On 25/04/2022 09:11, Laurent Pinchart wrote: > On Mon, Apr 25, 2022 at 08:57:07AM +0200, Hans Verkuil wrote: >> On 28/03/2022 16:13, Xavier Roumegue wrote: >>> The DW100 driver gets the dewarping mapping as a binary blob from the >>> userspace application through a custom control. >>> The blob format is hardware specific so create a dedicated control for >>> this purpose. >>> >>> Signed-off-by: Xavier Roumegue <xavier.roumegue@xxxxxxxxxxx> >>> --- >>> Documentation/userspace-api/media/drivers/dw100.rst | 12 ++++++++++++ >>> include/uapi/linux/dw100.h | 11 +++++++++++ >>> 2 files changed, 23 insertions(+) >>> create mode 100644 include/uapi/linux/dw100.h >>> >>> diff --git a/Documentation/userspace-api/media/drivers/dw100.rst b/Documentation/userspace-api/media/drivers/dw100.rst >>> index 4cd55c75628e..f6d684cadf26 100644 >>> --- a/Documentation/userspace-api/media/drivers/dw100.rst >>> +++ b/Documentation/userspace-api/media/drivers/dw100.rst >>> @@ -20,4 +20,16 @@ match the expected size inherited from the destination image resolution. >>> More details on the DW100 hardware operations can be found in >>> *chapter 13.15 DeWarp* of IMX8MP_ reference manuel. >>> >>> +The Vivante DW100 m2m driver implements the following driver-specific control: >>> + >>> +``V4L2_CID_DW100_DEWARPING_16x16_VERTEX_MAP (integer)`` >> >> (integer) -> (__u32 array) >> >> But should this be a __u32 array at all? Wouldn't a __u16 array make more sense? >> >>> + Specifies to DW100 driver its dewarping map (aka LUT) blob as described in >>> + *chapter 13.15.2.3 Dewarping Remap* of IMX8MP_ reference manual as an U32 >>> + dynamic array. The image is divided into many small 16x16 blocks. If the >>> + width of the image is not divisible by 16, the size of the rightmost block >>> + is the remainder. >> >> Isn't the same true for the height? >> >> The dewarping map only saves the vertex coordinates of the >>> + block. The dewarping grid map is comprised of vertex coordinates for x and y. >>> + Each x, y coordinate register uses 16 bits (UQ12.4) to record the coordinate >> >> As mentioned before, UQ12.4 is not necessarily a standard notation. 'unsigned 12.4 >> fixed point' is better, but you also need to specify exactly where the bits are >> stored inside the __u16. I.e.: 'the integer part is stored in the 12 most significant >> bits, and the fractional part is stored in the 4 least significant bits of the __u16.' > > Isn't that implied ? I've never seen fixed-point numbers stored the > other way around. True, perhaps that's overkill. > > Regarding the Q notation, while it was coined by TI, I think it's > widespread enough to be used here. I don't mind much though. I had to look it up :-) That might say more about me, though... I think the key phrase that is missing here is "fixed point". Regards, Hans > >>> + address, with the Y coordinate in the upper bits and X in the lower bits. >> >> And with a __u16 array this becomes: 'The array contains pairs of X, Y coordinates.' >> Or something along those lines. >> >>> + >>> .. _IMX8MP: https://www.nxp.com/webapp/Download?colCode=IMX8MPRM >>> diff --git a/include/uapi/linux/dw100.h b/include/uapi/linux/dw100.h >>> new file mode 100644 >>> index 000000000000..7fdcf2bf42e5 >>> --- /dev/null >>> +++ b/include/uapi/linux/dw100.h >>> @@ -0,0 +1,11 @@ >>> +/* SPDX-License-Identifier: GPL-2.0-only WITH Linux-syscall-note */ >>> +/* Copyright 2022 NXP */ >>> + >>> +#ifndef __UAPI_DW100_H__ >>> +#define __UAPI_DW100_H__ >>> + >>> +#include <linux/v4l2-controls.h> >>> + >> >> Add a comment referring to the Documentation/userspace-api/media/drivers/dw100.rst >> documentation so users of this control know where to find the associated >> documentation. >> >>> +#define V4L2_CID_DW100_DEWARPING_16x16_VERTEX_MAP (V4L2_CID_USER_DW100_BASE + 1) >>> + >>> +#endif >