mirrors/qemu
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
19c417ec87
qemu/tcg/s390x/tcg-target-con-set.h

49 lines
1.1 KiB
C
Raw Normal View History

tcg/s390: Split out constraint sets to tcg-target-con-set.h Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-10-17 21:15:40 +03:00
/* SPDX-License-Identifier: MIT */
/*
* Define S390 target-specific constraint sets.
* Copyright (c) 2021 Linaro
*/
/*
* C_On_Im(...) defines a constraint set with <n> outputs and <m> inputs.
* Each operand should be a sequence of constraint letters as defined by
* tcg-target-con-str.h; the constraint combination is inclusive or.
tcg/{i386, s390x}: Add earlyclobber to the op_add2's first output i386 and s390x implementations of op_add2 require an earlyclobber, which is currently missing. This breaks VCKSM in s390x guests. E.g., on x86_64 the following op: add2_i32 tmp2,tmp3,tmp2,tmp3,tmp3,tmp2 dead: 0 2 3 4 5 pref=none,0xffff is translated to: addl %ebx, %r12d adcl %r12d, %ebx Introduce a new C_N1_O1_I4 constraint, and make sure that earlyclobber of aliased outputs is honored. Cc: qemu-stable@nongnu.org Fixes: 82790a870992 ("tcg: Add markup for output requires new register") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20230719221310.1968845-7-iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-07-20 01:11:18 +03:00
*
* C_Nn_Om_Ik(...) defines a constraint set with <n + m> outputs and <k>
* inputs, except that the first <n> outputs must use new registers.
tcg/s390: Split out constraint sets to tcg-target-con-set.h Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-10-17 21:15:40 +03:00
*/
C_O0_I1(r)
C_O0_I2(r, r)
C_O0_I2(r, ri)
tcg/s390x: Tighten constraints for 64-bit compare Give 64-bit comparison second operand a signed 33-bit immediate. This is the smallest superset of uint32_t and int32_t, as used by CLGFI and CGFI respectively. The rest of the 33-bit space can be loaded into TCG_TMP0. Drop use of the constant pool. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-12-09 00:48:00 +03:00
C_O0_I2(r, rA)
tcg/s390x: Add host vector framework Add registers and function stubs. The functionality is disabled via squashing s390_facilities[2] to 0. We must still include results for the mandatory opcodes in tcg_target_op_def, as all opcodes are checked during tcg init. Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 02:33:14 +03:00
C_O0_I2(v, r)
tcg/s390x: Support 128-bit load/store Use LPQ/STPQ when 16-byte atomicity is required. Note that these instructions require 16-byte alignment. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-04-19 18:58:23 +03:00
C_O0_I3(o, m, r)
tcg/s390: Split out constraint sets to tcg-target-con-set.h Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-10-17 21:15:40 +03:00
C_O1_I1(r, r)
tcg/s390x: Add host vector framework Add registers and function stubs. The functionality is disabled via squashing s390_facilities[2] to 0. We must still include results for the mandatory opcodes in tcg_target_op_def, as all opcodes are checked during tcg init. Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 02:33:14 +03:00
C_O1_I1(v, r)
tcg/s390x: Implement andc, orc, abs, neg, not vector operations These logical and arithmetic operations are optional but trivial. Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 04:23:16 +03:00
C_O1_I1(v, v)
tcg/s390x: Add host vector framework Add registers and function stubs. The functionality is disabled via squashing s390_facilities[2] to 0. We must still include results for the mandatory opcodes in tcg_target_op_def, as all opcodes are checked during tcg init. Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 02:33:14 +03:00
C_O1_I1(v, vr)
tcg/s390: Split out constraint sets to tcg-target-con-set.h Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-10-17 21:15:40 +03:00
C_O1_I2(r, 0, ri)
C_O1_I2(r, 0, rI)
C_O1_I2(r, 0, rJ)
tcg/s390x: Support MIE3 logical operations This is andc, orc, nand, nor, eqv. We can use nor for implementing not. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-02-24 03:04:25 +03:00
C_O1_I2(r, r, r)
tcg/s390: Split out constraint sets to tcg-target-con-set.h Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-10-17 21:15:40 +03:00
C_O1_I2(r, r, ri)
tcg/s390x: Tighten constraints for 64-bit compare Give 64-bit comparison second operand a signed 33-bit immediate. This is the smallest superset of uint32_t and int32_t, as used by CLGFI and CGFI respectively. The rest of the 33-bit space can be loaded into TCG_TMP0. Drop use of the constant pool. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-12-09 00:48:00 +03:00
C_O1_I2(r, r, rA)
tcg/s390x: Use tgen_movcond_int in tgen_clz Reuse code from movcond to conditionally copy a2 to dest, based on the condition codes produced by FLOGR. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-02-24 06:03:20 +03:00
C_O1_I2(r, r, rI)
tcg/s390x: Support MIE2 multiply single instructions The MIE2 facility adds 3-operand versions of multiply. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-02-24 04:28:44 +03:00
C_O1_I2(r, r, rJ)
tcg/s390x: Tighten constraints for or_i64 and xor_i64 Drop support for sequential OR and XOR, as the serial dependency is slower than loading the constant first. Let the register allocator handle such immediates by matching only what one insn can achieve. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-12-08 20:28:19 +03:00
C_O1_I2(r, r, rK)
tcg/s390x: Support MIE3 logical operations This is andc, orc, nand, nor, eqv. We can use nor for implementing not. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-02-24 03:04:25 +03:00
C_O1_I2(r, r, rKR)
C_O1_I2(r, r, rNK)
tcg/s390x: Tighten constraints for and_i64 Let the register allocator handle such immediates by matching only what one insn can achieve. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-12-08 21:06:03 +03:00
C_O1_I2(r, r, rNKR)
tcg/s390: Split out constraint sets to tcg-target-con-set.h Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-10-17 21:15:40 +03:00
C_O1_I2(r, rZ, r)
tcg/s390x: Implement vector shift operations Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 05:00:38 +03:00
C_O1_I2(v, v, r)
tcg/s390x: Add host vector framework Add registers and function stubs. The functionality is disabled via squashing s390_facilities[2] to 0. We must still include results for the mandatory opcodes in tcg_target_op_def, as all opcodes are checked during tcg init. Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 02:33:14 +03:00
C_O1_I2(v, v, v)
tcg/s390x: Implement TCG_TARGET_HAS_bitsel_vec Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2020-09-15 06:36:36 +03:00
C_O1_I3(v, v, v, v)
tcg/s390x: Generalize movcond implementation Generalize movcond to support pre-computed conditions, and the same set of arguments at all times. This will be assumed by a following patch, which needs to reuse tgen_movcond_int. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-02-24 03:24:05 +03:00
C_O1_I4(r, r, ri, rI, r)
tcg/s390x: Tighten constraints for 64-bit compare Give 64-bit comparison second operand a signed 33-bit immediate. This is the smallest superset of uint32_t and int32_t, as used by CLGFI and CGFI respectively. The rest of the 33-bit space can be loaded into TCG_TMP0. Drop use of the constant pool. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-12-09 00:48:00 +03:00
C_O1_I4(r, r, rA, rI, r)
tcg/s390x: Support 128-bit load/store Use LPQ/STPQ when 16-byte atomicity is required. Note that these instructions require 16-byte alignment. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-04-19 18:58:23 +03:00
C_O2_I1(o, m, r)
tcg/s390x: Use register pair allocation for div and mulu2 Previously we hard-coded R2 and R3. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-10-10 18:41:15 +03:00
C_O2_I2(o, m, 0, r)
tcg/s390x: Support MIE2 MGRK instruction The MIE2 facility adds a 3-operand signed 64x64->128 multiply. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-02-24 04:35:43 +03:00
C_O2_I2(o, m, r, r)
tcg/s390x: Use register pair allocation for div and mulu2 Previously we hard-coded R2 and R3. Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2022-10-10 18:41:15 +03:00
C_O2_I3(o, m, 0, 1, r)
tcg/{i386, s390x}: Add earlyclobber to the op_add2's first output i386 and s390x implementations of op_add2 require an earlyclobber, which is currently missing. This breaks VCKSM in s390x guests. E.g., on x86_64 the following op: add2_i32 tmp2,tmp3,tmp2,tmp3,tmp3,tmp2 dead: 0 2 3 4 5 pref=none,0xffff is translated to: addl %ebx, %r12d adcl %r12d, %ebx Introduce a new C_N1_O1_I4 constraint, and make sure that earlyclobber of aliased outputs is honored. Cc: qemu-stable@nongnu.org Fixes: 82790a870992 ("tcg: Add markup for output requires new register") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20230719221310.1968845-7-iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-07-20 01:11:18 +03:00
C_N1_O1_I4(r, r, 0, 1, ri, r)
C_N1_O1_I4(r, r, 0, 1, rA, r)
Reference in New Issue Copy Permalink