Hello. David Brownell wrote:
Following this are three patches from Sergei, which provide some very nice TX DMA improvements (which is the case that should work particularly smoothly), mostly for host side:
- bugfixes for multi-packet TXDMA support - sanitize clearing TXCSR DMA bits (take 2) - fix isochronous TXDMA (take 2)
RXDMA is still a bit problematic, since it's hard to use
Notte that ISO Rx DMA is still broken for non-Inventra DMA (and the recent patch fixing it for Inventra even added to that breakage). In particular, multi-frame transfers shouldn't work. But since audio recording reportedly works even without this, I still haven't looked into fixing...
multipacket DMA mode ... but it's easy to use that "mode 1" for all relevant TX transfers, reducing the CPU load in at least one side of USB protocol interactions.
I've updated the patch comments (e.g. para #1 from the third patch became one para and five bullet points, much easier to parse), removed some CamelCasing,
What CamelCasing, where?!
and made simple cosmetic changes to maintain current driver style.
That so called style makes me sick... :-/
Verified on DaVinci (with CPPI DMA) and OMAP3 (Inventra DMA) in both PIO and DMA modes.
Thanks. Dmitry has verified it with Inventra DMA too but not with the recent kernel, so hasn't reported to the list...
WBR, Sergei -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
