| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <3154909.aeNJFYEL58@steina-w>
Date: Mon, 07 Feb 2022 09:59:48 +0100
From: Alexander Stein <alexander.stein@...tq-group.com>
To: Laurent Pinchart <laurent.pinchart@...asonboard.com>
Cc: Steve Longerbeam <slongerbeam@...il.com>,
Philipp Zabel <p.zabel@...gutronix.de>,
Mauro Carvalho Chehab <mchehab@...nel.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Shawn Guo <shawnguo@...nel.org>,
Sascha Hauer <s.hauer@...gutronix.de>,
Pengutronix Kernel Team <kernel@...gutronix.de>,
Fabio Estevam <festevam@...il.com>,
NXP Linux Team <linux-imx@....com>,
Rui Miguel Silva <rmfrfs@...il.com>,
Dorota Czaplejewicz <dorota.czaplejewicz@...i.sm>,
linux-media@...r.kernel.org, linux-staging@...ts.linux.dev,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: (EXT) Re: [PATCH 0/8] imx7/imx8mm media / csi patches
Hi Laurent,
Am Samstag, 5. Februar 2022, 04:16:54 CET schrieb Laurent Pinchart:
> On Fri, Feb 04, 2022 at 01:15:06PM +0100, Alexander Stein wrote:
> > Hey everyone,
> >
> > this is a set of patch for imx[7] media drivers based on next-20220203.
> > With this set I was able to capture video frames from a MIPI CSI-2 camera
> > in my TQMa8MQML + MBA8MX hardware. The actual camera used is a Vision
> > Components 'VC MIPI IMX327 C' camera. IMX327 is compatible to IMX290.
> > Patch 8 shows the DT overlay I'm using, not suitable for merging.
>
> You may be interested in
> https://gitlab.com/ideasonboard/nxp/linux/-/commits/pinchartl/v5.17/sensors
> /
Thanks for your feedback. Working on imx290 driver for proper imx327 support
is on my todo. For the records, there are also patches at [1].
> > Please ignore the FPGA part, this is mainly for power supply and GPIO
> > reset
> > line. This is currently a custom driver I'm working on, but I do not want
> > to focus on in this series.
> >
> > Please note I tested this only on this imx8 platform.
> >
> > First thanks to Dorota for the patchset at [1] (patches 1-4) which is
> > necessary to capture correct images. I integrated Hans' review into it. I
> > hope the I didn't make a mistake and the original author is kept along. I
> > used v4 for that patchset, it is v1 again in this set. I hope this is not
> > confusing.
> >
> > Starting from patch 5 there are small fixes which allows me to configure
> > my
> > media device.
> >
> > Device configuration:
> > ```
> > media-ctl -l "'imx290 2-001a':0->'csis-32e30000.mipi-csi':0 [1]"
> > media-ctl -V "'imx290 2-001a':0 [fmt:SRGGB10_1X10/1920x1080 field:none
> > colorspace:raw]" media-ctl -V "'csi':0 [fmt:SRGGB10_1X10/1920x1080
> > field:none colorspace:raw]" v4l2-ctl -d0 --set-fmt-video
> > width=1920,height=1080,pixelformat='RG10',field=none media-ctl -p
> > ```
> >
> > The media-ctl topology is:
> > ```
> > # media-ctl -p
> > Media controller API version 5.17.0
> >
> > Media device information
> > ------------------------
> > driver imx7-csi
> > model imx-media
> > serial
> > bus info platform:32e20000.csi
> > hw revision 0x0
> > driver version 5.17.0
> >
> > Device topology
> > - entity 1: csi (2 pads, 2 links)
> >
> > type V4L2 subdev subtype Unknown flags 0
> > device node name /dev/v4l-subdev0
> >
> > pad0: Sink
> >
> > [fmt:SRGGB10_1X10/1920x1080 field:none colorspace:raw
> > xfer:none ycbcr:601 quantization:full-range] <-
> > "csis-32e30000.mipi-csi":1 [ENABLED,IMMUTABLE]
> >
> > pad1: Source
> >
> > [fmt:SRGGB10_1X10/1920x1080 field:none colorspace:raw
> > xfer:none ycbcr:601 quantization:full-range] -> "csi
> > capture":0 [ENABLED,IMMUTABLE]
> >
> > - entity 4: csi capture (1 pad, 1 link)
> >
> > type Node subtype V4L flags 0
> > device node name /dev/video0
> >
> > pad0: Sink
> >
> > <- "csi":1 [ENABLED,IMMUTABLE]
> >
> > - entity 10: csis-32e30000.mipi-csi (2 pads, 2 links)
> >
> > type V4L2 subdev subtype Unknown flags 0
> > device node name /dev/v4l-subdev1
> >
> > pad0: Sink
> >
> > [fmt:SRGGB10_1X10/1920x1080 field:none colorspace:raw
> > xfer:709 ycbcr:601 quantization:lim-range] <- "imx290
> > 2-001a":0 [ENABLED]
> >
> > pad1: Source
> >
> > [fmt:SRGGB10_1X10/1920x1080 field:none colorspace:raw
> > xfer:709 ycbcr:601 quantization:lim-range] -> "csi":0
> > [ENABLED,IMMUTABLE]
> >
> > - entity 15: imx290 2-001a (1 pad, 1 link)
> >
> > type V4L2 subdev subtype Sensor flags 0
> > device node name /dev/v4l-subdev2
> >
> > pad0: Source
> >
> > [fmt:SRGGB10_1X10/1920x1080 field:none colorspace:raw]
> > -> "csis-32e30000.mipi-csi":0 [ENABLED]
> >
> > ```
> >
> > Note: MIPI CSI settle times are not calculated correctly right now and
> > need a manual overwrite:
> > echo 13 > /sys/kernel/debug/32e30000.mipi-csi/ths_settle
> > echo 2 > /sys/kernel/debug/32e30000.mipi-csi/tclk_settle
>
> The reference manual isn't very prolix on tclk_settle :-S That's
> something I'd love to figure out one day.
>
> For ths_settle, is the issue on the CSIS driver side, or the sensor
> driver side ?
I can't answer that yet, but during my work on a fslc linux-5.10 I noticed
that the link frequency and/or some other clock was wrong, so the calculated
value didn't match.
Anyway I get the impression that the final values might also be affected by
the actual hardware. But I do not know any details regarding this.
> > I ignored the settings for xfer, ycbcr and quantization for now. I do
> > neither know how they will affect me nor what it should be.
> > Also I did not focus on v4l2-compliance tool, this is a further task as
> > well. IMHO imx7-mipi-csis.c should also create an immutable link to the
> > camera sensor.
>
> That makes sense, but note that, while CSI-2 is meant to be a
> point-to-point bus, there are boards designed with multiple sensors
> connected to the same CSI-2 RX without any hardware multiplexer. They
> keep one of the two sensors in reset at all times and are lucky that the
> signal reflections don't mess things up.
Still, even if only one device is actually powered, it is a immutable link
from v4l2 perspective. You don't switch the sensors on the fly, or do you
really do that? In the end immutable links are only a minor issue (to me).
Regards,
Alexander
[1] https://github.com/raspberrypi/linux/commits/rpi-5.10.y/drivers/media/i2c/
imx290.c
Powered by blists - more mailing lists