Re: [RFC PATCH v3 00/18] Support for Tegra video capture from external sensor

[Sowjanya Komatineni]

On 7/17/20 10:48 AM, Sowjanya Komatineni wrote:
> On 7/17/20 10:38 AM, Sowjanya Komatineni wrote:
> > On 7/17/20 10:23 AM, Sowjanya Komatineni wrote:
> > > On 7/17/20 10:08 AM, Hans Verkuil wrote:
> > > > On 17/07/2020 18:34, Sowjanya Komatineni wrote:
> > > > > On 7/17/20 3:54 AM, Hans Verkuil wrote:
> > > > > > Hi Sowjanya,
> > > > > >
> > > > > > On 15/07/2020 06:20, Sowjanya Komatineni wrote:
> > > > > > > This series adds support for video capture from external camera 
> > > > > > > sensor to
> > > > > > > Tegra video driver.
> > > > > > >
> > > > > > > Jetson TX1 has camera expansion connector and supports custom 
> > > > > > > camera module
> > > > > > > designed as per TX1 design specification.
> > > > > > >
> > > > > > > This series also enables camera capture support for Jetson Nano 
> > > > > > > which has
> > > > > > > Raspberry PI camera header.
> > > > > > >
> > > > > > > This series is tested with IMX219 camera sensor.
> > > > > > >
> > > > > > > This series include,
> > > > > > >
> > > > > > > VI I2C related fixes
> > > > > > > - Camera sensor programming happens through VI I2C which is on 
> > > > > > > host1x bus.
> > > > > > > - These patches includes device tree and I2C driver fixes for VI 
> > > > > > > I2C.
> > > > > > >
> > > > > > > Tegra video driver updates
> > > > > > > - TPG Vs Non-TPG based on Kconfig
> > > > > > > - Support for external sensor video capture based on device 
> > > > > > > graph from DT.
> > > > > > > - Support for selection ioctl operations
> > > > > > > - Tegra MIPI CSI pads calibration
> > > > > > > - CSI T-CLK and T-HS settle time computation based on clock rates.
> > > > > > >
> > > > > > > Host1x driver updates
> > > > > > > - Adds API to allow creating mipi device for specific device node.
> > > > > > > - Splits MIPI pads calibrate start and waiting for calibration 
> > > > > > > to be done.
> > > > > > >
> > > > > > > Device tree updates
> > > > > > > - Adds camera connector 2V8, 1V8, 1V2 regulator supplies to 
> > > > > > > Jetson TX1 DT.
> > > > > > > - Enabled VI and CSI support in Jetson Nano DT.
> > > > > > I'm doing a bit of stress testing with:
> > > > > >
> > > > > > while true; do v4l2-ctl --stream-mmap --stream-count=1; done
> > > > > >
> > > > > > and I see that the imx274 has often streaming failures:
> > > > > >
> > > > > > [Â 172.025144] IMX274 8-001a: s_stream failed
> > > > > > [Â 179.025192] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 3132 = 870 (2 bytes)
> > > > > > [Â 179.033575] IMX274 8-001a: s_stream failed
> > > > > > [Â 226.525378] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 3130 = 878 (2 bytes)
> > > > > > [Â 226.533761] IMX274 8-001a: s_stream failed
> > > > > > [Â 227.029325] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 30f6 = 107 (2 bytes)
> > > > > > [Â 227.037758] IMX274 8-001a: s_stream failed
> > > > > > [Â 247.025218] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 30f6 = 107 (2 bytes)
> > > > > > [Â 247.033658] IMX274 8-001a: s_stream failed
> > > > > > [Â 293.025517] IMX274 8-001a: s_stream failed
> > > > > > [Â 309.024727] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 30e0 = 0 (2 bytes)
> > > > > > [Â 309.032969] IMX274 8-001a: s_stream failed
> > > > > > [Â 309.529506] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 30f8 = 11d2 (3 bytes)
> > > > > > [Â 309.538103] IMX274 8-001a: imx274_set_frame_length error = -121
> > > > > > [Â 309.544102] IMX274 8-001a: imx274_set_frame_interval error = -121
> > > > > > [Â 309.550243] IMX274 8-001a: s_stream failed
> > > > > > [Â 314.025561] IMX274 8-001a: s_stream failed
> > > > > > [Â 329.025586] IMX274 8-001a: s_stream failed
> > > > > > [Â 340.529567] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 303a = f0c (2 bytes)
> > > > > > [Â 340.538009] IMX274 8-001a: s_stream failed
> > > > > > [Â 347.525627] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 30f6 = 107 (2 bytes)
> > > > > > [Â 347.534008] IMX274 8-001a: s_stream failed
> > > > > > [Â 365.033640] IMX274 8-001a: s_stream failed
> > > > > > [Â 437.525788] IMX274 8-001a: imx274_write_mbreg : i2c bulk write 
> > > > > > failed, 3038 = c (2 bytes)
> > > > > > [Â 437.533997] IMX274 8-001a: s_stream failed
> > > > > > [Â 456.029780] IMX274 8-001a: s_stream failed
> > > > > > [Â 472.025862] IMX274 8-001a: s_stream failed
> > > > > > [Â 498.025861] IMX274 8-001a: s_stream failed
> > > > > > [Â 500.025905] IMX274 8-001a: s_stream failed
> > > > > >
> > > > > > where v4l2-ctl returns:
> > > > > >
> > > > > > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ VIDIOC_STREAMON returned -1 (Remote I/O error)
> > > > > >
> > > > > > I don't see this with the imx219.
> > > > > >
> > > > > > I also see this occasionally:
> > > > > >
> > > > > > [Fri Jul 17 12:51:42 2020] video4linux video1: failed to run 
> > > > > > capture start kthread: -4
> > > > > >
> > > > > > Something is not stable here.
> > > > > >
> > > > > > Regards,
> > > > > >
> > > > > > ÂÂÂÂHans
> > > > > Hi Hans,
> > > > >
> > > > > Running the same single frame continuous loop for more than 2 
> > > > > hours now
> > > > > and I don't see any failure.
> > > > >
> > > > > Above failure shows i2c bulk writes to IMX274 failure due to which
> > > > > s_stream also failed.
> > > > >
> > > > > Not sure if its due to i2c mux in the path to sensor on your module
> > > > > causing some issue when there is more i2c write traffic as we are 
> > > > > doing
> > > > > single stream in continuous loop. Also IMX219 does not show on 
> > > > > your side
> > > > > so something specific to IMX274 setup probably.
> > > > I'll take a closer look next week. Good to know that it works fine 
> > > > for you.
> > > >
> > > > > Regarding kthread_run failure where kthread_run() returned -EINTR 
> > > > > during
> > > > > capture start thread, I always see this happen at the point of 
> > > > > stopping
> > > > > the continuous single stream while loop by pressing ctrl+c after few
> > > > > loops of execution.
> > > > Hmm, if this is normal behavior, then should this message be a 
> > > > debug message
> > > > only? Or perhaps only show the message if the error code != EINTR.
> > >
> > > I believe its good to still show this as its reported by kthread_run 
> > > -> kthread_create_on_node.
> > >
> > > But not sure in real usecase we will ever use while true like this 
> > > and we should use script to also break while loop along with 
> > > v4l2-ctl termination when ctrl-c terminate request happens.
> > Hi Hans, As this happens only during this type of case, I can update 
> > to show message only when error code != EINTR.
> >
> > Thanks
> >
> > Sowjanya
>
>
> Sorry, Was thinking to not mask debug message for -EINTR in case if it 
> happens in any other valid scenarios. If you still want to mask, will 
> update in next version.
As we are running application that uses threads depending on when ctrl-c 
gets hit I see sigkill happens during kthread run even with break 
causing this.

Good way I see it working all the time is to use crtl-z and then kill 
application as below

Command to run

while true; do ./v4l2-ctl --stream-mmap --stream-count=1; done

To stop ,

press ctrl-z and then execute kill %%


> > > > Regards,
> > > >
> > > > ÂÂÂÂHans
> > > >
> > > > > while true; do v4l2-ctl --stream-mmap --stream-count=1; done
> > > > >
> > > > > when we stop loop with ctrl+c, v4l2-ctl terminates but loop does not
> > > > > terminate immediately and probably SIGKILLed is seen prior to 
> > > > > complete.
> > > > >
> > > > > Using below can help to terminate loop as well when we stop ctrl-c 
> > > > > and
> > > > > with this I don't see any repro of EINTR error from kthread_run 
> > > > > when run
> > > > > in infinite loop.
> > > > >
> > > > > while true; do ./v4l2-ctl --stream-mmap --stream-count=1 || break; 
> > > > > done
> > > > >
> > > > >
> > > > >
> > > > > > > Delta between patch versions:
> > > > > > >
> > > > > > > [v3]:ÂÂÂ Includes v2 feedback
> > > > > > > ÂÂÂÂ- Uses separate helper function for retrieving remote csi 
> > > > > > > subdevice
> > > > > > > ÂÂÂÂÂ and source subdevice.
> > > > > > > ÂÂÂÂ- Added check for presence of subdevice ops set/get_selection
> > > > > > > ÂÂÂÂ- dropped vb2_queue_release from driver and using
> > > > > > > ÂÂÂÂÂ vb2_video_unregister_device instead of 
> > > > > > > video_unregister_device.
> > > > > > > ÂÂÂÂ- video device register should happen in the last after all 
> > > > > > > video
> > > > > > > ÂÂÂÂÂ device related setup is done in the driver. This is being 
> > > > > > > addressed
> > > > > > > ÂÂÂÂÂ in below RFC patch. Once proper implementation of this is 
> > > > > > > available
> > > > > > > ÂÂÂÂÂ will update Tegra video driver to use split APIs and do 
> > > > > > > all setup
> > > > > > > ÂÂÂÂÂ prior to device register. Added this as TODO in the driver.
> > > > > > > https://www.spinics.net/lists/linux-media/msg172761.html
> > > > > > >
> > > > > > > ÂÂÂÂNote:
> > > > > > > ÂÂÂÂPatch-0012 has compilation dependency on
> > > > > > > ÂÂÂÂhttps://patchwork.kernel.org/patch/11659521/
> > > > > > >
> > > > > > >
> > > > > > > [v2]:ÂÂÂ Includes below changes based on v1 feedback
> > > > > > > ÂÂÂÂ- dt-binding document and the driver update for device graph 
> > > > > > > to use
> > > > > > > ÂÂÂÂÂ separate ports for sink endpoint and source endpoint for csi.
> > > > > > > ÂÂÂÂ- Use data-lanes endpoint property for csi.
> > > > > > > ÂÂÂÂ- Update tegra_mipi_request() to take device node pointer 
> > > > > > > argument
> > > > > > > ÂÂÂÂÂ rather than adding extra API.
> > > > > > > ÂÂÂÂ- Remove checking for clk pointer before clk_disable.
> > > > > > >
> > > > > > >
> > > > > > > Sowjanya Komatineni (18):
> > > > > > > ÂÂÂ dt-bindings: i2c: tegra: Document Tegra210 VI I2C clocks and
> > > > > > > ÂÂÂÂÂ power-domains
> > > > > > > ÂÂÂ arm64: tegra: Add missing clocks and power-domains to 
> > > > > > > Tegra210 VI I2C
> > > > > > > ÂÂÂ i2c: tegra: Don't mark VI I2C as IRQ safe runtime PM
> > > > > > > ÂÂÂ i2c: tegra: Remove NULL pointer check before
> > > > > > > ÂÂÂÂÂ clk_enable/disable/prepare/unprepare
> > > > > > > ÂÂÂ i2c: tegra: Fix the error path in tegra_i2c_runtime_resume
> > > > > > > ÂÂÂ i2c: tegra: Fix runtime resume to re-init VI I2C
> > > > > > > ÂÂÂ i2c: tegra: Avoid tegra_i2c_init_dma() for Tegra210 vi i2c
> > > > > > > ÂÂÂ media: tegra-video: Fix channel format alignment
> > > > > > > ÂÂÂ media: tegra-video: Enable TPG based on kernel config
> > > > > > > ÂÂÂ media: tegra-video: Update format lookup to offset based
> > > > > > > ÂÂÂ dt-bindings: tegra: Update VI and CSI bindings with port info
> > > > > > > ÂÂÂ media: tegra-video: Add support for external sensor capture
> > > > > > > ÂÂÂ media: tegra-video: Add support for selection ioctl ops
> > > > > > > ÂÂÂ gpu: host1x: mipi: Update tegra_mipi_request() to be node based
> > > > > > > ÂÂÂ gpu: host1x: mipi: Use readl_relaxed_poll_timeout in 
> > > > > > > tegra_mipi_wait
> > > > > > > ÂÂÂ gpu: host1x: mipi: Split tegra_mipi_calibrate and 
> > > > > > > tegra_mipi_wait
> > > > > > > ÂÂÂ media: tegra-video: Add CSI MIPI pads calibration
> > > > > > > ÂÂÂ media: tegra-video: Compute settle times based on the clock 
> > > > > > > rate
> > > > > > >
> > > > > > > ÂÂ .../display/tegra/nvidia,tegra20-host1x.txt | 92 ++-
> > > > > > > ÂÂ .../devicetree/bindings/i2c/nvidia,tegra20-i2c.txt | 19 +-
> > > > > > > ÂÂ arch/arm64/boot/dts/nvidia/tegra210.dtsi | 6 +
> > > > > > > ÂÂ drivers/gpu/drm/tegra/dsi.c | 9 +-
> > > > > > > ÂÂ drivers/gpu/host1x/mipi.c | 37 +-
> > > > > > > ÂÂ drivers/i2c/busses/i2c-tegra.c | 101 +--
> > > > > > > ÂÂ drivers/staging/media/tegra-video/Kconfig | 7 +
> > > > > > > ÂÂ drivers/staging/media/tegra-video/csi.c | 247 ++++++-
> > > > > > > ÂÂ drivers/staging/media/tegra-video/csi.h | 8 +
> > > > > > > ÂÂ drivers/staging/media/tegra-video/tegra210.c | 25 +-
> > > > > > > ÂÂ drivers/staging/media/tegra-video/vi.c | 793 
> > > > > > > +++++++++++++++++++--
> > > > > > > ÂÂ drivers/staging/media/tegra-video/vi.h | 25 +-
> > > > > > > ÂÂ drivers/staging/media/tegra-video/video.c | 23 +-
> > > > > > > ÂÂ include/linux/host1x.h | 4 +-
> > > > > > > ÂÂ 14 files changed, 1242 insertions(+), 154 deletions(-)