nordic-dev.net/nixpkgs
nordic-dev.net/nixpkgs
2
0
Fork 0
mirror of https://github.com/NixOS/nixpkgs.git synced 2024-12-29 17:14:33 +00:00
Code Issues Packages Projects Releases Wiki Activity
c86cd65e9b
nixpkgs/nixos/modules/hardware/video/nvidia.nix

365 lines
15 KiB
Nix
Raw Normal View History

Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
# This module provides the proprietary NVIDIA X11 / OpenGL drivers.
treewide: remove unused variables (#63177) * treewide: remove unused variables * making ofborg happy
2019-06-16 19:59:06 +00:00
{ config, lib, pkgs, ... }:
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
with lib;
let
nixos: hardware.nvidia.package option for selecting nvidia package Make it possible to select a derivation to use. This also moves some logic that belongs to the derivation back into it.
2018-11-03 18:50:12 +00:00
nvidia_x11 = let
drivers = config.services.xserver.videoDrivers;
isDeprecated = str: (hasPrefix "nvidia" str) && (str != "nvidia");
hasDeprecated = drivers: any isDeprecated drivers;
in if (hasDeprecated drivers) then
throw ''
Selecting an nvidia driver has been modified for NixOS 19.03. The version is now set using `hardware.nvidia.package`.
''
else if (elem "nvidia" drivers) then cfg.package else null;
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nixos/nvidia: use correct 32-bit legacy libs; refactor Fixes #6127. Now we avoid multiple switches among the version possibilities.
2015-02-21 15:29:36 +00:00
enabled = nvidia_x11 != null;
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
cfg = config.hardware.nvidia;
nixos/nvidia : added nvidia-persistenced
2020-03-09 08:47:57 +00:00
nixos/nvidia: prime.sync.{intel,nvidia}BusId -> prime.{intel,nvidia}BusId
2019-08-14 12:18:28 +00:00
pCfg = cfg.prime;
syncCfg = pCfg.sync;
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
offloadCfg = pCfg.offload;
primeEnabled = syncCfg.enable || offloadCfg.enable;
nixos/nvidia : added nvidia-persistenced
2020-03-09 08:47:57 +00:00
nvidiaPersistencedEnabled = cfg.nvidiaPersistenced;
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
in
{
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
imports =
[
(mkRenamedOptionModule [ "hardware" "nvidia" "optimus_prime" "enable" ] [ "hardware" "nvidia" "prime" "sync" "enable" ])
(mkRenamedOptionModule [ "hardware" "nvidia" "optimus_prime" "allowExternalGpu" ] [ "hardware" "nvidia" "prime" "sync" "allowExternalGpu" ])
nixos/nvidia: prime.sync.{intel,nvidia}BusId -> prime.{intel,nvidia}BusId
2019-08-14 12:18:28 +00:00
(mkRenamedOptionModule [ "hardware" "nvidia" "optimus_prime" "nvidiaBusId" ] [ "hardware" "nvidia" "prime" "nvidiaBusId" ])
(mkRenamedOptionModule [ "hardware" "nvidia" "optimus_prime" "intelBusId" ] [ "hardware" "nvidia" "prime" "intelBusId" ])
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
];
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
options = {
nixos/nvidia: include systemd power management
2019-11-16 19:27:16 +00:00
hardware.nvidia.powerManagement.enable = mkOption {
type = types.bool;
default = false;
description = ''
Experimental power management through systemd. For more information, see
the NVIDIA docs, on Chapter 21. Configuring Power Management Support.
'';
};
nVidia: Support dynamic (fine-grained) power management
2020-10-14 15:17:53 +00:00
hardware.nvidia.powerManagement.finegrained = mkOption {
type = types.bool;
default = false;
description = ''
Experimental power management of PRIME offload. For more information, see
the NVIDIA docs, chapter 22. PCI-Express runtime power management.
'';
};
nixos/nvidia: remove lib references
2019-08-14 02:39:26 +00:00
hardware.nvidia.modesetting.enable = mkOption {
type = types.bool;
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
default = false;
description = ''
Enable kernel modesetting when using the NVIDIA proprietary driver.
Enabling this fixes screen tearing when using Optimus via PRIME (see
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
<option>hardware.nvidia.prime.sync.enable</option>. This is not enabled
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
by default because it is not officially supported by NVIDIA and would not
work with SLI.
'';
};
nixos/nvidia: prime.sync.{intel,nvidia}BusId -> prime.{intel,nvidia}BusId
2019-08-14 12:18:28 +00:00
hardware.nvidia.prime.nvidiaBusId = mkOption {
type = types.str;
default = "";
example = "PCI:1:0:0";
description = ''
Bus ID of the NVIDIA GPU. You can find it using lspci; for example if lspci
shows the NVIDIA GPU at "01:00.0", set this option to "PCI:1:0:0".
'';
};
hardware.nvidia.prime.intelBusId = mkOption {
type = types.str;
default = "";
example = "PCI:0:2:0";
description = ''
Bus ID of the Intel GPU. You can find it using lspci; for example if lspci
shows the Intel GPU at "00:02.0", set this option to "PCI:0:2:0".
'';
};
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
hardware.nvidia.prime.amdgpuBusId = mkOption {
type = types.str;
default = "";
example = "PCI:4:0:0";
description = ''
Bus ID of the AMD APU. You can find it using lspci; for example if lspci
shows the AMD APU at "04:00.0", set this option to "PCI:4:0:0".
'';
};
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
hardware.nvidia.prime.sync.enable = mkOption {
nixos/nvidia: remove lib references
2019-08-14 02:39:26 +00:00
type = types.bool;
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
default = false;
description = ''
Enable NVIDIA Optimus support using the NVIDIA proprietary driver via PRIME.
If enabled, the NVIDIA GPU will be always on and used for all rendering,
while enabling output to displays attached only to the integrated Intel GPU
without a multiplexer.
Note that this option only has any effect if the "nvidia" driver is specified
in <option>services.xserver.videoDrivers</option>, and it should preferably
be the only driver there.
If this is enabled, then the bus IDs of the NVIDIA and Intel GPUs have to be
nixos/nvidia: prime.sync.{intel,nvidia}BusId -> prime.{intel,nvidia}BusId
2019-08-14 12:18:28 +00:00
specified (<option>hardware.nvidia.prime.nvidiaBusId</option> and
<option>hardware.nvidia.prime.intelBusId</option>).
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
If you enable this, you may want to also enable kernel modesetting for the
NVIDIA driver (<option>hardware.nvidia.modesetting.enable</option>) in order
to prevent tearing.
Note that this configuration will only be successful when a display manager
for which the <option>services.xserver.displayManager.setupCommands</option>
nixos/slim: remove The SLIM project is abandoned and their last release was in 2013. Because of this it poses a security risk to systems, no one is working on it or picked up maintenance. It also lacks compatibility with systemd and logind sessions. For users, there liikely isn't anything like slim that's as lightweight in terms of dependencies.
2019-11-11 21:18:23 +00:00
option is supported is used.
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
'';
};
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
hardware.nvidia.prime.sync.allowExternalGpu = mkOption {
nixos/nvidia: remove lib references
2019-08-14 02:39:26 +00:00
type = types.bool;
nixos/nvidia: Add NVIDIA optimus option to allow external GPUs Without this option - NVIDIA refuses to use an external GPU.
2019-04-18 22:51:35 +00:00
default = false;
description = ''
Configure X to allow external NVIDIA GPUs when using optimus.
'';
};
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
hardware.nvidia.prime.offload.enable = mkOption {
type = types.bool;
default = false;
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
description = ''
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
Enable render offload support using the NVIDIA proprietary driver via PRIME.
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
If this is enabled, then the bus IDs of the NVIDIA and Intel GPUs have to be
specified (<option>hardware.nvidia.prime.nvidiaBusId</option> and
<option>hardware.nvidia.prime.intelBusId</option>).
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
'';
};
nixos/nvidia : added nvidia-persistenced
2020-03-09 08:47:57 +00:00
hardware.nvidia.nvidiaPersistenced = mkOption {
default = false;
type = types.bool;
description = ''
Update for NVIDA GPU headless mode, i.e. nvidia-persistenced. It ensures all
GPUs stay awake even during headless mode.
'';
};
nixos: hardware.nvidia.package option for selecting nvidia package Make it possible to select a derivation to use. This also moves some logic that belongs to the derivation back into it.
2018-11-03 18:50:12 +00:00
hardware.nvidia.package = lib.mkOption {
type = lib.types.package;
default = config.boot.kernelPackages.nvidiaPackages.stable;
defaultText = "config.boot.kernelPackages.nvidiaPackages.stable";
description = ''
The NVIDIA X11 derivation to use.
'';
example = "config.boot.kernelPackages.nvidiaPackages.legacy340";
};
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
};
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
config = let
igpuDriver = if pCfg.intelBusId != "" then "modesetting" else "amdgpu";
igpuBusId = if pCfg.intelBusId != "" then pCfg.intelBusId else pCfg.amdgpuBusId;
in mkIf enabled {
nixos/gdm,nvidia: new options to enable GDM on Wayland and disabling it for nvidia drivers.
2017-05-07 20:27:02 +00:00
assertions = [
nixos/nvidia: assert that gdm.nvidiaWayland needs modesetting
2019-10-03 00:07:19 +00:00
{
assertion = with config.services.xserver.displayManager; gdm.nvidiaWayland -> cfg.modesetting.enable;
message = "You cannot use wayland with GDM without modesetting enabled for NVIDIA drivers, set `hardware.nvidia.modesetting.enable = true`";
}
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
{
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
assertion = primeEnabled -> pCfg.intelBusId == "" || pCfg.amdgpuBusId == "";
message = ''
You cannot configure both an Intel iGPU and an AMD APU. Pick the one corresponding to your processor.
'';
}
{
assertion = primeEnabled -> pCfg.nvidiaBusId != "" && (pCfg.intelBusId != "" || pCfg.amdgpuBusId != "");
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
message = ''
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
When NVIDIA PRIME is enabled, the GPU bus IDs must configured.
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
'';
}
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
{
assertion = offloadCfg.enable -> versionAtLeast nvidia_x11.version "435.21";
message = "NVIDIA PRIME render offload is currently only supported on versions >= 435.21.";
}
{
assertion = !(syncCfg.enable && offloadCfg.enable);
message = "Only one NVIDIA PRIME solution may be used at a time.";
}
nVidia: Support dynamic (fine-grained) power management
2020-10-14 15:17:53 +00:00
{
assertion = !(syncCfg.enable && cfg.powerManagement.finegrained);
message = "Sync precludes powering down the NVIDIA GPU.";
}
{
assertion = cfg.powerManagement.enable -> offloadCfg.enable;
message = "Fine-grained power management requires offload to be enabled.";
}
nixos/gdm,nvidia: new options to enable GDM on Wayland and disabling it for nvidia drivers.
2017-05-07 20:27:02 +00:00
];
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
# If Optimus/PRIME is enabled, we:
# - Specify the configured NVIDIA GPU bus ID in the Device section for the
# "nvidia" driver.
# - Add the AllowEmptyInitialConfiguration option to the Screen section for the
# "nvidia" driver, in order to allow the X server to start without any outputs.
# - Add a separate Device section for the Intel GPU, using the "modesetting"
# driver and with the configured BusID.
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
# - OR add a separate Device section for the AMD APU, using the "amdgpu"
# driver and with the configures BusID.
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
# - Reference that Device section from the ServerLayout section as an inactive
# device.
# - Configure the display manager to run specific `xrandr` commands which will
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
# configure/enable displays connected to the Intel iGPU / AMD APU.
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
services.xserver.useGlamor = mkDefault offloadCfg.enable;
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
services.xserver.drivers = let
in optional primeEnabled {
name = igpuDriver;
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
display = offloadCfg.enable;
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
modules = optional (igpuDriver == "amdgpu") [ pkgs.xorg.xf86videoamdgpu ];
nixos/xserver: introduce attr display to xserver.drivers Specifically for NVIDIA so that only the device section would be created
2019-08-14 16:37:03 +00:00
deviceSection = ''
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
BusID "${igpuBusId}"
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
${optionalString syncCfg.enable ''Option "AccelMethod" "none"''}
nixos/xserver: introduce attr display to xserver.drivers Specifically for NVIDIA so that only the device section would be created
2019-08-14 16:37:03 +00:00
'';
} ++ singleton {
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
name = "nvidia";
modules = [ nvidia_x11.bin ];
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
display = !offloadCfg.enable;
deviceSection = optionalString primeEnabled
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
''
nixos/nvidia: prime.sync.{intel,nvidia}BusId -> prime.{intel,nvidia}BusId
2019-08-14 12:18:28 +00:00
BusID "${pCfg.nvidiaBusId}"
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
${optionalString syncCfg.allowExternalGpu "Option \"AllowExternalGpus\""}
nVidia: Support dynamic (fine-grained) power management
2020-10-14 15:17:53 +00:00
${optionalString cfg.powerManagement.finegrained "Option \"NVreg_DynamicPowerManagement=0x02\""}
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
'';
screenSection =
''
Option "RandRRotation" "on"
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
${optionalString syncCfg.enable "Option \"AllowEmptyInitialConfiguration\""}
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
'';
};
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nixos/xserver: introduce attr display to xserver.drivers Specifically for NVIDIA so that only the device section would be created
2019-08-14 16:37:03 +00:00
services.xserver.serverLayoutSection = optionalString syncCfg.enable ''
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
Inactive "Device-${igpuDriver}[0]"
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
'' + optionalString offloadCfg.enable ''
Option "AllowNVIDIAGPUScreens"
nixos/xserver: introduce attr display to xserver.drivers Specifically for NVIDIA so that only the device section would be created
2019-08-14 16:37:03 +00:00
'';
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nixos/nvidia: optimus_prime -> prime.sync
2019-08-14 12:15:43 +00:00
services.xserver.displayManager.setupCommands = optionalString syncCfg.enable ''
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
# Added by nvidia configuration module for Optimus/PRIME.
nvidia: Add AMD APU option for Prime
2020-10-14 14:24:48 +00:00
${pkgs.xorg.xrandr}/bin/xrandr --setprovideroutputsource ${igpuDriver} NVIDIA-0
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
${pkgs.xorg.xrandr}/bin/xrandr --auto
'';
nvidia: support application profiles Closes #22666.
2017-02-11 14:11:18 +00:00
environment.etc."nvidia/nvidia-application-profiles-rc" = mkIf nvidia_x11.useProfiles {
source = "${nvidia_x11.bin}/share/nvidia/nvidia-application-profiles-rc";
};
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
hardware.opengl.package = mkIf (!offloadCfg.enable) nvidia_x11.out;
nixos: hardware.nvidia.package option for selecting nvidia package Make it possible to select a derivation to use. This also moves some logic that belongs to the derivation back into it.
2018-11-03 18:50:12 +00:00
hardware.opengl.package32 = mkIf (!offloadCfg.enable) nvidia_x11.lib32;
nixos/nvidia: implement prime render offload
2019-08-14 16:41:15 +00:00
hardware.opengl.extraPackages = optional offloadCfg.enable nvidia_x11.out;
nixos: hardware.nvidia.package option for selecting nvidia package Make it possible to select a derivation to use. This also moves some logic that belongs to the derivation back into it.
2018-11-03 18:50:12 +00:00
hardware.opengl.extraPackages32 = optional offloadCfg.enable nvidia_x11.lib32;
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nvidia module: allow .persistenced == null Fixes #26250. This is fallout from PR #22304. It's null for 304 and 173 legacy drivers.
2017-05-30 18:56:34 +00:00
environment.systemPackages = [ nvidia_x11.bin nvidia_x11.settings ]
nixos/nvidia: fix optionals usage
2020-10-18 18:42:05 +00:00
++ optionals nvidiaPersistencedEnabled [ nvidia_x11.persistenced ];
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
nixos/nvidia: include systemd power management
2019-11-16 19:27:16 +00:00
systemd.packages = optional cfg.powerManagement.enable nvidia_x11.out;
systemd.services = let
baseNvidiaService = state: {
description = "NVIDIA system ${state} actions";
path = with pkgs; [ kbd ];
serviceConfig = {
Type = "oneshot";
ExecStart = "${nvidia_x11.out}/bin/nvidia-sleep.sh '${state}'";
};
};
nvidiaService = sleepState: (baseNvidiaService sleepState) // {
before = [ "systemd-${sleepState}.service" ];
requiredBy = [ "systemd-${sleepState}.service" ];
};
services = (builtins.listToAttrs (map (t: nameValuePair "nvidia-${t}" (nvidiaService t)) ["hibernate" "suspend"]))
// {
nvidia-resume = (baseNvidiaService "resume") // {
after = [ "systemd-suspend.service" "systemd-hibernate.service" ];
requiredBy = [ "systemd-suspend.service" "systemd-hibernate.service" ];
};
};
in optionalAttrs cfg.powerManagement.enable services
// optionalAttrs nvidiaPersistencedEnabled {
"nvidia-persistenced" = mkIf nvidiaPersistencedEnabled {
description = "NVIDIA Persistence Daemon";
wantedBy = [ "multi-user.target" ];
serviceConfig = {
Type = "forking";
Restart = "always";
PIDFile = "/var/run/nvidia-persistenced/nvidia-persistenced.pid";
ExecStart = "${nvidia_x11.persistenced}/bin/nvidia-persistenced --verbose";
ExecStopPost = "${pkgs.coreutils}/bin/rm -rf /var/run/nvidia-persistenced";
};
};
};
nixos/docker: add enableNvidia option
2019-02-27 08:44:37 +00:00
systemd.tmpfiles.rules = optional config.virtualisation.docker.enableNvidia
"L+ /run/nvidia-docker/bin - - - - ${nvidia_x11.bin}/origBin"
++ optional (nvidia_x11.persistenced != null && config.virtualisation.docker.enableNvidia)
"L+ /run/nvidia-docker/extras/bin/nvidia-persistenced - - - - ${nvidia_x11.persistenced}/origBin/nvidia-persistenced";
linuxPackages.nvidia_x11: refactor, build more from source * Use libglvnd; * Compile nvidia-settings, nvidia-persistenced from source; * Generalize builder.
2017-01-29 21:29:39 +00:00
boot.extraModulePackages = [ nvidia_x11.bin ];
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
Create /dev/nvidia-uvm for CUDA support
2015-03-03 19:15:38 +00:00
# nvidia-uvm is required by CUDA applications.
nixos/nvidia: populating /dev with nvidia devices at boot
2017-05-07 19:37:56 +00:00
boot.kernelModules = [ "nvidia-uvm" ] ++
nixos/nvidia: remove lib references
2019-08-14 02:39:26 +00:00
optionals config.services.xserver.enable [ "nvidia" "nvidia_modeset" "nvidia_drm" ];
nixos/nvidia: populating /dev with nvidia devices at boot
2017-05-07 19:37:56 +00:00
nixos/xserver: Implement configuration of NVIDIA Optimus via PRIME This adds configuration options which automate the configuration of NVIDIA Optimus using PRIME. This allows using the NVIDIA proprietary driver on Optimus laptops, in order to render using the NVIDIA GPU while outputting to displays connected only to the integrated Intel GPU. It also adds an option for enabling kernel modesetting for the NVIDIA driver (via a kernel command line flag); this is particularly useful together with Optimus/PRIME because it fixes tearing on PRIME-connected screens. The user still needs to enable the Optimus/PRIME feature and specify the bus IDs of the Intel and NVIDIA GPUs, but this is still much easier for users and more reliable. The implementation handles both the X configuration file as well as getting display managers to run certain necessary `xrandr` commands just after X has started. Configuration of commands run after X startup is done using a new configuration option `services.xserver.displayManager.setupCommands`. Support for this option is implemented for LightDM, GDM and SDDM; all of these have been tested with this feature including logging into a Plasma session. Note: support of `setupCommands` for GDM is implemented by making GDM run the session executable via a wrapper; the wrapper will run the `setupCommands` before execing. This seemed like the simplest and most reliable approach, and solves running these commands both for GDM's X server and user X servers (GDM starts separate X servers for itself and user sessions). An alternative approach would be with autostart files but that seems harder to set up and less reliable. Note that some simple features for X configuration file generation (in `xserver.nix`) are added which are used in the implementation: - `services.xserver.extraConfig`: Allows adding arbitrary new sections. This is used to add the Device section for the Intel GPU. - `deviceSection` and `screenSection` within `services.xserver.drivers`. This allows the nvidia configuration module to add additional contents into the `Device` and `Screen` sections of the "nvidia" driver, and not into such sections for other drivers that may be enabled.
2018-06-30 07:33:45 +00:00
# If requested enable modesetting via kernel parameter.
nixos/nvidia: include systemd power management
2019-11-16 19:27:16 +00:00
boot.kernelParams = optional (offloadCfg.enable || cfg.modesetting.enable) "nvidia-drm.modeset=1"
++ optional cfg.powerManagement.enable "nvidia.NVreg_PreserveVideoMemoryAllocations=1";
Create /dev/nvidia-uvm for CUDA support
2015-03-03 19:15:38 +00:00
services.udev.extraRules =
''
nVidia: Support dynamic (fine-grained) power management
2020-10-14 15:17:53 +00:00
# Create /dev/nvidia-uvm when the nvidia-uvm module is loaded.
nvidia: fix $ substitution in udev rules This fixes a harmless error from systemd-udevd that looks like: Dec 23 15:35:23 dellbook systemd-udevd[696]: /nix/store/iixya3ni5whybpq9zz1h7f4pyw7nhd19-udev-rules/99-local.rules:25 Invalid value "..." for RUN (char 101: invalid substitution type), ignoring, but please fix it. Using $$ fixes it using the escaping documented at https://www.freedesktop.org/software/systemd/man/udev.html.
2019-12-26 17:52:03 +00:00
KERNEL=="nvidia", RUN+="${pkgs.runtimeShell} -c 'mknod -m 666 /dev/nvidiactl c $$(grep nvidia-frontend /proc/devices | cut -d \ -f 1) 255'"
KERNEL=="nvidia_modeset", RUN+="${pkgs.runtimeShell} -c 'mknod -m 666 /dev/nvidia-modeset c $$(grep nvidia-frontend /proc/devices | cut -d \ -f 1) 254'"
KERNEL=="card*", SUBSYSTEM=="drm", DRIVERS=="nvidia", RUN+="${pkgs.runtimeShell} -c 'mknod -m 666 /dev/nvidia%n c $$(grep nvidia-frontend /proc/devices | cut -d \ -f 1) %n'"
KERNEL=="nvidia_uvm", RUN+="${pkgs.runtimeShell} -c 'mknod -m 666 /dev/nvidia-uvm c $$(grep nvidia-uvm /proc/devices | cut -d \ -f 1) 0'"
nixos/nvidia: add nvidia-uvm-tools device Related to https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-390/+bug/1767777
2019-12-30 18:15:10 +00:00
KERNEL=="nvidia_uvm", RUN+="${pkgs.runtimeShell} -c 'mknod -m 666 /dev/nvidia-uvm-tools c $$(grep nvidia-uvm /proc/devices | cut -d \ -f 1) 0'"
nVidia: Support dynamic (fine-grained) power management
2020-10-14 15:17:53 +00:00
'' + optionalString cfg.powerManagement.finegrained ''
# Remove NVIDIA USB xHCI Host Controller devices, if present
ACTION=="add", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x0c0330", ATTR{remove}="1"
# Remove NVIDIA USB Type-C UCSI devices, if present
ACTION=="add", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x0c8000", ATTR{remove}="1"
# Remove NVIDIA Audio devices, if present
ACTION=="add", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x040300", ATTR{remove}="1"
# Enable runtime PM for NVIDIA VGA/3D controller devices on driver bind
ACTION=="bind", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x030000", TEST=="power/control", ATTR{power/control}="auto"
ACTION=="bind", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x030200", TEST=="power/control", ATTR{power/control}="auto"
# Disable runtime PM for NVIDIA VGA/3D controller devices on driver unbind
ACTION=="unbind", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x030000", TEST=="power/control", ATTR{power/control}="on"
ACTION=="unbind", SUBSYSTEM=="pci", ATTR{vendor}=="0x10de", ATTR{class}=="0x030200", TEST=="power/control", ATTR{power/control}="on"
Create /dev/nvidia-uvm for CUDA support
2015-03-03 19:15:38 +00:00
'';
nVidia: Support dynamic (fine-grained) power management
2020-10-14 15:17:53 +00:00
boot.extraModprobeConfig = mkIf cfg.powerManagement.finegrained ''
options nvidia "NVreg_DynamicPowerManagement=0x02"
'';
Move the NVIDIA support into its own module Previously all card-specific stuff was scattered across xserver.nix and opengl.nix, which is ugly. Now it can be kept together in a single card-specific module. This required the addition of a few internal options: - services.xserver.drivers: A list of { name, driverName, modules, libPath } sets. - hardware.opengl.package: The OpenGL implementation. Note that there can be only one OpenGL implementation at a time in a system configuration (i.e. no dynamic detection). - hardware.opengl.package32: The 32-bit OpenGL implementation.
2014-04-29 12:16:34 +00:00
boot.blacklistedKernelModules = [ "nouveau" "nvidiafb" ];
services.acpid.enable = true;
};
}
Reference in New Issue Copy Permalink