Xorg doesn't start and some other issues with the RC1 of kernel 6.10

[Greg KH]

On Fri, May 31, 2024 at 11:19:34AM +0200, Thorsten Leemhuis wrote:
> On 31.05.24 11:03, Michael Ellerman wrote:
> > Michael Ellerman <mpe at ellerman.id.au> writes:
> >> Christian Zigotzky <chzigotzky at xenosoft.de> writes:
> >>> On 28.05.24 22:00, Christian Zigotzky wrote:
> >>>> Hi All,
> >>>>
> >>>> Xorg doesn't start anymore since the RC1 of kernel 6.10. We tested it 
> >>>> with the VirtIO GPU and with some Radeon cards.
> >>>>
> >>>> Another error message: Failed to start Setup Virtual Console.
> >>>>
> >>>> Maybe this is the issue: + CONFIG_ARCH_HAS_KERNEL_FPU_SUPPORT=y
> >>>>
> >>>> Tested with FSL P5040, FSL P5020, and PASEMI boards.
> >>>>
> >>>> Could you please test Xorg on your PowerPC machines?
> >>>>
> >>>> Thanks,
> >>>> Christian
> >>> I tested the RC1 in a virtual e5500 QEMU PowerPC machine with Bochs VGA 
> >>> (-device VGA,vgamem_mb=256) and Xorg doesn't start either.
> >>>
> >>> Error message: xf86OpenConsole: KDSETMODE KD_GRAPHICS failed 
> >>> Inappropriate ioctl for device.
> >>
> >> That is presumably because of this:
> >>   https://lore.kernel.org/all/0da9785e-ba44-4718-9d08-4e96c1ba7ab2@kernel.org/
> > 
> > Attempting to regzbot this.
> > 
> > #regzbot introduced: 8c467f330059
> > #regzbot monitor: https://lore.kernel.org/all/0da9785e-ba44-4718-9d08-4e96c1ba7ab2@kernel.org/
> 
> Thx, I already had an eye on this, but thought tracking would not be
> needed, as Greg (now CCed) wanted to revert 8c467f3300591a ("VT: Use
> macros to define ioctls") two days ago:
> https://lore.kernel.org/all/2024052901-police-trash-e9f9@gregkh/
> 
> But that commit is not yet in -next afaics. :-/
> 
> /me meanwhile wonders if it would be wise to fix this before -rc2

I do, sorry, been traveling this week with geen vrije tijd.  Will get to
it tomorrow.

greg k-h