* [PULL 00/23] tcg patch queue
@ 2023-05-25 18:10 Richard Henderson
2023-05-25 18:10 ` [PULL 01/23] tcg/mips: Move TCG_AREG0 to S8 Richard Henderson
` (23 more replies)
0 siblings, 24 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
The following changes since commit b300c134465465385045ab705b68a42699688332:
Merge tag 'pull-vfio-20230524' of https://github.com/legoater/qemu into staging (2023-05-24 14:23:41 -0700)
are available in the Git repository at:
https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20230525
for you to fetch changes up to a30498fcea5a8b9c544324ccfb0186090104b229:
tcg/riscv: Support CTZ, CLZ from Zbb (2023-05-25 15:29:36 +0000)
----------------------------------------------------------------
tcg/mips:
- Constant formation improvements
- Replace MIPS_BE with HOST_BIG_ENDIAN
- General cleanups
tcg/riscv:
- Improve setcond
- Support movcond
- Support Zbb, Zba
----------------------------------------------------------------
Richard Henderson (23):
tcg/mips: Move TCG_AREG0 to S8
tcg/mips: Move TCG_GUEST_BASE_REG to S7
tcg/mips: Unify TCG_GUEST_BASE_REG tests
tcg/mips: Create and use TCG_REG_TB
tcg/mips: Split out tcg_out_movi_one
tcg/mips: Split out tcg_out_movi_two
tcg/mips: Use the constant pool for 64-bit constants
tcg/mips: Aggressively use the constant pool for n64 calls
tcg/mips: Try tb-relative addresses in tcg_out_movi
tcg/mips: Try three insns with shift and add in tcg_out_movi
tcg/mips: Use qemu_build_not_reached for LO/HI_OFF
tcg/mips: Replace MIPS_BE with HOST_BIG_ENDIAN
disas/riscv: Decode czero.{eqz,nez}
tcg/riscv: Probe for Zba, Zbb, Zicond extensions
tcg/riscv: Support ANDN, ORN, XNOR from Zbb
tcg/riscv: Support ADD.UW, SEXT.B, SEXT.H, ZEXT.H from Zba+Zbb
tcg/riscv: Use ADD.UW for guest address generation
tcg/riscv: Support rotates from Zbb
tcg/riscv: Support REV8 from Zbb
tcg/riscv: Support CPOP from Zbb
tcg/riscv: Improve setcond expansion
tcg/riscv: Implement movcond
tcg/riscv: Support CTZ, CLZ from Zbb
tcg/mips/tcg-target.h | 3 +-
tcg/riscv/tcg-target-con-set.h | 3 +
tcg/riscv/tcg-target-con-str.h | 1 +
tcg/riscv/tcg-target.h | 48 ++--
disas/riscv.c | 6 +
tcg/mips/tcg-target.c.inc | 308 ++++++++++++++++-----
tcg/riscv/tcg-target.c.inc | 612 ++++++++++++++++++++++++++++++++++++-----
7 files changed, 825 insertions(+), 156 deletions(-)
^ permalink raw reply [flat|nested] 25+ messages in thread
* [PULL 01/23] tcg/mips: Move TCG_AREG0 to S8
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 02/23] tcg/mips: Move TCG_GUEST_BASE_REG to S7 Richard Henderson
` (22 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
No functional change; just moving the saved reserved regs to the end.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.h | 2 +-
tcg/mips/tcg-target.c.inc | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/tcg/mips/tcg-target.h b/tcg/mips/tcg-target.h
index 7277a117ef..46b63e59cc 100644
--- a/tcg/mips/tcg-target.h
+++ b/tcg/mips/tcg-target.h
@@ -76,7 +76,7 @@ typedef enum {
TCG_REG_RA,
TCG_REG_CALL_STACK = TCG_REG_SP,
- TCG_AREG0 = TCG_REG_S0,
+ TCG_AREG0 = TCG_REG_S8,
} TCGReg;
/* used for function call generation */
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index ef146b193c..ee6c2eb872 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -2183,7 +2183,7 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
}
static const int tcg_target_callee_save_regs[] = {
- TCG_REG_S0, /* used for the global env (TCG_AREG0) */
+ TCG_REG_S0,
TCG_REG_S1,
TCG_REG_S2,
TCG_REG_S3,
@@ -2191,7 +2191,7 @@ static const int tcg_target_callee_save_regs[] = {
TCG_REG_S5,
TCG_REG_S6,
TCG_REG_S7,
- TCG_REG_S8,
+ TCG_REG_S8, /* used for the global env (TCG_AREG0) */
TCG_REG_RA, /* should be last for ABI compliance */
};
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 02/23] tcg/mips: Move TCG_GUEST_BASE_REG to S7
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
2023-05-25 18:10 ` [PULL 01/23] tcg/mips: Move TCG_AREG0 to S8 Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 03/23] tcg/mips: Unify TCG_GUEST_BASE_REG tests Richard Henderson
` (21 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
No functional change; just moving the saved reserved regs to the end.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index ee6c2eb872..f322891797 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -86,7 +86,7 @@ static const char * const tcg_target_reg_names[TCG_TARGET_NB_REGS] = {
#define TCG_TMP3 TCG_REG_T7
#ifndef CONFIG_SOFTMMU
-#define TCG_GUEST_BASE_REG TCG_REG_S1
+#define TCG_GUEST_BASE_REG TCG_REG_S7
#endif
/* check if we really need so many registers :P */
@@ -2190,7 +2190,7 @@ static const int tcg_target_callee_save_regs[] = {
TCG_REG_S4,
TCG_REG_S5,
TCG_REG_S6,
- TCG_REG_S7,
+ TCG_REG_S7, /* used for guest_base */
TCG_REG_S8, /* used for the global env (TCG_AREG0) */
TCG_REG_RA, /* should be last for ABI compliance */
};
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 03/23] tcg/mips: Unify TCG_GUEST_BASE_REG tests
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
2023-05-25 18:10 ` [PULL 01/23] tcg/mips: Move TCG_AREG0 to S8 Richard Henderson
2023-05-25 18:10 ` [PULL 02/23] tcg/mips: Move TCG_GUEST_BASE_REG to S7 Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 04/23] tcg/mips: Create and use TCG_REG_TB Richard Henderson
` (20 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
In tcg_out_qemu_ld/st, we already check for guest_base matching int16_t.
Mirror that when setting up TCG_GUEST_BASE_REG in the prologue.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index f322891797..ccb3a1cd9a 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -2312,7 +2312,7 @@ static void tcg_target_qemu_prologue(TCGContext *s)
}
#ifndef CONFIG_SOFTMMU
- if (guest_base) {
+ if (guest_base != (int16_t)guest_base) {
tcg_out_movi(s, TCG_TYPE_PTR, TCG_GUEST_BASE_REG, guest_base);
tcg_regset_set_reg(s->reserved_regs, TCG_GUEST_BASE_REG);
}
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 04/23] tcg/mips: Create and use TCG_REG_TB
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (2 preceding siblings ...)
2023-05-25 18:10 ` [PULL 03/23] tcg/mips: Unify TCG_GUEST_BASE_REG tests Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 05/23] tcg/mips: Split out tcg_out_movi_one Richard Henderson
` (19 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
This vastly reduces the size of code generated for 64-bit addresses.
The code for exit_tb, for instance, where we load a (tagged) pointer
to the current TB, goes from
0x400aa9725c: li v0,64
0x400aa97260: dsll v0,v0,0x10
0x400aa97264: ori v0,v0,0xaa9
0x400aa97268: dsll v0,v0,0x10
0x400aa9726c: j 0x400aa9703c
0x400aa97270: ori v0,v0,0x7083
to
0x400aa97240: j 0x400aa97040
0x400aa97244: daddiu v0,s6,-189
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 69 +++++++++++++++++++++++++++++++++------
1 file changed, 59 insertions(+), 10 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index ccb3a1cd9a..6f03b44ac0 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -88,6 +88,11 @@ static const char * const tcg_target_reg_names[TCG_TARGET_NB_REGS] = {
#ifndef CONFIG_SOFTMMU
#define TCG_GUEST_BASE_REG TCG_REG_S7
#endif
+#if TCG_TARGET_REG_BITS == 64
+#define TCG_REG_TB TCG_REG_S6
+#else
+#define TCG_REG_TB (qemu_build_not_reached(), TCG_REG_ZERO)
+#endif
/* check if we really need so many registers :P */
static const int tcg_target_reg_alloc_order[] = {
@@ -1547,27 +1552,61 @@ static void tcg_out_clz(TCGContext *s, MIPSInsn opcv2, MIPSInsn opcv6,
static void tcg_out_exit_tb(TCGContext *s, uintptr_t a0)
{
- TCGReg b0 = TCG_REG_ZERO;
+ TCGReg base = TCG_REG_ZERO;
+ int16_t lo = 0;
- if (a0 & ~0xffff) {
- tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_V0, a0 & ~0xffff);
- b0 = TCG_REG_V0;
+ if (a0) {
+ intptr_t ofs;
+ if (TCG_TARGET_REG_BITS == 64) {
+ ofs = tcg_tbrel_diff(s, (void *)a0);
+ lo = ofs;
+ if (ofs == lo) {
+ base = TCG_REG_TB;
+ } else {
+ base = TCG_REG_V0;
+ tcg_out_movi(s, TCG_TYPE_PTR, base, ofs - lo);
+ tcg_out_opc_reg(s, ALIAS_PADD, base, base, TCG_REG_TB);
+ }
+ } else {
+ ofs = a0;
+ lo = ofs;
+ base = TCG_REG_V0;
+ tcg_out_movi(s, TCG_TYPE_PTR, base, ofs - lo);
+ }
}
if (!tcg_out_opc_jmp(s, OPC_J, tb_ret_addr)) {
tcg_out_movi(s, TCG_TYPE_PTR, TCG_TMP0, (uintptr_t)tb_ret_addr);
tcg_out_opc_reg(s, OPC_JR, 0, TCG_TMP0, 0);
}
- tcg_out_opc_imm(s, OPC_ORI, TCG_REG_V0, b0, a0 & 0xffff);
+ /* delay slot */
+ tcg_out_opc_imm(s, ALIAS_PADDI, TCG_REG_V0, base, lo);
}
static void tcg_out_goto_tb(TCGContext *s, int which)
{
+ intptr_t ofs = get_jmp_target_addr(s, which);
+ TCGReg base, dest;
+
/* indirect jump method */
- tcg_out_ld(s, TCG_TYPE_PTR, TCG_TMP0, TCG_REG_ZERO,
- get_jmp_target_addr(s, which));
- tcg_out_opc_reg(s, OPC_JR, 0, TCG_TMP0, 0);
+ if (TCG_TARGET_REG_BITS == 64) {
+ dest = TCG_REG_TB;
+ base = TCG_REG_TB;
+ ofs = tcg_tbrel_diff(s, (void *)ofs);
+ } else {
+ dest = TCG_TMP0;
+ base = TCG_REG_ZERO;
+ }
+ tcg_out_ld(s, TCG_TYPE_PTR, dest, base, ofs);
+ tcg_out_opc_reg(s, OPC_JR, 0, dest, 0);
+ /* delay slot */
tcg_out_nop(s);
+
set_jmp_reset_offset(s, which);
+ if (TCG_TARGET_REG_BITS == 64) {
+ /* For the unlinked case, need to reset TCG_REG_TB. */
+ tcg_out_ldst(s, ALIAS_PADDI, TCG_REG_TB, TCG_REG_TB,
+ -tcg_current_code_size(s));
+ }
}
void tb_target_set_jmp_target(const TranslationBlock *tb, int n,
@@ -1598,7 +1637,11 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
case INDEX_op_goto_ptr:
/* jmp to the given host address (could be epilogue) */
tcg_out_opc_reg(s, OPC_JR, 0, a0, 0);
- tcg_out_nop(s);
+ if (TCG_TARGET_REG_BITS == 64) {
+ tcg_out_mov(s, TCG_TYPE_PTR, TCG_REG_TB, a0);
+ } else {
+ tcg_out_nop(s);
+ }
break;
case INDEX_op_br:
tcg_out_brcond(s, TCG_COND_EQ, TCG_REG_ZERO, TCG_REG_ZERO,
@@ -2189,7 +2232,7 @@ static const int tcg_target_callee_save_regs[] = {
TCG_REG_S3,
TCG_REG_S4,
TCG_REG_S5,
- TCG_REG_S6,
+ TCG_REG_S6, /* used for the tb base (TCG_REG_TB) */
TCG_REG_S7, /* used for guest_base */
TCG_REG_S8, /* used for the global env (TCG_AREG0) */
TCG_REG_RA, /* should be last for ABI compliance */
@@ -2317,6 +2360,9 @@ static void tcg_target_qemu_prologue(TCGContext *s)
tcg_regset_set_reg(s->reserved_regs, TCG_GUEST_BASE_REG);
}
#endif
+ if (TCG_TARGET_REG_BITS == 64) {
+ tcg_out_mov(s, TCG_TYPE_PTR, TCG_REG_TB, tcg_target_call_iarg_regs[1]);
+ }
/* Call generated code */
tcg_out_opc_reg(s, OPC_JR, 0, tcg_target_call_iarg_regs[1], 0);
@@ -2498,6 +2544,9 @@ static void tcg_target_init(TCGContext *s)
tcg_regset_set_reg(s->reserved_regs, TCG_REG_RA); /* return address */
tcg_regset_set_reg(s->reserved_regs, TCG_REG_SP); /* stack pointer */
tcg_regset_set_reg(s->reserved_regs, TCG_REG_GP); /* global pointer */
+ if (TCG_TARGET_REG_BITS == 64) {
+ tcg_regset_set_reg(s->reserved_regs, TCG_REG_TB); /* tc->tc_ptr */
+ }
}
typedef struct {
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 05/23] tcg/mips: Split out tcg_out_movi_one
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (3 preceding siblings ...)
2023-05-25 18:10 ` [PULL 04/23] tcg/mips: Create and use TCG_REG_TB Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 06/23] tcg/mips: Split out tcg_out_movi_two Richard Henderson
` (18 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
Emit all constants that can be loaded in exactly one insn.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 26 ++++++++++++++++++++------
1 file changed, 20 insertions(+), 6 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index 6f03b44ac0..bbff510c46 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -510,20 +510,34 @@ static bool tcg_out_mov(TCGContext *s, TCGType type, TCGReg ret, TCGReg arg)
return true;
}
+static bool tcg_out_movi_one(TCGContext *s, TCGReg ret, tcg_target_long arg)
+{
+ if (arg == (int16_t)arg) {
+ tcg_out_opc_imm(s, OPC_ADDIU, ret, TCG_REG_ZERO, arg);
+ return true;
+ }
+ if (arg == (uint16_t)arg) {
+ tcg_out_opc_imm(s, OPC_ORI, ret, TCG_REG_ZERO, arg);
+ return true;
+ }
+ if (arg == (int32_t)arg && (arg & 0xffff) == 0) {
+ tcg_out_opc_imm(s, OPC_LUI, ret, TCG_REG_ZERO, arg >> 16);
+ return true;
+ }
+ return false;
+}
+
static void tcg_out_movi(TCGContext *s, TCGType type,
TCGReg ret, tcg_target_long arg)
{
if (TCG_TARGET_REG_BITS == 64 && type == TCG_TYPE_I32) {
arg = (int32_t)arg;
}
- if (arg == (int16_t)arg) {
- tcg_out_opc_imm(s, OPC_ADDIU, ret, TCG_REG_ZERO, arg);
- return;
- }
- if (arg == (uint16_t)arg) {
- tcg_out_opc_imm(s, OPC_ORI, ret, TCG_REG_ZERO, arg);
+
+ if (tcg_out_movi_one(s, ret, arg)) {
return;
}
+
if (TCG_TARGET_REG_BITS == 32 || arg == (int32_t)arg) {
tcg_out_opc_imm(s, OPC_LUI, ret, TCG_REG_ZERO, arg >> 16);
} else {
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 06/23] tcg/mips: Split out tcg_out_movi_two
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (4 preceding siblings ...)
2023-05-25 18:10 ` [PULL 05/23] tcg/mips: Split out tcg_out_movi_one Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 07/23] tcg/mips: Use the constant pool for 64-bit constants Richard Henderson
` (17 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
Emit all 32-bit signed constants, which can be loaded in two insns.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 35 ++++++++++++++++++++++++-----------
1 file changed, 24 insertions(+), 11 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index bbff510c46..7a19f8db1d 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -527,6 +527,22 @@ static bool tcg_out_movi_one(TCGContext *s, TCGReg ret, tcg_target_long arg)
return false;
}
+static bool tcg_out_movi_two(TCGContext *s, TCGReg ret, tcg_target_long arg)
+{
+ /*
+ * All signed 32-bit constants are loadable with two immediates,
+ * and everything else requires more work.
+ */
+ if (arg == (int32_t)arg) {
+ if (!tcg_out_movi_one(s, ret, arg)) {
+ tcg_out_opc_imm(s, OPC_LUI, ret, TCG_REG_ZERO, arg >> 16);
+ tcg_out_opc_imm(s, OPC_ORI, ret, ret, arg & 0xffff);
+ }
+ return true;
+ }
+ return false;
+}
+
static void tcg_out_movi(TCGContext *s, TCGType type,
TCGReg ret, tcg_target_long arg)
{
@@ -534,21 +550,18 @@ static void tcg_out_movi(TCGContext *s, TCGType type,
arg = (int32_t)arg;
}
- if (tcg_out_movi_one(s, ret, arg)) {
+ /* Load all 32-bit constants. */
+ if (tcg_out_movi_two(s, ret, arg)) {
return;
}
- if (TCG_TARGET_REG_BITS == 32 || arg == (int32_t)arg) {
- tcg_out_opc_imm(s, OPC_LUI, ret, TCG_REG_ZERO, arg >> 16);
+ tcg_out_movi(s, TCG_TYPE_I32, ret, arg >> 31 >> 1);
+ if (arg & 0xffff0000ull) {
+ tcg_out_dsll(s, ret, ret, 16);
+ tcg_out_opc_imm(s, OPC_ORI, ret, ret, arg >> 16);
+ tcg_out_dsll(s, ret, ret, 16);
} else {
- tcg_out_movi(s, TCG_TYPE_I32, ret, arg >> 31 >> 1);
- if (arg & 0xffff0000ull) {
- tcg_out_dsll(s, ret, ret, 16);
- tcg_out_opc_imm(s, OPC_ORI, ret, ret, arg >> 16);
- tcg_out_dsll(s, ret, ret, 16);
- } else {
- tcg_out_dsll(s, ret, ret, 32);
- }
+ tcg_out_dsll(s, ret, ret, 32);
}
if (arg & 0xffff) {
tcg_out_opc_imm(s, OPC_ORI, ret, ret, arg & 0xffff);
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 07/23] tcg/mips: Use the constant pool for 64-bit constants
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (5 preceding siblings ...)
2023-05-25 18:10 ` [PULL 06/23] tcg/mips: Split out tcg_out_movi_two Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 08/23] tcg/mips: Aggressively use the constant pool for n64 calls Richard Henderson
` (16 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
During normal processing, the constant pool is accessible via
TCG_REG_TB. During the prologue, it is accessible via TCG_REG_T9.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.h | 1 +
tcg/mips/tcg-target.c.inc | 65 +++++++++++++++++++++++++++++----------
2 files changed, 49 insertions(+), 17 deletions(-)
diff --git a/tcg/mips/tcg-target.h b/tcg/mips/tcg-target.h
index 46b63e59cc..8fbb6c6507 100644
--- a/tcg/mips/tcg-target.h
+++ b/tcg/mips/tcg-target.h
@@ -208,5 +208,6 @@ extern bool use_mips32r2_instructions;
#define TCG_TARGET_DEFAULT_MO 0
#define TCG_TARGET_NEED_LDST_LABELS
+#define TCG_TARGET_NEED_POOL_LABELS
#endif
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index 7a19f8db1d..3b840ecc4c 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -25,6 +25,7 @@
*/
#include "../tcg-ldst.c.inc"
+#include "../tcg-pool.c.inc"
#if HOST_BIG_ENDIAN
# define MIPS_BE 1
@@ -168,9 +169,18 @@ static bool reloc_pc16(tcg_insn_unit *src_rw, const tcg_insn_unit *target)
static bool patch_reloc(tcg_insn_unit *code_ptr, int type,
intptr_t value, intptr_t addend)
{
- tcg_debug_assert(type == R_MIPS_PC16);
- tcg_debug_assert(addend == 0);
- return reloc_pc16(code_ptr, (const tcg_insn_unit *)value);
+ value += addend;
+ switch (type) {
+ case R_MIPS_PC16:
+ return reloc_pc16(code_ptr, (const tcg_insn_unit *)value);
+ case R_MIPS_16:
+ if (value != (int16_t)value) {
+ return false;
+ }
+ *code_ptr = deposit32(*code_ptr, 0, 16, value);
+ return true;
+ }
+ g_assert_not_reached();
}
#define TCG_CT_CONST_ZERO 0x100
@@ -486,6 +496,11 @@ static void tcg_out_nop(TCGContext *s)
tcg_out32(s, 0);
}
+static void tcg_out_nop_fill(tcg_insn_unit *p, int count)
+{
+ memset(p, 0, count * sizeof(tcg_insn_unit));
+}
+
static void tcg_out_dsll(TCGContext *s, TCGReg rd, TCGReg rt, TCGArg sa)
{
tcg_out_opc_sa64(s, OPC_DSLL, OPC_DSLL32, rd, rt, sa);
@@ -543,8 +558,15 @@ static bool tcg_out_movi_two(TCGContext *s, TCGReg ret, tcg_target_long arg)
return false;
}
-static void tcg_out_movi(TCGContext *s, TCGType type,
- TCGReg ret, tcg_target_long arg)
+static void tcg_out_movi_pool(TCGContext *s, TCGReg ret,
+ tcg_target_long arg, TCGReg tbreg)
+{
+ new_pool_label(s, arg, R_MIPS_16, s->code_ptr, tcg_tbrel_diff(s, NULL));
+ tcg_out_opc_imm(s, OPC_LD, ret, tbreg, 0);
+}
+
+static void tcg_out_movi_int(TCGContext *s, TCGType type, TCGReg ret,
+ tcg_target_long arg, TCGReg tbreg)
{
if (TCG_TARGET_REG_BITS == 64 && type == TCG_TYPE_I32) {
arg = (int32_t)arg;
@@ -554,18 +576,17 @@ static void tcg_out_movi(TCGContext *s, TCGType type,
if (tcg_out_movi_two(s, ret, arg)) {
return;
}
+ assert(TCG_TARGET_REG_BITS == 64);
- tcg_out_movi(s, TCG_TYPE_I32, ret, arg >> 31 >> 1);
- if (arg & 0xffff0000ull) {
- tcg_out_dsll(s, ret, ret, 16);
- tcg_out_opc_imm(s, OPC_ORI, ret, ret, arg >> 16);
- tcg_out_dsll(s, ret, ret, 16);
- } else {
- tcg_out_dsll(s, ret, ret, 32);
- }
- if (arg & 0xffff) {
- tcg_out_opc_imm(s, OPC_ORI, ret, ret, arg & 0xffff);
- }
+ /* Otherwise, put 64-bit constants into the constant pool. */
+ tcg_out_movi_pool(s, ret, arg, tbreg);
+}
+
+static void tcg_out_movi(TCGContext *s, TCGType type,
+ TCGReg ret, tcg_target_long arg)
+{
+ TCGReg tbreg = TCG_TARGET_REG_BITS == 64 ? TCG_REG_TB : 0;
+ tcg_out_movi_int(s, type, ret, arg, tbreg);
}
static void tcg_out_ext8s(TCGContext *s, TCGType type, TCGReg rd, TCGReg rs)
@@ -2383,10 +2404,20 @@ static void tcg_target_qemu_prologue(TCGContext *s)
#ifndef CONFIG_SOFTMMU
if (guest_base != (int16_t)guest_base) {
- tcg_out_movi(s, TCG_TYPE_PTR, TCG_GUEST_BASE_REG, guest_base);
+ /*
+ * The function call abi for n32 and n64 will have loaded $25 (t9)
+ * with the address of the prologue, so we can use that instead
+ * of TCG_REG_TB.
+ */
+#if TCG_TARGET_REG_BITS == 64 && !defined(__mips_abicalls)
+# error "Unknown mips abi"
+#endif
+ tcg_out_movi_int(s, TCG_TYPE_PTR, TCG_GUEST_BASE_REG, guest_base,
+ TCG_TARGET_REG_BITS == 64 ? TCG_REG_T9 : 0);
tcg_regset_set_reg(s->reserved_regs, TCG_GUEST_BASE_REG);
}
#endif
+
if (TCG_TARGET_REG_BITS == 64) {
tcg_out_mov(s, TCG_TYPE_PTR, TCG_REG_TB, tcg_target_call_iarg_regs[1]);
}
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 08/23] tcg/mips: Aggressively use the constant pool for n64 calls
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (6 preceding siblings ...)
2023-05-25 18:10 ` [PULL 07/23] tcg/mips: Use the constant pool for 64-bit constants Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 09/23] tcg/mips: Try tb-relative addresses in tcg_out_movi Richard Henderson
` (15 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
Repeated calls to a single helper are common -- especially
the ones for softmmu memory access. Prefer the constant pool
to longer sequences to increase sharing.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 16 +++++++++++++---
1 file changed, 13 insertions(+), 3 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index 3b840ecc4c..068deab8c9 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -1101,9 +1101,19 @@ static void tcg_out_movcond(TCGContext *s, TCGCond cond, TCGReg ret,
static void tcg_out_call_int(TCGContext *s, const tcg_insn_unit *arg, bool tail)
{
- /* Note that the ABI requires the called function's address to be
- loaded into T9, even if a direct branch is in range. */
- tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_T9, (uintptr_t)arg);
+ /*
+ * Note that __mips_abicalls requires the called function's address
+ * to be loaded into $25 (t9), even if a direct branch is in range.
+ *
+ * For n64, always drop the pointer into the constant pool.
+ * We can re-use helper addresses often and do not want any
+ * of the longer sequences tcg_out_movi may try.
+ */
+ if (sizeof(uintptr_t) == 8) {
+ tcg_out_movi_pool(s, TCG_REG_T9, (uintptr_t)arg, TCG_REG_TB);
+ } else {
+ tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_T9, (uintptr_t)arg);
+ }
/* But do try a direct branch, allowing the cpu better insn prefetch. */
if (tail) {
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 09/23] tcg/mips: Try tb-relative addresses in tcg_out_movi
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (7 preceding siblings ...)
2023-05-25 18:10 ` [PULL 08/23] tcg/mips: Aggressively use the constant pool for n64 calls Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 10/23] tcg/mips: Try three insns with shift and add " Richard Henderson
` (14 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
These addresses are often loaded by the qemu_ld/st slow path,
for loading the retaddr value.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index 068deab8c9..9fab424ecc 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -568,6 +568,8 @@ static void tcg_out_movi_pool(TCGContext *s, TCGReg ret,
static void tcg_out_movi_int(TCGContext *s, TCGType type, TCGReg ret,
tcg_target_long arg, TCGReg tbreg)
{
+ tcg_target_long tmp;
+
if (TCG_TARGET_REG_BITS == 64 && type == TCG_TYPE_I32) {
arg = (int32_t)arg;
}
@@ -578,6 +580,17 @@ static void tcg_out_movi_int(TCGContext *s, TCGType type, TCGReg ret,
}
assert(TCG_TARGET_REG_BITS == 64);
+ /* Load addresses within 2GB of TB with 1 or 3 insns. */
+ tmp = tcg_tbrel_diff(s, (void *)arg);
+ if (tmp == (int16_t)tmp) {
+ tcg_out_opc_imm(s, OPC_DADDIU, ret, tbreg, tmp);
+ return;
+ }
+ if (tcg_out_movi_two(s, ret, tmp)) {
+ tcg_out_opc_reg(s, OPC_DADDU, ret, ret, tbreg);
+ return;
+ }
+
/* Otherwise, put 64-bit constants into the constant pool. */
tcg_out_movi_pool(s, ret, arg, tbreg);
}
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 10/23] tcg/mips: Try three insns with shift and add in tcg_out_movi
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (8 preceding siblings ...)
2023-05-25 18:10 ` [PULL 09/23] tcg/mips: Try tb-relative addresses in tcg_out_movi Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 11/23] tcg/mips: Use qemu_build_not_reached for LO/HI_OFF Richard Henderson
` (13 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
These sequences are inexpensive to test. Maxing out at three insns
results in the same space as a load plus the constant pool entry.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 44 +++++++++++++++++++++++++++++++++++++++
1 file changed, 44 insertions(+)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index 9fab424ecc..b86a0679af 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -569,6 +569,7 @@ static void tcg_out_movi_int(TCGContext *s, TCGType type, TCGReg ret,
tcg_target_long arg, TCGReg tbreg)
{
tcg_target_long tmp;
+ int sh, lo;
if (TCG_TARGET_REG_BITS == 64 && type == TCG_TYPE_I32) {
arg = (int32_t)arg;
@@ -591,6 +592,49 @@ static void tcg_out_movi_int(TCGContext *s, TCGType type, TCGReg ret,
return;
}
+ /*
+ * Load bitmasks with a right-shift. This is good for things
+ * like 0x0fff_ffff_ffff_fff0: ADDUI r,0,0xff00 + DSRL r,r,4.
+ * or similarly using LUI. For this to work, bit 31 must be set.
+ */
+ if (arg > 0 && (int32_t)arg < 0) {
+ sh = clz64(arg);
+ if (tcg_out_movi_one(s, ret, arg << sh)) {
+ tcg_out_dsrl(s, ret, ret, sh);
+ return;
+ }
+ }
+
+ /*
+ * Load slightly larger constants using left-shift.
+ * Limit this sequence to 3 insns to avoid too much expansion.
+ */
+ sh = ctz64(arg);
+ if (sh && tcg_out_movi_two(s, ret, arg >> sh)) {
+ tcg_out_dsll(s, ret, ret, sh);
+ return;
+ }
+
+ /*
+ * Load slightly larger constants using left-shift and add/or.
+ * Prefer addi with a negative immediate when that would produce
+ * a larger shift. For this to work, bits 15 and 16 must be set.
+ */
+ lo = arg & 0xffff;
+ if (lo) {
+ if ((arg & 0x18000) == 0x18000) {
+ lo = (int16_t)arg;
+ }
+ tmp = arg - lo;
+ sh = ctz64(tmp);
+ tmp >>= sh;
+ if (tcg_out_movi_one(s, ret, tmp)) {
+ tcg_out_dsll(s, ret, ret, sh);
+ tcg_out_opc_imm(s, lo < 0 ? OPC_DADDIU : OPC_ORI, ret, ret, lo);
+ return;
+ }
+ }
+
/* Otherwise, put 64-bit constants into the constant pool. */
tcg_out_movi_pool(s, ret, arg, tbreg);
}
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 11/23] tcg/mips: Use qemu_build_not_reached for LO/HI_OFF
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (9 preceding siblings ...)
2023-05-25 18:10 ` [PULL 10/23] tcg/mips: Try three insns with shift and add " Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 12/23] tcg/mips: Replace MIPS_BE with HOST_BIG_ENDIAN Richard Henderson
` (12 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel
The new(ish) macro produces a compile-time error instead
of a link-time error.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index b86a0679af..fd92cc30ca 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -37,11 +37,9 @@
# define LO_OFF (MIPS_BE * 4)
# define HI_OFF (4 - LO_OFF)
#else
-/* To assert at compile-time that these values are never used
- for TCG_TARGET_REG_BITS == 64. */
-int link_error(void);
-# define LO_OFF link_error()
-# define HI_OFF link_error()
+/* Assert at compile-time that these values are never used for 64-bit. */
+# define LO_OFF ({ qemu_build_not_reached(); 0; })
+# define HI_OFF ({ qemu_build_not_reached(); 0; })
#endif
#ifdef CONFIG_DEBUG_TCG
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 12/23] tcg/mips: Replace MIPS_BE with HOST_BIG_ENDIAN
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (10 preceding siblings ...)
2023-05-25 18:10 ` [PULL 11/23] tcg/mips: Use qemu_build_not_reached for LO/HI_OFF Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 13/23] disas/riscv: Decode czero.{eqz,nez} Richard Henderson
` (11 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
Since e03b56863d2b, which replaced HOST_WORDS_BIGENDIAN
with HOST_BIG_ENDIAN, there is no need to define a second
symbol which is [0,1].
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/mips/tcg-target.c.inc | 46 +++++++++++++++++----------------------
1 file changed, 20 insertions(+), 26 deletions(-)
diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc
index fd92cc30ca..3274d9aace 100644
--- a/tcg/mips/tcg-target.c.inc
+++ b/tcg/mips/tcg-target.c.inc
@@ -27,14 +27,8 @@
#include "../tcg-ldst.c.inc"
#include "../tcg-pool.c.inc"
-#if HOST_BIG_ENDIAN
-# define MIPS_BE 1
-#else
-# define MIPS_BE 0
-#endif
-
#if TCG_TARGET_REG_BITS == 32
-# define LO_OFF (MIPS_BE * 4)
+# define LO_OFF (HOST_BIG_ENDIAN * 4)
# define HI_OFF (4 - LO_OFF)
#else
/* Assert at compile-time that these values are never used for 64-bit. */
@@ -1439,7 +1433,7 @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg lo, TCGReg hi,
/* Prefer to load from offset 0 first, but allow for overlap. */
if (TCG_TARGET_REG_BITS == 64) {
tcg_out_opc_imm(s, OPC_LD, lo, base, 0);
- } else if (MIPS_BE ? hi != base : lo == base) {
+ } else if (HOST_BIG_ENDIAN ? hi != base : lo == base) {
tcg_out_opc_imm(s, OPC_LW, hi, base, HI_OFF);
tcg_out_opc_imm(s, OPC_LW, lo, base, LO_OFF);
} else {
@@ -1455,10 +1449,10 @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg lo, TCGReg hi,
static void tcg_out_qemu_ld_unalign(TCGContext *s, TCGReg lo, TCGReg hi,
TCGReg base, MemOp opc, TCGType type)
{
- const MIPSInsn lw1 = MIPS_BE ? OPC_LWL : OPC_LWR;
- const MIPSInsn lw2 = MIPS_BE ? OPC_LWR : OPC_LWL;
- const MIPSInsn ld1 = MIPS_BE ? OPC_LDL : OPC_LDR;
- const MIPSInsn ld2 = MIPS_BE ? OPC_LDR : OPC_LDL;
+ const MIPSInsn lw1 = HOST_BIG_ENDIAN ? OPC_LWL : OPC_LWR;
+ const MIPSInsn lw2 = HOST_BIG_ENDIAN ? OPC_LWR : OPC_LWL;
+ const MIPSInsn ld1 = HOST_BIG_ENDIAN ? OPC_LDL : OPC_LDR;
+ const MIPSInsn ld2 = HOST_BIG_ENDIAN ? OPC_LDR : OPC_LDL;
bool sgn = opc & MO_SIGN;
switch (opc & MO_SIZE) {
@@ -1497,10 +1491,10 @@ static void tcg_out_qemu_ld_unalign(TCGContext *s, TCGReg lo, TCGReg hi,
tcg_out_opc_imm(s, ld1, lo, base, 0);
tcg_out_opc_imm(s, ld2, lo, base, 7);
} else {
- tcg_out_opc_imm(s, lw1, MIPS_BE ? hi : lo, base, 0 + 0);
- tcg_out_opc_imm(s, lw2, MIPS_BE ? hi : lo, base, 0 + 3);
- tcg_out_opc_imm(s, lw1, MIPS_BE ? lo : hi, base, 4 + 0);
- tcg_out_opc_imm(s, lw2, MIPS_BE ? lo : hi, base, 4 + 3);
+ tcg_out_opc_imm(s, lw1, HOST_BIG_ENDIAN ? hi : lo, base, 0 + 0);
+ tcg_out_opc_imm(s, lw2, HOST_BIG_ENDIAN ? hi : lo, base, 0 + 3);
+ tcg_out_opc_imm(s, lw1, HOST_BIG_ENDIAN ? lo : hi, base, 4 + 0);
+ tcg_out_opc_imm(s, lw2, HOST_BIG_ENDIAN ? lo : hi, base, 4 + 3);
}
break;
@@ -1550,8 +1544,8 @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg lo, TCGReg hi,
if (TCG_TARGET_REG_BITS == 64) {
tcg_out_opc_imm(s, OPC_SD, lo, base, 0);
} else {
- tcg_out_opc_imm(s, OPC_SW, MIPS_BE ? hi : lo, base, 0);
- tcg_out_opc_imm(s, OPC_SW, MIPS_BE ? lo : hi, base, 4);
+ tcg_out_opc_imm(s, OPC_SW, HOST_BIG_ENDIAN ? hi : lo, base, 0);
+ tcg_out_opc_imm(s, OPC_SW, HOST_BIG_ENDIAN ? lo : hi, base, 4);
}
break;
default:
@@ -1562,10 +1556,10 @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg lo, TCGReg hi,
static void tcg_out_qemu_st_unalign(TCGContext *s, TCGReg lo, TCGReg hi,
TCGReg base, MemOp opc)
{
- const MIPSInsn sw1 = MIPS_BE ? OPC_SWL : OPC_SWR;
- const MIPSInsn sw2 = MIPS_BE ? OPC_SWR : OPC_SWL;
- const MIPSInsn sd1 = MIPS_BE ? OPC_SDL : OPC_SDR;
- const MIPSInsn sd2 = MIPS_BE ? OPC_SDR : OPC_SDL;
+ const MIPSInsn sw1 = HOST_BIG_ENDIAN ? OPC_SWL : OPC_SWR;
+ const MIPSInsn sw2 = HOST_BIG_ENDIAN ? OPC_SWR : OPC_SWL;
+ const MIPSInsn sd1 = HOST_BIG_ENDIAN ? OPC_SDL : OPC_SDR;
+ const MIPSInsn sd2 = HOST_BIG_ENDIAN ? OPC_SDR : OPC_SDL;
switch (opc & MO_SIZE) {
case MO_16:
@@ -1584,10 +1578,10 @@ static void tcg_out_qemu_st_unalign(TCGContext *s, TCGReg lo, TCGReg hi,
tcg_out_opc_imm(s, sd1, lo, base, 0);
tcg_out_opc_imm(s, sd2, lo, base, 7);
} else {
- tcg_out_opc_imm(s, sw1, MIPS_BE ? hi : lo, base, 0 + 0);
- tcg_out_opc_imm(s, sw2, MIPS_BE ? hi : lo, base, 0 + 3);
- tcg_out_opc_imm(s, sw1, MIPS_BE ? lo : hi, base, 4 + 0);
- tcg_out_opc_imm(s, sw2, MIPS_BE ? lo : hi, base, 4 + 3);
+ tcg_out_opc_imm(s, sw1, HOST_BIG_ENDIAN ? hi : lo, base, 0 + 0);
+ tcg_out_opc_imm(s, sw2, HOST_BIG_ENDIAN ? hi : lo, base, 0 + 3);
+ tcg_out_opc_imm(s, sw1, HOST_BIG_ENDIAN ? lo : hi, base, 4 + 0);
+ tcg_out_opc_imm(s, sw2, HOST_BIG_ENDIAN ? lo : hi, base, 4 + 3);
}
break;
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 13/23] disas/riscv: Decode czero.{eqz,nez}
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (11 preceding siblings ...)
2023-05-25 18:10 ` [PULL 12/23] tcg/mips: Replace MIPS_BE with HOST_BIG_ENDIAN Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 14/23] tcg/riscv: Probe for Zba, Zbb, Zicond extensions Richard Henderson
` (10 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
disas/riscv.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/disas/riscv.c b/disas/riscv.c
index e61bda5674..d597161d46 100644
--- a/disas/riscv.c
+++ b/disas/riscv.c
@@ -962,6 +962,8 @@ typedef enum {
rv_op_cm_mvsa01 = 786,
rv_op_cm_jt = 787,
rv_op_cm_jalt = 788,
+ rv_op_czero_eqz = 789,
+ rv_op_czero_nez = 790,
} rv_op;
/* structures */
@@ -2119,6 +2121,8 @@ const rv_opcode_data opcode_data[] = {
{ "cm.mvsa01", rv_codec_zcmp_cm_mv, rv_fmt_rd_rs2, NULL, 0, 0, 0 },
{ "cm.jt", rv_codec_zcmt_jt, rv_fmt_zcmt_index, NULL, 0 },
{ "cm.jalt", rv_codec_zcmt_jt, rv_fmt_zcmt_index, NULL, 0 },
+ { "czero.eqz", rv_codec_r, rv_fmt_rd_rs1_rs2, NULL, 0, 0, 0 },
+ { "czero.nez", rv_codec_r, rv_fmt_rd_rs1_rs2, NULL, 0, 0, 0 },
};
/* CSR names */
@@ -2914,6 +2918,8 @@ static void decode_inst_opcode(rv_decode *dec, rv_isa isa)
case 45: op = rv_op_minu; break;
case 46: op = rv_op_max; break;
case 47: op = rv_op_maxu; break;
+ case 075: op = rv_op_czero_eqz; break;
+ case 077: op = rv_op_czero_nez; break;
case 130: op = rv_op_sh1add; break;
case 132: op = rv_op_sh2add; break;
case 134: op = rv_op_sh3add; break;
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 14/23] tcg/riscv: Probe for Zba, Zbb, Zicond extensions
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (12 preceding siblings ...)
2023-05-25 18:10 ` [PULL 13/23] disas/riscv: Decode czero.{eqz,nez} Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 15/23] tcg/riscv: Support ANDN, ORN, XNOR from Zbb Richard Henderson
` (9 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Define a useful subset of the extensions. Probe for them
via compiler pre-processor feature macros and SIGILL.
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.h | 6 +++
tcg/riscv/tcg-target.c.inc | 96 ++++++++++++++++++++++++++++++++++++++
2 files changed, 102 insertions(+)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index 494c986b49..863ac8ba2f 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -90,6 +90,12 @@ typedef enum {
#define TCG_TARGET_CALL_ARG_I128 TCG_CALL_ARG_NORMAL
#define TCG_TARGET_CALL_RET_I128 TCG_CALL_RET_NORMAL
+#if defined(__riscv_arch_test) && defined(__riscv_zbb)
+# define have_zbb true
+#else
+extern bool have_zbb;
+#endif
+
/* optional instructions */
#define TCG_TARGET_HAS_movcond_i32 0
#define TCG_TARGET_HAS_div_i32 1
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index ff6334980f..eb3e2e9eb0 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -113,6 +113,20 @@ static const int tcg_target_call_iarg_regs[] = {
TCG_REG_A7,
};
+#ifndef have_zbb
+bool have_zbb;
+#endif
+#if defined(__riscv_arch_test) && defined(__riscv_zba)
+# define have_zba true
+#else
+static bool have_zba;
+#endif
+#if defined(__riscv_arch_test) && defined(__riscv_zicond)
+# define have_zicond true
+#else
+static bool have_zicond;
+#endif
+
static TCGReg tcg_target_call_oarg_reg(TCGCallReturnKind kind, int slot)
{
tcg_debug_assert(kind == TCG_CALL_RET_NORMAL);
@@ -234,6 +248,34 @@ typedef enum {
OPC_FENCE = 0x0000000f,
OPC_NOP = OPC_ADDI, /* nop = addi r0,r0,0 */
+
+ /* Zba: Bit manipulation extension, address generation */
+ OPC_ADD_UW = 0x0800003b,
+
+ /* Zbb: Bit manipulation extension, basic bit manipulaton */
+ OPC_ANDN = 0x40007033,
+ OPC_CLZ = 0x60001013,
+ OPC_CLZW = 0x6000101b,
+ OPC_CPOP = 0x60201013,
+ OPC_CPOPW = 0x6020101b,
+ OPC_CTZ = 0x60101013,
+ OPC_CTZW = 0x6010101b,
+ OPC_ORN = 0x40006033,
+ OPC_REV8 = 0x6b805013,
+ OPC_ROL = 0x60001033,
+ OPC_ROLW = 0x6000103b,
+ OPC_ROR = 0x60005033,
+ OPC_RORW = 0x6000503b,
+ OPC_RORI = 0x60005013,
+ OPC_RORIW = 0x6000501b,
+ OPC_SEXT_B = 0x60401013,
+ OPC_SEXT_H = 0x60501013,
+ OPC_XNOR = 0x40004033,
+ OPC_ZEXT_H = 0x0800403b,
+
+ /* Zicond: integer conditional operations */
+ OPC_CZERO_EQZ = 0x0e005033,
+ OPC_CZERO_NEZ = 0x0e007033,
} RISCVInsn;
/*
@@ -1619,8 +1661,62 @@ static void tcg_target_qemu_prologue(TCGContext *s)
tcg_out_opc_imm(s, OPC_JALR, TCG_REG_ZERO, TCG_REG_RA, 0);
}
+static volatile sig_atomic_t got_sigill;
+
+static void sigill_handler(int signo, siginfo_t *si, void *data)
+{
+ /* Skip the faulty instruction */
+ ucontext_t *uc = (ucontext_t *)data;
+ uc->uc_mcontext.__gregs[REG_PC] += 4;
+
+ got_sigill = 1;
+}
+
+static void tcg_target_detect_isa(void)
+{
+#if !defined(have_zba) || !defined(have_zbb) || !defined(have_zicond)
+ /*
+ * TODO: It is expected that this will be determinable via
+ * linux riscv_hwprobe syscall, not yet merged.
+ * In the meantime, test via sigill.
+ */
+
+ struct sigaction sa_old, sa_new;
+
+ memset(&sa_new, 0, sizeof(sa_new));
+ sa_new.sa_flags = SA_SIGINFO;
+ sa_new.sa_sigaction = sigill_handler;
+ sigaction(SIGILL, &sa_new, &sa_old);
+
+#ifndef have_zba
+ /* Probe for Zba: add.uw zero,zero,zero. */
+ got_sigill = 0;
+ asm volatile(".insn r 0x3b, 0, 0x04, zero, zero, zero" : : : "memory");
+ have_zba = !got_sigill;
+#endif
+
+#ifndef have_zbb
+ /* Probe for Zba: andn zero,zero,zero. */
+ got_sigill = 0;
+ asm volatile(".insn r 0x33, 7, 0x20, zero, zero, zero" : : : "memory");
+ have_zbb = !got_sigill;
+#endif
+
+#ifndef have_zicond
+ /* Probe for Zicond: czero.eqz zero,zero,zero. */
+ got_sigill = 0;
+ asm volatile(".insn r 0x33, 5, 0x07, zero, zero, zero" : : : "memory");
+ have_zicond = !got_sigill;
+#endif
+
+ sigaction(SIGILL, &sa_old, NULL);
+#endif
+}
+
static void tcg_target_init(TCGContext *s)
{
+ tcg_target_detect_isa();
+
tcg_target_available_regs[TCG_TYPE_I32] = 0xffffffff;
tcg_target_available_regs[TCG_TYPE_I64] = 0xffffffff;
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 15/23] tcg/riscv: Support ANDN, ORN, XNOR from Zbb
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (13 preceding siblings ...)
2023-05-25 18:10 ` [PULL 14/23] tcg/riscv: Probe for Zba, Zbb, Zicond extensions Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 16/23] tcg/riscv: Support ADD.UW, SEXT.B, SEXT.H, ZEXT.H from Zba+Zbb Richard Henderson
` (8 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target-con-set.h | 1 +
tcg/riscv/tcg-target-con-str.h | 1 +
tcg/riscv/tcg-target.h | 12 +++++-----
tcg/riscv/tcg-target.c.inc | 41 ++++++++++++++++++++++++++++++++++
4 files changed, 49 insertions(+), 6 deletions(-)
diff --git a/tcg/riscv/tcg-target-con-set.h b/tcg/riscv/tcg-target-con-set.h
index d88888d3ac..1a33ece98f 100644
--- a/tcg/riscv/tcg-target-con-set.h
+++ b/tcg/riscv/tcg-target-con-set.h
@@ -15,6 +15,7 @@ C_O0_I2(rZ, rZ)
C_O1_I1(r, r)
C_O1_I2(r, r, ri)
C_O1_I2(r, r, rI)
+C_O1_I2(r, r, rJ)
C_O1_I2(r, rZ, rN)
C_O1_I2(r, rZ, rZ)
C_O2_I4(r, r, rZ, rZ, rM, rM)
diff --git a/tcg/riscv/tcg-target-con-str.h b/tcg/riscv/tcg-target-con-str.h
index 6f1cfb976c..d5c419dff1 100644
--- a/tcg/riscv/tcg-target-con-str.h
+++ b/tcg/riscv/tcg-target-con-str.h
@@ -15,6 +15,7 @@ REGS('r', ALL_GENERAL_REGS)
* CONST(letter, TCG_CT_CONST_* bit set)
*/
CONST('I', TCG_CT_CONST_S12)
+CONST('J', TCG_CT_CONST_J12)
CONST('N', TCG_CT_CONST_N12)
CONST('M', TCG_CT_CONST_M12)
CONST('Z', TCG_CT_CONST_ZERO)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index 863ac8ba2f..9f58d46208 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -120,9 +120,9 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_bswap32_i32 0
#define TCG_TARGET_HAS_not_i32 1
#define TCG_TARGET_HAS_neg_i32 1
-#define TCG_TARGET_HAS_andc_i32 0
-#define TCG_TARGET_HAS_orc_i32 0
-#define TCG_TARGET_HAS_eqv_i32 0
+#define TCG_TARGET_HAS_andc_i32 have_zbb
+#define TCG_TARGET_HAS_orc_i32 have_zbb
+#define TCG_TARGET_HAS_eqv_i32 have_zbb
#define TCG_TARGET_HAS_nand_i32 0
#define TCG_TARGET_HAS_nor_i32 0
#define TCG_TARGET_HAS_clz_i32 0
@@ -154,9 +154,9 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_bswap64_i64 0
#define TCG_TARGET_HAS_not_i64 1
#define TCG_TARGET_HAS_neg_i64 1
-#define TCG_TARGET_HAS_andc_i64 0
-#define TCG_TARGET_HAS_orc_i64 0
-#define TCG_TARGET_HAS_eqv_i64 0
+#define TCG_TARGET_HAS_andc_i64 have_zbb
+#define TCG_TARGET_HAS_orc_i64 have_zbb
+#define TCG_TARGET_HAS_eqv_i64 have_zbb
#define TCG_TARGET_HAS_nand_i64 0
#define TCG_TARGET_HAS_nor_i64 0
#define TCG_TARGET_HAS_clz_i64 0
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index eb3e2e9eb0..edfe4c8f8d 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -138,6 +138,7 @@ static TCGReg tcg_target_call_oarg_reg(TCGCallReturnKind kind, int slot)
#define TCG_CT_CONST_S12 0x200
#define TCG_CT_CONST_N12 0x400
#define TCG_CT_CONST_M12 0x800
+#define TCG_CT_CONST_J12 0x1000
#define ALL_GENERAL_REGS MAKE_64BIT_MASK(0, 32)
@@ -174,6 +175,13 @@ static bool tcg_target_const_match(int64_t val, TCGType type, int ct)
if ((ct & TCG_CT_CONST_M12) && val >= -0x7ff && val <= 0x7ff) {
return 1;
}
+ /*
+ * Inverse of sign extended from 12 bits: ~[-0x800, 0x7ff].
+ * Used to map ANDN back to ANDI, etc.
+ */
+ if ((ct & TCG_CT_CONST_J12) && ~val >= -0x800 && ~val <= 0x7ff) {
+ return 1;
+ }
return 0;
}
@@ -1305,6 +1313,31 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
}
break;
+ case INDEX_op_andc_i32:
+ case INDEX_op_andc_i64:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_ANDI, a0, a1, ~a2);
+ } else {
+ tcg_out_opc_reg(s, OPC_ANDN, a0, a1, a2);
+ }
+ break;
+ case INDEX_op_orc_i32:
+ case INDEX_op_orc_i64:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_ORI, a0, a1, ~a2);
+ } else {
+ tcg_out_opc_reg(s, OPC_ORN, a0, a1, a2);
+ }
+ break;
+ case INDEX_op_eqv_i32:
+ case INDEX_op_eqv_i64:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_XORI, a0, a1, ~a2);
+ } else {
+ tcg_out_opc_reg(s, OPC_XNOR, a0, a1, a2);
+ }
+ break;
+
case INDEX_op_not_i32:
case INDEX_op_not_i64:
tcg_out_opc_imm(s, OPC_XORI, a0, a1, -1);
@@ -1539,6 +1572,14 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_xor_i64:
return C_O1_I2(r, r, rI);
+ case INDEX_op_andc_i32:
+ case INDEX_op_andc_i64:
+ case INDEX_op_orc_i32:
+ case INDEX_op_orc_i64:
+ case INDEX_op_eqv_i32:
+ case INDEX_op_eqv_i64:
+ return C_O1_I2(r, r, rJ);
+
case INDEX_op_sub_i32:
case INDEX_op_sub_i64:
return C_O1_I2(r, rZ, rN);
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 16/23] tcg/riscv: Support ADD.UW, SEXT.B, SEXT.H, ZEXT.H from Zba+Zbb
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (14 preceding siblings ...)
2023-05-25 18:10 ` [PULL 15/23] tcg/riscv: Support ANDN, ORN, XNOR from Zbb Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 17/23] tcg/riscv: Use ADD.UW for guest address generation Richard Henderson
` (7 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.c.inc | 32 ++++++++++++++++++++++++--------
1 file changed, 24 insertions(+), 8 deletions(-)
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index edfe4c8f8d..297119817b 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -593,26 +593,42 @@ static void tcg_out_ext8u(TCGContext *s, TCGReg ret, TCGReg arg)
static void tcg_out_ext16u(TCGContext *s, TCGReg ret, TCGReg arg)
{
- tcg_out_opc_imm(s, OPC_SLLIW, ret, arg, 16);
- tcg_out_opc_imm(s, OPC_SRLIW, ret, ret, 16);
+ if (have_zbb) {
+ tcg_out_opc_reg(s, OPC_ZEXT_H, ret, arg, TCG_REG_ZERO);
+ } else {
+ tcg_out_opc_imm(s, OPC_SLLIW, ret, arg, 16);
+ tcg_out_opc_imm(s, OPC_SRLIW, ret, ret, 16);
+ }
}
static void tcg_out_ext32u(TCGContext *s, TCGReg ret, TCGReg arg)
{
- tcg_out_opc_imm(s, OPC_SLLI, ret, arg, 32);
- tcg_out_opc_imm(s, OPC_SRLI, ret, ret, 32);
+ if (have_zba) {
+ tcg_out_opc_reg(s, OPC_ADD_UW, ret, arg, TCG_REG_ZERO);
+ } else {
+ tcg_out_opc_imm(s, OPC_SLLI, ret, arg, 32);
+ tcg_out_opc_imm(s, OPC_SRLI, ret, ret, 32);
+ }
}
static void tcg_out_ext8s(TCGContext *s, TCGType type, TCGReg ret, TCGReg arg)
{
- tcg_out_opc_imm(s, OPC_SLLIW, ret, arg, 24);
- tcg_out_opc_imm(s, OPC_SRAIW, ret, ret, 24);
+ if (have_zbb) {
+ tcg_out_opc_imm(s, OPC_SEXT_B, ret, arg, 0);
+ } else {
+ tcg_out_opc_imm(s, OPC_SLLIW, ret, arg, 24);
+ tcg_out_opc_imm(s, OPC_SRAIW, ret, ret, 24);
+ }
}
static void tcg_out_ext16s(TCGContext *s, TCGType type, TCGReg ret, TCGReg arg)
{
- tcg_out_opc_imm(s, OPC_SLLIW, ret, arg, 16);
- tcg_out_opc_imm(s, OPC_SRAIW, ret, ret, 16);
+ if (have_zbb) {
+ tcg_out_opc_imm(s, OPC_SEXT_H, ret, arg, 0);
+ } else {
+ tcg_out_opc_imm(s, OPC_SLLIW, ret, arg, 16);
+ tcg_out_opc_imm(s, OPC_SRAIW, ret, ret, 16);
+ }
}
static void tcg_out_ext32s(TCGContext *s, TCGReg ret, TCGReg arg)
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 17/23] tcg/riscv: Use ADD.UW for guest address generation
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (15 preceding siblings ...)
2023-05-25 18:10 ` [PULL 16/23] tcg/riscv: Support ADD.UW, SEXT.B, SEXT.H, ZEXT.H from Zba+Zbb Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 18/23] tcg/riscv: Support rotates from Zbb Richard Henderson
` (6 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
The instruction is a combined zero-extend and add.
Use it for exactly that.
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.c.inc | 33 ++++++++++++++++++++++-----------
1 file changed, 22 insertions(+), 11 deletions(-)
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index 297119817b..2fdd450da3 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -1038,14 +1038,18 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase,
tcg_out_opc_branch(s, OPC_BNE, TCG_REG_TMP0, TCG_REG_TMP1, 0);
/* TLB Hit - translate address using addend. */
- addr_adj = addr_reg;
- if (TARGET_LONG_BITS == 32) {
- addr_adj = TCG_REG_TMP0;
- tcg_out_ext32u(s, addr_adj, addr_reg);
+ if (TARGET_LONG_BITS == 64) {
+ tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP0, addr_reg, TCG_REG_TMP2);
+ } else if (have_zba) {
+ tcg_out_opc_reg(s, OPC_ADD_UW, TCG_REG_TMP0, addr_reg, TCG_REG_TMP2);
+ } else {
+ tcg_out_ext32u(s, TCG_REG_TMP0, addr_reg);
+ tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP0, TCG_REG_TMP0, TCG_REG_TMP2);
}
- tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP0, TCG_REG_TMP2, addr_adj);
*pbase = TCG_REG_TMP0;
#else
+ TCGReg base;
+
if (a_mask) {
ldst = new_ldst_label(s);
ldst->is_ld = is_ld;
@@ -1060,14 +1064,21 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase,
tcg_out_opc_branch(s, OPC_BNE, TCG_REG_TMP1, TCG_REG_ZERO, 0);
}
- TCGReg base = addr_reg;
- if (TARGET_LONG_BITS == 32) {
- tcg_out_ext32u(s, TCG_REG_TMP0, base);
- base = TCG_REG_TMP0;
- }
if (guest_base != 0) {
- tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP0, TCG_GUEST_BASE_REG, base);
base = TCG_REG_TMP0;
+ if (TARGET_LONG_BITS == 64) {
+ tcg_out_opc_reg(s, OPC_ADD, base, addr_reg, TCG_GUEST_BASE_REG);
+ } else if (have_zba) {
+ tcg_out_opc_reg(s, OPC_ADD_UW, base, addr_reg, TCG_GUEST_BASE_REG);
+ } else {
+ tcg_out_ext32u(s, base, addr_reg);
+ tcg_out_opc_reg(s, OPC_ADD, base, base, TCG_GUEST_BASE_REG);
+ }
+ } else if (TARGET_LONG_BITS == 64) {
+ base = addr_reg;
+ } else {
+ base = TCG_REG_TMP0;
+ tcg_out_ext32u(s, base, addr_reg);
}
*pbase = base;
#endif
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 18/23] tcg/riscv: Support rotates from Zbb
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (16 preceding siblings ...)
2023-05-25 18:10 ` [PULL 17/23] tcg/riscv: Use ADD.UW for guest address generation Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 19/23] tcg/riscv: Support REV8 " Richard Henderson
` (5 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.h | 4 ++--
tcg/riscv/tcg-target.c.inc | 34 ++++++++++++++++++++++++++++++++++
2 files changed, 36 insertions(+), 2 deletions(-)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index 9f58d46208..317d385924 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -101,7 +101,7 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_div_i32 1
#define TCG_TARGET_HAS_rem_i32 1
#define TCG_TARGET_HAS_div2_i32 0
-#define TCG_TARGET_HAS_rot_i32 0
+#define TCG_TARGET_HAS_rot_i32 have_zbb
#define TCG_TARGET_HAS_deposit_i32 0
#define TCG_TARGET_HAS_extract_i32 0
#define TCG_TARGET_HAS_sextract_i32 0
@@ -136,7 +136,7 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_div_i64 1
#define TCG_TARGET_HAS_rem_i64 1
#define TCG_TARGET_HAS_div2_i64 0
-#define TCG_TARGET_HAS_rot_i64 0
+#define TCG_TARGET_HAS_rot_i64 have_zbb
#define TCG_TARGET_HAS_deposit_i64 0
#define TCG_TARGET_HAS_extract_i64 0
#define TCG_TARGET_HAS_sextract_i64 0
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index 2fdd450da3..cc96425413 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -1457,6 +1457,36 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
}
break;
+ case INDEX_op_rotl_i32:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_RORIW, a0, a1, -a2 & 0x1f);
+ } else {
+ tcg_out_opc_reg(s, OPC_ROLW, a0, a1, a2);
+ }
+ break;
+ case INDEX_op_rotl_i64:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_RORI, a0, a1, -a2 & 0x3f);
+ } else {
+ tcg_out_opc_reg(s, OPC_ROL, a0, a1, a2);
+ }
+ break;
+
+ case INDEX_op_rotr_i32:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_RORIW, a0, a1, a2 & 0x1f);
+ } else {
+ tcg_out_opc_reg(s, OPC_RORW, a0, a1, a2);
+ }
+ break;
+ case INDEX_op_rotr_i64:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_RORI, a0, a1, a2 & 0x3f);
+ } else {
+ tcg_out_opc_reg(s, OPC_ROR, a0, a1, a2);
+ }
+ break;
+
case INDEX_op_add2_i32:
tcg_out_addsub2(s, a0, a1, a2, args[3], args[4], args[5],
const_args[4], const_args[5], false, true);
@@ -1632,9 +1662,13 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_shl_i32:
case INDEX_op_shr_i32:
case INDEX_op_sar_i32:
+ case INDEX_op_rotl_i32:
+ case INDEX_op_rotr_i32:
case INDEX_op_shl_i64:
case INDEX_op_shr_i64:
case INDEX_op_sar_i64:
+ case INDEX_op_rotl_i64:
+ case INDEX_op_rotr_i64:
return C_O1_I2(r, r, ri);
case INDEX_op_brcond_i32:
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 19/23] tcg/riscv: Support REV8 from Zbb
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (17 preceding siblings ...)
2023-05-25 18:10 ` [PULL 18/23] tcg/riscv: Support rotates from Zbb Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 20/23] tcg/riscv: Support CPOP " Richard Henderson
` (4 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.h | 10 +++++-----
tcg/riscv/tcg-target.c.inc | 29 +++++++++++++++++++++++++++++
2 files changed, 34 insertions(+), 5 deletions(-)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index 317d385924..8e327afc3a 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -116,8 +116,8 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_ext16s_i32 1
#define TCG_TARGET_HAS_ext8u_i32 1
#define TCG_TARGET_HAS_ext16u_i32 1
-#define TCG_TARGET_HAS_bswap16_i32 0
-#define TCG_TARGET_HAS_bswap32_i32 0
+#define TCG_TARGET_HAS_bswap16_i32 have_zbb
+#define TCG_TARGET_HAS_bswap32_i32 have_zbb
#define TCG_TARGET_HAS_not_i32 1
#define TCG_TARGET_HAS_neg_i32 1
#define TCG_TARGET_HAS_andc_i32 have_zbb
@@ -149,9 +149,9 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_ext8u_i64 1
#define TCG_TARGET_HAS_ext16u_i64 1
#define TCG_TARGET_HAS_ext32u_i64 1
-#define TCG_TARGET_HAS_bswap16_i64 0
-#define TCG_TARGET_HAS_bswap32_i64 0
-#define TCG_TARGET_HAS_bswap64_i64 0
+#define TCG_TARGET_HAS_bswap16_i64 have_zbb
+#define TCG_TARGET_HAS_bswap32_i64 have_zbb
+#define TCG_TARGET_HAS_bswap64_i64 have_zbb
#define TCG_TARGET_HAS_not_i64 1
#define TCG_TARGET_HAS_neg_i64 1
#define TCG_TARGET_HAS_andc_i64 have_zbb
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index cc96425413..cb4afb4733 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -1487,6 +1487,30 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
}
break;
+ case INDEX_op_bswap64_i64:
+ tcg_out_opc_imm(s, OPC_REV8, a0, a1, 0);
+ break;
+ case INDEX_op_bswap32_i32:
+ a2 = 0;
+ /* fall through */
+ case INDEX_op_bswap32_i64:
+ tcg_out_opc_imm(s, OPC_REV8, a0, a1, 0);
+ if (a2 & TCG_BSWAP_OZ) {
+ tcg_out_opc_imm(s, OPC_SRLI, a0, a0, 32);
+ } else {
+ tcg_out_opc_imm(s, OPC_SRAI, a0, a0, 32);
+ }
+ break;
+ case INDEX_op_bswap16_i64:
+ case INDEX_op_bswap16_i32:
+ tcg_out_opc_imm(s, OPC_REV8, a0, a1, 0);
+ if (a2 & TCG_BSWAP_OZ) {
+ tcg_out_opc_imm(s, OPC_SRLI, a0, a0, 48);
+ } else {
+ tcg_out_opc_imm(s, OPC_SRAI, a0, a0, 48);
+ }
+ break;
+
case INDEX_op_add2_i32:
tcg_out_addsub2(s, a0, a1, a2, args[3], args[4], args[5],
const_args[4], const_args[5], false, true);
@@ -1608,6 +1632,11 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_extrl_i64_i32:
case INDEX_op_extrh_i64_i32:
case INDEX_op_ext_i32_i64:
+ case INDEX_op_bswap16_i32:
+ case INDEX_op_bswap32_i32:
+ case INDEX_op_bswap16_i64:
+ case INDEX_op_bswap32_i64:
+ case INDEX_op_bswap64_i64:
return C_O1_I1(r, r);
case INDEX_op_st8_i32:
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 20/23] tcg/riscv: Support CPOP from Zbb
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (18 preceding siblings ...)
2023-05-25 18:10 ` [PULL 19/23] tcg/riscv: Support REV8 " Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 21/23] tcg/riscv: Improve setcond expansion Richard Henderson
` (3 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.h | 4 ++--
tcg/riscv/tcg-target.c.inc | 9 +++++++++
2 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index 8e327afc3a..e0b23006c4 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -127,7 +127,7 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_nor_i32 0
#define TCG_TARGET_HAS_clz_i32 0
#define TCG_TARGET_HAS_ctz_i32 0
-#define TCG_TARGET_HAS_ctpop_i32 0
+#define TCG_TARGET_HAS_ctpop_i32 have_zbb
#define TCG_TARGET_HAS_brcond2 1
#define TCG_TARGET_HAS_setcond2 1
#define TCG_TARGET_HAS_qemu_st8_i32 0
@@ -161,7 +161,7 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_nor_i64 0
#define TCG_TARGET_HAS_clz_i64 0
#define TCG_TARGET_HAS_ctz_i64 0
-#define TCG_TARGET_HAS_ctpop_i64 0
+#define TCG_TARGET_HAS_ctpop_i64 have_zbb
#define TCG_TARGET_HAS_add2_i64 1
#define TCG_TARGET_HAS_sub2_i64 1
#define TCG_TARGET_HAS_mulu2_i64 0
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index cb4afb4733..05ea9fead8 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -1511,6 +1511,13 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
}
break;
+ case INDEX_op_ctpop_i32:
+ tcg_out_opc_imm(s, OPC_CPOPW, a0, a1, 0);
+ break;
+ case INDEX_op_ctpop_i64:
+ tcg_out_opc_imm(s, OPC_CPOP, a0, a1, 0);
+ break;
+
case INDEX_op_add2_i32:
tcg_out_addsub2(s, a0, a1, a2, args[3], args[4], args[5],
const_args[4], const_args[5], false, true);
@@ -1637,6 +1644,8 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_bswap16_i64:
case INDEX_op_bswap32_i64:
case INDEX_op_bswap64_i64:
+ case INDEX_op_ctpop_i32:
+ case INDEX_op_ctpop_i64:
return C_O1_I1(r, r);
case INDEX_op_st8_i32:
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 21/23] tcg/riscv: Improve setcond expansion
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (19 preceding siblings ...)
2023-05-25 18:10 ` [PULL 20/23] tcg/riscv: Support CPOP " Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 22/23] tcg/riscv: Implement movcond Richard Henderson
` (2 subsequent siblings)
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Split out a helper function, tcg_out_setcond_int, which does not
always produce the complete boolean result, but returns a set of
flags to do so.
Based on 21af16198425, the same improvement for loongarch64.
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.c.inc | 164 +++++++++++++++++++++++++++----------
1 file changed, 121 insertions(+), 43 deletions(-)
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index 05ea9fead8..db328ddc2d 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -812,50 +812,128 @@ static void tcg_out_brcond(TCGContext *s, TCGCond cond, TCGReg arg1,
tcg_out_opc_branch(s, op, arg1, arg2, 0);
}
-static void tcg_out_setcond(TCGContext *s, TCGCond cond, TCGReg ret,
- TCGReg arg1, TCGReg arg2)
+#define SETCOND_INV TCG_TARGET_NB_REGS
+#define SETCOND_NEZ (SETCOND_INV << 1)
+#define SETCOND_FLAGS (SETCOND_INV | SETCOND_NEZ)
+
+static int tcg_out_setcond_int(TCGContext *s, TCGCond cond, TCGReg ret,
+ TCGReg arg1, tcg_target_long arg2, bool c2)
{
+ int flags = 0;
+
switch (cond) {
- case TCG_COND_EQ:
- tcg_out_opc_reg(s, OPC_SUB, ret, arg1, arg2);
- tcg_out_opc_imm(s, OPC_SLTIU, ret, ret, 1);
- break;
- case TCG_COND_NE:
- tcg_out_opc_reg(s, OPC_SUB, ret, arg1, arg2);
- tcg_out_opc_reg(s, OPC_SLTU, ret, TCG_REG_ZERO, ret);
- break;
- case TCG_COND_LT:
- tcg_out_opc_reg(s, OPC_SLT, ret, arg1, arg2);
- break;
- case TCG_COND_GE:
- tcg_out_opc_reg(s, OPC_SLT, ret, arg1, arg2);
- tcg_out_opc_imm(s, OPC_XORI, ret, ret, 1);
- break;
- case TCG_COND_LE:
- tcg_out_opc_reg(s, OPC_SLT, ret, arg2, arg1);
- tcg_out_opc_imm(s, OPC_XORI, ret, ret, 1);
- break;
- case TCG_COND_GT:
- tcg_out_opc_reg(s, OPC_SLT, ret, arg2, arg1);
- break;
- case TCG_COND_LTU:
- tcg_out_opc_reg(s, OPC_SLTU, ret, arg1, arg2);
- break;
- case TCG_COND_GEU:
- tcg_out_opc_reg(s, OPC_SLTU, ret, arg1, arg2);
- tcg_out_opc_imm(s, OPC_XORI, ret, ret, 1);
- break;
- case TCG_COND_LEU:
- tcg_out_opc_reg(s, OPC_SLTU, ret, arg2, arg1);
- tcg_out_opc_imm(s, OPC_XORI, ret, ret, 1);
- break;
- case TCG_COND_GTU:
- tcg_out_opc_reg(s, OPC_SLTU, ret, arg2, arg1);
+ case TCG_COND_EQ: /* -> NE */
+ case TCG_COND_GE: /* -> LT */
+ case TCG_COND_GEU: /* -> LTU */
+ case TCG_COND_GT: /* -> LE */
+ case TCG_COND_GTU: /* -> LEU */
+ cond = tcg_invert_cond(cond);
+ flags ^= SETCOND_INV;
break;
default:
- g_assert_not_reached();
- break;
- }
+ break;
+ }
+
+ switch (cond) {
+ case TCG_COND_LE:
+ case TCG_COND_LEU:
+ /*
+ * If we have a constant input, the most efficient way to implement
+ * LE is by adding 1 and using LT. Watch out for wrap around for LEU.
+ * We don't need to care for this for LE because the constant input
+ * is constrained to signed 12-bit, and 0x800 is representable in the
+ * temporary register.
+ */
+ if (c2) {
+ if (cond == TCG_COND_LEU) {
+ /* unsigned <= -1 is true */
+ if (arg2 == -1) {
+ tcg_out_movi(s, TCG_TYPE_REG, ret, !(flags & SETCOND_INV));
+ return ret;
+ }
+ cond = TCG_COND_LTU;
+ } else {
+ cond = TCG_COND_LT;
+ }
+ tcg_debug_assert(arg2 <= 0x7ff);
+ if (++arg2 == 0x800) {
+ tcg_out_movi(s, TCG_TYPE_REG, TCG_REG_TMP0, arg2);
+ arg2 = TCG_REG_TMP0;
+ c2 = false;
+ }
+ } else {
+ TCGReg tmp = arg2;
+ arg2 = arg1;
+ arg1 = tmp;
+ cond = tcg_swap_cond(cond); /* LE -> GE */
+ cond = tcg_invert_cond(cond); /* GE -> LT */
+ flags ^= SETCOND_INV;
+ }
+ break;
+ default:
+ break;
+ }
+
+ switch (cond) {
+ case TCG_COND_NE:
+ flags |= SETCOND_NEZ;
+ if (!c2) {
+ tcg_out_opc_reg(s, OPC_XOR, ret, arg1, arg2);
+ } else if (arg2 == 0) {
+ ret = arg1;
+ } else {
+ tcg_out_opc_imm(s, OPC_XORI, ret, arg1, arg2);
+ }
+ break;
+
+ case TCG_COND_LT:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_SLTI, ret, arg1, arg2);
+ } else {
+ tcg_out_opc_reg(s, OPC_SLT, ret, arg1, arg2);
+ }
+ break;
+
+ case TCG_COND_LTU:
+ if (c2) {
+ tcg_out_opc_imm(s, OPC_SLTIU, ret, arg1, arg2);
+ } else {
+ tcg_out_opc_reg(s, OPC_SLTU, ret, arg1, arg2);
+ }
+ break;
+
+ default:
+ g_assert_not_reached();
+ }
+
+ return ret | flags;
+}
+
+static void tcg_out_setcond(TCGContext *s, TCGCond cond, TCGReg ret,
+ TCGReg arg1, tcg_target_long arg2, bool c2)
+{
+ int tmpflags = tcg_out_setcond_int(s, cond, ret, arg1, arg2, c2);
+
+ if (tmpflags != ret) {
+ TCGReg tmp = tmpflags & ~SETCOND_FLAGS;
+
+ switch (tmpflags & SETCOND_FLAGS) {
+ case SETCOND_INV:
+ /* Intermediate result is boolean: simply invert. */
+ tcg_out_opc_imm(s, OPC_XORI, ret, tmp, 1);
+ break;
+ case SETCOND_NEZ:
+ /* Intermediate result is zero/non-zero: test != 0. */
+ tcg_out_opc_reg(s, OPC_SLTU, ret, TCG_REG_ZERO, tmp);
+ break;
+ case SETCOND_NEZ | SETCOND_INV:
+ /* Intermediate result is zero/non-zero: test == 0. */
+ tcg_out_opc_imm(s, OPC_SLTIU, ret, tmp, 1);
+ break;
+ default:
+ g_assert_not_reached();
+ }
+ }
}
static void tcg_out_call_int(TCGContext *s, const tcg_insn_unit *arg, bool tail)
@@ -1542,7 +1620,7 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
case INDEX_op_setcond_i32:
case INDEX_op_setcond_i64:
- tcg_out_setcond(s, args[3], a0, a1, a2);
+ tcg_out_setcond(s, args[3], a0, a1, a2, c2);
break;
case INDEX_op_qemu_ld_a32_i32:
@@ -1665,6 +1743,8 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_and_i64:
case INDEX_op_or_i64:
case INDEX_op_xor_i64:
+ case INDEX_op_setcond_i32:
+ case INDEX_op_setcond_i64:
return C_O1_I2(r, r, rI);
case INDEX_op_andc_i32:
@@ -1686,7 +1766,6 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_divu_i32:
case INDEX_op_rem_i32:
case INDEX_op_remu_i32:
- case INDEX_op_setcond_i32:
case INDEX_op_mul_i64:
case INDEX_op_mulsh_i64:
case INDEX_op_muluh_i64:
@@ -1694,7 +1773,6 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_divu_i64:
case INDEX_op_rem_i64:
case INDEX_op_remu_i64:
- case INDEX_op_setcond_i64:
return C_O1_I2(r, rZ, rZ);
case INDEX_op_shl_i32:
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 22/23] tcg/riscv: Implement movcond
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (20 preceding siblings ...)
2023-05-25 18:10 ` [PULL 21/23] tcg/riscv: Improve setcond expansion Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 18:10 ` [PULL 23/23] tcg/riscv: Support CTZ, CLZ from Zbb Richard Henderson
2023-05-25 19:32 ` [PULL 00/23] tcg patch queue Richard Henderson
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Implement with and without Zicond. Without Zicond, we were letting
the middle-end expand to a 5 insn sequence; better to use a branch
over a single insn.
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target-con-set.h | 1 +
tcg/riscv/tcg-target.h | 4 +-
tcg/riscv/tcg-target.c.inc | 139 ++++++++++++++++++++++++++++++++-
3 files changed, 141 insertions(+), 3 deletions(-)
diff --git a/tcg/riscv/tcg-target-con-set.h b/tcg/riscv/tcg-target-con-set.h
index 1a33ece98f..a5cadd303f 100644
--- a/tcg/riscv/tcg-target-con-set.h
+++ b/tcg/riscv/tcg-target-con-set.h
@@ -18,4 +18,5 @@ C_O1_I2(r, r, rI)
C_O1_I2(r, r, rJ)
C_O1_I2(r, rZ, rN)
C_O1_I2(r, rZ, rZ)
+C_O1_I4(r, r, rI, rM, rM)
C_O2_I4(r, r, rZ, rZ, rM, rM)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index e0b23006c4..e9e84be9a5 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -97,7 +97,7 @@ extern bool have_zbb;
#endif
/* optional instructions */
-#define TCG_TARGET_HAS_movcond_i32 0
+#define TCG_TARGET_HAS_movcond_i32 1
#define TCG_TARGET_HAS_div_i32 1
#define TCG_TARGET_HAS_rem_i32 1
#define TCG_TARGET_HAS_div2_i32 0
@@ -132,7 +132,7 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_setcond2 1
#define TCG_TARGET_HAS_qemu_st8_i32 0
-#define TCG_TARGET_HAS_movcond_i64 0
+#define TCG_TARGET_HAS_movcond_i64 1
#define TCG_TARGET_HAS_div_i64 1
#define TCG_TARGET_HAS_rem_i64 1
#define TCG_TARGET_HAS_div2_i64 0
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index db328ddc2d..811b84d152 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -169,7 +169,7 @@ static bool tcg_target_const_match(int64_t val, TCGType type, int ct)
}
/*
* Sign extended from 12 bits, +/- matching: [-0x7ff, 0x7ff].
- * Used by addsub2, which may need the negative operation,
+ * Used by addsub2 and movcond, which may need the negative value,
* and requires the modified constant to be representable.
*/
if ((ct & TCG_CT_CONST_M12) && val >= -0x7ff && val <= 0x7ff) {
@@ -936,6 +936,133 @@ static void tcg_out_setcond(TCGContext *s, TCGCond cond, TCGReg ret,
}
}
+static void tcg_out_movcond_zicond(TCGContext *s, TCGReg ret, TCGReg test_ne,
+ int val1, bool c_val1,
+ int val2, bool c_val2)
+{
+ if (val1 == 0) {
+ if (c_val2) {
+ tcg_out_movi(s, TCG_TYPE_REG, TCG_REG_TMP1, val2);
+ val2 = TCG_REG_TMP1;
+ }
+ tcg_out_opc_reg(s, OPC_CZERO_NEZ, ret, val2, test_ne);
+ return;
+ }
+
+ if (val2 == 0) {
+ if (c_val1) {
+ tcg_out_movi(s, TCG_TYPE_REG, TCG_REG_TMP1, val1);
+ val1 = TCG_REG_TMP1;
+ }
+ tcg_out_opc_reg(s, OPC_CZERO_EQZ, ret, val1, test_ne);
+ return;
+ }
+
+ if (c_val2) {
+ if (c_val1) {
+ tcg_out_movi(s, TCG_TYPE_REG, TCG_REG_TMP1, val1 - val2);
+ } else {
+ tcg_out_opc_imm(s, OPC_ADDI, TCG_REG_TMP1, val1, -val2);
+ }
+ tcg_out_opc_reg(s, OPC_CZERO_EQZ, ret, TCG_REG_TMP1, test_ne);
+ tcg_out_opc_imm(s, OPC_ADDI, ret, ret, val2);
+ return;
+ }
+
+ if (c_val1) {
+ tcg_out_opc_imm(s, OPC_ADDI, TCG_REG_TMP1, val2, -val1);
+ tcg_out_opc_reg(s, OPC_CZERO_NEZ, ret, TCG_REG_TMP1, test_ne);
+ tcg_out_opc_imm(s, OPC_ADDI, ret, ret, val1);
+ return;
+ }
+
+ tcg_out_opc_reg(s, OPC_CZERO_NEZ, TCG_REG_TMP1, val2, test_ne);
+ tcg_out_opc_reg(s, OPC_CZERO_EQZ, TCG_REG_TMP0, val1, test_ne);
+ tcg_out_opc_reg(s, OPC_OR, ret, TCG_REG_TMP0, TCG_REG_TMP1);
+}
+
+static void tcg_out_movcond_br1(TCGContext *s, TCGCond cond, TCGReg ret,
+ TCGReg cmp1, TCGReg cmp2,
+ int val, bool c_val)
+{
+ RISCVInsn op;
+ int disp = 8;
+
+ tcg_debug_assert((unsigned)cond < ARRAY_SIZE(tcg_brcond_to_riscv));
+ op = tcg_brcond_to_riscv[cond].op;
+ tcg_debug_assert(op != 0);
+
+ if (tcg_brcond_to_riscv[cond].swap) {
+ tcg_out_opc_branch(s, op, cmp2, cmp1, disp);
+ } else {
+ tcg_out_opc_branch(s, op, cmp1, cmp2, disp);
+ }
+ if (c_val) {
+ tcg_out_opc_imm(s, OPC_ADDI, ret, TCG_REG_ZERO, val);
+ } else {
+ tcg_out_opc_imm(s, OPC_ADDI, ret, val, 0);
+ }
+}
+
+static void tcg_out_movcond_br2(TCGContext *s, TCGCond cond, TCGReg ret,
+ TCGReg cmp1, TCGReg cmp2,
+ int val1, bool c_val1,
+ int val2, bool c_val2)
+{
+ TCGReg tmp;
+
+ /* TCG optimizer reorders to prefer ret matching val2. */
+ if (!c_val2 && ret == val2) {
+ cond = tcg_invert_cond(cond);
+ tcg_out_movcond_br1(s, cond, ret, cmp1, cmp2, val1, c_val1);
+ return;
+ }
+
+ if (!c_val1 && ret == val1) {
+ tcg_out_movcond_br1(s, cond, ret, cmp1, cmp2, val2, c_val2);
+ return;
+ }
+
+ tmp = (ret == cmp1 || ret == cmp2 ? TCG_REG_TMP1 : ret);
+ if (c_val1) {
+ tcg_out_movi(s, TCG_TYPE_REG, tmp, val1);
+ } else {
+ tcg_out_mov(s, TCG_TYPE_REG, tmp, val1);
+ }
+ tcg_out_movcond_br1(s, cond, tmp, cmp1, cmp2, val2, c_val2);
+ tcg_out_mov(s, TCG_TYPE_REG, ret, tmp);
+}
+
+static void tcg_out_movcond(TCGContext *s, TCGCond cond, TCGReg ret,
+ TCGReg cmp1, int cmp2, bool c_cmp2,
+ TCGReg val1, bool c_val1,
+ TCGReg val2, bool c_val2)
+{
+ int tmpflags;
+ TCGReg t;
+
+ if (!have_zicond && (!c_cmp2 || cmp2 == 0)) {
+ tcg_out_movcond_br2(s, cond, ret, cmp1, cmp2,
+ val1, c_val1, val2, c_val2);
+ return;
+ }
+
+ tmpflags = tcg_out_setcond_int(s, cond, TCG_REG_TMP0, cmp1, cmp2, c_cmp2);
+ t = tmpflags & ~SETCOND_FLAGS;
+
+ if (have_zicond) {
+ if (tmpflags & SETCOND_INV) {
+ tcg_out_movcond_zicond(s, ret, t, val2, c_val2, val1, c_val1);
+ } else {
+ tcg_out_movcond_zicond(s, ret, t, val1, c_val1, val2, c_val2);
+ }
+ } else {
+ cond = tmpflags & SETCOND_INV ? TCG_COND_EQ : TCG_COND_NE;
+ tcg_out_movcond_br2(s, cond, ret, t, TCG_REG_ZERO,
+ val1, c_val1, val2, c_val2);
+ }
+}
+
static void tcg_out_call_int(TCGContext *s, const tcg_insn_unit *arg, bool tail)
{
TCGReg link = tail ? TCG_REG_ZERO : TCG_REG_RA;
@@ -1623,6 +1750,12 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
tcg_out_setcond(s, args[3], a0, a1, a2, c2);
break;
+ case INDEX_op_movcond_i32:
+ case INDEX_op_movcond_i64:
+ tcg_out_movcond(s, args[5], a0, a1, a2, c2,
+ args[3], const_args[3], args[4], const_args[4]);
+ break;
+
case INDEX_op_qemu_ld_a32_i32:
case INDEX_op_qemu_ld_a64_i32:
tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I32);
@@ -1791,6 +1924,10 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_brcond_i64:
return C_O0_I2(rZ, rZ);
+ case INDEX_op_movcond_i32:
+ case INDEX_op_movcond_i64:
+ return C_O1_I4(r, r, rI, rM, rM);
+
case INDEX_op_add2_i32:
case INDEX_op_add2_i64:
case INDEX_op_sub2_i32:
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [PULL 23/23] tcg/riscv: Support CTZ, CLZ from Zbb
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (21 preceding siblings ...)
2023-05-25 18:10 ` [PULL 22/23] tcg/riscv: Implement movcond Richard Henderson
@ 2023-05-25 18:10 ` Richard Henderson
2023-05-25 19:32 ` [PULL 00/23] tcg patch queue Richard Henderson
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 18:10 UTC (permalink / raw)
To: qemu-devel; +Cc: Alistair Francis, Daniel Henrique Barboza
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target-con-set.h | 1 +
tcg/riscv/tcg-target.h | 8 ++++----
tcg/riscv/tcg-target.c.inc | 35 ++++++++++++++++++++++++++++++++++
3 files changed, 40 insertions(+), 4 deletions(-)
diff --git a/tcg/riscv/tcg-target-con-set.h b/tcg/riscv/tcg-target-con-set.h
index a5cadd303f..aac5ceee2b 100644
--- a/tcg/riscv/tcg-target-con-set.h
+++ b/tcg/riscv/tcg-target-con-set.h
@@ -18,5 +18,6 @@ C_O1_I2(r, r, rI)
C_O1_I2(r, r, rJ)
C_O1_I2(r, rZ, rN)
C_O1_I2(r, rZ, rZ)
+C_N1_I2(r, r, rM)
C_O1_I4(r, r, rI, rM, rM)
C_O2_I4(r, r, rZ, rZ, rM, rM)
diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h
index e9e84be9a5..62fe61af7b 100644
--- a/tcg/riscv/tcg-target.h
+++ b/tcg/riscv/tcg-target.h
@@ -125,8 +125,8 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_eqv_i32 have_zbb
#define TCG_TARGET_HAS_nand_i32 0
#define TCG_TARGET_HAS_nor_i32 0
-#define TCG_TARGET_HAS_clz_i32 0
-#define TCG_TARGET_HAS_ctz_i32 0
+#define TCG_TARGET_HAS_clz_i32 have_zbb
+#define TCG_TARGET_HAS_ctz_i32 have_zbb
#define TCG_TARGET_HAS_ctpop_i32 have_zbb
#define TCG_TARGET_HAS_brcond2 1
#define TCG_TARGET_HAS_setcond2 1
@@ -159,8 +159,8 @@ extern bool have_zbb;
#define TCG_TARGET_HAS_eqv_i64 have_zbb
#define TCG_TARGET_HAS_nand_i64 0
#define TCG_TARGET_HAS_nor_i64 0
-#define TCG_TARGET_HAS_clz_i64 0
-#define TCG_TARGET_HAS_ctz_i64 0
+#define TCG_TARGET_HAS_clz_i64 have_zbb
+#define TCG_TARGET_HAS_ctz_i64 have_zbb
#define TCG_TARGET_HAS_ctpop_i64 have_zbb
#define TCG_TARGET_HAS_add2_i64 1
#define TCG_TARGET_HAS_sub2_i64 1
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index 811b84d152..c0257124fa 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -1063,6 +1063,22 @@ static void tcg_out_movcond(TCGContext *s, TCGCond cond, TCGReg ret,
}
}
+static void tcg_out_cltz(TCGContext *s, TCGType type, RISCVInsn insn,
+ TCGReg ret, TCGReg src1, int src2, bool c_src2)
+{
+ tcg_out_opc_imm(s, insn, ret, src1, 0);
+
+ if (!c_src2 || src2 != (type == TCG_TYPE_I32 ? 32 : 64)) {
+ /*
+ * The requested zero result does not match the insn, so adjust.
+ * Note that constraints put 'ret' in a new register, so the
+ * computation above did not clobber either 'src1' or 'src2'.
+ */
+ tcg_out_movcond(s, TCG_COND_EQ, ret, src1, 0, true,
+ src2, c_src2, ret, false);
+ }
+}
+
static void tcg_out_call_int(TCGContext *s, const tcg_insn_unit *arg, bool tail)
{
TCGReg link = tail ? TCG_REG_ZERO : TCG_REG_RA;
@@ -1723,6 +1739,19 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
tcg_out_opc_imm(s, OPC_CPOP, a0, a1, 0);
break;
+ case INDEX_op_clz_i32:
+ tcg_out_cltz(s, TCG_TYPE_I32, OPC_CLZW, a0, a1, a2, c2);
+ break;
+ case INDEX_op_clz_i64:
+ tcg_out_cltz(s, TCG_TYPE_I64, OPC_CLZ, a0, a1, a2, c2);
+ break;
+ case INDEX_op_ctz_i32:
+ tcg_out_cltz(s, TCG_TYPE_I32, OPC_CTZW, a0, a1, a2, c2);
+ break;
+ case INDEX_op_ctz_i64:
+ tcg_out_cltz(s, TCG_TYPE_I64, OPC_CTZ, a0, a1, a2, c2);
+ break;
+
case INDEX_op_add2_i32:
tcg_out_addsub2(s, a0, a1, a2, args[3], args[4], args[5],
const_args[4], const_args[5], false, true);
@@ -1920,6 +1949,12 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
case INDEX_op_rotr_i64:
return C_O1_I2(r, r, ri);
+ case INDEX_op_clz_i32:
+ case INDEX_op_clz_i64:
+ case INDEX_op_ctz_i32:
+ case INDEX_op_ctz_i64:
+ return C_N1_I2(r, r, rM);
+
case INDEX_op_brcond_i32:
case INDEX_op_brcond_i64:
return C_O0_I2(rZ, rZ);
--
2.34.1
^ permalink raw reply related [flat|nested] 25+ messages in thread
* Re: [PULL 00/23] tcg patch queue
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
` (22 preceding siblings ...)
2023-05-25 18:10 ` [PULL 23/23] tcg/riscv: Support CTZ, CLZ from Zbb Richard Henderson
@ 2023-05-25 19:32 ` Richard Henderson
23 siblings, 0 replies; 25+ messages in thread
From: Richard Henderson @ 2023-05-25 19:32 UTC (permalink / raw)
To: qemu-devel
On 5/25/23 11:10, Richard Henderson wrote:
> The following changes since commit b300c134465465385045ab705b68a42699688332:
>
> Merge tag 'pull-vfio-20230524' ofhttps://github.com/legoater/qemu into staging (2023-05-24 14:23:41 -0700)
>
> are available in the Git repository at:
>
> https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20230525
>
> for you to fetch changes up to a30498fcea5a8b9c544324ccfb0186090104b229:
>
> tcg/riscv: Support CTZ, CLZ from Zbb (2023-05-25 15:29:36 +0000)
>
> ----------------------------------------------------------------
> tcg/mips:
> - Constant formation improvements
> - Replace MIPS_BE with HOST_BIG_ENDIAN
> - General cleanups
> tcg/riscv:
> - Improve setcond
> - Support movcond
> - Support Zbb, Zba
Applied, thanks. Please update https://wiki.qemu.org/ChangeLog/8.1 as appropriate.
r~
^ permalink raw reply [flat|nested] 25+ messages in thread
end of thread, other threads:[~2023-05-25 19:33 UTC | newest]
Thread overview: 25+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-05-25 18:10 [PULL 00/23] tcg patch queue Richard Henderson
2023-05-25 18:10 ` [PULL 01/23] tcg/mips: Move TCG_AREG0 to S8 Richard Henderson
2023-05-25 18:10 ` [PULL 02/23] tcg/mips: Move TCG_GUEST_BASE_REG to S7 Richard Henderson
2023-05-25 18:10 ` [PULL 03/23] tcg/mips: Unify TCG_GUEST_BASE_REG tests Richard Henderson
2023-05-25 18:10 ` [PULL 04/23] tcg/mips: Create and use TCG_REG_TB Richard Henderson
2023-05-25 18:10 ` [PULL 05/23] tcg/mips: Split out tcg_out_movi_one Richard Henderson
2023-05-25 18:10 ` [PULL 06/23] tcg/mips: Split out tcg_out_movi_two Richard Henderson
2023-05-25 18:10 ` [PULL 07/23] tcg/mips: Use the constant pool for 64-bit constants Richard Henderson
2023-05-25 18:10 ` [PULL 08/23] tcg/mips: Aggressively use the constant pool for n64 calls Richard Henderson
2023-05-25 18:10 ` [PULL 09/23] tcg/mips: Try tb-relative addresses in tcg_out_movi Richard Henderson
2023-05-25 18:10 ` [PULL 10/23] tcg/mips: Try three insns with shift and add " Richard Henderson
2023-05-25 18:10 ` [PULL 11/23] tcg/mips: Use qemu_build_not_reached for LO/HI_OFF Richard Henderson
2023-05-25 18:10 ` [PULL 12/23] tcg/mips: Replace MIPS_BE with HOST_BIG_ENDIAN Richard Henderson
2023-05-25 18:10 ` [PULL 13/23] disas/riscv: Decode czero.{eqz,nez} Richard Henderson
2023-05-25 18:10 ` [PULL 14/23] tcg/riscv: Probe for Zba, Zbb, Zicond extensions Richard Henderson
2023-05-25 18:10 ` [PULL 15/23] tcg/riscv: Support ANDN, ORN, XNOR from Zbb Richard Henderson
2023-05-25 18:10 ` [PULL 16/23] tcg/riscv: Support ADD.UW, SEXT.B, SEXT.H, ZEXT.H from Zba+Zbb Richard Henderson
2023-05-25 18:10 ` [PULL 17/23] tcg/riscv: Use ADD.UW for guest address generation Richard Henderson
2023-05-25 18:10 ` [PULL 18/23] tcg/riscv: Support rotates from Zbb Richard Henderson
2023-05-25 18:10 ` [PULL 19/23] tcg/riscv: Support REV8 " Richard Henderson
2023-05-25 18:10 ` [PULL 20/23] tcg/riscv: Support CPOP " Richard Henderson
2023-05-25 18:10 ` [PULL 21/23] tcg/riscv: Improve setcond expansion Richard Henderson
2023-05-25 18:10 ` [PULL 22/23] tcg/riscv: Implement movcond Richard Henderson
2023-05-25 18:10 ` [PULL 23/23] tcg/riscv: Support CTZ, CLZ from Zbb Richard Henderson
2023-05-25 19:32 ` [PULL 00/23] tcg patch queue Richard Henderson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).