summaryrefslogtreecommitdiffstats
path: root/src/intel/compiler
diff options
context:
space:
mode:
authorMatt Turner <[email protected]>2017-06-14 16:04:07 -0700
committerMatt Turner <[email protected]>2017-10-20 15:00:17 -0700
commitf11fa5ac6c5f955097c79e6bf27cd1a437c90ca5 (patch)
tree1547cbd52265a8ecd4ec0ab75c73d9e61d301411 /src/intel/compiler
parent6c7fc9b73a2919dbc246a888002210d7e4b9818b (diff)
i965: Add align1 ternary instruction disassembler support
Reviewed-by: Scott D Phillips <[email protected]>
Diffstat (limited to 'src/intel/compiler')
-rw-r--r--src/intel/compiler/brw_disasm.c352
-rw-r--r--src/intel/compiler/brw_eu_defines.h11
2 files changed, 288 insertions, 75 deletions
diff --git a/src/intel/compiler/brw_disasm.c b/src/intel/compiler/brw_disasm.c
index ef7f623ceb0..1a94ed39540 100644
--- a/src/intel/compiler/brw_disasm.c
+++ b/src/intel/compiler/brw_disasm.c
@@ -30,6 +30,7 @@
#include "brw_reg.h"
#include "brw_inst.h"
#include "brw_eu.h"
+#include "util/half_float.h"
static bool
has_jip(const struct gen_device_info *devinfo, enum opcode opcode)
@@ -237,13 +238,6 @@ static const char *const access_mode[2] = {
[1] = "align16",
};
-static const char *const three_source_reg_encoding[] = {
- [BRW_3SRC_TYPE_F] = "F",
- [BRW_3SRC_TYPE_D] = "D",
- [BRW_3SRC_TYPE_UD] = "UD",
- [BRW_3SRC_TYPE_DF] = "DF",
-};
-
static const char *const reg_file[4] = {
[0] = "A",
[1] = "g",
@@ -762,14 +756,15 @@ dest(FILE *file, const struct gen_device_info *devinfo, const brw_inst *inst)
static int
dest_3src(FILE *file, const struct gen_device_info *devinfo, const brw_inst *inst)
{
+ bool is_align1 = brw_inst_3src_access_mode(devinfo, inst) == BRW_ALIGN_1;
int err = 0;
uint32_t reg_file;
- enum brw_reg_type type = brw_inst_3src_a16_dst_type(devinfo, inst);
- unsigned dst_subreg_nr =
- brw_inst_3src_a16_dst_subreg_nr(devinfo, inst) * 4 /
- brw_reg_type_to_size(type);
+ unsigned subreg_nr;
+ enum brw_reg_type type;
- if (devinfo->gen == 6 && brw_inst_3src_a16_dst_reg_file(devinfo, inst))
+ if (is_align1 && brw_inst_3src_a1_dst_reg_file(devinfo, inst))
+ reg_file = BRW_ARCHITECTURE_REGISTER_FILE;
+ else if (devinfo->gen == 6 && brw_inst_3src_a16_dst_reg_file(devinfo, inst))
reg_file = BRW_MESSAGE_REGISTER_FILE;
else
reg_file = BRW_GENERAL_REGISTER_FILE;
@@ -777,13 +772,25 @@ dest_3src(FILE *file, const struct gen_device_info *devinfo, const brw_inst *ins
err |= reg(file, reg_file, brw_inst_3src_dst_reg_nr(devinfo, inst));
if (err == -1)
return 0;
- if (dst_subreg_nr)
- format(file, ".%u", dst_subreg_nr);
+
+ if (is_align1) {
+ type = brw_inst_3src_a1_dst_type(devinfo, inst);
+ subreg_nr = brw_inst_3src_a1_dst_subreg_nr(devinfo, inst);
+ } else {
+ type = brw_inst_3src_a16_dst_type(devinfo, inst);
+ subreg_nr = brw_inst_3src_a16_dst_subreg_nr(devinfo, inst) * 4;
+ }
+ subreg_nr /= brw_reg_type_to_size(type);
+
+ if (subreg_nr)
+ format(file, ".%u", subreg_nr);
string(file, "<1>");
- err |= control(file, "writemask", writemask,
- brw_inst_3src_a16_dst_writemask(devinfo, inst), NULL);
- err |= control(file, "dest reg encoding", three_source_reg_encoding,
- brw_inst_3src_a16_dst_hw_type(devinfo, inst), NULL);
+
+ if (!is_align1) {
+ err |= control(file, "writemask", writemask,
+ brw_inst_3src_a16_dst_writemask(devinfo, inst), NULL);
+ }
+ string(file, brw_reg_type_to_letters(type));
return 0;
}
@@ -928,33 +935,161 @@ src_da16(FILE *file,
return err;
}
+static enum brw_vertical_stride
+vstride_from_align1_3src_vstride(enum gen10_align1_3src_vertical_stride vstride)
+{
+ switch (vstride) {
+ case BRW_ALIGN1_3SRC_VERTICAL_STRIDE_0: return BRW_VERTICAL_STRIDE_0;
+ case BRW_ALIGN1_3SRC_VERTICAL_STRIDE_2: return BRW_VERTICAL_STRIDE_2;
+ case BRW_ALIGN1_3SRC_VERTICAL_STRIDE_4: return BRW_VERTICAL_STRIDE_4;
+ case BRW_ALIGN1_3SRC_VERTICAL_STRIDE_8: return BRW_VERTICAL_STRIDE_8;
+ default:
+ unreachable("not reached");
+ }
+}
+
+static enum brw_horizontal_stride
+hstride_from_align1_3src_hstride(enum gen10_align1_3src_src_horizontal_stride hstride)
+{
+ switch (hstride) {
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_0: return BRW_HORIZONTAL_STRIDE_0;
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_1: return BRW_HORIZONTAL_STRIDE_1;
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_2: return BRW_HORIZONTAL_STRIDE_2;
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_4: return BRW_HORIZONTAL_STRIDE_4;
+ default:
+ unreachable("not reached");
+ }
+}
+
+static enum brw_vertical_stride
+vstride_from_align1_3src_hstride(enum gen10_align1_3src_src_horizontal_stride hstride)
+{
+ switch (hstride) {
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_0: return BRW_VERTICAL_STRIDE_0;
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_1: return BRW_VERTICAL_STRIDE_1;
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_2: return BRW_VERTICAL_STRIDE_2;
+ case BRW_ALIGN1_3SRC_SRC_HORIZONTAL_STRIDE_4: return BRW_VERTICAL_STRIDE_4;
+ default:
+ unreachable("not reached");
+ }
+}
+
+/* From "GEN10 Regioning Rules for Align1 Ternary Operations" in the
+ * "Register Region Restrictions" documentation
+ */
+static enum brw_width
+implied_width(enum brw_vertical_stride _vert_stride,
+ enum brw_horizontal_stride _horiz_stride)
+{
+ /* "1. Width is 1 when Vertical and Horizontal Strides are both zero." */
+ if (_vert_stride == BRW_VERTICAL_STRIDE_0 &&
+ _horiz_stride == BRW_HORIZONTAL_STRIDE_0) {
+ return BRW_WIDTH_1;
+
+ /* "2. Width is equal to vertical stride when Horizontal Stride is zero." */
+ } else if (_horiz_stride == BRW_HORIZONTAL_STRIDE_0) {
+ switch (_vert_stride) {
+ case BRW_VERTICAL_STRIDE_2: return BRW_WIDTH_2;
+ case BRW_VERTICAL_STRIDE_4: return BRW_WIDTH_4;
+ case BRW_VERTICAL_STRIDE_8: return BRW_WIDTH_8;
+ case BRW_VERTICAL_STRIDE_0:
+ default:
+ unreachable("not reached");
+ }
+
+ } else {
+ /* FINISHME: Implement these: */
+
+ /* "3. Width is equal to Vertical Stride/Horizontal Stride when both
+ * Strides are non-zero.
+ *
+ * 4. Vertical Stride must not be zero if Horizontal Stride is non-zero.
+ * This implies Vertical Stride is always greater than Horizontal
+ * Stride."
+ *
+ * Given these statements and the knowledge that the stride and width
+ * values are encoded in logarithmic form, we can perform the division
+ * by just subtracting.
+ */
+ return _vert_stride - _horiz_stride;
+ }
+}
+
static int
src0_3src(FILE *file, const struct gen_device_info *devinfo, const brw_inst *inst)
{
int err = 0;
- enum brw_reg_type type = brw_inst_3src_a16_src_type(devinfo, inst);
- unsigned src0_subreg_nr =
- brw_inst_3src_a16_src0_subreg_nr(devinfo, inst) * 4 /
- brw_reg_type_to_size(type);
+ unsigned reg_nr, subreg_nr;
+ enum brw_reg_file _file;
+ enum brw_reg_type type;
+ enum brw_vertical_stride _vert_stride;
+ enum brw_width _width;
+ enum brw_horizontal_stride _horiz_stride;
+ bool is_scalar_region;
+ bool is_align1 = brw_inst_3src_access_mode(devinfo, inst) == BRW_ALIGN_1;
+
+ if (is_align1) {
+ if (brw_inst_3src_a1_src0_reg_file(devinfo, inst) ==
+ BRW_ALIGN1_3SRC_GENERAL_REGISTER_FILE) {
+ _file = BRW_GENERAL_REGISTER_FILE;
+ reg_nr = brw_inst_3src_src0_reg_nr(devinfo, inst);
+ subreg_nr = brw_inst_3src_a1_src0_subreg_nr(devinfo, inst);
+ type = brw_inst_3src_a1_src0_type(devinfo, inst);
+ } else {
+ _file = BRW_IMMEDIATE_VALUE;
+ uint16_t imm_val = brw_inst_3src_a1_src0_imm(devinfo, inst);
+ enum brw_reg_type type = brw_inst_3src_a1_src0_type(devinfo, inst);
+
+ if (type == BRW_REGISTER_TYPE_W) {
+ format(file, "%dW", imm_val);
+ } else if (type == BRW_REGISTER_TYPE_UW) {
+ format(file, "0x%04xUW", imm_val);
+ } else if (type == BRW_REGISTER_TYPE_HF) {
+ format(file, "%-gF", _mesa_half_to_float(imm_val));
+ }
+ return 0;
+ }
+
+ _vert_stride = vstride_from_align1_3src_vstride(
+ brw_inst_3src_a1_src0_vstride(devinfo, inst));
+ _horiz_stride = hstride_from_align1_3src_hstride(
+ brw_inst_3src_a1_src0_hstride(devinfo, inst));
+ _width = implied_width(_vert_stride, _horiz_stride);
+ } else {
+ _file = BRW_GENERAL_REGISTER_FILE;
+ reg_nr = brw_inst_3src_src0_reg_nr(devinfo, inst);
+ subreg_nr = brw_inst_3src_a16_src0_subreg_nr(devinfo, inst) * 4;
+ type = brw_inst_3src_a16_src_type(devinfo, inst);
+
+ if (brw_inst_3src_a16_src0_rep_ctrl(devinfo, inst)) {
+ _vert_stride = BRW_VERTICAL_STRIDE_0;
+ _width = BRW_WIDTH_1;
+ _horiz_stride = BRW_HORIZONTAL_STRIDE_0;
+ } else {
+ _vert_stride = BRW_VERTICAL_STRIDE_4;
+ _width = BRW_WIDTH_4;
+ _horiz_stride = BRW_HORIZONTAL_STRIDE_1;
+ }
+ }
+ is_scalar_region = _vert_stride == BRW_VERTICAL_STRIDE_0 &&
+ _width == BRW_WIDTH_1 &&
+ _horiz_stride == BRW_HORIZONTAL_STRIDE_0;
+
+ subreg_nr /= brw_reg_type_to_size(type);
err |= control(file, "negate", m_negate,
brw_inst_3src_src0_negate(devinfo, inst), NULL);
err |= control(file, "abs", _abs, brw_inst_3src_src0_abs(devinfo, inst), NULL);
- err |= reg(file, BRW_GENERAL_REGISTER_FILE,
- brw_inst_3src_src0_reg_nr(devinfo, inst));
+ err |= reg(file, _file, reg_nr);
if (err == -1)
return 0;
- if (src0_subreg_nr || brw_inst_3src_a16_src0_rep_ctrl(devinfo, inst))
- format(file, ".%d", src0_subreg_nr);
- if (brw_inst_3src_a16_src0_rep_ctrl(devinfo, inst))
- string(file, "<0,1,0>");
- else {
- string(file, "<4,4,1>");
+ if (subreg_nr || is_scalar_region)
+ format(file, ".%d", subreg_nr);
+ src_align1_region(file, _vert_stride, _width, _horiz_stride);
+ if (!is_scalar_region && !is_align1)
err |= src_swizzle(file, brw_inst_3src_a16_src0_swizzle(devinfo, inst));
- }
- err |= control(file, "src da16 reg type", three_source_reg_encoding,
- brw_inst_3src_a16_src_hw_type(devinfo, inst), NULL);
+ string(file, brw_reg_type_to_letters(type));
return err;
}
@@ -962,60 +1097,149 @@ static int
src1_3src(FILE *file, const struct gen_device_info *devinfo, const brw_inst *inst)
{
int err = 0;
- enum brw_reg_type type = brw_inst_3src_a16_src_type(devinfo, inst);
- unsigned src1_subreg_nr =
- brw_inst_3src_a16_src1_subreg_nr(devinfo, inst) * 4 /
- brw_reg_type_to_size(type);
+ unsigned reg_nr, subreg_nr;
+ enum brw_reg_file _file;
+ enum brw_reg_type type;
+ enum brw_vertical_stride _vert_stride;
+ enum brw_width _width;
+ enum brw_horizontal_stride _horiz_stride;
+ bool is_scalar_region;
+ bool is_align1 = brw_inst_3src_access_mode(devinfo, inst) == BRW_ALIGN_1;
+
+ if (is_align1) {
+ if (brw_inst_3src_a1_src1_reg_file(devinfo, inst) ==
+ BRW_ALIGN1_3SRC_GENERAL_REGISTER_FILE) {
+ _file = BRW_GENERAL_REGISTER_FILE;
+ } else {
+ _file = BRW_ARCHITECTURE_REGISTER_FILE;
+ }
+
+ reg_nr = brw_inst_3src_src1_reg_nr(devinfo, inst);
+ subreg_nr = brw_inst_3src_a1_src1_subreg_nr(devinfo, inst);
+ type = brw_inst_3src_a1_src1_type(devinfo, inst);
+
+ _vert_stride = vstride_from_align1_3src_vstride(
+ brw_inst_3src_a1_src1_vstride(devinfo, inst));
+ _horiz_stride = hstride_from_align1_3src_hstride(
+ brw_inst_3src_a1_src1_hstride(devinfo, inst));
+ _width = implied_width(_vert_stride, _horiz_stride);
+ } else {
+ _file = BRW_GENERAL_REGISTER_FILE;
+ reg_nr = brw_inst_3src_src1_reg_nr(devinfo, inst);
+ subreg_nr = brw_inst_3src_a16_src1_subreg_nr(devinfo, inst) * 4;
+ type = brw_inst_3src_a16_src_type(devinfo, inst);
+
+ if (brw_inst_3src_a16_src1_rep_ctrl(devinfo, inst)) {
+ _vert_stride = BRW_VERTICAL_STRIDE_0;
+ _width = BRW_WIDTH_1;
+ _horiz_stride = BRW_HORIZONTAL_STRIDE_0;
+ } else {
+ _vert_stride = BRW_VERTICAL_STRIDE_4;
+ _width = BRW_WIDTH_4;
+ _horiz_stride = BRW_HORIZONTAL_STRIDE_1;
+ }
+ }
+ is_scalar_region = _vert_stride == BRW_VERTICAL_STRIDE_0 &&
+ _width == BRW_WIDTH_1 &&
+ _horiz_stride == BRW_HORIZONTAL_STRIDE_0;
+
+ subreg_nr /= brw_reg_type_to_size(type);
err |= control(file, "negate", m_negate,
brw_inst_3src_src1_negate(devinfo, inst), NULL);
err |= control(file, "abs", _abs, brw_inst_3src_src1_abs(devinfo, inst), NULL);
- err |= reg(file, BRW_GENERAL_REGISTER_FILE,
- brw_inst_3src_src1_reg_nr(devinfo, inst));
+ err |= reg(file, _file, reg_nr);
if (err == -1)
return 0;
- if (src1_subreg_nr || brw_inst_3src_a16_src1_rep_ctrl(devinfo, inst))
- format(file, ".%d", src1_subreg_nr);
- if (brw_inst_3src_a16_src1_rep_ctrl(devinfo, inst))
- string(file, "<0,1,0>");
- else {
- string(file, "<4,4,1>");
+ if (subreg_nr || is_scalar_region)
+ format(file, ".%d", subreg_nr);
+ src_align1_region(file, _vert_stride, _width, _horiz_stride);
+ if (!is_scalar_region && !is_align1)
err |= src_swizzle(file, brw_inst_3src_a16_src1_swizzle(devinfo, inst));
- }
- err |= control(file, "src da16 reg type", three_source_reg_encoding,
- brw_inst_3src_a16_src_hw_type(devinfo, inst), NULL);
+ string(file, brw_reg_type_to_letters(type));
return err;
}
-
static int
src2_3src(FILE *file, const struct gen_device_info *devinfo, const brw_inst *inst)
{
int err = 0;
- enum brw_reg_type type = brw_inst_3src_a16_src_type(devinfo, inst);
- unsigned src2_subreg_nr =
- brw_inst_3src_a16_src2_subreg_nr(devinfo, inst) * 4 /
- brw_reg_type_to_size(type);
+ unsigned reg_nr, subreg_nr;
+ enum brw_reg_file _file;
+ enum brw_reg_type type;
+ enum brw_vertical_stride _vert_stride;
+ enum brw_width _width;
+ enum brw_horizontal_stride _horiz_stride;
+ bool is_scalar_region;
+ bool is_align1 = brw_inst_3src_access_mode(devinfo, inst) == BRW_ALIGN_1;
+
+ if (is_align1) {
+ if (brw_inst_3src_a1_src2_reg_file(devinfo, inst) ==
+ BRW_ALIGN1_3SRC_GENERAL_REGISTER_FILE) {
+ _file = BRW_GENERAL_REGISTER_FILE;
+ reg_nr = brw_inst_3src_src2_reg_nr(devinfo, inst);
+ subreg_nr = brw_inst_3src_a1_src2_subreg_nr(devinfo, inst);
+ type = brw_inst_3src_a1_src2_type(devinfo, inst);
+ } else {
+ _file = BRW_IMMEDIATE_VALUE;
+ uint16_t imm_val = brw_inst_3src_a1_src2_imm(devinfo, inst);
+ enum brw_reg_type type = brw_inst_3src_a1_src2_type(devinfo, inst);
+
+ if (type == BRW_REGISTER_TYPE_W) {
+ format(file, "%dW", imm_val);
+ } else if (type == BRW_REGISTER_TYPE_UW) {
+ format(file, "0x%04xUW", imm_val);
+ } else if (type == BRW_REGISTER_TYPE_HF) {
+ format(file, "%-gF", _mesa_half_to_float(imm_val));
+ }
+ return 0;
+ }
+
+ /* FINISHME: No vertical stride on src2. Is using the hstride in place
+ * correct? Doesn't seem like it, since there's hstride=1 but
+ * no vstride=1.
+ */
+ _vert_stride = vstride_from_align1_3src_hstride(
+ brw_inst_3src_a1_src2_hstride(devinfo, inst));
+ _horiz_stride = hstride_from_align1_3src_hstride(
+ brw_inst_3src_a1_src2_hstride(devinfo, inst));
+ _width = implied_width(_vert_stride, _horiz_stride);
+ } else {
+ _file = BRW_GENERAL_REGISTER_FILE;
+ reg_nr = brw_inst_3src_src2_reg_nr(devinfo, inst);
+ subreg_nr = brw_inst_3src_a16_src2_subreg_nr(devinfo, inst) * 4;
+ type = brw_inst_3src_a16_src_type(devinfo, inst);
+
+ if (brw_inst_3src_a16_src2_rep_ctrl(devinfo, inst)) {
+ _vert_stride = BRW_VERTICAL_STRIDE_0;
+ _width = BRW_WIDTH_1;
+ _horiz_stride = BRW_HORIZONTAL_STRIDE_0;
+ } else {
+ _vert_stride = BRW_VERTICAL_STRIDE_4;
+ _width = BRW_WIDTH_4;
+ _horiz_stride = BRW_HORIZONTAL_STRIDE_1;
+ }
+ }
+ is_scalar_region = _vert_stride == BRW_VERTICAL_STRIDE_0 &&
+ _width == BRW_WIDTH_1 &&
+ _horiz_stride == BRW_HORIZONTAL_STRIDE_0;
+
+ subreg_nr /= brw_reg_type_to_size(type);
err |= control(file, "negate", m_negate,
brw_inst_3src_src2_negate(devinfo, inst), NULL);
err |= control(file, "abs", _abs, brw_inst_3src_src2_abs(devinfo, inst), NULL);
- err |= reg(file, BRW_GENERAL_REGISTER_FILE,
- brw_inst_3src_src2_reg_nr(devinfo, inst));
+ err |= reg(file, _file, reg_nr);
if (err == -1)
return 0;
- if (src2_subreg_nr || brw_inst_3src_a16_src2_rep_ctrl(devinfo, inst))
- format(file, ".%d", src2_subreg_nr);
- if (brw_inst_3src_a16_src2_rep_ctrl(devinfo, inst))
- string(file, "<0,1,0>");
- else {
- string(file, "<4,4,1>");
+ if (subreg_nr || is_scalar_region)
+ format(file, ".%d", subreg_nr);
+ src_align1_region(file, _vert_stride, _width, _horiz_stride);
+ if (!is_scalar_region && !is_align1)
err |= src_swizzle(file, brw_inst_3src_a16_src2_swizzle(devinfo, inst));
- }
- err |= control(file, "src da16 reg type", three_source_reg_encoding,
- brw_inst_3src_a16_src_hw_type(devinfo, inst), NULL);
+ string(file, brw_reg_type_to_letters(type));
return err;
}
diff --git a/src/intel/compiler/brw_eu_defines.h b/src/intel/compiler/brw_eu_defines.h
index 3af55e830cf..291dd361a29 100644
--- a/src/intel/compiler/brw_eu_defines.h
+++ b/src/intel/compiler/brw_eu_defines.h
@@ -837,17 +837,6 @@ enum PACKED gen10_align1_3src_reg_file {
BRW_ALIGN1_3SRC_ACCUMULATOR = 1, /* dest, src1 */
};
-/* SNB adds 3-src instructions (MAD and LRP) that only operate on floats, so
- * the types were implied. IVB adds BFE and BFI2 that operate on doublewords
- * and unsigned doublewords, so a new field is also available in the da3src
- * struct (part of struct brw_instruction.bits1 in brw_structs.h) to select
- * dst and shared-src types. The values are different from BRW_REGISTER_TYPE_*.
- */
-#define BRW_3SRC_TYPE_F 0
-#define BRW_3SRC_TYPE_D 1
-#define BRW_3SRC_TYPE_UD 2
-#define BRW_3SRC_TYPE_DF 3
-
/* CNL adds Align1 support for 3-src instructions. Bit 35 of the instruction
* word is "Execution Datatype" which controls whether the instruction operates
* on float or integer types. The register arguments have fields that offer