target/s390x: Fix LAE setting a wrong access register
cover target/s390x: Fix LAE setting a wrong access register
Hi,
Ido has noticed that LAE sets a wrong access register and proposed a fix. This series fixes the issue and adds a test.
Best regards,
Ilya
Ilya Leoshkevich (2):
target/s390x: Fix LAE setting a wrong access register tests/tcg/s390x: Test LOAD ADDRESS EXTENDED
target/s390x/tcg/translate.c | 3 ++- tests/tcg/s390x/Makefile.target | 1 + tests/tcg/s390x/lae.c | 25 +++++++++++++++++++++++++ 3 files changed, 28 insertions(+), 1 deletion(-) create mode 100644 tests/tcg/s390x/lae.c
Signature · 1 line
2.43.0
PATCH 1/2 target/s390x: Fix LAE setting a wrong access register
LAE should set the access register corresponding to the first operand, instead, it always modifies access register 1.
--- target/s390x/tcg/translate.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
target/s390x/tcg/translate.c +2 -1
@@ -3221,6 +3221,7 @@ static DisasJumpType op_mov2e(DisasContext *s, DisasOps *o) {int b2 =get_field (s, b2);TCGv ar1 =tcg_temp_new_i64 (); +int r1 =get_field (s, r1); o->out = o->in2; o->in2 =NULL ;
@@ -3244,7 +3245,7 @@ static DisasJumpType op_mov2e(DisasContext *s, DisasOps *o)break ; } - tcg_gen_st32_i64(ar1, tcg_env, offsetof(CPUS390XState, aregs[1])); + tcg_gen_st32_i64(ar1, tcg_env, offsetof(CPUS390XState, aregs[r1]));return DISAS_NEXT ; }
Signature · 1 line
2.43.0
PATCH 2/2 tests/tcg/s390x: Test LOAD ADDRESS EXTENDED
In reply to Ilya LeoshkevichAdd a small test to prevent regressions. Userspace runs in primary mode, so LAE should always set the access register to 0.
--- tests/tcg/s390x/Makefile.target | 1 + tests/tcg/s390x/lae.c | 25 +++++++++++++++++++++++++ 2 files changed, 26 insertions(+) create mode 100644 tests/tcg/s390x/lae.c
tests/tcg/s390x/Makefile.target +1 -0
@@ -44,6 +44,7 @@ TESTS+=clgebr TESTS+=clc TESTS+=laalg TESTS+=add-logical-with-carry +TESTS+=lae cdsg: CFLAGS+=-pthread cdsg: LDFLAGS+=-pthread
tests/tcg/s390x/lae.c new file +25 -0
@@ -0,0 +1,25 @@ +/* +* Test the LOAD ADDRESS EXTENDED instruction. +* +* SPDX-License-Identifier: GPL-2.0-or-later +*/ +#include <assert.h> +#include <stdlib.h> + +int main (void ) +{ +unsigned long long ar =-1 , b2 =100000 , r, x2 =500 ; +int tmp; + + asm("ear %[tmp],%[r]\n" +"lae %[r],42(%[x2],%[b2])\n" +"ear %[ar],%[r]\n" +"sar %[r],%[tmp]" + : [tmp]"=&r" (tmp), [r]"=&r" (r), [ar]"+r" (ar) + : [b2]"r" (b2), [x2]"r" (x2) + :"memory" ); +assert (ar ==0xffffffff00000000ULL ); +assert (r ==100542 ); + +return EXIT_SUCCESS ; +}
Signature · 1 line
2.43.0
Quoted 24 lines · Ilya Leoshkevich
LAE should set the access register corresponding to the first operand, instead, it always modifies access register 1.
Co-developed-by Ido Plat <[email protected]>Fixes a1c7610a6879 ("target-s390x: implement LAY and LAEY instructions")Signed-off-by Ilya Leoshkevich <[email protected]>--- target/s390x/tcg/translate.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)target/s390x/tcg/translate.c +1
-0@@ -3221,6 +3221,7 @@ static DisasJumpType op_mov2e(DisasContext *s, DisasOps *o) {int b2 =get_field (s, b2);TCGv ar1 =tcg_temp_new_i64 (); +int r1 =get_field (s, r1); o->out = o->in2; o->in2 =NULL ;
target/s390x/tcg/translate.c static DisasJumpType op_mov2e(DisasContext *s, DisasOps *o) · line 3251
Quoted 8 lines
target/s390x/tcg/translate.c +1
-1@@ -3244,7 +3245,7 @@ static DisasJumpType op_mov2e(DisasContext *s, DisasOps *o)break ; } - tcg_gen_st32_i64(ar1, tcg_env, offsetof(CPUS390XState, aregs[1])); + tcg_gen_st32_i64(ar1, tcg_env, offsetof(CPUS390XState, aregs[r1]));return DISAS_NEXT ; }
Signature · 3 lines
Cheers, David / dhildenb
Re: [PATCH 2/2] tests/tcg/s390x: Test LOAD ADDRESS EXTENDED
In reply to Ilya LeoshkevichQuoted 9 lines · Ilya Leoshkevich
Add a small test to prevent regressions. Userspace runs in primary mode, so LAE should always set the access register to 0.
Signed-off-by Ilya Leoshkevich <[email protected]>--- tests/tcg/s390x/Makefile.target | 1 + tests/tcg/s390x/lae.c | 25 +++++++++++++++++++++++++ 2 files changed, 26 insertions(+)create mode 100644
tests/tcg/s390x/lae.c
Re: [PATCH 2/2] tests/tcg/s390x: Test LOAD ADDRESS EXTENDED
In reply to Ilya LeoshkevichQuoted 43 lines · Ilya Leoshkevich
Add a small test to prevent regressions. Userspace runs in primary mode, so LAE should always set the access register to 0.
Signed-off-by Ilya Leoshkevich <[email protected]>--- tests/tcg/s390x/Makefile.target | 1 + tests/tcg/s390x/lae.c | 25 +++++++++++++++++++++++++ 2 files changed, 26 insertions(+)create mode 100644
tests/tcg/s390x/lae.ctests/tcg/s390x/Makefile.target +1
-0@@ -44,6 +44,7 @@ TESTS+=clgebr TESTS+=clc TESTS+=laalg TESTS+=add-logical-with-carry +TESTS+=lae cdsg: CFLAGS+=-pthread cdsg: LDFLAGS+=-pthreadtests/tcg/s390x/lae.c new file +15
-0@@ -0,0 +1,25 @@ +/* +* Test the LOAD ADDRESS EXTENDED instruction. +* +* SPDX-License-Identifier: GPL-2.0-or-later +*/ +#include <assert.h> +#include <stdlib.h> + +int main (void ) +{ +unsigned long long ar =-1 , b2 =100000 , r, x2 =500 ; +int tmp; + + asm("ear %[tmp],%[r]\n" + "lae %[r],42(%[x2],%[b2])\n"
tests/tcg/s390x/lae.c line 25
Quoted 10 lines
tests/tcg/s390x/lae.c new file +10
-0+"ear %[ar],%[r]\n" +"sar %[r],%[tmp]" + : [tmp] "=&r" (tmp), [r] "=&r" (r), [ar] "+r" (ar) + : [b2] "r" (b2), [x2] "r" (x2) + : "memory"); + assert(ar ==0xffffffff00000000ULL ); +assert (r ==100542 ); + +return EXIT_SUCCESS ; +}
I'm sorry, but it fails when building with Clang (version 17):
.../qemu/tests/tcg/s390x/lae.c:14:9: error: invalid operand for instruction
14 | asm("ear %[tmp],%[r]\n"
| ^
<inline asm>:1:10: note: instantiated into assembly here
1 | ear %r2,%r1
| ^.../qemu/tests/tcg/s390x/lae.c:16:10: error: invalid operand for instruction
16 | "ear %[ar],%[r]\n"
| ^
<inline asm>:3:9: note: instantiated into assembly here
3 | ear %r0,%r1
| ^.../qemu/tests/tcg/s390x/lae.c:17:10: error: invalid operand for instruction
17 | "sar %[r],%[tmp]"
| ^
<inline asm>:4:5: note: instantiated into assembly here
4 | sar %r1,%r2
| ^3 errors generated.
Any suggestions how to fix it best?
Thomas
Re: [PATCH 2/2] tests/tcg/s390x: Test LOAD ADDRESS EXTENDED
Quoted 81 lines · Thomas Huth
Quoted 44 lines · Ilya Leoshkevich
Add a small test to prevent regressions. Userspace runs in primary mode, so LAE should always set the access register to 0.
Signed-off-by Ilya Leoshkevich <[email protected]>---
tests/tcg/s390x/Makefile.target | 1 + tests/tcg/s390x/lae.c | 25 +++++++++++++++++++++++++2 files changed, 26 insertions(+)create mode 100644
tests/tcg/s390x/lae.ctests/tcg/s390x/Makefile.target +0
-0@@ -44,6 +44,7 @@ TESTS+=clgebrTESTS+=clc
TESTS+=laalg
TESTS+=add-logical-with-carry
+TESTS+=lae
cdsg: CFLAGS+=-pthread
cdsg: LDFLAGS+=-pthreadtests/tcg/s390x/lae.c new file +15
-0@@ -0,0 +1,25 @@ +/* +* Test the LOAD ADDRESS EXTENDED instruction. +* +* SPDX-License-Identifier: GPL-2.0-or-later +*/ +#include <assert.h> +#include <stdlib.h> + +int main (void ) +{ + unsigned long long ar = -1, b2 = 100000, r, x2 = 500; + int tmp; + + asm("ear %[tmp],%[r]\n" + "lae %[r],42(%[x2],%[b2])\n"tests/tcg/s390x/lae.c line 25
Quoted 10 lines
tests/tcg/s390x/lae.c new file +10
-0+ "ear %[ar],%[r]\n" + "sar %[r],%[tmp]" + : [tmp] "=&r" (tmp), [r] "=&r" (r), [ar] "+r" (ar) + : [b2] "r" (b2), [x2] "r" (x2) + : "memory"); + assert(ar == 0xffffffff00000000ULL); + assert(r == 100542); + + return EXIT_SUCCESS; +}I'm sorry, but it fails when building with Clang (version 17):
.../qemu/tests/tcg/s390x/lae.c:14:9: error: invalid operand for instruction
14 | asm("ear %[tmp],%[r]\n"
| ^
<inline asm>:1:10: note: instantiated into assembly here 1 | ear %r2,%r1
| ^.../qemu/tests/tcg/s390x/lae.c:16:10: error: invalid operand for instruction
16 | "ear %[ar],%[r]\n"
| ^
<inline asm>:3:9: note: instantiated into assembly here 3 | ear %r0,%r1
| ^.../qemu/tests/tcg/s390x/lae.c:17:10: error: invalid operand for instruction
17 | "sar %[r],%[tmp]"
| ^
<inline asm>:4:5: note: instantiated into assembly here 4 | sar %r1,%r2
| ^
3 errors generated.
Quoted 4 lines
Any suggestions how to fix it best?
Thomas
clang wants %aN there, and I don't see a way to convert %rN to %aN. Seems like I'll have to hardcode the register number. I'll send a v2.