Questions - ESP32 Camera beyond JPEG

[jhinkle]

I have been studying your ESP32 Camera project on github to get an understanding on how I2S in camera mode and DMA works.

I have several questions: More than three since there are multiple questions embedded in each below.

#1. During the init process of DMA descr - fields EOF and OWNER are set or cleared (depending on what source you look at). Looking at the I2S driver code, I can find nothing that looks at or changes those values so are they ignored?

#2. DMA buffer is limited to 4096 bytes. If the camera data line is more (as in 1600 pixels) - I am assuming the DMA will fill one buffer and then go to the next one in the link list (correct?). If the line takes less than two buffers, will the DMA controller issue an IRQ indicating that HREF is now low? Will the DMA controller move to the next buffer in the linklist or will it wait for the next line and continue filling until the buffer is full? (This is were I would think the EOF field would be useful).

#3. The code in the camera project acquires 8 DMA descr/buffers. I never see any code to imply in-use/empty/ready - etc. Is this because the design is such that with 8 buffers linked in a circular queue that one will always be processed and ready by the time the DMA needs a new one?

Thanks in advance for any comments / insights.

Joe