > Freeing pipeline ...
>
> I did a test of my own here, and did get kernel splat. The warning
> indicate that the state machine is no longer respected. This needs to be
> address in v2, we added these check, since the locking is bound to legal
> use of the state machine.

Since applying the performance patch, device_run and wave5_vpu_dec_finish_decode is not synchronized any more.
How about removing this warning message ?

>
> [ 401.018648] Execution of a job in state STOP illegal.
> [ 401.023761] WARNING: CPU: 0 PID: 635 at
> drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c:1715
> wave5_vpu_dec_device_run+0x448/0x83c [wave5] [ 401.036890] Modules linked
> in: rfkill ip6table_filter ip6_tables iptable_filter ip_tables x_tables
> rpmsg_ctrl rpmsg_char phy_cadence_torrent rtc_tps6594 tps6594_esm
> tps6594_regulator tps6594_pfsm pinctrl_tps6594 gpio_regmap ti_am335x_adc
> cdns3 kfifo_buf cdns_usb_common qrtr mux_gpio omap_mailbox phy_j721e_wiz
> wave5 phy_can_transceiver ti_k3_r5_remoteproc v4l2_mem2mem
> videobuf2_dma_contig videobuf2_memops tps6594_i2c videobuf2_v4l2
> tps6594_core at24 k3_j72xx_bandgap sa2ul videodev m_can_platform
> videobuf2_common authenc m_can ti_k3_dsp_remoteproc mc cdns3_ti
> ti_am335x_tscadc can_dev rti_wdt fuse drm backlight dm_mod ipv6
> [ 401.091795] CPU: 0 UID: 1000 PID: 635 Comm: h264parse0:sink Tainted:
> G W 6.15.0-rc1-jacinto+ #1 PREEMPT
> [ 401.102731] Tainted: [W]=WARN
> [ 401.105687] Hardware name: Texas Instruments J721S2 EVM (DT)
> [ 401.111330] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS
> BTYPE=--)
> [ 401.118277] pc : wave5_vpu_dec_device_run+0x448/0x83c [wave5]
> [ 401.124015] lr : wave5_vpu_dec_device_run+0x448/0x83c [wave5]
> [ 401.129749] sp : ffff800084a0ba40 [ 401.133051] x29: ffff800084a0baf0
> x28: ffff00080a64c130 x27:
> ffff800084a0bc08
> [ 401.140175] x26: 00000000c058560f x25: 0000000000000000 x24:
> ffff00081331c000
> [ 401.147297] x23: ffff00081331c010 x22: ffff00080ed059a8 x21:
> ffff00081331dbc0
> [ 401.154418] x20: ffff00081331d000 x19: 0000000000000000 x18:
> 0000000000000006
> [ 401.161540] x17: 0000000000000000 x16: 0000000000000000 x15:
> 072e076c07610767
> [ 401.168662] x14: ffff00080ee0b500 x13: 072e076c07610767 x12:
> ffff800082107128
> [ 401.175783] x11: 0000000000000058 x10: 0000000000000018 x9 :
> ffff00080ee0b500
> [ 401.182907] x8 : 00000000000004c7 x7 : ffff00080ee0b500 x6 :
> ffff80008215f128
> [ 401.190028] x5 : 0000000000000000 x4 : 0000000000000000 x3 :
> 0000000000000001
> [ 401.197149] x2 : 0000000000000000 x1 : 0000000000000000 x0 :
> ffff00080ee0b480
> [ 401.204272] Call trace:
> [ 401.206709] wave5_vpu_dec_device_run+0x448/0x83c [wave5] (P)
> [ 401.212448] v4l2_m2m_try_run+0x84/0x134 [v4l2_mem2mem] [ 401.217667]
> v4l2_m2m_qbuf+0x184/0x240 [v4l2_mem2mem] [ 401.222709]
> v4l2_m2m_ioctl_qbuf+0x18/0x4e0 [v4l2_mem2mem] [ 401.228184]
> v4l_qbuf+0x48/0x70 [videodev] [ 401.232292] __video_do_ioctl+0x40c/0x4a0
> [videodev] [ 401.237260] video_usercopy+0x1e0/0x688 [videodev]
> [ 401.242054] video_ioctl2+0x18/0x38 [videodev] [ 401.246500]
> v4l2_ioctl+0x40/0x60 [videodev] [ 401.250774] __arm64_sys_ioctl+0xb4/0xf4
> [ 401.254690] invoke_syscall+0x48/0x104 [ 401.258433]
> el0_svc_common.constprop.0+0x40/0xe0
> [ 401.263125] do_el0_svc+0x1c/0x28
> [ 401.266431] el0_svc+0x30/0xcc
> [ 401.269480] el0t_64_sync_handler+0x10c/0x138 [ 401.273827]
> el0t_64_sync+0x198/0x19c [ 401.277480] ---[ end trace 0000000000000000 ]-
> -- [ 401.316876] ------------[ cut here ]------------ [ 401.321523]
> Execution of a job in state STOP illegal.
> [ 401.326992] WARNING: CPU: 0 PID: 635 at
> drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c:1715
> wave5_vpu_dec_device_run+0x448/0x83c [wave5]
>
>
> To test:
> gst-launch-1.0 videotestsrc num-buffers=3 ! video/x-
> raw,format=NV12,width=320,height=240 ! v4l2h264enc ! filesink
> location=340x240.h264
> gst-launch-1.0 videotestsrc num-buffers=3 ! video/x-
> raw,format=NV12,width=640,height=480 ! v4l2h264enc ! filesink
> location=640x480.h264 cat 340x240.h264 640x480.h264 340x240.h264
> 640x480.h264 > drc.h264
> gst-launch-1.0 filesrc location=drc.h264 ! parsebin ! v4l2h264dec !
> fakevideosink -v
>
>
> >
> > Change since v0:
> > ===================
> > * For [PATCH v1 2/7] media: chips-media: wave5: Improve performance of
> > decoder
> > - separates the previous patch to a few patches
> >
> > * For [PATCH v1 3/7] media: chips-media: wave5: Fix not to be closed
> > - separated from the previous patch of performance improvement of
> >    decoder
> >
> > * For [PATCH v1 4/7] media: chips-media: wave5: Use spinlock whenever
> > state is changed
> > - separated from the previous patch of performance improvement of
> >    decoder
> >
> > * For [PATCH v1 5/7] media: chips-media: wave5: Fix not to free
> > resources normally when
> >     instance was destroyed
> > - separated from the previous patch of performance improvement of
> >    decoder
> >
> > * For [PATCH v1 7/7] media: chips-media: wave5: Fix SError of kernel
> > panic when closed
> > - separated from the previous patch of performance improvement of
> >    decoder
> >
> > Jackson Lee (7):
> > media: chips-media: wave5: Fix Null reference while testing fluster
> > media: chips-media: wave5: Improve performance of decoder
> > media: chips-media: wave5: Fix not to be closed
> > media: chips-media: wave5: Use spinlock whenever state is changed
> > media: chips-media: wave5: Fix not to free resources normally when
> >     instance was destroyed
> > media: chips-media: wave5: Reduce high CPU load
> > media: chips-media: wave5: Fix SError of kernel panic when closed
> >
> > .../platform/chips-media/wave5/wave5-helper.c | 10 +-
> > .../chips-media/wave5/wave5-vpu-dec.c         | 116 +++++++++++-----
> > --
> > .../chips-media/wave5/wave5-vpu-enc.c         |   8 +-
> > .../platform/chips-media/wave5/wave5-vpu.c    | 70 +++++++++--
> > .../platform/chips-media/wave5/wave5-vpuapi.c | 36 +++---
> > .../platform/chips-media/wave5/wave5-vpuapi.h | 10 ++
> > .../chips-media/wave5/wave5-vpuconfig.h       |   1 +
> > 7 files changed, 179 insertions(+), 72 deletions(-)
>
> --
> Nicolas Dufresne
> Principal Engineer at Collabora

Next message: yohan.joung: "[PATCH v2] f2fs: Improve large section GC by locating valid block segments"
Previous message: Shuai Xue: "Re: [PATCH v10 10/11] arm64: idle: export arch_cpu_idle()"
In reply to: Nicolas Dufresne: "Re: [RESEND PATCH v1 0/7] Performance improvement of decoder"
Next in thread: Nicolas Dufresne: "Re: [RESEND PATCH v1 0/7] Performance improvement of decoder"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]