[PATCH v4 09/17] drm/i915/vm_bind: Add out fence support
Matthew Auld
matthew.auld at intel.com
Wed Oct 19 14:24:54 UTC 2022
On 19/10/2022 03:43, Niranjana Vishwanathapura wrote:
> On Tue, Oct 18, 2022 at 04:28:07PM +0100, Matthew Auld wrote:
>> On 18/10/2022 08:16, Niranjana Vishwanathapura wrote:
>>> Add support for handling out fence for vm_bind call.
>>>
>>> v2: Reset vma->vm_bind_fence.syncobj to NULL at the end
>>> of vm_bind call.
>>> v3: Remove vm_unbind out fence uapi which is not supported yet.
>>>
>>> Signed-off-by: Niranjana Vishwanathapura
>>> <niranjana.vishwanathapura at intel.com>
>>> Signed-off-by: Andi Shyti <andi.shyti at linux.intel.com>
>>> ---
>>> drivers/gpu/drm/i915/gem/i915_gem_vm_bind.h | 4 +
>>> .../drm/i915/gem/i915_gem_vm_bind_object.c | 82 +++++++++++++++++++
>>> drivers/gpu/drm/i915/i915_vma.c | 7 +-
>>> drivers/gpu/drm/i915/i915_vma_types.h | 7 ++
>>> include/uapi/drm/i915_drm.h | 49 ++++++++++-
>>> 5 files changed, 146 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_vm_bind.h
>>> b/drivers/gpu/drm/i915/gem/i915_gem_vm_bind.h
>>> index 36262a6357b5..b70e900e35ab 100644
>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_vm_bind.h
>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_vm_bind.h
>>> @@ -8,6 +8,7 @@
>>> #include <linux/types.h>
>>> +struct dma_fence;
>>> struct drm_device;
>>> struct drm_file;
>>> struct i915_address_space;
>>> @@ -23,4 +24,7 @@ int i915_gem_vm_unbind_ioctl(struct drm_device
>>> *dev, void *data,
>>> void i915_gem_vm_unbind_all(struct i915_address_space *vm);
>>> +void i915_vm_bind_signal_fence(struct i915_vma *vma,
>>> + struct dma_fence * const fence);
>>> +
>>> #endif /* __I915_GEM_VM_BIND_H */
>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_vm_bind_object.c
>>> b/drivers/gpu/drm/i915/gem/i915_gem_vm_bind_object.c
>>> index 3ea3cb3ed97e..63889ba00183 100644
>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_vm_bind_object.c
>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_vm_bind_object.c
>>> @@ -7,6 +7,8 @@
>>> #include <linux/interval_tree_generic.h>
>>> +#include <drm/drm_syncobj.h>
>>> +
>>> #include "gem/i915_gem_context.h"
>>> #include "gem/i915_gem_vm_bind.h"
>>> @@ -100,6 +102,76 @@ static void i915_gem_vm_bind_remove(struct
>>> i915_vma *vma, bool release_obj)
>>> i915_gem_object_put(vma->obj);
>>> }
>>> +static int i915_vm_bind_add_fence(struct drm_file *file, struct
>>> i915_vma *vma,
>>> + u32 handle, u64 point)
>>> +{
>>> + struct drm_syncobj *syncobj;
>>> +
>>> + syncobj = drm_syncobj_find(file, handle);
>>> + if (!syncobj) {
>>> + DRM_DEBUG("Invalid syncobj handle provided\n");
>>> + return -ENOENT;
>>> + }
>>> +
>>> + /*
>>> + * For timeline syncobjs we need to preallocate chains for
>>> + * later signaling.
>>> + */
>>> + if (point) {
>>> + vma->vm_bind_fence.chain_fence = dma_fence_chain_alloc();
>>> + if (!vma->vm_bind_fence.chain_fence) {
>>> + drm_syncobj_put(syncobj);
>>> + return -ENOMEM;
>>> + }
>>> + } else {
>>> + vma->vm_bind_fence.chain_fence = NULL;
>>> + }
>>> + vma->vm_bind_fence.syncobj = syncobj;
>>> + vma->vm_bind_fence.value = point;
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static void i915_vm_bind_put_fence(struct i915_vma *vma)
>>> +{
>>> + if (!vma->vm_bind_fence.syncobj)
>>> + return;
>>> +
>>> + drm_syncobj_put(vma->vm_bind_fence.syncobj);
>>> + dma_fence_chain_free(vma->vm_bind_fence.chain_fence);
>>> + vma->vm_bind_fence.syncobj = NULL;
>>> +}
>>> +
>>> +/**
>>> + * i915_vm_bind_signal_fence() - Add fence to vm_bind syncobj
>>> + * @vma: vma mapping requiring signaling
>>> + * @fence: fence to be added
>>> + *
>>> + * Associate specified @fence with the @vma's syncobj to be
>>> + * signaled after the @fence work completes.
>>> + */
>>> +void i915_vm_bind_signal_fence(struct i915_vma *vma,
>>> + struct dma_fence * const fence)
>>> +{
>>> + struct drm_syncobj *syncobj = vma->vm_bind_fence.syncobj;
>>> +
>>> + if (!syncobj)
>>> + return;
>>> +
>>> + if (vma->vm_bind_fence.chain_fence) {
>>> + drm_syncobj_add_point(syncobj,
>>> + vma->vm_bind_fence.chain_fence,
>>> + fence, vma->vm_bind_fence.value);
>>> + /*
>>> + * The chain's ownership is transferred to the
>>> + * timeline.
>>> + */
>>> + vma->vm_bind_fence.chain_fence = NULL;
>>> + } else {
>>> + drm_syncobj_replace_fence(syncobj, fence);
>>> + }
>>> +}
>>> +
>>> static int i915_gem_vm_unbind_vma(struct i915_address_space *vm,
>>> struct drm_i915_gem_vm_unbind *va)
>>> {
>>> @@ -237,6 +309,13 @@ static int i915_gem_vm_bind_obj(struct
>>> i915_address_space *vm,
>>> goto unlock_vm;
>>> }
>>> + if (va->fence.flags & I915_TIMELINE_FENCE_SIGNAL) {
>>> + ret = i915_vm_bind_add_fence(file, vma, va->fence.handle,
>>> + va->fence.value);
>>> + if (ret)
>>> + goto put_vma;
>>> + }
>>> +
>>> pin_flags = va->start | PIN_OFFSET_FIXED | PIN_USER | PIN_VALIDATE;
>>> for_i915_gem_ww(&ww, ret, true) {
>>> @@ -258,6 +337,9 @@ static int i915_gem_vm_bind_obj(struct
>>> i915_address_space *vm,
>>> i915_gem_object_get(vma->obj);
>>> }
>>> + if (va->fence.flags & I915_TIMELINE_FENCE_SIGNAL)
>>> + i915_vm_bind_put_fence(vma);
>>> +put_vma:
>>> if (ret)
>>> i915_vma_destroy(vma);
>>> unlock_vm:
>>> diff --git a/drivers/gpu/drm/i915/i915_vma.c
>>> b/drivers/gpu/drm/i915/i915_vma.c
>>> index 04abdb92c2b2..eaa13e9ba966 100644
>>> --- a/drivers/gpu/drm/i915/i915_vma.c
>>> +++ b/drivers/gpu/drm/i915/i915_vma.c
>>> @@ -29,6 +29,7 @@
>>> #include "display/intel_frontbuffer.h"
>>> #include "gem/i915_gem_lmem.h"
>>> #include "gem/i915_gem_tiling.h"
>>> +#include "gem/i915_gem_vm_bind.h"
>>> #include "gt/intel_engine.h"
>>> #include "gt/intel_engine_heartbeat.h"
>>> #include "gt/intel_gt.h"
>>> @@ -1567,8 +1568,12 @@ int i915_vma_pin_ww(struct i915_vma *vma,
>>> struct i915_gem_ww_ctx *ww,
>>> err_vma_res:
>>> i915_vma_resource_free(vma_res);
>>> err_fence:
>>> - if (work)
>>> + if (work) {
>>> + if (i915_vma_is_persistent(vma))
>>> + i915_vm_bind_signal_fence(vma, &work->base.dma);
>>> +
>>> dma_fence_work_commit_imm(&work->base);
>>> + }
>>> err_rpm:
>>> if (wakeref)
>>> intel_runtime_pm_put(&vma->vm->i915->runtime_pm, wakeref);
>>> diff --git a/drivers/gpu/drm/i915/i915_vma_types.h
>>> b/drivers/gpu/drm/i915/i915_vma_types.h
>>> index d32c72e8d242..2c740500ac1b 100644
>>> --- a/drivers/gpu/drm/i915/i915_vma_types.h
>>> +++ b/drivers/gpu/drm/i915/i915_vma_types.h
>>> @@ -308,6 +308,13 @@ struct i915_vma {
>>> /* @vm_rebind_link: link to vm_rebind_list and protected by
>>> vm_rebind_lock */
>>> struct list_head vm_rebind_link; /* Link in vm_rebind_list */
>>> + /** Timeline fence for vm_bind completion notification */
>>> + struct {
>>> + struct dma_fence_chain *chain_fence;
>>> + struct drm_syncobj *syncobj;
>>> + u64 value;
>>> + } vm_bind_fence;
>>> +
>>> /** Interval tree structures for persistent vma */
>>> /** @rb: node for the interval tree of vm for persistent vmas */
>>> diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
>>> index 4383ed85f10a..87f5c2a470f5 100644
>>> --- a/include/uapi/drm/i915_drm.h
>>> +++ b/include/uapi/drm/i915_drm.h
>>> @@ -1527,6 +1527,41 @@ struct drm_i915_gem_execbuffer2 {
>>> #define i915_execbuffer2_get_context_id(eb2) \
>>> ((eb2).rsvd1 & I915_EXEC_CONTEXT_ID_MASK)
>>> +/**
>>> + * struct drm_i915_gem_timeline_fence - An input or output timeline
>>> fence.
>>> + *
>>> + * The operation will wait for input fence to signal.
>>> + *
>>> + * The returned output fence will be signaled after the completion
>>> of the
>>> + * operation.
>>> + */
>>> +struct drm_i915_gem_timeline_fence {
>>> + /** @handle: User's handle for a drm_syncobj to wait on or
>>> signal. */
>>> + __u32 handle;
>>> +
>>> + /**
>>> + * @flags: Supported flags are:
>>> + *
>>> + * I915_TIMELINE_FENCE_WAIT:
>>> + * Wait for the input fence before the operation.
>>> + *
>>> + * I915_TIMELINE_FENCE_SIGNAL:
>>> + * Return operation completion fence as output.
>>> + */
>>> + __u32 flags;
>>> +#define I915_TIMELINE_FENCE_WAIT (1 << 0)
>>> +#define I915_TIMELINE_FENCE_SIGNAL (1 << 1)
>>
>> So this is the out-fence for bind completion, which then gets fed into
>> execbuf? Is the idea here to always get an out fence, and feed that
>> into execbuf to correctly order the bind(s) vs submit? i.e that's left
>> to userspace. IIRC we now do await_bind() in execbuf3 which I guess
>> forces the synchronisation, so just wondering if we need that, or
>> what's the story here with the out-fence for vm_bind?
>>
>
> Yah, vm_bind and execbuf3 are independent paths and user has to
> resolve any dependencies.
> I will be adding support for synchronous bind if vm_bind out fence is
> not requested (more below).
> Execbuf3 path will only await_bind() for those mappings which it
> started the binding for (the evicted/invalidated mappings which it
> scooped up). It won't await_bind() for mappings bound through vm_bind
> ioctl call.
Ah, got it. Thanks.
>
>>> +#define __I915_TIMELINE_FENCE_UNKNOWN_FLAGS
>>> (-(I915_TIMELINE_FENCE_SIGNAL << 1))
>>> +
>>> + /**
>>> + * @value: A point in the timeline.
>>> + * Value must be 0 for a binary drm_syncobj. A Value of 0 for a
>>> + * timeline drm_syncobj is invalid as it turns a drm_syncobj into a
>>> + * binary one.
>>> + */
>>> + __u64 value;
>>> +};
>>> +
>>> struct drm_i915_gem_pin {
>>> /** Handle of the buffer to be pinned. */
>>> __u32 handle;
>>> @@ -3826,8 +3861,18 @@ struct drm_i915_gem_vm_bind {
>>> */
>>> __u64 flags;
>>> - /** @rsvd: Reserved, MBZ */
>>> - __u64 rsvd[2];
>>> + /**
>>> + * @fence: Timeline fence for bind completion signaling.
>>> + *
>>> + * Timeline fence is of format struct drm_i915_gem_timeline_fence.
>>> + *
>>> + * It is an out fence, hence using I915_TIMELINE_FENCE_WAIT flag
>>> + * is invalid, and an error will be returned.
>>
>> Where we do reject that? Maybe I'm blind...
>>
>
> No you are not :)...will add the check in vm_bind call.
>
>>> + *
>>> + * If I915_TIMELINE_FENCE_SIGNAL flag is not set, then out fence
>>> + * is not requested and binding is completed synchronously.
>>
>> "completed synchronously", where do we do that currently?
>>
>
> Hmm...looks like are are not. Will add the wait call in vm_bind path.
>
> Niranjana
>
>>> + */
>>> + struct drm_i915_gem_timeline_fence fence;
>>> /**
>>> * @extensions: Zero-terminated chain of extensions.
More information about the dri-devel
mailing list