On 17/05/11 21:29, Alan Stern wrote:
> On Tue, 17 May 2011, Jan Andersson wrote:
>
>> This patch adds support for big endian GRUSBHC UHCI controllers.
>> The HCD bus glue will probe the register interface to determine
>> the endianness of the controller.
>>
>> Tested on GR-LEON4-ITX board which has a controller with little endian
>> interface and on custom LEON3 board with a BE controller.
>
> ...
>
>> --- a/drivers/usb/host/uhci-grlib.c
>> +++ b/drivers/usb/host/uhci-grlib.c
>> @@ -25,6 +25,22 @@ static int uhci_grlib_init(struct usb_hcd *hcd)
>> {
>> struct uhci_hcd *uhci = hcd_to_uhci(hcd);
>>
>> + /* Probe to see if we have a big or little endian controller. Bit 7
>> + * of PORTSC is always 1 and 15:13 are always zero, so we have:
>> + *
>> + * 1------- 000----- -------- 000----- => Little endian mode HC
>> + * 000----- 1------- 000----- -------- => Big endian mode HC
>> + *
>> + * We also need to take into account that readl will byte swap.
>> + *
>> + * If we detect a big endian register interface we also assume that
>> + * the controller uses big endian descriptors.
>> + */
>> + if (!(uhci_readl(uhci, USBPORTSC1) & 0x80)) {
>> + uhci->big_endian_mmio = 1;
>> + uhci->big_endian_desc = 1;
>> + }
>
> Sorry for not noticing this earlier. This is a 16-bit register, so
> shouldn't you be accessing it with uhci_readw() instead of
> uhci_readl()?
The hardware allows reading both registers in one go with a 32-bit
access so the code above works. I used a 32-bit access since I had used
the same technique for a debug monitor where I was limited to 32-bit
accesses.
>
> Then the comment could be shortened too, since only the first two byte
> values on each line would be needed and the line about readl could go
> away.
You are right that the comment could be shortened, but readw also
byteswaps so the remark about readl would just be changed to readw :-)
>
> Apart from this minor detail, the patch series is fine with me. You
> can resubmit this last patch if you want to fix it up, and Greg can add
>
> Acked-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
>
> to patches 1/5, 3/5, and 5/5.
>
Thank you. I want to test if I change to readw so that I don't make a
mistake with all the swaps that take place. I will resubmit this after
testing tomorrow - unless you can accept the use of readl() above.
Best regards,
Jan
--
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
