diff options
author | Roger Quadros <rogerq@ti.com> | 2012-08-07 13:09:56 (GMT) |
---|---|---|
committer | Felipe Balbi <balbi@ti.com> | 2012-08-07 13:16:34 (GMT) |
commit | 4f3e8d263d34e52e75b5adfa14811467d3033d8e (patch) | |
tree | 4316c907e4fcc85fa01bb80ed800a4180f1adc1b /drivers | |
parent | 8e8a55165469c99af0c24a276d997f9473dc89ab (diff) | |
download | linux-fsl-qoriq-4f3e8d263d34e52e75b5adfa14811467d3033d8e.tar.xz |
usb: musb: use DMA mode 1 whenever possible
Do not rely on any hints from gadget drivers and use DMA mode 1
whenever we expect data of at least the endpoint's packet size and
have not yet received a short packet.
The last packet if short is always transferred using DMA mode 0.
This patch fixes USB throughput issues in mass storage mode for
host to device transfers.
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Diffstat (limited to 'drivers')
-rw-r--r-- | drivers/usb/musb/musb_gadget.c | 30 |
1 files changed, 4 insertions, 26 deletions
diff --git a/drivers/usb/musb/musb_gadget.c b/drivers/usb/musb/musb_gadget.c index 8d2cce1..5c4392b 100644 --- a/drivers/usb/musb/musb_gadget.c +++ b/drivers/usb/musb/musb_gadget.c @@ -707,12 +707,11 @@ static void rxstate(struct musb *musb, struct musb_request *req) fifo_count = musb_readw(epio, MUSB_RXCOUNT); /* - * Enable Mode 1 on RX transfers only when short_not_ok flag - * is set. Currently short_not_ok flag is set only from - * file_storage and f_mass_storage drivers + * use mode 1 only if we expect data of at least ep packet_sz + * and have not yet received a short packet */ - - if (request->short_not_ok && fifo_count == musb_ep->packet_sz) + if ((request->length - request->actual >= musb_ep->packet_sz) && + (fifo_count >= musb_ep->packet_sz)) use_mode_1 = 1; else use_mode_1 = 0; @@ -727,27 +726,6 @@ static void rxstate(struct musb *musb, struct musb_request *req) c = musb->dma_controller; channel = musb_ep->dma; - /* We use DMA Req mode 0 in rx_csr, and DMA controller operates in - * mode 0 only. So we do not get endpoint interrupts due to DMA - * completion. We only get interrupts from DMA controller. - * - * We could operate in DMA mode 1 if we knew the size of the tranfer - * in advance. For mass storage class, request->length = what the host - * sends, so that'd work. But for pretty much everything else, - * request->length is routinely more than what the host sends. For - * most these gadgets, end of is signified either by a short packet, - * or filling the last byte of the buffer. (Sending extra data in - * that last pckate should trigger an overflow fault.) But in mode 1, - * we don't get DMA completion interrupt for short packets. - * - * Theoretically, we could enable DMAReq irq (MUSB_RXCSR_DMAMODE = 1), - * to get endpoint interrupt on every DMA req, but that didn't seem - * to work reliably. - * - * REVISIT an updated g_file_storage can set req->short_not_ok, which - * then becomes usable as a runtime "use mode 1" hint... - */ - /* Experimental: Mode1 works with mass storage use cases */ if (use_mode_1) { csr |= MUSB_RXCSR_AUTOCLEAR; |