rpms/libpfm
6
0
Fork 0
Code Issues Pull Requests Releases Activity

Rebase on libpfm-4.13.0.

Browse Source
This commit is contained in:
William Cohen 2023-03-28 23:31:46 -04:00
parent a5d8146854
commit face84c751
5 changed files with 7 additions and 362 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@ -11,3 +11,4 @@
/libpfm-4.10.0.tar.gz
/libpfm-4.10.1.tar.gz
/libpfm-4.11.0.tar.gz
/libpfm-4.13.0.tar.gz

89
libpfm-gcc12.patch
View File

@ -1,89 +0,0 @@
commit a7b26272d8327ad1c001456a18518a0ac65dc2bb
Author: Stephane Eranian <eranian@gmail.com>
Date: Wed Jun 8 06:55:36 2022 -0700
avoid GCC-12 use-after-free warnings
gcc-12 seems to complain about bogus use-after-free situations in the
libpfm4 code:
p = realloc(q, ...)
if (!p)
return NULL
s = p + (q - z)
It complains because of the use of q after realloc in this case.
Yet q - z is just pointer artihmetic and is not dereferencing any
memory through the pointer q which may have been freed by realloc.
Fix is to pre-computer the delta before realloc to avoid using the
pointer after the call.
Reported-by: Vitaly Chikunov <vt@altlinux.org>
Signed-off-by: Stephane Eranian <eranian@gmail.com>
diff --git a/lib/pfmlib_perf_event_pmu.c b/lib/pfmlib_perf_event_pmu.c
index c3386aa..637c5b1 100644
--- a/lib/pfmlib_perf_event_pmu.c
+++ b/lib/pfmlib_perf_event_pmu.c
@@ -268,6 +268,7 @@ perf_table_alloc_event(void)
perf_table_alloc_event(void)
{
perf_event_t *new_pe;
+ size_t num_free;
retry:
if (perf_pe_free < perf_pe_end)
@@ -286,11 +287,20 @@ retry:
perf_pe_count += PERF_ALLOC_EVENT_COUNT;
+ /*
+ * compute number of free events left
+ * before realloc() to avoid compiler warning (use-after-free)
+ * even though we are simply doing pointer arithmetic and not
+ * dereferencing the perf_pe after realloc when it may be stale
+ * in case the memory was moved.
+ */
+ num_free = perf_pe_free - perf_pe;
+
new_pe = realloc(perf_pe, perf_pe_count * sizeof(perf_event_t));
if (!new_pe)
return NULL;
- perf_pe_free = new_pe + (perf_pe_free - perf_pe);
+ perf_pe_free = new_pe + num_free;
perf_pe_end = perf_pe_free + PERF_ALLOC_EVENT_COUNT;
perf_pe = new_pe;
@@ -315,18 +325,27 @@ static perf_umask_t *
perf_table_alloc_umask(void)
{
perf_umask_t *new_um;
+ size_t num_free;
retry:
if (perf_um_free < perf_um_end)
return perf_um_free++;
perf_um_count += PERF_ALLOC_UMASK_COUNT;
-
+
+ /*
+ * compute number of free unmasks left
+ * before realloc() to avoid compiler warning (use-after-free)
+ * even though we are simply doing pointer arithmetic and not
+ * dereferencing the perf_um after realloc when it may be stale
+ * in case the memory was moved.
+ */
+ num_free = perf_um_free - perf_um;
new_um = realloc(perf_um, perf_um_count * sizeof(*new_um));
if (!new_um)
return NULL;
- perf_um_free = new_um + (perf_um_free - perf_um);
+ perf_um_free = new_um + num_free;
perf_um_end = perf_um_free + PERF_ALLOC_UMASK_COUNT;
perf_um = new_um;

266
libpfm-kernel518.patch
View File

@ -1,266 +0,0 @@
commit 8c606bc2f2d186c2797d9f013283c9150f594f93
Author: Masahiko, Yamada <yamada.masahiko@fujitsu.com>
Date: Tue Sep 20 14:04:31 2022 +0900
update perf_event.h to Linux 5.18
The perf_events interface for directly accessing PMU registers
from userspace for arm64 has been formally implemented in the
kernel v5.18.
Update perf_event.h header used to build perf_event based examples.
Signed-off-by: Masahiko Yamada <yamada.masahiko@fujitsu.com>
diff --git a/include/perfmon/perf_event.h b/include/perfmon/perf_event.h
index 81e02a2..a3bbb14 100644
--- a/include/perfmon/perf_event.h
+++ b/include/perfmon/perf_event.h
@@ -143,7 +143,12 @@ enum perf_event_sample_format {
PERF_SAMPLE_TRANSACTION = 1U << 17,
PERF_SAMPLE_REGS_INTR = 1U << 18,
PERF_SAMPLE_PHYS_ADDR = 1U << 19,
- PERF_SAMPLE_MAX = 1U << 19,
+ PERF_SAMPLE_AUX = 1U << 20,
+ PERF_SAMPLE_CGROUP = 1U << 21,
+ PERF_SAMPLE_DATA_PAGE_SIZE = 1U << 22,
+ PERF_SAMPLE_CODE_PAGE_SIZE = 1U << 23,
+ PERF_SAMPLE_WEIGHT_STRUCT = 1U << 24,
+ PERF_SAMPLE_MAX = 1U << 25,
};
enum {
PERF_TXN_ELISION = (1 << 0),
@@ -180,6 +185,7 @@ enum perf_branch_sample_type_shift {
PERF_SAMPLE_BRANCH_NO_FLAGS_SHIFT = 14,
PERF_SAMPLE_BRANCH_NO_CYCLES_SHIFT = 15,
PERF_SAMPLE_BRANCH_TYPE_SAVE_SHIFT = 16,
+ PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT = 17,
PERF_SAMPLE_BRANCH_MAX_SHIFT /* non-ABI */
};
@@ -204,6 +210,7 @@ enum perf_branch_sample_type {
PERF_SAMPLE_BRANCH_NO_FLAGS = 1U << PERF_SAMPLE_BRANCH_NO_FLAGS_SHIFT,
PERF_SAMPLE_BRANCH_NO_CYCLES = 1U << PERF_SAMPLE_BRANCH_NO_CYCLES_SHIFT,
PERF_SAMPLE_BRANCH_TYPE_SAVE = 1U << PERF_SAMPLE_BRANCH_TYPE_SAVE_SHIFT,
+ PERF_SAMPLE_BRANCH_HW_INDEX = 1U << PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT,
PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT,
};
@@ -232,6 +239,8 @@ enum perf_event_read_format {
/* add: sample_stack_user */
#define PERF_ATTR_SIZE_VER4 104 /* add: sample_regs_intr */
#define PERF_ATTR_SIZE_VER5 112 /* add: aux_watermark */
+#define PERF_ATTR_SIZE_VER6 120 /* add: aux_sample_size */
+#define PERF_ATTR_SIZE_VER7 128 /* add: sig_data */
/*
@@ -289,7 +298,16 @@ typedef struct perf_event_attr {
context_switch : 1,
write_backward : 1,
namespaces : 1,
- __reserved_1 : 35;
+ ksymbol : 1,
+ bpf_event : 1,
+ aux_output : 1,
+ cgroup : 1,
+ text_poke : 1,
+ build_id : 1,
+ inherit_thread : 1,
+ remove_on_exec : 1,
+ sigtrap : 1,
+ __reserved_1 : 26;
union {
uint32_t wakeup_events;
@@ -311,7 +329,11 @@ typedef struct perf_event_attr {
int32_t clockid;
uint64_t sample_regs_intr;
uint32_t aux_watermark;
- uint32_t __reserved_2;
+ uint16_t sample_max_stack;
+ uint16_t __reserved_2;
+ uint32_t aux_sample_size;
+ uint32_t __reserved_3;
+ uint64_t sig_data;
} perf_event_attr_t;
struct perf_branch_entry {
@@ -340,19 +362,32 @@ struct perf_branch_stack {
struct perf_branch_entry entries[0];
};
+/*
+ * Structure used by below PERF_EVENT_IOC_QUERY_BPF command
+ * to query bpf programs attached to the same perf tracepoint
+ * as the given perf event.
+ */
+struct perf_event_query_bpf {
+ uint32_t ids_len;
+ uint32_t prog_cnt;
+ uint32_t ids[0];
+};
+
/*
* perf_events ioctl commands, use with event fd
*/
-#define PERF_EVENT_IOC_ENABLE _IO ('$', 0)
-#define PERF_EVENT_IOC_DISABLE _IO ('$', 1)
-#define PERF_EVENT_IOC_REFRESH _IO ('$', 2)
-#define PERF_EVENT_IOC_RESET _IO ('$', 3)
-#define PERF_EVENT_IOC_PERIOD _IOW('$', 4, uint64_t)
-#define PERF_EVENT_IOC_SET_OUTPUT _IO ('$', 5)
-#define PERF_EVENT_IOC_SET_FILTER _IOW('$', 6, char *)
-#define PERF_EVENT_IOC_ID _IOR('$', 7, uint64_t *)
-#define PERF_EVENT_IOC_SET_BPF _IOW('$', 8, uint32_t)
-#define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW('$', 9, __u32)
+#define PERF_EVENT_IOC_ENABLE _IO ('$', 0)
+#define PERF_EVENT_IOC_DISABLE _IO ('$', 1)
+#define PERF_EVENT_IOC_REFRESH _IO ('$', 2)
+#define PERF_EVENT_IOC_RESET _IO ('$', 3)
+#define PERF_EVENT_IOC_PERIOD _IOW('$', 4, uint64_t)
+#define PERF_EVENT_IOC_SET_OUTPUT _IO ('$', 5)
+#define PERF_EVENT_IOC_SET_FILTER _IOW('$', 6, char *)
+#define PERF_EVENT_IOC_ID _IOR('$', 7, uint64_t *)
+#define PERF_EVENT_IOC_SET_BPF _IOW('$', 8, uint32_t)
+#define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW('$', 9, __u32)
+#define PERF_EVENT_IOC_QUERY_BPF _IOWR('$', 10, struct perf_event_query_bpf *)
+#define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW('$', 11, struct perf_event_attr *)
/*
* ioctl() 3rd argument
@@ -381,7 +416,8 @@ struct perf_event_mmap_page {
cap_usr_rdpmc:1,
cap_user_time:1,
cap_user_time_zero:1,
- cap_____res:59;
+ cap_user_time_short:1,
+ cap_____res:58;
} SWIG_NAME(rdmap_cap_s);
} SWIG_NAME(rdmap_cap_u);
uint16_t pmc_width;
@@ -391,7 +427,10 @@ struct perf_event_mmap_page {
uint64_t time_zero;
uint32_t size;
- uint8_t __reserved[118*8+4];
+ uint32_t __reserved_1;
+ uint64_t time_cycles;
+ uint64_t time_mask;
+ uint8_t __reserved[116*8];
uint64_t data_head;
uint64_t data_tail;
@@ -450,9 +489,38 @@ enum perf_event_type {
PERF_RECORD_SWITCH = 14,
PERF_RECORD_SWITCH_CPU_WIDE = 15,
PERF_RECORD_NAMESPACES = 16,
+ PERF_RECORD_KSYMBOL = 17,
+ PERF_RECORD_BPF_EVENT = 18,
+ PERF_RECORD_CGROUP = 19,
+ PERF_RECORD_TEXT_POKE = 20,
+ PERF_RECORD_AUX_OUTPUT_HW_ID = 21,
PERF_RECORD_MAX
};
+enum perf_record_ksymbol_type {
+ PERF_RECORD_KSYMBOL_TYPE_UNKNOWN = 0,
+ PERF_RECORD_KSYMBOL_TYPE_BPF = 1,
+ /*
+ * Out of line code such as kprobe-replaced instructions or optimized
+ * kprobes or ftrace trampolines.
+ */
+ PERF_RECORD_KSYMBOL_TYPE_OOL = 2,
+ PERF_RECORD_KSYMBOL_TYPE_MAX /* non-ABI */
+};
+
+#define PERF_RECORD_KSYMBOL_FLAGS_UNREGISTER (1 << 0)
+
+enum perf_bpf_event_type {
+ PERF_BPF_EVENT_UNKNOWN = 0,
+ PERF_BPF_EVENT_PROG_LOAD = 1,
+ PERF_BPF_EVENT_PROG_UNLOAD = 2,
+ PERF_BPF_EVENT_MAX, /* non-ABI */
+};
+
+#define PERF_MAX_STACK_DEPTH 127
+#define PERF_MAX_CONTEXTS_PER_STACK 8
+
+
enum perf_callchain_context {
PERF_CONTEXT_HV = (uint64_t)-32,
PERF_CONTEXT_KERNEL = (uint64_t)-128,
@@ -465,8 +533,16 @@ enum perf_callchain_context {
PERF_CONTEXT_MAX = (uint64_t)-4095,
};
-#define PERF_AUX_FLAG_TRUNCATED 0x01
-#define PERF_AUX_FLAG_OVERWRITE 0x02
+#define PERF_AUX_FLAG_TRUNCATED 0x01
+#define PERF_AUX_FLAG_OVERWRITE 0x02
+#define PERF_AUX_FLAG_PARTIAL 0x04
+#define PERF_AUX_FLAG_COLLISION 0x08
+#define PERF_AUX_FLAG_PMU_FORMAT_TYPE_MASK 0xff00
+
+/* CoreSight PMU AUX buffer formats */
+#define PERF_AUX_FLAG_CORESIGHT_FORMAT_CORESIGHT 0x0000
+#define PERF_AUX_FLAG_CORESIGHT_FORMAT_RAW 0x0100
+
/*
* flags for perf_event_open()
@@ -544,12 +620,17 @@ perf_event_open(
union perf_mem_data_src {
uint64_t val;
struct {
- uint64_t mem_op:5, /* type of opcode */
- mem_lvl:14, /* memory hierarchy level */
- mem_snoop:5, /* snoop mode */
- mem_lock:2, /* lock instr */
- mem_dtlb:7, /* tlb access */
- mem_rsvd:31;
+ uint64_t mem_op:5, /* type of opcode */
+ mem_lvl:14, /* memory hierarchy level */
+ mem_snoop:5, /* snoop mode */
+ mem_lock:2, /* lock instr */
+ mem_dtlb:7, /* tlb access */
+ mem_lvl_num:4, /* memory hierarchy level number */
+ mem_remote:1, /* remote */
+ mem_snoopx:2, /* snoop mode, ext */
+ mem_blk:3, /* access blocked */
+ mem_hops:3, /* hop level */
+ mem_rsvd:18;
};
};
@@ -590,7 +671,8 @@ union perf_mem_data_src {
#define PERF_MEM_SNOOP_SHIFT 19
#define PERF_MEM_SNOOPX_FWD 0x01 /* forward */
-#define PERF_MEM_SNOOPX_SHIFT 37
+/* 1 free */
+#define PERF_MEM_SNOOPX_SHIFT 38
/* locked instruction */
#define PERF_MEM_LOCK_NA 0x01 /* not available */
@@ -607,6 +689,20 @@ union perf_mem_data_src {
#define PERF_MEM_TLB_OS 0x40 /* OS fault handler */
#define PERF_MEM_TLB_SHIFT 26
+/* Access blocked */
+#define PERF_MEM_BLK_NA 0x01 /* not available */
+#define PERF_MEM_BLK_DATA 0x02 /* data could not be forwarded */
+#define PERF_MEM_BLK_ADDR 0x04 /* address conflict */
+#define PERF_MEM_BLK_SHIFT 40
+
+/* hop level */
+#define PERF_MEM_HOPS_0 0x01 /* remote core, same node */
+#define PERF_MEM_HOPS_1 0x02 /* remote node, same socket */
+#define PERF_MEM_HOPS_2 0x03 /* remote socket, same board */
+#define PERF_MEM_HOPS_3 0x04 /* remote board */
+/* 5-7 available */
+#define PERF_MEM_HOPS_SHIFT 43
+
#define PERF_MEM_S(a, s) \
(((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)

11
libpfm.spec
View File

@ -11,8 +11,8 @@
%endif
Name: libpfm
Version: 4.11.0
Release: 12%{?dist}
Version: 4.13.0
Release: 1%{?dist}
Summary: Library to encode performance events for use by perf tool
@ -20,8 +20,6 @@ License: MIT
URL: http://perfmon2.sourceforge.net/
Source0: http://sourceforge.net/projects/perfmon2/files/libpfm4/%{name}-%{version}.tar.gz
Patch2: libpfm-python3-setup.patch
Patch3: libpfm-gcc12.patch
Patch4: libpfm-kernel518.patch
BuildRequires: make
BuildRequires: gcc
@ -73,8 +71,6 @@ Python bindings for libpfm4 and perf_event_open system call.
%prep
%setup -q
%patch2 -p1 -b .python3
%patch3 -p1 -b .gcc12
%patch4 -p1 -b .kernel518
%build
%if %{with python}
@ -128,6 +124,9 @@ rm $RPM_BUILD_ROOT%{_libdir}/lib*.a
%endif
%changelog
* Tue Mar 28 2023 William Cohen <wcohen@redhat.com> - 4.13.0-1
- Rebase on libpfm-4.13.0.
* Tue Mar 14 2023 William Cohen <wcohen@redhat.com> - 4.11.0-12
- Add libpfm upstream patch to allow papi-7.0.1 to build.

2
sources
View File

@ -1 +1 @@
SHA512 (libpfm-4.11.0.tar.gz) = 633035b8a7b35973437572095cdc80d422b2a1a61e74e14f106db95fa8e44e4518e591699cc457f828b8f2fb63f60eef6d0c7535c6b4c9a6c3a70d4550b3c3c7
SHA512 (libpfm-4.13.0.tar.gz) = e61b210aa2ce80f0e47603c88eee2e4f2fe30ca2c0e194a5472b6a8de3bf9dc1085e5261bbb9ddbe5b6531c4b391fb34f20d038e5ebd8e6f4c14c2112aee508f