Xe Configfs

Overview

Configfs is a filesystem-based manager of kernel objects. Xe KMD registers aconfigfs subsystem calledxe that creates a directory in the mountedconfigfs directory. The user can create devices under this directory andconfigure them as necessary. SeeConfigfs - Userspace-driven Kernel Object Configuration formore information about how configfs works.

Create devices

To create a device, thexe module should already be loaded, but someattributes can only be set before binding the device. It can be accomplishedby blocking the driver autoprobe:

# echo 0 > /sys/bus/pci/drivers_autoprobe# modprobe xe

In order to create a device, the user has to create a directory insidexe:

# mkdir /sys/kernel/config/xe/0000:03:00.0/

Every device created is populated by the driver with entries that can beused to configure it:

/sys/kernel/config/xe/├── 0000:00:02.0│   └── ...├── 0000:00:02.1│   └── ...:└── 0000:03:00.0    ├── survivability_mode    ├── gt_types_allowed    ├── engines_allowed    └── enable_psmi

After configuring the attributes as per next section, the device can beprobed with:

# echo 0000:03:00.0 > /sys/bus/pci/drivers/xe/bind# # or# echo 0000:03:00.0 > /sys/bus/pci/drivers_probe

Configure Attributes

Survivability mode:

Enable survivability mode on supported cards. This setting only takeseffect when probing the device. Example to enable it:

# echo 1 > /sys/kernel/config/xe/0000:03:00.0/survivability_mode

This attribute can only be set before binding to the device.

Allowed GT types:

Allow only specific types of GTs to be detected and initialized by thedriver. Any combination of GT types can be enabled/disabled, althoughsome settings will cause the device to fail to probe.

Writes support both comma- and newline-separated input format. Readswill always return one GT type per line. “primary” and “media” are theGT type names supported by this interface.

This attribute can only be set before binding to the device.

Examples:

Allow both primary and media GTs to be initialized and used. This matchesthe driver’s default behavior:

# echo 'primary,media' > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed

Allow only the primary GT of each tile to be initialized and used,effectively disabling the media GT if it exists on the platform:

# echo 'primary' > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed

Allow only the media GT of each tile to be initialized and used,effectively disabling the primary GT.This configuration will causedevice probe failure on all current platforms, but may be allowed onigpu platforms in the future:

# echo 'media' > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed

Disable all GTs. Only other GPU IP (such as display) is potentially usable.This configuration will cause device probe failure on all currentplatforms, but may be allowed on igpu platforms in the future:

# echo '' > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed

Allowed engines:

Allow only a set of engine(s) to be available, disabling the other engineseven if they are available in hardware. This is applied after HW fuses areconsidered on each tile. Examples:

Allow only one render and one copy engines, nothing else:

# echo 'rcs0,bcs0' > /sys/kernel/config/xe/0000:03:00.0/engines_allowed

Allow only compute engines and first copy engine:

# echo 'ccs*,bcs0' > /sys/kernel/config/xe/0000:03:00.0/engines_allowed

Note that the engine names are the per-GT hardware names. On multi-tileplatforms, writingrcs0,bcs0 to this file would allow the first renderand copy engines on each tile.

The requested configuration may not be supported by the platform and drivermay fail to probe. For example: if at least one copy engine is expected to beavailable for migrations, but it’s disabled. This is intended for debuggingpurposes only.

This attribute can only be set before binding to the device.

PSMI

Enable extra debugging capabilities to trace engine execution. Only usefulduring early platform enabling and requires additional hardware connected.Once it’s enabled, additionals WAs are added and runtime configuration isdone via debugfs. Example to enable it:

# echo 1 > /sys/kernel/config/xe/0000:03:00.0/enable_psmi

This attribute can only be set before binding to the device.

Context restore BB

Allow to execute a batch buffer during any context switches. When theGPU is restoring the context, it executes additional commands. It’s usefulfor testing additional workarounds and validating certain HW behaviors: it’snot intended for normal execution and will taint the kernel with TAINT_TESTwhen used.

The syntax allows to pass straight instructions to be executed by the enginein a batch buffer or set specific registers.

  1. Generic instruction:

    <engine-class> cmd <instr> [[dword0] [dword1] [...]]
  2. Simple register setting:

    <engine-class> reg <address> <value>

Commands are saved per engine class: all instances of that class will executethose commands during context switch. The instruction, dword arguments,addresses and values are in hex format like in the examples below.

  1. Execute a LRI command to write 0xDEADBEEF to register 0x4f10 after thenormal context restore:

    # echo 'rcs cmd 11000001 4F100 DEADBEEF' \        > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_post_bb
  2. Execute a LRI command to write 0xDEADBEEF to register 0x4f10 at thebeginning of the context restore:

    # echo 'rcs cmd 11000001 4F100 DEADBEEF' \        > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_mid_bb
  3. Load certain values in a couple of registers (it can be used as a simpleralternative to thecmd) action:

    # cat > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_post_bb <<EOFrcs reg 4F100 DEADBEEFrcs reg 4F104 FFFFFFFFEOF

    Note

    When using multiple lines, make sure to use a command that isimplemented with a single write syscall, like HEREDOC.

Currently this is implemented only for post and mid context restore andthese attributes can only be set before binding to the device.

Max SR-IOV Virtual Functions

This config allows to limit number of the Virtual Functions (VFs) that canbe managed by the Physical Function (PF) driver, where value 0 disables thePF mode (no VFs).

The default max_vfs config value is taken from the max_vfs modparam.

How to enable PF with support with unlimited (up to HW limit) number of VFs:

# echo unlimited > /sys/kernel/config/xe/0000:00:02.0/sriov/max_vfs# echo 0000:00:02.0 > /sys/bus/pci/drivers/xe/bind

How to enable PF with support up to 3 VFs:

# echo 3 > /sys/kernel/config/xe/0000:00:02.0/sriov/max_vfs# echo 0000:00:02.0 > /sys/bus/pci/drivers/xe/bind

How to disable PF mode and always run as native:

# echo 0 > /sys/kernel/config/xe/0000:00:02.0/sriov/max_vfs# echo 0000:00:02.0 > /sys/bus/pci/drivers/xe/bind

This setting only takes effect when probing the device.

Remove devices

The created device directories can be removed usingrmdir:

# rmdir /sys/kernel/config/xe/0000:03:00.0/

Internal API

voidxe_configfs_check_device(structpci_dev*pdev)

Test if device was configured by configfs

Parameters

structpci_dev*pdev

thepci_dev device to test

Description

Try to find the configfs group that belongs to the specified pci deviceand print a diagnostic message if different than the default value.

boolxe_configfs_get_survivability_mode(structpci_dev*pdev)

get configfs survivability mode attribute

Parameters

structpci_dev*pdev

pci device

Return

survivability_mode attribute in configfs

boolxe_configfs_primary_gt_allowed(structpci_dev*pdev)

determine whether primary GTs are supported

Parameters

structpci_dev*pdev

pci device

Return

True if primary GTs are enabled, false if they have been disabled viaconfigfs.

boolxe_configfs_media_gt_allowed(structpci_dev*pdev)

determine whether media GTs are supported

Parameters

structpci_dev*pdev

pci device

Return

True if the media GTs are enabled, false if they have been disabledvia configfs.

u64xe_configfs_get_engines_allowed(structpci_dev*pdev)

get engine allowed mask from configfs

Parameters

structpci_dev*pdev

pci device

Return

engine mask with allowed engines set in configfs

boolxe_configfs_get_psmi_enabled(structpci_dev*pdev)

get configfs enable_psmi setting

Parameters

structpci_dev*pdev

pci device

Return

enable_psmi setting in configfs

u32xe_configfs_get_ctx_restore_mid_bb(structpci_dev*pdev,enumxe_engine_classclass,constu32**cs)

get configfs ctx_restore_mid_bb setting

Parameters

structpci_dev*pdev

pci device

enumxe_engine_classclass

hw engine class

constu32**cs

pointer to the bb to use - only valid during probe

Return

Number of dwords used in the mid_ctx_restore setting in configfs

u32xe_configfs_get_ctx_restore_post_bb(structpci_dev*pdev,enumxe_engine_classclass,constu32**cs)

get configfs ctx_restore_post_bb setting

Parameters

structpci_dev*pdev

pci device

enumxe_engine_classclass

hw engine class

constu32**cs

pointer to the bb to use - only valid during probe

Return

Number of dwords used in the post_ctx_restore setting in configfs

boolxe_configfs_admin_only_pf(structpci_dev*pdev)

Get PF’s operational mode.

Parameters

structpci_dev*pdev

thepci_dev device

Description

Find the configfs group that belongs to the PCI device and return a flagwhether the PF driver should be dedicated for VFs management only.

If configfs group is not present, use driver’s default value.

Return

true if PF driver is dedicated for VFs administration only.

unsignedintxe_configfs_get_max_vfs(structpci_dev*pdev)

Get number of VFs that could be managed

Parameters

structpci_dev*pdev

thepci_dev device

Description

Find the configfs group that belongs to the PCI device and return maximumnumber of Virtual Functions (VFs) that could be managed by this device.If configfs group is not present, use value of max_vfs module parameter.

Return

maximum number of VFs that could be managed.