------------------------------------------------------------------- Fri Nov 22 23:36:28 UTC 2024 - RN - Added prime-run binary that just links process to the nvidia libraries ------------------------------------------------------------------- Mon Nov 18 16:44:53 UTC 2024 - RN - Added upstream GPU detection patch ------------------------------------------------------------------- Mon Nov 18 11:07:05 UTC 2024 - RN - Added patch for user groups - Removed `users`, `sudo`, `adm` and `wheel` groups - Added `video` group and separated `supergfxctl` ------------------------------------------------------------------- Thu Oct 03 16:16:22 UTC 2024 - RN - Update to version 5.2.4 release: * Update CHANGELOG.md * Fix dGPU detection for FA401WV ------------------------------------------------------------------- Sun Sep 22 06:50:15 UTC 2024 - RN - Update to version 5.2.4: * Bump version and deps * fix: check alternate path for egpu_enable flag * Bump deps and version * Bump deps * Remove the nvidia/false config stuff as too problematic * Further tweaks to asus dgpu safety check * asus egpu fixes * Update deps * Update changelog ------------------------------------------------------------------- Wed Jun 12 17:39:23 UTC 2024 - RN - Fixing update process bug ------------------------------------------------------------------- Thu Mar 28 19:59:05 UTC 2024 - RN - Update to 5.2.1 - Better sanity check for booting without egpu after it was previously set - Added Vulkan ICD profile switching ------------------------------------------------------------------- Thu Dec 28 16:37:31 UTC 2023 - RN - Switched to _service usage, because cargo_config is part of the vendor.tar now - Cleaned up .spec ------------------------------------------------------------------- Wed Dec 27 22:56:20 UTC 2023 - Jan Engelhardt - Avoid bashisms in build recipe ------------------------------------------------------------------- Tue Nov 14 05:46:34 UTC 2023 - RN - Update to version 5.1.2: * Bump deps * Add missing entry in changelog * Adjust makefile to use vender-filter * Ensure supergfx tries to fallback if change fails * Small fixes for egpu * Bump version * Update deps * Update deps * Fix: Don't continue with action list if the error is a logout timeout * Fix pipeline ------------------------------------------------------------------- Fri Nov 18 07:57:22 UTC 2022 - luke@ljones.dev - Update to version 5.0.1: * Add dgpu status notification on dbus * Fix: Check for GA401I, remove compute mode, set log level info * Release 5.0.0 * Prep tag RC15 * RC15, final * RC13 with various corrections * Update dependencies * Enable hotplug use, fixes to dgpu_disable, fixes to integrated * Do pci rescan even if no missing device * Update changelog * Switch to async with tokio. Use asus switches if avilable. Bugfixes. * Ignore possible i/o error as invalid * Bump zbus version * Add 99-nvidia-ac.rules udev rule * fix: correction in changelog * fix: make sure to use asus_use_dgpu_disable * Add pipeline, cleanup with clippy * 5.0.0-rc1 * Update dbus API docs (introspectable) * Add more debug logging, default service to debug for now * Switch to async runtime * Slap Arc> around gpu device * Switch to udev. Add ASUS methods. Optimise. * Remove dedicated mode * Added pipe to sh command to install Rust Taken from the official Rust website -- https://www.rust-lang.org/tools/install * fix AMD detection for pci-ids version 2020.03.20 ------------------------------------------------------------------- Tue Jul 05 08:23:07 UTC 2022 - luke@ljones.dev - Update to version 4.0.5: * Prep new release * Fix interaction with newer lspci version * add "Quadro" to the lspci parsing for NVIDIA cards ------------------------------------------------------------------- Wed May 11 16:17:34 UTC 2022 - Enrico Belleri - Use obs services for source and vendor ------------------------------------------------------------------- Fri Feb 4 22:35:35 UTC 2022 - Luke Jones - Version 4.0.4 - Adjust the kernel cmdline arg code path ------------------------------------------------------------------- Thu Feb 3 21:37:52 UTC 2022 - Luke Jones - Version 4.0.3 - Add config option `no_logind`: Don't use logind to see if all sessions are logged out and therefore safe to change mode. This will be useful for people not using a login manager, however it is not guaranteed to work unless all graphical sessions are ended and nothing is hooking the drivers. Ignored if `always_reboot` is set. - Add config option `logout_timeout_s`: The timeout in seconds to wait for all user graphical sessions to end. Default is 3 minutes, 0 = infinite. Ignored if `no_logind` or `always_reboot` is set. - Add new dbus method: `PendingMode`, to check if a mode change is required - Add new dbus method: `PendingUserAction`, to check if the user is required to perform an action - Add new dbus method: `Config`, to get the current base config - Add new dbus method: `SetConfig`, to set the base config - Add `-p, --pend-action` CLI arg to get the pending user action if any - Add `-P, --pend-mode` CLI arg to get the pending mode change if any` - Add ability to read `supergfxd.mode=` from kernel cmdline on startup and set the mode appropriately - CLI option `--force` was unused, it is now removed. ------------------------------------------------------------------- Sat Jan 22 09:35:11 UTC 2022 - Luke Jones - Version 4.0.2 ------------------------------------------------------------------- Wed Jan 19 23:12:03 UTC 2022 - Luke Jones - Version 4.0.1 - Fix version upgrade of config - Recreate the config if parsing fails - Only write the (reboot) mode change to config file, don't update live config - AMD dedicated + hybrid config for xorg - "AllowExternalGpus" added to xorg for Nvidia Egpu mode ------------------------------------------------------------------- Tue Jan 18 07:17:49 UTC 2022 - Luke Jones - Version 4.0.0 - Added - Add new dbus method: `Version` to get supergfxd version - Add new dbus method: `Vendor` to get dGPU vendor name - Add new dbus method: `Supported` to get list of supported modes - Add `-v, --version` CLI arg to get supergfxd version - Add `-V, --vendor` CLI arg to get dGPU vendor name - Add `-s, --supported` CLI arg to get list of supported modes - Add new config option: `vfio_save` to reload VFIO on boot - Add new config option: `compute_save` to reload compute on boot - Add new config option: `always_reboot` reboot to change modes - Changed - Adjust startup to check for ASUS eGPU and dGPU enablement if no modes supported - If nvidia-drm.modeset=1 is set then save mode and require a reboot by default\ - Add extra check for Nvidia dGPU (fixes Flow 13") - Properly check the correct device for power status - Breaking - Rename Vendor, GetVendor to Mode, GetMode to better reflect their results ------------------------------------------------------------------- Mon Jan 10 00:37:46 UTC 2022 - Luke Jones - Version 3.0.0 Sun Dec 19 07:43:11 UTC 2021 - Luke Jones - Version 2.0.5 ------------------------------------------------------------------- Tue Nov 2 20:23:58 UTC 2021 - Luke Jones - Version 2.0.4 ------------------------------------------------------------------- Tue Nov 2 03:56:07 UTC 2021 - Luke Jones - Version 2.0.3