Re: [PATCH 2/2] hw/usb/hcd-xhci.c: allow unaligned access to Capability Registers

[Tomoyuki Hirose]

On Tue, Dec 12, 2023 at 7:26 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Tue, 12 Dec 2023 at 01:43, Tomoyuki Hirose
> <tomoyuki.hirose@igel.co.jp> wrote:
> >
> > Thanks for comment.
> >
> > On Mon, Dec 11, 2023 at 10:57 PM Peter Maydell <peter.maydell@linaro.org> 
> > wrote:
> > > We should definitely look at fixing the unaligned access
> > > stuff, but the linked bug report is not trying to do an
> > > unaligned access -- it wants to do a 2-byte read from offset 2,
> > > which is aligned. The capability registers in the xHCI spec
> > > are also all at offsets and sizes that mean that a natural
> > > read of them is not unaligned.
> >
> > Shouldn't I link this bug report?
> > Or is it not appropriate to allow unaligned access?
>
> The bug report is definitely relevant. But depending
> on how tricky the unaligned access handling turns out to
> be to get right, we might be able to fix the bug by
> permitting aligned-but-not-4-bytes accesses. (I'm
> a bit surprised that doesn't work already, in fact:
> we use it in other devices.)
>
> thanks
> -- PMM

Thank you for answering my question.
The unaligned access handling of my patch is not so tricky.
If the access is unaligned, just correct the access size
and address and read the value as before.
Also, it is allowed by the specifications, and byte access
was possible even on real devices.

Regards,
Tomoyuki HIROSE