question: CSI on imx8mq with (any) CSI2 camera / experience with mx6s_capture?

Laurent Pinchart laurent.pinchart at
Thu Dec 10 04:23:59 EST 2020

Hi Martin,

On Thu, Dec 10, 2020 at 09:17:48AM +0100, Martin Kepplinger wrote:
> hi,
> TL;DR: did you use the NXP "mx6s_capture" csi bridge driver with other 
> cameras?

I've recently worked on camera support for i.MX8MM (whose camera IP
cores are, if not identical, very similar to the i.MX8MQ's). The least I
can say is that it was painful :-(

I'm using an MT9M114 sensor, which can produce RAW8, RAW10 and YUV and
has a CSI-2 interface. My first use case is to capture RAW10, which
isn't supported by the mx6s_capture driver.

> I try to use a CSI2 camera (hi846 I'm writing a driver for) on imx8mq: 
> Using NXP's CSI bridge driver
> as well as the CSI driver itself:
> works fine when using the ov5640 camera with this driver:
> (I realize there is a CSI bridge driver in staging, but that need more 
> work to be actually used. Of course after this the goal is to fix and 
> use it; and mainline a CSI phy driver too.)

I have lots of patches for this driver, which I've developed on i.MX7D
for a separate project. I'd like to mainline them, but this is blocked
by one last issue that I haven't been able to solve yet. In a nutshell,
the CSI writes two consecutive frames in each buffer, overflowing the
allocated memory. The registers that control the buffer size seem to be
programmed correctly as far as I can tell. I've reported this issue to
NXP but haven't received any feedback yet.

I've also added support for i.MX8MM to this driver, but haven't been
able to capture RAW8, RAW10 or YUV successfully. YUV produces the "best
images", but seems to drop 3 out of 4 consecutive pixels. The sensor
driver has been tested successfully on i.MX6 so I don't think it's the
most likely cause of issues, but I can't rule out bugs on that side
either. The i.MX8 reference manuals describe the MIPI_CSI and CSI IP
cores but doesn't tell much about how the two are connected, so I have
lots of unanswered questions about the register fields that control the
interface between those IP cores. I'm pretty sure there are issues in
that area of the drivers, but I've tested all combinations I could think
of, without luck.

> Now I use said NXP's CSI drivers on mainline without problems with the 
> ov5640. For the hi846 (as an example), I want to use a different 
> pixelformat and CSI media bus format. Describing, say, 
> MEDIA_BUS_FMT_SBGGR10_1X10 in the driver lets mx6s_capture find it; but 
> as soon as I try to add a different pixelformat than V4L2_PIX_FMT_UYVY in:
> streaming won't start anymore: "not negotiated, -4" and the 
> "mx6s_vidioc_enum_framesizes" ioctl is not called anymore. Why so?

That will likely be the easy part of the problem. The NXP driver doesn't
support RAW10 at all, so you will have to add that, and I don't think it
will be easy.

I'll push my development branch shortly if you're interested in trying
the mainline driver.

> I didn't find much when comparing strace from "gst-launch-1.0 v4l2src ! 
> video/x-raw,width=1280,height=720 ! videoconvert ! xvimagesink" (which 
> is how I test).
> When I simply use V4L2_PIX_FMT_UYVY I *do* get interrupts (in mx6s) but 
> only "FIELD0_INT" (and drawing at this point suggests that buffer 1 is 
> not yet full and "distorted" in a wrong format, which could explain why 
> DMA is never completed in order to create a full frame).
> Now this is details, but the hi846 camera uses a 10bit CSI format and 
> therefore I need to set PIXEL_BIT in cr1 too, but when I do that for 
> example, I don't get "FIELD0_INT" anymore (only SFF_OR_INT and of course 
> the "base address switching Change Err").
> Do you have experience with using this driver with other cameras and can 
> point me in a direction that might help me?
> If you want to look at the work-in-progress driver:


Laurent Pinchart

More information about the linux-arm-kernel mailing list