On Tue, 16 Jul 2024 at 15:37, Philippe Mathieu-Daudé <phi...@linaro.org> wrote: > > On 16/7/24 12:35, Peter Maydell wrote: > > On Sat, 13 Jul 2024 at 17:04, SamJakob <m...@samjakob.com> wrote: > >> > >> It is common practice when implementing double-buffering > >> on VideoCore to do so by multiplying the height of the > >> virtual buffer by the number of virtual screens desired > >> (i.e., two - in the case of double-bufferring). > >> > >> At present, this won't work in QEMU because the logic in > >> fb_use_offsets require that both the virtual width and > >> height exceed their physical counterparts. > >> > >> This appears to be unintentional/a typo and indeed the > >> comment states; "Experimentally, the hardware seems to > >> do this only if the viewport size is larger than the > >> physical screen". The viewport/virtual size would be > >> larger than the physical size if either virtual dimension > >> were larger than their physical counterparts and not > >> necessarily both. > >> > >> Signed-off-by: SamJakob <m...@samjakob.com> > > > > Thanks for this bugfix; I've applied it to my target-arm.next > > queue and it should get upstream within a week or so. > > Since I'm seeing 2 times the same patch, adding R-b again on > this one: > > Reviewed-by: Philippe Mathieu-Daudé <phi...@linaro.org> > > (BTW Peter the commit description is mis-aligned, if you > don't mind, correcting it while applying would be appreciated!)
Yep, I already picked up your r-by and rewrapped the commit message. -- PMM
