Re: [PATCH for-9.2 00/53] arm: Drop deprecated boards

[Peter Maydell <peter.maydell@linaro.org>]

On Tue, 3 Sept 2024 at 17:55, Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> On 9/3/24 18:06, Peter Maydell wrote:
> > This patchset removes the various Arm machines which we deprecated
> > for the 9.0 release and are therefore allowed to remove for the 9.2
> > release:
> >   akita, borzoi, cheetah, connex, mainstone, n800, n810,
> >   spitz, terrier, tosa, verdex, z2
> > We get to drop over 30,000 lines of unmaintained code. So it's
> > a big patchset but it's almost all deletions.
> >
> > We have some command line options which were documented as only used
> > by the pxa2xx LCD display driver: -portrait and -rotate.  These
> > allowed the user to tell the display device to rotate its output by
> > 90/180/270 degrees (and the ui input layer to correspondingly rotate
> > mouse event coordinates to match).  I didn't realize these existed
> > when we deprecated the pxa2xx machines -- do we need a separate
> > deprecate-and-drop period to remove the command line options?  (If
> > so, I can drop the relevant patch from this series.)
>
> They are not specific to PXA; PXA LCD is the only one that supports
> passing it to the guest, but the logic is generic: if (for whatever
> reason) your VM generates output that is rotated, you can use the option
> to rotate mouse input.
>
> It's okay to remove it without deprecation notice, but also to keep it.
> Your choice, it's not a lot of code.

Well, the documentation for them says:
 -portrait       rotate graphical output 90 deg left (only PXA LCD)
 -rotate <deg>   rotate graphical output some deg left (only PXA LCD)

so the original intent was clearly (a) that the main effect
was rotation of the graphical output and (b) that these were
only effective for PXA.

The PXA display device doesn't pass anything through to the guest,
by the way -- it just draws the pixels in the guest framebuffer
in a different place in the UI window. As the FIXME comment in
pxa2xx_lcd.c notes, this should really have been done in common
code, not in a specific display driver.

As we both note, the UI input layer part *is* generic code so it
will do the rotation regardless of whether the display device is
also rotating the guest output. But that seems to me more of
an accident than an intentional feature.

I don't care very much because there's not that much code
involved. It's all separated out into patch 24 of this series:
https://patchew.org/QEMU/20240903160751.4100218-1-peter.maydell@linaro.org/20240903160751.4100218-25-peter.maydell@linaro.org/

-- PMM