On the blitter (and in test code), we see long sequences of repeated
commands, e.g. XY_PIXEL_BLT, XY_SCANLINE_BLT, or XY_SRC_COPY. For these,
we can skip the hashtable lookup by remembering the previous command
descriptor and doing a straightforward compare of the command header.
The corollary is that we need to do one extra comparison before lookup
up new commands.
v2: Less magic mask (ok, it is still magic, but now you cannot see!)
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-36-chris@chris-wilson.co.uk
#define STD_3D_OPCODE_SHIFT (32 - 16)
#define STD_2D_OPCODE_SHIFT (32 - 10)
#define STD_MFX_OPCODE_SHIFT (32 - 16)
#define STD_3D_OPCODE_SHIFT (32 - 16)
#define STD_2D_OPCODE_SHIFT (32 - 10)
#define STD_MFX_OPCODE_SHIFT (32 - 16)
+#define MIN_OPCODE_SHIFT 16
#define CMD(op, opm, f, lm, fl, ...) \
{ \
#define CMD(op, opm, f, lm, fl, ...) \
{ \
CMD( MI_LOAD_SCAN_LINES_EXCL, SMI, !F, 0x3F, R ),
};
CMD( MI_LOAD_SCAN_LINES_EXCL, SMI, !F, 0x3F, R ),
};
+static const struct drm_i915_cmd_descriptor noop_desc =
+ CMD(MI_NOOP, SMI, F, 1, S);
+
#undef CMD
#undef SMI
#undef S3D
#undef CMD
#undef SMI
#undef S3D
static const struct drm_i915_cmd_descriptor*
find_cmd(struct intel_engine_cs *engine,
u32 cmd_header,
static const struct drm_i915_cmd_descriptor*
find_cmd(struct intel_engine_cs *engine,
u32 cmd_header,
+ const struct drm_i915_cmd_descriptor *desc,
struct drm_i915_cmd_descriptor *default_desc)
{
struct drm_i915_cmd_descriptor *default_desc)
{
- const struct drm_i915_cmd_descriptor *desc;
+ if (((cmd_header ^ desc->cmd.value) & desc->cmd.mask) == 0)
+ return desc;
+
desc = find_cmd_in_table(engine, cmd_header);
if (desc)
return desc;
desc = find_cmd_in_table(engine, cmd_header);
if (desc)
return desc;
- BUG_ON(!default_desc);
- default_desc->flags = CMD_DESC_SKIP;
+ default_desc->cmd.value = cmd_header;
+ default_desc->cmd.mask = ~0u << MIN_OPCODE_SHIFT;
default_desc->length.mask = mask;
default_desc->length.mask = mask;
+ default_desc->flags = CMD_DESC_SKIP;
bool is_master)
{
u32 *cmd, *batch_end;
bool is_master)
{
u32 *cmd, *batch_end;
- struct drm_i915_cmd_descriptor default_desc = { 0 };
+ struct drm_i915_cmd_descriptor default_desc = noop_desc;
+ const struct drm_i915_cmd_descriptor *desc = &default_desc;
bool oacontrol_set = false; /* OACONTROL tracking. See check_cmd() */
bool needs_clflush_after = false;
int ret = 0;
bool oacontrol_set = false; /* OACONTROL tracking. See check_cmd() */
bool needs_clflush_after = false;
int ret = 0;
*/
batch_end = cmd + (batch_len / sizeof(*batch_end));
while (cmd < batch_end) {
*/
batch_end = cmd + (batch_len / sizeof(*batch_end));
while (cmd < batch_end) {
- const struct drm_i915_cmd_descriptor *desc;
u32 length;
if (*cmd == MI_BATCH_BUFFER_END)
break;
u32 length;
if (*cmd == MI_BATCH_BUFFER_END)
break;
- desc = find_cmd(engine, *cmd, &default_desc);
+ desc = find_cmd(engine, *cmd, desc, &default_desc);
if (!desc) {
DRM_DEBUG_DRIVER("CMD: Unrecognized command: 0x%08X\n",
*cmd);
if (!desc) {
DRM_DEBUG_DRIVER("CMD: Unrecognized command: 0x%08X\n",
*cmd);