[PATCH 40/74] i915/HACK: Add support for fence objects
Chris Wilson
chris at chris-wilson.co.uk
Sun Jul 16 19:00:41 UTC 2017
From: Jason Ekstrand <jason.ekstrand at intel.com>
---
drivers/gpu/drm/i915/i915_drv.c | 3 +-
drivers/gpu/drm/i915/i915_gem_execbuffer.c | 148 ++++++++++++++++++++++++++++-
include/uapi/drm/i915_drm.h | 32 ++++++-
3 files changed, 175 insertions(+), 8 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 6cb31e55a4dd..e4a5e5a6c65c 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -388,6 +388,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
case I915_PARAM_HAS_EXEC_FENCE:
case I915_PARAM_HAS_EXEC_CAPTURE:
case I915_PARAM_HAS_EXEC_BATCH_FIRST:
+ case I915_PARAM_HAS_EXEC_FENCE_ARRAY:
/* For the time being all of these are always true;
* if some supported hardware does not have one of these
* features this value needs to be provided from
@@ -2738,7 +2739,7 @@ static struct drm_driver driver = {
*/
.driver_features =
DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM | DRIVER_PRIME |
- DRIVER_RENDER | DRIVER_MODESET | DRIVER_ATOMIC,
+ DRIVER_RENDER | DRIVER_MODESET | DRIVER_ATOMIC | DRIVER_SYNCOBJ,
.release = i915_driver_release,
.open = i915_driver_open,
.lastclose = i915_driver_lastclose,
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 900a333c9881..1df4f809cb10 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -33,6 +33,7 @@
#include <drm/drmP.h>
#include <drm/i915_drm.h>
+#include <drm/drm_syncobj.h>
#include "i915_drv.h"
#include "i915_gem_clflush.h"
@@ -1809,8 +1810,10 @@ static bool i915_gem_check_execbuffer(struct drm_i915_gem_execbuffer2 *exec)
return false;
/* Kernel clipping was a DRI1 misfeature */
- if (exec->num_cliprects || exec->cliprects_ptr)
- return false;
+ if (!(exec->flags & I915_EXEC_FENCE_ARRAY)) {
+ if (exec->num_cliprects || exec->cliprects_ptr)
+ return false;
+ }
if (exec->DR4 == 0xffffffff) {
DRM_DEBUG("UXA submitting garbage DR4, fixing up\n");
@@ -2038,11 +2041,129 @@ eb_select_engine(struct drm_i915_private *dev_priv,
return engine;
}
+static void __free_syncobj_array(struct drm_syncobj **syncobjs, unsigned int n)
+{
+ while (n--)
+ drm_syncobj_put(ptr_mask_bits(syncobjs[n], 3));
+
+ kvfree(syncobjs);
+}
+
+static struct drm_syncobj **
+get_syncobj_array(struct drm_i915_gem_execbuffer2 *args, struct drm_file *file)
+{
+ const unsigned int nfences = args->num_cliprects;
+ union {
+ struct drm_i915_gem_exec_fence fence;
+ u64 x;
+ } __user *user;
+ struct drm_syncobj **syncobjs;
+ unsigned int n;
+ int err;
+
+ BUILD_BUG_ON(sizeof(user->fence) != sizeof(user->x));
+
+ if (!(args->flags & I915_EXEC_FENCE_ARRAY))
+ return NULL;
+
+ if (nfences > SIZE_MAX / sizeof(*syncobjs))
+ return ERR_PTR(-EINVAL);
+
+ user = u64_to_user_ptr(args->cliprects_ptr);
+ if (!access_ok(VERIFY_READ, user, nfences * 2 * sizeof(u32)))
+ return ERR_PTR(-EFAULT);
+
+ syncobjs = kvmalloc_array(nfences, sizeof(*syncobjs), GFP_TEMPORARY);
+ if (!syncobjs)
+ return ERR_PTR(-ENOMEM);
+
+ user_access_begin();
+ for (n = 0; n < nfences; n++) {
+ union {
+ struct drm_i915_gem_exec_fence fence;
+ u64 x;
+ } u;
+ struct drm_syncobj *sync;
+
+ unsafe_get_user(u.x, &user->x, err);
+ user++;
+
+ sync = drm_syncobj_find(file, u.fence.handle);
+ if (!sync) {
+ err = -ENOENT;
+ goto err;
+ }
+ syncobjs[n] = ptr_pack_bits(sync, u.fence.flags, 3);
+ }
+ user_access_end();
+
+ return syncobjs;
+
+err:
+ user_access_end();
+ __free_syncobj_array(syncobjs, n);
+ return ERR_PTR(err);
+}
+
+static void put_syncobj_array(struct drm_i915_gem_execbuffer2 *args,
+ struct drm_syncobj **syncobjs)
+{
+ if (!syncobjs)
+ return;
+
+ __free_syncobj_array(syncobjs, args->num_cliprects);
+}
+
+static int await_syncobj_array(struct i915_execbuffer *eb,
+ struct drm_syncobj **syncobjs)
+{
+ const unsigned int nfences = eb->args->num_cliprects;
+ unsigned int n;
+ int err;
+
+ for (n = 0; n < nfences; n++) {
+ struct drm_syncobj *sync;
+ unsigned int flags;
+
+ sync = ptr_unpack_bits(syncobjs[n], &flags, 3);
+ if (!(flags & I915_EXEC_FENCE_WAIT))
+ continue;
+
+ if (!sync->fence)
+ return -EINVAL;
+
+ err = i915_gem_request_await_dma_fence(eb->request,
+ sync->fence);
+ if (err < 0)
+ return err;
+ }
+
+ return 0;
+}
+
+static void signal_syncobj_array(struct i915_execbuffer *eb,
+ struct drm_syncobj **syncobjs)
+{
+ const unsigned int nfences = eb->args->num_cliprects;
+ struct dma_fence * const fence = &eb->request->fence;
+ unsigned int n;
+
+ for (n = 0; n < nfences; n++) {
+ struct drm_syncobj *sync;
+ unsigned int flags;
+
+ sync = ptr_unpack_bits(syncobjs[n], &flags, 3);
+ if (flags & I915_EXEC_FENCE_SIGNAL)
+ drm_syncobj_replace_fence(sync, fence);
+ }
+}
+
static int
i915_gem_do_execbuffer(struct drm_device *dev,
struct drm_file *file,
struct drm_i915_gem_execbuffer2 *args,
- struct drm_i915_gem_exec_object2 *exec)
+ struct drm_i915_gem_exec_object2 *exec,
+ struct drm_syncobj **syncobjs)
{
struct i915_execbuffer eb;
struct dma_fence *in_fence = NULL;
@@ -2233,6 +2354,12 @@ i915_gem_do_execbuffer(struct drm_device *dev,
goto err_request;
}
+ if (syncobjs) {
+ err = await_syncobj_array(&eb, syncobjs);
+ if (err)
+ goto err_request;
+ }
+
if (out_fence_fd != -1) {
out_fence = sync_file_create(&eb.request->fence);
if (!out_fence) {
@@ -2256,6 +2383,9 @@ i915_gem_do_execbuffer(struct drm_device *dev,
__i915_add_request(eb.request, err == 0);
add_to_client(eb.request, file);
+ if (err == 0 && syncobjs)
+ signal_syncobj_array(&eb, syncobjs);
+
if (out_fence) {
if (err == 0) {
fd_install(out_fence_fd, out_fence->file);
@@ -2359,7 +2489,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
exec2_list[i].flags = 0;
}
- err = i915_gem_do_execbuffer(dev, file, &exec2, exec2_list);
+ err = i915_gem_do_execbuffer(dev, file, &exec2, exec2_list, NULL);
if (exec2.flags & __EXEC_HAS_RELOC) {
struct drm_i915_gem_exec_object __user *user_exec_list =
u64_to_user_ptr(args->buffers_ptr);
@@ -2393,6 +2523,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
sizeof(unsigned int));
struct drm_i915_gem_execbuffer2 *args = data;
struct drm_i915_gem_exec_object2 *exec2_list;
+ struct drm_syncobj **syncobjs;
int err;
if (args->buffer_count < 1 || args->buffer_count > SIZE_MAX / sz - 1) {
@@ -2419,7 +2550,13 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
return -EFAULT;
}
- err = i915_gem_do_execbuffer(dev, file, args, exec2_list);
+ syncobjs = get_syncobj_array(args, file);
+ if (IS_ERR(syncobjs)) {
+ kvfree(exec2_list);
+ return PTR_ERR(syncobjs);
+ }
+
+ err = i915_gem_do_execbuffer(dev, file, args, exec2_list, syncobjs);
/*
* Now that we have begun execution of the batchbuffer, we ignore
@@ -2449,6 +2586,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
}
args->flags &= ~__I915_EXEC_UNKNOWN_FLAGS;
+ put_syncobj_array(args, syncobjs);
kvfree(exec2_list);
return err;
}
diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
index 5eb9f315f513..bccc9787f1ef 100644
--- a/include/uapi/drm/i915_drm.h
+++ b/include/uapi/drm/i915_drm.h
@@ -434,6 +434,11 @@ typedef struct drm_i915_irq_wait {
*/
#define I915_PARAM_HAS_EXEC_BATCH_FIRST 48
+/* Query whether DRM_I915_GEM_EXECBUFFER2 supports supplying an array of
+ * drm_i915_gem_exec_fence structures. See I915_EXEC_FENCE_ARRAY.
+ */
+#define I915_PARAM_HAS_EXEC_FENCE_ARRAY 49
+
typedef struct drm_i915_getparam {
__s32 param;
/*
@@ -815,6 +820,19 @@ struct drm_i915_gem_exec_object2 {
__u64 rsvd2;
};
+struct drm_i915_gem_exec_fence {
+ /**
+ * User's handle for a dma-fence to wait on or signal.
+ */
+ __u32 handle;
+
+#define I915_EXEC_FENCE_WAIT (1<<0)
+#define I915_EXEC_FENCE_SIGNAL (1<<1)
+#define I915_EXEC_FENCE_UNSIGNAL (1<<2)
+#define I915_EXEC_FENCE_FD (1<<3)
+ __u32 flags;
+};
+
struct drm_i915_gem_execbuffer2 {
/**
* List of gem_exec_object2 structs
@@ -829,7 +847,10 @@ struct drm_i915_gem_execbuffer2 {
__u32 DR1;
__u32 DR4;
__u32 num_cliprects;
- /** This is a struct drm_clip_rect *cliprects */
+ /** This is a struct drm_clip_rect *cliprects if I915_EXEC_FENCE_ARRAY
+ * is not set. If I915_EXEC_FENCE_ARRAY is set, then this is a
+ * struct drm_i915_gem_exec_fence *fences.
+ */
__u64 cliprects_ptr;
#define I915_EXEC_RING_MASK (7<<0)
#define I915_EXEC_DEFAULT (0<<0)
@@ -930,7 +951,14 @@ struct drm_i915_gem_execbuffer2 {
* element).
*/
#define I915_EXEC_BATCH_FIRST (1<<18)
-#define __I915_EXEC_UNKNOWN_FLAGS (-(I915_EXEC_BATCH_FIRST<<1))
+
+/* Setting I915_FENCE_ARRAY implies that num_cliprects and cliprects_ptr
+ * define an array of i915_gem_exec_fence structures which specify a set of
+ * dma fences to wait upon or signal.
+ */
+#define I915_EXEC_FENCE_ARRAY (1<<19)
+
+#define __I915_EXEC_UNKNOWN_FLAGS (-(I915_EXEC_FENCE_ARRAY<<1))
#define I915_EXEC_CONTEXT_ID_MASK (0xffffffff)
#define i915_execbuffer2_set_context_id(eb2, context) \
--
2.13.2
More information about the Intel-gfx-trybot
mailing list