[Mesa-dev] [PATCH] winsys/radeon: add buffer_get_reloc_offset

[Nicolai Hähnle]

From: Nicolai Hähnle <nicolai.haehnle at amd.com>

Really fix the bug that was supposed to be fixed by commits 3e7cced4b and
a48bf02d: even when virtual addresses are used, the legacy relocation-based
method with offsets relative to the kernel's buffer object are used for
video submissions.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97969
---
 src/gallium/drivers/radeon/radeon_uvd.c       |  2 +-
 src/gallium/drivers/radeon/radeon_vce.c       |  2 +-
 src/gallium/drivers/radeon/radeon_winsys.h    | 12 ++++++++++++
 src/gallium/winsys/radeon/drm/radeon_drm_bo.c | 11 +++++++++++
 4 files changed, 25 insertions(+), 2 deletions(-)

diff --git a/src/gallium/drivers/radeon/radeon_uvd.c b/src/gallium/drivers/radeon/radeon_uvd.c
index 9c376cb..fb1491a 100644
--- a/src/gallium/drivers/radeon/radeon_uvd.c
+++ b/src/gallium/drivers/radeon/radeon_uvd.c
@@ -116,21 +116,21 @@ static void send_cmd(struct ruvd_decoder *dec, unsigned cmd,
 	reloc_idx = dec->ws->cs_add_buffer(dec->cs, buf, usage | RADEON_USAGE_SYNCHRONIZED,
 					   domain,
 					  RADEON_PRIO_UVD);
 	if (!dec->use_legacy) {
 		uint64_t addr;
 		addr = dec->ws->buffer_get_virtual_address(buf);
 		addr = addr + off;
 		set_reg(dec, RUVD_GPCOM_VCPU_DATA0, addr);
 		set_reg(dec, RUVD_GPCOM_VCPU_DATA1, addr >> 32);
 	} else {
-		off += dec->ws->buffer_get_virtual_address(buf);
+		off += dec->ws->buffer_get_reloc_offset(buf);
 		set_reg(dec, RUVD_GPCOM_VCPU_DATA0, off);
 		set_reg(dec, RUVD_GPCOM_VCPU_DATA1, reloc_idx * 4);
 	}
 	set_reg(dec, RUVD_GPCOM_VCPU_CMD, cmd << 1);
 }
 
 /* do the codec needs an IT buffer ?*/
 static bool have_it(struct ruvd_decoder *dec)
 {
 	return dec->stream_type == RUVD_CODEC_H264_PERF ||
diff --git a/src/gallium/drivers/radeon/radeon_vce.c b/src/gallium/drivers/radeon/radeon_vce.c
index 30705c1..ef93e46 100644
--- a/src/gallium/drivers/radeon/radeon_vce.c
+++ b/src/gallium/drivers/radeon/radeon_vce.c
@@ -542,15 +542,15 @@ void rvce_add_buffer(struct rvce_encoder *enc, struct pb_buffer *buf,
 
 	reloc_idx = enc->ws->cs_add_buffer(enc->cs, buf, usage | RADEON_USAGE_SYNCHRONIZED,
 					   domain, RADEON_PRIO_VCE);
 	if (enc->use_vm) {
 		uint64_t addr;
 		addr = enc->ws->buffer_get_virtual_address(buf);
 		addr = addr + offset;
 		RVCE_CS(addr >> 32);
 		RVCE_CS(addr);
 	} else {
-		offset += enc->ws->buffer_get_virtual_address(buf);
+		offset += enc->ws->buffer_get_reloc_offset(buf);
 		RVCE_CS(reloc_idx * 4);
 		RVCE_CS(offset);
 	}
 }
diff --git a/src/gallium/drivers/radeon/radeon_winsys.h b/src/gallium/drivers/radeon/radeon_winsys.h
index e8fc6d6..7146737 100644
--- a/src/gallium/drivers/radeon/radeon_winsys.h
+++ b/src/gallium/drivers/radeon/radeon_winsys.h
@@ -508,20 +508,32 @@ struct radeon_winsys {
      *
      * When virtual memory is not in use, this is the offset relative to the
      * relocation base (non-zero for sub-allocated buffers).
      *
      * \param buf       A winsys buffer object
      * \return          virtual address
      */
     uint64_t (*buffer_get_virtual_address)(struct pb_buffer *buf);
 
     /**
+     * Return the offset of this buffer relative to the relocation base.
+     * This is only non-zero for sub-allocated buffers.
+     *
+     * This is only supported in the radeon winsys, since amdgpu uses virtual
+     * addresses in submissions even for the video engines.
+     *
+     * \param buf      A winsys buffer object
+     * \return         the offset for relocations
+     */
+    unsigned (*buffer_get_reloc_offset)(struct pb_buffer *buf);
+
+    /**
      * Query the initial placement of the buffer from the kernel driver.
      */
     enum radeon_bo_domain (*buffer_get_initial_domain)(struct pb_buffer *buf);
 
     /**************************************************************************
      * Command submission.
      *
      * Each pipe context should create its own command stream and submit
      * commands independently of other contexts.
      *************************************************************************/
diff --git a/src/gallium/winsys/radeon/drm/radeon_drm_bo.c b/src/gallium/winsys/radeon/drm/radeon_drm_bo.c
index 5818006..a15d559 100644
--- a/src/gallium/winsys/radeon/drm/radeon_drm_bo.c
+++ b/src/gallium/winsys/radeon/drm/radeon_drm_bo.c
@@ -1332,25 +1332,36 @@ static bool radeon_winsys_bo_get_handle(struct pb_buffer *buffer,
 static bool radeon_winsys_bo_is_user_ptr(struct pb_buffer *buf)
 {
    return ((struct radeon_bo*)buf)->user_ptr != NULL;
 }
 
 static uint64_t radeon_winsys_bo_va(struct pb_buffer *buf)
 {
     return ((struct radeon_bo*)buf)->va;
 }
 
+static unsigned radeon_winsys_bo_get_reloc_offset(struct pb_buffer *buf)
+{
+    struct radeon_bo *bo = radeon_bo(buf);
+
+    if (bo->handle)
+        return 0;
+
+    return bo->va - bo->u.slab.real->va;
+}
+
 void radeon_drm_bo_init_functions(struct radeon_drm_winsys *ws)
 {
     ws->base.buffer_set_metadata = radeon_bo_set_metadata;
     ws->base.buffer_get_metadata = radeon_bo_get_metadata;
     ws->base.buffer_map = radeon_bo_map;
     ws->base.buffer_unmap = radeon_bo_unmap;
     ws->base.buffer_wait = radeon_bo_wait;
     ws->base.buffer_create = radeon_winsys_bo_create;
     ws->base.buffer_from_handle = radeon_winsys_bo_from_handle;
     ws->base.buffer_from_ptr = radeon_winsys_bo_from_ptr;
     ws->base.buffer_is_user_ptr = radeon_winsys_bo_is_user_ptr;
     ws->base.buffer_get_handle = radeon_winsys_bo_get_handle;
     ws->base.buffer_get_virtual_address = radeon_winsys_bo_va;
+    ws->base.buffer_get_reloc_offset = radeon_winsys_bo_get_reloc_offset;
     ws->base.buffer_get_initial_domain = radeon_bo_get_initial_domain;
 }
-- 
2.7.4