r600g: avoid using pb* helper we are loosing previous cpu cycle with it
r600g is up to a point where all small CPU cycle matter and pb* turn
high on profile. It's mostly because pb try to be generic and thus
trigger unecessary check for r600g driver. To avoid having too much
abstraction & too much depth in the call embedded everythings into
r600_bo. Make code simpler & faster. The performance win highly depend
on the CPU & application considered being more important on slower CPU
and marginal/unoticeable on faster one.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
diff --git a/src/gallium/winsys/r600/drm/r600_priv.h b/src/gallium/winsys/r600/drm/r600_priv.h
index 193af98..056d025 100644
--- a/src/gallium/winsys/r600/drm/r600_priv.h
+++ b/src/gallium/winsys/r600/drm/r600_priv.h
@@ -30,24 +30,24 @@
#include <stdint.h>
#include <stdlib.h>
#include <assert.h>
-#include <pipebuffer/pb_bufmgr.h>
-#include "util/u_double_list.h"
+#include <util/u_double_list.h>
+#include <util/u_inlines.h>
+#include <os/os_thread.h>
#include "r600.h"
+struct r600_bomgr;
+
struct radeon {
int fd;
int refcount;
unsigned device;
unsigned family;
enum chip_class chip_class;
- struct pb_manager *kman; /* kernel bo manager */
- struct pb_manager *cman; /* cached bo manager */
- struct r600_tiling_info tiling_info;
+ struct r600_tiling_info tiling_info;
+ struct r600_bomgr *bomgr;
+ unsigned *cfence;
};
-struct radeon *r600_new(int fd, unsigned device);
-void r600_delete(struct radeon *r600);
-
struct r600_reg {
unsigned opcode;
unsigned offset_base;
@@ -75,25 +75,49 @@
struct r600_bo {
struct pipe_reference reference;
- struct pb_buffer *pb;
unsigned size;
unsigned tiling_flags;
unsigned kernel_pitch;
unsigned domains;
+ struct radeon_bo *bo;
+ unsigned fence;
+ /* manager data */
+ struct list_head list;
+ unsigned manager_id;
+ unsigned alignment;
+ unsigned offset;
+ int64_t start;
+ int64_t end;
};
+struct r600_bomgr {
+ struct radeon *radeon;
+ unsigned usecs;
+ pipe_mutex mutex;
+ struct list_head delayed;
+ unsigned num_delayed;
+};
-/* radeon_pciid.c */
+/*
+ * r600_drm.c
+ */
+struct radeon *r600_new(int fd, unsigned device);
+void r600_delete(struct radeon *r600);
+
+/*
+ * radeon_pciid.c
+ */
unsigned radeon_family_from_device(unsigned device);
-/* radeon_bo.c */
+/*
+ * radeon_bo.c
+ */
struct radeon_bo *radeon_bo(struct radeon *radeon, unsigned handle,
unsigned size, unsigned alignment);
void radeon_bo_reference(struct radeon *radeon, struct radeon_bo **dst,
struct radeon_bo *src);
int radeon_bo_wait(struct radeon *radeon, struct radeon_bo *bo);
int radeon_bo_busy(struct radeon *radeon, struct radeon_bo *bo, uint32_t *domain);
-void radeon_bo_pbmgr_flush_maps(struct pb_manager *_mgr);
int radeon_bo_fencelist(struct radeon *radeon, struct radeon_bo **bolist, uint32_t num_bo);
int radeon_bo_get_tiling_flags(struct radeon *radeon,
struct radeon_bo *bo,
@@ -103,13 +127,9 @@
struct radeon_bo *bo,
uint32_t *name);
-/* radeon_bo_pb.c */
-struct radeon_bo *radeon_bo_pb_get_bo(struct pb_buffer *_buf);
-struct pb_manager *radeon_bo_pbmgr_create(struct radeon *radeon);
-struct pb_buffer *radeon_bo_pb_create_buffer_from_handle(struct pb_manager *_mgr,
- uint32_t handle);
-
-/* r600_hw_context.c */
+/*
+ * r600_hw_context.c
+ */
int r600_context_init_fence(struct r600_context *ctx);
void r600_context_bo_reloc(struct r600_context *ctx, u32 *pm4, struct r600_bo *rbo);
void r600_context_bo_flush(struct r600_context *ctx, unsigned flush_flags,
@@ -117,14 +137,27 @@
struct r600_bo *r600_context_reg_bo(struct r600_context *ctx, unsigned offset);
int r600_context_add_block(struct r600_context *ctx, const struct r600_reg *reg, unsigned nreg);
-/* r600_bo.c */
-unsigned r600_bo_get_handle(struct r600_bo *bo);
-unsigned r600_bo_get_size(struct r600_bo *bo);
-static INLINE struct radeon_bo *r600_bo_get_bo(struct r600_bo *bo)
-{
- return radeon_bo_pb_get_bo(bo->pb);
-}
+/*
+ * r600_bo.c
+ */
+void r600_bo_destroy(struct radeon *radeon, struct r600_bo *bo);
+/*
+ * r600_bomgr.c
+ */
+struct r600_bomgr *r600_bomgr_create(struct radeon *radeon, unsigned usecs);
+void r600_bomgr_destroy(struct r600_bomgr *mgr);
+bool r600_bomgr_bo_destroy(struct r600_bomgr *mgr, struct r600_bo *bo);
+void r600_bomgr_bo_init(struct r600_bomgr *mgr, struct r600_bo *bo);
+struct r600_bo *r600_bomgr_bo_create(struct r600_bomgr *mgr,
+ unsigned size,
+ unsigned alignment,
+ unsigned cfence);
+
+
+/*
+ * helpers
+ */
#define CTX_RANGE_ID(ctx, offset) (((offset) >> (ctx)->hash_shift) & 255)
#define CTX_BLOCK_ID(ctx, offset) ((offset) & ((1 << (ctx)->hash_shift) - 1))
@@ -172,6 +205,9 @@
LIST_DELINIT(&block->list);
}
+/*
+ * radeon_bo.c
+ */
static inline int radeon_bo_map(struct radeon *radeon, struct radeon_bo *bo)
{
bo->map_count++;
@@ -184,4 +220,35 @@
assert(bo->map_count >= 0);
}
+/*
+ * r600_bo
+ */
+static inline struct radeon_bo *r600_bo_get_bo(struct r600_bo *bo)
+{
+ return bo->bo;
+}
+
+static unsigned inline r600_bo_get_handle(struct r600_bo *bo)
+{
+ return bo->bo->handle;
+}
+
+static unsigned inline r600_bo_get_size(struct r600_bo *bo)
+{
+ return bo->size;
+}
+
+/*
+ * fence
+ */
+static inline bool fence_is_after(unsigned fence, unsigned ofence)
+{
+ /* handle wrap around */
+ if (fence < 0x80000000 && ofence > 0x80000000)
+ return TRUE;
+ if (fence > ofence)
+ return TRUE;
+ return FALSE;
+}
+
#endif