All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2 0/2] selftests/powerpc: Add test for rfi_flush
@ 2018-05-21 15:13 Naveen N. Rao
  2018-05-21 15:13 ` [PATCH v2 1/2] selftests/powerpc: Move UCONTEXT_NIA() into utils.h Naveen N. Rao
  2018-05-21 15:13 ` [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call Naveen N. Rao
  0 siblings, 2 replies; 8+ messages in thread
From: Naveen N. Rao @ 2018-05-21 15:13 UTC (permalink / raw)
  To: Michael Ellerman; +Cc: linuxppc-dev

This is v2 of:
https://patchwork.ozlabs.org/patch/895765/

Changes:
- Patch 1 is new.
- Patch 2 uses the new macro introduced in the first patch. The output 
  message has also been slightly tweaked to make things clearer.


- Naveen

Naveen N. Rao (2):
  selftests/powerpc: Move UCONTEXT_NIA() into utils.h
  selftests/powerpc: Add test to verify rfi flush across a system call

 tools/testing/selftests/powerpc/Makefile      |   3 +-
 .../testing/selftests/powerpc/include/utils.h |  18 +++
 .../primitives/load_unaligned_zeropad.c       |   8 -
 .../selftests/powerpc/security/Makefile       |   7 +
 .../selftests/powerpc/security/rfi_flush.c    | 132 +++++++++++++++
 tools/testing/selftests/powerpc/utils.c       | 153 ++++++++++++++++++
 6 files changed, 312 insertions(+), 9 deletions(-)
 create mode 100644 tools/testing/selftests/powerpc/security/Makefile
 create mode 100644 tools/testing/selftests/powerpc/security/rfi_flush.c

-- 
2.17.0

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [PATCH v2 1/2] selftests/powerpc: Move UCONTEXT_NIA() into utils.h
  2018-05-21 15:13 [PATCH v2 0/2] selftests/powerpc: Add test for rfi_flush Naveen N. Rao
@ 2018-05-21 15:13 ` Naveen N. Rao
  2018-10-22  9:34   ` [v2,1/2] " Michael Ellerman
  2018-05-21 15:13 ` [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call Naveen N. Rao
  1 sibling, 1 reply; 8+ messages in thread
From: Naveen N. Rao @ 2018-05-21 15:13 UTC (permalink / raw)
  To: Michael Ellerman; +Cc: linuxppc-dev

... so that it can be used by others.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
---
 tools/testing/selftests/powerpc/include/utils.h           | 8 ++++++++
 .../selftests/powerpc/primitives/load_unaligned_zeropad.c | 8 --------
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/utils.h b/tools/testing/selftests/powerpc/include/utils.h
index 735815b3ad7f..b8f9cc6c9ff2 100644
--- a/tools/testing/selftests/powerpc/include/utils.h
+++ b/tools/testing/selftests/powerpc/include/utils.h
@@ -78,4 +78,12 @@ do {								\
 #define PPC_FEATURE2_ARCH_3_00 0x00800000
 #endif
 
+#if defined(__powerpc64__)
+#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.gp_regs[PT_NIP]
+#elif defined(__powerpc__)
+#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.uc_regs->gregs[PT_NIP]
+#else
+#error implement UCONTEXT_NIA
+#endif
+
 #endif /* _SELFTESTS_POWERPC_UTILS_H */
diff --git a/tools/testing/selftests/powerpc/primitives/load_unaligned_zeropad.c b/tools/testing/selftests/powerpc/primitives/load_unaligned_zeropad.c
index ed3239bbfae2..ee1e9ca22f0d 100644
--- a/tools/testing/selftests/powerpc/primitives/load_unaligned_zeropad.c
+++ b/tools/testing/selftests/powerpc/primitives/load_unaligned_zeropad.c
@@ -65,14 +65,6 @@ static int unprotect_region(void)
 extern char __start___ex_table[];
 extern char __stop___ex_table[];
 
-#if defined(__powerpc64__)
-#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.gp_regs[PT_NIP]
-#elif defined(__powerpc__)
-#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.uc_regs->gregs[PT_NIP]
-#else
-#error implement UCONTEXT_NIA
-#endif
-
 struct extbl_entry {
 	int insn;
 	int fixup;
-- 
2.17.0

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call
  2018-05-21 15:13 [PATCH v2 0/2] selftests/powerpc: Add test for rfi_flush Naveen N. Rao
  2018-05-21 15:13 ` [PATCH v2 1/2] selftests/powerpc: Move UCONTEXT_NIA() into utils.h Naveen N. Rao
@ 2018-05-21 15:13 ` Naveen N. Rao
  2018-05-24 13:20   ` Michael Ellerman
  1 sibling, 1 reply; 8+ messages in thread
From: Naveen N. Rao @ 2018-05-21 15:13 UTC (permalink / raw)
  To: Michael Ellerman; +Cc: linuxppc-dev

This adds a test to verify proper functioning of the rfi flush
capability implemented to mitigate meltdown. The test works by measuring
the number of L1d cache misses encountered while loading data from
memory. Across a system call, since the L1d cache is flushed when
rfi_flush is enabled, the number of cache misses is expected to be
relative to the number of cachelines corresponding to the data being
loaded.

The current system setting is reflected via powerpc/rfi_flush under
debugfs (assumed to be /sys/kernel/debug/). This test verifies the
expected result with rfi_flush enabled as well as when it is disabled.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
---
 tools/testing/selftests/powerpc/Makefile      |   3 +-
 .../testing/selftests/powerpc/include/utils.h |  10 ++
 .../selftests/powerpc/security/Makefile       |   7 +
 .../selftests/powerpc/security/rfi_flush.c    | 132 +++++++++++++++
 tools/testing/selftests/powerpc/utils.c       | 153 ++++++++++++++++++
 5 files changed, 304 insertions(+), 1 deletion(-)
 create mode 100644 tools/testing/selftests/powerpc/security/Makefile
 create mode 100644 tools/testing/selftests/powerpc/security/rfi_flush.c

diff --git a/tools/testing/selftests/powerpc/Makefile b/tools/testing/selftests/powerpc/Makefile
index 201b598558b9..b3ad909aefbc 100644
--- a/tools/testing/selftests/powerpc/Makefile
+++ b/tools/testing/selftests/powerpc/Makefile
@@ -28,7 +28,8 @@ SUB_DIRS = alignment		\
 	   tm			\
 	   vphn         \
 	   math		\
-	   ptrace
+	   ptrace	\
+	   security
 
 endif
 
diff --git a/tools/testing/selftests/powerpc/include/utils.h b/tools/testing/selftests/powerpc/include/utils.h
index b8f9cc6c9ff2..684b9d554639 100644
--- a/tools/testing/selftests/powerpc/include/utils.h
+++ b/tools/testing/selftests/powerpc/include/utils.h
@@ -11,6 +11,7 @@
 #include <stdint.h>
 #include <stdbool.h>
 #include <linux/auxvec.h>
+#include <linux/perf_event.h>
 #include "reg.h"
 
 /* Avoid headaches with PRI?64 - just use %ll? always */
@@ -31,6 +32,15 @@ void *get_auxv_entry(int type);
 
 int pick_online_cpu(void);
 
+int read_debugfs_file(char *debugfs_file, int *result);
+int write_debugfs_file(char *debugfs_file, int result);
+void set_dscr(unsigned long val);
+int perf_event_open_counter(unsigned int type,
+			    unsigned long config, int group_fd);
+int perf_event_enable(int fd);
+int perf_event_disable(int fd);
+int perf_event_reset(int fd);
+
 static inline bool have_hwcap(unsigned long ftr)
 {
 	return ((unsigned long)get_auxv_entry(AT_HWCAP) & ftr) == ftr;
diff --git a/tools/testing/selftests/powerpc/security/Makefile b/tools/testing/selftests/powerpc/security/Makefile
new file mode 100644
index 000000000000..8a472e13950a
--- /dev/null
+++ b/tools/testing/selftests/powerpc/security/Makefile
@@ -0,0 +1,7 @@
+TEST_GEN_PROGS := rfi_flush
+
+CFLAGS += -I../../../../../usr/include
+
+include ../../lib.mk
+
+$(TEST_GEN_PROGS): ../harness.c ../utils.c
diff --git a/tools/testing/selftests/powerpc/security/rfi_flush.c b/tools/testing/selftests/powerpc/security/rfi_flush.c
new file mode 100644
index 000000000000..a20fe8eca161
--- /dev/null
+++ b/tools/testing/selftests/powerpc/security/rfi_flush.c
@@ -0,0 +1,132 @@
+#include <sys/types.h>
+#include <stdint.h>
+#include <malloc.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stdio.h>
+#include "utils.h"
+
+#define CACHELINE_SIZE	128
+
+struct perf_event_read {
+	uint64_t nr;
+	uint64_t l1d_misses;
+};
+
+static inline uint64_t load(void *addr)
+{
+	uint64_t tmp;
+
+	asm volatile("ld %0,0(%1)" : "=r"(tmp) : "b" (addr));
+
+	return tmp;
+}
+
+static void syscall_loop(char *p, unsigned long iterations, unsigned long zero_size)
+{
+	for (unsigned long i = 0; i < iterations; i++) {
+		for (unsigned long j = 0; j < zero_size; j += CACHELINE_SIZE)
+			load(p + j);
+		getppid();
+	}
+}
+
+int rfi_flush_test(void)
+{
+	char *p;
+	int repetitions = 10;
+	int fd, passes = 0, iter, rc = 0;
+	struct perf_event_read v;
+	uint64_t l1d_misses_total = 0;
+	unsigned long iterations = 100000, zero_size = 24*1024;
+	int rfi_flush_org, rfi_flush;
+
+	SKIP_IF(geteuid() != 0);
+
+	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
+		perror("error reading powerpc/rfi_flush debugfs file");
+	        printf("unable to determine current rfi_flush setting");
+		return 1;
+	}
+
+	rfi_flush = rfi_flush_org;
+
+	FAIL_IF((fd = perf_event_open_counter(PERF_TYPE_RAW,
+					  0x400f0, /* L1d miss */
+					  -1)) < 0);
+
+	p = (char *)memalign(zero_size, CACHELINE_SIZE);
+
+	FAIL_IF(perf_event_enable(fd));
+
+	set_dscr(1);
+
+	iter = repetitions;
+
+again:
+	FAIL_IF(perf_event_reset(fd));
+
+	syscall_loop(p, iterations, zero_size);
+
+	FAIL_IF(read(fd, &v, sizeof(v)) != sizeof(v));
+
+	/* Expect at least zero_size/CACHELINE_SIZE misses per iteration */
+	if (v.l1d_misses >= (iterations * zero_size / CACHELINE_SIZE) &&
+			rfi_flush)
+		passes++;
+	else if (v.l1d_misses < iterations && !rfi_flush)
+		passes++;
+
+	l1d_misses_total += v.l1d_misses;
+
+	while (--iter)
+		goto again;
+
+	if (passes < repetitions) {
+		printf("FAIL (L1D misses with rfi_flush=%d: %lu %c %lu) [%d/%d failures]\n",
+			rfi_flush,
+			l1d_misses_total,
+			rfi_flush ? '<' : '>',
+			rfi_flush ? (repetitions * iterations * zero_size/CACHELINE_SIZE) :
+				iterations,
+			repetitions - passes, repetitions);
+		rc = 1;
+	} else
+		printf("PASS (L1D misses with rfi_flush=%d: %lu %c %lu) [%d/%d pass]\n",
+			rfi_flush,
+			l1d_misses_total,
+			rfi_flush ? '>' : '<',
+			rfi_flush ? (repetitions * iterations * zero_size/CACHELINE_SIZE) :
+				iterations,
+			passes, repetitions);
+
+	if (rfi_flush == rfi_flush_org) {
+		rfi_flush = !rfi_flush_org;
+		if (write_debugfs_file("powerpc/rfi_flush", rfi_flush) < 0) {
+			perror("error writing to powerpc/rfi_flush debugfs file");
+			return 1;
+		}
+		iter = repetitions;
+		l1d_misses_total = 0;
+		passes = 0;
+		goto again;
+	}
+
+	perf_event_disable(fd);
+	close(fd);
+
+	set_dscr(0);
+
+	if (write_debugfs_file("powerpc/rfi_flush", rfi_flush_org) < 0) {
+		perror("unable to restore original value of powerpc/rfi_flush debugfs file");
+		return 1;
+	}
+
+	return rc;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(rfi_flush_test, "rfi_flush_test");
+}
diff --git a/tools/testing/selftests/powerpc/utils.c b/tools/testing/selftests/powerpc/utils.c
index d46916867a6f..7f8b7de08c75 100644
--- a/tools/testing/selftests/powerpc/utils.c
+++ b/tools/testing/selftests/powerpc/utils.c
@@ -11,13 +11,20 @@
 #include <link.h>
 #include <sched.h>
 #include <stdio.h>
+#include <signal.h>
+#include <stdlib.h>
+#include <string.h>
 #include <sys/stat.h>
 #include <sys/types.h>
+#include <sys/ioctl.h>
 #include <unistd.h>
+#include <asm/unistd.h>
+#include <linux/limits.h>
 
 #include "utils.h"
 
 static char auxv[4096];
+extern unsigned int dscr_insn[];
 
 int read_auxv(char *buf, ssize_t buf_size)
 {
@@ -104,3 +111,149 @@ int pick_online_cpu(void)
 	printf("No cpus in affinity mask?!\n");
 	return -1;
 }
+
+int read_debugfs_file(char *debugfs_file, int *result)
+{
+	int rc = -1, fd;
+	char path[PATH_MAX];
+	char value[16];
+
+	strcpy(path, "/sys/kernel/debug/");
+	strncat(path, debugfs_file, PATH_MAX - strlen(path) - 1);
+
+	if ((fd = open(path, O_RDONLY)) < 0)
+		return rc;
+
+	if ((rc = read(fd, value, sizeof(value))) < 0)
+		return rc;
+
+	value[15] = 0;
+	*result = atoi(value);
+	close(fd);
+
+	return 0;
+}
+
+int write_debugfs_file(char *debugfs_file, int result)
+{
+	int rc = -1, fd;
+	char path[PATH_MAX];
+	char value[16];
+
+	strcpy(path, "/sys/kernel/debug/");
+	strncat(path, debugfs_file, PATH_MAX - strlen(path) - 1);
+
+	if ((fd = open(path, O_WRONLY)) < 0)
+		return rc;
+
+	snprintf(value, 16, "%d", result);
+
+	if ((rc = write(fd, value, strlen(value))) < 0)
+		return rc;
+
+	close(fd);
+
+	return 0;
+}
+
+static long perf_event_open(struct perf_event_attr *hw_event, pid_t pid,
+		int cpu, int group_fd, unsigned long flags)
+{
+	return syscall(__NR_perf_event_open, hw_event, pid, cpu,
+		      group_fd, flags);
+}
+
+static void perf_event_attr_init(struct perf_event_attr *event_attr,
+					unsigned int type,
+					unsigned long config)
+{
+	memset(event_attr, 0, sizeof(*event_attr));
+
+	event_attr->type = type;
+	event_attr->size = sizeof(struct perf_event_attr);
+	event_attr->config = config;
+	event_attr->read_format = PERF_FORMAT_GROUP;
+	event_attr->disabled = 1;
+	event_attr->exclude_kernel = 1;
+	event_attr->exclude_hv = 1;
+	event_attr->exclude_guest = 1;
+}
+
+int perf_event_open_counter(unsigned int type,
+			    unsigned long config, int group_fd)
+{
+	int fd;
+	struct perf_event_attr event_attr;
+
+	perf_event_attr_init(&event_attr, type, config);
+
+	fd = perf_event_open(&event_attr, 0, -1, group_fd, 0);
+
+	if (fd < 0)
+		perror("perf_event_open() failed");
+
+	return fd;
+}
+
+int perf_event_enable(int fd)
+{
+	if (ioctl(fd, PERF_EVENT_IOC_ENABLE, PERF_IOC_FLAG_GROUP) == -1) {
+		perror("error while enabling perf events");
+		return -1;
+	}
+
+	return 0;
+}
+
+int perf_event_disable(int fd)
+{
+	if (ioctl(fd, PERF_EVENT_IOC_DISABLE, PERF_IOC_FLAG_GROUP) == -1) {
+		perror("error disabling perf events");
+		return -1;
+	}
+
+	return 0;
+}
+
+int perf_event_reset(int fd)
+{
+	if (ioctl(fd, PERF_EVENT_IOC_RESET, PERF_IOC_FLAG_GROUP) == -1) {
+		perror("error resetting perf events");
+		return -1;
+	}
+
+	return 0;
+}
+
+static void sigill_handler(int signr, siginfo_t *info, void *unused)
+{
+	static int warned = 0;
+	ucontext_t *ctx = (ucontext_t *)unused;
+	unsigned long *pc = &UCONTEXT_NIA(ctx);
+
+	if (*pc == (unsigned long)&dscr_insn) {
+		if (!warned++)
+			printf("WARNING: Skipping over dscr setup. Consider running 'ppc64_cpu --dscr=1' manually.\n");
+		*pc += 4;
+	} else {
+		printf("SIGILL at %p\n", pc);
+		abort();
+	}
+}
+
+void set_dscr(unsigned long val)
+{
+	static int init = 0;
+	struct sigaction sa;
+
+	if (!init) {
+		memset(&sa, 0, sizeof(sa));
+		sa.sa_sigaction = sigill_handler;
+		sa.sa_flags = SA_SIGINFO;
+		if (sigaction(SIGILL, &sa, NULL))
+			perror("sigill_handler");
+		init = 1;
+	}
+
+	asm volatile("dscr_insn: mtspr %1,%0" : : "r" (val), "i" (SPRN_DSCR));
+}
-- 
2.17.0

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call
  2018-05-21 15:13 ` [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call Naveen N. Rao
@ 2018-05-24 13:20   ` Michael Ellerman
  2018-05-24 18:29     ` Naveen N. Rao
  0 siblings, 1 reply; 8+ messages in thread
From: Michael Ellerman @ 2018-05-24 13:20 UTC (permalink / raw)
  To: Naveen N. Rao; +Cc: linuxppc-dev

"Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
> diff --git a/tools/testing/selftests/powerpc/security/rfi_flush.c b/tools/testing/selftests/powerpc/security/rfi_flush.c
> new file mode 100644
> index 000000000000..a20fe8eca161
> --- /dev/null
> +++ b/tools/testing/selftests/powerpc/security/rfi_flush.c
> @@ -0,0 +1,132 @@
...
> +
> +int rfi_flush_test(void)
> +{
> +	char *p;
> +	int repetitions = 10;
> +	int fd, passes = 0, iter, rc = 0;
> +	struct perf_event_read v;
> +	uint64_t l1d_misses_total = 0;
> +	unsigned long iterations = 100000, zero_size = 24*1024;
> +	int rfi_flush_org, rfi_flush;
> +
> +	SKIP_IF(geteuid() != 0);
> +
> +	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
> +		perror("error reading powerpc/rfi_flush debugfs file");
> +	        printf("unable to determine current rfi_flush setting");
> +		return 1;
> +	}

This leads to a hard error on old kernels, which I don't want (breaks my CI).

So I changed it to:

	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
		perror("Unable to read powerpc/rfi_flush debugfs file");
		SKIP_IF(1);
	}

cheers

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call
  2018-05-24 13:20   ` Michael Ellerman
@ 2018-05-24 18:29     ` Naveen N. Rao
  2018-05-25 10:39       ` Michael Ellerman
  0 siblings, 1 reply; 8+ messages in thread
From: Naveen N. Rao @ 2018-05-24 18:29 UTC (permalink / raw)
  To: Michael Ellerman; +Cc: linuxppc-dev

Michael Ellerman wrote:
> "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
>> diff --git a/tools/testing/selftests/powerpc/security/rfi_flush.c b/tool=
s/testing/selftests/powerpc/security/rfi_flush.c
>> new file mode 100644
>> index 000000000000..a20fe8eca161
>> --- /dev/null
>> +++ b/tools/testing/selftests/powerpc/security/rfi_flush.c
>> @@ -0,0 +1,132 @@
> ...
>> +
>> +int rfi_flush_test(void)
>> +{
>> +	char *p;
>> +	int repetitions =3D 10;
>> +	int fd, passes =3D 0, iter, rc =3D 0;
>> +	struct perf_event_read v;
>> +	uint64_t l1d_misses_total =3D 0;
>> +	unsigned long iterations =3D 100000, zero_size =3D 24*1024;
>> +	int rfi_flush_org, rfi_flush;
>> +
>> +	SKIP_IF(geteuid() !=3D 0);
>> +
>> +	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
>> +		perror("error reading powerpc/rfi_flush debugfs file");
>> +	        printf("unable to determine current rfi_flush setting");
>> +		return 1;
>> +	}
>=20
> This leads to a hard error on old kernels, which I don't want (breaks my =
CI).

Hmm... didn't realize new tests would be run on older kernels. Are those=20
older stable/distro releases, or actually just earlier kernel versions?

>=20
> So I changed it to:
>=20
> 	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
> 		perror("Unable to read powerpc/rfi_flush debugfs file");
> 		SKIP_IF(1);
> 	}

Sure, thanks!
- Naveen

=

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call
  2018-05-24 18:29     ` Naveen N. Rao
@ 2018-05-25 10:39       ` Michael Ellerman
  2018-05-25 11:30         ` Naveen N. Rao
  0 siblings, 1 reply; 8+ messages in thread
From: Michael Ellerman @ 2018-05-25 10:39 UTC (permalink / raw)
  To: Naveen N. Rao; +Cc: linuxppc-dev

"Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
> Michael Ellerman wrote:
>> "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
>>> diff --git a/tools/testing/selftests/powerpc/security/rfi_flush.c b/tools/testing/selftests/powerpc/security/rfi_flush.c
>>> new file mode 100644
>>> index 000000000000..a20fe8eca161
>>> --- /dev/null
>>> +++ b/tools/testing/selftests/powerpc/security/rfi_flush.c
>>> @@ -0,0 +1,132 @@
>> ...
>>> +
>>> +int rfi_flush_test(void)
>>> +{
>>> +	char *p;
>>> +	int repetitions = 10;
>>> +	int fd, passes = 0, iter, rc = 0;
>>> +	struct perf_event_read v;
>>> +	uint64_t l1d_misses_total = 0;
>>> +	unsigned long iterations = 100000, zero_size = 24*1024;
>>> +	int rfi_flush_org, rfi_flush;
>>> +
>>> +	SKIP_IF(geteuid() != 0);
>>> +
>>> +	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
>>> +		perror("error reading powerpc/rfi_flush debugfs file");
>>> +	        printf("unable to determine current rfi_flush setting");
>>> +		return 1;
>>> +	}
>> 
>> This leads to a hard error on old kernels, which I don't want (breaks my CI).
>
> Hmm... didn't realize new tests would be run on older kernels. Are those 
> older stable/distro releases, or actually just earlier kernel versions?

All of the above :)

Obviously keeping all the tests working across all kernel versions ever
is impossible, but when it's relatively easy like in this case, it's
nice to do.

So that was mostly just an FYI, I don't expect test writers to get their
tests running across all kernel versions.

cheers

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call
  2018-05-25 10:39       ` Michael Ellerman
@ 2018-05-25 11:30         ` Naveen N. Rao
  0 siblings, 0 replies; 8+ messages in thread
From: Naveen N. Rao @ 2018-05-25 11:30 UTC (permalink / raw)
  To: Michael Ellerman; +Cc: linuxppc-dev

Michael Ellerman wrote:
> "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
>> Michael Ellerman wrote:
>>> "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
>>>> diff --git a/tools/testing/selftests/powerpc/security/rfi_flush.c b/to=
ols/testing/selftests/powerpc/security/rfi_flush.c
>>>> new file mode 100644
>>>> index 000000000000..a20fe8eca161
>>>> --- /dev/null
>>>> +++ b/tools/testing/selftests/powerpc/security/rfi_flush.c
>>>> @@ -0,0 +1,132 @@
>>> ...
>>>> +
>>>> +int rfi_flush_test(void)
>>>> +{
>>>> +	char *p;
>>>> +	int repetitions =3D 10;
>>>> +	int fd, passes =3D 0, iter, rc =3D 0;
>>>> +	struct perf_event_read v;
>>>> +	uint64_t l1d_misses_total =3D 0;
>>>> +	unsigned long iterations =3D 100000, zero_size =3D 24*1024;
>>>> +	int rfi_flush_org, rfi_flush;
>>>> +
>>>> +	SKIP_IF(geteuid() !=3D 0);
>>>> +
>>>> +	if (read_debugfs_file("powerpc/rfi_flush", &rfi_flush_org)) {
>>>> +		perror("error reading powerpc/rfi_flush debugfs file");
>>>> +	        printf("unable to determine current rfi_flush setting");
>>>> +		return 1;
>>>> +	}
>>>=20
>>> This leads to a hard error on old kernels, which I don't want (breaks m=
y CI).
>>
>> Hmm... didn't realize new tests would be run on older kernels. Are those=
=20
>> older stable/distro releases, or actually just earlier kernel versions?
>=20
> All of the above :)
>=20
> Obviously keeping all the tests working across all kernel versions ever
> is impossible, but when it's relatively easy like in this case, it's
> nice to do.
>=20
> So that was mostly just an FYI, I don't expect test writers to get their
> tests running across all kernel versions.

Good to know, thanks!
- Naveen

=

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [v2,1/2] selftests/powerpc: Move UCONTEXT_NIA() into utils.h
  2018-05-21 15:13 ` [PATCH v2 1/2] selftests/powerpc: Move UCONTEXT_NIA() into utils.h Naveen N. Rao
@ 2018-10-22  9:34   ` Michael Ellerman
  0 siblings, 0 replies; 8+ messages in thread
From: Michael Ellerman @ 2018-10-22  9:34 UTC (permalink / raw)
  To: Naveen N. Rao; +Cc: linuxppc-dev

On Mon, 2018-05-21 at 15:13:56 UTC, "Naveen N. Rao" wrote:
> ... so that it can be used by others.
> 
> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>

Series applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/db3848515aff369404c31ef5a63699

cheers

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2018-10-22  9:38 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-05-21 15:13 [PATCH v2 0/2] selftests/powerpc: Add test for rfi_flush Naveen N. Rao
2018-05-21 15:13 ` [PATCH v2 1/2] selftests/powerpc: Move UCONTEXT_NIA() into utils.h Naveen N. Rao
2018-10-22  9:34   ` [v2,1/2] " Michael Ellerman
2018-05-21 15:13 ` [PATCH v2 2/2] selftests/powerpc: Add test to verify rfi flush across a system call Naveen N. Rao
2018-05-24 13:20   ` Michael Ellerman
2018-05-24 18:29     ` Naveen N. Rao
2018-05-25 10:39       ` Michael Ellerman
2018-05-25 11:30         ` Naveen N. Rao

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.