[PATCH 3/7] usb: gadget: pxa25x_udc: use readl/writel for mmio

[Arnd Bergmann]

On Friday 29 January 2016 21:03:40 Sergei Shtylyov wrote:
> On 01/29/2016 07:18 PM, Krzysztof Hałasa wrote:
> 
> >> The unclear part here is for IXP4xx, which supports both big-endian
> >> and little-endian configurations. So far, the driver has done
> >> no byteswap in either case. I suspect that is wrong and it would
> >> actually need to swap in one or the other case, but I don't know
> >> which.
> >
> > If at all, I guess it should swap in LE mode. But it's far from certain.
> >
> >> It's also possible that there is some magic setting in
> >> the chip that makes the endianess of the MMIO register match the
> >> CPU, and in that case, the code actually does the right thing
> >> for all configurations, both before and after this patch.
> >
> > This is IMHO most probable.
> >
> > Actually, the IXP4xx is "natural" in BE mode (except for PCI) and
> > normally in LE mode it's order-coherent, meaning 32-bit "integer"
> > accesses need no swapping, but 8-bit and (mostly unused) 16-bit
> > transfers need swapping.
> >
> > Anyway, I think readl()/writel() do the right thing: in BE mode they
> > swap PCI accesses and don't swap normal registers,
> 
>     Alas, readl()/writel() don't know what registers you are calling them for, 
> they were designed for PCI which is little-endian, so they will swap in BE mode.
> 

In general, that is correct, but as Krzysztof said, ixp4xx is rather
special here, and it implements its own readl/writel functions
that behave differently for PCI spaces than for on-chip addresses,
See arch/arm/mach-ixp4xx/include/mach/io.h.

This platform evidently does its own tricks with byte swapping so
while a normal big-endian platform has to swap on readl (but not
readsl), ixp4xx never does any swaps when CONFIG_IXP4XX_INDIRECT_PCI
is set (neither readl nor readsl, neither PCI nor on-chip MMU)

However if CONFIG_IXP4XX_INDIRECT_PCI is disabled (which I think
is almost always the case), it basically behaves like all the other
platforms, and in big-endian configurations performsm swaps for
inl/readl/ioread32 but not readsl/ioread32_rep/insl (it swaps
twice for the last one).

	Arnd