[PATCH v1 0/3] kselftest/arm64: Test ptrace writing via FPSIMD and reading via SVE

[PATCH v1 0/3] kselftest/arm64: Test ptrace writing via FPSIMD and reading via SVE

[Mark Brown]

This series has a couple of minor fixes and cleanups for sve-ptrace plus
the addition of a new test which validates that we can write using the
FPSIMD regset and then read matching data back using the SVE regset -
previously we only validated writing SVE and reading FPSIMD data.

Mark Brown (3):
  kselftest/arm64: Fix comment for ptrace_sve_get_fpsimd_data()
  kselftest/arm64: Remove assumption that tasks start FPSIMD only
  kselftest/arm64: Validate setting via FPSIMD and read via SVE regsets

 tools/testing/selftests/arm64/fp/sve-ptrace.c | 164 +++++++++++++++---
 1 file changed, 139 insertions(+), 25 deletions(-)


base-commit: dfd42facf1e4ada021b939b4e19c935dcdd55566
-- 
2.30.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 1/3] kselftest/arm64: Fix comment for ptrace_sve_get_fpsimd_data()

[Mark Brown]

The comment for ptrace_sve_get_fpsimd_data() doesn't describe what the test
does at all, fix that.

Signed-off-by: Mark Brown <broonie@kernel.org>
---
 tools/testing/selftests/arm64/fp/sve-ptrace.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/arm64/fp/sve-ptrace.c b/tools/testing/selftests/arm64/fp/sve-ptrace.c
index a3c1e67441f9..6b139d5f67f1 100644
--- a/tools/testing/selftests/arm64/fp/sve-ptrace.c
+++ b/tools/testing/selftests/arm64/fp/sve-ptrace.c
@@ -397,7 +397,7 @@ static void ptrace_set_sve_get_sve_data(pid_t child,
 	free(write_buf);
 }
 
-/* Validate attempting to set SVE data and read SVE data */
+/* Validate attempting to set SVE data and read it via the FPSIMD regset */
 static void ptrace_set_sve_get_fpsimd_data(pid_t child,
 					   const struct vec_type *type,
 					   unsigned int vl)
-- 
2.30.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 1/3] kselftest/arm64: Fix comment for ptrace_sve_get_fpsimd_data()

[Shuah Khan]

On 3/9/22 4:59 AM, Mark Brown wrote:
> The comment for ptrace_sve_get_fpsimd_data() doesn't describe what the test
> does at all, fix that.
> 
> Signed-off-by: Mark Brown <broonie@kernel.org>
> ---
>   tools/testing/selftests/arm64/fp/sve-ptrace.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/arm64/fp/sve-ptrace.c b/tools/testing/selftests/arm64/fp/sve-ptrace.c
> index a3c1e67441f9..6b139d5f67f1 100644
> --- a/tools/testing/selftests/arm64/fp/sve-ptrace.c
> +++ b/tools/testing/selftests/arm64/fp/sve-ptrace.c
> @@ -397,7 +397,7 @@ static void ptrace_set_sve_get_sve_data(pid_t child,
>   	free(write_buf);
>   }
>   
> -/* Validate attempting to set SVE data and read SVE data */
> +/* Validate attempting to set SVE data and read it via the FPSIMD regset */
>   static void ptrace_set_sve_get_fpsimd_data(pid_t child,
>   					   const struct vec_type *type,
>   					   unsigned int vl)
> 

Looks good to me.

Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>

thanks,
-- Shuah

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 2/3] kselftest/arm64: Remove assumption that tasks start FPSIMD only

[Mark Brown]

Currently the sve-ptrace test for setting and reading FPSIMD data assumes
that the child will start off in FPSIMD only mode and that it can use this
to read some FPSIMD mode SVE ptrace data, skipping the test if it can't.
This isn't an assumption guaranteed by the ABI and also limits how we can
use this testcase within the program. Instead skip the initial read and
just generate a FPSIMD format buffer for the write part of the test, making
the coverage more robust in the face of future kernel and test program
changes.

Signed-off-by: Mark Brown <broonie@kernel.org>
---
 tools/testing/selftests/arm64/fp/sve-ptrace.c | 39 ++++++++-----------
 1 file changed, 16 insertions(+), 23 deletions(-)

diff --git a/tools/testing/selftests/arm64/fp/sve-ptrace.c b/tools/testing/selftests/arm64/fp/sve-ptrace.c
index 6b139d5f67f1..ccd3ff1e8c06 100644
--- a/tools/testing/selftests/arm64/fp/sve-ptrace.c
+++ b/tools/testing/selftests/arm64/fp/sve-ptrace.c
@@ -48,7 +48,7 @@ static const struct vec_type vec_types[] = {
 
 #define VL_TESTS (((SVE_VQ_MAX - SVE_VQ_MIN) + 1) * 3)
 #define FLAG_TESTS 2
-#define FPSIMD_TESTS 3
+#define FPSIMD_TESTS 2
 
 #define EXPECTED_TESTS ((VL_TESTS + FLAG_TESTS + FPSIMD_TESTS) * ARRAY_SIZE(vec_types))
 
@@ -242,28 +242,24 @@ static void check_u32(unsigned int vl, const char *reg,
 /* Access the FPSIMD registers via the SVE regset */
 static void ptrace_sve_fpsimd(pid_t child, const struct vec_type *type)
 {
-	void *svebuf = NULL;
-	size_t svebufsz = 0;
+	void *svebuf;
 	struct user_sve_header *sve;
 	struct user_fpsimd_state *fpsimd, new_fpsimd;
 	unsigned int i, j;
 	unsigned char *p;
+	int ret;
 
-	/* New process should start with FPSIMD registers only */
-	sve = get_sve(child, type, &svebuf, &svebufsz);
-	if (!sve) {
-		ksft_test_result_fail("get_sve(%s): %s\n",
-				      type->name, strerror(errno));
-
+	svebuf = malloc(SVE_PT_SIZE(0, SVE_PT_REGS_FPSIMD));
+	if (!svebuf) {
+		ksft_test_result_fail("Failed to allocate FPSIMD buffer\n");
 		return;
-	} else {
-		ksft_test_result_pass("get_sve(%s FPSIMD)\n", type->name);
 	}
 
-	ksft_test_result((sve->flags & SVE_PT_REGS_MASK) == SVE_PT_REGS_FPSIMD,
-			 "Got FPSIMD registers via %s\n", type->name);
-	if ((sve->flags & SVE_PT_REGS_MASK) != SVE_PT_REGS_FPSIMD)
-		goto out;
+	memset(svebuf, 0, SVE_PT_SIZE(0, SVE_PT_REGS_FPSIMD));
+	sve = svebuf;
+	sve->flags = SVE_PT_REGS_FPSIMD;
+	sve->size = SVE_PT_SIZE(0, SVE_PT_REGS_FPSIMD);
+	sve->vl = 16;  /* We don't care what the VL is */
 
 	/* Try to set a known FPSIMD state via PT_REGS_SVE */
 	fpsimd = (struct user_fpsimd_state *)((char *)sve +
@@ -275,12 +271,11 @@ static void ptrace_sve_fpsimd(pid_t child, const struct vec_type *type)
 			p[j] = j;
 	}
 
-	if (set_sve(child, type, sve)) {
-		ksft_test_result_fail("set_sve(%s FPSIMD): %s\n",
-				      type->name, strerror(errno));
-
+	ret = set_sve(child, type, sve);
+	ksft_test_result(ret == 0, "%s FPSIMD set via SVE: %d\n",
+			 type->name, ret);
+	if (ret)
 		goto out;
-	}
 
 	/* Verify via the FPSIMD regset */
 	if (get_fpsimd(child, &new_fpsimd)) {
@@ -548,11 +543,9 @@ static int do_parent(pid_t child)
 		if (getauxval(vec_types[i].hwcap_type) & vec_types[i].hwcap) {
 			ptrace_sve_fpsimd(child, &vec_types[i]);
 		} else {
-			ksft_test_result_skip("%s FPSIMD get via SVE\n",
-					      vec_types[i].name);
 			ksft_test_result_skip("%s FPSIMD set via SVE\n",
 					      vec_types[i].name);
-			ksft_test_result_skip("%s set read via FPSIMD\n",
+			ksft_test_result_skip("%s FPSIMD read\n",
 					      vec_types[i].name);
 		}
 
-- 
2.30.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 2/3] kselftest/arm64: Remove assumption that tasks start FPSIMD only

[Shuah Khan]

On 3/9/22 4:59 AM, Mark Brown wrote:
> Currently the sve-ptrace test for setting and reading FPSIMD data assumes
> that the child will start off in FPSIMD only mode and that it can use this
> to read some FPSIMD mode SVE ptrace data, skipping the test if it can't.
> This isn't an assumption guaranteed by the ABI and also limits how we can
> use this testcase within the program. Instead skip the initial read and
> just generate a FPSIMD format buffer for the write part of the test, making
> the coverage more robust in the face of future kernel and test program
> changes.
> 
> Signed-off-by: Mark Brown <broonie@kernel.org>
> ---
>   tools/testing/selftests/arm64/fp/sve-ptrace.c | 39 ++++++++-----------
>   1 file changed, 16 insertions(+), 23 deletions(-)
> 
> diff --git a/tools/testing/selftests/arm64/fp/sve-ptrace.c b/tools/testing/selftests/arm64/fp/sve-ptrace.c
> index 6b139d5f67f1..ccd3ff1e8c06 100644
> --- a/tools/testing/selftests/arm64/fp/sve-ptrace.c
> +++ b/tools/testing/selftests/arm64/fp/sve-ptrace.c
> @@ -48,7 +48,7 @@ static const struct vec_type vec_types[] = {
>   
>   #define VL_TESTS (((SVE_VQ_MAX - SVE_VQ_MIN) + 1) * 3)
>   #define FLAG_TESTS 2
> -#define FPSIMD_TESTS 3
> +#define FPSIMD_TESTS 2

Looks like there is one less test now?

>   
>   #define EXPECTED_TESTS ((VL_TESTS + FLAG_TESTS + FPSIMD_TESTS) * ARRAY_SIZE(vec_types))
>   
>

Looks good to me.

Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>

thanks,
-- Shuah


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 2/3] kselftest/arm64: Remove assumption that tasks start FPSIMD only

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 537 bytes --]

On Wed, Mar 09, 2022 at 02:36:26PM -0700, Shuah Khan wrote:
> On 3/9/22 4:59 AM, Mark Brown wrote:

> > Currently the sve-ptrace test for setting and reading FPSIMD data assumes
> > that the child will start off in FPSIMD only mode and that it can use this
> > to read some FPSIMD mode SVE ptrace data, skipping the test if it can't.

> > -#define FPSIMD_TESTS 3
> > +#define FPSIMD_TESTS 2

> Looks like there is one less test now?

Yes, the initial read which we assumed would return FPSIMD data was
being reported as a distinct test.

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 3/3] kselftest/arm64: Validate setting via FPSIMD and read via SVE regsets

[Mark Brown]

Currently we validate that we can set the floating point state via the SVE
regset and read the data via the FPSIMD regset but we do not valiate that
the opposite case works as expected. Add a test that covers this case,
noting that when reading via SVE regset the kernel has the option of
returning either SVE or FPSIMD data so we need to accept both formats.

Signed-off-by: Mark Brown <broonie@kernel.org>
---
 tools/testing/selftests/arm64/fp/sve-ptrace.c | 123 +++++++++++++++++-
 1 file changed, 122 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/arm64/fp/sve-ptrace.c b/tools/testing/selftests/arm64/fp/sve-ptrace.c
index ccd3ff1e8c06..cdfb5107fd0c 100644
--- a/tools/testing/selftests/arm64/fp/sve-ptrace.c
+++ b/tools/testing/selftests/arm64/fp/sve-ptrace.c
@@ -46,7 +46,7 @@ static const struct vec_type vec_types[] = {
 	},
 };
 
-#define VL_TESTS (((SVE_VQ_MAX - SVE_VQ_MIN) + 1) * 3)
+#define VL_TESTS (((SVE_VQ_MAX - SVE_VQ_MIN) + 1) * 4)
 #define FLAG_TESTS 2
 #define FPSIMD_TESTS 2
 
@@ -80,6 +80,15 @@ static int get_fpsimd(pid_t pid, struct user_fpsimd_state *fpsimd)
 	return ptrace(PTRACE_GETREGSET, pid, NT_PRFPREG, &iov);
 }
 
+static int set_fpsimd(pid_t pid, struct user_fpsimd_state *fpsimd)
+{
+	struct iovec iov;
+
+	iov.iov_base = fpsimd;
+	iov.iov_len = sizeof(*fpsimd);
+	return ptrace(PTRACE_SETREGSET, pid, NT_PRFPREG, &iov);
+}
+
 static struct user_sve_header *get_sve(pid_t pid, const struct vec_type *type,
 				       void **buf, size_t *size)
 {
@@ -475,6 +484,115 @@ static void ptrace_set_sve_get_fpsimd_data(pid_t child,
 	free(write_buf);
 }
 
+/* Validate attempting to set FPSIMD data and read it via the SVE regset */
+static void ptrace_set_fpsimd_get_sve_data(pid_t child,
+					   const struct vec_type *type,
+					   unsigned int vl)
+{
+	void *read_buf = NULL;
+	unsigned char *p;
+	struct user_sve_header *read_sve;
+	unsigned int vq = sve_vq_from_vl(vl);
+	struct user_fpsimd_state write_fpsimd;
+	int ret, i, j;
+	size_t read_sve_size = 0;
+	size_t expected_size;
+	int errors = 0;
+
+	if (__BYTE_ORDER == __BIG_ENDIAN) {
+		ksft_test_result_skip("Big endian not supported\n");
+		return;
+	}
+
+	for (i = 0; i < 32; ++i) {
+		p = (unsigned char *)&write_fpsimd.vregs[i];
+
+		for (j = 0; j < sizeof(write_fpsimd.vregs[i]); ++j)
+			p[j] = j;
+	}
+
+	ret = set_fpsimd(child, &write_fpsimd);
+	if (ret != 0) {
+		ksft_test_result_fail("Failed to set FPSIMD state: %d\n)",
+				      ret);
+		return;
+	}
+
+	if (!get_sve(child, type, (void **)&read_buf, &read_sve_size)) {
+		ksft_test_result_fail("Failed to read %s VL %u data\n",
+				      type->name, vl);
+		return;
+	}
+	read_sve = read_buf;
+
+	if (read_sve->vl != vl) {
+		ksft_test_result_fail("Child VL != expected VL %d\n",
+				      read_sve->vl, vl);
+		goto out;
+	}
+
+	/* The kernel may return either SVE or FPSIMD format */
+	switch (read_sve->flags & SVE_PT_REGS_MASK) {
+	case SVE_PT_REGS_FPSIMD:
+		expected_size = SVE_PT_FPSIMD_SIZE(vq, SVE_PT_REGS_FPSIMD);
+		if (read_sve_size < expected_size) {
+			ksft_test_result_fail("Read %d bytes, expected %d\n",
+					      read_sve_size, expected_size);
+			goto out;
+		}
+
+		ret = memcmp(&write_fpsimd, read_buf + SVE_PT_FPSIMD_OFFSET,
+			     sizeof(write_fpsimd));
+		if (ret != 0) {
+			ksft_print_msg("Read FPSIMD data mismatch\n");
+			errors++;
+		}
+		break;
+
+	case SVE_PT_REGS_SVE:
+		expected_size = SVE_PT_SVE_SIZE(vq, SVE_PT_REGS_SVE);
+		if (read_sve_size < expected_size) {
+			ksft_test_result_fail("Read %d bytes, expected %d\n",
+					      read_sve_size, expected_size);
+			goto out;
+		}
+
+		for (i = 0; i < __SVE_NUM_ZREGS; i++) {
+			__uint128_t tmp = 0;
+
+			/*
+			 * Z regs are stored endianness invariant, this won't
+			 * work for big endian
+			 */
+			memcpy(&tmp, read_buf + SVE_PT_SVE_ZREG_OFFSET(vq, i),
+			       sizeof(tmp));
+
+			if (tmp != write_fpsimd.vregs[i]) {
+				ksft_print_msg("Mismatch in FPSIMD for %s VL %u Z%d/V%d\n",
+					       type->name, vl, i, i);
+				errors++;
+			}
+		}
+
+		check_u32(vl, "FPSR", &write_fpsimd.fpsr,
+			  read_buf + SVE_PT_SVE_FPSR_OFFSET(vq), &errors);
+		check_u32(vl, "FPCR", &write_fpsimd.fpcr,
+			  read_buf + SVE_PT_SVE_FPCR_OFFSET(vq), &errors);
+		break;
+	default:
+		ksft_print_msg("Unexpected regs type %d\n",
+			       read_sve->flags & SVE_PT_REGS_MASK);
+		errors++;
+		break;
+	}
+
+	ksft_test_result(errors == 0, "Set FPSIMD, read via SVE for %s VL %u\n",
+			 type->name, vl);
+
+out:
+	free(read_buf);
+}
+
 static int do_parent(pid_t child)
 {
 	int ret = EXIT_FAILURE;
@@ -578,11 +696,14 @@ static int do_parent(pid_t child)
 			if (vl_supported) {
 				ptrace_set_sve_get_sve_data(child, &vec_types[i], vl);
 				ptrace_set_sve_get_fpsimd_data(child, &vec_types[i], vl);
+				ptrace_set_fpsimd_get_sve_data(child, &vec_types[i], vl);
 			} else {
 				ksft_test_result_skip("%s set SVE get SVE for VL %d\n",
 						      vec_types[i].name, vl);
 				ksft_test_result_skip("%s set SVE get FPSIMD for VL %d\n",
 						      vec_types[i].name, vl);
+				ksft_test_result_skip("%s set FPSIMD get SVE for VL %d\n",
+						      vec_types[i].name, vl);
 			}
 		}
 	}
-- 
2.30.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 3/3] kselftest/arm64: Validate setting via FPSIMD and read via SVE regsets

[Shuah Khan]

On 3/9/22 4:59 AM, Mark Brown wrote:
> Currently we validate that we can set the floating point state via the SVE
> regset and read the data via the FPSIMD regset but we do not valiate that
> the opposite case works as expected. Add a test that covers this case,
> noting that when reading via SVE regset the kernel has the option of
> returning either SVE or FPSIMD data so we need to accept both formats.
> 
> Signed-off-by: Mark Brown <broonie@kernel.org>
> ---
>   tools/testing/selftests/arm64/fp/sve-ptrace.c | 123 +++++++++++++++++-
>   1 file changed, 122 insertions(+), 1 deletion(-)
> 

Looks good to me.

Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>

thanks,
-- Shuah

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel