Hi Dmitry, Comments inlined. swami ________________________________________ From: Dmitry Krivoschekov [dmitry.krivoschekov@xxxxxxxxx] Sent: Thursday, August 21, 2008 8:39 PM To: Subbrathnam, Swaminathan Cc: linux-omap@xxxxxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx; felipe.balbi@xxxxxxxxx Subject: Re: Patch [2/2] DaVinci CPPI TX DMA tasklet Hi Swami, Swaminathan S wrote: > Signed-off-by: Swaminathan S <swami.iyer@xxxxxx> > > This patch implements a tasklet to poll for TX Endpoint FIFO > empty status before completing a transfer. This is required > in the context of DaVinci CPPI DMA Tx transfers. CPPI > DMA engine generates completion interrupt before the data > actually moves out of TX FIFO I guess the DMA interrupt is generated after the data have been moved from a s/w buffer to TX FIFO. <Swami> Yes. This being an external DMA (unlike Mentor DMA) that was the designed behavior. > and the driver needs to save the > Data toggle after ensuring that the FIFO is empty. why don't rely on an endpoint interrupt which is generated after a USB transaction is complete? <Swami> When in DMA mode (mode 1) Endpoint interrupt does not get generated (only DMA completion interrupt) and hence cannot rely on that. Thanks, Dmitry-- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
