Re: [PATCH] drm/ssd130x: Set the address window in ssd132x_clear_screen()

[Fabio Piparo <holofermes@gmail.com>]

I can confirm that one of Sashiko's findings is correct, and it's not
limited to my use case: with the early return on !plane_state->crtc,
the clear in this path never runs at all.

I also have to retract the testing note. I did more testing, and the
blackout I had attributed to clear_screen during bring-up was
something else entirely, not the clear writing RAM. So "Exercised on
an SPI SSD1322" is wrong. The asymmetry itself is still valid I think
(the sibling ssd130x_clear_screen() sets the window, this one
doesn't), and the patch stands on inspection and compile testing
only. Happy to reword the commit message in a v2 if the cleanup is
still wanted, or drop it.

The other finding, the out-of-bounds read in ssd132x_update_rect(),
looks real too, but it's pre-existing and separate from this patch.

Thanks,
Fabio