Skip to contentSkip to conversation

selftests/arm64: Fix build warnings for abi

PATCH In review Dev Jain 3 participants 8 messages Aug 27, 2024 – Aug 29, 2024

Subsystems (2)
lore
Share & tools
lore

Dev Jain

author
#0 lore raw

PATCH 1/1 selftests/arm64: Fix build warnings for abi

A "%s" is missing in ksft_exit_fail_msg(); instead, use the newly introduced ksft_exit_fail_perror(). Also, uint64_t corresponds to unsigned 64-bit integer, so use %lx instead of %llx.

---
The changes in ptrace.c were earlier a part of the following:
https://lore.kernel.org/all/[email protected]/
which were reviewed by Mark.

 tools/testing/selftests/arm64/abi/ptrace.c      | 4 ++--
 tools/testing/selftests/arm64/abi/syscall-abi.c | 8 ++++----
 2 files changed, 6 insertions(+), 6 deletions(-)
tools/testing/selftests/arm64/abi/ptrace.c +2 -2
index e4fa507cbdd0..b51d21f78cf9 100644
@@ -163,10 +163,10 @@ static void test_hw_debug(pid_t child, int type, const char *type_name)
 static int do_child(void)
 {
 	if (ptrace(PTRACE_TRACEME, -1, NULL, NULL))
-		ksft_exit_fail_msg("PTRACE_TRACEME", strerror(errno));
+		ksft_exit_fail_perror("PTRACE_TRACEME");
 
 	if (raise(SIGSTOP))
-		ksft_exit_fail_msg("raise(SIGSTOP)", strerror(errno));
+		ksft_exit_fail_perror("raise(SIGSTOP)");
 
 	return EXIT_SUCCESS;
 }
tools/testing/selftests/arm64/abi/syscall-abi.c +4 -4
index d704511a0955..5ec9a18ec802 100644
@@ -81,7 +81,7 @@ static int check_gpr(struct syscall_cfg *cfg, int sve_vl, int sme_vl, uint64_t s
 	 */
 	for (i = 9; i < ARRAY_SIZE(gpr_in); i++) {
 		if (gpr_in[i] != gpr_out[i]) {
-			ksft_print_msg("%s SVE VL %d mismatch in GPR %d: %llx != %llx\n",
+			ksft_print_msg("%s SVE VL %d mismatch in GPR %d: %lx != %lx\n",
 				       cfg->name, sve_vl, i,
 				       gpr_in[i], gpr_out[i]);
 			errors++;
@@ -112,7 +112,7 @@ static int check_fpr(struct syscall_cfg *cfg, int sve_vl, int sme_vl,
 	if (!sve_vl && !(svcr & SVCR_SM_MASK)) {
 		for (i = 0; i < ARRAY_SIZE(fpr_in); i++) {
 			if (fpr_in[i] != fpr_out[i]) {
-				ksft_print_msg("%s Q%d/%d mismatch %llx != %llx\n",
+				ksft_print_msg("%s Q%d/%d mismatch %lx != %lx\n",
 					       cfg->name,
 					       i / 2, i % 2,
 					       fpr_in[i], fpr_out[i]);
@@ -294,13 +294,13 @@ static int check_svcr(struct syscall_cfg *cfg, int sve_vl, int sme_vl,
 	int errors = 0;
 
 	if (svcr_out & SVCR_SM_MASK) {
-		ksft_print_msg("%s Still in SM, SVCR %llx\n",
+		ksft_print_msg("%s Still in SM, SVCR %lx\n",
 			       cfg->name, svcr_out);
 		errors++;
 	}
 
 	if ((svcr_in & SVCR_ZA_MASK) != (svcr_out & SVCR_ZA_MASK)) {
-		ksft_print_msg("%s PSTATE.ZA changed, SVCR %llx != %llx\n",
+		ksft_print_msg("%s PSTATE.ZA changed, SVCR %lx != %lx\n",
 			       cfg->name, svcr_in, svcr_out);
 		errors++;
 	}
Signature · 1 line
2.30.2
maintainer
#1 lore raw
Quoted 3 lines · Dev Jain

A "%s" is missing in ksft_exit_fail_msg(); instead, use the newly introduced ksft_exit_fail_perror(). Also, uint64_t corresponds to unsigned 64-bit integer, so use %lx instead of %llx.

What's wrong with using %llx for a uint64_t? I think that part of the code is fine as-is.

Will

author
#2 lore raw
Quoted 6 lines · Will Deacon
Quoted 3 lines · Dev Jain

A "%s" is missing in ksft_exit_fail_msg(); instead, use the newly introduced ksft_exit_fail_perror(). Also, uint64_t corresponds to unsigned 64-bit integer, so use %lx instead of %llx.

What's wrong with using %llx for a uint64_t? I think that part of the code is fine as-is.

I get a lot of warnings like these:
https://www.codedump.xyz/cpp/Zs6d7saC1IRuWV2O

Quoted 2 lines

Will

#5 lore raw
In reply to Dev Jain

tools/testing/selftests/arm64/abi/ptrace.c static void test_hw_debug(pid_t child, int type, const char *type_name) · line 169

Quoted 6 lines · Dev Jain
tools/testing/selftests/arm64/abi/ptrace.c +2 -2
-		ksft_exit_fail_msg("PTRACE_TRACEME", strerror(errno));
+		ksft_exit_fail_perror("PTRACE_TRACEME");
 
 	if (raise(SIGSTOP))
-		ksft_exit_fail_msg("raise(SIGSTOP)", strerror(errno));
+		ksft_exit_fail_perror("raise(SIGSTOP)");

The idea with these is to include the error code as well so adding the %s would be better.

tools/testing/selftests/arm64/abi/syscall-abi.c static int check_gpr(struct syscall_cfg *cfg, int sve_vl, int sme_vl, uint64_t s · line 87

Quoted 7 lines
tools/testing/selftests/arm64/abi/syscall-abi.c +1 -1
 	for (i = 9; i < ARRAY_SIZE(gpr_in); i++) {
 		if (gpr_in[i] != gpr_out[i]) {
-			ksft_print_msg("%s SVE VL %d mismatch in GPR %d: %llx != %llx\n",
+			ksft_print_msg("%s SVE VL %d mismatch in GPR %d: %lx != %lx\n",
 				       cfg->name, sve_vl, i,
 				       gpr_in[i], gpr_out[i]);
 			errors++;

This is a different pattern of warning, it'd be easier to review if each patch only followed one pattern. There's no code overlap between the changes.

#3 lore raw
In reply to Will Deacon
Quoted 7 lines · Will Deacon
Quoted 3 lines · Dev Jain

A "%s" is missing in ksft_exit_fail_msg(); instead, use the newly introduced ksft_exit_fail_perror(). Also, uint64_t corresponds to unsigned 64-bit integer, so use %lx instead of %llx.

What's wrong with using %llx for a uint64_t? I think that part of the code is fine as-is.

IIRC there's some printf format specifiers specific to uint64_t which are probably a better choice here.

author
#6 lore raw
In reply to Mark Brown
Quoted 10 lines · Mark Brown

tools/testing/selftests/arm64/abi/ptrace.c static void test_hw_debug(pid_t child, int type, const char *type_name) · line 169

Quoted 6 lines · Dev Jain
tools/testing/selftests/arm64/abi/ptrace.c +2 -2
-		ksft_exit_fail_msg("PTRACE_TRACEME", strerror(errno));
+		ksft_exit_fail_perror("PTRACE_TRACEME");
 
  	if (raise(SIGSTOP))
-		ksft_exit_fail_msg("raise(SIGSTOP)", strerror(errno));
+		ksft_exit_fail_perror("raise(SIGSTOP)");

The idea with these is to include the error code as well so adding the %s would be better.

ksft_exit_fail_perror() can do that thing for us.

Quoted 11 lines

tools/testing/selftests/arm64/abi/syscall-abi.c static int check_gpr(struct syscall_cfg *cfg, int sve_vl, int sme_vl, uint64_t s · line 87

Quoted 8 lines
  	for (i = 9; i < ARRAY_SIZE(gpr_in); i++) {
  		if (gpr_in[i] != gpr_out[i]) {
tools/testing/selftests/arm64/abi/syscall-abi.c +1 -1
-			ksft_print_msg("%s SVE VL %d mismatch in GPR %d: %llx != %llx\n",
+			ksft_print_msg("%s SVE VL %d mismatch in GPR %d: %lx != %lx\n",
  				       cfg->name, sve_vl, i,
  				       gpr_in[i], gpr_out[i]);
  			errors++;

This is a different pattern of warning, it'd be easier to review if each patch only followed one pattern. There's no code overlap between the changes.

author
#4 lore raw
In reply to Mark Brown
Quoted 9 lines · Mark Brown
Quoted 6 lines · Will Deacon
Quoted 3 lines · Dev Jain

A "%s" is missing in ksft_exit_fail_msg(); instead, use the newly introduced ksft_exit_fail_perror(). Also, uint64_t corresponds to unsigned 64-bit integer, so use %lx instead of %llx.

What's wrong with using %llx for a uint64_t? I think that part of the code is fine as-is.

IIRC there's some printf format specifiers specific to uint64_t which are probably a better choice here.

Well, then I will just send a patch fixing only the %s thing; I don't want to get lost into format specifiers right now :)

#7 lore raw
In reply to Dev Jain
Quoted 9 lines · Dev Jain
Quoted 5 lines · Mark Brown

tools/testing/selftests/arm64/abi/ptrace.c static void test_hw_debug(pid_t child, int type, const char *type_name) · line 169

Quoted 2 lines
tools/testing/selftests/arm64/abi/ptrace.c +1 -1
-		ksft_exit_fail_msg("raise(SIGSTOP)", strerror(errno));
+		ksft_exit_fail_perror("raise(SIGSTOP)");

The idea with these is to include the error code as well so adding the %s would be better.

ksft_exit_fail_perror() can do that thing for us.

Oh, sorry - I missed that it was changing to _perror() as well as removing the argument. Yes, that's fine.