Dear technical support:
I use mx8_v4l2_cap_drm.c to test and collect image data, and add the code in the static int save_to_file(int ch, struct video_channel *video_ch) function:
memcpy(cpu_buf,video_ch[ch].buffers[buf_id].planes[i].start,1280*720*2);
It takes 20 milliseconds to execute this copy.
Is there any way to shorten the copy time? Such as DMA?
Thank you!
Solved! Go to Solution.
Hi 783435100
what processor revision used in the case, recommended to use latest rev.C0,
it improved image susystem performance as described in AN12770
i.MX 8X C0 Migration Guide Migrating from i.MX 8X B0 to i.MX 8X C0
Recommended to try copying on i.MX8QXP MEK board using latest linux release
https://source.codeaurora.org/external/imx/linux-imx/tree/?h=imx_5.4.47_2.2.0
Best regards
igor
Hi 783435100
what processor revision used in the case, recommended to use latest rev.C0,
it improved image susystem performance as described in AN12770
i.MX 8X C0 Migration Guide Migrating from i.MX 8X B0 to i.MX 8X C0
Recommended to try copying on i.MX8QXP MEK board using latest linux release
https://source.codeaurora.org/external/imx/linux-imx/tree/?h=imx_5.4.47_2.2.0
Best regards
igor
Is there a way?
one can try copy with edma, M4 SDK has example in
..memory_to_memory /edma_memory_to_memory.c
https://mcuxpresso.nxp.com/en/welcome
Best regards
igor
I just use imx-yocto-L5.4.47_2.2.0