Re: [PATCH v2 5/5] drm/i915: Introduce intel_bo_fbdev_bios_fb_takeover()
Maarten Lankhorst <[email protected]> Thu, 23 Jul 2026 15:37:02 +0200
| Newsgroups | org.freedesktop.lists.intel-gfx,org.freedesktop.lists.intel-xe |
|---|---|
| Message-ID | <[email protected]> |
Hey, On 7/23/26 11:06, Shankar, Uma wrote: > > >> -----Original Message----- >> From: Intel-xe <[email protected]> On Behalf Of Maarten >> Lankhorst >> Sent: Tuesday, July 21, 2026 7:56 PM >> To: [email protected] >> Cc: [email protected]; Maarten Lankhorst <[email protected]>; Ville >> Syrjälä <[email protected]> >> Subject: [PATCH v2 5/5] drm/i915: Introduce intel_bo_fbdev_bios_fb_takeover() >> >> Pull the "do we want to use stolen for the BIOS FB?" checks into a new >> intel_bo_fbdev_bios_fb_takeover() helper, and defer that decision until we're ready >> to reallocate the fbdev FB. This way even if we don't want to ultimately use the >> BIOS FB we'll keep the plane enabled until the replacement FB is ready. Should >> hopefully result in fewer display blinks during boot. >> >> Additionally, on some platforms this allows us to re-create the framebuffer when >> stolen memory is not suitable, using the blitter engine to copy the contents to >> system memory without causing any flickering by not preserving the stolen >> memory allocation. >> >> Based on a similar patch by Ville, that introduced intel_bo_fbdev_bios_fb_ok(). > > Looks Good to me. > Reviewed-by: Uma Shankar <[email protected]> > > Please get an ack from someone from Core KMD as well. > > @B S, Karthik Can we ask someone to check and give a Tested-by from validation > side for this series. I've been testing a bit locally, but I see a small flicker while transitioning from stolen to system memory. I need to figure out why that happens, as that should have been prevented by the call to drm_gem_plane_helper_prepare_fb() in prepare_plane_fb() before the atomic flip.