On Thu, Jun 11, 2026 at 11:29 AM Srinivasan Shanmugam <[email protected]> wrote: > > The current EVENTFD UAPI uses a userspace-defined event identifier. > > For render-node event notifications, userspace should subscribe to > kernel-defined event types instead, allowing the kernel to define event > semantics and signaling sources. > > Add a shared AMDGPU EVENTFD event type enum and update the EVENTFD UAPI > to use event_type instead of event_id. > > Value 0 is reserved and rejected by the existing event_type validation. > > queue_id remains part of the UAPI and is used to distinguish > queue-scoped subscriptions from GPU/device-scoped subscriptions. > > Eventfd signaling remains notification-only. > > v10: (per Christian) > - Add a comment clarifying that the flags field is currently unused, > reserved for future UAPI extensions, and must be zero. > - Introduce event_type directly in the EVENTFD UAPI. > - Remove the INVALID event type; value 0 remains rejected by validation. > - Squash the event_id to event_type rename into the original UAPI patch. > > v11: (per Christian) > - Keep the event_id to event_type rename in this patch so that later > patches focus only on queue reference handling. > - Move queue reference conversion and queue-scoped EVENTFD handling > into the subsequent patch to reduce noise in this change. > > Cc: Alex Deucher <[email protected]> > Cc: Christian König <[email protected]> > Signed-off-by: Srinivasan Shanmugam <[email protected]> > Reviewed-by: Christian König <[email protected]> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu.h | 5 + > drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 1 + > drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.c | 107 +++++++++----------- > drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.h | 17 ++-- > drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 37 +++++++ > include/uapi/drm/amdgpu_drm.h | 43 ++++++++ > 6 files changed, 145 insertions(+), 65 deletions(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h > b/drivers/gpu/drm/amd/amdgpu/amdgpu.h > index a75c68195df9..54bc31ee795a 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h > @@ -460,6 +460,11 @@ struct amdgpu_fpriv { > struct amdgpu_eventfd_mgr eventfd_mgr; > }; > > +struct drm_device; > +struct drm_file; > + > +int amdgpu_eventfd_ioctl(struct drm_device *dev, void *data, struct drm_file > *file_priv); > + > int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv); > > /* > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c > index 1781c0c3d010..aaa4dd57099f 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c > @@ -3076,6 +3076,7 @@ const struct drm_ioctl_desc amdgpu_ioctls_kms[] = { > DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_SIGNAL, amdgpu_userq_signal_ioctl, > DRM_AUTH|DRM_RENDER_ALLOW), > DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_WAIT, amdgpu_userq_wait_ioctl, > DRM_AUTH|DRM_RENDER_ALLOW), > DRM_IOCTL_DEF_DRV(AMDGPU_GEM_LIST_HANDLES, > amdgpu_gem_list_handles_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), > + DRM_IOCTL_DEF_DRV(AMDGPU_EVENTFD, amdgpu_eventfd_ioctl, > DRM_RENDER_ALLOW), > }; > > static const struct drm_driver amdgpu_kms_driver = { > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.c > index 4955ee942a1c..c777d563f1b5 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.c > @@ -24,14 +24,10 @@ > /* > * Render-node eventfd subscription infrastructure. > * > - * This module provides a simple event notification mechanism for render-node > - * clients using Linux eventfd objects. > + * EVENTFD is notification-only. It wakes userspace when a GPU event happens. > + * Event metadata/details are expected to be consumed separately through the > + * corresponding wait/event path. > * > - * Userspace can bind an eventfd to a userspace-defined event_id. When the > - * driver signals that event_id, all eventfds bound to it are notified. > - * > - * This mechanism is intended to support lightweight GPU event notifications > - * without polling from userspace. > */ > > #include <linux/slab.h> > @@ -42,19 +38,19 @@ > #define AMDGPU_EVENTFD_MAX_BINDS 4096 > > /** > - * amdgpu_eventfd_id_alloc - allocate an event id container > - * @event_id: userspace-defined event identifier > + * amdgpu_eventfd_id_alloc - allocate an event type container > + * @event_type: kernel-defined AMDGPU event type > * > - * Each event_id represents a notification category. Multiple eventfds can > - * be bound to the same event_id. > + * Each event type represents a notification category. Multiple eventfds > + * can be bound to the same event type. > * > * This function allocates the container which stores the list of eventfds > - * associated with that event_id. > + * associated with that event type. > * > * Return: > * Pointer to the newly allocated structure or NULL on failure. > */ > -static struct amdgpu_eventfd_id *amdgpu_eventfd_id_alloc(u32 event_id) > +static struct amdgpu_eventfd_id *amdgpu_eventfd_id_alloc(u32 event_type) > { > struct amdgpu_eventfd_id *id; > > @@ -62,43 +58,41 @@ static struct amdgpu_eventfd_id > *amdgpu_eventfd_id_alloc(u32 event_id) > if (!id) > return NULL; > > - id->event_id = event_id; > + id->event_type = event_type; > INIT_HLIST_HEAD(&id->entries); > id->n_entries = 0; > + > return id; > } > > /** > - * amdgpu_eventfd_id_get_or_create - find or create an event_id entry > + * amdgpu_eventfd_id_get_or_create - lookup or create an event type > * @mgr: eventfd manager > - * @event_id: event identifier > - * > - * This helper returns the container associated with the given event_id. > - * If it does not exist, it will create one. > + * @event_type: kernel-defined AMDGPU event type > * > - * The function is designed to be callable without holding any locks. > - * Memory allocation is done outside the xarray lock to avoid blocking > - * inside critical sections. > + * Find the container associated with the specified event type. If none > + * exists, allocate and insert a new one. > * > * Return: > - * Pointer to the event_id structure or NULL on failure. > + * Existing event type container or a newly allocated one. > */ > static struct amdgpu_eventfd_id * > -amdgpu_eventfd_id_get_or_create(struct amdgpu_eventfd_mgr *mgr, u32 event_id) > +amdgpu_eventfd_id_get_or_create(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type) > { > struct amdgpu_eventfd_id *id; > struct amdgpu_eventfd_id *new_id; > - XA_STATE(xas, &mgr->ids, event_id); > + XA_STATE(xas, &mgr->ids, event_type); > unsigned long flags; > int r; > > xa_lock_irqsave(&mgr->ids, flags); > - id = xa_load(&mgr->ids, event_id); > + id = xa_load(&mgr->ids, event_type); > xa_unlock_irqrestore(&mgr->ids, flags); > if (id) > return id; > > - new_id = amdgpu_eventfd_id_alloc(event_id); > + new_id = amdgpu_eventfd_id_alloc(event_type); > if (!new_id) > return NULL; > > @@ -140,7 +134,7 @@ amdgpu_eventfd_id_get_or_create(struct amdgpu_eventfd_mgr > *mgr, u32 event_id) > * > * event_id -> list of eventfds > * > - * The xarray provides efficient lookup of event_id containers. > + * The xarray provides efficient lookup of event type containers. > */ > void amdgpu_eventfd_mgr_init(struct amdgpu_eventfd_mgr *mgr) > { > @@ -186,23 +180,19 @@ void amdgpu_eventfd_mgr_fini(struct amdgpu_eventfd_mgr > *mgr) > } > > /** > - * amdgpu_eventfd_bind - bind eventfd to an event_id > + * amdgpu_eventfd_bind - bind an eventfd to an event type > * @mgr: eventfd manager > - * @event_id: userspace event identifier > + * @event_type: kernel-defined AMDGPU event type > * @eventfd: eventfd file descriptor > * > - * This function allows userspace to subscribe to notifications for a > - * specific event_id. > - * > - * Multiple eventfds can be bound to the same event_id. > - * > - * Duplicate bindings of the same eventfd are treated as success and do > - * not create additional entries. > + * Register an eventfd subscription for the specified event type. > + * Duplicate subscriptions are treated as success. > * > * Return: > - * 0 on success, negative error code on failure. > + * 0 on success or a negative error code. > */ > -int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, u32 event_id, int > eventfd) > +int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type, int eventfd) > { > struct amdgpu_eventfd_id *id; > struct amdgpu_eventfd_entry *e = NULL, *it; > @@ -211,7 +201,7 @@ int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int eventf > bool dup = false; > int r = 0; /* duplicate binds are treated as success */ > > - if (!mgr || !event_id || eventfd < 0) > + if (!mgr || !event_type || eventfd < 0) > return -EINVAL; > > /* > @@ -229,7 +219,7 @@ int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int eventf > return PTR_ERR(ctx); > } > > - id = amdgpu_eventfd_id_get_or_create(mgr, event_id); > + id = amdgpu_eventfd_id_get_or_create(mgr, event_type); > if (!id) { > r = -ENOMEM; > goto out_put_ctx; > @@ -259,7 +249,7 @@ int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int eventf > > /* > * Re-check duplicate under lock to close the race with another bind() > - * for the same (event_id, ctx). > + * for the same event_type and eventfd context. > */ > xa_lock_irqsave(&mgr->ids, flags); > hlist_for_each_entry(it, &id->entries, hnode) { > @@ -291,17 +281,18 @@ int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int eventf > } > > /** > - * amdgpu_eventfd_unbind - remove eventfd binding > + * amdgpu_eventfd_unbind - remove an eventfd subscription > * @mgr: eventfd manager > - * @event_id: event identifier > + * @event_type: kernel-defined AMDGPU event type > * @eventfd: eventfd file descriptor > * > - * Removes an existing binding between an event_id and an eventfd. > + * Remove an existing subscription for the specified event type. > * > * Return: > - * 0 if removed, -ENOENT if binding does not exist. > + * 0 on success or a negative error code. > */ > -int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, u32 event_id, int > eventfd) > +int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type, int eventfd) > { > struct amdgpu_eventfd_id *id; > struct amdgpu_eventfd_entry *e; > @@ -310,7 +301,7 @@ int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int even > unsigned long flags; > bool removed = false; > > - if (!mgr || !event_id || eventfd < 0) > + if (!mgr || !event_type || eventfd < 0) > return -EINVAL; > > ctx = eventfd_ctx_fdget(eventfd); > @@ -319,7 +310,7 @@ int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int even > > xa_lock_irqsave(&mgr->ids, flags); > > - id = xa_load(&mgr->ids, event_id); > + id = xa_load(&mgr->ids, event_type); > if (!id) > goto out_unlock; > > @@ -337,7 +328,7 @@ int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, > u32 event_id, int even > atomic_dec(&mgr->bind_count); > > if (!id->n_entries) { > - __xa_erase(&mgr->ids, event_id); > + __xa_erase(&mgr->ids, event_type); > kfree(id); > } > > @@ -352,22 +343,22 @@ int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr > *mgr, u32 event_id, int even > } > > /** > - * amdgpu_eventfd_signal - notify all eventfds bound to event_id > + * amdgpu_eventfd_signal - signal all eventfds bound to an event type > * @mgr: eventfd manager > - * @event_id: event identifier > + * @event_type: kernel-defined AMDGPU event type > * > - * This function is typically called from interrupt context. > + * Notify all subscribers associated with the specified event type. > * > - * All eventfds registered for the given event_id will be signaled. > - * Userspace processes waiting on those eventfds will wake up. > + * Signaling is notification-only and does not carry additional payload. > */ > -void amdgpu_eventfd_signal(struct amdgpu_eventfd_mgr *mgr, u32 event_id) > +void amdgpu_eventfd_signal(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type) > { > struct amdgpu_eventfd_id *id; > struct amdgpu_eventfd_entry *e; > unsigned long flags; > > - if (!mgr || !event_id) > + if (!mgr || !event_type) > return; > > /* > @@ -376,7 +367,7 @@ void amdgpu_eventfd_signal(struct amdgpu_eventfd_mgr > *mgr, u32 event_id) > */ > xa_lock_irqsave(&mgr->ids, flags); > > - id = xa_load(&mgr->ids, event_id); > + id = xa_load(&mgr->ids, event_type); > if (id) { > hlist_for_each_entry(e, &id->entries, hnode) > eventfd_signal(e->ctx); > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.h > b/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.h > index 248afb1f2f14..e13b86240afc 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.h > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eventfd.h > @@ -28,9 +28,9 @@ > #ifndef __AMDGPU_EVENTFD_H__ > #define __AMDGPU_EVENTFD_H__ > > +#include <linux/atomic.h> > #include <linux/eventfd.h> > #include <linux/xarray.h> > -#include <linux/atomic.h> > > struct amdgpu_eventfd_entry { > struct eventfd_ctx *ctx; > @@ -38,22 +38,25 @@ struct amdgpu_eventfd_entry { > }; > > struct amdgpu_eventfd_id { > - u32 event_id; > + u32 event_type; > struct hlist_head entries; > u32 n_entries; > }; > > struct amdgpu_eventfd_mgr { > - struct xarray ids; /* event_id -> struct amdgpu_eventfd_id* > */ > - atomic_t bind_count; /* total binds across all event_ids */ > + struct xarray ids; /* event_type -> struct amdgpu_eventfd_id* */ > + atomic_t bind_count; /* total binds across all event types */ > }; > > void amdgpu_eventfd_mgr_init(struct amdgpu_eventfd_mgr *mgr); > void amdgpu_eventfd_mgr_fini(struct amdgpu_eventfd_mgr *mgr); > > -int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, u32 event_id, int > eventfd); > -int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, u32 event_id, int > eventfd); > +int amdgpu_eventfd_bind(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type, int eventfd); > +int amdgpu_eventfd_unbind(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type, int eventfd); > > -void amdgpu_eventfd_signal(struct amdgpu_eventfd_mgr *mgr, u32 event_id); > +void amdgpu_eventfd_signal(struct amdgpu_eventfd_mgr *mgr, > + u32 event_type); > > #endif /* __AMDGPU_EVENTFD_H__ */ > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > index f5719500527f..7cc54b7e3b81 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > @@ -634,6 +634,43 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, > return 0; > } > > +int amdgpu_eventfd_ioctl(struct drm_device *dev, void *data, > + struct drm_file *file_priv) > +{ > + struct amdgpu_fpriv *fpriv = file_priv->driver_priv; > + struct drm_amdgpu_eventfd *args = data; > + > + if (!fpriv) > + return -EINVAL; > + > + /* > + * flags are reserved for future UAPI extensions and must be zero. > + */ > + if (args->flags || !args->event_type || args->eventfd < 0) > + return -EINVAL; > + > + /* > + * Queue-scoped subscriptions are enabled by the later > + * queue-reference routing patch. Until then, keep > + * queue_id zero. > + */ > + if (args->queue_id) > + return -EINVAL; > + > + switch (args->op) { > + case DRM_AMDGPU_EVENTFD_OP_BIND: > + return amdgpu_eventfd_bind(&fpriv->eventfd_mgr, > + args->event_type, > + args->eventfd); > + case DRM_AMDGPU_EVENTFD_OP_UNBIND: > + return amdgpu_eventfd_unbind(&fpriv->eventfd_mgr, > + args->event_type, > + args->eventfd); > + default: > + return -EINVAL; > + } > +} > + > /* > * Userspace get information ioctl > */ > diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h > index 9f3090db2f16..bf4fc61cf0c7 100644 > --- a/include/uapi/drm/amdgpu_drm.h > +++ b/include/uapi/drm/amdgpu_drm.h > @@ -39,6 +39,7 @@ extern "C" { > #endif > > #define DRM_AMDGPU_GEM_CREATE 0x00 > +#define DRM_AMDGPU_EVENTFD 0x1A
Move this to the end so we keep these in order. Other than that, looks good to me. Reviewed-by: Alex Deucher <[email protected]> Alex > #define DRM_AMDGPU_GEM_MMAP 0x01 > #define DRM_AMDGPU_CTX 0x02 > #define DRM_AMDGPU_BO_LIST 0x03 > @@ -79,6 +80,8 @@ extern "C" { > #define DRM_IOCTL_AMDGPU_USERQ_SIGNAL DRM_IOWR(DRM_COMMAND_BASE + > DRM_AMDGPU_USERQ_SIGNAL, struct drm_amdgpu_userq_signal) > #define DRM_IOCTL_AMDGPU_USERQ_WAIT DRM_IOWR(DRM_COMMAND_BASE + > DRM_AMDGPU_USERQ_WAIT, struct drm_amdgpu_userq_wait) > #define DRM_IOCTL_AMDGPU_GEM_LIST_HANDLES DRM_IOWR(DRM_COMMAND_BASE + > DRM_AMDGPU_GEM_LIST_HANDLES, struct drm_amdgpu_gem_list_handles) > +#define DRM_IOCTL_AMDGPU_EVENTFD \ > + DRM_IOW(DRM_COMMAND_BASE + DRM_AMDGPU_EVENTFD, struct > drm_amdgpu_eventfd) > > /** > * DOC: memory domains > @@ -204,6 +207,46 @@ union drm_amdgpu_gem_create { > struct drm_amdgpu_gem_create_out out; > }; > > +enum drm_amdgpu_event_type { > + DRM_AMDGPU_EVENT_TYPE_USERQ_EOP = 1, > + DRM_AMDGPU_EVENT_TYPE_QUEUE_RESET = 2, > + DRM_AMDGPU_EVENT_TYPE_MEMORY_EXCEPTION = 3, > + DRM_AMDGPU_EVENT_TYPE_SCRATCH = 4, > + DRM_AMDGPU_EVENT_TYPE_GPU_RESET = 5, > +}; > + > +enum drm_amdgpu_eventfd_op { > + DRM_AMDGPU_EVENTFD_OP_BIND = 0, > + DRM_AMDGPU_EVENTFD_OP_UNBIND = 1, > +}; > + > +/** > + * struct drm_amdgpu_eventfd - bind or unbind an eventfd to an AMDGPU event > + * @op: operation type, see &enum drm_amdgpu_eventfd_op > + * @event_type: kernel-defined AMDGPU event type > + * @eventfd: eventfd file descriptor > + * @queue_id: queue identifier for queue-scoped subscriptions, or 0 for > + * device/GPU-scoped subscriptions > + * @flags: must be 0 > + * > + * This ioctl lets userspace register or unregister eventfd notifications > + * for a render-node event. > + * > + * Eventfd signaling is notification-only. > + * > + * USERQ_EOP, QUEUE_RESET, and SCRATCH are queue-scoped events. > + * Userspace specifies @queue_id when registering these subscriptions. > + * > + * MEMORY_EXCEPTION is currently GPU-scoped and requires @queue_id = 0. > + */ > +struct drm_amdgpu_eventfd { > + __u32 op; > + __u32 event_type; > + __s32 eventfd; > + __u32 queue_id; > + __u32 flags; > +}; > + > /** Opcode to create new residency list. */ > #define AMDGPU_BO_LIST_OP_CREATE 0 > /** Opcode to destroy previously created residency list */ > -- > 2.34.1 >
