There is a poll loop for max 25us for HS devices. Now guess what, I tested it in gadget mode and forgot about the little detail. Nobody seem to have it noticed… This patch adds the missing logic for hostmode so it is recognized in host and device mode properly. Fixes: 50aea6fca771 ("usb: musb: cppi41: fire hrtimer according to programmed channel length") Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> --- v1…v2: proper ident. Felipe, I've been told that his is no risk while patch two makes things wrose for ISO. I added a fixes tag since this should improve performance for HDD behind an USB bridge. For a proper/better #2 I need to take of ISO first… drivers/usb/musb/musb_cppi41.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/usb/musb/musb_cppi41.c b/drivers/usb/musb/musb_cppi41.c index 5a9b977fbc19..f64fd964dc6d 100644 --- a/drivers/usb/musb/musb_cppi41.c +++ b/drivers/usb/musb/musb_cppi41.c @@ -253,6 +253,7 @@ static void cppi41_dma_callback(void *private_data) cppi41_trans_done(cppi41_channel); } else { struct cppi41_dma_controller *controller; + int is_hs = 0; /* * On AM335x it has been observed that the TX interrupt fires * too early that means the TXFIFO is not yet empty but the DMA @@ -265,7 +266,14 @@ static void cppi41_dma_callback(void *private_data) */ controller = cppi41_channel->controller; - if (musb->g.speed == USB_SPEED_HIGH) { + if (is_host_active(musb)) { + if (musb->port1_status & USB_PORT_STAT_HIGH_SPEED) + is_hs = 1; + } else { + if (musb->g.speed == USB_SPEED_HIGH) + is_hs = 1; + } + if (is_hs) { unsigned wait = 25; do { -- 2.1.3 -- 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