Environment Variables

Normally, no environment variables need to be set. Most of the environment variables used by Mesa/Gallium are for debugging purposes, but they can sometimes be useful for debugging end-user issues.

LibGL environment variables

LIBGL_DEBUG

If defined debug information will be printed to stderr. If set to verbose additional information will be printed.

LIBGL_ALWAYS_INDIRECT

if set to true, forces an indirect rendering context/connection.

LIBGL_ALWAYS_SOFTWARE

if set to true, always use software rendering

LIBGL_NO_DRAWARRAYS

if set to true, do not use DrawArrays GLX protocol (for debugging)

LIBGL_DRI2_DISABLE

disable DRI2 if set to true.

LIBGL_DRI3_DISABLE

disable DRI3 if set to true.

LIBGL_KOPPER_DISABLE

disable Vulkan swapchains with Zink if set to true. In general, this should not be used unless you know what you are doing. Some examples of “knowing what you are doing” include: - using a VK driver which has no WSI implementation for your display server - profiling the DRI frontend against your VK driver’s WSI implementation

LIBGL_KOPPER_DRI2

disable DRI3 with Zink if set to true. In general, this should not be used unless you know what you are doing. Some examples of “knowing what you are doing” include: - running xrdp - using a VK driver which doesn’t support modifiers

Core Mesa environment variables

MESA_NO_ERROR

if set to 1, error checking is disabled as per GL_KHR_no_error. This will result in undefined behavior for invalid use of the API, but can reduce CPU use for apps that are known to be error free.

MESA_DEBUG

if set, error messages are printed to stderr. For example, if the application generates a GL_INVALID_ENUM error, a corresponding error message indicating where the error occurred, and possibly why, will be printed to stderr. For release builds, MESA_DEBUG defaults to off (no debug output). MESA_DEBUG accepts the following comma-separated list of named flags, which adds extra behavior to just set MESA_DEBUG to 1:

silent

turn off debug messages. Only useful for debug builds.

flush

flush after each drawing command

incomplete_tex

extra debug messages when a texture is incomplete

incomplete_fbo

extra debug messages when a FBO is incomplete

context

create a debug context (see GLX_CONTEXT_DEBUG_BIT_ARB) and print error and performance messages to stderr (or MESA_LOG_FILE).

MESA_PROCESS_NAME

if set, overrides the process name string used internally for various purposes (e.g. for driconf option matching, logging, artifact storage, etc.).

MESA_LOG_FILE

specifies a file name for logging all errors, warnings, etc., rather than stderr

MESA_EXTENSION_OVERRIDE

can be used to enable/disable extensions. A value such as GL_EXT_foo -GL_EXT_bar will enable the GL_EXT_foo extension and disable the GL_EXT_bar extension. Note that this will override extensions override configured using driconf.

MESA_EXTENSION_MAX_YEAR

The GL_EXTENSIONS string returned by Mesa is sorted by extension year. If this variable is set to year X, only extensions defined on or before year X will be reported. This is to work-around a bug in some games where the extension string is copied into a fixed-size buffer without truncating. If the extension string is too long, the buffer overrun can cause the game to crash. This is a work-around for that.

MESA_GL_VERSION_OVERRIDE

changes the value returned by glGetString(GL_VERSION) and possibly the GL API type.

  • The format should be MAJOR.MINOR[FC|COMPAT]

  • FC is an optional suffix that indicates a forward compatible context. This is only valid for versions >= 3.0.

  • COMPAT is an optional suffix that indicates a compatibility context or GL_ARB_compatibility support. This is only valid for versions >= 3.1.

  • GL versions <= 3.0 are set to a compatibility (non-Core) profile

  • GL versions = 3.1, depending on the driver, it may or may not have the GL_ARB_compatibility extension enabled.

  • GL versions >= 3.2 are set to a Core profile

  • Examples:

    2.1

    select a compatibility (non-Core) profile with GL version 2.1.

    3.0

    select a compatibility (non-Core) profile with GL version 3.0.

    3.0FC

    select a Core+Forward Compatible profile with GL version 3.0.

    3.1

    select GL version 3.1 with GL_ARB_compatibility enabled per the driver default.

    3.1FC

    select GL version 3.1 with forward compatibility and GL_ARB_compatibility disabled.

    3.1COMPAT

    select GL version 3.1 with GL_ARB_compatibility enabled.

    X.Y

    override GL version to X.Y without changing the profile.

    X.YFC

    select a Core+Forward Compatible profile with GL version X.Y.

    X.YCOMPAT

    select a Compatibility profile with GL version X.Y.

  • Mesa may not really implement all the features of the given version. (for developers only)

MESA_GLES_VERSION_OVERRIDE

changes the value returned by glGetString(GL_VERSION) for OpenGL ES.

  • The format should be MAJOR.MINOR

  • Examples: 2.0, 3.0, 3.1

  • Mesa may not really implement all the features of the given version. (for developers only)

MESA_GLSL_VERSION_OVERRIDE

changes the value returned by glGetString(GL_SHADING_LANGUAGE_VERSION). Valid values are integers, such as 130. Mesa will not really implement all the features of the given language version if it’s higher than what’s normally reported. (for developers only)

MESA_DRICONF_EXECUTABLE_OVERRIDE

if set, overrides the “executable” string used specifically for driconf option matching. This takes higher precedence over more general process name override (e.g. MESA_PROCESS_NAME).

MESA_SHADER_CACHE_DISABLE

if set to true, disables the on-disk shader cache. If set to false, enables the on-disk shader cache when it is disabled by default. Note that EGL_ANDROID_blob_cache is still enabled even if on-disk shader cache is disabled.

MESA_SHADER_CACHE_MAX_SIZE

if set, determines the maximum size of the on-disk cache of compiled shader programs. Should be set to a number optionally followed by K, M, or G to specify a size in kilobytes, megabytes, or gigabytes. By default, gigabytes will be assumed. And if unset, a maximum size of 1GB will be used.

Note

A separate cache might be created for each architecture that Mesa is installed for on your system. For example under the default settings you may end up with a 1GB cache for x86_64 and another 1GB cache for i386.

MESA_SHADER_CACHE_DIR

if set, determines the directory to be used for the on-disk cache of compiled shader programs. If this variable is not set, then the cache will be stored in $XDG_CACHE_HOME/mesa_shader_cache_db (if that variable is set), or else within .cache/mesa_shader_cache_db within the user’s home directory.

MESA_SHADER_CACHE_SHOW_STATS

if set to true, keeps hit/miss statistics for the shader cache. These statistics are printed when the app terminates.

MESA_DISK_CACHE_SINGLE_FILE

if set to 1, enables the single file Fossilize DB on-disk shader cache implementation instead of the default Mesa-DB cache implementation. This implementation allows for loading of precompiled cache DBs via MESA_DISK_CACHE_READ_ONLY_FOZ_DBS or MESA_DISK_CACHE_READ_ONLY_FOZ_DBS_DYNAMIC_LIST. This implementation does not support cache size limits via MESA_SHADER_CACHE_MAX_SIZE. If MESA_SHADER_CACHE_DIR is not set, the cache will be stored in $XDG_CACHE_HOME/mesa_shader_cache_sf (if that variable is set) or else within .cache/mesa_shader_cache_sf within the user’s home directory.

MESA_DISK_CACHE_MULTI_FILE

if set to 1, enables the multi file on-disk shader cache implementation instead of the default Mesa-DB cache implementation. This implementation increases the overall disk usage. If MESA_SHADER_CACHE_DIR is not set, the cache will be stored in $XDG_CACHE_HOME/mesa_shader_cache (if that variable is set) or else within .cache/mesa_shader_cache within the user’s home directory.

MESA_DISK_CACHE_READ_ONLY_FOZ_DBS

if set with MESA_DISK_CACHE_SINGLE_FILE enabled, references a string of comma separated file paths to read only Fossilize DB shader caches for loading at initialization. The file paths are relative to the cache directory and do not include suffixes, referencing both the cache DB and its index file. E.g. MESA_DISK_CACHE_SINGLE_FILE=filename1 refers to filename1.foz and filename1_idx.foz. A limit of 8 DBs can be loaded and this limit is shared with MESA_DISK_CACHE_READ_ONLY_FOZ_DBS_DYNAMIC_LIST.

MESA_DISK_CACHE_DATABASE_NUM_PARTS

specifies number of mesa-db cache parts, default is 50.

MESA_DISK_CACHE_DATABASE_EVICTION_SCORE_2X_PERIOD

Mesa-DB cache eviction algorithm calculates weighted score for the cache items. The weight is doubled based on the last access time of cache entry. By default period of weight doubling is set to one month. Period value is given in seconds.

MESA_DISK_CACHE_READ_ONLY_FOZ_DBS_DYNAMIC_LIST

if set with MESA_DISK_CACHE_SINGLE_FILE enabled, references a text file that contains a new-line separated list of read only Fossilize DB shader caches to load. The list file is modifiable at runtime to allow for loading read only caches after initialization unlike MESA_DISK_CACHE_READ_ONLY_FOZ_DBS. This variable takes an absolute path to the list file. The list file must exist at initialization for updating to occur. Cache files in the list take relative paths to the current cache directory like MESA_DISK_CACHE_READ_ONLY_FOZ_DBS. A limit of 8 DBs can be loaded and this limit is shared with MESA_DISK_CACHE_READ_ONLY_FOZ_DBS.

MESA_DISK_CACHE_COMBINE_RW_WITH_RO_FOZ

if set to 1, enables simultaneous use of RW and RW Fossilize DB caches. At first, data will be retrieved from the RO Fossilize cache. If data isn’t found in the RO cache, then it will be retrieved from the RW cache.

MESA_GLSL

shading language compiler options

MESA_NO_MINMAX_CACHE

when set, the minmax index cache is globally disabled.

MESA_SHADER_CAPTURE_PATH

see Capturing Shaders

MESA_SHADER_DUMP_PATH

see Experimenting with Shader Replacements

MESA_SHADER_READ_PATH

see Experimenting with Shader Replacements

MESA_VK_VERSION_OVERRIDE

changes the Vulkan physical device version as returned in VkPhysicalDeviceProperties::apiVersion.

  • The format should be MAJOR.MINOR[.PATCH]

  • This will not let you force a version higher than the driver’s instance version as advertised by vkEnumerateInstanceVersion

  • This can be very useful for debugging but some features may not be implemented correctly. (For developers only)

MESA_VK_WSI_PRESENT_MODE

overrides the WSI present mode clients specify in VkSwapchainCreateInfoKHR::presentMode. Values can be fifo, relaxed, mailbox or immediate.

MESA_VK_WSI_HEADLESS_SWAPCHAIN

Forces all swapchains to be headless (no rendering will be display in the swapchain’s window).

MESA_VK_ABORT_ON_DEVICE_LOSS

causes the Vulkan driver to call abort() immediately after detecting a lost device. This is extremely useful when testing as it prevents the test suite from continuing on with a lost device.

MESA_VK_ENABLE_SUBMIT_THREAD

for Vulkan drivers which support real timeline semaphores, this forces them to use a submit thread from the beginning, regardless of whether or not they ever see a wait-before-signal condition.

MESA_VK_DEVICE_SELECT_DEBUG

print debug info about device selection decision-making

MESA_VK_TRACE

A comma-separated list of trace types used for offline analysis. The option names are equal to the file extension. Traces are dumped into /tmp. Captures can be triggered by pressing F1 with the application window focused (Currently X11 only) or via MESA_VK_TRACE_FRAME and MESA_VK_TRACE_TRIGGER.

File extension

Offline analysis tool

Supported drivers

rmv

Radeon Memory Visualizer

RADV

rgp

Radeon GPU Profiler

RADV

rra

Radeon Raytracing Analyzer

RADV

ctxroll

Context rolls

RADV

  • Creating RMV captures requires the scripts/setup.sh script in the Radeon Developer Tools folder to be run beforehand

MESA_VK_TRACE_FRAME

Specifies a frame index at which a trace capture is automatically triggered.

MESA_VK_TRACE_TRIGGER

Specifies a trigger file. Creating the file triggers the capture. (e.g. export MESA_VK_TRACE_TRIGGER=/tmp/trigger and then touch /tmp/trigger)

MESA_LOADER_DRIVER_OVERRIDE

chooses a different driver binary such as etnaviv or zink.

DRI_PRIME_DEBUG

print debug info about device selection decision-making

DRI_PRIME

the default GPU is the one used by Wayland/Xorg or the one connected to a display. This variable allows to select a different GPU. It applies to OpenGL and Vulkan (in this case “select” means the GPU will be first in the reported physical devices list). The supported syntaxes are:

  • DRI_PRIME=N: selects the Nth non-default GPU (N > 0).

  • DRI_PRIME=pci-0000_02_00_0: selects the GPU connected to this PCIe bus

  • DRI_PRIME=vendor_id:device_id: selects the first GPU matching these ids.

For Vulkan it’s possible to append !, in which case only the selected GPU will be exposed to the application (e.g.: DRI_PRIME=1!).

Note

lspci -nn | grep VGA can be used to know the PCIe bus or ids to use.

DRIRC_CONFIGDIR

If set, overrides the global search-directories used when searching for drirc config files. The user-local one will still be used. Mostly useful for internal debugging.

NIR passes environment variables

The following are only applicable for drivers that uses NIR, as they modify the behavior for the common NIR_PASS and NIR_PASS_V macros, that wrap calls to NIR lowering/optimizations.

NIR_DEBUG

a comma-separated list of debug options to apply to NIR shaders. Use NIR_DEBUG=help to print a list of available options.

NIR_SKIP

a comma-separated list of optimization/lowering passes to skip.

Mesa Xlib driver environment variables

The following are only applicable to the Mesa Xlib software driver. See the Xlib software driver page for details.

MESA_RGB_VISUAL

specifies the X visual and depth for RGB mode

MESA_BACK_BUFFER

specifies how to implement the back color buffer, either pixmap or ximage

MESA_XSYNC

enable synchronous X behavior (for debugging only)

MESA_GLX_FORCE_ALPHA

if set, forces RGB windows to have an alpha channel.

MESA_GLX_DEPTH_BITS

specifies default number of bits for depth buffer.

MESA_GLX_ALPHA_BITS

specifies default number of bits for alpha channel.

Mesa WGL driver environment variables

The following are only applicable to the Mesa WGL driver, which is in use on Windows.

WGL_FORCE_MSAA

if set to a positive value, specifies the number of MSAA samples to force when choosing the display configuration.

WGL_DISABLE_ERROR_DIALOGS

if set to 1, true or yes, disables Win32 error dialogs. Useful for automated test-runs.

Intel driver environment variables

INTEL_BLACKHOLE_DEFAULT

if set to 1, true or yes, then the OpenGL implementation will default GL_BLACKHOLE_RENDER_INTEL to true, thus disabling any rendering.

INTEL_DEBUG

a comma-separated list of named flags, which do various things:

ann

annotate IR in assembly dumps

bat

emit batch information. Can control in which frames batches get dumped using INTEL_DEBUG_BATCH_FRAME_*, where INTEL_DEBUG_BATCH_FRAME_START <= frame < INTEL_DEBUG_BATCH_FRAME_STOP

blit

emit messages about blit operations

blorp

emit messages about the blorp operations (blits & clears)

buf

emit messages about buffer objects

bt

emit messages binding tables

capture-all

flag all buffers to be captured by the kernel driver when generating an error stage after a GPU hang

clip

emit messages about the clip unit (for old gens, includes the CLIP program)

color

use color in output

cs

dump shader assembly for compute shaders

do32

generate compute shader SIMD32 programs even if workgroup size doesn’t exceed the SIMD16 limit

draw_bkp

Add semaphore wait before/after draw call count. INTEL_DEBUG_BKP_BEFORE_DRAW_COUNT or INTEL_DEBUG_BKP_AFTER_DRAW_COUNT can control draw call number. To make test wait forever, we need to set preempt_timeout_ms and i915.enable_hangcheck to zero.

fall

emit messages about performance issues (same as perf)

fs

dump shader assembly for fragment shaders

gs

dump shader assembly for geometry shaders

heaps

print information about the driver’s heaps (ANV only)

hex

print instruction hex dump with the disassembly

l3

emit messages about the new L3 state during transitions

mesh

dump shader assembly for mesh shaders

no8

don’t generate SIMD8 fragment shader

no16

suppress generation of 16-wide fragment shaders. useful for debugging broken shaders

no32

suppress generation of 32-wide fragment shaders. useful for debugging broken shaders

no-oaconfig

disable HW performance metric configuration, and anything related to i915-perf (useful when running on simulation)

nocompact

disable instruction compaction

nodualobj

suppress generation of dual-object geometry shader code

nofc

disable fast clears

noccs

disable lossless color compression

optimizer

dump shader assembly to files at each optimization pass and iteration that make progress

pc

emit messages about PIPE_CONTROL instruction usage

perf

emit messages about performance issues

perfmon

emit messages about GL_AMD_performance_monitor

perf-symbol-names

use performance counter symbols instead of the counter name (counter symbols are like variable names, it’s sometimes easier to work with when you have lots of metrics to collect)

reemit

mark all state dirty on each draw call

rt

dump shader assembly for ray tracing shaders

sf

emit messages about the strips & fans unit (for old gens, includes the SF program)

shader-print

allow developer print traces added by brw_nir_printf to be printed out on the console

soft64

enable implementation of software 64bit floating point support

sparse

dump usage of sparse resources

spill_fs

force spilling of all registers in the scalar backend (useful to debug spilling code)

spill_vec4

force spilling of all registers in the vec4 backend (useful to debug spilling code)

stall

inserts a stall on the GPU after each draw/dispatch command to wait for it to finish before starting any new work.

submit

emit batchbuffer usage statistics

sync

after sending each batch, wait on the CPU for that batch to finish rendering

swsb-stall

Insert sync NOP after each instruction. This is only valid for Gfx12+.

task

dump shader assembly for task shaders

tcs

dump shader assembly for tessellation control shaders

tcs8

force usage of 8-patches tessellation control shaders (only for gfx 9-11)

tes

dump shader assembly for tessellation evaluation shaders

tex

emit messages about textures.

urb

emit messages about URB setup

vs

dump shader assembly for vertex shaders

wm

dump shader assembly for fragment shaders (same as fs)

cl-quiet

quiets the OpenCL warnings recommending use of Intel compute-runtime

INTEL_DECODE

a comma-separated list of enable/disable flags configuring the output produced by INTEL_DEBUG=bat (use with INTEL_DECODE=+color,-floats) :

color

print colored output

floats

try to decode floating point data in buffers

full

print additional custom information for instructions (usually pulling more information by inspecting memory)

offsets

print offsets of instructions

INTEL_EXTENDED_METRICS

By default, only a standard set of GPU metrics are advertised. This reduces time to collect metrics and hides infrequently used metrics. To enable all metrics, set value to 1.

INTEL_FORCE_PROBE

A comma-separated list of device probe override values. The basic format is <pci-id>[,<pci-id>,...]. The wildcard value of * will specify all known PCI IDs. If ! precedes a PCI ID, or the wildcard value, then the device support will be disabled. All numbers are interpreted in base 16, and a 0x prefix is optional. Values specified later take precedence, so the wildcard probably should only be used at the beginning.

Some examples :

1234,!abcd

Device 0x1234 would be forced on and 0xabcd would be disabled.

1234,!*

All devices are disabled since the wildcard appears later.

!*,0x1234

All devices disabled except 0x1234 which is forced on.

*,!0x1234

All devices are forced on, except 0x1234 which is disabled.

!0x1234,1234

Support for device 0x1234 is forced on since the enable appears later.

Note

If a device requires using INTEL_FORCE_PROBE to force it to load, then Mesa does not have full support for the device. It may have limited, or possibly no functionality within Mesa at this point. It is recommended to upgrade to a Mesa which does not require INTEL_FORCE_PROBE for the device as soon as it is available.

INTEL_MEASURE

Collects GPU timestamps over common intervals, and generates a CSV report to show how long rendering took. The overhead of collection is limited to the flushing that is required at the interval boundaries for accurate timestamps. By default, timing data is sent to stderr. To direct output to a file:

INTEL_MEASURE=file=/tmp/measure.csv {workload}

To begin capturing timestamps at a particular frame:

INTEL_MEASURE=file=/tmp/measure.csv,start=15 {workload}

To capture only 23 frames:

INTEL_MEASURE=count=23 {workload}

To capture frames 15-37, stopping before frame 38:

INTEL_MEASURE=start=15,count=23 {workload}

Designate an asynchronous control file with:

INTEL_MEASURE=control=path/to/control.fifo {workload}

As the workload runs, enable capture for 5 frames with:

$ echo 5 > path/to/control.fifo

Enable unbounded capture:

$ echo -1 > path/to/control.fifo

and disable with:

$ echo 0 > path/to/control.fifo

Select the boundaries of each snapshot with:

INTEL_MEASURE=draw

Collects timings for every render (DEFAULT)

INTEL_MEASURE=rt

Collects timings when the render target changes

INTEL_MEASURE=batch

Collects timings when batches are submitted

INTEL_MEASURE=frame

Collects timings at frame boundaries

With INTEL_MEASURE=interval=5, the duration of 5 events will be combined into a single record in the output. When possible, a single start and end event will be submitted to the GPU to minimize stalling. Combined events will not span batches, except in the case of INTEL_MEASURE=frame.

Collect CPU timestamps instead of GPU timestamps. Prints results immediately instead of waiting for GPU execution. Useful when used with interactive debug to know which frame, or where in frame, you are currently in.

INTEL_MEASURE=cpu {workload}

INTEL_MODIFIER_OVERRIDE

if set, determines the single DRM modifier reported back to (Vulkan) applications, in order to make selecting modifier deterministic between Vulkan driver and applications. The value can be one of the supported modifiers on a platform, but other values are also acceptable for debug purposes.

INTEL_NO_HW

if set to 1, true or yes, prevents batches from being submitted to the hardware. This is useful for debugging hangs, etc.

INTEL_PRECISE_TRIG

if set to 1, true or yes, then the driver prefers accuracy over performance in trig functions.

INTEL_SHADER_OPTIMIZER_PATH

if set, determines the directory to be used for overriding shader assembly. The binaries with custom assembly should be placed in this folder and have a name formatted as sha1_of_assembly.bin. The SHA-1 of a shader assembly is printed when assembly is dumped via corresponding INTEL_DEBUG flag (e.g. vs for vertex shader). A binary could be generated from a dumped assembly by i965_asm. For INTEL_SHADER_ASM_READ_PATH to work it is necessary to enable dumping of corresponding shader stages via INTEL_DEBUG. It is advised to use nocompact flag of INTEL_DEBUG when dumping and overriding shader assemblies. The success of assembly override would be signified by “Successfully overrode shader with sha1 <SHA-1>” in stderr replacing the original assembly.

INTEL_SHADER_ASM_READ_PATH

if set, determines the directory to be used for overriding shader assembly. The binaries with custom assembly should be placed in this folder and have a name formatted as sha1_of_assembly.bin. The SHA-1 of a shader assembly is printed when assembly is dumped via corresponding INTEL_DEBUG flag (e.g. vs for vertex shader). A binary could be generated from a dumped assembly by i965_asm. For INTEL_SHADER_ASM_READ_PATH to work it is necessary to enable dumping of corresponding shader stages via INTEL_DEBUG. It is advised to use nocompact flag of INTEL_DEBUG when dumping and overriding shader assemblies. The success of assembly override would be signified by “Successfully overrode shader with sha1 <SHA-1>” in stderr replacing the original assembly.

INTEL_SHADER_BIN_DUMP_PATH

if set, determines the directory to which the compiled shaders will be dumped. They will be dumped as sha1_of_assembly.bin, where the sha1 values will be the same as can be found in the INTEL_DEBUG output, and can be used for INTEL_SHADER_ASM_READ_PATH input.

Note

Unlike the text form of shader dumping, INTEL_DEBUG does not affect on the list of shaders to dump. All generated shaders are always dumped if INTEL_SHADER_BIN_DUMP_PATH variable is set.

INTEL_SIMD_DEBUG

a comma-separated list of named flags, which control simd dispatch widths:

fs8

allow generation of SIMD8 fragment shader

fs16

allow generation of SIMD16 fragment shader

fs32

allow generation of SIMD32 fragment shader

cs8

allow generation of SIMD8 compute shader

cs16

allow generation of SIMD16 compute shader

cs32

allow generation of SIMD32 compute shader

ts8

allow generation of SIMD8 task shader

ts16

allow generation of SIMD16 task shader

ts32

allow generation of SIMD32 task shader

ms8

allow generation of SIMD8 mesh shader

ms16

allow generation of SIMD16 mesh shader

ms32

allow generation of SIMD32 mesh shader

rt8

allow generation of SIMD8 ray-tracing shader

rt16

allow generation of SIMD16 ray-tracing shader

rt32

allow generation of SIMD32 ray-tracing shader

If none of widths for particular shader stage was specified, then all widths are allowed.

Anvil(ANV) driver environment variables

ANV_ENABLE_PIPELINE_CACHE

If defined to 0 or false, this will disable pipeline caching, forcing ANV to reparse and recompile any VkShaderModule (SPIRV) it is given.

ANV_DISABLE_SECONDARY_CMD_BUFFER_CALLS

If defined to 1 or true, this will prevent usage of self modifying command buffers to implement vkCmdExecuteCommands. As a result of this, it will also disable VK_KHR_performance_query.

ANV_ALWAYS_BINDLESS

If defined to 1 or true, this forces all descriptor sets to use the internal Binding Model.

ANV_QUEUE_THREAD_DISABLE

If defined to 1 or true, this disables support for timeline semaphores.

ANV_USERSPACE_RELOCS

If defined to 1 or true, this forces ANV to always do kernel relocations in command buffers. This should only have an effect on hardware that doesn’t support soft-pinning (Ivybridge, Haswell, Cherryview).

ANV_PRIMITIVE_REPLICATION_MAX_VIEWS

Specifies up to how many view shaders can be lowered to handle VK_KHR_multiview. Beyond this number, multiview is implemented using instanced rendering. If unspecified, the value default to 2.

ANV_NO_GPL

If set to 1, true, or yes, then VK_EXT_graphics_pipeline_library will be disabled.

ANV_SPARSE

By default, the sparse resources feature is enabled. However, if set to 0, false, or no, it will be disabled. Platforms older than Tiger Lake do not support this feature.

ANV_SPARSE_USE_TRTT

On platforms supported by Xe KMD (Lunar Lake and newer) this parameter changes the implementation of sparse resources feature. For i915 there is no option, sparse resources is always implemented with TRTT.

DRI environment variables

DRI_NO_MSAA

disable MSAA for GLX/EGL MSAA visuals

Vulkan mesa device select layer environment variables

MESA_VK_DEVICE_SELECT

when set to “list” prints the list of devices. when set to “vid:did” number from PCI device. That PCI device is selected as default. The default device is returned as the first device in vkEnumeratePhysicalDevices API. Using “vid:did!” will have the same effect as using the MESA_VK_DEVICE_SELECT_FORCE_DEFAULT_DEVICE variable.

MESA_VK_DEVICE_SELECT_FORCE_DEFAULT_DEVICE

when set to 1, the device identified as default will be the only one returned in vkEnumeratePhysicalDevices API.

EGL environment variables

Mesa EGL supports different sets of environment variables. See the Mesa EGL page for the details.

Gallium environment variables

GALLIUM_HUD

draws various information on the screen, like framerate, CPU load, driver statistics, performance counters, etc. Set GALLIUM_HUD to help and run e.g. glxgears for more info.

GALLIUM_HUD_PERIOD

sets the HUD update rate in seconds (float). Use zero to update every frame. The default period is 1/2 second.

GALLIUM_HUD_VISIBLE

control default visibility, defaults to true.

GALLIUM_HUD_OPACITY

control background opacity as an integer percentage (1-100), defaults to 66%.

GALLIUM_HUD_TOGGLE_SIGNAL

toggle visibility via user specified signal. Especially useful to toggle HUD at specific points of application and disable for unencumbered viewing the rest of the time. For example, set GALLIUM_HUD_VISIBLE to false and GALLIUM_HUD_TOGGLE_SIGNAL to 10 (SIGUSR1). Use kill -10 <pid> to toggle the HUD as desired.

GALLIUM_HUD_SCALE

Scale HUD by an integer factor, for high DPI displays. Default is 1.

GALLIUM_HUD_ROTATION

Rotate the HUD by an integer number of degrees, the specified value must be a multiple of 90. Default is 0.

GALLIUM_HUD_DUMP_DIR

specifies a directory for writing the displayed HUD values into files.

GALLIUM_DRIVER

useful in combination with LIBGL_ALWAYS_SOFTWARE = true for choosing one of the software renderers softpipe or llvmpipe.

GALLIUM_LOG_FILE

specifies a file for logging all errors, warnings, etc. rather than stderr.

GALLIUM_PIPE_SEARCH_DIR

specifies an alternate search directory for pipe-loader which overrides the compile-time path based on the install location.

GALLIUM_PRINT_OPTIONS

if non-zero, print all the Gallium environment variables which are used, and their current values.

GALLIUM_TRACE

If set, this variable will cause the Trace output to be written to the specified file. Paths may be relative or absolute; relative paths are relative to the working directory. For example, setting it to “trace.xml” will cause the trace to be written to a file of the same name in the working directory.

GALLIUM_TRACE_TC

If enabled while Trace is active, this variable specifies that the threaded context should be traced for drivers which implement it. By default, the driver thread is traced, which will include any reordering of the command stream from threaded context.

GALLIUM_TRACE_TRIGGER

If set while Trace is active, this variable specifies a filename to monitor. Once the file exists (e.g., from the user running ‘touch /path/to/file’), a single frame will be recorded into the trace output. Paths may be relative or absolute; relative paths are relative to the working directory.

GALLIUM_DUMP_CPU

if non-zero, print information about the CPU on start-up

TGSI_PRINT_SANITY

if set, do extra sanity checking on TGSI shaders and print any errors to stderr.

DRAW_FSE

Enable fetch-shade-emit middle-end even though its not correct (e.g. for Softpipe)

DRAW_NO_FSE

Disable fetch-shade-emit middle-end even when it is correct

DRAW_USE_LLVM

if set to zero, the draw module will not use LLVM to execute shaders, vertex fetch, etc.

ST_DEBUG

controls debug output from the Mesa/Gallium state tracker. Setting to tgsi, for example, will print all the TGSI shaders. See src/mesa/state_tracker/st_debug.c for other options.

GALLIUM_OVERRIDE_CPU_CAPS

Override CPU capabilities for LLVMpipe and Softpipe, possible values for x86: nosse sse sse2 sse3 ssse3 sse4.1 avx

Clover environment variables

CLOVER_DEVICE_TYPE

allows to overwrite the device type of devices. Possible values are accelerator, cpu, custom and gpu

CLOVER_DEVICE_VERSION_OVERRIDE

overwrites the auto detected OpenCL version of a device. Possible values: 1.0 1.1 1.2 2.0 2.1 2.2 3.0

CLOVER_DEVICE_CLC_VERSION_OVERRIDE

overwrites the auto detected CLC version. Possible values: 1.0 1.1 1.2 2.0 2.1 2.2 3.0

CLOVER_EXTRA_BUILD_OPTIONS

allows specifying additional compiler and linker options. Specified options are appended after the options set by the OpenCL program in clBuildProgram.

CLOVER_EXTRA_COMPILE_OPTIONS

allows specifying additional compiler options. Specified options are appended after the options set by the OpenCL program in clCompileProgram.

allows specifying additional linker options. Specified options are appended after the options set by the OpenCL program in clLinkProgram.

IRIS_ENABLE_CLOVER

allows to enable experimental Clover NIR support with the iris driver if set to 1 or true.

Rusticl environment variables

RUSTICL_DEVICE_TYPE

allows to overwrite the device type of devices. Possible values are accelerator, cpu, custom and gpu

RUSTICL_CL_VERSION

overwrites the auto detected OpenCL version of all devices. Specified as major.minor.

RUSTICL_ENABLE

a comma-separated list of drivers to enable CL on. An optional list of comma-separated integers can be passed per driver to specify which devices to enable. Examples:

  • RUSTICL_ENABLE=iris (enables all iris devices)

  • RUSTICL_ENABLE=iris:1,radeonsi:0,2 (enables second iris and first and third radeonsi device)

Supported drivers (decent support with maybe a few conformance issues or bugs): iris, llvmpipe, nouveau, panfrost, radeonsi, Experimental drivers (unknown level of support, expect conformance issues or major bugs): r600

RUSTICL_FEATURES

a comma-separated list of features to enable. Those are disabled by default as they might not be stable enough or break OpenCL conformance.

  • fp16 enables OpenCL half support

  • fp64 enables OpenCL double support

RUSTICL_DEBUG

a comma-separated list of debug channels to enable.

  • allow_invalid_spirv disables validation of any input SPIR-V

  • clc dumps all OpenCL C source being compiled

  • nir dumps nirs in various compilation stages. Might print nothing if shader caching is

    enabled.

  • no_reuse_context pipe_contexts are not recycled

  • no_variants disable kernel variants (e.g. specialized binaries for offsets == 0)

  • perf prints a warning when hitting slow paths once

  • perfspam same as perf, but doesn’t skip same warnings

  • program dumps compilation logs to stderr

  • sync waits on the GPU to complete after every event

  • validate validates any internally generated SPIR-Vs, e.g. through compiling OpenCL C code

RUSTICL_MAX_WORK_GROUPS

Limits the amount of threads per dimension in a work-group. Useful for splitting up long running tasks to increase responsiveness or to simulate the lowering of huge global sizes for testing.

clc environment variables

CLC_DEBUG

a comma-separated list of debug channels to enable.

  • dump_llvm Dumps all generated LLVM IRs

  • dump_spirv Dumps all compiled, linked and specialized SPIR-Vs

  • verbose Enable debug logging of clc code

Nine frontend environment variables

D3D_ALWAYS_SOFTWARE

an integer, which forces Nine to use the CPU instead of GPU acceleration.

NINE_DEBUG

a comma-separated list of named flags that do debugging things. Use NINE_DEBUG=help to print a list of available options.

NINE_FF_DUMP

a boolean, which dumps shaders generated by a fixed function (FF).

NINE_SHADER

a comma-separated list of named flags, which do alternate shader handling. Use NINE_SHADER=help to print a list of available options.

NINE_QUIRKS

a comma-separated list of named flags that do various things. Use NINE_DEBUG=help to print a list of available options.

Softpipe driver environment variables

SOFTPIPE_DEBUG

a comma-separated list of named flags, which do various things:

vs

Dump vertex shader assembly to stderr

fs

Dump fragment shader assembly to stderr

gs

Dump geometry shader assembly to stderr

cs

Dump compute shader assembly to stderr

no_rast

rasterization is disabled. For profiling purposes.

use_llvm

the Softpipe driver will try to use LLVM JIT for vertex shading processing.

LLVMpipe driver environment variables

LP_NO_RAST

if set LLVMpipe will no-op rasterization

LP_DEBUG

a comma-separated list of debug options is accepted. See the source code for details.

LP_PERF

a comma-separated list of options to selectively no-op various parts of the driver. See the source code for details.

LP_NUM_THREADS

an integer indicating how many threads to use for rendering. Zero turns off threading completely. The default value is the number of CPU cores present.

VMware SVGA driver environment variables

SVGA_FORCE_SWTNL

force use of software vertex transformation

SVGA_NO_SWTNL

don’t allow software vertex transformation fallbacks (will often result in incorrect rendering).

SVGA_DEBUG

for dumping shaders, constant buffers, etc. See the code for details.

SVGA_EXTRA_LOGGING

if set, enables extra logging to the vmware.log file, such as the OpenGL program’s name and command line arguments.

SVGA_NO_LOGGING

if set, disables logging to the vmware.log file. This is useful when using Valgrind because it otherwise crashes when initializing the host log feature.

See the driver code for other, lesser-used variables.

WGL environment variables

WGL_SWAP_INTERVAL

to set a swap interval, equivalent to calling wglSwapIntervalEXT() in an application. If this environment variable is set, application calls to wglSwapIntervalEXT() will have no effect.

VA-API environment variables

VAAPI_MPEG4_ENABLED

enable MPEG4 for VA-API, disabled by default.

VC4 driver environment variables

VC4_DEBUG

a comma-separated list of named flags, which do various things. Use VC4_DEBUG=help to print a list of available options.

V3D/V3DV driver environment variables

V3D_DEBUG

a comma-separated list of debug options. Use V3D_DEBUG=help to print a list of available options.

RADV driver environment variables

RADV_DEBUG

a comma-separated list of named flags, which do various things:

llvm

enable LLVM compiler backend

allbos

force all allocated buffers to be referenced in submissions

checkir

validate the LLVM IR before LLVM compiles the shader

epilogs

dump fragment shader epilogs

extra_md

add extra information in bo metadata to help tools (umr)

forcecompress

Enables DCC,FMASK,CMASK,HTILE in situations where the driver supports it but normally does not deem it beneficial.

hang

enable GPU hangs detection and dump a report to $HOME/radv_dumps_<pid>_<time> if a GPU hang is detected

img

Print image info

info

show GPU-related information

invariantgeom

Mark geometry-affecting outputs as invariant. This works around a common class of application bugs appearing as flickering.

metashaders

dump internal meta shaders

noatocdithering

disable dithering for alpha to coverage

nobinning

disable primitive binning

nocache

disable shaders cache

nocompute

disable compute queue

nodcc

disable Delta Color Compression (DCC) on images

nodisplaydcc

disable Delta Color Compression (DCC) on displayable images

nodynamicbounds

do not check OOB access for dynamic descriptors

noeso

disable VK_EXT_shader_object

nofastclears

disable fast color/depthstencil clears

nofmask

disable FMASK compression on MSAA images (GFX6-GFX10.3)

nogpl

disable VK_EXT_graphics_pipeline_library

nohiz

disable HIZ for depthstencil images

noibs

disable directly recording command buffers in GPU-visible memory

nomeshshader

disable mesh shader support on GFX10.3+

nongg

disable NGG for GFX10 and GFX10.3

nonggc

disable NGG culling on GPUs where it’s enabled by default (GFX10.3 only).

nongg_gs

disable NGG GS for GFX10 and GFX10.3

nort

skip executing vkCmdTraceRays and ray queries (RT extensions will still be advertised)

notccompatcmask

disable TC-compat CMASK for MSAA surfaces

noumr

disable UMR dumps during GPU hang detection (only with RADV_DEBUG = hang)

novrsflatshading

disable VRS for flat shading (only on GFX10.3+)

preoptir

dump LLVM IR before any optimizations

prologs

dump vertex shader prologs

shaders

dump shaders

shaderstats

dump shader statistics

shadowregs

enable register shadowing

spirv

dump SPIR-V

splitfma

split application-provided fused multiply-add in geometry stages

startup

display info at startup

syncshaders

synchronize shaders after all draws/dispatches

zerovram

initialize all memory allocated in VRAM as zero

RADV_FORCE_FAMILY

create a null device to compile shaders without a AMD GPU (e.g. VEGA10)

RADV_FORCE_VRS

allow to force per-pipeline vertex VRS rates on GFX10.3+. This is only forced for pipelines that don’t explicitly use VRS or flat shading. The supported values are 2x2, 1x2, 2x1 and 1x1. Only for testing purposes.

RADV_FORCE_VRS_CONFIG_FILE

similar to RADV_FORCE_VRS but allow to configure from a file. If present, this supersedes RADV_FORCE_VRS.

RADV_PERFTEST

a comma-separated list of named flags, which do various things:

bolist

enable the global BO list

cswave32

enable wave32 for compute shaders (GFX10+)

dccmsaa

enable DCC for MSAA images

dmashaders

upload shaders to invisible VRAM (might be useful for non-resizable BAR systems)

emulate_rt

forces ray-tracing to be emulated in software on GFX10_3+ and enables rt extensions with older hardware.

gewave32

enable wave32 for vertex/tess/geometry shaders (GFX10+)

localbos

enable local BOs

nggc

enable NGG culling on GPUs where it’s not enabled by default (GFX10.1 only).

nircache

cache per-stage NIR for graphics pipelines

nosam

disable optimizations that get enabled when all VRAM is CPU visible.

pswave32

enable wave32 for pixel shaders (GFX10+)

rtwave32

enable wave32 for ray tracing shaders (GFX11+)

rtwave64

enable wave64 for ray tracing shaders (GFX10-10.3)

sam

enable optimizations to move more driver internal objects to VRAM.

transfer_queue

enable experimental transfer queue support (GFX9+, not yet spec compliant)

video_decode

enable experimental video decoding support

RADV_TEX_ANISO

force anisotropy filter (up to 16)

RADV_THREAD_TRACE_BUFFER_SIZE

set the SQTT/RGP buffer size in bytes (default value is 32MiB, the buffer is automatically resized if too small)

RADV_THREAD_TRACE_CACHE_COUNTERS

enable/disable SQTT/RGP cache counters on GFX10+ (enabled by default)

RADV_THREAD_TRACE_INSTRUCTION_TIMING

enable/disable SQTT/RGP instruction timing (enabled by default)

RADV_THREAD_TRACE_QUEUE_EVENTS

enable/disable SQTT/RGP queue events (enabled by default)

RADV_RRA_TRACE_VALIDATE

enable validation of captured acceleration structures. Can be useful if RRA crashes upon opening a trace.

RADV_RRA_TRACE_HISTORY_SIZE

set the ray history buffer size when capturing RRA traces (default value is 100MiB, small buffers may result in incomplete traces)

RADV_RRA_TRACE_RESOLUTION_SCALE

decrease the resolution used for dumping the ray history resolution when capturing RRA traces. This allows for dumping every Nth invocation along each dispatch dimension.

RADV_PROFILE_PSTATE

choose the specific pstate to enter when using thread tracing or when acquiring the profiling lock for performance queries.

standard

force GPU clocks to an arbitrary fixed level

min_sclk

force the shader clock to its minimum level

min_mclk

force the memory clock to its minimum level

peak

force GPU clocks to their maximum level, this is the default value

ACO_DEBUG

a comma-separated list of named flags, which do various things:

validateir

validate the ACO IR at various points of compilation (enabled by default for debug/debugoptimized builds)

novalidateir

disable ACO IR validation in debug/debugoptimized builds

validatera

validate register assignment of ACO IR and catches many RA bugs

force-waitcnt

force emitting waitcnt states if there is something to wait for

force-waitdeps

force emitting waitcnt dependencies for debugging hazards on GFX10+

novn

disable value numbering

noopt

disable various optimizations

nosched

disable pre-RA, ILP and VOPD instruction scheduling

nosched-ilp

disable ILP instruction scheduling

nosched-vopd

disable VOPD instruction scheduling

perfinfo

print information used to calculate some pipeline statistics

liveinfo

print liveness and register demand information before scheduling

RadeonSI driver environment variables

radeonsi_no_infinite_interp

Kill PS with infinite interp coeff (might fix hangs)

radeonsi_clamp_div_by_zero

Clamp div by zero (x / 0 becomes FLT_MAX instead of NaN) (might fix rendering corruptions)

radeonsi_zerovram

Clear all allocated memory to 0 before usage (might fix rendering corruptions)

AMD_DEBUG

a comma-separated list of named flags, which do various things:

nodcc

Disable DCC.

nodccclear

Disable DCC fast clear

nodisplaydcc

disable Delta Color Compression (DCC) on displayable images

nodccmsaa

Disable DCC for MSAA

nodpbb

Disable DPBB. Overrules the dpbb enable option.

noefc

Disable hardware based encoder color format conversion

lowlatencyenc

Enable low latency encoding

notiling

Disable tiling

nofmask

Disable MSAA compression

nohyperz

Disable Hyper-Z

no2d

Disable 2D tiling

info

Print driver information

tex

Print texture info

compute

Print compute info

vm

Print virtual addresses when creating resources

vs

Print vertex shaders

ps

Print pixel shaders

gs

Print geometry shaders

tcs

Print tessellation control shaders

tes

Print tessellation evaluation shaders

cs

Print compute shaders

noir

Don’t print the LLVM IR

nonir

Don’t print NIR when printing shaders

noasm

Don’t print disassembled shaders

preoptir

Print the LLVM IR before initial optimizations

w32ge

Use Wave32 for vertex, tessellation, and geometry shaders.

w32ps

Use Wave32 for pixel shaders.

w32cs

Use Wave32 for computes shaders.

w64ge

Use Wave64 for vertex, tessellation, and geometry shaders.

w64ps

Use Wave64 for pixel shaders.

w64cs

Use Wave64 for computes shaders.

checkir

Enable additional sanity checks on shader IR

mono

Use old-style monolithic shaders compiled on demand

nooptvariant

Disable compiling optimized shader variants.

useaco

Use ACO as shader compiler when possible

nowc

Disable GTT write combining

check_vm

Check VM faults and dump debug info.

reserve_vmid

Force VMID reservation per context.

nongg

Disable NGG and use the legacy pipeline.

nggc

Always use NGG culling even when it can hurt.

nonggc

Disable NGG culling.

switch_on_eop

Program WD/IA to switch on end-of-packet.

nooutoforder

Disable out-of-order rasterization

dpbb

Enable DPBB. Enable DPBB for gfx9 dGPU. Default enabled for gfx9 APU and >= gfx10.

extra_md

add extra information in bo metadata to help tools (umr)

r600 driver environment variables

R600_DEBUG

a comma-separated list of named flags, which do various things:

nocpdma

Disable CP DMA

nosb

Disable sb backend for graphics shaders

sbcl

Enable sb backend for compute shaders

sbdry

Don’t use optimized bytecode (just print the dumps)

sbstat

Print optimization statistics for shaders

sbdump

Print IR dumps after some optimization passes

sbnofallback

Abort on errors instead of fallback

sbdisasm

Use sb disassembler for shader dumps

sbsafemath

Disable unsafe math optimizations

nirsb

Enable NIR with SB optimizer

tex

Print texture info

nir

Enable experimental NIR shaders

compute

Print compute info

vm

Print virtual addresses when creating resources

info

Print driver information

fs

Print fetch shaders

vs

Print vertex shaders

gs

Print geometry shaders

ps

Print pixel shaders

cs

Print compute shaders

tcs

Print tessellation control shaders

tes

Print tessellation evaluation shaders

noir

Don’t print the LLVM IR

notgsi

Don’t print the TGSI

noasm

Don’t print disassembled shaders

preoptir

Print the LLVM IR before initial optimizations

checkir

Enable additional sanity checks on shader IR

nooptvariant

Disable compiling optimized shader variants.

testdma

Invoke SDMA tests and exit.

testvmfaultcp

Invoke a CP VM fault test and exit.

testvmfaultsdma

Invoke a SDMA VM fault test and exit.

testvmfaultshader

Invoke a shader VM fault test and exit.

nodma

Disable asynchronous DMA

nohyperz

Disable Hyper-Z

noinvalrange

Disable handling of INVALIDATE_RANGE map flags

no2d

Disable 2D tiling

notiling

Disable tiling

switch_on_eop

Program WD/IA to switch on end-of-packet.

forcedma

Use asynchronous DMA for all operations when possible.

precompile

Compile one shader variant at shader creation.

nowc

Disable GTT write combining

check_vm

Check VM faults and dump debug info.

unsafemath

Enable unsafe math shader optimizations

R600_DEBUG_COMPUTE

if set to true, various compute-related debug information will be printed to stderr. Defaults to false.

R600_DUMP_SHADERS

if set to true, NIR shaders will be printed to stderr. Defaults to false.

R600_HYPERZ

If set to false, disables HyperZ optimizations. Defaults to true.

R600_NIR_DEBUG

a comma-separated list of named flags, which do various things:

instr

Log all consumed nir instructions

ir

Log created R600 IR

cc

Log R600 IR to assembly code creation

noerr

Don’t log shader conversion errors

si

Log shader info (non-zero values)

reg

Log register allocation and lookup

io

Log shader in and output

ass

Log IR to assembly conversion

flow

Log control flow instructions

merge

Log register merge operations

nomerge

Skip register merge step

tex

Log texture ops

trans

Log generic translation messages

r300 driver environment variables

RADEON_DEBUG

a comma-separated list of named flags, which do various things:

info

Print hardware info (printed by default on debug builds

fp

Log fragment program compilation

vp

Log vertex program compilation

draw

Log draw calls

swtcl

Log SWTCL-specific info

rsblock

Log rasterizer registers

psc

Log vertex stream registers

tex

Log basic info about textures

texalloc

Log texture mipmap tree info

rs

Log rasterizer

fb

Log framebuffer

cbzb

Log fast color clear info

hyperz

Log HyperZ info

scissor

Log scissor info

msaa

Log MSAA resources

anisohq

Use high quality anisotropic filtering

notiling

Disable tiling

noimmd

Disable immediate mode

noopt

Disable shader optimizations

nocbzb

Disable fast color clear

nozmask

Disable zbuffer compression

nohiz

Disable hierarchical zbuffer

nocmask

Disable AA compression and fast AA clear

notcl

Disable hardware accelerated Transform/Clip/Lighting

Asahi driver environment variables

ASAHI_MESA_DEBUG

a comma-separated list of named flags, which do various things:

trace

Trace work submitted to the GPU to files, using the agxdecode infrastructure. This produces a large volume of data, so should be used with caution. The traces are written to agxdecode.dump, but this can be overridden using AGXDECODE_DUMP_FILE.

no16

Disable 16-bit floating point support. This may workaround application bugs in certain OpenGL ES applications originally written for desktops. If such applications are found in the wild, they should be fixed upstream (if possible) or added in the Mesa-wide driconf (if closed source).

dirty

In debug builds only: disable dirty tracking optimizations.

nowc

Disable write-combining (force all allocations to be write-through). This may be useful for diagnosing certain performance issues. Note imported buffers may still be write-combined.

AGX_MESA_DEBUG

a comma-separated list of named flags, which do various things:

shaders

Print shaders being compiled at various stages in the pipeline.

shaderdb

Print statistics about compiled shaders.

verbose

Disassemble in verbose mode, including additional information that may be useful for debugging.

internal

Include even internal shaders (as produced for clears, blits, and such) when printing shaders. Without this flag, internal shaders are ignored by the shaders and shaderdb flags.

novalidate

In debug builds only: skip internal intermediate representation validation.

noopt

Disable various backend optimizations.

PowerVR driver environment variables

PVR_DEBUG

A comma-separated list of debug options. Use PVR_DEBUG=help to print a list of available options.

ROGUE_DEBUG

a comma-separated list of named flags for the Rogue compiler, which do various things:

nir

Print the input NIR to stdout.

nir_passes

Print the output of each NIR pass to stdout.

ir

Print the input Rogue IR to stdout.

ir_passes

Print the output of each Rogue IR pass to stdout.

ir_details

Includes additional details when printing Rogue IR.

vld_skip

Skips the compiler validation step.

vld_nonfatal

Prints all the validation errors instead of stopping after the first.

ROGUE_COLOR

if set to auto Rogue IR will be colorized if stdout is not a pipe. Color is forced off if set to off/0 or on if set to on/1. Defaults to auto.

i915 driver environment variables

I915_DEBUG

Debug flags for the i915 driver.

I915_NO_HW

Stop the i915 driver from submitting commands to the hardware.

I915_DUMP_CMD

Dump all commands going to the hardware.

Freedreno driver environment variables

FD_MESA_DEBUG

Debug flags for the Freedreno driver.


Other Gallium drivers have their own environment variables. These may change frequently so the source code should be consulted for details.

Vulkan loader environment variables

These variable are handled by Khronos’ Vulkan loader, not by Mesa, but they are documented here as we reference them in other places in our docs.

VK_DRIVER_FILES

Force the loader to use the specific driver JSON files. The value contains a list of delimited full path listings to driver JSON Manifest files and/or paths to folders containing driver JSON files.

See Vulkan loader docs on environment variables.

VK_LOADER_LAYERS_ENABLE

A comma-delimited list of globs to search for in known layers and used to select only the layers whose layer name matches one or more of the provided globs. Known layers are those which are found by the loader taking into account default search paths and other environment variables (like VK_LAYER_PATH).

See Vulkan loader docs on environment variables.

VK_ICD_FILENAMES

Deprecated, replaced by VK_DRIVER_FILES.

VK_INSTANCE_LAYERS

Deprecated, replaced by VK_LOADER_LAYERS_ENABLE.