On Thu, 20 Nov 2025, Lucas De Marchi <[email protected]> wrote: > On Wed, Nov 19, 2025 at 08:52:52PM +0200, Jani Nikula wrote: >>Drop the #ifdef I915, and use the same intel_gmch_vga_set_decode() for >>both i915 and xe. > > I hope this doesn't regress our display side on other archs. We are very > close to having display working, but this messing with VGA is likely to > break it. > > See "drm/i915/display: Stop touching vga on post enable", which is > needed for xe to use a DG2/BMG with a raspberry pi (+pci/resources > branch + a few other patches).
I'm feeling confident [1] that the patch at hand does not have similar issues. This is about VGA arbitration on the PCI bridge, not about actually poking at VGA registers. I went ahead and merged the series. If there are any ill effects with the last patch, we can revert with a low bar. BR, Jani. [1] https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect -- Jani Nikula, Intel
