[PATCH] selftests/powerpc: Fix TM resched DSCR test with some compilers

[PATCH] selftests/powerpc: Fix TM resched DSCR test with some compilers

[Michael Ellerman]

The tm-resched-dscr test has started failing sometimes, depending on
what compiler it's built with, eg:

  test: tm_resched_dscr
  Check DSCR TM context switch: tm-resched-dscr: tm-resched-dscr.c:76: test_body: Assertion `rv' failed.
  !! child died by signal 6

When it fails we see that the compiler doesn't initialise rv to 1 before
entering the inline asm block. Although that's counter intuitive, it
is allowed because we tell the compiler that the inline asm will write
to rv (using "=r"), meaning the original value is irrelevant.

Marking it as a read/write parameter would presumably work, but it seems
simpler to fix it by setting the initial value of rv in the inline asm.

Fixes: 96d016108640 ("powerpc: Correct DSCR during TM context switch")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/tm/tm-resched-dscr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c b/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
index d9c49f41515e..e79ccd6aada1 100644
--- a/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
+++ b/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
@@ -42,12 +42,12 @@ int test_body(void)
 	printf("Check DSCR TM context switch: ");
 	fflush(stdout);
 	for (;;) {
-		rv = 1;
 		asm __volatile__ (
 			/* set a known value into the DSCR */
 			"ld      3, %[dscr1];"
 			"mtspr   %[sprn_dscr], 3;"
 
+			"li      %[rv], 1;"
 			/* start and suspend a transaction */
 			"tbegin.;"
 			"beq     1f;"
-- 
2.7.4

Re: [PATCH] selftests/powerpc: Fix TM resched DSCR test with some compilers

[Michael Neuling]

On Fri, 2017-05-19 at 11:37 +1000, Michael Ellerman wrote:
> The tm-resched-dscr test has started failing sometimes, depending on
> what compiler it's built with, eg:
>=20
> =C2=A0 test: tm_resched_dscr
> =C2=A0 Check DSCR TM context switch: tm-resched-dscr: tm-resched-dscr.c:7=
6:
> test_body: Assertion `rv' failed.
> =C2=A0 !! child died by signal 6
>=20
> When it fails we see that the compiler doesn't initialise rv to 1 before
> entering the inline asm block. Although that's counter intuitive, it
> is allowed because we tell the compiler that the inline asm will write
> to rv (using "=3Dr"), meaning the original value is irrelevant.
>=20
> Marking it as a read/write parameter would presumably work, but it seems
> simpler to fix it by setting the initial value of rv in the inline asm.
>=20
> Fixes: 96d016108640 ("powerpc: Correct DSCR during TM context switch")
> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Thanks.

Acked-by: Michael Neuling <mikey@neuling.org>

> ---
> =C2=A0tools/testing/selftests/powerpc/tm/tm-resched-dscr.c | 2 +-
> =C2=A01 file changed, 1 insertion(+), 1 deletion(-)
>=20
> diff --git a/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> b/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> index d9c49f41515e..e79ccd6aada1 100644
> --- a/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> +++ b/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> @@ -42,12 +42,12 @@ int test_body(void)
> =C2=A0	printf("Check DSCR TM context switch: ");
> =C2=A0	fflush(stdout);
> =C2=A0	for (;;) {
> -		rv =3D 1;
> =C2=A0		asm __volatile__ (
> =C2=A0			/* set a known value into the DSCR */
> =C2=A0			"ld=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A03, %[dscr1];"
> =C2=A0			"mtspr=C2=A0=C2=A0=C2=A0%[sprn_dscr], 3;"
> =C2=A0
> +			"li=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0%[rv], 1;"
> =C2=A0			/* start and suspend a transaction */
> =C2=A0			"tbegin.;"
> =C2=A0			"beq=C2=A0=C2=A0=C2=A0=C2=A0=C2=A01f;"

Re: selftests/powerpc: Fix TM resched DSCR test with some compilers

[Michael Ellerman]

On Fri, 2017-05-19 at 01:37:45 UTC, Michael Ellerman wrote:
> The tm-resched-dscr test has started failing sometimes, depending on
> what compiler it's built with, eg:
> 
>   test: tm_resched_dscr
>   Check DSCR TM context switch: tm-resched-dscr: tm-resched-dscr.c:76: test_body: Assertion `rv' failed.
>   !! child died by signal 6
> 
> When it fails we see that the compiler doesn't initialise rv to 1 before
> entering the inline asm block. Although that's counter intuitive, it
> is allowed because we tell the compiler that the inline asm will write
> to rv (using "=r"), meaning the original value is irrelevant.
> 
> Marking it as a read/write parameter would presumably work, but it seems
> simpler to fix it by setting the initial value of rv in the inline asm.
> 
> Fixes: 96d016108640 ("powerpc: Correct DSCR during TM context switch")
> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
> Acked-by: Michael Neuling <mikey@neuling.org>

Applied to powerpc fixes.

https://git.kernel.org/powerpc/c/fe06fe860250a4f01d0eaf70a2563b

cheers