Fix crashes and improve hybrid mode handling in EnvyControl

[vjaykrsna]

This pull request makes several improvements to the GPU detection and mode switching logic in envycontrol.py, focusing on more robust handling of cache creation, error handling, and mode detection. The changes ensure better reliability when switching GPU modes, especially in hybrid configurations.

Improvements to GPU mode switching and cache management:

• Changed cache file creation logic in the adapter method so that the cache is only recreated when leaving hybrid mode, preventing unnecessary cache updates and handling cases where the GPU may not be visible due to old udev rules.

Error handling and robustness:

• Added a return None statement in get_amd_igpu_name() if the xrandr command fails, ensuring the function exits gracefully on error.

Mode detection logic:

• Updated get_current_mode() to consider the system to be in 'integrated' mode if any of the relevant files exist (using logical ORs), instead of requiring both BLACKLIST_PATH and one of the udev rules to exist, making the detection logic more accurate and flexible.

[Copilot]

Pull request overview

This PR improves EnvyControl’s robustness around GPU mode detection and hybrid-mode transitions by adjusting when the Nvidia PCI BusID cache is written, hardening AMD iGPU name detection, and loosening integrated-mode detection to handle partially-present config artifacts.

Changes:

• Return None from get_amd_igpu_name() when xrandr execution fails to avoid a crash path.
• Adjust CachedConfig.adapter() cache creation conditions to avoid unnecessary cache recreation when switching to hybrid.
• Update get_current_mode() to classify integrated when any integrated-mode marker file exists.

Comments suppressed due to low confidence (1)

envycontrol.py:658

• use_cache is computed before create_cache_file() runs. If the cache file didn’t exist initially, this path will create it but still skip read_cache_file() and won’t rebind get_nvidia_gpu_pci_bus, so later calls during the same run may still hit live lspci and fail if the GPU becomes hidden. Consider recomputing use_cache after cache creation, or directly setting use_cache = True / rebinding after create_cache_file() succeeds.

        use_cache = os.path.exists(CACHE_FILE_PATH)

        # Only cache the GPU PCI bus when leaving hybrid mode.
        # Switching *to* hybrid never needs the cached value and the GPU
        # may not be visible via lspci if old udev removal rules are still
        # in place, which would make create_cache_file() exit early.
        if self.is_hybrid() and self.app_args.switch != 'hybrid':
            self.create_cache_file()

        if use_cache:
            self.read_cache_file()  # might not be in hybrid mode

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

if self.is_hybrid() and self.app_args.switch != 'hybrid': will also run when --reset / --reset-sddm are used (since app_args.switch is None), causing an unnecessary cache (re)creation and potential failure on systems where the dGPU isn’t visible even though get_current_mode() reports hybrid. If the intent is “only when leaving hybrid mode”, it would be safer to gate on an explicit switch target (e.g., only when switch is 'integrated' or 'nvidia').

Suggested change

	if self.is_hybrid() and self.app_args.switch != 'hybrid':
	if self.is_hybrid() and self.app_args.switch in ('integrated', 'nvidia'):