From 0923dad90a3859d4136d483f3038fda9f616d1a8 Mon Sep 17 00:00:00 2001 From: Kenneth Graunke Date: Thu, 6 Dec 2012 22:36:26 -0800 Subject: i965: Add a new representation for Broadwell shader instructions. Broadwell significantly changes the EU instruction encoding. Many of the fields got moved to different bit positions; some even got split in two. With so many changes, it was infeasible to continue using struct brw_instruction. We needed a new representation. This new approach is a bit different: rather than a struct, I created a class that has four DWords, and helper functions that read/write various bits. This has several advantages: 1. We can create several different names for the same bits. For example, conditional modifiers, SFID for SEND instructions, and the MATH instruction's function opcode are all stored in bits 27:24. In each situation, we can use the appropriate setter function: set_sfid(), set_math_function(), or set_cond_modifier(). This is much easier to follow. 2. Since the fields are expressed using the original 128-bit numbers, the code to create the getter/setter functions follows the table in the documentation very closely. To aid in debugging, I've enabled -fkeep-inline-functions when building gen8_instruction.c. Otherwise, these functions cannot be called by gdb, making it insanely difficult to print out anything. Kenneth Graunke wrote most of this code. Damien Lespiau ported it to C99. Xiang Haihao added media fields. Zhao Yakui added indirect addressing support. Eric Anholt added an assertion to make sure that values fit in the alloted number of bits. v2: Update for brw_reg_type_to_hw_type(), which necessitates passing brw_context pointers around everywhere. Signed-off-by: Kenneth Graunke Signed-off-by: Damien Lespiau Signed-off-by: Xiang, Haihao Signed-off-by: Zhao Yakui Reviewed-by: Eric Anholt Acked-by: Matt Turner --- src/mesa/drivers/dri/i965/gen8_instruction.c | 433 +++++++++++++++++++++++++++ 1 file changed, 433 insertions(+) create mode 100644 src/mesa/drivers/dri/i965/gen8_instruction.c (limited to 'src/mesa/drivers/dri/i965/gen8_instruction.c') diff --git a/src/mesa/drivers/dri/i965/gen8_instruction.c b/src/mesa/drivers/dri/i965/gen8_instruction.c new file mode 100644 index 00000000000..476b3a91130 --- /dev/null +++ b/src/mesa/drivers/dri/i965/gen8_instruction.c @@ -0,0 +1,433 @@ +/* + * Copyright © 2012 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ + +/** + * @file gen8_instruction.c + * + * A representation of a Gen8+ EU instruction, with helper methods to get + * and set various fields. This is the actual hardware format. + */ + +#include "brw_defines.h" +#include "gen8_instruction.h" + +static void +gen8_convert_mrf_to_grf(struct brw_reg *reg) +{ + /* From the Ivybridge PRM, Volume 4 Part 3, page 218 ("send"): + * "The send with EOT should use register space R112-R127 for . This is + * to enable loading of a new thread into the same slot while the message + * with EOT for current thread is pending dispatch." + * + * Since we're pretending to have 16 MRFs anyway, we may as well use the + * registers required for messages with EOT. + */ + if (reg->file == BRW_MESSAGE_REGISTER_FILE) { + reg->file = BRW_GENERAL_REGISTER_FILE; + reg->nr += GEN7_MRF_HACK_START; + } +} + +void +gen8_set_dst(const struct brw_context *brw, + struct gen8_instruction *inst, + struct brw_reg reg) +{ + gen8_convert_mrf_to_grf(®); + + if (reg.file == BRW_GENERAL_REGISTER_FILE) + assert(reg.nr < BRW_MAX_GRF); + + gen8_set_dst_reg_file(inst, reg.file); + gen8_set_dst_reg_type(inst, brw_reg_type_to_hw_type(brw, reg.type, reg.file)); + gen8_set_dst_address_mode(inst, reg.address_mode); + + if (reg.address_mode == BRW_ADDRESS_DIRECT) { + gen8_set_dst_da_reg_nr(inst, reg.nr); + + if (gen8_access_mode(inst) == BRW_ALIGN_1) { + /* Set Dst.SubRegNum[4:0] */ + gen8_set_dst_da1_subreg_nr(inst, reg.subnr); + + /* Set Dst.HorzStride */ + if (reg.hstride == BRW_HORIZONTAL_STRIDE_0) + reg.hstride = BRW_HORIZONTAL_STRIDE_1; + gen8_set_dst_da1_hstride(inst, reg.hstride); + } else { + /* Align16 SubRegNum only has a single bit (bit 4; bits 3:0 MBZ). */ + assert(reg.subnr == 0 || reg.subnr == 16); + gen8_set_dst_da16_subreg_nr(inst, reg.subnr >> 4); + gen8_set_da16_writemask(inst, reg.dw1.bits.writemask); + } + } else { + /* Indirect addressing */ + assert(gen8_access_mode(inst) == BRW_ALIGN_1); + + /* Set Dst.HorzStride */ + if (reg.hstride == BRW_HORIZONTAL_STRIDE_0) + reg.hstride = BRW_HORIZONTAL_STRIDE_1; + gen8_set_dst_da1_hstride(inst, reg.hstride); + gen8_set_dst_ia1_subreg_nr(inst, reg.subnr); + gen8_set_dst_ia1_addr_imm(inst, reg.dw1.bits.indirect_offset); + } + + /* Generators should set a default exec_size of either 8 (SIMD4x2 or SIMD8) + * or 16 (SIMD16), as that's normally correct. However, when dealing with + * small registers, we automatically reduce it to match the register size. + */ + if (reg.width < BRW_EXECUTE_8) + gen8_set_exec_size(inst, reg.width); +} + +static void +gen8_validate_reg(struct gen8_instruction *inst, struct brw_reg reg) +{ + int hstride_for_reg[] = {0, 1, 2, 4}; + int vstride_for_reg[] = {0, 1, 2, 4, 8, 16, 32, 64, 128, 256}; + int width_for_reg[] = {1, 2, 4, 8, 16}; + int execsize_for_reg[] = {1, 2, 4, 8, 16}; + int width, hstride, vstride, execsize; + + if (reg.file == BRW_IMMEDIATE_VALUE) { + /* TODO: check immediate vectors */ + return; + } + + if (reg.file == BRW_ARCHITECTURE_REGISTER_FILE) + return; + + assert(reg.hstride >= 0 && reg.hstride < ARRAY_SIZE(hstride_for_reg)); + hstride = hstride_for_reg[reg.hstride]; + + if (reg.vstride == 0xf) { + vstride = -1; + } else { + assert(reg.vstride >= 0 && reg.vstride < ARRAY_SIZE(vstride_for_reg)); + vstride = vstride_for_reg[reg.vstride]; + } + + assert(reg.width >= 0 && reg.width < ARRAY_SIZE(width_for_reg)); + width = width_for_reg[reg.width]; + + assert(gen8_exec_size(inst) >= 0 && + gen8_exec_size(inst) < ARRAY_SIZE(execsize_for_reg)); + execsize = execsize_for_reg[gen8_exec_size(inst)]; + + /* Restrictions from 3.3.10: Register Region Restrictions. */ + /* 3. */ + assert(execsize >= width); + + /* 4. */ + if (execsize == width && hstride != 0) { + assert(vstride == -1 || vstride == width * hstride); + } + + /* 5. */ + if (execsize == width && hstride == 0) { + /* no restriction on vstride. */ + } + + /* 6. */ + if (width == 1) { + assert(hstride == 0); + } + + /* 7. */ + if (execsize == 1 && width == 1) { + assert(hstride == 0); + assert(vstride == 0); + } + + /* 8. */ + if (vstride == 0 && hstride == 0) { + assert(width == 1); + } + + /* 10. Check destination issues. */ +} + +void +gen8_set_src0(const struct brw_context *brw, + struct gen8_instruction *inst, + struct brw_reg reg) +{ + gen8_convert_mrf_to_grf(®); + + if (reg.file == BRW_GENERAL_REGISTER_FILE) + assert(reg.nr < BRW_MAX_GRF); + + gen8_validate_reg(inst, reg); + + gen8_set_src0_reg_file(inst, reg.file); + gen8_set_src0_reg_type(inst, + brw_reg_type_to_hw_type(brw, reg.type, reg.file)); + gen8_set_src0_abs(inst, reg.abs); + gen8_set_src0_negate(inst, reg.negate); + + if (reg.file == BRW_IMMEDIATE_VALUE) { + inst->data[3] = reg.dw1.ud; + + /* Required to set some fields in src1 as well: */ + gen8_set_src1_reg_file(inst, BRW_ARCHITECTURE_REGISTER_FILE); + gen8_set_src1_reg_type(inst, + brw_reg_type_to_hw_type(brw, reg.type, reg.file)); + return; + } + + gen8_set_src0_address_mode(inst, reg.address_mode); + + if (reg.address_mode == BRW_ADDRESS_DIRECT) { + gen8_set_src0_da_reg_nr(inst, reg.nr); + + if (gen8_access_mode(inst) == BRW_ALIGN_1) { + /* Set Src0.SubRegNum[4:0] */ + gen8_set_src0_da1_subreg_nr(inst, reg.subnr); + + if (reg.width == BRW_WIDTH_1 && gen8_exec_size(inst) == BRW_EXECUTE_1) { + gen8_set_src0_da1_hstride(inst, BRW_HORIZONTAL_STRIDE_0); + gen8_set_src0_vert_stride(inst, BRW_VERTICAL_STRIDE_0); + } else { + gen8_set_src0_da1_hstride(inst, reg.hstride); + gen8_set_src0_vert_stride(inst, reg.vstride); + } + gen8_set_src0_da1_width(inst, reg.width); + + } else { + /* Align16 SubRegNum only has a single bit (bit 4; bits 3:0 MBZ). */ + assert(reg.subnr == 0 || reg.subnr == 16); + gen8_set_src0_da16_subreg_nr(inst, reg.subnr >> 4); + + gen8_set_src0_da16_swiz_x(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_X)); + gen8_set_src0_da16_swiz_y(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_Y)); + gen8_set_src0_da16_swiz_z(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_Z)); + gen8_set_src0_da16_swiz_w(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_W)); + + /* This is an oddity of the fact that we're using the same + * descriptions for registers in both Align16 and Align1 modes. + */ + if (reg.vstride == BRW_VERTICAL_STRIDE_8) + gen8_set_src0_vert_stride(inst, BRW_VERTICAL_STRIDE_4); + else + gen8_set_src0_vert_stride(inst, reg.vstride); + } + } else { + /* Indirect addressing */ + assert(gen8_access_mode(inst) == BRW_ALIGN_1); + if (reg.width == BRW_WIDTH_1 && + gen8_exec_size(inst) == BRW_EXECUTE_1) { + gen8_set_src0_da1_hstride(inst, BRW_HORIZONTAL_STRIDE_0); + gen8_set_src0_vert_stride(inst, BRW_VERTICAL_STRIDE_0); + } else { + gen8_set_src0_da1_hstride(inst, reg.hstride); + gen8_set_src0_vert_stride(inst, reg.vstride); + } + + gen8_set_src0_da1_width(inst, reg.width); + gen8_set_src0_ia1_subreg_nr(inst, reg.subnr); + gen8_set_src0_ia1_addr_imm(inst, reg.dw1.bits.indirect_offset); + } +} + +void +gen8_set_src1(const struct brw_context *brw, + struct gen8_instruction *inst, + struct brw_reg reg) +{ + gen8_convert_mrf_to_grf(®); + + if (reg.file == BRW_GENERAL_REGISTER_FILE) + assert(reg.nr < BRW_MAX_GRF); + + gen8_validate_reg(inst, reg); + + gen8_set_src1_reg_file(inst, reg.file); + gen8_set_src1_reg_type(inst, + brw_reg_type_to_hw_type(brw, reg.type, reg.file)); + gen8_set_src1_abs(inst, reg.abs); + gen8_set_src1_negate(inst, reg.negate); + + /* Only src1 can be an immediate in two-argument instructions. */ + assert(gen8_src0_reg_file(inst) != BRW_IMMEDIATE_VALUE); + + if (reg.file == BRW_IMMEDIATE_VALUE) { + inst->data[3] = reg.dw1.ud; + return; + } + + gen8_set_src1_address_mode(inst, reg.address_mode); + + if (reg.address_mode == BRW_ADDRESS_DIRECT) { + gen8_set_src1_da_reg_nr(inst, reg.nr); + + if (gen8_access_mode(inst) == BRW_ALIGN_1) { + /* Set Src0.SubRegNum[4:0] */ + gen8_set_src1_da1_subreg_nr(inst, reg.subnr); + + if (reg.width == BRW_WIDTH_1 && gen8_exec_size(inst) == BRW_EXECUTE_1) { + gen8_set_src1_da1_hstride(inst, BRW_HORIZONTAL_STRIDE_0); + gen8_set_src1_vert_stride(inst, BRW_VERTICAL_STRIDE_0); + } else { + gen8_set_src1_da1_hstride(inst, reg.hstride); + gen8_set_src1_vert_stride(inst, reg.vstride); + } + gen8_set_src1_da1_width(inst, reg.width); + } else { + /* Align16 SubRegNum only has a single bit (bit 4; bits 3:0 MBZ). */ + assert(reg.subnr == 0 || reg.subnr == 16); + gen8_set_src1_da16_subreg_nr(inst, reg.subnr >> 4); + + gen8_set_src1_da16_swiz_x(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_X)); + gen8_set_src1_da16_swiz_y(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_Y)); + gen8_set_src1_da16_swiz_z(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_Z)); + gen8_set_src1_da16_swiz_w(inst, + BRW_GET_SWZ(reg.dw1.bits.swizzle, + BRW_CHANNEL_W)); + + /* This is an oddity of the fact that we're using the same + * descriptions for registers in both Align16 and Align1 modes. + */ + if (reg.vstride == BRW_VERTICAL_STRIDE_8) + gen8_set_src1_vert_stride(inst, BRW_VERTICAL_STRIDE_4); + else + gen8_set_src1_vert_stride(inst, reg.vstride); + } + } else { + /* Indirect addressing */ + assert(gen8_access_mode(inst) == BRW_ALIGN_1); + if (reg.width == BRW_WIDTH_1 && gen8_exec_size(inst) == BRW_EXECUTE_1) { + gen8_set_src1_da1_hstride(inst, BRW_HORIZONTAL_STRIDE_0); + gen8_set_src1_vert_stride(inst, BRW_VERTICAL_STRIDE_0); + } else { + gen8_set_src1_da1_hstride(inst, reg.hstride); + gen8_set_src1_vert_stride(inst, reg.vstride); + } + + gen8_set_src1_da1_width(inst, reg.width); + gen8_set_src1_ia1_subreg_nr(inst, reg.subnr); + gen8_set_src1_ia1_addr_imm(inst, reg.dw1.bits.indirect_offset); + } +} + +/** + * Set the Message Descriptor and Extended Message Descriptor fields + * for SEND messages. + * + * \note This zeroes out the Function Control bits, so it must be called + * \b before filling out any message-specific data. Callers can + * choose not to fill in irrelevant bits; they will be zero. + */ +static void +gen8_set_message_descriptor(const struct brw_context *brw, + struct gen8_instruction *inst, + enum brw_message_target sfid, + unsigned msg_length, + unsigned response_length, + bool header_present, + bool end_of_thread) +{ + gen8_set_src1(brw, inst, brw_imm_d(0)); + + gen8_set_sfid(inst, sfid); + gen8_set_mlen(inst, msg_length); + gen8_set_rlen(inst, response_length); + gen8_set_header_present(inst, header_present); + gen8_set_eot(inst, end_of_thread); +} + +void +gen8_set_urb_message(const struct brw_context *brw, + struct gen8_instruction *inst, + enum brw_urb_write_flags flags, + unsigned msg_length, + unsigned response_length, + unsigned offset, + bool interleave) +{ + gen8_set_message_descriptor(brw, inst, BRW_SFID_URB, + msg_length, response_length, + true, flags & BRW_URB_WRITE_EOT); + gen8_set_src0(brw, inst, brw_vec8_grf(GEN7_MRF_HACK_START + 1, 0)); + if (flags & BRW_URB_WRITE_OWORD) { + assert(msg_length == 2); + gen8_set_urb_opcode(inst, BRW_URB_OPCODE_WRITE_OWORD); + } else { + gen8_set_urb_opcode(inst, BRW_URB_OPCODE_WRITE_HWORD); + } + gen8_set_urb_global_offset(inst, offset); + gen8_set_urb_interleave(inst, interleave); + gen8_set_urb_per_slot_offset(inst, + flags & BRW_URB_WRITE_PER_SLOT_OFFSET ? 1 : 0); +} + +void +gen8_set_sampler_message(const struct brw_context *brw, + struct gen8_instruction *inst, + unsigned binding_table_index, + unsigned sampler, + unsigned msg_type, + unsigned response_length, + unsigned msg_length, + bool header_present, + unsigned simd_mode) +{ + gen8_set_message_descriptor(brw, inst, BRW_SFID_SAMPLER, msg_length, + response_length, header_present, false); + + gen8_set_binding_table_index(inst, binding_table_index); + gen8_set_sampler(inst, sampler); + gen8_set_sampler_msg_type(inst, msg_type); + gen8_set_sampler_simd_mode(inst, simd_mode); +} + +void +gen8_set_dp_message(const struct brw_context *brw, + struct gen8_instruction *inst, + enum brw_message_target sfid, + unsigned binding_table_index, + unsigned msg_type, + unsigned msg_control, + unsigned mlen, + unsigned rlen, + bool header_present, + bool end_of_thread) +{ + gen8_set_message_descriptor(brw, inst, sfid, mlen, rlen, header_present, + end_of_thread); + gen8_set_binding_table_index(inst, binding_table_index); + gen8_set_dp_message_type(inst, msg_type); + gen8_set_dp_message_control(inst, msg_control); +} -- cgit v1.2.3