PF: Add sriov_admin sysfs tree
Subsystems (2)
cover PF: Add sriov_admin sysfs tree
To allow the admin to provision VFs using production quality vGPU profiles, we need to expose some of the SR-IOV knobs at the sysfs level as existing debugfs entries may not be available.
Start with some basic scheduling parameters and apply them to all tiles/GTs where VF can be running. Provisioning of the other hard resources (like GGTT, VRAM) is still fully done by the PF, which uses fair auto-provisioning, but soon we will try to expose that to the admin too (after this series).
Below is an example of the new sysfs attributes tree on the PF with 2 of 3 VFs enabled (VFs were limited by max_vfs config):
/sys/bus/pci/drivers/xe/0000:00:02.0/sriov_admin
├── .bulk_profile
│ ├── exec_quantum_ms
│ ├── preempt_timeout_us
│ └── sched_priority
├── pf
│ ├── device -> ../../../0000:00:02.0
│ └── profile
│ ├── exec_quantum_ms
│ ├── preempt_timeout_us
│ └── sched_priority
├── vf1
│ ├── device -> ../../../0000:00:02.1
│ ├── profile
│ │ ├── exec_quantum_ms
│ │ ├── preempt_timeout_us
│ │ └── sched_priority
│ ├── reset
│ └── stop
├── vf2
│ ├── device -> ../../../0000:00:02.2
│ ├── profile
│ │ ├── exec_quantum_ms
│ │ ├── preempt_timeout_us
│ │ └── sched_priority
│ ├── reset
│ └── stop
└── vf3
├── profile
│ ├── exec_quantum_ms
│ ├── preempt_timeout_us
│ └── sched_priority
├── reset
└── stopMichal Wajdeczko (14):
drm/xe/pf: Prepare sysfs for SR-IOV admin attributes drm/xe/pf: Take RPM during calls to SR-IOV attr.store() drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs drm/xe/pf: Relax report helper to accept PF in bulk configs drm/xe/pf: Add functions to bulk configure EQ/PT on GT drm/xe/pf: Add functions to bulk provision EQ/PT drm/xe/pf: Allow bulk change all VFs EQ/PT using sysfs drm/xe/pf: Add functions to provision scheduling priority drm/xe/pf: Allow bulk change all VFs priority using sysfs drm/xe/pf: Allow change PF scheduling priority using sysfs drm/xe/pf: Promote xe_pci_sriov_get_vf_pdev drm/xe/pf: Add sysfs device symlinks to enabled VFs drm/xe/pf: Allow to stop and reset VF using sysfs drm/xe/pf: Add documentation for sriov_admin attributes
.../ABI/testing/sysfs-driver-intel-xe-sriov | 164 +++++ drivers/gpu/drm/xe/Makefile | 1 + drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c | 69 +- drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h | 2 + drivers/gpu/drm/xe/xe_pci_sriov.c | 42 +- drivers/gpu/drm/xe/xe_pci_sriov.h | 1 + drivers/gpu/drm/xe/xe_sriov_pf.c | 5 + drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 261 +++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 14 + drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 658 ++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h | 16 + drivers/gpu/drm/xe/xe_sriov_pf_types.h | 11 + 12 files changed, 1226 insertions(+), 18 deletions(-) create mode 100644 Documentation/ABI/testing/sysfs-driver-intel-xe-sriov create mode 100644 drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c create mode 100644 drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h
Signature · 1 line
2.47.1
PATCH 1/14 drm/xe/pf: Prepare sysfs for SR-IOV admin attributes
We already have some SR-IOV specific knobs exposed as debugfs files to allow low level tuning of the SR-IOV configurations, but those files are mainly for the use by the developers and debugfs might not be available on the production builds.
Start building dedicated sysfs sub-tree under xe device, where in upcoming patches we will add selected attributes that will help provision and manage PF and all VFs:
/sys/bus/pci/drivers/xe/BDF/
├── sriov_admin/
├── pf/
├── vf1/
├── vf2/
:
└── vfN/Add all required data types and helper macros that will be used by upcoming patches to define actual attributes.
--- drivers/gpu/drm/xe/Makefile | 1 + drivers/gpu/drm/xe/xe_sriov_pf.c | 5 + drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 295 +++++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h | 13 ++ drivers/gpu/drm/xe/xe_sriov_pf_types.h | 11 + 5 files changed, 325 insertions(+) create mode 100644 drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c create mode 100644 drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h
drivers/gpu/drm/xe/Makefile +1 -0
@@ -178,6 +178,7 @@ xe-$(CONFIG_PCI_IOV) += \ xe_sriov_pf_debugfs.o \ xe_sriov_pf_provision.o \ xe_sriov_pf_service.o \ + xe_sriov_pf_sysfs.o \ xe_tile_sriov_pf_debugfs.o # include helpers for tests even when XE is built-in
drivers/gpu/drm/xe/xe_sriov_pf.c +5 -0
@@ -16,6 +16,7 @@#include "xe_sriov_pf.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_service.h" +#include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" static unsigned int wanted_max_vfs (struct xe_device *xe)
@@ -128,6 +129,10 @@ int xe_sriov_pf_init_late(struct xe_device *xe)return err; } + err =xe_sriov_pf_sysfs_init (xe); +if (err) +return err; +return 0 ; }
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +295 -0
295 lines not on this page · raw message
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h new file +13 -0
@@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: MIT */ +/* +* Copyright © 2025 Intel Corporation +*/ + +#ifndef _XE_SRIOV_PF_SYSFS_H_ +#define _XE_SRIOV_PF_SYSFS_H_ + +struct xe_device ; + +int xe_sriov_pf_sysfs_init (struct xe_device *xe); + +#endif
drivers/gpu/drm/xe/xe_sriov_pf_types.h +11 -0
@@ -12,10 +12,15 @@#include "xe_sriov_pf_provision_types.h" #include "xe_sriov_pf_service_types.h" +struct kobject ; +/** * struct xe_sriov_metadata - per-VF device level metadata */ struct xe_sriov_metadata { +/** @kobj: kobject representing VF in PF's SR-IOV sysfs tree. */ +struct kobject *kobj; +/** @version: negotiated VF/PF ABI version */ struct xe_sriov_pf_service_version version; };
@@ -42,6 +47,12 @@ struct xe_device_pf {/** @service: device level service data. */ struct xe_sriov_pf_service service; +/** @sysfs: device level sysfs data. */ +struct { +/** @sysfs.root: the root kobject for all SR-IOV entries in sysfs. */ +struct kobject *root; + } sysfs; +/** @vfs: metadata for all VFs. */ struct xe_sriov_metadata *vfs; };
Signature · 1 line
2.47.1
Re: [PATCH 01/14] drm/xe/pf: Prepare sysfs for SR-IOV admin attributes
Quoted 9 lines · Michal Wajdeczko
We already have some SR-IOV specific knobs exposed as debugfs files to allow low level tuning of the SR-IOV configurations, but those files are mainly for the use by the developers and debugfs might not be available on the production builds.
Start building dedicated sysfs sub-tree under xe device, where in upcoming patches we will add selected attributes that will help provision and manage PF and all VFs:
Quoted 7 lines
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── pf/ ├── vf1/ ├── vf2/ : └── vfN/
Quoted 408 lines
Add all required data types and helper macros that will be used by upcoming patches to define actual attributes.
Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/Makefile | 1 + drivers/gpu/drm/xe/xe_sriov_pf.c | 5 + drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 295 +++++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h | 13 ++ drivers/gpu/drm/xe/xe_sriov_pf_types.h | 11 + 5 files changed, 325 insertions(+) create mode 100644 drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c create mode 100644 drivers/gpu/drm/xe/xe_sriov_pf_sysfs.hdrivers/gpu/drm/xe/Makefile +1
-0@@ -178,6 +178,7 @@ xe-$(CONFIG_PCI_IOV) += \ xe_sriov_pf_debugfs.o \ xe_sriov_pf_provision.o \ xe_sriov_pf_service.o \ + xe_sriov_pf_sysfs.o \ xe_tile_sriov_pf_debugfs.o # include helpers for tests even when XE is built-indrivers/gpu/drm/xe/xe_sriov_pf.c +5
-0@@ -16,6 +16,7 @@#include "xe_sriov_pf.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_service.h" +#include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" static unsigned int wanted_max_vfs (struct xe_device *xe)@@ -128,6 +129,10 @@ int xe_sriov_pf_init_late(struct xe_device *xe)return err; } + err =xe_sriov_pf_sysfs_init (xe); +if (err) +return err; +return 0 ; }drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +295
-0@@ -0,0 +1,295 @@ +// SPDX-License-Identifier: MIT +/* +* Copyright © 2025 Intel Corporation +*/ + +#include <linux/kobject.h> +#include <linux/sysfs.h> + +#include <drm/drm_managed.h> + +#include "xe_assert.h" +#include "xe_sriov.h" +#include "xe_sriov_pf_helpers.h" +#include "xe_sriov_pf_sysfs.h" +#include "xe_sriov_printk.h" + +/* +* /sys/bus/pci/drivers/xe/BDF/ +* : +* ├── sriov_admin/ +* ├── ... +* ├── pf/ +* │ ├── ... +* │ └── ... +* ├── vf1/ +* │ ├── ... +* │ └── ... +* ├── vf2/ +* : +* └── vfN/ +*/ + +struct xe_sriov_kobj { +struct kobject base; +struct xe_device *xe; +unsigned int vfid; +}; +#define to_xe_sriov_kobj (p) container_of_const((p), struct xe_sriov_kobj, base) + +struct xe_sriov_dev_attr { +struct attribute attr; +ssize_t (*show)(struct xe_device *xe,char *buf); +ssize_t (*store)(struct xe_device *xe,const char *buf,size_t count); +}; +#define to_xe_sriov_dev_attr (p) container_of_const((p), struct xe_sriov_dev_attr, attr) + +#define XE_SRIOV_DEV_ATTR (NAME ) \ +struct xe_sriov_dev_attr xe_sriov_dev_attr_##NAME = \ + __ATTR(NAME, 0644, xe_sriov_dev_attr_##NAME##_show, xe_sriov_dev_attr_##NAME##_store) + +#define XE_SRIOV_DEV_ATTR_RO (NAME ) \ +struct xe_sriov_dev_attr xe_sriov_dev_attr_##NAME = \ + __ATTR(NAME, 0444, xe_sriov_dev_attr_##NAME##_show, NULL) + +#define XE_SRIOV_DEV_ATTR_WO (NAME ) \ +struct xe_sriov_dev_attr xe_sriov_dev_attr_##NAME = \ + __ATTR(NAME, 0200, NULL, xe_sriov_dev_attr_##NAME##_store) + +struct xe_sriov_vf_attr { +struct attribute attr; +ssize_t (*show)(struct xe_device *xe,unsigned int vfid,char *buf); +ssize_t (*store)(struct xe_device *xe,unsigned int vfid,const char *buf,size_t count); +}; +#define to_xe_sriov_vf_attr (p) container_of_const((p), struct xe_sriov_vf_attr, attr) + +#define XE_SRIOV_VF_ATTR (NAME ) \ +struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \ + __ATTR(NAME, 0644, xe_sriov_vf_attr_##NAME##_show, xe_sriov_vf_attr_##NAME##_store) + +#define XE_SRIOV_VF_ATTR_RO (NAME ) \ +struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \ + __ATTR(NAME, 0444, xe_sriov_vf_attr_##NAME##_show, NULL) + +#define XE_SRIOV_VF_ATTR_WO (NAME ) \ +struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \ + __ATTR(NAME, 0200, NULL, xe_sriov_vf_attr_##NAME##_store) + +/* device level attributes go here */ + +static const struct attribute_group *xe_sriov_dev_attr_groups[] = { +NULL +}; + +/* and VF-level attributes go here */ + +static const struct attribute_group *xe_sriov_vf_attr_groups[] = { +NULL +}; + +/* no user serviceable parts below */ + +static struct kobject *new_xe_sriov_kobj (struct xe_device *xe,unsigned int vfid) +{ +struct xe_sriov_kobj *vkobj; + +xe_sriov_pf_assert_vfid (xe, vfid); + + vkobj =kzalloc (sizeof (*vkobj),GFP_KERNEL ); +if (!vkobj) +return NULL ; + + vkobj->xe = xe; + vkobj->vfid = vfid; +return &vkobj->base; +} + +static void release_xe_sriov_kobj (struct kobject *kobj) +{ +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); + +kfree (vkobj); +} + +static ssize_t xe_sriov_dev_attr_show (struct kobject *kobj,struct attribute *attr,char *buf) +{ +struct xe_sriov_dev_attr *vattr =to_xe_sriov_dev_attr (attr); +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); +struct xe_device *xe = vkobj->xe; + +if (!vattr->show) { +xe_sriov_dbg (xe,"attribute '%s' was not intended to be %s\n" , + attr->name,"readable" ); +return -EPERM ; + } +return vattr->show (xe, buf); +} + +static ssize_t xe_sriov_dev_attr_store (struct kobject *kobj,struct attribute *attr, +const char *buf,size_t count) +{ +struct xe_sriov_dev_attr *vattr =to_xe_sriov_dev_attr (attr); +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); +struct xe_device *xe = vkobj->xe; + +if (!vattr->store) { +xe_sriov_dbg (xe,"attribute '%s' was not intended to be %s\n" , + attr->name,"writable" ); +return -EPERM ; + } +return vattr->store (xe, buf, count); +} + +static ssize_t xe_sriov_vf_attr_show (struct kobject *kobj,struct attribute *attr,char *buf) +{ +struct xe_sriov_vf_attr *vattr =to_xe_sriov_vf_attr (attr); +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); +struct xe_device *xe = vkobj->xe; +unsigned int vfid = vkobj->vfid; + +xe_sriov_pf_assert_vfid (xe, vfid); + +if (!vattr->show) { +xe_sriov_dbg (xe,"attribute '%s' was not intended to be %s\n" , + attr->name,"readable" ); +return -EPERM ; + } +return vattr->show (xe, vfid, buf); +} + +static ssize_t xe_sriov_vf_attr_store (struct kobject *kobj,struct attribute *attr, +const char *buf,size_t count) +{ +struct xe_sriov_vf_attr *vattr =to_xe_sriov_vf_attr (attr); +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); +struct xe_device *xe = vkobj->xe; +unsigned int vfid = vkobj->vfid; + +xe_sriov_pf_assert_vfid (xe, vfid); + +if (!vattr->store) { +xe_sriov_dbg (xe,"attribute '%s' was not intended to be %s\n" , + attr->name,"writable" ); +return -EPERM ; + } +return vattr->store (xe, vfid, buf, count); +} + +static const struct sysfs_ops xe_sriov_dev_sysfs_ops = { + .show = xe_sriov_dev_attr_show, + .store = xe_sriov_dev_attr_store, +}; + +static const struct sysfs_ops xe_sriov_vf_sysfs_ops = { + .show = xe_sriov_vf_attr_show, + .store = xe_sriov_vf_attr_store, +}; + +static const struct kobj_type xe_sriov_dev_ktype = { + .release = release_xe_sriov_kobj, + .sysfs_ops = &xe_sriov_dev_sysfs_ops, + .default_groups = xe_sriov_dev_attr_groups, +}; + +static const struct kobj_type xe_sriov_vf_ktype = { + .release = release_xe_sriov_kobj, + .sysfs_ops = &xe_sriov_vf_sysfs_ops, + .default_groups = xe_sriov_vf_attr_groups, +}; + +static int pf_sysfs_error (struct xe_device *xe,int err,const char *what) +{ +if (IS_ENABLED (CONFIG_DRM_XE_DEBUG )) +xe_sriov_dbg (xe,"Failed to setup sysfs %s (%pe)\n" , what,ERR_PTR (err)); +return err; +} + +static void action_put_kobject (void *arg) +{ +struct kobject *kobj = arg; + +kobject_put (kobj); +} + +static int pf_setup_root (struct xe_device *xe) +{ +struct kobject *parent = &xe->drm.dev->kobj; +struct kobject *root; +int err; + + root =new_xe_sriov_kobj (xe,PFID ); +if (!root) +return pf_sysfs_error (xe, -ENOMEM ,"root obj" ); + + err =devm_add_action_or_reset (xe->drm.dev, action_put_kobject, root); +if (err) +return pf_sysfs_error (xe, err,"root action" ); + + err =kobject_init_and_add (root, &xe_sriov_dev_ktype, parent,"sriov_admin" ); +if (err) +return pf_sysfs_error (xe, err,"root init" ); + +xe_assert (xe,IS_SRIOV_PF (xe)); +xe_assert (xe, !xe->sriov.pf.sysfs.root); + xe->sriov.pf.sysfs.root = root; +return 0 ; +} + +static int pf_setup_tree (struct xe_device *xe) +{ +unsigned int totalvfs =xe_sriov_pf_get_totalvfs (xe); +struct kobject *root, *kobj; +unsigned int n; +int err; + +xe_assert (xe,IS_SRIOV_PF (xe)); + root = xe->sriov.pf.sysfs.root; + +for (n =0 ; n <= totalvfs; n++) { + kobj =new_xe_sriov_kobj (xe,VFID (n)); +if (!kobj) +return pf_sysfs_error (xe, -ENOMEM ,"tree obj" ); + + err =devm_add_action_or_reset (xe->drm.dev, action_put_kobject, root); +if (err) +return pf_sysfs_error (xe, err,"tree action" ); + +if (n) + err =kobject_init_and_add (kobj, &xe_sriov_vf_ktype, + root,"vf%u" , n); +else + err =kobject_init_and_add (kobj, &xe_sriov_vf_ktype, + root,"pf" ); +if (err) +return pf_sysfs_error (xe, err,"tree init" ); + +xe_assert (xe, !xe->sriov.pf.vfs[n].kobj); + xe->sriov.pf.vfs[n].kobj = kobj; + } + +return 0 ; +} + +/** +* xe_sriov_pf_sysfs_init() - Setup PF's SR-IOV sysfs tree. +* @xe: the PF &xe_device to setup sysfs +* +* This function will create additional nodes that will represent PF and VFs +* devices, each populated with SR-IOV Xe specific attributes. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_sysfs_init (struct xe_device *xe) +{ +int err; + + err =pf_setup_root (xe); +if (err) +return err; + + err =pf_setup_tree (xe); +if (err) +return err; + +return 0 ; +}drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h new file +13
-0@@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: MIT */ +/* +* Copyright © 2025 Intel Corporation +*/ + +#ifndef _XE_SRIOV_PF_SYSFS_H_ +#define _XE_SRIOV_PF_SYSFS_H_ + +struct xe_device ; + +int xe_sriov_pf_sysfs_init (struct xe_device *xe); + +#endif drivers/gpu/drm/xe/xe_sriov_pf_types.h +11
-0@@ -12,10 +12,15 @@#include "xe_sriov_pf_provision_types.h" #include "xe_sriov_pf_service_types.h" +struct kobject ; +/** * struct xe_sriov_metadata - per-VF device level metadata */ struct xe_sriov_metadata { +/** @kobj: kobject representing VF in PF's SR-IOV sysfs tree. */ +struct kobject *kobj; +/** @version: negotiated VF/PF ABI version */ struct xe_sriov_pf_service_version version; };@@ -42,6 +47,12 @@ struct xe_device_pf {/** @service: device level service data. */ struct xe_sriov_pf_service service; +/** @sysfs: device level sysfs data. */ +struct { +/** @sysfs.root: the root kobject for all SR-IOV entries in sysfs. */ +struct kobject *root; + } sysfs; +/** @vfs: metadata for all VFs. */ struct xe_sriov_metadata *vfs; };Signature · 1 line
2.47.1
Re: [PATCH 01/14] drm/xe/pf: Prepare sysfs for SR-IOV admin attributes
In reply to Michal Wajdeczkodrivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 152
Quoted 10 lines · Michal Wajdeczko
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +10
-0+static ssize_t xe_sriov_vf_attr_show (struct kobject *kobj,struct attribute *attr,char *buf) +{ +struct xe_sriov_vf_attr *vattr =to_xe_sriov_vf_attr (attr); +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); +struct xe_device *xe = vkobj->xe; +unsigned int vfid = vkobj->vfid; + +xe_sriov_pf_assert_vfid (xe, vfid); + +if (!vattr->show) {
nit: I don't like hiding this programmer error behind a dbg. If the perm was set to read, the attribute was supposed to be defined with a show method.
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 190
Quoted 3 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +3
-0+static const struct kobj_type xe_sriov_dev_ktype = { + .release = release_xe_sriov_kobj, + .sysfs_ops = &xe_sriov_dev_sysfs_ops,
(general comment, feel free to ignore): I wonder if we could inline the ops definition since it was not supposed to be used outside.
Quoted 20 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +20
-0+ .default_groups = xe_sriov_dev_attr_groups, +}; + +static const struct kobj_type xe_sriov_vf_ktype = { + .release = release_xe_sriov_kobj, + .sysfs_ops = &xe_sriov_vf_sysfs_ops, + .default_groups = xe_sriov_vf_attr_groups, +}; + +static int pf_sysfs_error (struct xe_device *xe,int err,const char *what) +{ +if (IS_ENABLED (CONFIG_DRM_XE_DEBUG )) +xe_sriov_dbg (xe,"Failed to setup sysfs %s (%pe)\n" , what,ERR_PTR (err)); +return err; +} + +static void action_put_kobject (void *arg) +{ +struct kobject *kobj = arg; +
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 220
Quoted 10 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +10
-0+kobject_put (kobj); +} + +static int pf_setup_root (struct xe_device *xe) +{ +struct kobject *parent = &xe->drm.dev->kobj; +struct kobject *root; +int err; + + root =new_xe_sriov_kobj (xe,PFID );
Slight preference to follow the _create() name convention in xe, so this would be:
root = xe_sriov_kobj_create(xe, PFID);
but won't block on it.
thanks,
Lucas De Marchi
Re: [PATCH 01/14] drm/xe/pf: Prepare sysfs for SR-IOV admin attributes
Quoted 15 lines · Lucas De Marchi
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 152
Quoted 10 lines · Michal Wajdeczko
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +10
-0+static ssize_t xe_sriov_vf_attr_show (struct kobject *kobj,struct attribute *attr,char *buf) +{ + struct xe_sriov_vf_attr *vattr = to_xe_sriov_vf_attr(attr); + struct xe_sriov_kobj *vkobj = to_xe_sriov_kobj(kobj); + struct xe_device *xe = vkobj->xe; + unsigned int vfid = vkobj->vfid; + + xe_sriov_pf_assert_vfid(xe, vfid); + + if (!vattr->show) {nit: I don't like hiding this programmer error behind a dbg. If the perm was set to read, the attribute was supposed to be defined with a show method.
it is not trying to hide programmer error, it is trying to protect kernel
if bad admin will change perm of the attribute using chmod a+rw and then will try to read this we will get NPD here
Quoted 7 lines
Quoted 4 lines
+static const struct
kobj_typexe_sriov_dev_ktype= { + .release =release_xe_sriov_kobj,
+ .sysfs_ops =&xe_sriov_dev_sysfs_ops,(general comment, feel free to ignore): I wonder if we could inline the ops definition since it was not supposed to be used outside.
it's all static already, or I get your comment wrong?
Quoted 25 lines
Quoted 25 lines
+ .default_groups =
xe_sriov_dev_attr_groups,drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +23
-0+}; + +static const struct kobj_type xe_sriov_vf_ktype = { + .release = release_xe_sriov_kobj, + .sysfs_ops = &xe_sriov_vf_sysfs_ops, + .default_groups = xe_sriov_vf_attr_groups, +}; + +static int pf_sysfs_error (struct xe_device *xe,int err,const char *what) +{ + if (IS_ENABLED(CONFIG_DRM_XE_DEBUG)) + xe_sriov_dbg(xe, "Failed to setup sysfs %s (%pe)\n", what, ERR_PTR(err)); + return err; +} + +static void action_put_kobject (void *arg) +{ + struct kobject *kobj = arg; + + kobject_put(kobj); +} + +static int pf_setup_root (struct xe_device *xe)
Quoted 8 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 220
Quoted 6 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +6
-0+{ + struct kobject *parent = &xe->drm.dev->kobj; + struct kobject *root; + int err; + + root = new_xe_sriov_kobj(xe, PFID);Slight preference to follow the
_create()name convention in xe, so this
hmm, I just didn't want to break the other rule, that function name should start with or match the first param name ;)
in this case it is a factory style, that takes xe, so started with a verb, also somehow matches release_xe_sriov_kobj() that takes plain kobj
would create_xe_sriov_kobj() work for you any better?
Quoted 7 lines
would be:
root =
xe_sriov_kobj_create(xe, PFID);but won't block on it.
Reviewed-by Lucas De Marchi <[email protected]>
thanks!
Quoted 3 lines
thanks,
Lucas De Marchi
Re: [PATCH 01/14] drm/xe/pf: Prepare sysfs for SR-IOV admin attributes
Quoted 19 lines · Michal Wajdeczko
Quoted 15 lines · Lucas De Marchi
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 152
Quoted 10 lines · Michal Wajdeczko
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +10
-0+static ssize_t xe_sriov_vf_attr_show (struct kobject *kobj,struct attribute *attr,char *buf) +{ + struct xe_sriov_vf_attr *vattr = to_xe_sriov_vf_attr(attr); + struct xe_sriov_kobj *vkobj = to_xe_sriov_kobj(kobj); + struct xe_device *xe = vkobj->xe; + unsigned int vfid = vkobj->vfid; + + xe_sriov_pf_assert_vfid(xe, vfid); + + if (!vattr->show) {nit: I don't like hiding this programmer error behind a dbg. If the perm was set to read, the attribute was supposed to be defined with a show method.
Quoted 4 lines
it is not trying to hide programmer error, it is trying to protect kernel
if bad admin will change perm of the attribute using chmod a+rw and then will try to read this we will get NPD here
humn... right, I was thinking that the show would be NULL and kobj already checks for that. But yes, with the sysfs_ops define below we do need it.
However samples/kobject/kset-example.c uses the following style and return code:
if (!attribute->show)
return -EIO;
return attribute->show(foo, attribute, buf);Quoted 10 lines
Quoted 8 lines
Quoted 5 lines
+static const struct
kobj_typexe_sriov_dev_ktype= { + .release =release_xe_sriov_kobj,
+ .sysfs_ops =&xe_sriov_dev_sysfs_ops,(general comment, feel free to ignore): I wonder if we could inline the ops definition since it was not supposed to be used outside.
it's all static already, or I get your comment wrong?
since it's only to be used here, I was thinking on inlining the definition like .sysfs_ops = &(struct ...){ };
let's ignore that.
Quoted 35 lines
Quoted 26 lines
Quoted 26 lines
+ .default_groups =
xe_sriov_dev_attr_groups,drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +23
-0+}; + +static const struct kobj_type xe_sriov_vf_ktype = { + .release = release_xe_sriov_kobj, + .sysfs_ops = &xe_sriov_vf_sysfs_ops, + .default_groups = xe_sriov_vf_attr_groups, +}; + +static int pf_sysfs_error (struct xe_device *xe,int err,const char *what) +{ + if (IS_ENABLED(CONFIG_DRM_XE_DEBUG)) + xe_sriov_dbg(xe, "Failed to setup sysfs %s (%pe)\n", what, ERR_PTR(err)); + return err; +} + +static void action_put_kobject (void *arg) +{ + struct kobject *kobj = arg; + + kobject_put(kobj); +} + +static int pf_setup_root (struct xe_device *xe)Quoted 8 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c line 220
Quoted 6 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c new file +6
-0+{ + struct kobject *parent = &xe->drm.dev->kobj; + struct kobject *root; + int err; + + root = new_xe_sriov_kobj(xe, PFID);Slight preference to follow the
_create()name convention in xe, so this
Quoted 5 lines
hmm, I just didn't want to break the other rule, that function name should start with or match the first param name ;)
in this case it is a factory style, that takes xe, so started with a verb, also somehow matches
release_xe_sriov_kobj()that takes plain kobj
ok
Quoted 2 lines
would
create_xe_sriov_kobj()work for you any better?
yes
Lucas De Marchi
Quoted 15 lines
Quoted 8 lines
would be:
root =
xe_sriov_kobj_create(xe, PFID);but won't block on it.
Reviewed-by Lucas De Marchi <[email protected]>thanks!
Quoted 3 lines
thanks,
Lucas De Marchi
PATCH 2/14 drm/xe/pf: Take RPM during calls to SR-IOV attr.store()
In reply to Michal WajdeczkoWe expect that all SR-IOV attr.store() handlers will require active runtime PM reference. To simplify implementation of those handlers, take an implicit RPM reference on their behalf. Also wait until PF completes its restart.
--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-)
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +16 -2
@@ -9,7 +9,9 @@#include <drm/drm_managed.h> #include "xe_assert.h" +#include "xe_pm.h" #include "xe_sriov.h" +#include "xe_sriov_pf.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h"
@@ -131,13 +133,19 @@ static ssize_t xe_sriov_dev_attr_store(struct kobject *kobj, struct attribute *astruct xe_sriov_dev_attr *vattr =to_xe_sriov_dev_attr (attr);struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj);struct xe_device *xe = vkobj->xe; +ssize_t ret;if (!vattr->store) {xe_sriov_dbg (xe,"attribute '%s' was not intended to be %s\n" , attr->name,"writable" );return -EPERM ; } -return vattr->store (xe, buf, count); + +xe_pm_runtime_get (xe); + ret =xe_sriov_pf_wait_ready (xe) ?: vattr->store (xe, buf, count); +xe_pm_runtime_put (xe); + +return ret; }static ssize_t xe_sriov_vf_attr_show (struct kobject *kobj,struct attribute *attr,char *buf)
@@ -164,6 +172,7 @@ static ssize_t xe_sriov_vf_attr_store(struct kobject *kobj, struct attribute *atstruct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj);struct xe_device *xe = vkobj->xe;unsigned int vfid = vkobj->vfid; +ssize_t ret;xe_sriov_pf_assert_vfid (xe, vfid);
@@ -172,7 +181,12 @@ static ssize_t xe_sriov_vf_attr_store(struct kobject *kobj, struct attribute *at attr->name, "writable");return -EPERM ; } -return vattr->store (xe, vfid, buf, count); + +xe_pm_runtime_get (xe); + ret =xe_sriov_pf_wait_ready (xe) ?: vattr->store (xe, vfid, buf, count); +xe_pm_runtime_get (xe); + +return ret; }static const struct sysfs_ops xe_sriov_dev_sysfs_ops = {
Signature · 1 line
2.47.1
Re: [PATCH 02/14] drm/xe/pf: Take RPM during calls to SR-IOV attr.store()
Quoted 6 lines · Michal Wajdeczko
We expect that all SR-IOV
attr.store()handlers will require active runtime PM reference. To simplify implementation of those handlers, take an implicit RPM reference on their behalf. Also wait until PF completes its restart.Signed-off-by Michal Wajdeczko <[email protected]>
thanks,
Lucas De Marchi
PATCH 3/14 drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs
In reply to Michal WajdeczkoOn current platforms, in SR-IOV virtualization, the GPU is shared between VFs on the time-slice basis. The 'execution quantum' (EQ) and 'preemption timeout' (PT) are two main scheduling parameters that could be set individually per each VF.
Add EQ/PT read-write attributes for the PF and all VFs.
By exposing those two parameters over sysfs, the admin can change their default values (infinity) and let the GuC scheduler enforce that settings.
/sys/bus/pci/drivers/xe/BDF/
├── sriov_admin/
├── pf/
│ └── profile
│ ├── exec_quantum_ms [RW] unsigned integer
│ └── preempt_timeout_us [RW] unsigned integer
├── vf1/
│ └── profile
│ ├── exec_quantum_ms [RW] unsigned integer
│ └── preempt_timeout_us [RW] unsigned integerWriting 0 to these files will set infinity EQ/PT for the VF on all tiles/GTs. This is a default value. Writing non-zero integers to these files will change EQ/PT to new value (in their respective
units: msec or usec).
Reading from these files will return EQ/PT as previously set on all tiles/GTs. In case of inconsistent values detected, due to errors or low-level configuration done using debugfs, -EUCLEAN error will be returned.
--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 116 +++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 8 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 54 +++++++++- 3 files changed, 176 insertions(+), 2 deletions(-)
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +116 -0
@@ -152,3 +152,119 @@ int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision xe->sriov.pf.provision.mode = mode;return 0 ; } + +/** +* xe_sriov_pf_provision_apply_vf_eq() - Change VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: execution quantum in [ms] to set +* +* Change VF's execution quantum (EQ) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_exec_quantum (gt, vfid, eq); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_query_vf_eq() - Query VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: placeholder for the returned execution quantum in [ms] +* +* Query VF's execution quantum (EQ) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_query_vf_eq (struct xe_device *xe,unsigned int vfid,u32 *eq) +{ +struct xe_gt *gt; +unsigned int id; +int count =0 ; +u32 value; + +for_each_gt (gt, xe, id) { + value =xe_gt_sriov_pf_config_get_exec_quantum (gt, vfid); +if (!count++) + *eq = value; +else if (value != *eq) +return -EUCLEAN ; + } + +return !count ? -ENODATA :0 ; +} + +/** +* xe_sriov_pf_provision_apply_vf_pt() - Change VF's preemption timeout. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @pt: preemption timeout in [us] to set +* +* Change VF's preemption timeout (PT) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_preempt_timeout (gt, vfid, pt); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_query_vf_pt() - Query VF's preemption timeout. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @pt: placeholder for the returned preemption timeout in [us] +* +* Query VF's preemption timeout (PT) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt) +{ +struct xe_gt *gt; +unsigned int id; +int count =0 ; +u32 value; + +for_each_gt (gt, xe, id) { + value =xe_gt_sriov_pf_config_get_preempt_timeout (gt, vfid); +if (!count++) + *pt = value; +else if (value != *pt) +return -EUCLEAN ; + } + +return !count ? -ENODATA :0 ; +}
drivers/gpu/drm/xe/xe_sriov_pf_provision.h +8 -0
@@ -6,10 +6,18 @@#ifndef _XE_SRIOV_PF_PROVISION_H_#define _XE_SRIOV_PF_PROVISION_H_ +#include <linux/types.h> +#include "xe_sriov_pf_provision_types.h" struct xe_device ; +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq); +int xe_sriov_pf_provision_query_vf_eq (struct xe_device *xe,unsigned int vfid,u32 *eq); + +int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt); +int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt); +int xe_sriov_pf_provision_vfs (struct xe_device *xe,unsigned int num_vfs);int xe_sriov_pf_unprovision_vfs (struct xe_device *xe,unsigned int num_vfs);
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +52 -2
75 lines not on this page · raw message
Signature · 1 line
2.47.1
Re: [PATCH 03/14] drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs
Quoted 12 lines · Michal Wajdeczko
On current platforms, in SR-IOV virtualization, the GPU is shared between VFs on the time-slice basis. The 'execution quantum' (EQ) and 'preemption timeout' (PT) are two main scheduling parameters that could be set individually per each VF.
Add EQ/PT read-write attributes for the PF and all VFs.
By exposing those two parameters over sysfs, the admin can change their default values (infinity) and let the GuC scheduler enforce that settings.
/sys/bus/pci/drivers/xe/BDF/
Quoted 9 lines
├── sriov_admin/
├── pf/
│ └── profile
│ ├── exec_quantum_ms [RW] unsigned integer
│ └── preempt_timeout_us [RW] unsigned integer
├── vf1/
│ └── profile
│ ├── exec_quantum_ms [RW] unsigned integer
│ └── preempt_timeout_us [RW] unsigned integerQuoted 252 lines
Writing 0 to these files will set infinity EQ/PT for the VF on all tiles/GTs. This is a default value. Writing non-zero integers to these files will change EQ/PT to new value (in their respective
units: msec or usec).Reading from these files will return EQ/PT as previously set on all tiles/GTs. In case of inconsistent values detected, due to errors or low-level configuration done using debugfs,
-EUCLEANerror will be returned.Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 116 +++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 8 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 54 +++++++++- 3 files changed, 176 insertions(+), 2 deletions(-)drivers/gpu/drm/xe/xe_sriov_pf_provision.c +116
-0@@ -152,3 +152,119 @@ int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision xe->sriov.pf.provision.mode = mode;return 0 ; } + +/** +* xe_sriov_pf_provision_apply_vf_eq() - Change VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: execution quantum in [ms] to set +* +* Change VF's execution quantum (EQ) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_exec_quantum (gt, vfid, eq); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_query_vf_eq() - Query VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: placeholder for the returned execution quantum in [ms] +* +* Query VF's execution quantum (EQ) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_query_vf_eq (struct xe_device *xe,unsigned int vfid,u32 *eq) +{ +struct xe_gt *gt; +unsigned int id; +int count =0 ; +u32 value; + +for_each_gt (gt, xe, id) { + value =xe_gt_sriov_pf_config_get_exec_quantum (gt, vfid); +if (!count++) + *eq = value; +else if (value != *eq) +return -EUCLEAN ; + } + +return !count ? -ENODATA :0 ; +} + +/** +* xe_sriov_pf_provision_apply_vf_pt() - Change VF's preemption timeout. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @pt: preemption timeout in [us] to set +* +* Change VF's preemption timeout (PT) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_preempt_timeout (gt, vfid, pt); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_query_vf_pt() - Query VF's preemption timeout. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @pt: placeholder for the returned preemption timeout in [us] +* +* Query VF's preemption timeout (PT) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt) +{ +struct xe_gt *gt; +unsigned int id; +int count =0 ; +u32 value; + +for_each_gt (gt, xe, id) { + value =xe_gt_sriov_pf_config_get_preempt_timeout (gt, vfid); +if (!count++) + *pt = value; +else if (value != *pt) +return -EUCLEAN ; + } + +return !count ? -ENODATA :0 ; +}drivers/gpu/drm/xe/xe_sriov_pf_provision.h +8
-0@@ -6,10 +6,18 @@#ifndef _XE_SRIOV_PF_PROVISION_H_#define _XE_SRIOV_PF_PROVISION_H_ +#include <linux/types.h> +#include "xe_sriov_pf_provision_types.h" struct xe_device ; +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq); +int xe_sriov_pf_provision_query_vf_eq (struct xe_device *xe,unsigned int vfid,u32 *eq); + +int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt); +int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt); +int xe_sriov_pf_provision_vfs (struct xe_device *xe,unsigned int num_vfs);int xe_sriov_pf_unprovision_vfs (struct xe_device *xe,unsigned int num_vfs);drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +52
-2@@ -13,6 +13,7 @@#include "xe_sriov.h" #include "xe_sriov_pf.h" #include "xe_sriov_pf_helpers.h" +#include "xe_sriov_pf_provision.h" #include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" @@ -23,10 +24,14 @@* ├── ... * ├── pf/ * │ ├── ... -* │ └── ... +* │ └── profile +* │ ├── exec_quantum_ms +* │ └── preempt_timeout_us * ├── vf1/ * │ ├── ... -* │ └── ... +* │ └── profile +* │ ├── exec_quantum_ms +* │ └── preempt_timeout_us * ├── vf2/ * : * └── vfN/ @@ -85,7 +90,52 @@ static const struct attribute_group *xe_sriov_dev_attr_groups[] = {/* and VF-level attributes go here */ +#define DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (NAME ,ITEM ,TYPE ,FORMAT ) \ +static ssize_t xe_sriov_vf_attr_##NAME##_show(struct xe_device *xe, unsigned int vfid, \ + char *buf) \ +{ \ + TYPE value = 0; \ + int err; \ + \ + err = xe_sriov_pf_provision_query_vf_##ITEM(xe, vfid, &value); \ + if (err) \ + return err; \ + \ + return sysfs_emit(buf, FORMAT, value); \ +} \ + \ +static ssize_t xe_sriov_vf_attr_##NAME##_store(struct xe_device *xe, unsigned int vfid, \ + const char *buf, size_t count) \ +{ \ + TYPE value; \ + int err; \ + \ + err = kstrto##TYPE(buf, 0, &value); \ + if (err) \ + return err; \ + \ + err = xe_sriov_pf_provision_apply_vf_##ITEM(xe, vfid, value); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_VF_ATTR(NAME) + +DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (exec_quantum_ms, eq, u32,"%u\n" ); +DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (preempt_timeout_us, pt, u32,"%u\n" ); + +static struct attribute *profile_vf_attrs[] = { + &xe_sriov_vf_attr_exec_quantum_ms.attr, + &xe_sriov_vf_attr_preempt_timeout_us.attr, +NULL +}; + +static const struct attribute_group profile_vf_attr_group = { + .name ="profile" , + .attrs = profile_vf_attrs, +}; +static const struct attribute_group *xe_sriov_vf_attr_groups[] = { + &profile_vf_attr_group,NULL };Signature · 1 line
2.47.1
Re: [PATCH 03/14] drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs
In reply to Michal WajdeczkoQuoted 61 lines · Michal Wajdeczko
On current platforms, in SR-IOV virtualization, the GPU is shared between VFs on the time-slice basis. The 'execution quantum' (EQ) and 'preemption timeout' (PT) are two main scheduling parameters that could be set individually per each VF.
Add EQ/PT read-write attributes for the PF and all VFs.
By exposing those two parameters over sysfs, the admin can change their default values (infinity) and let the GuC scheduler enforce that settings.
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── pf/ │ └── profile │ ├── exec_quantum_ms [RW] unsigned integer │ └── preempt_timeout_us [RW] unsigned integer ├── vf1/ │ └── profile │ ├── exec_quantum_ms [RW] unsigned integer │ └── preempt_timeout_us [RW] unsigned integerWriting 0 to these files will set infinity EQ/PT for the VF on all tiles/GTs. This is a default value. Writing non-zero integers to these files will change EQ/PT to new value (in their respective
units: msec or usec).Reading from these files will return EQ/PT as previously set on all tiles/GTs. In case of inconsistent values detected, due to errors or low-level configuration done using debugfs,
-EUCLEANerror will be returned.Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>---
drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 116 +++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 8 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 54 +++++++++- 3 files changed, 176 insertions(+), 2 deletions(-)drivers/gpu/drm/xe/xe_sriov_pf_provision.c +12
-0@@ -152,3 +152,119 @@ int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision xe->sriov.pf.provision.mode = mode;return 0 ; } + +/** +* xe_sriov_pf_provision_apply_vf_eq() - Change VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: execution quantum in [ms] to set +* +* Change VF's execution quantum (EQ) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure.
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision · line 176
Quoted 10 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +10
-0+*/ +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_exec_quantum (gt, vfid, eq);
I find it weird that a race on multiple writes would basically give undeterministic results for the device. We are locking/releaseing xe_gt_sriov_pf_master_mutex on the gt, but shouldn't we syncronize multiple writes at the device level?
Quoted 4 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +4
-0+ result = result ?: err; + } + +return result;
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision · line 190
Quoted 10 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +10
-0+} + +/** +* xe_sriov_pf_provision_query_vf_eq() - Query VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: placeholder for the returned execution quantum in [ms] +* +* Query VF's execution quantum (EQ) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned.
or if you had multiple userspace components trying to set the value creating the race above.
Lucas De Marchi
Re: [PATCH 03/14] drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs
Quoted 70 lines · Lucas De Marchi
Quoted 69 lines · Michal Wajdeczko
On current platforms, in SR-IOV virtualization, the GPU is shared between VFs on the time-slice basis. The 'execution quantum' (EQ) and 'preemption timeout' (PT) are two main scheduling parameters that could be set individually per each VF.
Add EQ/PT read-write attributes for the PF and all VFs.
By exposing those two parameters over sysfs, the admin can change their default values (infinity) and let the GuC scheduler enforce that settings.
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── pf/ │ └── profile │ ├── exec_quantum_ms [RW] unsigned integer │ └── preempt_timeout_us [RW] unsigned integer ├── vf1/ │ └── profile │ ├── exec_quantum_ms [RW] unsigned integer │ └── preempt_timeout_us [RW] unsigned integerWriting 0 to these files will set infinity EQ/PT for the VF on all tiles/GTs. This is a default value. Writing non-zero integers to these files will change EQ/PT to new value (in their respective
units: msec or usec).Reading from these files will return EQ/PT as previously set on all tiles/GTs. In case of inconsistent values detected, due to errors or low-level configuration done using debugfs,
-EUCLEANerror will be returned.Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>---
drivers/gpu/drm/xe/xe_sriov_pf_provision.c| 116 +++++++++++++++++++++drivers/gpu/drm/xe/xe_sriov_pf_provision.h| 8 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 54 +++++++++-3 files changed, 176 insertions(+), 2 deletions(-)drivers/gpu/drm/xe/xe_sriov_pf_provision.c +20
-0@@ -152,3 +152,119 @@ int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision xe->sriov.pf.provision.mode = mode; return 0; } + +/** +* xe_sriov_pf_provision_apply_vf_eq() - Change VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: execution quantum in [ms] to set +* +* Change VF's execution quantum (EQ) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq) +{ + struct xe_gt *gt; + unsigned int id; + int result = 0; + int err; +
Quoted 8 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision · line 176
Quoted 2 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +2
-0+ for_each_gt(gt, xe, id) { + err = xe_gt_sriov_pf_config_set_exec_quantum(gt, vfid, eq);I find it weird that a race on multiple writes would basically give undeterministic results for the device. We are locking/releaseing
xe_gt_sriov_pf_master_mutexon the gt, but shouldn't we syncronize multiple writes at the device level?
but we still expose more granular attributes over debugfs, so even if we would be successful in this protected round here, the actual configuration could became "unclean" soon anyway, so I'm not sure there is any benefit trying to add this extra protection
Quoted 10 lines
Quoted 10 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +9
-0+ result = result ?: err; + } + + return result; +} + +/** +* xe_sriov_pf_provision_query_vf_eq() - Query VF's execution quantum. +* @xe: the PF &xe_device
Quoted 8 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision · line 190
Quoted 5 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +5
-0+* @vfid: the VF identifier +* @eq: placeholder for the returned execution quantum in [ms] +* +* Query VF's execution quantum (EQ) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. or if you had multiple userspace components trying to set the value creating the race above.
yes, this "inconsistency" can be mostly introduced by the userspace and concurrency / race is not only sysfs vs sysfs but also sysfs vs debugfs
Quoted 2 lines
Lucas De Marchi
Re: [PATCH 03/14] drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs
Quoted 106 lines · Michal Wajdeczko
Quoted 70 lines · Lucas De Marchi
Quoted 69 lines · Michal Wajdeczko
On current platforms, in SR-IOV virtualization, the GPU is shared between VFs on the time-slice basis. The 'execution quantum' (EQ) and 'preemption timeout' (PT) are two main scheduling parameters that could be set individually per each VF.
Add EQ/PT read-write attributes for the PF and all VFs.
By exposing those two parameters over sysfs, the admin can change their default values (infinity) and let the GuC scheduler enforce that settings.
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── pf/ │ └── profile │ ├── exec_quantum_ms [RW] unsigned integer │ └── preempt_timeout_us [RW] unsigned integer ├── vf1/ │ └── profile │ ├── exec_quantum_ms [RW] unsigned integer │ └── preempt_timeout_us [RW] unsigned integerWriting 0 to these files will set infinity EQ/PT for the VF on all tiles/GTs. This is a default value. Writing non-zero integers to these files will change EQ/PT to new value (in their respective
units: msec or usec).Reading from these files will return EQ/PT as previously set on all tiles/GTs. In case of inconsistent values detected, due to errors or low-level configuration done using debugfs,
-EUCLEANerror will be returned.Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>---
drivers/gpu/drm/xe/xe_sriov_pf_provision.c| 116 +++++++++++++++++++++drivers/gpu/drm/xe/xe_sriov_pf_provision.h| 8 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 54 +++++++++-3 files changed, 176 insertions(+), 2 deletions(-)drivers/gpu/drm/xe/xe_sriov_pf_provision.c +20
-0@@ -152,3 +152,119 @@ int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision xe->sriov.pf.provision.mode = mode; return 0; } + +/** +* xe_sriov_pf_provision_apply_vf_eq() - Change VF's execution quantum. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @eq: execution quantum in [ms] to set +* +* Change VF's execution quantum (EQ) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq) +{ + struct xe_gt *gt; + unsigned int id; + int result = 0; + int err; +Quoted 8 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision · line 176
Quoted 2 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +2
-0+ for_each_gt(gt, xe, id) { + err = xe_gt_sriov_pf_config_set_exec_quantum(gt, vfid, eq);I find it weird that a race on multiple writes would basically give undeterministic results for the device. We are locking/releaseing
xe_gt_sriov_pf_master_mutexon the gt, but shouldn't we syncronize multiple writes at the device level?but we still expose more granular attributes over debugfs, so even if we would be successful in this protected round here, the actual configuration could became "unclean" soon anyway, so I'm not sure there is any benefit trying to add this extra protection
Quoted 10 lines
Quoted 10 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +9
-0+ result = result ?: err; + } + + return result; +} + +/** +* xe_sriov_pf_provision_query_vf_eq() - Query VF's execution quantum. +* @xe: the PF &xe_device Quoted 8 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provision · line 190
Quoted 5 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +5
-0+* @vfid: the VF identifier +* @eq: placeholder for the returned execution quantum in [ms] +* +* Query VF's execution quantum (EQ) provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. or if you had multiple userspace components trying to set the value creating the race above.
Quoted 2 lines
yes, this "inconsistency" can be mostly introduced by the userspace and concurrency / race is not only sysfs vs sysfs but also sysfs vs debugfs
right, but the point is that a normal (non-developer, production) user shouldn't be messing with our debugfs. At that point, the only concurrency is sysfs vs sysfs.
Lucas De Marchi
Quoted 4 lines
Quoted 3 lines
Lucas De Marchi
PATCH 4/14 drm/xe/pf: Relax report helper to accept PF in bulk configs
In reply to Michal WajdeczkoOur current bulk configuration requests are only about VFs, but we want to add new functions that will also include PF configs. Update our bulk report helper to accept also PFID as first VFID.
--- drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-)
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +8 -5
@@ -924,7 +924,8 @@ static int pf_config_bulk_set_u32_done(struct xe_gt *gt, unsigned int first, uns const char *what, const char *(*unit)(u32), unsigned int last, int err) { -xe_gt_assert (gt ,first ); + char name[8]; +xe_gt_assert (gt, num_vfs);xe_gt_assert (gt, first <= last);
@@ -932,8 +933,9 @@ static int pf_config_bulk_set_u32_done(struct xe_gt *gt, unsigned int first, unsreturn pf_config_set_u32_done (gt, first, value,get (gt, first), what, unit, err);if (unlikely (err)) { -xe_gt_sriov_notice (gt,"Failed to bulk provision VF%u..VF%u with %s\n" , - first, first + num_vfs -1 , what); +xe_gt_sriov_notice (gt,"Failed to bulk provision %s..VF%u with %s\n" , +xe_sriov_function_name (first, name,sizeof (name)), + first + num_vfs -1 , what);if (last > first)pf_config_bulk_set_u32_done (gt, first, last - first, value, get, what, unit, last,0 );
@@ -942,8 +944,9 @@ static int pf_config_bulk_set_u32_done(struct xe_gt *gt, unsigned int first, uns/* pick actual value from first VF - bulk provisioning shall be equal across all VFs */ value =get (gt, first); -xe_gt_sriov_info (gt,"VF%u..VF%u provisioned with %u%s %s\n" , - first, first + num_vfs -1 , value,unit (value), what); +xe_gt_sriov_info (gt,"%s..VF%u provisioned with %u%s %s\n" , +xe_sriov_function_name (first, name,sizeof (name)), + first + num_vfs -1 , value,unit (value), what);return 0 ; }
Signature · 1 line
2.47.1
Re: [PATCH 04/14] drm/xe/pf: Relax report helper to accept PF in bulk configs
Quoted 5 lines · Michal Wajdeczko
Our current bulk configuration requests are only about VFs, but we want to add new functions that will also include PF configs. Update our bulk report helper to accept also PFID as first VFID.
Signed-off-by Michal Wajdeczko <[email protected]>
Lucas De Marchi
PATCH 5/14 drm/xe/pf: Add functions to bulk configure EQ/PT on GT
In reply to Michal WajdeczkoWe already have functions to bulk configure 'hard' resources like GGTT, LMEM or GuC context/doorbells IDs. Now add functions for the 'soft' scheduling parameters, as we will need them soon in the upcoming patches.
--- drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c | 56 ++++++++++++++++++++++ drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h | 2 + 2 files changed, 58 insertions(+)
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +56 -0
@@ -1778,6 +1778,34 @@ u32 xe_gt_sriov_pf_config_get_exec_quantum(struct xe_gt *gt, unsigned int vfid)return exec_quantum; } +/** +* xe_gt_sriov_pf_config_bulk_set_exec_quantum() - Configure execution quantum for PF and VFs. +* @gt: the &xe_gt to configure +* @exec_quantum: requested execution quantum in milliseconds (0 is infinity) +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_gt_sriov_pf_config_bulk_set_exec_quantum (struct xe_gt *gt,u32 exec_quantum) +{ +unsigned int totalvfs =xe_gt_sriov_pf_get_totalvfs (gt); +unsigned int n; +int err =0 ; + +mutex_lock (xe_gt_sriov_pf_master_mutex (gt)); +for (n =0 ; n <= totalvfs; n++) { + err =pf_provision_exec_quantum (gt,VFID (n), exec_quantum); +if (err) +break ; + } +mutex_unlock (xe_gt_sriov_pf_master_mutex (gt)); + +return pf_config_bulk_set_u32_done (gt,0 ,1 + totalvfs, exec_quantum, + xe_gt_sriov_pf_config_get_exec_quantum, +"execution quantum" , exec_quantum_unit, n, err); +} +static const char *preempt_timeout_unit (u32 preempt_timeout) {return preempt_timeout ?"us" :"(infinity)" ;
@@ -1849,6 +1877,34 @@ u32 xe_gt_sriov_pf_config_get_preempt_timeout(struct xe_gt *gt, unsigned int vfireturn preempt_timeout; } +/** +* xe_gt_sriov_pf_config_bulk_set_preempt_timeout() - Configure preemption timeout for PF and VFs. +* @gt: the &xe_gt to configure +* @preempt_timeout: requested preemption timeout in microseconds (0 is infinity) +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_gt_sriov_pf_config_bulk_set_preempt_timeout (struct xe_gt *gt,u32 preempt_timeout) +{ +unsigned int totalvfs =xe_gt_sriov_pf_get_totalvfs (gt); +unsigned int n; +int err =0 ; + +mutex_lock (xe_gt_sriov_pf_master_mutex (gt)); +for (n =0 ; n <= totalvfs; n++) { + err =pf_provision_preempt_timeout (gt,VFID (n), preempt_timeout); +if (err) +break ; + } +mutex_unlock (xe_gt_sriov_pf_master_mutex (gt)); + +return pf_config_bulk_set_u32_done (gt,0 ,1 + totalvfs, preempt_timeout, + xe_gt_sriov_pf_config_get_preempt_timeout, +"preemption timeout" , preempt_timeout_unit, n, err); +} +static const char *sched_priority_unit (u32 priority) { return priority == GUC_SCHED_PRIORITY_LOW ? "(low)" :
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h +2 -0
@@ -39,10 +39,12 @@ int xe_gt_sriov_pf_config_bulk_set_lmem(struct xe_gt *gt, unsigned int vfid, unsu32 xe_gt_sriov_pf_config_get_exec_quantum (struct xe_gt *gt,unsigned int vfid);int xe_gt_sriov_pf_config_set_exec_quantum (struct xe_gt *gt,unsigned int vfid,u32 exec_quantum); +int xe_gt_sriov_pf_config_bulk_set_exec_quantum (struct xe_gt *gt,u32 exec_quantum);u32 xe_gt_sriov_pf_config_get_preempt_timeout (struct xe_gt *gt,unsigned int vfid);int xe_gt_sriov_pf_config_set_preempt_timeout (struct xe_gt *gt,unsigned int vfid,u32 preempt_timeout); +int xe_gt_sriov_pf_config_bulk_set_preempt_timeout (struct xe_gt *gt,u32 preempt_timeout);u32 xe_gt_sriov_pf_config_get_sched_priority (struct xe_gt *gt,unsigned int vfid);int xe_gt_sriov_pf_config_set_sched_priority (struct xe_gt *gt,unsigned int vfid,u32 priority);
Signature · 1 line
2.47.1
Re: [PATCH 05/14] drm/xe/pf: Add functions to bulk configure EQ/PT on GT
Quoted 27 lines · Michal Wajdeczko
We already have functions to bulk configure 'hard' resources like GGTT, LMEM or GuC context/doorbells IDs. Now add functions for the 'soft' scheduling parameters, as we will need them soon in the upcoming patches.
Signed-off-by Michal Wajdeczko <[email protected]>---
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c| 56 ++++++++++++++++++++++drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h| 2 +2 files changed, 58 insertions(+)drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +8
-0@@ -1778,6 +1778,34 @@ u32 xe_gt_sriov_pf_config_get_exec_quantum(struct xe_gt *gt, unsigned int vfid)return exec_quantum; } +/** +* xe_gt_sriov_pf_config_bulk_set_exec_quantum() - Configure execution quantum for PF and VFs. +* @gt: the &xe_gt to configure +* @exec_quantum: requested execution quantum in milliseconds (0 is infinity) +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure.
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c u32 xe_gt_sriov_pf_config_get_exec_quantum(struct xe_gt *gt, unsigned int vfid) · line 1798
Quoted 10 lines
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +10
-0+*/ +int xe_gt_sriov_pf_config_bulk_set_exec_quantum (struct xe_gt *gt,u32 exec_quantum) +{ +unsigned int totalvfs =xe_gt_sriov_pf_get_totalvfs (gt); +unsigned int n; +int err =0 ; + +mutex_lock (xe_gt_sriov_pf_master_mutex (gt)); +for (n =0 ; n <= totalvfs; n++) { + err =pf_provision_exec_quantum (gt,VFID (n), exec_quantum);
Given now the function accepts PFID, it's weird to pass VFID(n). Any advantage on using that nop macro?
Lucas De Marchi
Re: [PATCH 05/14] drm/xe/pf: Add functions to bulk configure EQ/PT on GT
Quoted 32 lines · Lucas De Marchi
Quoted 31 lines · Michal Wajdeczko
We already have functions to bulk configure 'hard' resources like GGTT, LMEM or GuC context/doorbells IDs. Now add functions for the 'soft' scheduling parameters, as we will need them soon in the upcoming patches.
Signed-off-by Michal Wajdeczko <[email protected]>---
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c| 56 ++++++++++++++++++++++drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h| 2 +2 files changed, 58 insertions(+)drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +12
-0@@ -1778,6 +1778,34 @@ u32 xe_gt_sriov_pf_config_get_exec_quantum(struct xe_gt *gt, unsigned int vfid) return exec_quantum; } +/** +* xe_gt_sriov_pf_config_bulk_set_exec_quantum() - Configure execution quantum for PF and VFs. +* @gt: the &xe_gt to configure +* @exec_quantum: requested execution quantum in milliseconds (0 is infinity) +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_gt_sriov_pf_config_bulk_set_exec_quantum (struct xe_gt *gt,u32 exec_quantum) +{ + unsigned int totalvfs = xe_gt_sriov_pf_get_totalvfs(gt);
Quoted 8 lines
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c u32 xe_gt_sriov_pf_config_get_exec_quantum(struct xe_gt *gt, unsigned int vfid) · line 1798
Quoted 6 lines
drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +6
-0+ unsigned int n; + int err = 0; + + mutex_lock(xe_gt_sriov_pf_master_mutex(gt)); + for (n = 0; n <= totalvfs; n++) { + err = pf_provision_exec_quantum(gt, VFID(n), exec_quantum);Given now the function accepts PFID, it's weird to pass VFID(n).
PFID is still VFID(0)
Quoted 1 line
Any advantage on using that nop macro?
the function pf_provision_exec_quantum() expects "vfid" anyway so use of the macro was supposed to be self-documenting that we are not passing any random integer here
but I can drop this macro here if you want
Quoted 3 lines
Lucas De Marchi
PATCH 6/14 drm/xe/pf: Add functions to bulk provision EQ/PT
In reply to Michal WajdeczkoWe already have functions to configure EQ/PT for single VF across all tiles/GTs. Now add helper functions that will do that for all VFs (and the PF) at once.
--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 52 ++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 2 + 2 files changed, 54 insertions(+)
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +52 -0
@@ -153,6 +153,32 @@ int xe_sriov_pf_provision_set_mode(struct xe_device *xe, enum xe_sriov_provisionreturn 0 ; } +/** +* xe_sriov_pf_provision_bulk_apply_eq() - Change execution quantum for all VFs and PF. +* @xe: the PF &xe_device +* @eq: execution quantum in [ms] to set +* +* Change execution quantum (EQ) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_bulk_apply_eq (struct xe_device *xe,u32 eq) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_bulk_set_exec_quantum (gt, eq); + result = result ?: err; + } + +return result; +} +/** * xe_sriov_pf_provision_apply_vf_eq() - Change VF's execution quantum. * @xe: the PF &xe_device
@@ -211,6 +237,32 @@ int xe_sriov_pf_provision_query_vf_eq(struct xe_device *xe, unsigned int vfid, ureturn !count ? -ENODATA :0 ; } +/** +* xe_sriov_pf_provision_bulk_apply_pt() - Change preemption timeout for all VFs and PF. +* @xe: the PF &xe_device +* @pt: preemption timeout in [us] to set +* +* Change preemption timeout (PT) provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_bulk_apply_pt (struct xe_device *xe,u32 pt) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_bulk_set_preempt_timeout (gt, pt); + result = result ?: err; + } + +return result; +} +/** * xe_sriov_pf_provision_apply_vf_pt() - Change VF's preemption timeout. * @xe: the PF &xe_device
drivers/gpu/drm/xe/xe_sriov_pf_provision.h +2 -0
@@ -12,9 +12,11 @@struct xe_device ; +int xe_sriov_pf_provision_bulk_apply_eq (struct xe_device *xe,u32 eq);int xe_sriov_pf_provision_apply_vf_eq (struct xe_device *xe,unsigned int vfid,u32 eq);int xe_sriov_pf_provision_query_vf_eq (struct xe_device *xe,unsigned int vfid,u32 *eq); +int xe_sriov_pf_provision_bulk_apply_pt (struct xe_device *xe,u32 pt);int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt);int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt);
Signature · 1 line
2.47.1
Re: [PATCH 06/14] drm/xe/pf: Add functions to bulk provision EQ/PT
Quoted 5 lines · Michal Wajdeczko
We already have functions to configure EQ/PT for single VF across all tiles/GTs. Now add helper functions that will do that for all VFs (and the PF) at once.
Signed-off-by Michal Wajdeczko <[email protected]>
Lucas De Marchi
PATCH 7/14 drm/xe/pf: Allow bulk change all VFs EQ/PT using sysfs
In reply to Michal WajdeczkoIt is expected to be a common practice to configure the same values of execution quantum and preemption timeout parameters across all VFs.
Add write-only sysfs attributes that will apply required EQ/PT values globally, without forcing admin to update PF and each VF separately.
/sys/bus/pci/drivers/xe/BDF/
├── sriov_admin/
├── .bulk_profile
│ ├── exec_quantum_ms [WO] unsigned integer
│ └── preempt_timeout_us [WO] unsigned integer--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 36 ++++++++++++++++++++++++++ 1 file changed, 36 insertions(+)
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +36 -0
@@ -22,6 +22,9 @@* : * ├── sriov_admin/ * ├── ... +* ├── .bulk_profile +* │ ├── exec_quantum_ms +* │ └── preempt_timeout_us * ├── pf/ * │ ├── ... * │ └── profile
@@ -84,7 +87,40 @@ struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \/* device level attributes go here */ +#define DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (NAME ,ITEM ,TYPE ) \ + \ +static ssize_t xe_sriov_dev_attr_##NAME##_store(struct xe_device *xe, \ + const char *buf, size_t count) \ +{ \ + TYPE value; \ + int err; \ + \ + err = kstrto##TYPE(buf, 0, &value); \ + if (err) \ + return err; \ + \ + err = xe_sriov_pf_provision_bulk_apply_##ITEM(xe, value); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_DEV_ATTR_WO(NAME) + +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (exec_quantum_ms, eq, u32); +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (preempt_timeout_us, pt, u32); + +static struct attribute *bulk_profile_dev_attrs[] = { + &xe_sriov_dev_attr_exec_quantum_ms.attr, + &xe_sriov_dev_attr_preempt_timeout_us.attr, +NULL +}; + +static const struct attribute_group bulk_profile_dev_attr_group = { + .name =".bulk_profile" , + .attrs = bulk_profile_dev_attrs, +}; +static const struct attribute_group *xe_sriov_dev_attr_groups[] = { + &bulk_profile_dev_attr_group,NULL };
Signature · 1 line
2.47.1
Re: [PATCH 07/14] drm/xe/pf: Allow bulk change all VFs EQ/PT using sysfs
Quoted 11 lines · Michal Wajdeczko
It is expected to be a common practice to configure the same values of execution quantum and preemption timeout parameters across all VFs.
Add write-only sysfs attributes that will apply required EQ/PT values globally, without forcing admin to update PF and each VF separately.
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── .bulk_profile │ ├── exec_quantum_ms [WO] unsigned integer │ └── preempt_timeout_us [WO] unsigned integer
Quoted 66 lines
Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 36 ++++++++++++++++++++++++++ 1 file changed, 36 insertions(+)drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +36
-0@@ -22,6 +22,9 @@* : * ├── sriov_admin/ * ├── ... +* ├── .bulk_profile +* │ ├── exec_quantum_ms +* │ └── preempt_timeout_us * ├── pf/ * │ ├── ... * │ └── profile @@ -84,7 +87,40 @@ struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \/* device level attributes go here */ +#define DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (NAME ,ITEM ,TYPE ) \ + \ +static ssize_t xe_sriov_dev_attr_##NAME##_store(struct xe_device *xe, \ + const char *buf, size_t count) \ +{ \ + TYPE value; \ + int err; \ + \ + err = kstrto##TYPE(buf, 0, &value); \ + if (err) \ + return err; \ + \ + err = xe_sriov_pf_provision_bulk_apply_##ITEM(xe, value); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_DEV_ATTR_WO(NAME) + +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (exec_quantum_ms, eq, u32); +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (preempt_timeout_us, pt, u32); + +static struct attribute *bulk_profile_dev_attrs[] = { + &xe_sriov_dev_attr_exec_quantum_ms.attr, + &xe_sriov_dev_attr_preempt_timeout_us.attr, +NULL +}; + +static const struct attribute_group bulk_profile_dev_attr_group = { + .name =".bulk_profile" , + .attrs = bulk_profile_dev_attrs, +}; +static const struct attribute_group *xe_sriov_dev_attr_groups[] = { + &bulk_profile_dev_attr_group,NULL };Signature · 1 line
2.47.1
Re: [PATCH 07/14] drm/xe/pf: Allow bulk change all VFs EQ/PT using sysfs
In reply to Michal WajdeczkoQuoted 7 lines · Michal Wajdeczko
It is expected to be a common practice to configure the same values of execution quantum and preemption timeout parameters across all VFs.
Add write-only sysfs attributes that will apply required EQ/PT values globally, without forcing admin to update PF and each VF separately.
/sys/bus/pci/drivers/xe/BDF/
Quoted 8 lines
├── sriov_admin/ ├── .bulk_profile │ ├── exec_quantum_ms [WO] unsigned integer │ └── preempt_timeout_us [WO] unsigned integerSigned-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>
Are there plans for a admin-oriented Documentation/gpu/xe/xe_sriov.rst? Besides rendering the kernel-doc, we should give an overview of setting these attributes, provisioning, etc.
Lucas De Marchi
Quoted 61 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c| 36 ++++++++++++++++++++++++++1 file changed, 36 insertions(+)drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +36
-0@@ -22,6 +22,9 @@* : * ├── sriov_admin/ * ├── ... +* ├── .bulk_profile +* │ ├── exec_quantum_ms +* │ └── preempt_timeout_us * ├── pf/ * │ ├── ... * │ └── profile @@ -84,7 +87,40 @@ struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \/* device level attributes go here */ +#define DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (NAME ,ITEM ,TYPE ) \ + \ +static ssize_t xe_sriov_dev_attr_##NAME##_store(struct xe_device *xe, \ + const char *buf, size_t count) \ +{ \ + TYPE value; \ + int err; \ + \ + err = kstrto##TYPE(buf, 0, &value); \ + if (err) \ + return err; \ + \ + err = xe_sriov_pf_provision_bulk_apply_##ITEM(xe, value); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_DEV_ATTR_WO(NAME) + +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (exec_quantum_ms, eq, u32); +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (preempt_timeout_us, pt, u32); + +static struct attribute *bulk_profile_dev_attrs[] = { + &xe_sriov_dev_attr_exec_quantum_ms.attr, + &xe_sriov_dev_attr_preempt_timeout_us.attr, +NULL +}; + +static const struct attribute_group bulk_profile_dev_attr_group = { + .name =".bulk_profile" , + .attrs = bulk_profile_dev_attrs, +}; +static const struct attribute_group *xe_sriov_dev_attr_groups[] = { + &bulk_profile_dev_attr_group,NULL };Signature · 1 line
2.47.1
Re: [PATCH 07/14] drm/xe/pf: Allow bulk change all VFs EQ/PT using sysfs
Quoted 18 lines · Lucas De Marchi
Quoted 7 lines · Michal Wajdeczko
It is expected to be a common practice to configure the same values of execution quantum and preemption timeout parameters across all VFs.
Add write-only sysfs attributes that will apply required EQ/PT values globally, without forcing admin to update PF and each VF separately.
/sys/bus/pci/drivers/xe/BDF/
Quoted 8 lines
├── sriov_admin/ ├── .bulk_profile │ ├── exec_quantum_ms [WO] unsigned integer │ └── preempt_timeout_us [WO] unsigned integerSigned-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>
Quoted 5 lines
Reviewed-by Lucas De Marchi <[email protected]>Are there plans for a admin-oriented
Documentation/gpu/xe/xe_sriov.rst? Besides rendering the kernel-doc, we should give an overview of setting these attributes, provisioning, etc.
some initial draft was posted 2 years ago [1], I'm still waiting for some more official white-paper, that hopefully will include suggested values tailored for specific use-cases
[1] https://patchwork.freedesktop.org/patch/567127/?series=126289&rev=1
Quoted 65 lines
Lucas De Marchi
Quoted 61 lines
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c| 36 ++++++++++++++++++++++++++1 file changed, 36 insertions(+)drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +36
-0@@ -22,6 +22,9 @@ * : * ├── sriov_admin/ * ├── ... + * ├── .bulk_profile + * │ ├── exec_quantum_ms + * │ └── preempt_timeout_us * ├── pf/ * │ ├── ... * │ └── profile@@ -84,7 +87,40 @@ struct xe_sriov_vf_attr xe_sriov_vf_attr_##NAME = \/* device level attributes go here */ +#define DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (NAME ,ITEM ,TYPE ) \ + \ +static ssize_t xe_sriov_dev_attr_##NAME##_store(struct xe_device *xe, \ + const char *buf, size_t count) \ +{ \ + TYPE value; \ + int err; \ + \ + err = kstrto##TYPE(buf, 0, &value); \ + if (err) \ + return err; \ + \ + err = xe_sriov_pf_provision_bulk_apply_##ITEM(xe, value); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_DEV_ATTR_WO(NAME) + +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (exec_quantum_ms, eq, u32); +DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (preempt_timeout_us, pt, u32); + +static struct attribute *bulk_profile_dev_attrs[] = { + &xe_sriov_dev_attr_exec_quantum_ms.attr, + &xe_sriov_dev_attr_preempt_timeout_us.attr, + NULL +}; + +static const struct attribute_group bulk_profile_dev_attr_group = { + .name = ".bulk_profile", + .attrs = bulk_profile_dev_attrs, +}; +static const struct attribute_group *xe_sriov_dev_attr_groups[] = { + &bulk_profile_dev_attr_group,NULL
};--
2.47.1
PATCH 8/14 drm/xe/pf: Add functions to provision scheduling priority
In reply to Michal WajdeczkoWe already have function to configure PF (or VF) scheduling priority on single GT, but we also need function that will cover all tiles/GTs.
However, due to the current GuC FW limitation, we can't always rely on per-GT function as it actually only works for the PF case. The only way to change VFs scheduling priority is to use 'sched_if_idle' policy KLV that will change priorities for all VFs (and the PF).
We will all these new functions in the upcoming patches.
--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 92 ++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 4 + 2 files changed, 96 insertions(+)
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +92 -0
@@ -320,3 +320,95 @@ int xe_sriov_pf_provision_query_vf_pt(struct xe_device *xe, unsigned int vfid, ureturn !count ? -ENODATA :0 ; } + +/** +* xe_sriov_pf_provision_bulk_apply_priority() - Change scheduling priority of all VFs and PF. +* @xe: the PF &xe_device +* @prio: scheduling priority to set +* +* Change scheduling priority provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_bulk_apply_priority (struct xe_device *xe,u32 prio) +{ +bool sched_if_idle; +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +/* +* Currently, priority changes that involves VFs are only allowed using +* the 'sched_if_idle' policy KLV, so only LOW and NORMAL are supported. +*/ +xe_assert (xe, prio <GUC_SCHED_PRIORITY_HIGH ); + sched_if_idle = prio ==GUC_SCHED_PRIORITY_NORMAL ; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_policy_set_sched_if_idle (gt, sched_if_idle); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_apply_vf_priority() - Change VF's scheduling priority. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @prio: scheduling priority to set +* +* Change VF's scheduling priority provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_priority (struct xe_device *xe,unsigned int vfid,u32 prio) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_sched_priority (gt, vfid, prio); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_query_vf_priority() - Query VF's scheduling priority. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @prio: placeholder for the returned scheduling priority +* +* Query VF's scheduling priority provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_query_vf_priority (struct xe_device *xe,unsigned int vfid,u32 *prio) +{ +struct xe_gt *gt; +unsigned int id; +int count =0 ; +u32 value; + +for_each_gt (gt, xe, id) { + value =xe_gt_sriov_pf_config_get_sched_priority (gt, vfid); +if (!count++) + *prio = value; +else if (value != *prio) +return -EUCLEAN ; + } + +return !count ? -ENODATA :0 ; +}
drivers/gpu/drm/xe/xe_sriov_pf_provision.h +4 -0
@@ -20,6 +20,10 @@ int xe_sriov_pf_provision_bulk_apply_pt(struct xe_device *xe, u32 pt);int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt);int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt); +int xe_sriov_pf_provision_bulk_apply_priority (struct xe_device *xe,u32 prio); +int xe_sriov_pf_provision_apply_vf_priority (struct xe_device *xe,unsigned int vfid,u32 prio); +int xe_sriov_pf_provision_query_vf_priority (struct xe_device *xe,unsigned int vfid,u32 *prio); +int xe_sriov_pf_provision_vfs (struct xe_device *xe,unsigned int num_vfs);int xe_sriov_pf_unprovision_vfs (struct xe_device *xe,unsigned int num_vfs);
Signature · 1 line
2.47.1
Re: [PATCH 08/14] drm/xe/pf: Add functions to provision scheduling priority
Michal Wajdeczko <[email protected]> wrote on pon [2025-paź-20 20:24:08 +0200]:
Quoted 2 lines
We already have function to configure PF (or VF) scheduling priority on single GT, but we also need function that will cover all tiles/GTs.
... on a single GT ....
Quoted 7 lines
However, due to the current GuC FW limitation, we can't always rely on per-GT function as it actually only works for the PF case. The only way to change VFs scheduling priority is to use '
sched_if_idle' policy KLV that will change priorities for all VFs (and the PF).We will all these new functions in the upcoming patches.
We will use these new functions in the upcoming patches.
Quoted 12 lines
Signed-off-by Michal Wajdeczko <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 92 ++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_provision.h | 4 + 2 files changed, 96 insertions(+)drivers/gpu/drm/xe/xe_sriov_pf_provision.c +0
-0
drivers/gpu/drm/xe/xe_sriov_pf_provision.c int xe_sriov_pf_provision_query_vf_pt(struct xe_device *xe, unsigned int vfid, u · line 329
Quoted 10 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +7
-0@@ -320,3 +320,95 @@ int xe_sriov_pf_provision_query_vf_pt(struct xe_device *xe, unsigned int vfid, ureturn !count ? -ENODATA :0 ; } + +/** +* xe_sriov_pf_provision_bulk_apply_priority() - Change scheduling priority of all VFs and PF. +* @xe: the PF &xe_device +* @prio: scheduling priority to set +* +* Change scheduling priority provisioning on all tiles/GTs.
Change the scheduling ...
Quoted 90 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +85
-0+* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_bulk_apply_priority (struct xe_device *xe,u32 prio) +{ +bool sched_if_idle; +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +/* +* Currently, priority changes that involves VFs are only allowed using +* the 'sched_if_idle' policy KLV, so only LOW and NORMAL are supported. +*/ +xe_assert (xe, prio <GUC_SCHED_PRIORITY_HIGH ); + sched_if_idle = prio ==GUC_SCHED_PRIORITY_NORMAL ; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_policy_set_sched_if_idle (gt, sched_if_idle); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_apply_vf_priority() - Change VF's scheduling priority. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @prio: scheduling priority to set +* +* Change VF's scheduling priority provisioning on all tiles/GTs. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_apply_vf_priority (struct xe_device *xe,unsigned int vfid,u32 prio) +{ +struct xe_gt *gt; +unsigned int id; +int result =0 ; +int err; + +for_each_gt (gt, xe, id) { + err =xe_gt_sriov_pf_config_set_sched_priority (gt, vfid, prio); + result = result ?: err; + } + +return result; +} + +/** +* xe_sriov_pf_provision_query_vf_priority() - Query VF's scheduling priority. +* @xe: the PF &xe_device +* @vfid: the VF identifier +* @prio: placeholder for the returned scheduling priority +* +* Query VF's scheduling priority provisioning from all tiles/GTs. +* If values across tiles/GTs are inconsistent then -EUCLEAN error will be returned. +* +* This function can only be called on PF. +* +* Return: 0 on success or a negative error code on failure. +*/ +int xe_sriov_pf_provision_query_vf_priority (struct xe_device *xe,unsigned int vfid,u32 *prio) +{ +struct xe_gt *gt; +unsigned int id; +int count =0 ; +u32 value; + +for_each_gt (gt, xe, id) { + value =xe_gt_sriov_pf_config_get_sched_priority (gt, vfid); +if (!count++) + *prio = value; +else if (value != *prio) +return -EUCLEAN ; + } + +return !count ? -ENODATA :0 ; +}drivers/gpu/drm/xe/xe_sriov_pf_provision.h +0
-0
drivers/gpu/drm/xe/xe_sriov_pf_provision.h int xe_sriov_pf_provision_bulk_apply_pt(struct xe_device *xe, u32 pt); · line 28
Quoted 9 lines
drivers/gpu/drm/xe/xe_sriov_pf_provision.h +4
-0@@ -20,6 +20,10 @@ int xe_sriov_pf_provision_bulk_apply_pt(struct xe_device *xe, u32 pt);int xe_sriov_pf_provision_apply_vf_pt (struct xe_device *xe,unsigned int vfid,u32 pt);int xe_sriov_pf_provision_query_vf_pt (struct xe_device *xe,unsigned int vfid,u32 *pt); +int xe_sriov_pf_provision_bulk_apply_priority (struct xe_device *xe,u32 prio); +int xe_sriov_pf_provision_apply_vf_priority (struct xe_device *xe,unsigned int vfid,u32 prio); +int xe_sriov_pf_provision_query_vf_priority (struct xe_device *xe,unsigned int vfid,u32 *prio); +int xe_sriov_pf_provision_vfs (struct xe_device *xe,unsigned int num_vfs);int xe_sriov_pf_unprovision_vfs (struct xe_device *xe,unsigned int num_vfs);
Some comments in commit massage and docs, but overall it looks okay:
Quoted 4 lines
Signature · 2 lines
-- 2.47.1
PATCH 9/14 drm/xe/pf: Allow bulk change all VFs priority using sysfs
In reply to Michal WajdeczkoIt is expected to be a common practice to configure the same level of scheduling priority across all VFs and PF (at least as starting point). Due to current GuC FW limitations it is also the only way to change VFs priority.
Add write-only sysfs attribute that will apply required priority level to all VFs and PF at once.
/sys/bus/pci/drivers/xe/BDF/
├── sriov_admin/
├── .bulk_profile
│ └── sched_priority [WO] low, normalWriting "low" to this write-only attribute will change PF and VFs scheduling priority on all tiles/GTs to LOW (function will be scheduled only if it has work submitted). Similarly, writing "normal" will change functions priority to NORMAL (functions will be scheduled irrespective of whether there is a work or not).
--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 1 + drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 42 +++++++++++++++++++++- 2 files changed, 42 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_sriov_pf_provision.c +1 -0
@@ -6,6 +6,7 @@#include "xe_assert.h" #include "xe_device.h" #include "xe_gt_sriov_pf_config.h" +#include "xe_gt_sriov_pf_policy.h" #include "xe_sriov.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h"
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +41 -1
57 lines not on this page · raw message
Signature · 1 line
2.47.1
Re: [PATCH 09/14] drm/xe/pf: Allow bulk change all VFs priority using sysfs
Quoted 5 lines · Michal Wajdeczko
It is expected to be a common practice to configure the same level of scheduling priority across all VFs and PF (at least as starting point). Due to current GuC FW limitations it is also the only way to change VFs priority.
Quoted 7 lines
Add write-only sysfs attribute that will apply required priority level to all VFs and PF at once.
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── .bulk_profile │ └── sched_priority [WO] low, normal
Quoted 93 lines
Writing "low" to this write-only attribute will change PF and VFs scheduling priority on all tiles/GTs to LOW (function will be scheduled only if it has work submitted). Similarly, writing "normal" will change functions priority to NORMAL (functions will be scheduled irrespective of whether there is a work or not).
Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 1 + drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 42 +++++++++++++++++++++- 2 files changed, 42 insertions(+), 1 deletion(-)drivers/gpu/drm/xe/xe_sriov_pf_provision.c +1
-0@@ -6,6 +6,7 @@#include "xe_assert.h" #include "xe_device.h" #include "xe_gt_sriov_pf_config.h" +#include "xe_gt_sriov_pf_policy.h" #include "xe_sriov.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h" drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +41
-1@@ -24,7 +24,8 @@* ├── ... * ├── .bulk_profile * │ ├── exec_quantum_ms -* │ └── preempt_timeout_us +* │ ├── preempt_timeout_us +* │ └── sched_priority * ├── pf/ * │ ├── ... * │ └── profile @@ -108,9 +109,48 @@ static XE_SRIOV_DEV_ATTR_WO(NAME)DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (exec_quantum_ms, eq, u32);DEFINE_SIMPLE_BULK_PROVISIONING_SRIOV_DEV_ATTR_WO (preempt_timeout_us, pt, u32); +static const char *const sched_priority_names[] = { + [GUC_SCHED_PRIORITY_LOW ] ="low" , + [GUC_SCHED_PRIORITY_NORMAL ] ="normal" , + [GUC_SCHED_PRIORITY_HIGH ] ="high" , +}; + +static bool sched_priority_high_allowed (unsigned int vfid) +{ +/* As of today GuC FW allows to select 'high' priority only for the PF. */ +return vfid ==PFID ; +} + +static bool sched_priority_bulk_high_allowed (struct xe_device *xe) +{ +/* all VFs are equal - it's sufficient to check VF1 only */ +return sched_priority_high_allowed (VFID (1 )); +} + +static ssize_t xe_sriov_dev_attr_sched_priority_store (struct xe_device *xe, +const char *buf,size_t count) +{ +size_t num_priorities =ARRAY_SIZE (sched_priority_names); +int match; +int err; + +if (!sched_priority_bulk_high_allowed (xe)) + num_priorities--; + + match =__sysfs_match_string (sched_priority_names, num_priorities, buf); +if (match <0 ) +return -EINVAL ; + + err =xe_sriov_pf_provision_bulk_apply_priority (xe, match); +return err ?: count; +} + +static XE_SRIOV_DEV_ATTR_WO (sched_priority); +static struct attribute *bulk_profile_dev_attrs[] = { &xe_sriov_dev_attr_exec_quantum_ms.attr, &xe_sriov_dev_attr_preempt_timeout_us.attr, + &xe_sriov_dev_attr_sched_priority.attr,NULL };Signature · 1 line
2.47.1
PATCH 10/14 drm/xe/pf: Allow change PF scheduling priority using sysfs
In reply to Michal WajdeczkoWe have just added bulk change of the scheduling priority for all VFs and PF, but that only allow to select LOW and NORMAL priority.
Add read-write attribute under PF to allow changing its priority without impacting other VFs priority settings.
For completeness also add read-only attributes under VFs, to show currently selected priority levels used by the VFs.
/sys/bus/pci/drivers/xe/BDF/
├── sriov_admin/
├── pf/
│ └── profile
│ └── sched_priority [RW] low, normal, high
├── vf1/
│ └── profile
│ └── sched_priority [RO] low, normalWriting "high" to the PF read-write attribute will change PF priority on all tiles/GTs to HIGH (schedule function in the next time-slice after current one completes and it has work). Writing "low" or "normal" to change priority to LOW/NORMAL is supported.
When read, those files will display the current and available scheduling priorities. The currently active priority level will be enclosed in square brackets, default output will be like:
$ grep . -h sriov_admin/{pf,vf1,vf2}/profile/sched_priority
[low] normal high
[low] normal
[low] normal--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 82 +++++++++++++++++++++++++- 1 file changed, 80 insertions(+), 2 deletions(-)
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +80 -2
@@ -17,6 +17,21 @@#include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" +static int emit_choice (char *buf,int choice,const char *const *array,size_t size) +{ +int pos =0 ; +int n; + +for (n =0 ; n < size; n++) { + pos +=sysfs_emit_at (buf, pos,"%s%s%s%s" , + n ?" " :"" , + n == choice ?"[" :"" , + array[n], + n == choice ?"]" :"" ); + } +return pos +sysfs_emit_at (buf, pos,"\n" ); +} +/* * /sys/bus/pci/drivers/xe/BDF/ * :
@@ -30,12 +45,14 @@* │ ├── ... * │ └── profile * │ ├── exec_quantum_ms -* │ └── preempt_timeout_us +* │ ├── preempt_timeout_us +* │ └── sched_priority * ├── vf1/ * │ ├── ... * │ └── profile * │ ├── exec_quantum_ms -* │ └── preempt_timeout_us +* │ ├── preempt_timeout_us +* │ └── sched_priority * ├── vf2/ * : * └── vfN/
@@ -115,6 +132,12 @@ static const char * const sched_priority_names[] = { [GUC_SCHED_PRIORITY_HIGH] = "high", }; +static bool sched_priority_change_allowed (unsigned int vfid) +{ +/* As of today GuC FW allows to selectively change only the PF priority. */ +return vfid ==PFID ; +} +static bool sched_priority_high_allowed (unsigned int vfid) {/* As of today GuC FW allows to select 'high' priority only for the PF. */
@@ -199,15 +222,70 @@ static XE_SRIOV_VF_ATTR(NAME)DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (exec_quantum_ms, eq, u32,"%u\n" );DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (preempt_timeout_us, pt, u32,"%u\n" ); +static ssize_t xe_sriov_vf_attr_sched_priority_show (struct xe_device *xe,unsigned int vfid, +char *buf) +{ +size_t num_priorities =ARRAY_SIZE (sched_priority_names); +u32 priority; +int err; + + err =xe_sriov_pf_provision_query_vf_priority (xe, vfid, &priority); +if (err) +return err; + +if (!sched_priority_high_allowed (vfid)) + num_priorities--; + +xe_assert (xe, priority < num_priorities); +return emit_choice (buf, priority, sched_priority_names, num_priorities); +} + +static ssize_t xe_sriov_vf_attr_sched_priority_store (struct xe_device *xe,unsigned int vfid, +const char *buf,size_t count) +{ +size_t num_priorities =ARRAY_SIZE (sched_priority_names); +int match; +int err; + +if (!sched_priority_change_allowed (vfid)) +return -EOPNOTSUPP ; + +if (!sched_priority_high_allowed (vfid)) + num_priorities--; + + match =__sysfs_match_string (sched_priority_names, num_priorities, buf); +if (match <0 ) +return -EINVAL ; + + err =xe_sriov_pf_provision_apply_vf_priority (xe, vfid, match); +return err ?: count; +} + +static XE_SRIOV_VF_ATTR (sched_priority); +static struct attribute *profile_vf_attrs[] = { &xe_sriov_vf_attr_exec_quantum_ms.attr, &xe_sriov_vf_attr_preempt_timeout_us.attr, + &xe_sriov_vf_attr_sched_priority.attr,NULL }; +static umode_t profile_vf_attr_is_visible (struct kobject *kobj, +struct attribute *attr,int index) +{ +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); + +if (attr == &xe_sriov_vf_attr_sched_priority.attr && + !sched_priority_change_allowed (vkobj->vfid)) +return attr->mode &0444 ; + +return attr->mode; +} +static const struct attribute_group profile_vf_attr_group = { .name ="profile" , .attrs = profile_vf_attrs, + .is_visible = profile_vf_attr_is_visible, };static const struct attribute_group *xe_sriov_vf_attr_groups[] = {
Signature · 1 line
2.47.1
Re: [PATCH 10/14] drm/xe/pf: Allow change PF scheduling priority using sysfs
Quoted 9 lines · Michal Wajdeczko
We have just added bulk change of the scheduling priority for all VFs and PF, but that only allow to select LOW and NORMAL priority.
Add read-write attribute under PF to allow changing its priority without impacting other VFs priority settings.
For completeness also add read-only attributes under VFs, to show currently selected priority levels used by the VFs.
Quoted 9 lines
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── pf/ │ └── profile │ └── sched_priority [RW] low, normal, high ├── vf1/ │ └── profile │ └── sched_priority [RO] low, normal
Quoted 151 lines
Writing "high" to the PF read-write attribute will change PF priority on all tiles/GTs to HIGH (schedule function in the next time-slice after current one completes and it has work). Writing "low" or "normal" to change priority to LOW/NORMAL is supported.
When read, those files will display the current and available scheduling priorities. The currently active priority level will be enclosed in square brackets, default output will be like:
$ grep . -h sriov_admin/{pf,vf1,vf2}/profile/sched_priority [low] normal high [low] normal [low] normalSigned-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 82 +++++++++++++++++++++++++- 1 file changed, 80 insertions(+), 2 deletions(-)drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +80
-2@@ -17,6 +17,21 @@#include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" +static int emit_choice (char *buf,int choice,const char *const *array,size_t size) +{ +int pos =0 ; +int n; + +for (n =0 ; n < size; n++) { + pos +=sysfs_emit_at (buf, pos,"%s%s%s%s" , + n ?" " :"" , + n == choice ?"[" :"" , + array[n], + n == choice ?"]" :"" ); + } +return pos +sysfs_emit_at (buf, pos,"\n" ); +} +/* * /sys/bus/pci/drivers/xe/BDF/ * : @@ -30,12 +45,14 @@* │ ├── ... * │ └── profile * │ ├── exec_quantum_ms -* │ └── preempt_timeout_us +* │ ├── preempt_timeout_us +* │ └── sched_priority * ├── vf1/ * │ ├── ... * │ └── profile * │ ├── exec_quantum_ms -* │ └── preempt_timeout_us +* │ ├── preempt_timeout_us +* │ └── sched_priority * ├── vf2/ * : * └── vfN/ @@ -115,6 +132,12 @@ static const char * const sched_priority_names[] = { [GUC_SCHED_PRIORITY_HIGH] = "high", }; +static bool sched_priority_change_allowed (unsigned int vfid) +{ +/* As of today GuC FW allows to selectively change only the PF priority. */ +return vfid ==PFID ; +} +static bool sched_priority_high_allowed (unsigned int vfid) {/* As of today GuC FW allows to select 'high' priority only for the PF. */ @@ -199,15 +222,70 @@ static XE_SRIOV_VF_ATTR(NAME)DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (exec_quantum_ms, eq, u32,"%u\n" );DEFINE_SIMPLE_PROVISIONING_SRIOV_VF_ATTR (preempt_timeout_us, pt, u32,"%u\n" ); +static ssize_t xe_sriov_vf_attr_sched_priority_show (struct xe_device *xe,unsigned int vfid, +char *buf) +{ +size_t num_priorities =ARRAY_SIZE (sched_priority_names); +u32 priority; +int err; + + err =xe_sriov_pf_provision_query_vf_priority (xe, vfid, &priority); +if (err) +return err; + +if (!sched_priority_high_allowed (vfid)) + num_priorities--; + +xe_assert (xe, priority < num_priorities); +return emit_choice (buf, priority, sched_priority_names, num_priorities); +} + +static ssize_t xe_sriov_vf_attr_sched_priority_store (struct xe_device *xe,unsigned int vfid, +const char *buf,size_t count) +{ +size_t num_priorities =ARRAY_SIZE (sched_priority_names); +int match; +int err; + +if (!sched_priority_change_allowed (vfid)) +return -EOPNOTSUPP ; + +if (!sched_priority_high_allowed (vfid)) + num_priorities--; + + match =__sysfs_match_string (sched_priority_names, num_priorities, buf); +if (match <0 ) +return -EINVAL ; + + err =xe_sriov_pf_provision_apply_vf_priority (xe, vfid, match); +return err ?: count; +} + +static XE_SRIOV_VF_ATTR (sched_priority); +static struct attribute *profile_vf_attrs[] = { &xe_sriov_vf_attr_exec_quantum_ms.attr, &xe_sriov_vf_attr_preempt_timeout_us.attr, + &xe_sriov_vf_attr_sched_priority.attr,NULL }; +static umode_t profile_vf_attr_is_visible (struct kobject *kobj, +struct attribute *attr,int index) +{ +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); + +if (attr == &xe_sriov_vf_attr_sched_priority.attr && + !sched_priority_change_allowed (vkobj->vfid)) +return attr->mode &0444 ; + +return attr->mode; +} +static const struct attribute_group profile_vf_attr_group = { .name ="profile" , .attrs = profile_vf_attrs, + .is_visible = profile_vf_attr_is_visible, };static const struct attribute_group *xe_sriov_vf_attr_groups[] = {Signature · 1 line
2.47.1
PATCH 11/14 drm/xe/pf: Promote xe_pci_sriov_get_vf_pdev
In reply to Michal WajdeczkoIn the upcoming patch we would like to use this private helper during preparation of the sysfs links. Promote it.
--- drivers/gpu/drm/xe/xe_pci_sriov.c | 37 ++++++++++++++++++++----------- drivers/gpu/drm/xe/xe_pci_sriov.h | 1 + 2 files changed, 25 insertions(+), 13 deletions(-)
drivers/gpu/drm/xe/xe_pci_sriov.c +24 -13
@@ -30,18 +30,6 @@ static void pf_reset_vfs(struct xe_device *xe, unsigned int num_vfs)xe_sriov_pf_control_reset_vf (xe, n); } -static struct pci_dev *xe_pci_pf_get_vf_dev (struct xe_device *xe,unsigned int vf_id) -{ -struct pci_dev *pdev =to_pci_dev (xe->drm.dev); - -xe_assert (xe,IS_SRIOV_PF (xe)); - -/* caller must use pci_dev_put() */ -return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), - pdev->bus->number, -pci_iov_virtfn_devfn (pdev, vf_id)); -} -static void pf_link_vfs (struct xe_device *xe,int num_vfs) {struct pci_dev *pdev_pf =to_pci_dev (xe->drm.dev);
@@ -60,7 +48,7 @@ static void pf_link_vfs(struct xe_device *xe, int num_vfs)* enforce correct resume order. */ for (n =1 ; n <= num_vfs; n++) { - pdev_vf =xe_pci_pf_get_vf_dev (xe, n -1 ); + pdev_vf =xe_pci_sriov_get_vf_pdev (pdev_pf, n);/* unlikely, something weird is happening, abort */ if (!pdev_vf) {
@@ -228,3 +216,26 @@ int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs)return ret; } + +/** +* xe_pci_sriov_get_vf_pdev() - Lookup the VF's PCI device using the VF identifier. +* @pdev: the PF's &pci_dev +* @vfid: VF identifier (1-based) +* +* The caller must decrement the reference count by calling pci_dev_put(). +* +* Return: the VF's &pci_dev or NULL if the VF device was not found. +*/ +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid) +{ +struct xe_device *xe =pdev_to_xe_device (pdev); + +xe_assert (xe,dev_is_pf (&pdev->dev)); +xe_assert (xe, vfid); +xe_assert (xe, vfid <=pci_sriov_get_totalvfs (pdev)); + +/* caller must use pci_dev_put() */ +return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), + pdev->bus->number, +pci_iov_virtfn_devfn (pdev, vfid -1 )); +}
drivers/gpu/drm/xe/xe_pci_sriov.h +1 -0
@@ -10,6 +10,7 @@ struct pci_dev;#ifdef CONFIG_PCI_IOV int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs); +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid);#else static inline int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs) {
Signature · 1 line
2.47.1
Re: [PATCH 11/14] drm/xe/pf: Promote xe_pci_sriov_get_vf_pdev
Michal Wajdeczko <[email protected]> wrote on pon [2025-paź-20 20:24:11 +0200]:
Quoted 53 lines
In the upcoming patch we would like to use this private helper during preparation of the sysfs links. Promote it.
Signed-off-by Michal Wajdeczko <[email protected]>--- drivers/gpu/drm/xe/xe_pci_sriov.c | 37 ++++++++++++++++++++----------- drivers/gpu/drm/xe/xe_pci_sriov.h | 1 + 2 files changed, 25 insertions(+), 13 deletions(-)drivers/gpu/drm/xe/xe_pci_sriov.c +9
-13@@ -30,18 +30,6 @@ static void pf_reset_vfs(struct xe_device *xe, unsigned int num_vfs)xe_sriov_pf_control_reset_vf (xe, n); } -static struct pci_dev *xe_pci_pf_get_vf_dev (struct xe_device *xe,unsigned int vf_id) -{ -struct pci_dev *pdev =to_pci_dev (xe->drm.dev); - -xe_assert (xe,IS_SRIOV_PF (xe)); - -/* caller must use pci_dev_put() */ -return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), - pdev->bus->number, -pci_iov_virtfn_devfn (pdev, vf_id)); -} -static void pf_link_vfs (struct xe_device *xe,int num_vfs) {struct pci_dev *pdev_pf =to_pci_dev (xe->drm.dev);@@ -60,7 +48,7 @@ static void pf_link_vfs(struct xe_device *xe, int num_vfs)* enforce correct resume order. */ for (n =1 ; n <= num_vfs; n++) { - pdev_vf =xe_pci_pf_get_vf_dev (xe, n -1 ); + pdev_vf =xe_pci_sriov_get_vf_pdev (pdev_pf, n);/* unlikely, something weird is happening, abort */ if (!pdev_vf) {@@ -228,3 +216,26 @@ int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs)return ret; } + +/** +* xe_pci_sriov_get_vf_pdev() - Lookup the VF's PCI device using the VF identifier. +* @pdev: the PF's &pci_dev +* @vfid: VF identifier (1-based) +* +* The caller must decrement the reference count by calling pci_dev_put(). +*
drivers/gpu/drm/xe/xe_pci_sriov.c int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs) · line 236
Quoted 10 lines
drivers/gpu/drm/xe/xe_pci_sriov.c +10
-0+* Return: the VF's &pci_dev or NULL if the VF device was not found. +*/ +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid) +{ +struct xe_device *xe =pdev_to_xe_device (pdev); + +xe_assert (xe,dev_is_pf (&pdev->dev)); +xe_assert (xe, vfid); +xe_assert (xe, vfid <=pci_sriov_get_totalvfs (pdev)); +
NIT: You already mentioned this in the docs above. I don't see the point in repeating it here.
Quoted 10 lines
drivers/gpu/drm/xe/xe_pci_sriov.c +5
-0+/* caller must use pci_dev_put() */ +return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), + pdev->bus->number, +pci_iov_virtfn_devfn (pdev, vfid -1 )); +}drivers/gpu/drm/xe/xe_pci_sriov.h +0
-0
drivers/gpu/drm/xe/xe_pci_sriov.h struct pci_dev; · line 16
Quoted 7 lines
drivers/gpu/drm/xe/xe_pci_sriov.h +1
-0@@ -10,6 +10,7 @@ struct pci_dev;#ifdef CONFIG_PCI_IOV int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs); +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid);#else static inline int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs) {
Shouldn't you prepare a variant for case without CONFIG_PCI_IOV for the function xe_pci_sriov_get_vf_pdev ?
Thanks
Piotr
Quoted 3 lines
Signature · 1 line
2.47.1
Re: [PATCH 11/14] drm/xe/pf: Promote xe_pci_sriov_get_vf_pdev
Quoted 76 lines · Piotr Piórkowski
Michal Wajdeczko <[email protected]> wrote on pon [2025-paź-20 20:24:11 +0200]:
Quoted 53 lines
In the upcoming patch we would like to use this private helper during preparation of the sysfs links. Promote it.
Signed-off-by Michal Wajdeczko <[email protected]>--- drivers/gpu/drm/xe/xe_pci_sriov.c | 37 ++++++++++++++++++++----------- drivers/gpu/drm/xe/xe_pci_sriov.h | 1 + 2 files changed, 25 insertions(+), 13 deletions(-)drivers/gpu/drm/xe/xe_pci_sriov.c +9
-13@@ -30,18 +30,6 @@ static void pf_reset_vfs(struct xe_device *xe, unsigned int num_vfs)xe_sriov_pf_control_reset_vf (xe, n); } -static struct pci_dev *xe_pci_pf_get_vf_dev (struct xe_device *xe,unsigned int vf_id) -{ -struct pci_dev *pdev =to_pci_dev (xe->drm.dev); - -xe_assert (xe,IS_SRIOV_PF (xe)); - -/* caller must use pci_dev_put() */ -return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), - pdev->bus->number, -pci_iov_virtfn_devfn (pdev, vf_id)); -} -static void pf_link_vfs (struct xe_device *xe,int num_vfs) {struct pci_dev *pdev_pf =to_pci_dev (xe->drm.dev);@@ -60,7 +48,7 @@ static void pf_link_vfs(struct xe_device *xe, int num_vfs)* enforce correct resume order. */ for (n =1 ; n <= num_vfs; n++) { - pdev_vf =xe_pci_pf_get_vf_dev (xe, n -1 ); + pdev_vf =xe_pci_sriov_get_vf_pdev (pdev_pf, n);/* unlikely, something weird is happening, abort */ if (!pdev_vf) {@@ -228,3 +216,26 @@ int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs)return ret; } + +/** +* xe_pci_sriov_get_vf_pdev() - Lookup the VF's PCI device using the VF identifier. +* @pdev: the PF's &pci_dev +* @vfid: VF identifier (1-based) +* +* The caller must decrement the reference count by calling pci_dev_put(). +* drivers/gpu/drm/xe/xe_pci_sriov.c int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs) · line 236
Quoted 10 lines
drivers/gpu/drm/xe/xe_pci_sriov.c +10
-0+* Return: the VF's &pci_dev or NULL if the VF device was not found. +*/ +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid) +{ +struct xe_device *xe =pdev_to_xe_device (pdev); + +xe_assert (xe,dev_is_pf (&pdev->dev)); +xe_assert (xe, vfid); +xe_assert (xe, vfid <=pci_sriov_get_totalvfs (pdev)); +NIT: You already mentioned this in the docs above. I don't see the point in repeating it here.
Quoted 11 lines
drivers/gpu/drm/xe/xe_pci_sriov.c +5
-0+/* caller must use pci_dev_put() */ +return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), + pdev->bus->number, +pci_iov_virtfn_devfn (pdev, vfid -1 )); +}drivers/gpu/drm/xe/xe_pci_sriov.h +0
-0@@ -10,6 +10,7 @@ struct pci_dev;
Quoted 9 lines
drivers/gpu/drm/xe/xe_pci_sriov.h struct pci_dev; · line 16
Quoted 6 lines
drivers/gpu/drm/xe/xe_pci_sriov.h +1
-0#ifdef CONFIG_PCI_IOV int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs); +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid);#else static inline int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs) {Shouldn't you prepare a variant for case without
CONFIG_PCI_IOVfor the functionxe_pci_sriov_get_vf_pdev?
not really, as it will be only used by the PF code (which is already under this config, and without PCI_IOV config this will not work anyway ;)
the only reason why xe_pci_sriov_configure() has it's stub is because it is called from the core xe_pci.c (twice)
Quoted 7 lines
Thanks
PiotrQuoted 3 lines
Signature · 1 line
2.47.1
Piotr Piórkowski
Reviewed-byRe: [PATCH 11/14] drm/xe/pf: Promote xe_pci_sriov_get_vf_pdev
Michal Wajdeczko <[email protected]> wrote on wto [2025-paź-28 13:22:08 +0100]:
Quoted 89 lines · Piotr Piórkowski
Quoted 76 lines
Michal Wajdeczko <[email protected]> wrote on pon [2025-paź-20 20:24:11 +0200]:
Quoted 53 lines
In the upcoming patch we would like to use this private helper during preparation of the sysfs links. Promote it.
Signed-off-by Michal Wajdeczko <[email protected]>--- drivers/gpu/drm/xe/xe_pci_sriov.c | 37 ++++++++++++++++++++----------- drivers/gpu/drm/xe/xe_pci_sriov.h | 1 + 2 files changed, 25 insertions(+), 13 deletions(-)drivers/gpu/drm/xe/xe_pci_sriov.c +9
-13@@ -30,18 +30,6 @@ static void pf_reset_vfs(struct xe_device *xe, unsigned int num_vfs)xe_sriov_pf_control_reset_vf (xe, n); } -static struct pci_dev *xe_pci_pf_get_vf_dev (struct xe_device *xe,unsigned int vf_id) -{ -struct pci_dev *pdev =to_pci_dev (xe->drm.dev); - -xe_assert (xe,IS_SRIOV_PF (xe)); - -/* caller must use pci_dev_put() */ -return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), - pdev->bus->number, -pci_iov_virtfn_devfn (pdev, vf_id)); -} -static void pf_link_vfs (struct xe_device *xe,int num_vfs) {struct pci_dev *pdev_pf =to_pci_dev (xe->drm.dev);@@ -60,7 +48,7 @@ static void pf_link_vfs(struct xe_device *xe, int num_vfs)* enforce correct resume order. */ for (n =1 ; n <= num_vfs; n++) { - pdev_vf =xe_pci_pf_get_vf_dev (xe, n -1 ); + pdev_vf =xe_pci_sriov_get_vf_pdev (pdev_pf, n);/* unlikely, something weird is happening, abort */ if (!pdev_vf) {@@ -228,3 +216,26 @@ int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs)return ret; } + +/** +* xe_pci_sriov_get_vf_pdev() - Lookup the VF's PCI device using the VF identifier. +* @pdev: the PF's &pci_dev +* @vfid: VF identifier (1-based) +* +* The caller must decrement the reference count by calling pci_dev_put(). +* drivers/gpu/drm/xe/xe_pci_sriov.c int xe_pci_sriov_configure(struct pci_dev *pdev, int num_vfs) · line 236
Quoted 10 lines
drivers/gpu/drm/xe/xe_pci_sriov.c +10
-0+* Return: the VF's &pci_dev or NULL if the VF device was not found. +*/ +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid) +{ +struct xe_device *xe =pdev_to_xe_device (pdev); + +xe_assert (xe,dev_is_pf (&pdev->dev)); +xe_assert (xe, vfid); +xe_assert (xe, vfid <=pci_sriov_get_totalvfs (pdev)); +NIT: You already mentioned this in the docs above. I don't see the point in repeating it here.
Quoted 11 lines
drivers/gpu/drm/xe/xe_pci_sriov.c +5
-0+/* caller must use pci_dev_put() */ +return pci_get_domain_bus_and_slot (pci_domain_nr (pdev->bus), + pdev->bus->number, +pci_iov_virtfn_devfn (pdev, vfid -1 )); +}drivers/gpu/drm/xe/xe_pci_sriov.h +0
-0@@ -10,6 +10,7 @@ struct pci_dev;Quoted 9 lines
drivers/gpu/drm/xe/xe_pci_sriov.h struct pci_dev; · line 16
Quoted 6 lines
drivers/gpu/drm/xe/xe_pci_sriov.h +1
-0#ifdef CONFIG_PCI_IOV int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs); +struct pci_dev *xe_pci_sriov_get_vf_pdev (struct pci_dev *pdev,unsigned int vfid);#else static inline int xe_pci_sriov_configure (struct pci_dev *pdev,int num_vfs) {Shouldn't you prepare a variant for case without
CONFIG_PCI_IOVfor the functionxe_pci_sriov_get_vf_pdev?
Quoted 6 lines
not really, as it will be only used by the PF code (which is already under this config, and without
PCI_IOVconfig this will not work anyway ;)the only reason why
xe_pci_sriov_configure()has it's stub is because it is called from the corexe_pci.c(twice)
ok
Quoted 8 lines
Quoted 7 lines
Thanks
PiotrQuoted 3 lines
Signature · 1 line
2.47.1
PATCH 12/14 drm/xe/pf: Add sysfs device symlinks to enabled VFs
In reply to Michal WajdeczkoFor convenience, for every enabled VF add 'device' symlink from our SR-IOV admin VF folder to enabled sysfs PCI VF device entry. Remove all those links when disabling PCI VFs.
For completeness, add static 'device' symlink for the PF itself.
/sys/bus/pci/drivers/xe/BDF/sriov_admin/
├── pf │ └── device -> ../../../BDF # PF BDF ├── vf1 │ └── device -> ../../../BDF' # VF1 BDF ├── vf2 │ └── device -> ../../../BDF" # VF2 BDF
--- drivers/gpu/drm/xe/xe_pci_sriov.c | 5 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 93 ++++++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h | 3 + 3 files changed, 101 insertions(+)
drivers/gpu/drm/xe/xe_pci_sriov.c +5 -0
@@ -20,6 +20,7 @@#include "xe_sriov_pf_control.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h" +#include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" static void pf_reset_vfs (struct xe_device *xe,unsigned int num_vfs)
@@ -138,6 +139,8 @@ static int pf_enable_vfs(struct xe_device *xe, int num_vfs)xe_sriov_info (xe,"Enabled %u of %u VF%s\n" , num_vfs, total_vfs,str_plural (total_vfs)); +xe_sriov_pf_sysfs_link_vfs (xe, num_vfs); +pf_engine_activity_stats (xe, num_vfs, true);return num_vfs;
@@ -165,6 +168,8 @@ static int pf_disable_vfs(struct xe_device *xe)pf_engine_activity_stats (xe, num_vfs, false); +xe_sriov_pf_sysfs_unlink_vfs (xe, num_vfs); +pci_disable_sriov (pdev);pf_reset_vfs (xe, num_vfs);
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +93 -0
128 lines not on this page · raw message
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h +3 -0
@@ -10,4 +10,7 @@ struct xe_device;int xe_sriov_pf_sysfs_init (struct xe_device *xe); +void xe_sriov_pf_sysfs_link_vfs (struct xe_device *xe,unsigned int num_vfs); +void xe_sriov_pf_sysfs_unlink_vfs (struct xe_device *xe,unsigned int num_vfs); +#endif
Signature · 1 line
2.47.1
Re: [PATCH 12/14] drm/xe/pf: Add sysfs device symlinks to enabled VFs
Quoted 6 lines · Michal Wajdeczko
For convenience, for every enabled VF add 'device' symlink from our SR-IOV admin VF folder to enabled sysfs PCI VF device entry. Remove all those links when disabling PCI VFs.
For completeness, add static 'device' symlink for the PF itself.
Quoted 8 lines
/sys/bus/pci/drivers/xe/BDF/sriov_admin/├── pf │ └── device -> ../../../BDF # PF BDF ├── vf1 │ └── device -> ../../../BDF' # VF1 BDF ├── vf2 │ └── device -> ../../../BDF" # VF2 BDF
Quoted 191 lines
Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_pci_sriov.c | 5 ++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 93 ++++++++++++++++++++++++++ drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h | 3 + 3 files changed, 101 insertions(+)drivers/gpu/drm/xe/xe_pci_sriov.c +5
-0@@ -20,6 +20,7 @@#include "xe_sriov_pf_control.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h" +#include "xe_sriov_pf_sysfs.h" #include "xe_sriov_printk.h" static void pf_reset_vfs (struct xe_device *xe,unsigned int num_vfs)@@ -138,6 +139,8 @@ static int pf_enable_vfs(struct xe_device *xe, int num_vfs)xe_sriov_info (xe,"Enabled %u of %u VF%s\n" , num_vfs, total_vfs,str_plural (total_vfs)); +xe_sriov_pf_sysfs_link_vfs (xe, num_vfs); +pf_engine_activity_stats (xe, num_vfs, true);return num_vfs;@@ -165,6 +168,8 @@ static int pf_disable_vfs(struct xe_device *xe)pf_engine_activity_stats (xe, num_vfs, false); +xe_sriov_pf_sysfs_unlink_vfs (xe, num_vfs); +pci_disable_sriov (pdev);pf_reset_vfs (xe, num_vfs);drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +93
-0@@ -9,6 +9,7 @@#include <drm/drm_managed.h> #include "xe_assert.h" +#include "xe_pci_sriov.h" #include "xe_pm.h" #include "xe_sriov.h" #include "xe_sriov_pf.h" @@ -43,12 +44,14 @@ static int emit_choice(char *buf, int choice, const char * const *array, size_t* │ └── sched_priority * ├── pf/ * │ ├── ... +* │ ├── device -> ../../../BDF * │ └── profile * │ ├── exec_quantum_ms * │ ├── preempt_timeout_us * │ └── sched_priority * ├── vf1/ * │ ├── ... +* │ ├── device -> ../../../BDF.1 * │ └── profile * │ ├── exec_quantum_ms * │ ├── preempt_timeout_us @@ -422,6 +425,11 @@ static int pf_sysfs_error(struct xe_device *xe, int err, const char *what)return err; } +static void pf_sysfs_note (struct xe_device *xe,int err,const char *what) +{ +xe_sriov_dbg (xe,"Failed to setup sysfs %s (%pe)\n" , what,ERR_PTR (err)); +} +static void action_put_kobject (void *arg) {struct kobject *kobj = arg;@@ -488,6 +496,29 @@ static int pf_setup_tree(struct xe_device *xe)return 0 ; } +static void action_rm_device_link (void *arg) +{ +struct kobject *kobj = arg; + +sysfs_remove_link (kobj,"device" ); +} + +static int pf_link_pf_device (struct xe_device *xe) +{ +struct kobject *kobj = xe->sriov.pf.vfs[PFID ].kobj; +int err; + + err =sysfs_create_link (kobj, &xe->drm.dev->kobj,"device" ); +if (err) +return pf_sysfs_error (xe, err,"PF device link" ); + + err =devm_add_action_or_reset (xe->drm.dev, action_rm_device_link, kobj); +if (err) +return pf_sysfs_error (xe, err,"PF unlink action" ); + +return 0 ; +} +/** * xe_sriov_pf_sysfs_init() - Setup PF's SR-IOV sysfs tree. * @xe: the PF &xe_device to setup sysfs @@ -509,5 +540,67 @@ int xe_sriov_pf_sysfs_init(struct xe_device *xe)if (err)return err; + err =pf_link_pf_device (xe); +if (err) +return err; +return 0 ; } + +/** +* xe_sriov_pf_sysfs_link_vfs() - Add VF's links in SR-IOV sysfs tree. +* @xe: the &xe_device where to update sysfs +* @num_vfs: number of enabled VFs to link +* +* This function is specific for the PF driver. +* +* This function will add symbolic links between VFs represented in the SR-IOV +* sysfs tree maintained by the PF and enabled VF PCI devices. +* +* The @xe_sriov_pf_sysfs_unlink_vfs() shall be used to remove those links. +*/ +void xe_sriov_pf_sysfs_link_vfs (struct xe_device *xe,unsigned int num_vfs) +{ +unsigned int totalvfs =xe_sriov_pf_get_totalvfs (xe); +struct pci_dev *pf_pdev =to_pci_dev (xe->drm.dev); +struct pci_dev *vf_pdev =NULL ; +unsigned int n; +int err; + +xe_assert (xe,IS_SRIOV_PF (xe)); +xe_assert (xe, num_vfs <= totalvfs); + +for (n =1 ; n <= num_vfs; n++) { + vf_pdev =xe_pci_sriov_get_vf_pdev (pf_pdev,VFID (n)); +if (!vf_pdev) +return pf_sysfs_note (xe, -ENOENT ,"VF link" ); + + err =sysfs_create_link (xe->sriov.pf.vfs[VFID (n)].kobj, + &vf_pdev->dev.kobj,"device" ); + +/* must balance xe_pci_sriov_get_vf_pdev() */ +pci_dev_put (vf_pdev); + +if (err) +return pf_sysfs_note (xe, err,"VF link" ); + } +} + +/** +* xe_sriov_pf_sysfs_unlink_vfs() - Remove VF's links from SR-IOV sysfs tree. +* @xe: the &xe_device where to update sysfs +* @num_vfs: number of VFs to unlink +* +* This function shall be called only on the PF. +* This function will remove "device" links added by @xe_sriov_sysfs_link_vfs(). +*/ +void xe_sriov_pf_sysfs_unlink_vfs (struct xe_device *xe,unsigned int num_vfs) +{ +unsigned int n; + +xe_assert (xe,IS_SRIOV_PF (xe)); +xe_assert (xe, num_vfs <=xe_sriov_pf_get_totalvfs (xe)); + +for (n =1 ; n <= num_vfs; n++) +sysfs_remove_link (xe->sriov.pf.vfs[VFID (n)].kobj,"device" ); +}drivers/gpu/drm/xe/xe_sriov_pf_sysfs.h +3
-0@@ -10,4 +10,7 @@ struct xe_device;int xe_sriov_pf_sysfs_init (struct xe_device *xe); +void xe_sriov_pf_sysfs_link_vfs (struct xe_device *xe,unsigned int num_vfs); +void xe_sriov_pf_sysfs_unlink_vfs (struct xe_device *xe,unsigned int num_vfs); +#endif Signature · 1 line
2.47.1
PATCH 13/14 drm/xe/pf: Allow to stop and reset VF using sysfs
In reply to Michal WajdeczkoIt is expected that VFs activity will be monitored and in some cases admin might want to silence specific VF without killing the VM where it was attached.
Add write-only attributes to control GuC scheduling at VF level.
/sys/bus/pci/drivers/xe/BDF/
├── sriov_admin/
├── vf1/
│ ├── reset [WO] bool
│ └── stop [WO] bool
├── vf2/
│ ├── reset [WO] bool
│ └── stop [WO] boolWriting "1" or "y" (or whatever is recognized by the strtobool() function) to these files will trigger the change of the VF state to STOP (GuC will stop servicing the VF) or back to READY (GuC will start servicing this VF again but after completing the full sequence of the software initiated VF FLR).
--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 52 ++++++++++++++++++++++++++ 1 file changed, 52 insertions(+)
drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +52 -0
@@ -13,6 +13,7 @@#include "xe_pm.h" #include "xe_sriov.h" #include "xe_sriov_pf.h" +#include "xe_sriov_pf_control.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h" #include "xe_sriov_pf_sysfs.h"
@@ -52,6 +53,8 @@ static int emit_choice(char *buf, int choice, const char * const *array, size_t* ├── vf1/ * │ ├── ... * │ ├── device -> ../../../BDF.1 +* │ ├── reset +* │ ├── stop * │ └── profile * │ ├── exec_quantum_ms * │ ├── preempt_timeout_us
@@ -291,8 +294,57 @@ static const struct attribute_group profile_vf_attr_group = { .is_visible = profile_vf_attr_is_visible, }; +#define DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (NAME ) \ + \ +static ssize_t xe_sriov_vf_attr_##NAME##_store(struct xe_device *xe, unsigned int vfid, \ + const char *buf, size_t count) \ +{ \ + bool yes; \ + int err; \ + \ + if (!vfid) \ + return -EPERM; \ + \ + err = kstrtobool(buf, &yes); \ + if (err) \ + return err; \ + if (!yes) \ + return count; \ + \ + err = xe_sriov_pf_control_##NAME##_vf(xe, vfid); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_VF_ATTR_WO(NAME) + +DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (stop); +DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (reset); + +static struct attribute *control_vf_attrs[] = { + &xe_sriov_vf_attr_stop.attr, + &xe_sriov_vf_attr_reset.attr, +NULL +}; + +static umode_t control_vf_attr_is_visible (struct kobject *kobj, +struct attribute *attr,int index) +{ +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); + +if (vkobj->vfid ==PFID ) +return 0 ; + +return attr->mode; +} + +static const struct attribute_group control_vf_attr_group = { + .attrs = control_vf_attrs, + .is_visible = control_vf_attr_is_visible, +}; +static const struct attribute_group *xe_sriov_vf_attr_groups[] = { &profile_vf_attr_group, + &control_vf_attr_group,NULL };
Signature · 1 line
2.47.1
Rodrigo Vivi
maintainerRe: [PATCH 13/14] drm/xe/pf: Allow to stop and reset VF using sysfs
Quoted 10 lines · Michal Wajdeczko
It is expected that VFs activity will be monitored and in some cases admin might want to silence specific VF without killing the VM where it was attached.
Add write-only attributes to control GuC scheduling at VF level.
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── vf1/ │ ├── reset [WO] bool
This reset is the only thing in this series that I'm not 100% comfortable. in the device symlink you already have the reset that does the right thing.
If I understood correctly this one here has the same behavior, but with slightly different path... bypassing the pci subsystem.
I understand that your goal is to have it near the stop since you have to stop and then reset. But It is still a bit strange.
If you get our only PF FLR supported case, we do need to unbind the driver before echo 1 > /sys/../device/reset and still the unbind is not a file near the the reset.
So we can perhaps live without this double reset here.
Although not a big blocker from my side....
Lucas, Thomas, thoughts?
Quoted 100 lines
│ └── stop [WO] bool ├── vf2/ │ ├── reset [WO] bool │ └── stop [WO] boolWriting "1" or "y" (or whatever is recognized by the
strtobool()function) to these files will trigger the change of the VF state to STOP (GuC will stop servicing the VF) or back to READY (GuC will start servicing this VF again but after completing the full sequence of the software initiated VF FLR).Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 52 ++++++++++++++++++++++++++ 1 file changed, 52 insertions(+)drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +52
-0@@ -13,6 +13,7 @@#include "xe_pm.h" #include "xe_sriov.h" #include "xe_sriov_pf.h" +#include "xe_sriov_pf_control.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h" #include "xe_sriov_pf_sysfs.h" @@ -52,6 +53,8 @@ static int emit_choice(char *buf, int choice, const char * const *array, size_t* ├── vf1/ * │ ├── ... * │ ├── device -> ../../../BDF.1 +* │ ├── reset +* │ ├── stop * │ └── profile * │ ├── exec_quantum_ms * │ ├── preempt_timeout_us @@ -291,8 +294,57 @@ static const struct attribute_group profile_vf_attr_group = { .is_visible = profile_vf_attr_is_visible, }; +#define DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (NAME ) \ + \ +static ssize_t xe_sriov_vf_attr_##NAME##_store(struct xe_device *xe, unsigned int vfid, \ + const char *buf, size_t count) \ +{ \ + bool yes; \ + int err; \ + \ + if (!vfid) \ + return -EPERM; \ + \ + err = kstrtobool(buf, &yes); \ + if (err) \ + return err; \ + if (!yes) \ + return count; \ + \ + err = xe_sriov_pf_control_##NAME##_vf(xe, vfid); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_VF_ATTR_WO(NAME) + +DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (stop); +DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (reset); + +static struct attribute *control_vf_attrs[] = { + &xe_sriov_vf_attr_stop.attr, + &xe_sriov_vf_attr_reset.attr, +NULL +}; + +static umode_t control_vf_attr_is_visible (struct kobject *kobj, +struct attribute *attr,int index) +{ +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); + +if (vkobj->vfid ==PFID ) +return 0 ; + +return attr->mode; +} + +static const struct attribute_group control_vf_attr_group = { + .attrs = control_vf_attrs, + .is_visible = control_vf_attr_is_visible, +}; +static const struct attribute_group *xe_sriov_vf_attr_groups[] = { &profile_vf_attr_group, + &control_vf_attr_group,NULL };Signature · 1 line
2.47.1
Re: [PATCH 13/14] drm/xe/pf: Allow to stop and reset VF using sysfs
Quoted 7 lines · Rodrigo Vivi
Quoted 6 lines · Michal Wajdeczko
It is expected that VFs activity will be monitored and in some cases admin might want to silence specific VF without killing the VM where it was attached.
Add write-only attributes to control GuC scheduling at VF level.
Quoted 7 lines
Quoted 4 lines
/sys/bus/pci/drivers/xe/BDF/├── sriov_admin/ ├── vf1/ │ ├── reset [WO] boolThis reset is the only thing in this series that I'm not 100% comfortable. in the device symlink you already have the reset that does the right thing.
is it really the same thing? This will end up doing a xe_sriov_pf_control_reset_vf() and I don't see it being done via pci subsystem.
Quoted 11 lines
If I understood correctly this one here has the same behavior, but with slightly different path... bypassing the pci subsystem.
I understand that your goal is to have it near the stop since you have to stop and then reset. But It is still a bit strange.
If you get our only PF FLR supported case, we do need to unbind the driver before echo 1 >
/sys/../device/resetand still the unbind is not a file near the the reset.
Quoted 5 lines
So we can perhaps live without this double reset here.
Although not a big blocker from my side....
Lucas, Thomas, thoughts?
for a sysadmin what would be the alternative? This?
echo 1 > /sys/bus/pci/drivers/xe/<BDF>/sriov_admin/vf1/stop echo 1 > /sys/bus/pci/drivers/xe/<BDF>/sriov_admin/vf1/device/reset
If it's the same thing (which I'm not sure, see above), then I'd drop the reset attribute here and rely on the device symlink we are creating in the previous patch.
Lucas De Marchi
Quoted 101 lines
Quoted 101 lines
│ └── stop [WO] bool ├── vf2/ │ ├── reset [WO] bool │ └── stop [WO] boolWriting "1" or "y" (or whatever is recognized by the
strtobool()function) to these files will trigger the change of the VF state to STOP (GuC will stop servicing the VF) or back to READY (GuC will start servicing this VF again but after completing the full sequence of the software initiated VF FLR).Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c | 52 ++++++++++++++++++++++++++ 1 file changed, 52 insertions(+)drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c +52
-1@@ -13,6 +13,7 @@#include "xe_pm.h" #include "xe_sriov.h" #include "xe_sriov_pf.h" +#include "xe_sriov_pf_control.h" #include "xe_sriov_pf_helpers.h" #include "xe_sriov_pf_provision.h" #include "xe_sriov_pf_sysfs.h" @@ -52,6 +53,8 @@ static int emit_choice(char *buf, int choice, const char * const *array, size_t* ├── vf1/ * │ ├── ... * │ ├── device -> ../../../BDF.1 +* │ ├── reset +* │ ├── stop * │ └── profile * │ ├── exec_quantum_ms * │ ├── preempt_timeout_us @@ -291,8 +294,57 @@ static const struct attribute_group profile_vf_attr_group = { .is_visible = profile_vf_attr_is_visible, }; +#define DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (NAME ) \ + \ +static ssize_t xe_sriov_vf_attr_##NAME##_store(struct xe_device *xe, unsigned int vfid, \ + const char *buf, size_t count) \ +{ \ + bool yes; \ + int err; \ + \ + if (!vfid) \ + return -EPERM; \ + \ + err = kstrtobool(buf, &yes); \ + if (err) \ + return err; \ + if (!yes) \ + return count; \ + \ + err = xe_sriov_pf_control_##NAME##_vf(xe, vfid); \ + return err ?: count; \ +} \ + \ +static XE_SRIOV_VF_ATTR_WO(NAME) + +DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (stop); +DEFINE_SIMPLE_CONTROL_SRIOV_VF_ATTR (reset); + +static struct attribute *control_vf_attrs[] = { + &xe_sriov_vf_attr_stop.attr, + &xe_sriov_vf_attr_reset.attr, +NULL +}; + +static umode_t control_vf_attr_is_visible (struct kobject *kobj, +struct attribute *attr,int index) +{ +struct xe_sriov_kobj *vkobj =to_xe_sriov_kobj (kobj); + +if (vkobj->vfid ==PFID ) +return 0 ; + +return attr->mode; +} + +static const struct attribute_group control_vf_attr_group = { + .attrs = control_vf_attrs, + .is_visible = control_vf_attr_is_visible, +}; +static const struct attribute_group *xe_sriov_vf_attr_groups[] = { &profile_vf_attr_group, + &control_vf_attr_group,NULL }; --2.47.1
PATCH 14/14 drm/xe/pf: Add documentation for sriov_admin attributes
In reply to Michal WajdeczkoAdd initial documentation for all recently added Xe driver specific SR-IOV sysfs files located under device/sriov_admin.
--- .../ABI/testing/sysfs-driver-intel-xe-sriov | 164 ++++++++++++++++++ 1 file changed, 164 insertions(+) create mode 100644 Documentation/ABI/testing/sysfs-driver-intel-xe-sriov
Documentation/ABI/testing/sysfs-driver-intel-xe-sriov new file +164 -0
@@ -0,0 +1,164 @@ +What: /sys/bus/pci/drivers/xe/.../sriov_admin/ +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + This directory appears for the particular Intel Xe device when: + + - device supports SR-IOV, and + - device is a Physical Function (PF), and + - driver support for the SR-IOV PF is enabled on given device. + + This directory is used as a root for all attributes required to + manage both Physical Function (PF) and Virtual Functions (VFs). + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/ +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + This directory holds attributes related to the SR-IOV Physical + Function (PF). + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf1/ +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf2/ +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<N>/ +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These directories hold attributes related to the SR-IOV Virtual + Functions (VFs). + + Note that the VF number <N> is 1-based as described in PCI SR-IOV + specification as the Xe driver follows that naming schema. + + There could be "vf1", "vf2" and so on, up to "vf<N>", where <N> + matches value of the "sriov_totalvfs" attribute. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/profile/exec_quantum_ms +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/profile/preempt_timeout_us +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/profile/sched_priority +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/profile/exec_quantum_ms +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/profile/preempt_timeout_us +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/profile/sched_priority +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These files represent scheduling parameters of the PF or VFs and + are available only for Intel Xe platforms with GPU sharing based + on the time-slice basis. These scheduling parameters can be changed + even if VFs are enabled and running. Those parameters reflects + settings of all tiles/GTs assigned to the given function. + + exec_quantum_ms: (RW) unsigned integer + The GT execution quantum (EQ) in [ms] of the given function. + Actual quantum value might be aligned per HW/FW requirements. + + Default is 0 (unlimited). + + preempt_timeout_us: (RW) unsigned integer + The GT preemption timeout in [us] of the given function. + Actual timeout value might be aligned per HW/FW requirements. + + Default is 0 (unlimited). + + sched_priority: (RW/RO) string + The GT scheduling priority of the given function. + + "low" - function will be scheduled on the GPU for its EQ/PT + only if function has any work already submitted. + + "normal" - functions will be scheduled on the GPU for its EQ/PT + irrespective of whether it has submitted a work or not. + + "high" - function will be scheduled on the GPU for its EQ/PT + in the next time-slice after the current one completes + and function has a work submitted. + + Default is "low". + + When read, this file will display the current and available + scheduling priorities. The currently active priority level will + be enclosed in square brackets, like: + + [low] normal high + + This file can be read-only if changing is currently not supported + for given function due to any known HW/FW limitations. + + Writes to these attributes may fail with errors like: + -EINVAL if provided input is malformed or not recognized, + -EPERM if change is not applicable on given HW/FW, + -EIO if GuC refuses to change provisioning. + + Reads from these attributes may fail with: + -EUCLEAN if value is not consistent across all tiles/GTs. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/.bulk_profile/exec_quantum_ms +What: /sys/bus/pci/drivers/xe/.../sriov_admin/.bulk_profile/preempt_timeout_us +What: /sys/bus/pci/drivers/xe/.../sriov_admin/.bulk_profile/sched_priority +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These files allows bulk reconfiguration of the scheduling parameters + of the PF or VFs and are available only for Intel Xe platforms with + GPU sharing based on the time-slice basis. These scheduling parameters + can be changed even if VFs are enabled and running. + + exec_quantum_ms: (WO) unsigned integer + The GT execution quantum (EQ) in [ms] to be applied to all functions. + See sriov_admin/{pf,vf<N>}/profile/exec_quantum_ms for more details. + + preempt_timeout_us: (WO) unsigned integer + The GT preemption timeout (PT) in [us] to be applied to all functions. + See sriov_admin/{pf,vf<N>}/profile/preempt_timeout_us for more details. + + sched_priority: (RW/RO) string + The GT scheduling priority to be applied for all functions. + See sriov_admin/{pf,vf<N>}/profile/sched_priority for more details. + + Writes to these attributes may fail with errors like: + -EINVAL if provided input is malformed or not recognized, + -EPERM if change is not applicable on given HW/FW, + -EIO if GuC refuses to change provisioning. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/stop +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/reset +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These files allow to control scheduling of the VF on the Intel Xe GPU + platforms. They allow to implement custom policy mechanism in case VFs + are misbehaving or triggering adverse events above defined thresholds. + + stop: (WO) bool + All GT executions of given function shall be immediately stopped. + + reset: (WO) bool + The GT executions of the given function shall be reset after + completing software-initiated VF FLR. + + Writes to these attributes may fail with errors like: + -EINVAL if provided input is malformed or not recognized, + -EPERM if change is not applicable on given HW/FW, + -EIO if GuC refuses to change provisioning. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/device +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/device +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These are symlinks to the underlying PCI device entry representing + given Xe SR-IOV function. For the PF, this link is always present. + For VFs, this link is present only for currently enabled VFs.
Signature · 1 line
2.47.1
Re: [PATCH 14/14] drm/xe/pf: Add documentation for sriov_admin attributes
Quoted 166 lines · Michal Wajdeczko
Add initial documentation for all recently added Xe driver specific SR-IOV sysfs files located under device/sriov_admin.
Signed-off-by Michal Wajdeczko <[email protected]>Cc Lucas De Marchi <[email protected]>Cc Rodrigo Vivi <[email protected]>--- .../ABI/testing/sysfs-driver-intel-xe-sriov | 164 ++++++++++++++++++ 1 file changed, 164 insertions(+) create mode 100644 Documentation/ABI/testing/sysfs-driver-intel-xe-sriovDocumentation/ABI/testing/sysfs-driver-intel-xe-sriov new file +149
-0@@ -0,0 +1,164 @@ +What: /sys/bus/pci/drivers/xe/.../sriov_admin/ +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + This directory appears for the particular Intel Xe device when: + + - device supports SR-IOV, and + - device is a Physical Function (PF), and + - driver support for the SR-IOV PF is enabled on given device. + + This directory is used as a root for all attributes required to + manage both Physical Function (PF) and Virtual Functions (VFs). + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/ +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + This directory holds attributes related to the SR-IOV Physical + Function (PF). + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf1/ +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf2/ +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<N>/ +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These directories hold attributes related to the SR-IOV Virtual + Functions (VFs). + + Note that the VF number <N> is 1-based as described in PCI SR-IOV + specification as the Xe driver follows that naming schema. + + There could be "vf1", "vf2" and so on, up to "vf<N>", where <N> + matches value of the "sriov_totalvfs" attribute. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/profile/exec_quantum_ms +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/profile/preempt_timeout_us +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/profile/sched_priority +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/profile/exec_quantum_ms +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/profile/preempt_timeout_us +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/profile/sched_priority +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These files represent scheduling parameters of the PF or VFs and + are available only for Intel Xe platforms with GPU sharing based + on the time-slice basis. These scheduling parameters can be changed + even if VFs are enabled and running. Those parameters reflects + settings of all tiles/GTs assigned to the given function. + + exec_quantum_ms: (RW) unsigned integer + The GT execution quantum (EQ) in [ms] of the given function. + Actual quantum value might be aligned per HW/FW requirements. + + Default is 0 (unlimited). + + preempt_timeout_us: (RW) unsigned integer + The GT preemption timeout in [us] of the given function. + Actual timeout value might be aligned per HW/FW requirements. + + Default is 0 (unlimited). + + sched_priority: (RW/RO) string + The GT scheduling priority of the given function. + + "low" - function will be scheduled on the GPU for its EQ/PT + only if function has any work already submitted. + + "normal" - functions will be scheduled on the GPU for its EQ/PT + irrespective of whether it has submitted a work or not. + + "high" - function will be scheduled on the GPU for its EQ/PT + in the next time-slice after the current one completes + and function has a work submitted. + + Default is "low". + + When read, this file will display the current and available + scheduling priorities. The currently active priority level will + be enclosed in square brackets, like: + + [low] normal high + + This file can be read-only if changing is currently not supported + for given function due to any known HW/FW limitations. + + Writes to these attributes may fail with errors like: + -EINVAL if provided input is malformed or not recognized, + -EPERM if change is not applicable on given HW/FW, + -EIO if GuC refuses to change provisioning. + + Reads from these attributes may fail with: + -EUCLEAN if value is not consistent across all tiles/GTs. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/.bulk_profile/exec_quantum_ms +What: /sys/bus/pci/drivers/xe/.../sriov_admin/.bulk_profile/preempt_timeout_us +What: /sys/bus/pci/drivers/xe/.../sriov_admin/.bulk_profile/sched_priority +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These files allows bulk reconfiguration of the scheduling parameters + of the PF or VFs and are available only for Intel Xe platforms with + GPU sharing based on the time-slice basis. These scheduling parameters + can be changed even if VFs are enabled and running. + + exec_quantum_ms: (WO) unsigned integer + The GT execution quantum (EQ) in [ms] to be applied to all functions. + See sriov_admin/{pf,vf<N>}/profile/exec_quantum_ms for more details. + + preempt_timeout_us: (WO) unsigned integer + The GT preemption timeout (PT) in [us] to be applied to all functions. + See sriov_admin/{pf,vf<N>}/profile/preempt_timeout_us for more details. + + sched_priority: (RW/RO) string + The GT scheduling priority to be applied for all functions. + See sriov_admin/{pf,vf<N>}/profile/sched_priority for more details. + + Writes to these attributes may fail with errors like: + -EINVAL if provided input is malformed or not recognized, + -EPERM if change is not applicable on given HW/FW, + -EIO if GuC refuses to change provisioning. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/stop +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/reset +Date: October 2025 +KernelVersion: 6.18 +Contact: [email protected] +Description: + These files allow to control scheduling of the VF on the Intel Xe GPU + platforms. They allow to implement custom policy mechanism in case VFs + are misbehaving or triggering adverse events above defined thresholds. + + stop: (WO) bool + All GT executions of given function shall be immediately stopped. + + reset: (WO) bool + The GT executions of the given function shall be reset after + completing software-initiated VF FLR. +
Documentation/ABI/testing/sysfs-driver-intel-xe-sriov line 159
Quoted 10 lines
Documentation/ABI/testing/sysfs-driver-intel-xe-sriov new file +10
-0+ Writes to these attributes may fail with errors like: + -EINVAL if provided input is malformed or not recognized, + -EPERM if change is not applicable on given HW/FW, + -EIO if GuC refuses to change provisioning. + + +What: /sys/bus/pci/drivers/xe/.../sriov_admin/pf/device +What: /sys/bus/pci/drivers/xe/.../sriov_admin/vf<n>/device +Date: October 2025 +KernelVersion: 6.18
6.19
Quoted 8 lines
Documentation/ABI/testing/sysfs-driver-intel-xe-sriov new file +5
-0+Contact: [email protected] +Description: + These are symlinks to the underlying PCI device entry representing + given Xe SR-IOV function. For the PF, this link is always present. + For VFs, this link is present only for currently enabled VFs.Signature · 1 line
2.47.1
Patchwork
CI✗ CI.checkpatch: warning for PF: Add sriov_admin sysfs tree
In reply to Michal Wajdeczko== Series Details ==
== Series Details ==
Series: PF: Add sriov_admin sysfs tree
URL : https://patchwork.freedesktop.org/series/156220/
State : warning
== Summary ==
+ KERNEL=/kernel
+ git clone https://gitlab.freedesktop.org/drm/maintainer-tools mt Cloning into 'mt'...
warning: redirecting to https://gitlab.freedesktop.org/drm/maintainer-tools.git/
+ git -C mt rev-list -n1 origin/master fbd08a78c3a3bb17964db2a326514c69c1dca660
+ cd /kernel
+ git config --global --add safe.directory /kernel
+ git log -n1
commit 13f05e6c26216162a6efd86b0dc79bbaf0ebde65
Author: Michal Wajdeczko <[email protected]> Date: Mon Oct 20 20:24:14 2025 +0200
drm/xe/pf: Add documentation for sriov_admin attributes
Add initial documentation for all recently added Xe driver specific SR-IOV sysfs files located under device/sriov_admin.
+ /mt/dim checkpatch fb865fb539163448d03b35681ea275805a556b96 drm-intel 9aecdc1ff3c9 drm/xe/pf: Prepare sysfs for SR-IOV admin attributes
-:69: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating? #69:
new file mode 100644
-:111: CHECK:LINE_SPACING: Please use a blank line after function/struct/union/enum declarations #111: FILE: drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c:38:
+};
+#define to_xe_sriov_kobj(p) container_of_const((p), struct xe_sriov_kobj, base)
-:118: CHECK:LINE_SPACING: Please use a blank line after function/struct/union/enum declarations #118: FILE: drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c:45:
+};
+#define to_xe_sriov_dev_attr(p) container_of_const((p), struct xe_sriov_dev_attr, attr)
-:137: CHECK:LINE_SPACING: Please use a blank line after function/struct/union/enum declarations #137: FILE: drivers/gpu/drm/xe/xe_sriov_pf_sysfs.c:64:
+};
+#define to_xe_sriov_vf_attr(p) container_of_const((p), struct xe_sriov_vf_attr, attr)
total: 0 errors, 1 warnings, 3 checks, 359 lines checked679489b3549b drm/xe/pf: Take RPM during calls to SR-IOV attr.store()c625cd9787c0 drm/xe/pf: Allow change PF and VFs EQ/PT using sysfs0053259320e8 drm/xe/pf: Relax report helper to accept PF in bulk configs 4d89963f07e1 drm/xe/pf: Add functions to bulk configure EQ/PT on GTb274b73d482c drm/xe/pf: Add functions to bulk provision EQ/PT4ce8b6446285 drm/xe/pf: Allow bulk change all VFs EQ/PT using sysfs4fe594eac1eb drm/xe/pf: Add functions to provision scheduling priority 47c88211fdb8 drm/xe/pf: Allow bulk change all VFs priority using sysfs 0cab92b9e72f drm/xe/pf: Allow change PF scheduling priority using sysfs b06bdf8e7e91 drm/xe/pf: Promote xe_pci_sriov_get_vf_pdev0893681ba45e drm/xe/pf: Add sysfs device symlinks to enabled VFs2b7f0e0c6567 drm/xe/pf: Allow to stop and reset VF using sysfs13f05e6c2621 drm/xe/pf: Add documentation for sriov_admin attributes
-:14: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating? #14:
new file mode 100644
total: 0 errors, 1 warnings, 0 checks, 164 lines checked
Patchwork
CI✓ CI.KUnit: success for PF: Add sriov_admin sysfs tree
In reply to Michal Wajdeczko== Series Details ==
== Series Details ==
Series: PF: Add sriov_admin sysfs tree
URL : https://patchwork.freedesktop.org/series/156220/
State : success
== Summary ==
+ trap cleanup EXIT
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/xe/.kunitconfig [04:35:30] Configuring KUnit Kernel ...
Generating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [04:35:35] Building KUnit Kernel ... Populating config with: $ make ARCH=um O=.kunit olddefconfig Building with: $ make all compile_commands.json scripts_gdb ARCH=um O=.kunit --jobs=25 [04:36:12] Starting KUnit Kernel (1/1)... [04:36:12] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [04:36:12] ================== guc_buf (11 subtests) =================== [04:36:12] [PASSED] test_smallest [04:36:12] [PASSED] test_largest [04:36:12] [PASSED] test_granular [04:36:12] [PASSED] test_unique [04:36:12] [PASSED] test_overlap [04:36:12] [PASSED] test_reusable [04:36:12] [PASSED] test_too_big [04:36:12] [PASSED] test_flush [04:36:12] [PASSED] test_lookup [04:36:12] [PASSED] test_data [04:36:12] [PASSED] test_class [04:36:12] ===================== [PASSED] guc_buf ===================== [04:36:12] =================== guc_dbm (7 subtests) =================== [04:36:12] [PASSED] test_empty [04:36:12] [PASSED] test_default [04:36:12] ======================== test_size ======================== [04:36:12] [PASSED] 4 [04:36:12] [PASSED] 8 [04:36:12] [PASSED] 32 [04:36:12] [PASSED] 256 [04:36:12] ==================== [PASSED] test_size ==================== [04:36:12] ======================= test_reuse ======================== [04:36:12] [PASSED] 4 [04:36:12] [PASSED] 8 [04:36:12] [PASSED] 32 [04:36:12] [PASSED] 256 [04:36:12] =================== [PASSED] test_reuse ==================== [04:36:12] =================== test_range_overlap ==================== [04:36:12] [PASSED] 4 [04:36:12] [PASSED] 8 [04:36:12] [PASSED] 32 [04:36:12] [PASSED] 256 [04:36:12] =============== [PASSED] test_range_overlap ================ [04:36:12] =================== test_range_compact ==================== [04:36:12] [PASSED] 4 [04:36:12] [PASSED] 8 [04:36:12] [PASSED] 32 [04:36:12] [PASSED] 256 [04:36:12] =============== [PASSED] test_range_compact ================ [04:36:12] ==================== test_range_spare ===================== [04:36:12] [PASSED] 4 [04:36:12] [PASSED] 8 [04:36:12] [PASSED] 32 [04:36:12] [PASSED] 256 [04:36:12] ================ [PASSED] test_range_spare ================= [04:36:12] ===================== [PASSED] guc_dbm ===================== [04:36:12] =================== guc_idm (6 subtests) =================== [04:36:12] [PASSED] bad_init [04:36:12] [PASSED] no_init [04:36:12] [PASSED] init_fini [04:36:12] [PASSED] check_used [04:36:12] [PASSED] check_quota [04:36:12] [PASSED] check_all [04:36:12] ===================== [PASSED] guc_idm ===================== [04:36:12] ================== no_relay (3 subtests) =================== [04:36:12] [PASSED] xe_drops_guc2pf_if_not_ready [04:36:12] [PASSED] xe_drops_guc2vf_if_not_ready [04:36:12] [PASSED] xe_rejects_send_if_not_ready [04:36:12] ==================== [PASSED] no_relay ===================== [04:36:12] ================== pf_relay (14 subtests) ================== [04:36:12] [PASSED] pf_rejects_guc2pf_too_short [04:36:12] [PASSED] pf_rejects_guc2pf_too_long [04:36:12] [PASSED] pf_rejects_guc2pf_no_payload [04:36:12] [PASSED] pf_fails_no_payload [04:36:12] [PASSED] pf_fails_bad_origin [04:36:12] [PASSED] pf_fails_bad_type [04:36:12] [PASSED] pf_txn_reports_error [04:36:12] [PASSED] pf_txn_sends_pf2guc [04:36:12] [PASSED] pf_sends_pf2guc [04:36:12] [SKIPPED] pf_loopback_nop [04:36:12] [SKIPPED] pf_loopback_echo [04:36:12] [SKIPPED] pf_loopback_fail [04:36:12] [SKIPPED] pf_loopback_busy [04:36:12] [SKIPPED] pf_loopback_retry [04:36:12] ==================== [PASSED] pf_relay ===================== [04:36:12] ================== vf_relay (3 subtests) =================== [04:36:12] [PASSED] vf_rejects_guc2vf_too_short [04:36:12] [PASSED] vf_rejects_guc2vf_too_long [04:36:12] [PASSED] vf_rejects_guc2vf_no_payload [04:36:12] ==================== [PASSED] vf_relay ===================== [04:36:12] ===================== lmtt (1 subtest) ===================== [04:36:12] ======================== test_ops ========================= [04:36:12] [PASSED] 2-level [04:36:12] [PASSED] multi-level [04:36:12] ==================== [PASSED] test_ops ===================== [04:36:12] ====================== [PASSED] lmtt ======================= [04:36:12] ================= pf_service (11 subtests) ================= [04:36:12] [PASSED] pf_negotiate_any [04:36:12] [PASSED] pf_negotiate_base_match [04:36:12] [PASSED] pf_negotiate_base_newer [04:36:12] [PASSED] pf_negotiate_base_next [04:36:12] [SKIPPED] pf_negotiate_base_older [04:36:12] [PASSED] pf_negotiate_base_prev [04:36:12] [PASSED] pf_negotiate_latest_match [04:36:12] [PASSED] pf_negotiate_latest_newer [04:36:12] [PASSED] pf_negotiate_latest_next [04:36:12] [SKIPPED] pf_negotiate_latest_older [04:36:12] [SKIPPED] pf_negotiate_latest_prev [04:36:12] =================== [PASSED] pf_service ==================== [04:36:12] ================= xe_guc_g2g (2 subtests) ================== [04:36:12] ============== xe_live_guc_g2g_kunit_default ============== [04:36:12] ========= [SKIPPED] xe_live_guc_g2g_kunit_default ========== [04:36:12] ============== xe_live_guc_g2g_kunit_allmem =============== [04:36:12] ========== [SKIPPED] xe_live_guc_g2g_kunit_allmem ========== [04:36:12] =================== [SKIPPED] xe_guc_g2g =================== [04:36:12] =================== xe_mocs (2 subtests) =================== [04:36:12] ================ xe_live_mocs_kernel_kunit ================ [04:36:12] =========== [SKIPPED] xe_live_mocs_kernel_kunit ============ [04:36:12] ================ xe_live_mocs_reset_kunit ================= [04:36:12] ============ [SKIPPED] xe_live_mocs_reset_kunit ============ [04:36:12] ==================== [SKIPPED] xe_mocs ===================== [04:36:12] ================= xe_migrate (2 subtests) ================== [04:36:12] ================= xe_migrate_sanity_kunit ================= [04:36:12] ============ [SKIPPED] xe_migrate_sanity_kunit ============= [04:36:12] ================== xe_validate_ccs_kunit ================== [04:36:12] ============= [SKIPPED] xe_validate_ccs_kunit ============== [04:36:12] =================== [SKIPPED] xe_migrate =================== [04:36:12] ================== xe_dma_buf (1 subtest) ================== [04:36:12] ==================== xe_dma_buf_kunit ===================== [04:36:12] ================ [SKIPPED] xe_dma_buf_kunit ================ [04:36:12] =================== [SKIPPED] xe_dma_buf =================== [04:36:12] ================= xe_bo_shrink (1 subtest) ================= [04:36:12] =================== xe_bo_shrink_kunit ==================== [04:36:12] =============== [SKIPPED] xe_bo_shrink_kunit =============== [04:36:12] ================== [SKIPPED] xe_bo_shrink ================== [04:36:12] ==================== xe_bo (2 subtests) ==================== [04:36:12] ================== xe_ccs_migrate_kunit =================== [04:36:12] ============== [SKIPPED] xe_ccs_migrate_kunit ============== [04:36:12] ==================== xe_bo_evict_kunit ==================== [04:36:12] =============== [SKIPPED] xe_bo_evict_kunit ================ [04:36:12] ===================== [SKIPPED] xe_bo ====================== [04:36:12] ==================== args (11 subtests) ==================== [04:36:12] [PASSED] count_args_test [04:36:12] [PASSED] call_args_example [04:36:12] [PASSED] call_args_test [04:36:12] [PASSED] drop_first_arg_example [04:36:12] [PASSED] drop_first_arg_test [04:36:12] [PASSED] first_arg_example [04:36:12] [PASSED] first_arg_test [04:36:12] [PASSED] last_arg_example [04:36:12] [PASSED] last_arg_test [04:36:12] [PASSED] pick_arg_example [04:36:12] [PASSED] sep_comma_example [04:36:12] ====================== [PASSED] args ======================= [04:36:12] =================== xe_pci (3 subtests) ==================== [04:36:12] ==================== check_graphics_ip ==================== [04:36:12] [PASSED] 12.00 Xe_LP [04:36:12] [PASSED] 12.10 Xe_LP+ [04:36:12] [PASSED] 12.55 Xe_HPG [04:36:12] [PASSED] 12.60 Xe_HPC [04:36:12] [PASSED] 12.70 Xe_LPG [04:36:12] [PASSED] 12.71 Xe_LPG [04:36:12] [PASSED] 12.74 Xe_LPG+ [04:36:12] [PASSED] 20.01 Xe2_HPG [04:36:12] [PASSED] 20.02 Xe2_HPG [04:36:12] [PASSED] 20.04 Xe2_LPG [04:36:12] [PASSED] 30.00 Xe3_LPG [04:36:12] [PASSED] 30.01 Xe3_LPG [04:36:12] [PASSED] 30.03 Xe3_LPG [04:36:12] [PASSED] 30.04 Xe3_LPG [04:36:12] [PASSED] 30.05 Xe3_LPG [04:36:12] [PASSED] 35.11 Xe3p_XPC [04:36:12] ================ [PASSED] check_graphics_ip ================ [04:36:12] ===================== check_media_ip ====================== [04:36:12] [PASSED] 12.00 Xe_M [04:36:12] [PASSED] 12.55 Xe_HPM [04:36:12] [PASSED] 13.00 Xe_LPM+ [04:36:12] [PASSED] 13.01 Xe2_HPM [04:36:12] [PASSED] 20.00 Xe2_LPM [04:36:12] [PASSED] 30.00 Xe3_LPM [04:36:12] [PASSED] 30.02 Xe3_LPM [04:36:12] [PASSED] 35.00 Xe3p_LPM [04:36:12] [PASSED] 35.03 Xe3p_HPM [04:36:12] ================= [PASSED] check_media_ip ================== [04:36:12] ================= check_platform_gt_count ================= [04:36:12] [PASSED] 0x9A60 (TIGERLAKE) [04:36:12] [PASSED] 0x9A68 (TIGERLAKE) [04:36:12] [PASSED] 0x9A70 (TIGERLAKE) [04:36:12] [PASSED] 0x9A40 (TIGERLAKE) [04:36:12] [PASSED] 0x9A49 (TIGERLAKE) [04:36:12] [PASSED] 0x9A59 (TIGERLAKE) [04:36:12] [PASSED] 0x9A78 (TIGERLAKE) [04:36:12] [PASSED] 0x9AC0 (TIGERLAKE) [04:36:12] [PASSED] 0x9AC9 (TIGERLAKE) [04:36:12] [PASSED] 0x9AD9 (TIGERLAKE) [04:36:12] [PASSED] 0x9AF8 (TIGERLAKE) [04:36:12] [PASSED] 0x4C80 (ROCKETLAKE) [04:36:12] [PASSED] 0x4C8A (ROCKETLAKE) [04:36:12] [PASSED] 0x4C8B (ROCKETLAKE) [04:36:12] [PASSED] 0x4C8C (ROCKETLAKE) [04:36:12] [PASSED] 0x4C90 (ROCKETLAKE) [04:36:12] [PASSED] 0x4C9A (ROCKETLAKE) [04:36:12] [PASSED] 0x4680 (ALDERLAKE_S) [04:36:12] [PASSED] 0x4682 (ALDERLAKE_S) [04:36:12] [PASSED] 0x4688 (ALDERLAKE_S) [04:36:12] [PASSED] 0x468A (ALDERLAKE_S) [04:36:12] [PASSED] 0x468B (ALDERLAKE_S) [04:36:12] [PASSED] 0x4690 (ALDERLAKE_S) [04:36:12] [PASSED] 0x4692 (ALDERLAKE_S) [04:36:12] [PASSED] 0x4693 (ALDERLAKE_S) [04:36:12] [PASSED] 0x46A0 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46A1 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46A2 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46A3 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46A6 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46A8 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46AA (ALDERLAKE_P) [04:36:12] [PASSED] 0x462A (ALDERLAKE_P) [04:36:12] [PASSED] 0x4626 (ALDERLAKE_P) [04:36:12] [PASSED] 0x4628 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46B0 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46B1 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46B2 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46B3 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46C0 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46C1 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46C2 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46C3 (ALDERLAKE_P) [04:36:12] [PASSED] 0x46D0 (ALDERLAKE_N) [04:36:12] [PASSED] 0x46D1 (ALDERLAKE_N) [04:36:12] [PASSED] 0x46D2 (ALDERLAKE_N) [04:36:12] [PASSED] 0x46D3 (ALDERLAKE_N) [04:36:12] [PASSED] 0x46D4 (ALDERLAKE_N) [04:36:12] [PASSED] 0xA721 (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7A1 (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7A9 (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7AC (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7AD (ALDERLAKE_P) [04:36:12] [PASSED] 0xA720 (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7A0 (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7A8 (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7AA (ALDERLAKE_P) [04:36:12] [PASSED] 0xA7AB (ALDERLAKE_P) [04:36:12] [PASSED] 0xA780 (ALDERLAKE_S) [04:36:12] [PASSED] 0xA781 (ALDERLAKE_S) [04:36:12] [PASSED] 0xA782 (ALDERLAKE_S) [04:36:12] [PASSED] 0xA783 (ALDERLAKE_S) [04:36:12] [PASSED] 0xA788 (ALDERLAKE_S) [04:36:12] [PASSED] 0xA789 (ALDERLAKE_S) [04:36:12] [PASSED] 0xA78A (ALDERLAKE_S) [04:36:12] [PASSED] 0xA78B (ALDERLAKE_S) [04:36:12] [PASSED] 0x4905 (DG1) [04:36:12] [PASSED] 0x4906 (DG1) [04:36:12] [PASSED] 0x4907 (DG1) [04:36:12] [PASSED] 0x4908 (DG1) [04:36:12] [PASSED] 0x4909 (DG1) [04:36:12] [PASSED] 0x56C0 (DG2) [04:36:12] [PASSED] 0x56C2 (DG2) [04:36:12] [PASSED] 0x56C1 (DG2) [04:36:12] [PASSED] 0x7D51 (METEORLAKE) [04:36:12] [PASSED] 0x7DD1 (METEORLAKE) [04:36:12] [PASSED] 0x7D41 (METEORLAKE) [04:36:12] [PASSED] 0x7D67 (METEORLAKE) [04:36:12] [PASSED] 0xB640 (METEORLAKE) [04:36:12] [PASSED] 0x56A0 (DG2) [04:36:12] [PASSED] 0x56A1 (DG2) [04:36:12] [PASSED] 0x56A2 (DG2) [04:36:12] [PASSED] 0x56BE (DG2) [04:36:12] [PASSED] 0x56BF (DG2) [04:36:12] [PASSED] 0x5690 (DG2) [04:36:12] [PASSED] 0x5691 (DG2) [04:36:12] [PASSED] 0x5692 (DG2) [04:36:12] [PASSED] 0x56A5 (DG2) [04:36:12] [PASSED] 0x56A6 (DG2) [04:36:12] [PASSED] 0x56B0 (DG2) [04:36:12] [PASSED] 0x56B1 (DG2) [04:36:12] [PASSED] 0x56BA (DG2) [04:36:12] [PASSED] 0x56BB (DG2) [04:36:12] [PASSED] 0x56BC (DG2) [04:36:12] [PASSED] 0x56BD (DG2) [04:36:12] [PASSED] 0x5693 (DG2) [04:36:12] [PASSED] 0x5694 (DG2) [04:36:12] [PASSED] 0x5695 (DG2) [04:36:12] [PASSED] 0x56A3 (DG2) [04:36:12] [PASSED] 0x56A4 (DG2) [04:36:12] [PASSED] 0x56B2 (DG2) [04:36:12] [PASSED] 0x56B3 (DG2) [04:36:12] [PASSED] 0x5696 (DG2) [04:36:12] [PASSED] 0x5697 (DG2) [04:36:12] [PASSED] 0xB69 (PVC) [04:36:12] [PASSED] 0xB6E (PVC) [04:36:12] [PASSED] 0xBD4 (PVC) [04:36:12] [PASSED] 0xBD5 (PVC) [04:36:12] [PASSED] 0xBD6 (PVC) [04:36:12] [PASSED] 0xBD7 (PVC) [04:36:12] [PASSED] 0xBD8 (PVC) [04:36:12] [PASSED] 0xBD9 (PVC) [04:36:12] [PASSED] 0xBDA (PVC) [04:36:12] [PASSED] 0xBDB (PVC) [04:36:12] [PASSED] 0xBE0 (PVC) [04:36:12] [PASSED] 0xBE1 (PVC) [04:36:12] [PASSED] 0xBE5 (PVC) [04:36:12] [PASSED] 0x7D40 (METEORLAKE) [04:36:12] [PASSED] 0x7D45 (METEORLAKE) [04:36:12] [PASSED] 0x7D55 (METEORLAKE) [04:36:12] [PASSED] 0x7D60 (METEORLAKE) [04:36:12] [PASSED] 0x7DD5 (METEORLAKE) [04:36:12] [PASSED] 0x6420 (LUNARLAKE) [04:36:12] [PASSED] 0x64A0 (LUNARLAKE) [04:36:12] [PASSED] 0x64B0 (LUNARLAKE) [04:36:12] [PASSED] 0xE202 (BATTLEMAGE) [04:36:12] [PASSED] 0xE209 (BATTLEMAGE) [04:36:12] [PASSED] 0xE20B (BATTLEMAGE) [04:36:12] [PASSED] 0xE20C (BATTLEMAGE) [04:36:12] [PASSED] 0xE20D (BATTLEMAGE) [04:36:12] [PASSED] 0xE210 (BATTLEMAGE) [04:36:12] [PASSED] 0xE211 (BATTLEMAGE) [04:36:12] [PASSED] 0xE212 (BATTLEMAGE) [04:36:12] [PASSED] 0xE216 (BATTLEMAGE) [04:36:12] [PASSED] 0xE220 (BATTLEMAGE) [04:36:12] [PASSED] 0xE221 (BATTLEMAGE) [04:36:12] [PASSED] 0xE222 (BATTLEMAGE) [04:36:12] [PASSED] 0xE223 (BATTLEMAGE) [04:36:12] [PASSED] 0xB080 (PANTHERLAKE) [04:36:12] [PASSED] 0xB081 (PANTHERLAKE) [04:36:12] [PASSED] 0xB082 (PANTHERLAKE) [04:36:12] [PASSED] 0xB083 (PANTHERLAKE) [04:36:12] [PASSED] 0xB084 (PANTHERLAKE) [04:36:12] [PASSED] 0xB085 (PANTHERLAKE) [04:36:12] [PASSED] 0xB086 (PANTHERLAKE) [04:36:12] [PASSED] 0xB087 (PANTHERLAKE) [04:36:12] [PASSED] 0xB08F (PANTHERLAKE) [04:36:12] [PASSED] 0xB090 (PANTHERLAKE) [04:36:12] [PASSED] 0xB0A0 (PANTHERLAKE) [04:36:12] [PASSED] 0xB0B0 (PANTHERLAKE) [04:36:12] [PASSED] 0xFD80 (PANTHERLAKE) [04:36:12] [PASSED] 0xFD81 (PANTHERLAKE) [04:36:12] [PASSED] 0xD740 (NOVALAKE_S) [04:36:12] [PASSED] 0xD741 (NOVALAKE_S) [04:36:12] [PASSED] 0xD742 (NOVALAKE_S) [04:36:12] [PASSED] 0xD743 (NOVALAKE_S) [04:36:12] [PASSED] 0xD744 (NOVALAKE_S) [04:36:12] [PASSED] 0xD745 (NOVALAKE_S) [04:36:12] ============= [PASSED] check_platform_gt_count ============= [04:36:12] ===================== [PASSED] xe_pci ====================== [04:36:12] =================== xe_rtp (2 subtests) ==================== [04:36:12] =============== xe_rtp_process_to_sr_tests ================ [04:36:12] [PASSED] coalesce-same-reg [04:36:12] [PASSED] no-match-no-add [04:36:12] [PASSED] match-or [04:36:12] [PASSED] match-or-xfail [04:36:12] [PASSED] no-match-no-add-multiple-rules [04:36:12] [PASSED] two-regs-two-entries [04:36:12] [PASSED] clr-one-set-other [04:36:12] [PASSED] set-field [04:36:12] [PASSED] conflict-duplicate [04:36:12] [PASSED] conflict-not-disjoint [04:36:12] [PASSED] conflict-reg-type [04:36:12] =========== [PASSED] xe_rtp_process_to_sr_tests ============ [04:36:12] ================== xe_rtp_process_tests =================== [04:36:12] [PASSED] active1 [04:36:12] [PASSED] active2 [04:36:12] [PASSED] active-inactive [04:36:12] [PASSED] inactive-active [04:36:12] [PASSED] inactive-1st_or_active-inactive [04:36:12] [PASSED] inactive-2nd_or_active-inactive [04:36:12] [PASSED] inactive-last_or_active-inactive [04:36:12] [PASSED] inactive-no_or_active-inactive stty: 'standard input': Inappropriate ioctl for device [04:36:12] ============== [PASSED] xe_rtp_process_tests =============== [04:36:12] ===================== [PASSED] xe_rtp ====================== [04:36:12] ==================== xe_wa (1 subtest) ===================== [04:36:12] ======================== xe_wa_gt ========================= [04:36:12] [PASSED] TIGERLAKE B0 [04:36:12] [PASSED] DG1 A0 [04:36:12] [PASSED] DG1 B0 [04:36:12] [PASSED] ALDERLAKE_S A0 [04:36:12] [PASSED] ALDERLAKE_S B0 [04:36:12] [PASSED] ALDERLAKE_S C0 [04:36:12] [PASSED] ALDERLAKE_S D0 [04:36:12] [PASSED] ALDERLAKE_P A0 [04:36:12] [PASSED] ALDERLAKE_P B0 [04:36:12] [PASSED] ALDERLAKE_P C0 [04:36:12] [PASSED] ALDERLAKE_S RPLS D0 [04:36:12] [PASSED] ALDERLAKE_P RPLU E0 [04:36:12] [PASSED] DG2 G10 C0 [04:36:12] [PASSED] DG2 G11 B1 [04:36:12] [PASSED] DG2 G12 A1 [04:36:12] [PASSED] METEORLAKE 12.70(Xe_LPG) A0 13.00(Xe_LPM+) A0 [04:36:12] [PASSED] METEORLAKE 12.71(Xe_LPG) A0 13.00(Xe_LPM+) A0 [04:36:12] [PASSED] METEORLAKE 12.74(Xe_LPG+) A0 13.00(Xe_LPM+) A0 [04:36:12] [PASSED] LUNARLAKE 20.04(Xe2_LPG) A0 20.00(Xe2_LPM) A0 [04:36:12] [PASSED] LUNARLAKE 20.04(Xe2_LPG) B0 20.00(Xe2_LPM) A0 [04:36:12] [PASSED] BATTLEMAGE 20.01(Xe2_HPG) A0 13.01(Xe2_HPM) A1 [04:36:12] [PASSED] PANTHERLAKE 30.00(Xe3_LPG) A0 30.00(Xe3_LPM) A0 [04:36:12] ==================== [PASSED] xe_wa_gt ===================== [04:36:12] ====================== [PASSED] xe_wa ====================== [04:36:12] ============================================================ [04:36:12] Testing complete. Ran 317 tests: passed: 299, skipped: 18 [04:36:12] Elapsed time: 42.083s total, 4.772s configuring, 36.944s building, 0.337s running
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/tests/.kunitconfig [04:36:12] Configuring KUnit Kernel ...
Regenerating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [04:36:14] Building KUnit Kernel ... Populating config with: $ make ARCH=um O=.kunit olddefconfig Building with: $ make all compile_commands.json scripts_gdb ARCH=um O=.kunit --jobs=25 [04:36:43] Starting KUnit Kernel (1/1)... [04:36:43] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [04:36:43] ============ drm_test_pick_cmdline (2 subtests) ============ [04:36:43] [PASSED] drm_test_pick_cmdline_res_1920_1080_60 [04:36:43] =============== drm_test_pick_cmdline_named =============== [04:36:43] [PASSED] NTSC [04:36:43] [PASSED] NTSC-J [04:36:43] [PASSED] PAL [04:36:43] [PASSED] PAL-M [04:36:43] =========== [PASSED] drm_test_pick_cmdline_named =========== [04:36:43] ============== [PASSED] drm_test_pick_cmdline ============== [04:36:43] == drm_test_atomic_get_connector_for_encoder (1 subtest) === [04:36:43] [PASSED] drm_test_drm_atomic_get_connector_for_encoder [04:36:43] ==== [PASSED] drm_test_atomic_get_connector_for_encoder ==== [04:36:43] =========== drm_validate_clone_mode (2 subtests) =========== [04:36:43] ============== drm_test_check_in_clone_mode =============== [04:36:43] [PASSED] in_clone_mode [04:36:43] [PASSED] not_in_clone_mode [04:36:43] ========== [PASSED] drm_test_check_in_clone_mode =========== [04:36:43] =============== drm_test_check_valid_clones =============== [04:36:43] [PASSED] not_in_clone_mode [04:36:43] [PASSED] valid_clone [04:36:43] [PASSED] invalid_clone [04:36:43] =========== [PASSED] drm_test_check_valid_clones =========== [04:36:43] ============= [PASSED] drm_validate_clone_mode ============= [04:36:43] ============= drm_validate_modeset (1 subtest) ============= [04:36:43] [PASSED] drm_test_check_connector_changed_modeset [04:36:43] ============== [PASSED] drm_validate_modeset =============== [04:36:43] ====== drm_test_bridge_get_current_state (2 subtests) ====== [04:36:43] [PASSED] drm_test_drm_bridge_get_current_state_atomic [04:36:43] [PASSED] drm_test_drm_bridge_get_current_state_legacy [04:36:43] ======== [PASSED] drm_test_bridge_get_current_state ======== [04:36:43] ====== drm_test_bridge_helper_reset_crtc (3 subtests) ====== [04:36:43] [PASSED] drm_test_drm_bridge_helper_reset_crtc_atomic [04:36:43] [PASSED] drm_test_drm_bridge_helper_reset_crtc_atomic_disabled [04:36:43] [PASSED] drm_test_drm_bridge_helper_reset_crtc_legacy [04:36:43] ======== [PASSED] drm_test_bridge_helper_reset_crtc ======== [04:36:43] ============== drm_bridge_alloc (2 subtests) =============== [04:36:43] [PASSED] drm_test_drm_bridge_alloc_basic [04:36:43] [PASSED] drm_test_drm_bridge_alloc_get_put [04:36:43] ================ [PASSED] drm_bridge_alloc ================= [04:36:43] ================== drm_buddy (8 subtests) ================== [04:36:43] [PASSED] drm_test_buddy_alloc_limit [04:36:43] [PASSED] drm_test_buddy_alloc_optimistic [04:36:43] [PASSED] drm_test_buddy_alloc_pessimistic [04:36:43] [PASSED] drm_test_buddy_alloc_pathological [04:36:43] [PASSED] drm_test_buddy_alloc_contiguous [04:36:43] [PASSED] drm_test_buddy_alloc_clear [04:36:43] [PASSED] drm_test_buddy_alloc_range_bias [04:36:43] [PASSED] drm_test_buddy_fragmentation_performance [04:36:43] ==================== [PASSED] drm_buddy ==================== [04:36:43] ============= drm_cmdline_parser (40 subtests) ============= [04:36:43] [PASSED] drm_test_cmdline_force_d_only [04:36:43] [PASSED] drm_test_cmdline_force_D_only_dvi [04:36:43] [PASSED] drm_test_cmdline_force_D_only_hdmi [04:36:43] [PASSED] drm_test_cmdline_force_D_only_not_digital [04:36:43] [PASSED] drm_test_cmdline_force_e_only [04:36:43] [PASSED] drm_test_cmdline_res [04:36:43] [PASSED] drm_test_cmdline_res_vesa [04:36:43] [PASSED] drm_test_cmdline_res_vesa_rblank [04:36:43] [PASSED] drm_test_cmdline_res_rblank [04:36:43] [PASSED] drm_test_cmdline_res_bpp [04:36:43] [PASSED] drm_test_cmdline_res_refresh [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_interlaced [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_margins [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_force_off [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on_analog [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on_digital [04:36:43] [PASSED] drm_test_cmdline_res_bpp_refresh_interlaced_margins_force_on [04:36:43] [PASSED] drm_test_cmdline_res_margins_force_on [04:36:43] [PASSED] drm_test_cmdline_res_vesa_margins [04:36:43] [PASSED] drm_test_cmdline_name [04:36:43] [PASSED] drm_test_cmdline_name_bpp [04:36:43] [PASSED] drm_test_cmdline_name_option [04:36:43] [PASSED] drm_test_cmdline_name_bpp_option [04:36:43] [PASSED] drm_test_cmdline_rotate_0 [04:36:43] [PASSED] drm_test_cmdline_rotate_90 [04:36:43] [PASSED] drm_test_cmdline_rotate_180 [04:36:43] [PASSED] drm_test_cmdline_rotate_270 [04:36:43] [PASSED] drm_test_cmdline_hmirror [04:36:43] [PASSED] drm_test_cmdline_vmirror [04:36:43] [PASSED] drm_test_cmdline_margin_options [04:36:43] [PASSED] drm_test_cmdline_multiple_options [04:36:43] [PASSED] drm_test_cmdline_bpp_extra_and_option [04:36:43] [PASSED] drm_test_cmdline_extra_and_option [04:36:43] [PASSED] drm_test_cmdline_freestanding_options [04:36:43] [PASSED] drm_test_cmdline_freestanding_force_e_and_options [04:36:43] [PASSED] drm_test_cmdline_panel_orientation [04:36:43] ================ drm_test_cmdline_invalid ================= [04:36:43] [PASSED] margin_only [04:36:43] [PASSED] interlace_only [04:36:43] [PASSED] res_missing_x [04:36:43] [PASSED] res_missing_y [04:36:43] [PASSED] res_bad_y [04:36:43] [PASSED] res_missing_y_bpp [04:36:43] [PASSED] res_bad_bpp [04:36:43] [PASSED] res_bad_refresh [04:36:43] [PASSED] res_bpp_refresh_force_on_off [04:36:43] [PASSED] res_invalid_mode [04:36:43] [PASSED] res_bpp_wrong_place_mode [04:36:43] [PASSED] name_bpp_refresh [04:36:43] [PASSED] name_refresh [04:36:43] [PASSED] name_refresh_wrong_mode [04:36:43] [PASSED] name_refresh_invalid_mode [04:36:43] [PASSED] rotate_multiple [04:36:43] [PASSED] rotate_invalid_val [04:36:43] [PASSED] rotate_truncated [04:36:43] [PASSED] invalid_option [04:36:43] [PASSED] invalid_tv_option [04:36:43] [PASSED] truncated_tv_option [04:36:43] ============ [PASSED] drm_test_cmdline_invalid ============= [04:36:43] =============== drm_test_cmdline_tv_options =============== [04:36:43] [PASSED] NTSC [04:36:43] [PASSED] NTSC_443 [04:36:43] [PASSED] NTSC_J [04:36:43] [PASSED] PAL [04:36:43] [PASSED] PAL_M [04:36:43] [PASSED] PAL_N [04:36:43] [PASSED] SECAM [04:36:43] [PASSED] MONO_525 [04:36:43] [PASSED] MONO_625 [04:36:43] =========== [PASSED] drm_test_cmdline_tv_options =========== [04:36:43] =============== [PASSED] drm_cmdline_parser ================ [04:36:43] ========== drmm_connector_hdmi_init (20 subtests) ========== [04:36:43] [PASSED] drm_test_connector_hdmi_init_valid [04:36:43] [PASSED] drm_test_connector_hdmi_init_bpc_8 [04:36:43] [PASSED] drm_test_connector_hdmi_init_bpc_10 [04:36:43] [PASSED] drm_test_connector_hdmi_init_bpc_12 [04:36:43] [PASSED] drm_test_connector_hdmi_init_bpc_invalid [04:36:43] [PASSED] drm_test_connector_hdmi_init_bpc_null [04:36:43] [PASSED] drm_test_connector_hdmi_init_formats_empty [04:36:43] [PASSED] drm_test_connector_hdmi_init_formats_no_rgb [04:36:43] === drm_test_connector_hdmi_init_formats_yuv420_allowed === [04:36:43] [PASSED] supported_formats=0x9 yuv420_allowed=1 [04:36:43] [PASSED] supported_formats=0x9 yuv420_allowed=0 [04:36:43] [PASSED] supported_formats=0x3 yuv420_allowed=1 [04:36:43] [PASSED] supported_formats=0x3 yuv420_allowed=0 [04:36:43] === [PASSED] drm_test_connector_hdmi_init_formats_yuv420_allowed === [04:36:43] [PASSED] drm_test_connector_hdmi_init_null_ddc [04:36:43] [PASSED] drm_test_connector_hdmi_init_null_product [04:36:43] [PASSED] drm_test_connector_hdmi_init_null_vendor [04:36:43] [PASSED] drm_test_connector_hdmi_init_product_length_exact [04:36:43] [PASSED] drm_test_connector_hdmi_init_product_length_too_long [04:36:43] [PASSED] drm_test_connector_hdmi_init_product_valid [04:36:43] [PASSED] drm_test_connector_hdmi_init_vendor_length_exact [04:36:43] [PASSED] drm_test_connector_hdmi_init_vendor_length_too_long [04:36:43] [PASSED] drm_test_connector_hdmi_init_vendor_valid [04:36:43] ========= drm_test_connector_hdmi_init_type_valid ========= [04:36:43] [PASSED] HDMI-A [04:36:43] [PASSED] HDMI-B [04:36:43] ===== [PASSED] drm_test_connector_hdmi_init_type_valid ===== [04:36:43] ======== drm_test_connector_hdmi_init_type_invalid ======== [04:36:43] [PASSED] Unknown [04:36:43] [PASSED] VGA [04:36:43] [PASSED] DVI-I [04:36:43] [PASSED] DVI-D [04:36:43] [PASSED] DVI-A [04:36:43] [PASSED] Composite [04:36:43] [PASSED] SVIDEO [04:36:43] [PASSED] LVDS [04:36:43] [PASSED] Component [04:36:43] [PASSED] DIN [04:36:43] [PASSED] DP [04:36:43] [PASSED] TV [04:36:43] [PASSED] eDP [04:36:43] [PASSED] Virtual [04:36:43] [PASSED] DSI [04:36:43] [PASSED] DPI [04:36:43] [PASSED] Writeback [04:36:43] [PASSED] SPI [04:36:43] [PASSED] USB [04:36:43] ==== [PASSED] drm_test_connector_hdmi_init_type_invalid ==== [04:36:43] ============ [PASSED] drmm_connector_hdmi_init ============= [04:36:43] ============= drmm_connector_init (3 subtests) ============= [04:36:43] [PASSED] drm_test_drmm_connector_init [04:36:43] [PASSED] drm_test_drmm_connector_init_null_ddc [04:36:43] ========= drm_test_drmm_connector_init_type_valid ========= [04:36:43] [PASSED] Unknown [04:36:43] [PASSED] VGA [04:36:43] [PASSED] DVI-I [04:36:43] [PASSED] DVI-D [04:36:43] [PASSED] DVI-A [04:36:43] [PASSED] Composite [04:36:43] [PASSED] SVIDEO [04:36:43] [PASSED] LVDS [04:36:43] [PASSED] Component [04:36:43] [PASSED] DIN [04:36:43] [PASSED] DP [04:36:43] [PASSED] HDMI-A [04:36:43] [PASSED] HDMI-B [04:36:43] [PASSED] TV [04:36:43] [PASSED] eDP [04:36:43] [PASSED] Virtual [04:36:43] [PASSED] DSI [04:36:43] [PASSED] DPI [04:36:43] [PASSED] Writeback [04:36:43] [PASSED] SPI [04:36:43] [PASSED] USB [04:36:43] ===== [PASSED] drm_test_drmm_connector_init_type_valid ===== [04:36:43] =============== [PASSED] drmm_connector_init =============== [04:36:43] ========= drm_connector_dynamic_init (6 subtests) ========== [04:36:43] [PASSED] drm_test_drm_connector_dynamic_init [04:36:43] [PASSED] drm_test_drm_connector_dynamic_init_null_ddc [04:36:43] [PASSED] drm_test_drm_connector_dynamic_init_not_added [04:36:43] [PASSED] drm_test_drm_connector_dynamic_init_properties [04:36:43] ===== drm_test_drm_connector_dynamic_init_type_valid ====== [04:36:43] [PASSED] Unknown [04:36:43] [PASSED] VGA [04:36:43] [PASSED] DVI-I [04:36:43] [PASSED] DVI-D [04:36:43] [PASSED] DVI-A [04:36:43] [PASSED] Composite [04:36:43] [PASSED] SVIDEO [04:36:43] [PASSED] LVDS [04:36:43] [PASSED] Component [04:36:43] [PASSED] DIN [04:36:43] [PASSED] DP [04:36:43] [PASSED] HDMI-A [04:36:43] [PASSED] HDMI-B [04:36:43] [PASSED] TV [04:36:43] [PASSED] eDP [04:36:43] [PASSED] Virtual [04:36:43] [PASSED] DSI [04:36:43] [PASSED] DPI [04:36:43] [PASSED] Writeback [04:36:43] [PASSED] SPI [04:36:43] [PASSED] USB [04:36:43] = [PASSED] drm_test_drm_connector_dynamic_init_type_valid == [04:36:43] ======== drm_test_drm_connector_dynamic_init_name ========= [04:36:43] [PASSED] Unknown [04:36:43] [PASSED] VGA [04:36:43] [PASSED] DVI-I [04:36:43] [PASSED] DVI-D [04:36:43] [PASSED] DVI-A [04:36:43] [PASSED] Composite [04:36:43] [PASSED] SVIDEO [04:36:43] [PASSED] LVDS [04:36:43] [PASSED] Component [04:36:43] [PASSED] DIN [04:36:43] [PASSED] DP [04:36:43] [PASSED] HDMI-A [04:36:43] [PASSED] HDMI-B [04:36:43] [PASSED] TV [04:36:43] [PASSED] eDP [04:36:43] [PASSED] Virtual [04:36:43] [PASSED] DSI [04:36:43] [PASSED] DPI [04:36:43] [PASSED] Writeback [04:36:43] [PASSED] SPI [04:36:43] [PASSED] USB [04:36:43] ==== [PASSED] drm_test_drm_connector_dynamic_init_name ===== [04:36:43] =========== [PASSED] drm_connector_dynamic_init ============ [04:36:43] ==== drm_connector_dynamic_register_early (4 subtests) ===== [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_early_on_list [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_early_defer [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_early_no_init [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_early_no_mode_object [04:36:43] ====== [PASSED] drm_connector_dynamic_register_early ======= [04:36:43] ======= drm_connector_dynamic_register (7 subtests) ======== [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_on_list [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_no_defer [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_no_init [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_mode_object [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_sysfs [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_sysfs_name [04:36:43] [PASSED] drm_test_drm_connector_dynamic_register_debugfs [04:36:43] ========= [PASSED] drm_connector_dynamic_register ========== [04:36:43] = drm_connector_attach_broadcast_rgb_property (2 subtests) = [04:36:43] [PASSED] drm_test_drm_connector_attach_broadcast_rgb_property [04:36:43] [PASSED] drm_test_drm_connector_attach_broadcast_rgb_property_hdmi_connector [04:36:43] === [PASSED] drm_connector_attach_broadcast_rgb_property === [04:36:43] ========== drm_get_tv_mode_from_name (2 subtests) ========== [04:36:43] ========== drm_test_get_tv_mode_from_name_valid =========== [04:36:43] [PASSED] NTSC [04:36:43] [PASSED] NTSC-443 [04:36:43] [PASSED] NTSC-J [04:36:43] [PASSED] PAL [04:36:43] [PASSED] PAL-M [04:36:43] [PASSED] PAL-N [04:36:43] [PASSED] SECAM [04:36:43] [PASSED] Mono [04:36:43] ====== [PASSED] drm_test_get_tv_mode_from_name_valid ======= [04:36:43] [PASSED] drm_test_get_tv_mode_from_name_truncated [04:36:43] ============ [PASSED] drm_get_tv_mode_from_name ============ [04:36:43] = drm_test_connector_hdmi_compute_mode_clock (12 subtests) = [04:36:43] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb [04:36:43] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_10bpc [04:36:43] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_10bpc_vic_1 [04:36:43] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_12bpc [04:36:43] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_12bpc_vic_1 [04:36:43] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_double [04:36:43] = drm_test_connector_hdmi_compute_mode_clock_yuv420_valid = [04:36:43] [PASSED] VIC 96 [04:36:43] [PASSED] VIC 97 [04:36:43] [PASSED] VIC 101 [04:36:43] [PASSED] VIC 102 [04:36:43] [PASSED] VIC 106 [04:36:43] [PASSED] VIC 107 [04:36:43] === [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_valid === [04:36:43] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_10_bpc [04:36:43] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_12_bpc [04:36:43] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_8_bpc [04:36:43] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_10_bpc [04:36:43] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_12_bpc [04:36:43] === [PASSED] drm_test_connector_hdmi_compute_mode_clock ==== [04:36:43] == drm_hdmi_connector_get_broadcast_rgb_name (2 subtests) == [04:36:43] === drm_test_drm_hdmi_connector_get_broadcast_rgb_name ==== [04:36:43] [PASSED] Automatic [04:36:43] [PASSED] Full [04:36:43] [PASSED] Limited 16:235 [04:36:43] === [PASSED] drm_test_drm_hdmi_connector_get_broadcast_rgb_name === [04:36:43] [PASSED] drm_test_drm_hdmi_connector_get_broadcast_rgb_name_invalid [04:36:43] ==== [PASSED] drm_hdmi_connector_get_broadcast_rgb_name ==== [04:36:43] == drm_hdmi_connector_get_output_format_name (2 subtests) == [04:36:43] === drm_test_drm_hdmi_connector_get_output_format_name ==== [04:36:43] [PASSED] RGB [04:36:43] [PASSED] YUV 4:2:0 [04:36:43] [PASSED] YUV 4:2:2 [04:36:43] [PASSED] YUV 4:4:4 [04:36:43] === [PASSED] drm_test_drm_hdmi_connector_get_output_format_name === [04:36:43] [PASSED] drm_test_drm_hdmi_connector_get_output_format_name_invalid [04:36:43] ==== [PASSED] drm_hdmi_connector_get_output_format_name ==== [04:36:43] ============= drm_damage_helper (21 subtests) ============== [04:36:43] [PASSED] drm_test_damage_iter_no_damage [04:36:43] [PASSED] drm_test_damage_iter_no_damage_fractional_src [04:36:43] [PASSED] drm_test_damage_iter_no_damage_src_moved [04:36:43] [PASSED] drm_test_damage_iter_no_damage_fractional_src_moved [04:36:43] [PASSED] drm_test_damage_iter_no_damage_not_visible [04:36:43] [PASSED] drm_test_damage_iter_no_damage_no_crtc [04:36:43] [PASSED] drm_test_damage_iter_no_damage_no_fb [04:36:43] [PASSED] drm_test_damage_iter_simple_damage [04:36:43] [PASSED] drm_test_damage_iter_single_damage [04:36:43] [PASSED] drm_test_damage_iter_single_damage_intersect_src [04:36:43] [PASSED] drm_test_damage_iter_single_damage_outside_src [04:36:43] [PASSED] drm_test_damage_iter_single_damage_fractional_src [04:36:43] [PASSED] drm_test_damage_iter_single_damage_intersect_fractional_src [04:36:43] [PASSED] drm_test_damage_iter_single_damage_outside_fractional_src [04:36:43] [PASSED] drm_test_damage_iter_single_damage_src_moved [04:36:43] [PASSED] drm_test_damage_iter_single_damage_fractional_src_moved [04:36:43] [PASSED] drm_test_damage_iter_damage [04:36:43] [PASSED] drm_test_damage_iter_damage_one_intersect [04:36:43] [PASSED] drm_test_damage_iter_damage_one_outside [04:36:43] [PASSED] drm_test_damage_iter_damage_src_moved [04:36:43] [PASSED] drm_test_damage_iter_damage_not_visible [04:36:43] ================ [PASSED] drm_damage_helper ================ [04:36:43] ============== drm_dp_mst_helper (3 subtests) ============== [04:36:43] ============== drm_test_dp_mst_calc_pbn_mode ============== [04:36:43] [PASSED] Clock 154000 BPP 30 DSC disabled [04:36:43] [PASSED] Clock 234000 BPP 30 DSC disabled [04:36:43] [PASSED] Clock 297000 BPP 24 DSC disabled [04:36:43] [PASSED] Clock 332880 BPP 24 DSC enabled [04:36:43] [PASSED] Clock 324540 BPP 24 DSC enabled [04:36:43] ========== [PASSED] drm_test_dp_mst_calc_pbn_mode ========== [04:36:43] ============== drm_test_dp_mst_calc_pbn_div =============== [04:36:43] [PASSED] Link rate 2000000 lane count 4 [04:36:43] [PASSED] Link rate 2000000 lane count 2 [04:36:43] [PASSED] Link rate 2000000 lane count 1 [04:36:43] [PASSED] Link rate 1350000 lane count 4 [04:36:43] [PASSED] Link rate 1350000 lane count 2 [04:36:43] [PASSED] Link rate 1350000 lane count 1 [04:36:43] [PASSED] Link rate 1000000 lane count 4 [04:36:43] [PASSED] Link rate 1000000 lane count 2 [04:36:43] [PASSED] Link rate 1000000 lane count 1 [04:36:43] [PASSED] Link rate 810000 lane count 4 [04:36:43] [PASSED] Link rate 810000 lane count 2 [04:36:43] [PASSED] Link rate 810000 lane count 1 [04:36:43] [PASSED] Link rate 540000 lane count 4 [04:36:43] [PASSED] Link rate 540000 lane count 2 [04:36:43] [PASSED] Link rate 540000 lane count 1 [04:36:43] [PASSED] Link rate 270000 lane count 4 [04:36:43] [PASSED] Link rate 270000 lane count 2 [04:36:43] [PASSED] Link rate 270000 lane count 1 [04:36:43] [PASSED] Link rate 162000 lane count 4 [04:36:43] [PASSED] Link rate 162000 lane count 2 [04:36:43] [PASSED] Link rate 162000 lane count 1 [04:36:43] ========== [PASSED] drm_test_dp_mst_calc_pbn_div =========== [04:36:43] ========= drm_test_dp_mst_sideband_msg_req_decode ========= [04:36:43] [PASSED] DP_ENUM_PATH_RESOURCES with port number [04:36:43] [PASSED] DP_POWER_UP_PHY with port number [04:36:43] [PASSED] DP_POWER_DOWN_PHY with port number [04:36:43] [PASSED] DP_ALLOCATE_PAYLOAD with SDP stream sinks [04:36:43] [PASSED] DP_ALLOCATE_PAYLOAD with port number [04:36:43] [PASSED] DP_ALLOCATE_PAYLOAD with VCPI [04:36:43] [PASSED] DP_ALLOCATE_PAYLOAD with PBN [04:36:43] [PASSED] DP_QUERY_PAYLOAD with port number [04:36:43] [PASSED] DP_QUERY_PAYLOAD with VCPI [04:36:43] [PASSED] DP_REMOTE_DPCD_READ with port number [04:36:43] [PASSED] DP_REMOTE_DPCD_READ with DPCD address [04:36:43] [PASSED] DP_REMOTE_DPCD_READ with max number of bytes [04:36:43] [PASSED] DP_REMOTE_DPCD_WRITE with port number [04:36:43] [PASSED] DP_REMOTE_DPCD_WRITE with DPCD address [04:36:43] [PASSED] DP_REMOTE_DPCD_WRITE with data array [04:36:43] [PASSED] DP_REMOTE_I2C_READ with port number [04:36:43] [PASSED] DP_REMOTE_I2C_READ with I2C device ID [04:36:43] [PASSED] DP_REMOTE_I2C_READ with transactions array [04:36:43] [PASSED] DP_REMOTE_I2C_WRITE with port number [04:36:43] [PASSED] DP_REMOTE_I2C_WRITE with I2C device ID [04:36:43] [PASSED] DP_REMOTE_I2C_WRITE with data array [04:36:43] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream ID [04:36:43] [PASSED] DP_QUERY_STREAM_ENC_STATUS with client ID [04:36:43] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream event [04:36:43] [PASSED] DP_QUERY_STREAM_ENC_STATUS with valid stream event [04:36:43] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream behavior [04:36:43] [PASSED] DP_QUERY_STREAM_ENC_STATUS with a valid stream behavior [04:36:43] ===== [PASSED] drm_test_dp_mst_sideband_msg_req_decode ===== [04:36:43] ================ [PASSED] drm_dp_mst_helper ================ [04:36:43] ================== drm_exec (7 subtests) =================== [04:36:43] [PASSED] sanitycheck [04:36:43] [PASSED] test_lock [04:36:43] [PASSED] test_lock_unlock [04:36:43] [PASSED] test_duplicates [04:36:43] [PASSED] test_prepare [04:36:43] [PASSED] test_prepare_array [04:36:43] [PASSED] test_multiple_loops [04:36:43] ==================== [PASSED] drm_exec ===================== [04:36:43] =========== drm_format_helper_test (17 subtests) =========== [04:36:43] ============== drm_test_fb_xrgb8888_to_gray8 ============== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ========== [PASSED] drm_test_fb_xrgb8888_to_gray8 ========== [04:36:43] ============= drm_test_fb_xrgb8888_to_rgb332 ============== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb332 ========== [04:36:43] ============= drm_test_fb_xrgb8888_to_rgb565 ============== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb565 ========== [04:36:43] ============ drm_test_fb_xrgb8888_to_xrgb1555 ============= [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======== [PASSED] drm_test_fb_xrgb8888_to_xrgb1555 ========= [04:36:43] ============ drm_test_fb_xrgb8888_to_argb1555 ============= [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======== [PASSED] drm_test_fb_xrgb8888_to_argb1555 ========= [04:36:43] ============ drm_test_fb_xrgb8888_to_rgba5551 ============= [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======== [PASSED] drm_test_fb_xrgb8888_to_rgba5551 ========= [04:36:43] ============= drm_test_fb_xrgb8888_to_rgb888 ============== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb888 ========== [04:36:43] ============= drm_test_fb_xrgb8888_to_bgr888 ============== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ========= [PASSED] drm_test_fb_xrgb8888_to_bgr888 ========== [04:36:43] ============ drm_test_fb_xrgb8888_to_argb8888 ============= [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======== [PASSED] drm_test_fb_xrgb8888_to_argb8888 ========= [04:36:43] =========== drm_test_fb_xrgb8888_to_xrgb2101010 =========== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======= [PASSED] drm_test_fb_xrgb8888_to_xrgb2101010 ======= [04:36:43] =========== drm_test_fb_xrgb8888_to_argb2101010 =========== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======= [PASSED] drm_test_fb_xrgb8888_to_argb2101010 ======= [04:36:43] ============== drm_test_fb_xrgb8888_to_mono =============== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ========== [PASSED] drm_test_fb_xrgb8888_to_mono =========== [04:36:43] ==================== drm_test_fb_swab ===================== [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ================ [PASSED] drm_test_fb_swab ================= [04:36:43] ============ drm_test_fb_xrgb8888_to_xbgr8888 ============= [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======== [PASSED] drm_test_fb_xrgb8888_to_xbgr8888 ========= [04:36:43] ============ drm_test_fb_xrgb8888_to_abgr8888 ============= [04:36:43] [PASSED] single_pixel_source_buffer [04:36:43] [PASSED] single_pixel_clip_rectangle [04:36:43] [PASSED] well_known_colors [04:36:43] [PASSED] destination_pitch [04:36:43] ======== [PASSED] drm_test_fb_xrgb8888_to_abgr8888 ========= [04:36:43] ================= drm_test_fb_clip_offset ================= [04:36:43] [PASSED] pass through [04:36:43] [PASSED] horizontal offset [04:36:43] [PASSED] vertical offset [04:36:43] [PASSED] horizontal and vertical offset [04:36:43] [PASSED] horizontal offset (custom pitch) [04:36:43] [PASSED] vertical offset (custom pitch) [04:36:43] [PASSED] horizontal and vertical offset (custom pitch) [04:36:43] ============= [PASSED] drm_test_fb_clip_offset ============= [04:36:43] =================== drm_test_fb_memcpy ==================== [04:36:43] [PASSED] single_pixel_source_buffer: XR24 little-endian (0x34325258) [04:36:43] [PASSED] single_pixel_source_buffer: XRA8 little-endian (0x38415258) [04:36:43] [PASSED] single_pixel_source_buffer: YU24 little-endian (0x34325559) [04:36:43] [PASSED] single_pixel_clip_rectangle: XB24 little-endian (0x34324258) [04:36:43] [PASSED] single_pixel_clip_rectangle: XRA8 little-endian (0x38415258) [04:36:43] [PASSED] single_pixel_clip_rectangle: YU24 little-endian (0x34325559) [04:36:43] [PASSED] well_known_colors: XB24 little-endian (0x34324258) [04:36:43] [PASSED] well_known_colors: XRA8 little-endian (0x38415258) [04:36:43] [PASSED] well_known_colors: YU24 little-endian (0x34325559) [04:36:43] [PASSED] destination_pitch: XB24 little-endian (0x34324258) [04:36:43] [PASSED] destination_pitch: XRA8 little-endian (0x38415258) [04:36:43] [PASSED] destination_pitch: YU24 little-endian (0x34325559) [04:36:43] =============== [PASSED] drm_test_fb_memcpy ================ [04:36:43] ============= [PASSED] drm_format_helper_test ============== [04:36:43] ================= drm_format (18 subtests) ================= [04:36:43] [PASSED] drm_test_format_block_width_invalid [04:36:43] [PASSED] drm_test_format_block_width_one_plane [04:36:43] [PASSED] drm_test_format_block_width_two_plane [04:36:43] [PASSED] drm_test_format_block_width_three_plane [04:36:43] [PASSED] drm_test_format_block_width_tiled [04:36:43] [PASSED] drm_test_format_block_height_invalid [04:36:43] [PASSED] drm_test_format_block_height_one_plane [04:36:43] [PASSED] drm_test_format_block_height_two_plane [04:36:43] [PASSED] drm_test_format_block_height_three_plane [04:36:43] [PASSED] drm_test_format_block_height_tiled [04:36:43] [PASSED] drm_test_format_min_pitch_invalid [04:36:43] [PASSED] drm_test_format_min_pitch_one_plane_8bpp [04:36:43] [PASSED] drm_test_format_min_pitch_one_plane_16bpp [04:36:43] [PASSED] drm_test_format_min_pitch_one_plane_24bpp [04:36:43] [PASSED] drm_test_format_min_pitch_one_plane_32bpp [04:36:43] [PASSED] drm_test_format_min_pitch_two_plane [04:36:43] [PASSED] drm_test_format_min_pitch_three_plane_8bpp [04:36:43] [PASSED] drm_test_format_min_pitch_tiled [04:36:43] =================== [PASSED] drm_format ==================== [04:36:43] ============== drm_framebuffer (10 subtests) =============== [04:36:43] ========== drm_test_framebuffer_check_src_coords ========== [04:36:43] [PASSED] Success: source fits into fb [04:36:43] [PASSED] Fail: overflowing fb with x-axis coordinate [04:36:43] [PASSED] Fail: overflowing fb with y-axis coordinate [04:36:43] [PASSED] Fail: overflowing fb with source width [04:36:43] [PASSED] Fail: overflowing fb with source height [04:36:43] ====== [PASSED] drm_test_framebuffer_check_src_coords ====== [04:36:43] [PASSED] drm_test_framebuffer_cleanup [04:36:43] =============== drm_test_framebuffer_create =============== [04:36:43] [PASSED] ABGR8888 normal sizes [04:36:43] [PASSED] ABGR8888 max sizes [04:36:43] [PASSED] ABGR8888 pitch greater than min required [04:36:43] [PASSED] ABGR8888 pitch less than min required [04:36:43] [PASSED] ABGR8888 Invalid width [04:36:43] [PASSED] ABGR8888 Invalid buffer handle [04:36:43] [PASSED] No pixel format [04:36:43] [PASSED] ABGR8888 Width 0 [04:36:43] [PASSED] ABGR8888 Height 0 [04:36:43] [PASSED] ABGR8888 Out of bound height * pitch combination [04:36:43] [PASSED] ABGR8888 Large buffer offset [04:36:43] [PASSED] ABGR8888 Buffer offset for inexistent plane [04:36:43] [PASSED] ABGR8888 Invalid flag [04:36:43] [PASSED] ABGR8888 Set DRM_MODE_FB_MODIFIERS without modifiers [04:36:43] [PASSED] ABGR8888 Valid buffer modifier [04:36:43] [PASSED] ABGR8888 Invalid buffer modifier(DRM_FORMAT_MOD_SAMSUNG_64_32_TILE) [04:36:43] [PASSED] ABGR8888 Extra pitches without DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] ABGR8888 Extra pitches with DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] NV12 Normal sizes [04:36:43] [PASSED] NV12 Max sizes [04:36:43] [PASSED] NV12 Invalid pitch [04:36:43] [PASSED] NV12 Invalid modifier/missing DRM_MODE_FB_MODIFIERS flag [04:36:43] [PASSED] NV12 different modifier per-plane [04:36:43] [PASSED] NV12 with DRM_FORMAT_MOD_SAMSUNG_64_32_TILE [04:36:43] [PASSED] NV12 Valid modifiers without DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] NV12 Modifier for inexistent plane [04:36:43] [PASSED] NV12 Handle for inexistent plane [04:36:43] [PASSED] NV12 Handle for inexistent plane without DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] YVU420 DRM_MODE_FB_MODIFIERS set without modifier [04:36:43] [PASSED] YVU420 Normal sizes [04:36:43] [PASSED] YVU420 Max sizes [04:36:43] [PASSED] YVU420 Invalid pitch [04:36:43] [PASSED] YVU420 Different pitches [04:36:43] [PASSED] YVU420 Different buffer offsets/pitches [04:36:43] [PASSED] YVU420 Modifier set just for plane 0, without DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] YVU420 Modifier set just for planes 0, 1, without DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] YVU420 Modifier set just for plane 0, 1, with DRM_MODE_FB_MODIFIERS [04:36:43] [PASSED] YVU420 Valid modifier [04:36:43] [PASSED] YVU420 Different modifiers per plane [04:36:43] [PASSED] YVU420 Modifier for inexistent plane [04:36:43] [PASSED] YUV420_10BIT Invalid modifier(DRM_FORMAT_MOD_LINEAR) [04:36:43] [PASSED] X0L2 Normal sizes [04:36:43] [PASSED] X0L2 Max sizes [04:36:43] [PASSED] X0L2 Invalid pitch [04:36:43] [PASSED] X0L2 Pitch greater than minimum required [04:36:43] [PASSED] X0L2 Handle for inexistent plane [04:36:43] [PASSED] X0L2 Offset for inexistent plane, without DRM_MODE_FB_MODIFIERS set [04:36:43] [PASSED] X0L2 Modifier without DRM_MODE_FB_MODIFIERS set [04:36:43] [PASSED] X0L2 Valid modifier [04:36:43] [PASSED] X0L2 Modifier for inexistent plane [04:36:43] =========== [PASSED] drm_test_framebuffer_create =========== [04:36:43] [PASSED] drm_test_framebuffer_free [04:36:43] [PASSED] drm_test_framebuffer_init [04:36:43] [PASSED] drm_test_framebuffer_init_bad_format [04:36:43] [PASSED] drm_test_framebuffer_init_dev_mismatch [04:36:43] [PASSED] drm_test_framebuffer_lookup [04:36:43] [PASSED] drm_test_framebuffer_lookup_inexistent [04:36:43] [PASSED] drm_test_framebuffer_modifiers_not_supported [04:36:43] ================= [PASSED] drm_framebuffer ================= [04:36:43] ================ drm_gem_shmem (8 subtests) ================ [04:36:43] [PASSED] drm_gem_shmem_test_obj_create [04:36:43] [PASSED] drm_gem_shmem_test_obj_create_private [04:36:43] [PASSED] drm_gem_shmem_test_pin_pages [04:36:43] [PASSED] drm_gem_shmem_test_vmap [04:36:43] [PASSED] drm_gem_shmem_test_get_pages_sgt [04:36:43] [PASSED] drm_gem_shmem_test_get_sg_table [04:36:43] [PASSED] drm_gem_shmem_test_madvise [04:36:43] [PASSED] drm_gem_shmem_test_purge [04:36:43] ================== [PASSED] drm_gem_shmem ================== [04:36:43] === drm_atomic_helper_connector_hdmi_check (27 subtests) === [04:36:43] [PASSED] drm_test_check_broadcast_rgb_auto_cea_mode [04:36:43] [PASSED] drm_test_check_broadcast_rgb_auto_cea_mode_vic_1 [04:36:43] [PASSED] drm_test_check_broadcast_rgb_full_cea_mode [04:36:43] [PASSED] drm_test_check_broadcast_rgb_full_cea_mode_vic_1 [04:36:43] [PASSED] drm_test_check_broadcast_rgb_limited_cea_mode [04:36:43] [PASSED] drm_test_check_broadcast_rgb_limited_cea_mode_vic_1 [04:36:43] ====== drm_test_check_broadcast_rgb_cea_mode_yuv420 ======= [04:36:43] [PASSED] Automatic [04:36:43] [PASSED] Full [04:36:43] [PASSED] Limited 16:235 [04:36:43] == [PASSED] drm_test_check_broadcast_rgb_cea_mode_yuv420 === [04:36:43] [PASSED] drm_test_check_broadcast_rgb_crtc_mode_changed [04:36:43] [PASSED] drm_test_check_broadcast_rgb_crtc_mode_not_changed [04:36:43] [PASSED] drm_test_check_disable_connector [04:36:43] [PASSED] drm_test_check_hdmi_funcs_reject_rate [04:36:43] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_rgb [04:36:43] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_yuv420 [04:36:43] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_ignore_yuv422 [04:36:43] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_ignore_yuv420 [04:36:43] [PASSED] drm_test_check_driver_unsupported_fallback_yuv420 [04:36:43] [PASSED] drm_test_check_output_bpc_crtc_mode_changed [04:36:43] [PASSED] drm_test_check_output_bpc_crtc_mode_not_changed [04:36:43] [PASSED] drm_test_check_output_bpc_dvi [04:36:43] [PASSED] drm_test_check_output_bpc_format_vic_1 [04:36:43] [PASSED] drm_test_check_output_bpc_format_display_8bpc_only [04:36:43] [PASSED] drm_test_check_output_bpc_format_display_rgb_only [04:36:43] [PASSED] drm_test_check_output_bpc_format_driver_8bpc_only [04:36:43] [PASSED] drm_test_check_output_bpc_format_driver_rgb_only [04:36:43] [PASSED] drm_test_check_tmds_char_rate_rgb_8bpc [04:36:43] [PASSED] drm_test_check_tmds_char_rate_rgb_10bpc [04:36:43] [PASSED] drm_test_check_tmds_char_rate_rgb_12bpc [04:36:43] ===== [PASSED] drm_atomic_helper_connector_hdmi_check ====== [04:36:43] === drm_atomic_helper_connector_hdmi_reset (6 subtests) ==== [04:36:43] [PASSED] drm_test_check_broadcast_rgb_value [04:36:43] [PASSED] drm_test_check_bpc_8_value [04:36:43] [PASSED] drm_test_check_bpc_10_value [04:36:43] [PASSED] drm_test_check_bpc_12_value [04:36:43] [PASSED] drm_test_check_format_value [04:36:43] [PASSED] drm_test_check_tmds_char_value [04:36:43] ===== [PASSED] drm_atomic_helper_connector_hdmi_reset ====== [04:36:43] = drm_atomic_helper_connector_hdmi_mode_valid (4 subtests) = [04:36:43] [PASSED] drm_test_check_mode_valid [04:36:43] [PASSED] drm_test_check_mode_valid_reject [04:36:43] [PASSED] drm_test_check_mode_valid_reject_rate [04:36:43] [PASSED] drm_test_check_mode_valid_reject_max_clock [04:36:43] === [PASSED] drm_atomic_helper_connector_hdmi_mode_valid === [04:36:43] ================= drm_managed (2 subtests) ================= [04:36:43] [PASSED] drm_test_managed_release_action [04:36:43] [PASSED] drm_test_managed_run_action [04:36:43] =================== [PASSED] drm_managed =================== [04:36:43] =================== drm_mm (6 subtests) ==================== [04:36:43] [PASSED] drm_test_mm_init [04:36:43] [PASSED] drm_test_mm_debug [04:36:43] [PASSED] drm_test_mm_align32 [04:36:43] [PASSED] drm_test_mm_align64 [04:36:43] [PASSED] drm_test_mm_lowest [04:36:43] [PASSED] drm_test_mm_highest [04:36:43] ===================== [PASSED] drm_mm ====================== [04:36:43] ============= drm_modes_analog_tv (5 subtests) ============= [04:36:43] [PASSED] drm_test_modes_analog_tv_mono_576i [04:36:43] [PASSED] drm_test_modes_analog_tv_ntsc_480i [04:36:43] [PASSED] drm_test_modes_analog_tv_ntsc_480i_inlined [04:36:43] [PASSED] drm_test_modes_analog_tv_pal_576i [04:36:43] [PASSED] drm_test_modes_analog_tv_pal_576i_inlined [04:36:43] =============== [PASSED] drm_modes_analog_tv =============== [04:36:43] ============== drm_plane_helper (2 subtests) =============== [04:36:43] =============== drm_test_check_plane_state ================ [04:36:43] [PASSED] clipping_simple [04:36:43] [PASSED] clipping_rotate_reflect [04:36:43] [PASSED] positioning_simple [04:36:43] [PASSED] upscaling [04:36:43] [PASSED] downscaling [04:36:43] [PASSED] rounding1 [04:36:43] [PASSED] rounding2 [04:36:43] [PASSED] rounding3 [04:36:43] [PASSED] rounding4 [04:36:43] =========== [PASSED] drm_test_check_plane_state ============ [04:36:43] =========== drm_test_check_invalid_plane_state ============ [04:36:43] [PASSED] positioning_invalid [04:36:43] [PASSED] upscaling_invalid [04:36:43] [PASSED] downscaling_invalid [04:36:43] ======= [PASSED] drm_test_check_invalid_plane_state ======== [04:36:43] ================ [PASSED] drm_plane_helper ================= [04:36:43] ====== drm_connector_helper_tv_get_modes (1 subtest) ======= [04:36:43] ====== drm_test_connector_helper_tv_get_modes_check ======= [04:36:43] [PASSED] None [04:36:43] [PASSED] PAL [04:36:43] [PASSED] NTSC [04:36:43] [PASSED] Both, NTSC Default [04:36:43] [PASSED] Both, PAL Default [04:36:43] [PASSED] Both, NTSC Default, with PAL on command-line [04:36:43] [PASSED] Both, PAL Default, with NTSC on command-line [04:36:43] == [PASSED] drm_test_connector_helper_tv_get_modes_check === [04:36:43] ======== [PASSED] drm_connector_helper_tv_get_modes ======== [04:36:43] ================== drm_rect (9 subtests) =================== [04:36:43] [PASSED] drm_test_rect_clip_scaled_div_by_zero [04:36:43] [PASSED] drm_test_rect_clip_scaled_not_clipped [04:36:43] [PASSED] drm_test_rect_clip_scaled_clipped [04:36:43] [PASSED] drm_test_rect_clip_scaled_signed_vs_unsigned [04:36:43] ================= drm_test_rect_intersect ================= [04:36:43] [PASSED] top-left x bottom-right: 2x2+1+1 x 2x2+0+0 [04:36:43] [PASSED] top-right x bottom-left: 2x2+0+0 x 2x2+1-1 [04:36:43] [PASSED] bottom-left x top-right: 2x2+1-1 x 2x2+0+0 [04:36:43] [PASSED] bottom-right x top-left: 2x2+0+0 x 2x2+1+1 [04:36:43] [PASSED] right x left: 2x1+0+0 x 3x1+1+0 [04:36:43] [PASSED] left x right: 3x1+1+0 x 2x1+0+0 [04:36:43] [PASSED] up x bottom: 1x2+0+0 x 1x3+0-1 [04:36:43] [PASSED] bottom x up: 1x3+0-1 x 1x2+0+0 [04:36:43] [PASSED] touching corner: 1x1+0+0 x 2x2+1+1 [04:36:43] [PASSED] touching side: 1x1+0+0 x 1x1+1+0 [04:36:43] [PASSED] equal rects: 2x2+0+0 x 2x2+0+0 [04:36:43] [PASSED] inside another: 2x2+0+0 x 1x1+1+1 [04:36:43] [PASSED] far away: 1x1+0+0 x 1x1+3+6 [04:36:43] [PASSED] points intersecting: 0x0+5+10 x 0x0+5+10 [04:36:43] [PASSED] points not intersecting: 0x0+0+0 x 0x0+5+10 [04:36:43] ============= [PASSED] drm_test_rect_intersect ============= [04:36:43] ================ drm_test_rect_calc_hscale ================ [04:36:43] [PASSED] normal use [04:36:43] [PASSED] out of max range [04:36:43] [PASSED] out of min range [04:36:43] [PASSED] zero dst [04:36:43] [PASSED] negative src [04:36:43] [PASSED] negative dst [04:36:43] ============ [PASSED] drm_test_rect_calc_hscale ============ [04:36:43] ================ drm_test_rect_calc_vscale ================ [04:36:43] [PASSED] normal use stty: 'standard input': Inappropriate ioctl for device [04:36:43] [PASSED] out of max range [04:36:43] [PASSED] out of min range [04:36:43] [PASSED] zero dst [04:36:43] [PASSED] negative src [04:36:43] [PASSED] negative dst [04:36:43] ============ [PASSED] drm_test_rect_calc_vscale ============ [04:36:43] ================== drm_test_rect_rotate =================== [04:36:43] [PASSED] reflect-x [04:36:43] [PASSED] reflect-y [04:36:43] [PASSED] rotate-0 [04:36:43] [PASSED] rotate-90 [04:36:43] [PASSED] rotate-180 [04:36:43] [PASSED] rotate-270 [04:36:43] ============== [PASSED] drm_test_rect_rotate =============== [04:36:43] ================ drm_test_rect_rotate_inv ================= [04:36:43] [PASSED] reflect-x [04:36:43] [PASSED] reflect-y [04:36:43] [PASSED] rotate-0 [04:36:43] [PASSED] rotate-90 [04:36:43] [PASSED] rotate-180 [04:36:43] [PASSED] rotate-270 [04:36:43] ============ [PASSED] drm_test_rect_rotate_inv ============= [04:36:43] ==================== [PASSED] drm_rect ===================== [04:36:43] ============ drm_sysfb_modeset_test (1 subtest) ============ [04:36:43] ============ drm_test_sysfb_build_fourcc_list ============= [04:36:43] [PASSED] no native formats [04:36:43] [PASSED] XRGB8888 as native format [04:36:43] [PASSED] remove duplicates [04:36:43] [PASSED] convert alpha formats [04:36:43] [PASSED] random formats [04:36:43] ======== [PASSED] drm_test_sysfb_build_fourcc_list ========= [04:36:43] ============= [PASSED] drm_sysfb_modeset_test ============== [04:36:43] ============================================================ [04:36:43] Testing complete. Ran 622 tests: passed: 622 [04:36:44] Elapsed time: 31.318s total, 1.607s configuring, 29.193s building, 0.467s running
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/ttm/tests/.kunitconfig [04:36:44] Configuring KUnit Kernel ...
Regenerating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [04:36:45] Building KUnit Kernel ... Populating config with: $ make ARCH=um O=.kunit olddefconfig Building with: $ make all compile_commands.json scripts_gdb ARCH=um O=.kunit --jobs=25 [04:36:54] Starting KUnit Kernel (1/1)... [04:36:54] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [04:36:54] ================= ttm_device (5 subtests) ================== [04:36:54] [PASSED] ttm_device_init_basic [04:36:54] [PASSED] ttm_device_init_multiple [04:36:54] [PASSED] ttm_device_fini_basic [04:36:54] [PASSED] ttm_device_init_no_vma_man [04:36:54] ================== ttm_device_init_pools ================== [04:36:54] [PASSED] No DMA allocations, no DMA32 required [04:36:54] [PASSED] DMA allocations, DMA32 required [04:36:54] [PASSED] No DMA allocations, DMA32 required [04:36:54] [PASSED] DMA allocations, no DMA32 required [04:36:54] ============== [PASSED] ttm_device_init_pools ============== [04:36:54] =================== [PASSED] ttm_device ==================== [04:36:54] ================== ttm_pool (8 subtests) =================== [04:36:54] ================== ttm_pool_alloc_basic =================== [04:36:54] [PASSED] One page [04:36:54] [PASSED] More than one page [04:36:54] [PASSED] Above the allocation limit [04:36:54] [PASSED] One page, with coherent DMA mappings enabled [04:36:54] [PASSED] Above the allocation limit, with coherent DMA mappings enabled [04:36:54] ============== [PASSED] ttm_pool_alloc_basic =============== [04:36:54] ============== ttm_pool_alloc_basic_dma_addr ============== [04:36:54] [PASSED] One page [04:36:54] [PASSED] More than one page [04:36:54] [PASSED] Above the allocation limit [04:36:54] [PASSED] One page, with coherent DMA mappings enabled [04:36:54] [PASSED] Above the allocation limit, with coherent DMA mappings enabled [04:36:54] ========== [PASSED] ttm_pool_alloc_basic_dma_addr ========== [04:36:54] [PASSED] ttm_pool_alloc_order_caching_match [04:36:54] [PASSED] ttm_pool_alloc_caching_mismatch [04:36:54] [PASSED] ttm_pool_alloc_order_mismatch [04:36:54] [PASSED] ttm_pool_free_dma_alloc [04:36:54] [PASSED] ttm_pool_free_no_dma_alloc [04:36:54] [PASSED] ttm_pool_fini_basic [04:36:54] ==================== [PASSED] ttm_pool ===================== [04:36:54] ================ ttm_resource (8 subtests) ================= [04:36:54] ================= ttm_resource_init_basic ================= [04:36:54] [PASSED] Init resource in TTM_PL_SYSTEM [04:36:54] [PASSED] Init resource in TTM_PL_VRAM [04:36:54] [PASSED] Init resource in a private placement [04:36:54] [PASSED] Init resource in TTM_PL_SYSTEM, set placement flags [04:36:54] ============= [PASSED] ttm_resource_init_basic ============= [04:36:54] [PASSED] ttm_resource_init_pinned [04:36:54] [PASSED] ttm_resource_fini_basic [04:36:54] [PASSED] ttm_resource_manager_init_basic [04:36:54] [PASSED] ttm_resource_manager_usage_basic [04:36:54] [PASSED] ttm_resource_manager_set_used_basic [04:36:54] [PASSED] ttm_sys_man_alloc_basic [04:36:54] [PASSED] ttm_sys_man_free_basic [04:36:54] ================== [PASSED] ttm_resource =================== [04:36:54] =================== ttm_tt (15 subtests) =================== [04:36:54] ==================== ttm_tt_init_basic ==================== [04:36:54] [PASSED] Page-aligned size [04:36:54] [PASSED] Extra pages requested [04:36:54] ================ [PASSED] ttm_tt_init_basic ================ [04:36:54] [PASSED] ttm_tt_init_misaligned [04:36:54] [PASSED] ttm_tt_fini_basic [04:36:54] [PASSED] ttm_tt_fini_sg [04:36:54] [PASSED] ttm_tt_fini_shmem [04:36:54] [PASSED] ttm_tt_create_basic [04:36:54] [PASSED] ttm_tt_create_invalid_bo_type [04:36:54] [PASSED] ttm_tt_create_ttm_exists [04:36:54] [PASSED] ttm_tt_create_failed [04:36:54] [PASSED] ttm_tt_destroy_basic [04:36:54] [PASSED] ttm_tt_populate_null_ttm [04:36:54] [PASSED] ttm_tt_populate_populated_ttm [04:36:54] [PASSED] ttm_tt_unpopulate_basic [04:36:54] [PASSED] ttm_tt_unpopulate_empty_ttm [04:36:54] [PASSED] ttm_tt_swapin_basic [04:36:54] ===================== [PASSED] ttm_tt ====================== [04:36:54] =================== ttm_bo (14 subtests) =================== [04:36:54] =========== ttm_bo_reserve_optimistic_no_ticket =========== [04:36:54] [PASSED] Cannot be interrupted and sleeps [04:36:54] [PASSED] Cannot be interrupted, locks straight away [04:36:54] [PASSED] Can be interrupted, sleeps [04:36:54] ======= [PASSED] ttm_bo_reserve_optimistic_no_ticket ======= [04:36:54] [PASSED] ttm_bo_reserve_locked_no_sleep [04:36:54] [PASSED] ttm_bo_reserve_no_wait_ticket [04:36:54] [PASSED] ttm_bo_reserve_double_resv [04:36:54] [PASSED] ttm_bo_reserve_interrupted [04:36:54] [PASSED] ttm_bo_reserve_deadlock [04:36:54] [PASSED] ttm_bo_unreserve_basic [04:36:54] [PASSED] ttm_bo_unreserve_pinned [04:36:54] [PASSED] ttm_bo_unreserve_bulk [04:36:54] [PASSED] ttm_bo_fini_basic [04:36:54] [PASSED] ttm_bo_fini_shared_resv [04:36:54] [PASSED] ttm_bo_pin_basic [04:36:54] [PASSED] ttm_bo_pin_unpin_resource [04:36:54] [PASSED] ttm_bo_multiple_pin_one_unpin [04:36:54] ===================== [PASSED] ttm_bo ====================== [04:36:54] ============== ttm_bo_validate (21 subtests) =============== [04:36:54] ============== ttm_bo_init_reserved_sys_man =============== [04:36:54] [PASSED] Buffer object for userspace [04:36:54] [PASSED] Kernel buffer object [04:36:54] [PASSED] Shared buffer object [04:36:54] ========== [PASSED] ttm_bo_init_reserved_sys_man =========== [04:36:54] ============== ttm_bo_init_reserved_mock_man ============== [04:36:54] [PASSED] Buffer object for userspace [04:36:54] [PASSED] Kernel buffer object [04:36:54] [PASSED] Shared buffer object [04:36:54] ========== [PASSED] ttm_bo_init_reserved_mock_man ========== [04:36:54] [PASSED] ttm_bo_init_reserved_resv [04:36:54] ================== ttm_bo_validate_basic ================== [04:36:54] [PASSED] Buffer object for userspace [04:36:54] [PASSED] Kernel buffer object [04:36:54] [PASSED] Shared buffer object [04:36:54] ============== [PASSED] ttm_bo_validate_basic ============== [04:36:54] [PASSED] ttm_bo_validate_invalid_placement [04:36:54] ============= ttm_bo_validate_same_placement ============== [04:36:54] [PASSED] System manager [04:36:54] [PASSED] VRAM manager [04:36:54] ========= [PASSED] ttm_bo_validate_same_placement ========== [04:36:54] [PASSED] ttm_bo_validate_failed_alloc [04:36:54] [PASSED] ttm_bo_validate_pinned [04:36:54] [PASSED] ttm_bo_validate_busy_placement [04:36:54] ================ ttm_bo_validate_multihop ================= [04:36:54] [PASSED] Buffer object for userspace [04:36:54] [PASSED] Kernel buffer object [04:36:54] [PASSED] Shared buffer object [04:36:54] ============ [PASSED] ttm_bo_validate_multihop ============= [04:36:54] ========== ttm_bo_validate_no_placement_signaled ========== [04:36:54] [PASSED] Buffer object in system domain, no page vector [04:36:54] [PASSED] Buffer object in system domain with an existing page vector [04:36:54] ====== [PASSED] ttm_bo_validate_no_placement_signaled ====== [04:36:54] ======== ttm_bo_validate_no_placement_not_signaled ======== [04:36:54] [PASSED] Buffer object for userspace [04:36:54] [PASSED] Kernel buffer object [04:36:54] [PASSED] Shared buffer object [04:36:54] ==== [PASSED] ttm_bo_validate_no_placement_not_signaled ==== [04:36:54] [PASSED] ttm_bo_validate_move_fence_signaled [04:36:54] ========= ttm_bo_validate_move_fence_not_signaled ========= [04:36:54] [PASSED] Waits for GPU [04:36:54] [PASSED] Tries to lock straight away [04:36:54] ===== [PASSED] ttm_bo_validate_move_fence_not_signaled ===== [04:36:54] [PASSED] ttm_bo_validate_happy_evict [04:36:54] [PASSED] ttm_bo_validate_all_pinned_evict [04:36:54] [PASSED] ttm_bo_validate_allowed_only_evict [04:36:54] [PASSED] ttm_bo_validate_deleted_evict [04:36:54] [PASSED] ttm_bo_validate_busy_domain_evict [04:36:54] [PASSED] ttm_bo_validate_evict_gutting [04:36:54] [PASSED] ttm_bo_validate_recrusive_evict stty: 'standard input': Inappropriate ioctl for device [04:36:54] ================= [PASSED] ttm_bo_validate ================= [04:36:54] ============================================================ [04:36:54] Testing complete. Ran 101 tests: passed: 101 [04:36:54] Elapsed time: 10.872s total, 1.605s configuring, 9.050s building, 0.183s running
+ cleanup
++ stat -c %u:%g /kernel
+ chown -R 1003:1003 /kernel
Patchwork
CI✗ Xe.CI.Full: failure for PF: Add sriov_admin sysfs tree
In reply to Michal Wajdeczko== Series Details ==
== Series Details ==
Series: PF: Add sriov_admin sysfs tree
URL : https://patchwork.freedesktop.org/series/156220/
State : failure
== Summary ==
ERROR: The runconfig 'xe-3951-aae2e4df375e567f00c8d494004cf6a34a73d75a_FULL' does not exist in the database
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-156220v1/index.html