Skip to contentSkip to conversation

perf: Support AMD range breakpoints

PATCH In review Frederic Weisbecker 4 participants 17 messages May 29, 2014 – Aug 4, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover perf: Support AMD range breakpoints

Hi,

I've made it wait too much so I'm posting that for review now. This implements range breakpoints on some AMD machine by extending the length field.

Any comment?

Thanks.

Jacob Shin (4):
perf/x86/amd: AMD support for bp_len > HW_BREAKPOINT_LEN_8 perf tools: allow user to specify hardware breakpoint bp_len perf tools: add hardware breakpoint bp_len test cases perf/x86: Remove get_hbp_len and replace with bp_len

 arch/x86/include/asm/cpufeature.h        |  2 ++
 arch/x86/include/asm/debugreg.h          |  5 +++
 arch/x86/include/asm/hw_breakpoint.h     |  1 +
 arch/x86/include/uapi/asm/msr-index.h    |  4 +++
 arch/x86/kernel/cpu/amd.c                | 19 +++++++++++
 arch/x86/kernel/hw_breakpoint.c          | 45 ++++++++++---------------
 tools/perf/Documentation/perf-record.txt |  7 ++--
 tools/perf/tests/parse-events.c          | 58 ++++++++++++++++++++++++++++++++
 tools/perf/util/parse-events.c           | 21 ++++++------
 tools/perf/util/parse-events.h           |  2 +-
 tools/perf/util/parse-events.l           |  1 +
 tools/perf/util/parse-events.y           | 26 ++++++++++++--
 12 files changed, 148 insertions(+), 43 deletions(-)
Signature · 1 line
1.8.3.1
#1 lore raw

PATCH 1/4 perf/x86/amd: AMD support for bp_len > HW_BREAKPOINT_LEN_8

From: Jacob Shin <[email protected]>

Implement hardware breakpoint address mask for AMD Family 16h and above processors. CPUID feature bit indicates hardware support for DRn_ADDR_MASK MSRs. These masks further qualify DRn/DR7 hardware breakpoint addresses to allow matching of larger addresses ranges.

Valuable advice and pseudo code from Oleg Nesterov <[email protected]>

Cc Arnaldo Carvalho de Melo <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc xiakaixu <[email protected]>
---
 arch/x86/include/asm/cpufeature.h     |  2 ++
 arch/x86/include/asm/debugreg.h       |  5 +++++
 arch/x86/include/asm/hw_breakpoint.h  |  1 +
 arch/x86/include/uapi/asm/msr-index.h |  4 ++++
 arch/x86/kernel/cpu/amd.c             | 19 +++++++++++++++++++
 arch/x86/kernel/hw_breakpoint.c       | 20 ++++++++++++++++----
 6 files changed, 47 insertions(+), 4 deletions(-)
arch/x86/include/asm/cpufeature.h +2 -0
index e265ff9..6bd564d 100644
@@ -170,6 +170,7 @@
 #define X86_FEATURE_TOPOEXT	(6*32+22) /* topology extensions CPUID leafs */
 #define X86_FEATURE_PERFCTR_CORE (6*32+23) /* core performance counter extensions */
 #define X86_FEATURE_PERFCTR_NB  (6*32+24) /* NB performance counter extensions */
+#define X86_FEATURE_BPEXT	(6*32+26) /* data breakpoint extension */
 #define X86_FEATURE_PERFCTR_L2	(6*32+28) /* L2 performance counter extensions */
 
 /*
@@ -338,6 +339,7 @@ extern const char * const x86_power_flags[32];
 #define cpu_has_cx16		boot_cpu_has(X86_FEATURE_CX16)
 #define cpu_has_eager_fpu	boot_cpu_has(X86_FEATURE_EAGER_FPU)
 #define cpu_has_topoext		boot_cpu_has(X86_FEATURE_TOPOEXT)
+#define cpu_has_bpext		boot_cpu_has(X86_FEATURE_BPEXT)
 
 #ifdef CONFIG_X86_64
 
arch/x86/include/asm/debugreg.h +5 -0
index 4b528a9..145b009 100644
@@ -114,5 +114,10 @@ static inline void debug_stack_usage_inc(void) { }
 static inline void debug_stack_usage_dec(void) { }
 #endif /* X86_64 */
 
+#ifdef CONFIG_CPU_SUP_AMD
+extern void set_dr_addr_mask(unsigned long mask, int dr);
+#else
+static inline void set_dr_addr_mask(unsigned long mask, int dr) { }
+#endif
 
 #endif /* _ASM_X86_DEBUGREG_H */
arch/x86/include/asm/hw_breakpoint.h +1 -0
index ef1c4d2..6c98be8 100644
@@ -12,6 +12,7 @@
  */
 struct arch_hw_breakpoint {
 	unsigned long	address;
+	unsigned long	mask;
 	u8		len;
 	u8		type;
 };
arch/x86/include/uapi/asm/msr-index.h +4 -0
index c827ace..d6b767c 100644
@@ -208,6 +208,10 @@
 /* Fam 16h MSRs */
 #define MSR_F16H_L2I_PERF_CTL		0xc0010230
 #define MSR_F16H_L2I_PERF_CTR		0xc0010231
+#define MSR_F16H_DR1_ADDR_MASK		0xc0011019
+#define MSR_F16H_DR2_ADDR_MASK		0xc001101a
+#define MSR_F16H_DR3_ADDR_MASK		0xc001101b
+#define MSR_F16H_DR0_ADDR_MASK		0xc0011027
 
 /* Fam 15h MSRs */
 #define MSR_F15H_PERF_CTL		0xc0010200
arch/x86/kernel/cpu/amd.c +19 -0
index ce8b8ff..2d1fe7a 100644
@@ -887,3 +887,22 @@ static bool cpu_has_amd_erratum(struct cpuinfo_x86 *cpu, const int *erratum)
 
 	return false;
 }
+
+void set_dr_addr_mask(unsigned long mask, int dr)
+{
+	if (!cpu_has_bpext)
+		return;
+
+	switch (dr) {
+	case 0:
+		wrmsr(MSR_F16H_DR0_ADDR_MASK, mask, 0);
+		break;
+	case 1:
+	case 2:
+	case 3:
+		wrmsr(MSR_F16H_DR1_ADDR_MASK - 1 + dr, mask, 0);
+		break;
+	default:
+		break;
+	}
+}
arch/x86/kernel/hw_breakpoint.c +16 -4
index a67b47c..fa24eb5 100644
@@ -127,6 +127,8 @@ int arch_install_hw_breakpoint(struct perf_event *bp)
 	*dr7 |= encode_dr7(i, info->len, info->type);
 
 	set_debugreg(*dr7, 7);
+	if (info->mask)
+		set_dr_addr_mask(info->mask, i);
 
 	return 0;
 }
@@ -162,6 +164,8 @@ void arch_uninstall_hw_breakpoint(struct perf_event *bp)
 	*dr7 &= ~__encode_dr7(i, info->len, info->type);
 
 	set_debugreg(*dr7, 7);
+	if (info->mask)
+		set_dr_addr_mask(0, i);
 }
 
 static int get_hbp_len(u8 hbp_len)
@@ -278,6 +282,8 @@ static int arch_build_bp_info(struct perf_event *bp)
 	}
 
 	/* Len */
+	info->mask = 0;
+
 	switch (bp->attr.bp_len) {
 	case HW_BREAKPOINT_LEN_1:
 		info->len = X86_BREAKPOINT_LEN_1;
@@ -294,11 +300,17 @@ static int arch_build_bp_info(struct perf_event *bp)
 		break;
 #endif
 	default:
-		return -EINVAL;
+		if (!is_power_of_2(bp->attr.bp_len))
+			return -EINVAL;
+		if (!cpu_has_bpext)
+			return -EOPNOTSUPP;
+		info->mask = bp->attr.bp_len - 1;
+		info->len = X86_BREAKPOINT_LEN_1;
 	}
 
 	return 0;
 }
+
 /*
  * Validate the arch-specific HW Breakpoint register settings
  */
@@ -313,11 +325,11 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp)
 	if (ret)
 		return ret;
 
-	ret = -EINVAL;
-
 	switch (info->len) {
 	case X86_BREAKPOINT_LEN_1:
 		align = 0;
+		if (info->mask)
+			align = info->mask;
 		break;
 	case X86_BREAKPOINT_LEN_2:
 		align = 1;
@@ -331,7 +343,7 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp)
 		break;
 #endif
 	default:
-		return ret;
+		WARN_ON_ONCE(1);
 	}
 
 	/*
Signature · 1 line
1.8.3.1
#2 lore raw

Re: [PATCH 1/4] perf/x86/amd: AMD support for bp_len > HW_BREAKPOINT_LEN_8

Hi Jacob, Suravee,

Quoted 53 lines · Frederic Weisbecker

From: Jacob Shin <[email protected]>

Implement hardware breakpoint address mask for AMD Family 16h and above processors. CPUID feature bit indicates hardware support for DRn_ADDR_MASK MSRs. These masks further qualify DRn/DR7 hardware breakpoint addresses to allow matching of larger addresses ranges.

Valuable advice and pseudo code from Oleg Nesterov <[email protected]>

Cc Arnaldo Carvalho de Melo <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc xiakaixu <[email protected]>
---
 arch/x86/include/asm/cpufeature.h     |  2 ++
 arch/x86/include/asm/debugreg.h       |  5 +++++
 arch/x86/include/asm/hw_breakpoint.h  |  1 +
 arch/x86/include/uapi/asm/msr-index.h |  4 ++++
 arch/x86/kernel/cpu/amd.c             | 19 +++++++++++++++++++
 arch/x86/kernel/hw_breakpoint.c       | 20 ++++++++++++++++----
 6 files changed, 47 insertions(+), 4 deletions(-)
arch/x86/include/asm/cpufeature.h +2 -0
index e265ff9..6bd564d 100644
@@ -170,6 +170,7 @@
 #define X86_FEATURE_TOPOEXT	(6*32+22) /* topology extensions CPUID leafs */
 #define X86_FEATURE_PERFCTR_CORE (6*32+23) /* core performance counter extensions */
 #define X86_FEATURE_PERFCTR_NB  (6*32+24) /* NB performance counter extensions */
+#define X86_FEATURE_BPEXT	(6*32+26) /* data breakpoint extension */
 #define X86_FEATURE_PERFCTR_L2	(6*32+28) /* L2 performance counter extensions */
 
 /*
@@ -338,6 +339,7 @@ extern const char * const x86_power_flags[32];
 #define cpu_has_cx16		boot_cpu_has(X86_FEATURE_CX16)
 #define cpu_has_eager_fpu	boot_cpu_has(X86_FEATURE_EAGER_FPU)
 #define cpu_has_topoext		boot_cpu_has(X86_FEATURE_TOPOEXT)
+#define cpu_has_bpext		boot_cpu_has(X86_FEATURE_BPEXT)
 
 #ifdef CONFIG_X86_64
 
arch/x86/include/asm/debugreg.h +0 -0
index 4b528a9..145b009 100644

arch/x86/include/asm/debugreg.h static inline void debug_stack_usage_inc(void) { } · line 121

Quoted 8 lines
arch/x86/include/asm/debugreg.h +5 -0
@@ -114,5 +114,10 @@ static inline void debug_stack_usage_inc(void) { }
 static inline void debug_stack_usage_dec(void) { }
 #endif /* X86_64 */
 
+#ifdef CONFIG_CPU_SUP_AMD
+extern void set_dr_addr_mask(unsigned long mask, int dr);
+#else
+static inline void set_dr_addr_mask(unsigned long mask, int dr) { }
+#endif

I see this symbol in the code but it's not defined anywhere in a Kconfig file. Maybe you (or I) forgot to include a file in your patches?

Thanks.

#3 lore raw

Re: [PATCH 1/4] perf/x86/amd: AMD support for bp_len > HW_BREAKPOINT_LEN_8

Quoted 16 lines · Frederic Weisbecker
Quoted 8 lines
arch/x86/include/asm/debugreg.h +0 -0
Quoted 7 lines

index 4b528a9..145b009 100644

arch/x86/include/asm/debugreg.h +0 -0
@@ -114,5 +114,10 @@ static inline void debug_stack_usage_inc(void) { }
 static inline void debug_stack_usage_dec(void) { }
 #endif /* X86_64 */
 
Quoted 5 lines

arch/x86/include/asm/debugreg.h static inline void debug_stack_usage_inc(void) { } · line 121

Quoted 5 lines
arch/x86/include/asm/debugreg.h +5 -0
+#ifdef CONFIG_CPU_SUP_AMD
+extern void set_dr_addr_mask(unsigned long mask, int dr);
+#else
+static inline void set_dr_addr_mask(unsigned long mask, int dr) { }
+#endif

I see this symbol in the code but it's not defined anywhere in a Kconfig file. Maybe you (or I) forgot to include a file in your patches?

Quoted 1 line

Thanks.

If you are referring to the CONFIG_CPU_SUP_AMD, this is in arch/x86/Kconfig.cpu

Suravee

#4 lore raw

Re: [PATCH 1/4] perf/x86/amd: AMD support for bp_len > HW_BREAKPOINT_LEN_8

Quoted 19 lines · Suravee Suthikulpanit
Quoted 16 lines · Frederic Weisbecker
Quoted 8 lines
arch/x86/include/asm/debugreg.h +0 -0
Quoted 7 lines

index 4b528a9..145b009 100644

arch/x86/include/asm/debugreg.h +0 -0
@@ -114,5 +114,10 @@ static inline void debug_stack_usage_inc(void) { }
 static inline void debug_stack_usage_dec(void) { }
 #endif /* X86_64 */
 
Quoted 5 lines

arch/x86/include/asm/debugreg.h static inline void debug_stack_usage_inc(void) { } · line 121

Quoted 5 lines
arch/x86/include/asm/debugreg.h +5 -0
+#ifdef CONFIG_CPU_SUP_AMD
+extern void set_dr_addr_mask(unsigned long mask, int dr);
+#else
+static inline void set_dr_addr_mask(unsigned long mask, int dr) { }
+#endif

I see this symbol in the code but it's not defined anywhere in a Kconfig file. Maybe you (or I) forgot to include a file in your patches?

Quoted 4 lines
Quoted 1 line

Thanks.

If you are referring to the CONFIG_CPU_SUP_AMD, this is in arch/x86/Kconfig.cpu

Ah I thought it was created in the patchset. Nevermind.

#5 lore raw

PATCH 2/4 perf tools: allow user to specify hardware breakpoint bp_len

In reply to Frederic Weisbecker

From: Jacob Shin <[email protected]>

Currently bp_len is given a default value of 4. Allow user to override it:

  $ perf stat -e mem:0x1000/8
                            ^
                            bp_len

If no value is given, it will default to 4 as it did before.

Cc Arnaldo Carvalho de Melo <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc xiakaixu <[email protected]>
---
 tools/perf/Documentation/perf-record.txt |  7 +++++--
 tools/perf/util/parse-events.c           | 21 +++++++++++----------
 tools/perf/util/parse-events.h           |  2 +-
 tools/perf/util/parse-events.l           |  1 +
 tools/perf/util/parse-events.y           | 26 ++++++++++++++++++++++++--
 5 files changed, 42 insertions(+), 15 deletions(-)
tools/perf/Documentation/perf-record.txt +5 -2
index c71b0f3..5cac20d 100644
@@ -33,12 +33,15 @@ OPTIONS
         - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
 	  hexadecimal event descriptor.
 
-        - a hardware breakpoint event in the form of '\mem:addr[:access]'
+        - a hardware breakpoint event in the form of '\mem:addr[/len][:access]'
           where addr is the address in memory you want to break in.
           Access is the memory access type (read, write, execute) it can
-          be passed as follows: '\mem:addr[:[r][w][x]]'.
+          be passed as follows: '\mem:addr[:[r][w][x]]'. len is the range,
+          number of bytes from specified addr, which the breakpoint will cover.
           If you want to profile read-write accesses in 0x1000, just set
           'mem:0x1000:rw'.
+          If you want to profile write accesses in [0x1000~1008), just set
+          'mem:0x1000/8:w'.
 
 --filter=<filter>::
         Event filter.
tools/perf/util/parse-events.c +11 -10
index 1e15df1..8dd9287 100644
@@ -516,7 +516,7 @@ do {					\
 }
 
 int parse_events_add_breakpoint(struct list_head *list, int *idx,
-				void *ptr, char *type)
+				void *ptr, char *type, u64 len)
 {
 	struct perf_event_attr attr;
 
@@ -526,14 +526,15 @@ int parse_events_add_breakpoint(struct list_head *list, int *idx,
 	if (parse_breakpoint_type(type, &attr))
 		return -EINVAL;
 
-	/*
-	 * We should find a nice way to override the access length
-	 * Provide some defaults for now
-	 */
-	if (attr.bp_type == HW_BREAKPOINT_X)
-		attr.bp_len = sizeof(long);
-	else
-		attr.bp_len = HW_BREAKPOINT_LEN_4;
+	/* Provide some defaults if len is not specified */
+	if (!len) {
+		if (attr.bp_type == HW_BREAKPOINT_X)
+			len = sizeof(long);
+		else
+			len = HW_BREAKPOINT_LEN_4;
+	}
+
+	attr.bp_len = len;
 
 	attr.type = PERF_TYPE_BREAKPOINT;
 	attr.sample_period = 1;
@@ -1262,7 +1263,7 @@ void print_events(const char *event_glob, bool name_only)
 		printf("\n");
 
 		printf("  %-50s [%s]\n",
-		       "mem:<addr>[:access]",
+		       "mem:<addr>[/len][:access]",
 			event_type_descriptors[PERF_TYPE_BREAKPOINT]);
 		printf("\n");
 	}
tools/perf/util/parse-events.h +1 -1
index df094b4..00a5eb4 100644
@@ -92,7 +92,7 @@ int parse_events_add_numeric(struct list_head *list, int *idx,
 int parse_events_add_cache(struct list_head *list, int *idx,
 			   char *type, char *op_result1, char *op_result2);
 int parse_events_add_breakpoint(struct list_head *list, int *idx,
-				void *ptr, char *type);
+				void *ptr, char *type, u64 len);
 int parse_events_add_pmu(struct list_head *list, int *idx,
 			 char *pmu , struct list_head *head_config);
 void parse_events__set_leader(char *name, struct list_head *list);
tools/perf/util/parse-events.l +1 -0
index 3432995..199e18a 100644
@@ -141,6 +141,7 @@ branch_type		{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE
 <mem>{
 {modifier_bp}		{ return str(yyscanner, PE_MODIFIER_BP); }
 :			{ return ':'; }
+"/"			{ return '/'; }
 {num_dec}		{ return value(yyscanner, 10); }
 {num_hex}		{ return value(yyscanner, 16); }
 	/*
tools/perf/util/parse-events.y +24 -2
index 0bc87ba..456f4de 100644
@@ -276,6 +276,28 @@ PE_NAME_CACHE_TYPE
 }
 
 event_legacy_mem:
+PE_PREFIX_MEM PE_VALUE '/' PE_VALUE ':' PE_MODIFIER_BP sep_dc
+{
+	struct parse_events_evlist *data = _data;
+	struct list_head *list;
+
+	ALLOC_LIST(list);
+	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
+					     (void *) $2, $6, $4));
+	$$ = list;
+}
+|
+PE_PREFIX_MEM PE_VALUE '/' PE_VALUE sep_dc
+{
+	struct parse_events_evlist *data = _data;
+	struct list_head *list;
+
+	ALLOC_LIST(list);
+	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
+					     (void *) $2, NULL, $4));
+	$$ = list;
+}
+|
 PE_PREFIX_MEM PE_VALUE ':' PE_MODIFIER_BP sep_dc
 {
 	struct parse_events_evlist *data = _data;
@@ -283,7 +305,7 @@ PE_PREFIX_MEM PE_VALUE ':' PE_MODIFIER_BP sep_dc
 
 	ALLOC_LIST(list);
 	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
-					     (void *) $2, $4));
+					     (void *) $2, $4, 0));
 	$$ = list;
 }
 |
@@ -294,7 +316,7 @@ PE_PREFIX_MEM PE_VALUE sep_dc
 
 	ALLOC_LIST(list);
 	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
-					     (void *) $2, NULL));
+					     (void *) $2, NULL, 0));
 	$$ = list;
 }
 
Signature · 1 line
1.8.3.1
#6 lore raw

Re: [PATCH 2/4] perf tools: allow user to specify hardware breakpoint bp_len

Quoted 9 lines · Frederic Weisbecker

From: Jacob Shin <[email protected]>

Currently bp_len is given a default value of 4. Allow user to override it:

  $ perf stat -e mem:0x1000/8
                            ^
                            bp_len

If no value is given, it will default to 4 as it did before.

Namhyung,
both perf tols patches from this patchset mess up with hists tests.. I havent found any connection yet.. any idea? ;-)

[root@krava perf]# ./perf test
...

19: Test software clock events have valid period values    : Ok
20: Test converting perf time to TSC                       : Ok
21: Test object code reading                               : Ok
22: Test sample parsing                                    : Ok
23: Test using a dummy software event to keep tracking     : Ok
24: Test parsing with no sample_id_all bit set             : Ok
25: Test dwarf unwind                                      : Ok
26: Test filtering hist entries                            : FAILED!
27: Test mmap thread lookup                                : Ok
28: Test thread mg sharing                                 : Ok
29: Test output sorting of hist entries                    : FAILED!
30: Test cumulation of child hist entries                  : FAILED!

thanks,
jirka

Quoted 166 lines
Cc Arnaldo Carvalho de Melo <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc xiakaixu <[email protected]>
---
 tools/perf/Documentation/perf-record.txt |  7 +++++--
 tools/perf/util/parse-events.c           | 21 +++++++++++----------
 tools/perf/util/parse-events.h           |  2 +-
 tools/perf/util/parse-events.l           |  1 +
 tools/perf/util/parse-events.y           | 26 ++++++++++++++++++++++++--
 5 files changed, 42 insertions(+), 15 deletions(-)
tools/perf/Documentation/perf-record.txt +5 -2
index c71b0f3..5cac20d 100644
@@ -33,12 +33,15 @@ OPTIONS
         - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
 	  hexadecimal event descriptor.
 
-        - a hardware breakpoint event in the form of '\mem:addr[:access]'
+        - a hardware breakpoint event in the form of '\mem:addr[/len][:access]'
           where addr is the address in memory you want to break in.
           Access is the memory access type (read, write, execute) it can
-          be passed as follows: '\mem:addr[:[r][w][x]]'.
+          be passed as follows: '\mem:addr[:[r][w][x]]'. len is the range,
+          number of bytes from specified addr, which the breakpoint will cover.
           If you want to profile read-write accesses in 0x1000, just set
           'mem:0x1000:rw'.
+          If you want to profile write accesses in [0x1000~1008), just set
+          'mem:0x1000/8:w'.
 
 --filter=<filter>::
         Event filter.
tools/perf/util/parse-events.c +11 -10
index 1e15df1..8dd9287 100644
@@ -516,7 +516,7 @@ do {					\
 }
 
 int parse_events_add_breakpoint(struct list_head *list, int *idx,
-				void *ptr, char *type)
+				void *ptr, char *type, u64 len)
 {
 	struct perf_event_attr attr;
 
@@ -526,14 +526,15 @@ int parse_events_add_breakpoint(struct list_head *list, int *idx,
 	if (parse_breakpoint_type(type, &attr))
 		return -EINVAL;
 
-	/*
-	 * We should find a nice way to override the access length
-	 * Provide some defaults for now
-	 */
-	if (attr.bp_type == HW_BREAKPOINT_X)
-		attr.bp_len = sizeof(long);
-	else
-		attr.bp_len = HW_BREAKPOINT_LEN_4;
+	/* Provide some defaults if len is not specified */
+	if (!len) {
+		if (attr.bp_type == HW_BREAKPOINT_X)
+			len = sizeof(long);
+		else
+			len = HW_BREAKPOINT_LEN_4;
+	}
+
+	attr.bp_len = len;
 
 	attr.type = PERF_TYPE_BREAKPOINT;
 	attr.sample_period = 1;
@@ -1262,7 +1263,7 @@ void print_events(const char *event_glob, bool name_only)
 		printf("\n");
 
 		printf("  %-50s [%s]\n",
-		       "mem:<addr>[:access]",
+		       "mem:<addr>[/len][:access]",
 			event_type_descriptors[PERF_TYPE_BREAKPOINT]);
 		printf("\n");
 	}
tools/perf/util/parse-events.h +1 -1
index df094b4..00a5eb4 100644
@@ -92,7 +92,7 @@ int parse_events_add_numeric(struct list_head *list, int *idx,
 int parse_events_add_cache(struct list_head *list, int *idx,
 			   char *type, char *op_result1, char *op_result2);
 int parse_events_add_breakpoint(struct list_head *list, int *idx,
-				void *ptr, char *type);
+				void *ptr, char *type, u64 len);
 int parse_events_add_pmu(struct list_head *list, int *idx,
 			 char *pmu , struct list_head *head_config);
 void parse_events__set_leader(char *name, struct list_head *list);
tools/perf/util/parse-events.l +1 -0
index 3432995..199e18a 100644
@@ -141,6 +141,7 @@ branch_type		{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE
 <mem>{
 {modifier_bp}		{ return str(yyscanner, PE_MODIFIER_BP); }
 :			{ return ':'; }
+"/"			{ return '/'; }
 {num_dec}		{ return value(yyscanner, 10); }
 {num_hex}		{ return value(yyscanner, 16); }
 	/*
tools/perf/util/parse-events.y +24 -2
index 0bc87ba..456f4de 100644
@@ -276,6 +276,28 @@ PE_NAME_CACHE_TYPE
 }
 
 event_legacy_mem:
+PE_PREFIX_MEM PE_VALUE '/' PE_VALUE ':' PE_MODIFIER_BP sep_dc
+{
+	struct parse_events_evlist *data = _data;
+	struct list_head *list;
+
+	ALLOC_LIST(list);
+	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
+					     (void *) $2, $6, $4));
+	$$ = list;
+}
+|
+PE_PREFIX_MEM PE_VALUE '/' PE_VALUE sep_dc
+{
+	struct parse_events_evlist *data = _data;
+	struct list_head *list;
+
+	ALLOC_LIST(list);
+	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
+					     (void *) $2, NULL, $4));
+	$$ = list;
+}
+|
 PE_PREFIX_MEM PE_VALUE ':' PE_MODIFIER_BP sep_dc
 {
 	struct parse_events_evlist *data = _data;
@@ -283,7 +305,7 @@ PE_PREFIX_MEM PE_VALUE ':' PE_MODIFIER_BP sep_dc
 
 	ALLOC_LIST(list);
 	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
-					     (void *) $2, $4));
+					     (void *) $2, $4, 0));
 	$$ = list;
 }
 |
@@ -294,7 +316,7 @@ PE_PREFIX_MEM PE_VALUE sep_dc
 
 	ALLOC_LIST(list);
 	ABORT_ON(parse_events_add_breakpoint(list, &data->idx,
-					     (void *) $2, NULL));
+					     (void *) $2, NULL, 0));
 	$$ = list;
 }
 
Signature · 1 line
1.8.3.1
#7 lore raw

Re: [PATCH 2/4] perf tools: allow user to specify hardware breakpoint bp_len

Hi Jiri,

Quoted 9 lines · Jiri Olsa
Quoted 8 lines · Frederic Weisbecker

From: Jacob Shin <[email protected]>

Currently bp_len is given a default value of 4. Allow user to override it:

  $ perf stat -e mem:0x1000/8
                            ^
                            bp_len
Quoted 5 lines
Quoted 1 line

If no value is given, it will default to 4 as it did before.

Namhyung,
both perf tols patches from this patchset mess up with hists tests.. I havent found any connection yet.. any idea? ;-)

So you already found the problem in the hpp->elide change and that's the reason of the failure, right? :)

Thanks,
Namhyung

#8 lore raw

Re: [PATCH 2/4] perf tools: allow user to specify hardware breakpoint bp_len

Quoted 12 lines · Namhyung Kim

Hi Jiri,

Quoted 9 lines · Jiri Olsa
Quoted 8 lines · Frederic Weisbecker

From: Jacob Shin <[email protected]>

Currently bp_len is given a default value of 4. Allow user to override it:

  $ perf stat -e mem:0x1000/8
                            ^
                            bp_len
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

If no value is given, it will default to 4 as it did before.

Namhyung,
both perf tols patches from this patchset mess up with hists tests.. I havent found any connection yet.. any idea? ;-)

So you already found the problem in the hpp->elide change and that's the reason of the failure, right? :)

I haven't got a chance to test it yet.. but it looks like thats the case.. I'll retest soon ;-)

jirka

#9 lore raw

Re: [PATCH 2/4] perf tools: allow user to specify hardware breakpoint bp_len

Quoted 19 lines · Jiri Olsa
Quoted 18 lines · Namhyung Kim

Hi Jiri,

Quoted 14 lines · Jiri Olsa
Quoted 9 lines · Frederic Weisbecker

From: Jacob Shin <[email protected]>

Currently bp_len is given a default value of 4. Allow user to override it:

   $ perf stat -e mem:0x1000/8
                             ^
                             bp_len

If no value is given, it will default to 4 as it did before.

Namhyung,
both perf tols patches from this patchset mess up with hists tests.. I havent found any connection yet.. any idea? ;-)

Quoted 8 lines
Quoted 2 lines

So you already found the problem in the hpp->elide change and that's the reason of the failure, right? :)

I haven't got a chance to test it yet.. but it looks like thats the case.. I'll retest soon ;-)

jirka

Anything I can do to help here?

Thanks,

Suravee

#10 lore raw

Re: [PATCH 2/4] perf tools: allow user to specify hardware breakpoint bp_len

Frederic,

3.17 windows is now open. Do you think you could get this one in?

Thanks

Suravee

Quoted 37 lines · Suravee Suthikulanit

This message has been archived. View the original item
<http://ausev2.amd.com/EnterpriseVault/ViewMessage.asp?VaultId=1C7248B8A7CDE234D884C352F0CB2021B1110000amdvault.amd.com&SavesetId=201407115122720~201406061630160000~Z~F134EBF86540766B73DA6ADB45060FE1>

Quoted 8 lines · Jiri Olsa
Quoted 7 lines · Namhyung Kim

Hi Jiri,

Quoted 4 lines · Jiri Olsa
Quoted 3 lines · Frederic Weisbecker

From: Jacob Shin <[email protected]>

Currently bp_len is given a default value of 4. Allow user to

override it:

Quoted 11 lines
Quoted 11 lines
Quoted 10 lines
Quoted 6 lines
   $ perf stat -e mem:0x1000/8
                             ^
                             bp_len

If no value is given, it will default to 4 as it did before.

Namhyung,
both perf tols patches from this patchset mess up with hists tests.. I havent found any connection yet.. any idea? ;-)

Quoted 8 lines
Quoted 2 lines

So you already found the problem in the hpp->elide change and that's the reason of the failure, right? :)

I haven't got a chance to test it yet.. but it looks like thats the case.. I'll retest soon ;-)

jirka

Anything I can do to help here?

Thanks,

Suravee

#11 lore raw

PATCH 3/4 perf tools: add hardware breakpoint bp_len test cases

In reply to Frederic Weisbecker

From: Jacob Shin <[email protected]>

Cc Arnaldo Carvalho de Melo <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc xiakaixu <[email protected]>
---
 tools/perf/tests/parse-events.c | 58 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 58 insertions(+)
tools/perf/tests/parse-events.c +58 -0
index deba669..41ef91d 100644
@@ -1114,6 +1114,49 @@ static int test__pinned_group(struct perf_evlist *evlist)
 	return 0;
 }
 
+static int test__checkevent_breakpoint_len(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel = perf_evlist__first(evlist);
+
+	TEST_ASSERT_VAL("wrong number of entries", 1 == evlist->nr_entries);
+	TEST_ASSERT_VAL("wrong type", PERF_TYPE_BREAKPOINT == evsel->attr.type);
+	TEST_ASSERT_VAL("wrong config", 0 == evsel->attr.config);
+	TEST_ASSERT_VAL("wrong bp_type", (HW_BREAKPOINT_R | HW_BREAKPOINT_W) ==
+					 evsel->attr.bp_type);
+	TEST_ASSERT_VAL("wrong bp_len", HW_BREAKPOINT_LEN_1 ==
+					evsel->attr.bp_len);
+
+	return 0;
+}
+
+static int test__checkevent_breakpoint_len_w(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel = perf_evlist__first(evlist);
+
+	TEST_ASSERT_VAL("wrong number of entries", 1 == evlist->nr_entries);
+	TEST_ASSERT_VAL("wrong type", PERF_TYPE_BREAKPOINT == evsel->attr.type);
+	TEST_ASSERT_VAL("wrong config", 0 == evsel->attr.config);
+	TEST_ASSERT_VAL("wrong bp_type", HW_BREAKPOINT_W ==
+					 evsel->attr.bp_type);
+	TEST_ASSERT_VAL("wrong bp_len", HW_BREAKPOINT_LEN_2 ==
+					evsel->attr.bp_len);
+
+	return 0;
+}
+
+static int
+test__checkevent_breakpoint_len_rw_modifier(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel = perf_evlist__first(evlist);
+
+	TEST_ASSERT_VAL("wrong exclude_user", !evsel->attr.exclude_user);
+	TEST_ASSERT_VAL("wrong exclude_kernel", evsel->attr.exclude_kernel);
+	TEST_ASSERT_VAL("wrong exclude_hv", evsel->attr.exclude_hv);
+	TEST_ASSERT_VAL("wrong precise_ip", !evsel->attr.precise_ip);
+
+	return test__checkevent_breakpoint_rw(evlist);
+}
+
 static int count_tracepoints(void)
 {
 	char events_path[PATH_MAX];
@@ -1389,6 +1432,21 @@ static struct evlist_test test__events[] = {
 		.check = test__pinned_group,
 		.id    = 41,
 	},
+	{
+		.name  = "mem:0/1",
+		.check = test__checkevent_breakpoint_len,
+		.id    = 42,
+	},
+	{
+		.name  = "mem:0/2:w",
+		.check = test__checkevent_breakpoint_len_w,
+		.id    = 43,
+	},
+	{
+		.name  = "mem:0/4:rw:u",
+		.check = test__checkevent_breakpoint_len_rw_modifier,
+		.id    = 44
+	},
 #if defined(__s390x__)
 	{
 		.name  = "kvm-s390:kvm_s390_create_vm",
Signature · 1 line
1.8.3.1
#12 lore raw

Re: [PATCH 3/4] perf tools: add hardware breakpoint bp_len test cases

Quoted 12 lines · Frederic Weisbecker
tools/perf/tests/parse-events.c +8 -0
@@ -1389,6 +1432,21 @@ static struct evlist_test test__events[] = {
  		.check = test__pinned_group,
  		.id    = 41,
  	},
+	{
+		.name  = "mem:0/1",
+		.check = test__checkevent_breakpoint_len,
+		.id    = 42,
+	},
+	{
+		.name  = "mem:0/2:w",
+		.check = test__checkevent_breakpoint_len_w,

tools/perf/tests/parse-events.c static struct evlist_test test__events[] = { · line 1452

Quoted 10 lines
tools/perf/tests/parse-events.c +7 -0
+		.id    = 43,
+	},
+	{
+		.name  = "mem:0/4:rw:u",
+		.check = test__checkevent_breakpoint_len_rw_modifier,
+		.id    = 44
+	},
  #if defined(__s390x__)
  	{
  		.name  = "kvm-s390:kvm_s390_create_vm",

Frederic,

This hunk seems to fail to apply on the main tree. Which tree are you working on?

Suravee

#13 lore raw

Re: [PATCH 3/4] perf tools: add hardware breakpoint bp_len test cases

Quoted 25 lines · Suravee Suthikulanit
Quoted 12 lines · Frederic Weisbecker
tools/perf/tests/parse-events.c +8 -0
@@ -1389,6 +1432,21 @@ static struct evlist_test test__events[] = {
 		.check = test__pinned_group,
 		.id    = 41,
 	},
+	{
+		.name  = "mem:0/1",
+		.check = test__checkevent_breakpoint_len,
+		.id    = 42,
+	},
+	{
+		.name  = "mem:0/2:w",
+		.check = test__checkevent_breakpoint_len_w,

tools/perf/tests/parse-events.c static struct evlist_test test__events[] = { · line 1452

Quoted 10 lines
tools/perf/tests/parse-events.c +7 -0
+		.id    = 43,
+	},
+	{
+		.name  = "mem:0/4:rw:u",
+		.check = test__checkevent_breakpoint_len_rw_modifier,
+		.id    = 44
+	},
 #if defined(__s390x__)
 	{
 		.name  = "kvm-s390:kvm_s390_create_vm",

Frederic,

Quoted 2 lines

This hunk seems to fail to apply on the main tree. Which tree are you working on?

I forgot to say it's rebased on top of:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git
perf/core

Quoted 2 lines

Suravee

#14 lore raw

PATCH 4/4 perf/x86: Remove get_hbp_len and replace with bp_len

In reply to Frederic Weisbecker

From: Jacob Shin <[email protected]>

Clean up the logic for determining the breakpoint length

Cc Arnaldo Carvalho de Melo <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc xiakaixu <[email protected]>
---
 arch/x86/kernel/hw_breakpoint.c | 25 +------------------------
 1 file changed, 1 insertion(+), 24 deletions(-)
arch/x86/kernel/hw_breakpoint.c +1 -24
index fa24eb5..2a9000c 100644
@@ -168,29 +168,6 @@ void arch_uninstall_hw_breakpoint(struct perf_event *bp)
 		set_dr_addr_mask(0, i);
 }
 
-static int get_hbp_len(u8 hbp_len)
-{
-	unsigned int len_in_bytes = 0;
-
-	switch (hbp_len) {
-	case X86_BREAKPOINT_LEN_1:
-		len_in_bytes = 1;
-		break;
-	case X86_BREAKPOINT_LEN_2:
-		len_in_bytes = 2;
-		break;
-	case X86_BREAKPOINT_LEN_4:
-		len_in_bytes = 4;
-		break;
-#ifdef CONFIG_X86_64
-	case X86_BREAKPOINT_LEN_8:
-		len_in_bytes = 8;
-		break;
-#endif
-	}
-	return len_in_bytes;
-}
-
 /*
  * Check for virtual address in kernel space.
  */
@@ -201,7 +178,7 @@ int arch_check_bp_in_kernelspace(struct perf_event *bp)
 	struct arch_hw_breakpoint *info = counter_arch_bp(bp);
 
 	va = info->address;
-	len = get_hbp_len(info->len);
+	len = bp->attr.bp_len;
 
 	return (va >= TASK_SIZE) && ((va + len - 1) >= TASK_SIZE);
 }
Signature · 1 line
1.8.3.1
#15 lore raw
In reply to Frederic Weisbecker
Quoted 7 lines · Frederic Weisbecker

Hi,

I've made it wait too much so I'm posting that for review now. This implements range breakpoints on some AMD machine by extending the length field.

Any comment?

hi,
perf tools patches look ok to me.. let me know
if I should queue them

thanks,
jirka

#16 lore raw

I have also tested this patch set on the family16h system, and it seems to be working okay. Let me know if I can help with anything else to get this patch in.

Suravee

Quoted 15 lines · Jiri Olsa
Quoted 7 lines · Frederic Weisbecker

Hi,

I've made it wait too much so I'm posting that for review now. This implements range breakpoints on some AMD machine by extending the length field.

Any comment?

hi,
perf tools patches look ok to me.. let me know
if I should queue them

thanks,
jirka