-----Original Message-----
From: Lin, Shuicheng <shuicheng....@intel.com>
Sent: Wednesday, April 23, 2025 8:49 PM
To: Cavitt, Jonathan <jonathan.cav...@intel.com>; intel...@lists.freedesktop.org
Cc: Gupta, saurabhg <saurabhg.gu...@intel.com>; Zuo, Alex <alex....@intel.com>;
joonas.lahti...@linux.intel.com; Brost, Matthew <matthew.br...@intel.com>;
Zhang, Jianxun <jianxun.zh...@intel.com>; dri-devel@lists.freedesktop.org;
Wajdeczko, Michal <michal.wajdec...@intel.com>; Mrozek, Michal
<michal.mro...@intel.com>; Jadav, Raag <raag.ja...@intel.com>; Harrison, John C
<john.c.harri...@intel.com>; Briano, Ivan <ivan.bri...@intel.com>; Auld,
Matthew <matthew.a...@intel.com>
Subject: RE: [PATCH v21 4/5] drm/xe/xe_vm: Add per VM fault info
>
> On Wed, April 23, 2025 1:19 PM Cavitt, Jonathan wrote:
> > Add additional information to each VM so they can report up to the first
> > 50 seen faults. Only pagefaults are saved this way currently, though in the
> > future, all faults should be tracked by the VM for future reporting.
> >
> > Additionally, of the pagefaults reported, only failed pagefaults are saved
> > this
> > way, as successful pagefaults should recover silently and not need to be
> > reported to userspace.
> >
> > v2:
> > - Free vm after use (Shuicheng)
> > - Compress pf copy logic (Shuicheng)
> > - Update fault_unsuccessful before storing (Shuicheng)
> > - Fix old struct name in comments (Shuicheng)
> > - Keep first 50 pagefaults instead of last 50 (Jianxun)
> >
> > v3:
> > - Avoid unnecessary execution by checking MAX_PFS earlier (jcavitt)
> > - Fix double-locking error (jcavitt)
> > - Assert kmemdump is successful (Shuicheng)
> >
> > v4:
> > - Rename xe_vm.pfs to xe_vm.faults (jcavitt)
> > - Store fault data and not pagefault in xe_vm faults list (jcavitt)
> > - Store address, address type, and address precision per fault (jcavitt)
> > - Store engine class and instance data per fault (Jianxun)
> > - Add and fix kernel docs (Michal W)
> > - Properly handle kzalloc error (Michal W)
> > - s/MAX_PFS/MAX_FAULTS_SAVED_PER_VM (Michal W)
> > - Store fault level per fault (Micahl M)
> >
> > v5:
> > - Store fault and access type instead of address type (Jianxun)
> >
> > v6:
> > - Store pagefaults in non-fault-mode VMs as well (Jianxun)
> >
> > v7:
> > - Fix kernel docs and comments (Michal W)
> >
> > v8:
> > - Fix double-locking issue (Jianxun)
> >
> > v9:
> > - Do not report faults from reserved engines (Jianxun)
> >
> > v10:
> > - Remove engine class and instance (Ivan)
> >
> > v11:
> > - Perform kzalloc outside of lock (Auld)
> >
> > Signed-off-by: Jonathan Cavitt <jonathan.cav...@intel.com>
> > Suggested-by: Matthew Brost <matthew.br...@intel.com>
> > Cc: Shuicheng Lin <shuicheng....@intel.com>
> > Cc: Jianxun Zhang <jianxun.zh...@intel.com>
> > Cc: Michal Wajdeczko <michal.wajdec...@intel.com>
> > Cc: Michal Mzorek <michal.mzo...@intel.com>
> > Cc: Ivan Briano <ivan.bri...@intel.com>
> > Cc: Matthew Auld <matthew.a...@intel.com>
> > ---
> > drivers/gpu/drm/xe/xe_gt_pagefault.c | 26 ++++++++
> > drivers/gpu/drm/xe/xe_vm.c | 88 ++++++++++++++++++++++++++++
> > drivers/gpu/drm/xe/xe_vm.h | 9 +++
> > drivers/gpu/drm/xe/xe_vm_types.h | 28 +++++++++
> > 4 files changed, 151 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > b/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > index 93afa54c8780..a84f6247f8a2 100644
> > --- a/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > +++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > @@ -345,6 +345,31 @@ int xe_guc_pagefault_handler(struct xe_guc *guc,
> > u32 *msg, u32 len)
> > return full ? -ENOSPC : 0;
> > }
> >
> > +static void save_pagefault_to_vm(struct xe_device *xe, struct
> > +xe_gt_pagefault *pf) {
> > + struct xe_vm *vm;
> > +
> > + /*
> > + * Pagefault may be associated to VM that is not in fault mode.
> > + * Perform asid_to_vm behavior, except if vm is not in fault
> > + * mode, return the VM anyways.
> > + */
> > + down_read(&xe->usm.lock);
> > + vm = xa_load(&xe->usm.asid_to_vm, pf->asid);
> > + if (vm)
> > + xe_vm_get(vm);
> > + else
> > + vm = ERR_PTR(-EINVAL);
> > + up_read(&xe->usm.lock);
> > +
> > + if (IS_ERR(vm))
> > + return;
> > +
> > + xe_vm_add_fault_entry_pf(vm, pf);
> > +
> > + xe_vm_put(vm);
> > +}
> > +
> > #define USM_QUEUE_MAX_RUNTIME_MS 20
> >
> > static void pf_queue_work_func(struct work_struct *w) @@ -364,6 +389,7
> > @@ static void pf_queue_work_func(struct work_struct *w)
> > if (unlikely(ret)) {
> > print_pagefault(xe, &pf);
> > pf.fault_unsuccessful = 1;
> > + save_pagefault_to_vm(xe, &pf);
> > drm_dbg(&xe->drm, "Fault response:
> > Unsuccessful %d\n", ret);
> > }
> >
> > diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index
> > 0c69ef6b5ec5..107e397b4987 100644
> > --- a/drivers/gpu/drm/xe/xe_vm.c
> > +++ b/drivers/gpu/drm/xe/xe_vm.c
> > @@ -27,7 +27,9 @@
> > #include "xe_device.h"
> > #include "xe_drm_client.h"
> > #include "xe_exec_queue.h"
> > +#include "xe_gt.h"
> > #include "xe_gt_pagefault.h"
> > +#include "xe_gt_pagefault_types.h"
> > #include "xe_gt_tlb_invalidation.h"
> > #include "xe_migrate.h"
> > #include "xe_pat.h"
> > @@ -778,6 +780,87 @@ int xe_vm_userptr_check_repin(struct xe_vm *vm)
> > list_empty_careful(&vm->userptr.invalidated)) ? 0 : -
> > EAGAIN; }
> >
> > +static struct xe_hw_engine *
> > +hw_engine_lookup_class_instance(struct xe_vm *vm,
> > + enum xe_engine_class class,
> > + u16 instance)
> > +{
> > + struct xe_device *xe = vm->xe;
> > + struct xe_hw_engine *hwe;
> > + enum xe_hw_engine_id id;
> > + struct xe_gt *gt;
> > + u8 gt_id;
> > +
> > + for_each_gt(gt, xe, gt_id)
> > + for_each_hw_engine(hwe, gt, id)
> > + if (hwe->class == class && hwe->instance == instance)
> > + return hwe;
> > + return NULL;
> > +}
>
> There is a similar function xe_gt_hw_engine() to lookup to hwe.
> Anyway, I am ok with the function.
This is correct. The problem is that xe_gt_hw_engine takes a
struct xe_gt *gt, which is not provided by the PF structure.
Though, I suppose this wasn't necessary to explain since you're
already okay with the function.
>
> Shuicheng
>
> > +
> > +/**
> > + * xe_vm_add_fault_entry_pf() - Add pagefault to vm fault list
> > + * @vm: The VM.
> > + * @pf: The pagefault.
> > + *
> > + * This function takes the data from the pagefault @pf and saves it to @vm-
> > >faults.list.
> > + *
> > + * The function exits silently if the list is full, and reports a
> > +warning if the pagefault
> > + * could not be saved to the list.
> > + */
> > +void xe_vm_add_fault_entry_pf(struct xe_vm *vm, struct xe_gt_pagefault
> > +*pf) {
> > + struct xe_vm_fault_entry *e = NULL;
> > + struct xe_hw_engine *hwe;
> > +
> > + /* Do not report faults on reserved engines */
> > + hwe = hw_engine_lookup_class_instance(vm, pf->engine_class, pf-
> > >engine_instance);
> > + if (!hwe || xe_hw_engine_is_reserved(hwe))
> > + return;
> > +
> > + e = kzalloc(sizeof(*e), GFP_KERNEL);
> > + if (!e) {
> > + drm_warn(&vm->xe->drm,
> > + "Could not allocate memory for fault!\n");
> > + return;
> > + }
> > +
> > + spin_lock(&vm->faults.lock);
> > +
> > + /*
> > + * Limit the number of faults in the fault list to prevent
> > + * memory overuse.
> > + */
> > + if (vm->faults.len >= MAX_FAULTS_SAVED_PER_VM) {
> > + kfree(e);
> > + goto out;
> > + }
> > +
> > + e->address = pf->page_addr;
> > + e->address_precision = 1;
> > + e->access_type = pf->access_type;
> > + e->fault_type = pf->fault_type;
> > + e->fault_level = pf->fault_level;
> > +
> > + list_add_tail(&e->list, &vm->faults.list);
> > + vm->faults.len++;
> > +out:
> > + spin_unlock(&vm->faults.lock);
> > +}
> > +
> > +static void xe_vm_clear_fault_entries(struct xe_vm *vm) {
> > + struct xe_vm_fault_entry *e, *tmp;
> > +
> > + spin_lock(&vm->faults.lock);
> > + list_for_each_entry_safe(e, tmp, &vm->faults.list, list) {
> > + list_del(&e->list);
> > + kfree(e);
> > + }
> > + vm->faults.len = 0;
> > + spin_unlock(&vm->faults.lock);
> > +}
> > +
> > static int xe_vma_ops_alloc(struct xe_vma_ops *vops, bool array_of_binds)
> > {
> > int i;
> > @@ -1660,6 +1743,9 @@ struct xe_vm *xe_vm_create(struct xe_device *xe,
> > u32 flags)
> > init_rwsem(&vm->userptr.notifier_lock);
> > spin_lock_init(&vm->userptr.invalidated_lock);
> >
> > + INIT_LIST_HEAD(&vm->faults.list);
> > + spin_lock_init(&vm->faults.lock);
> > +
> > ttm_lru_bulk_move_init(&vm->lru_bulk_move);
> >
> > INIT_WORK(&vm->destroy_work, vm_destroy_work_func); @@ -
> > 1930,6 +2016,8 @@ void xe_vm_close_and_put(struct xe_vm *vm)
> > }
> > up_write(&xe->usm.lock);
> >
> > + xe_vm_clear_fault_entries(vm);
> > +
> > for_each_tile(tile, xe, id)
> > xe_range_fence_tree_fini(&vm->rftree[id]);
> >
> > diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h index
> > 0ef811fc2bde..9bd7e93824da 100644
> > --- a/drivers/gpu/drm/xe/xe_vm.h
> > +++ b/drivers/gpu/drm/xe/xe_vm.h
> > @@ -12,6 +12,12 @@
> > #include "xe_map.h"
> > #include "xe_vm_types.h"
> >
> > +/**
> > + * MAX_FAULTS_SAVED_PER_VM - Maximum number of faults each vm can
> > store
> > +before future
> > + * faults are discarded to prevent memory overuse */
> > +#define MAX_FAULTS_SAVED_PER_VM 50
> > +
> > struct drm_device;
> > struct drm_printer;
> > struct drm_file;
> > @@ -22,6 +28,7 @@ struct dma_fence;
> >
> > struct xe_exec_queue;
> > struct xe_file;
> > +struct xe_gt_pagefault;
> > struct xe_sync_entry;
> > struct xe_svm_range;
> > struct drm_exec;
> > @@ -257,6 +264,8 @@ int xe_vma_userptr_pin_pages(struct
> > xe_userptr_vma *uvma);
> >
> > int xe_vma_userptr_check_repin(struct xe_userptr_vma *uvma);
> >
> > +void xe_vm_add_fault_entry_pf(struct xe_vm *vm, struct xe_gt_pagefault
> > +*pf);
> > +
> > bool xe_vm_validate_should_retry(struct drm_exec *exec, int err, ktime_t
> > *end);
> >
> > int xe_vm_lock_vma(struct drm_exec *exec, struct xe_vma *vma); diff --git
> > a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
> > index 1662604c4486..ff67d3c7145d 100644
> > --- a/drivers/gpu/drm/xe/xe_vm_types.h
> > +++ b/drivers/gpu/drm/xe/xe_vm_types.h
> > @@ -19,6 +19,7 @@
> > #include "xe_range_fence.h"
> >
> > struct xe_bo;
> > +struct xe_pagefault;
> > struct xe_svm_range;
> > struct xe_sync_entry;
> > struct xe_user_fence;
> > @@ -142,6 +143,23 @@ struct xe_userptr_vma {
> >
> > struct xe_device;
> >
> > +/**
> > + * struct xe_vm_fault_entry - Elements of vm->faults.list
> > + * @list: link into @xe_vm.faults.list
> > + * @address: address of the fault
> > + * @address_type: type of address access that resulted in fault
>
> Should be access_type? And fault_type is missing.
> Also, the sequence is different from the structure definition.
I think this was a fragment from an earlier version of the patch
that never got updated to the most recent revision. I'll update
these kernel docs shortly.
-Jonathan Cavitt
>
> Shuicheng
>
> > + * @address_precision: precision of faulted address
> > + * @fault_level: fault level of the fault */ struct xe_vm_fault_entry
> > +{
> > + struct list_head list;
> > + u64 address;
> > + u32 address_precision;
> > + u8 access_type;
> > + u8 fault_type;
> > + u8 fault_level;
> > +};
> > +
> > struct xe_vm {
> > /** @gpuvm: base GPUVM used to track VMAs */
> > struct drm_gpuvm gpuvm;
> > @@ -305,6 +323,16 @@ struct xe_vm {
> > bool capture_once;
> > } error_capture;
> >
> > + /** @faults: List of all faults associated with this VM */
> > + struct {
> > + /** @faults.lock: lock protecting @faults.list */
> > + spinlock_t lock;
> > + /** @faults.list: list of xe_vm_fault_entry entries */
> > + struct list_head list;
> > + /** @faults.len: length of @faults.list */
> > + unsigned int len;
> > + } faults;
> > +
> > /**
> > * @tlb_flush_seqno: Required TLB flush seqno for the next exec.
> > * protected by the vm resv.
> > --
> > 2.43.0
>
>
