On Fri, 04 Dec 2009 09:34:06 +0000 David Vrabel <david.vrabel@xxxxxxx> wrote:
> Andrew Morton wrote:
> > On Mon, 24 Aug 2009 15:54:17 +0000 David Vrabel <david.vrabel@xxxxxxx> wrote:
> >
> >> + if (std->len + dma_len > QTD_MAX_XFER_SIZE) {
> >> + dma_len = QTD_MAX_XFER_SIZE - std->len;
> >> + ep = ((dma_addr + dma_len) / qset->max_packet) * qset->max_packet;
> >> + dma_len = ep - dma_addr;
> >> + }
> >
> > Generates a call to __udivdi3 on i386, which doesn't exist.
>
> This only happens if CONFIG_HIGHMEM_64G is enabled which isn't a very
> interesting configuration so fixing this is low priority.
Who said?
allmodconfig is commonly used for compilation coverage testing.
> I do have a patch but it needs more testing.
Please don't merge this code into mainline until that is completed.
> > I'm reluctant to convert it to do_div() because I can't conceive of any
> > situation in which it makes sense to perform division on a physical bus
> > address :(
>
> I don't understand this comment. Can you think of a way to implement
> this in another way?
Well I don't know what the code's trying to do. It takes a physical
bus address then rounds it down to some multiple of `max_packet'. I
can't see how that makes sense in terms of physical hardware unless
perhaps max_packet is a power of two, in which case all this code
should be using ands, ors and shifts.
--
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
