scan only cards and not outputs, safeguard against null renderer #106
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
no need to scan for card[0-9]* and probe card0-eDP etc if they are kms, bootvga and rendernodes etc. skip the wildcard and remove a unused size_t variable.
setting certain env vars to force egl implentations makes the render creation fail on the second gpu. instead of causing a coredump, safeguard commitState and let the monitor turn blank instead.
|
After this commit I can no longer start Hyprland in a virtual machine. In case this helps: |
Doubt it, both are links to same device, card0 exist. logs?, nvm didnt scroll right enough yeah indeed, 00:08.0/drm/card0/card0-Virtual-1 |
Technically virtual1 should be your "output" yeah, logs? |
ah so i was focusing on the wrong commit out of the two that merged, you are hitting the other one.
|
UjinT34
pushed a commit
to UjinT34/aquamarine
that referenced
this pull request
Jan 9, 2025
…hyprwm#106) * drm: dont scan card outputs no need to scan for card[0-9]* and probe card0-eDP etc if they are kms, bootvga and rendernodes etc. skip the wildcard and remove a unused size_t variable. * drm: dont commit state if renderer is missing setting certain env vars to force egl implentations makes the render creation fail on the second gpu. instead of causing a coredump, safeguard commitState and let the monitor turn blank instead.
vaxerski
added a commit
that referenced
this pull request
Jan 10, 2025
* do not set cursor planeProps unless smth changed * do not skip cursor state flag setting * drm: scan only cards and not outputs, safeguard against null renderer (#106) * drm: dont scan card outputs no need to scan for card[0-9]* and probe card0-eDP etc if they are kms, bootvga and rendernodes etc. skip the wildcard and remove a unused size_t variable. * drm: dont commit state if renderer is missing setting certain env vars to force egl implentations makes the render creation fail on the second gpu. instead of causing a coredump, safeguard commitState and let the monitor turn blank instead. * props: bump version to 0.5.0 * drm: Validate conn before dereference in CDRMAtomicRequest::commit() (#108) During startup, CDRMAtomicImpl::reset() may emit a call to method commit of a CDRMAtomicRequest instance with member "conn" uninitialized, leading to a segfault. Validate the the pointer before dereference it as a workaround. Fixes: 55ac962 ("DRM: preliminary atomic support") Closes: #107 Signed-off-by: Yao Zi <[email protected]> * buffer: remove useless forward def * drm: clearer flow when rescanning connectors (#113) * consolidates into checkOutput for clearer flow when rescanning connectors * add error log * drm: allow multigpu blit from explicit to implicit (#114) * version: bump to 0.5.1 * flake.lock: update * flake.nix: gcc13 -> gcc14 (#118) * drm: udev scan only drm_minor, not connectors (#119) * drm: log errno set by drmModeAtomicCommit (#120) * drm: moved null check for renderer to shouldBlit() (#109) (#121) * drm: only fail INVALID format when enabled (#122) * flake.lock: update * drm: only clear buffers when fullReconfigure succeeds (#124) * core/drm: Add HDR Support (#112) * version: bump to 0.6.0 * drm: limit udev drm_minor to Linux after a132fa4 (#129) Not implemented by libudev-devd yet: [ERR] [AQ] drm: No gpus in scanGPUs. [ERR] [AQ] drm: Found no gpus to use, cannot continue [ERR] [AQ] DRM Backend failed * do not set cursor planeProps unless smth changed * test separate cursor commits * do not change hdr blob unless asked to * rebase * split atomic commit processing and move hdr & colorspace into modeset * fix wide color gamut flag & cleanup * remove unused debug var --------- Signed-off-by: Yao Zi <[email protected]> Co-authored-by: Tom Englund <[email protected]> Co-authored-by: Vaxry <[email protected]> Co-authored-by: Ziyao <[email protected]> Co-authored-by: Ikalco <[email protected]> Co-authored-by: Mihai Fufezan <[email protected]> Co-authored-by: Austin Horstman <[email protected]> Co-authored-by: Richard Henninger <[email protected]> Co-authored-by: Jan Beich <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
seems pointless scanning the card[0-9]-eDP etc outputs for kms/bootvga/rendernodes just use the [0-9] cards, also safegaurd in commitstate for missing renderer prevents a coredump if __EGL_VENDOR_LIBRARY_FILENAMES is set to mesa on a laptop with dual gpus and a monitor attached to the nvidia dgpu. it might be a way to make this handled better but now it doesnt segfault atleast with a null dereference to renderer.