Dear Marek Vasut, > Dear Benoît Thébaudeau, > > > Signed-off-by: Benoît Thébaudeau <benoit.thebaud...@advansee.com> > > Cc: Marek Vasut <ma...@denx.de> > > Cc: Ilya Yanok <ilya.ya...@cogentembedded.com> > > Cc: Stefan Herbrechtsmeier <ste...@herbrechtsmeier.net> > > --- > > Changes for v2: N/A. > > Changes for v3: > > - New patch. > > > > .../drivers/usb/host/ehci-hcd.c | 135 > > +++++++++++--------- .../drivers/usb/host/ehci.h > > | > > 75 ++++++++++- 2 files changed, 150 insertions(+), 60 deletions(-) > > > > diff --git u-boot-usb-8d5fb14.orig/drivers/usb/host/ehci-hcd.c > > u-boot-usb-8d5fb14/drivers/usb/host/ehci-hcd.c index > > 5b3b906..1977c28 > > 100644 > > --- u-boot-usb-8d5fb14.orig/drivers/usb/host/ehci-hcd.c > > +++ u-boot-usb-8d5fb14/drivers/usb/host/ehci-hcd.c > > @@ -163,7 +163,7 @@ static int ehci_reset(void) > > > > #ifdef CONFIG_USB_EHCI_TXFIFO_THRESH > > cmd = ehci_readl(&hcor->or_txfilltuning); > > - cmd &= ~TXFIFO_THRESH(0x3f); > > + cmd &= ~TXFIFO_THRESH_MASK; > > cmd |= TXFIFO_THRESH(CONFIG_USB_EHCI_TXFIFO_THRESH); > > ehci_writel(&hcor->or_txfilltuning, cmd); > > #endif > > @@ -186,7 +186,7 @@ static int ehci_td_buffer(struct qTD *td, void > > *buf, > > size_t sz) while (idx < QT_BUFFER_CNT) { > > td->qt_buffer[idx] = cpu_to_hc32(addr); > > td->qt_buffer_hi[idx] = 0; > > - next = (addr + 4096) & ~4095; > > + next = (addr + EHCI_PAGE_SIZE) & ~(EHCI_PAGE_SIZE - 1); > > delta = next - addr; > > if (delta >= sz) > > break; > > @@ -208,7 +208,8 @@ ehci_submit_async(struct usb_device *dev, > > unsigned long > > pipe, void *buffer, int length, struct devrequest *req) > > { > > ALLOC_ALIGN_BUFFER(struct QH, qh, 1, USB_DMA_MINALIGN); > > - ALLOC_ALIGN_BUFFER(struct qTD, qtd, 3, USB_DMA_MINALIGN); > > +#define QTD_COUNT 3 > > + ALLOC_ALIGN_BUFFER(struct qTD, qtd, QTD_COUNT, USB_DMA_MINALIGN); > > int qtd_counter = 0; > > > > volatile struct qTD *vtd; > > @@ -230,7 +231,7 @@ ehci_submit_async(struct usb_device *dev, > > unsigned long > > pipe, void *buffer, le16_to_cpu(req->index)); > > > > memset(qh, 0, sizeof(struct QH)); > > - memset(qtd, 0, 3 * sizeof(*qtd)); > > + memset(qtd, 0, QTD_COUNT * sizeof(*qtd)); > > > > toggle = usb_gettoggle(dev, usb_pipeendpoint(pipe), > > usb_pipeout(pipe)); > > > > @@ -246,19 +247,20 @@ ehci_submit_async(struct usb_device *dev, > > unsigned > > long pipe, void *buffer, */ > > qh->qh_link = cpu_to_hc32((uint32_t)qh_list | QH_LINK_TYPE_QH); > > c = (usb_pipespeed(pipe) != USB_SPEED_HIGH && > > - usb_pipeendpoint(pipe) == 0) ? 1 : 0; > > - endpt = (8 << 28) | > > - (c << 27) | > > - (usb_maxpacket(dev, pipe) << 16) | > > - (0 << 15) | > > - (1 << 14) | > > - (usb_pipespeed(pipe) << 12) | > > - (usb_pipeendpoint(pipe) << 8) | > > - (0 << 7) | (usb_pipedevice(pipe) << 0); > > + usb_pipeendpoint(pipe) == 0); > > + endpt = (8 << QH_ENDPT1_RL) | > > + (c << QH_ENDPT1_C) | > > Maybe define it as #deifne QH_ENDPT1(x) ((x) << SEOMTHING) ? > [...]
For all of these? > > @@ -398,50 +408,53 @@ ehci_submit_async(struct usb_device *dev, > > unsigned > > long pipe, void *buffer, ALIGN((uint32_t)buffer + length, > > ARCH_DMA_MINALIGN)); > > > > /* Check that the TD processing happened */ > > - if (token & 0x80) { > > + if (token & (QT_TOKEN_STATUS_ACTIVE << QT_TOKEN_STATUS)) > > printf("EHCI timed out on TD - token=%#x\n", token); > > - } > > > > /* Disable async schedule. */ > > cmd = ehci_readl(&hcor->or_usbcmd); > > cmd &= ~CMD_ASE; > > ehci_writel(&hcor->or_usbcmd, cmd); > > > > - ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, 0, > > + ret = handshake((uint32_t *)&hcor->or_usbsts, STS_ASS, 0, > > Ooooh, nice catch :) > > [...] > The rest is cool. > > btw when (I hope you will) resubmitting next time, just submit the > whole series > under 0/8 patch (or 1/8) of the old one to make it a nice thread. OK, so with 2/8 removed since you have applied it. What is the rule here? I thought that a new version of a patch should be posted as a reply to the previous version so that patchwork could mark the old version as superseded (even if this feature is not yet available for U-Boot). Does it apply only to single patches while series should be replied to the previous 0/n? And should 3/n be a reply to 2/n, to 0/n (or to 1/n if no 0/n), or to nothing? I will wait for your review of 8/8. Best regards, Benoît _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot