From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Ingo Molnar <mingo@kernel.org>
Cc: linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
Jin Yao <yao.jin@linux.intel.com>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Andi Kleen <ak@linux.intel.com>, Kan Liang <kan.liang@intel.com>,
Michael Ellerman <mpe@ellerman.id.au>,
Arnaldo Carvalho de Melo <acme@redhat.com>
Subject: [PATCH 81/86] perf/x86/intel: Record branch type
Date: Wed, 19 Jul 2017 11:29:45 -0300 [thread overview]
Message-ID: <20170719142950.3747-82-acme@kernel.org> (raw)
In-Reply-To: <20170719142950.3747-1-acme@kernel.org>
From: Jin Yao <yao.jin@linux.intel.com>
Perf already has support for disassembling the branch instruction
and using the branch type for filtering. The patch just records
the branch type in perf_branch_entry.
Before recording, the patch converts the x86 branch type to
common branch type.
Change log:
v10: Set the branch_map array to be static. The previous version
has it on stack then makes the compiler to create it every
time when the function gets called.
v9: Use __ffs() to find first bit in type in common_branch_type().
It lets the code be clear.
v8: Change PERF_BR_NONE to PERF_BR_UNKNOWN.
v7: Just convert following x86 branch types to common branch types.
X86_BR_CALL -> PERF_BR_CALL
X86_BR_RET -> PERF_BR_RET
X86_BR_JCC -> PERF_BR_COND
X86_BR_JMP -> PERF_BR_UNCOND
X86_BR_IND_CALL -> PERF_BR_IND_CALL
X86_BR_ZERO_CALL -> PERF_BR_CALL
X86_BR_IND_JMP -> PERF_BR_IND
X86_BR_SYSCALL -> PERF_BR_SYSCALL
X86_BR_SYSRET -> PERF_BR_SYSRET
Others are set to PERF_BR_NONE
v6: Not changed.
v5: Just fix the merge error. No other update.
v4: Comparing to previous version, the major changes are:
1. Uses a lookup table to convert x86 branch type to common branch
type.
2. Move the JCC forward/JCC backward and cross page computing to
user space.
3. Initialize branch type to 0 in intel_pmu_lbr_read_32 and
intel_pmu_lbr_read_64
Signed-off-by: Yao Jin <yao.jin@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Link: http://lkml.kernel.org/r/1500379995-6449-3-git-send-email-yao.jin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
arch/x86/events/intel/lbr.c | 52 ++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 51 insertions(+), 1 deletion(-)
diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index eb261656a320..0edda489cf36 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -109,6 +109,9 @@ enum {
X86_BR_ZERO_CALL = 1 << 15,/* zero length call */
X86_BR_CALL_STACK = 1 << 16,/* call stack */
X86_BR_IND_JMP = 1 << 17,/* indirect jump */
+
+ X86_BR_TYPE_SAVE = 1 << 18,/* indicate to save branch type */
+
};
#define X86_BR_PLM (X86_BR_USER | X86_BR_KERNEL)
@@ -510,6 +513,7 @@ static void intel_pmu_lbr_read_32(struct cpu_hw_events *cpuc)
cpuc->lbr_entries[i].in_tx = 0;
cpuc->lbr_entries[i].abort = 0;
cpuc->lbr_entries[i].cycles = 0;
+ cpuc->lbr_entries[i].type = 0;
cpuc->lbr_entries[i].reserved = 0;
}
cpuc->lbr_stack.nr = i;
@@ -596,6 +600,7 @@ static void intel_pmu_lbr_read_64(struct cpu_hw_events *cpuc)
cpuc->lbr_entries[out].in_tx = in_tx;
cpuc->lbr_entries[out].abort = abort;
cpuc->lbr_entries[out].cycles = cycles;
+ cpuc->lbr_entries[out].type = 0;
cpuc->lbr_entries[out].reserved = 0;
out++;
}
@@ -673,6 +678,10 @@ static int intel_pmu_setup_sw_lbr_filter(struct perf_event *event)
if (br_type & PERF_SAMPLE_BRANCH_CALL)
mask |= X86_BR_CALL | X86_BR_ZERO_CALL;
+
+ if (br_type & PERF_SAMPLE_BRANCH_TYPE_SAVE)
+ mask |= X86_BR_TYPE_SAVE;
+
/*
* stash actual user request into reg, it may
* be used by fixup code for some CPU
@@ -926,6 +935,43 @@ static int branch_type(unsigned long from, unsigned long to, int abort)
return ret;
}
+#define X86_BR_TYPE_MAP_MAX 16
+
+static int branch_map[X86_BR_TYPE_MAP_MAX] = {
+ PERF_BR_CALL, /* X86_BR_CALL */
+ PERF_BR_RET, /* X86_BR_RET */
+ PERF_BR_SYSCALL, /* X86_BR_SYSCALL */
+ PERF_BR_SYSRET, /* X86_BR_SYSRET */
+ PERF_BR_UNKNOWN, /* X86_BR_INT */
+ PERF_BR_UNKNOWN, /* X86_BR_IRET */
+ PERF_BR_COND, /* X86_BR_JCC */
+ PERF_BR_UNCOND, /* X86_BR_JMP */
+ PERF_BR_UNKNOWN, /* X86_BR_IRQ */
+ PERF_BR_IND_CALL, /* X86_BR_IND_CALL */
+ PERF_BR_UNKNOWN, /* X86_BR_ABORT */
+ PERF_BR_UNKNOWN, /* X86_BR_IN_TX */
+ PERF_BR_UNKNOWN, /* X86_BR_NO_TX */
+ PERF_BR_CALL, /* X86_BR_ZERO_CALL */
+ PERF_BR_UNKNOWN, /* X86_BR_CALL_STACK */
+ PERF_BR_IND, /* X86_BR_IND_JMP */
+};
+
+static int
+common_branch_type(int type)
+{
+ int i;
+
+ type >>= 2; /* skip X86_BR_USER and X86_BR_KERNEL */
+
+ if (type) {
+ i = __ffs(type);
+ if (i < X86_BR_TYPE_MAP_MAX)
+ return branch_map[i];
+ }
+
+ return PERF_BR_UNKNOWN;
+}
+
/*
* implement actual branch filter based on user demand.
* Hardware may not exactly satisfy that request, thus
@@ -942,7 +988,8 @@ intel_pmu_lbr_filter(struct cpu_hw_events *cpuc)
bool compress = false;
/* if sampling all branches, then nothing to filter */
- if ((br_sel & X86_BR_ALL) == X86_BR_ALL)
+ if (((br_sel & X86_BR_ALL) == X86_BR_ALL) &&
+ ((br_sel & X86_BR_TYPE_SAVE) != X86_BR_TYPE_SAVE))
return;
for (i = 0; i < cpuc->lbr_stack.nr; i++) {
@@ -963,6 +1010,9 @@ intel_pmu_lbr_filter(struct cpu_hw_events *cpuc)
cpuc->lbr_entries[i].from = 0;
compress = true;
}
+
+ if ((br_sel & X86_BR_TYPE_SAVE) == X86_BR_TYPE_SAVE)
+ cpuc->lbr_entries[i].type = common_branch_type(type);
}
if (!compress)
--
2.9.4
next prev parent reply other threads:[~2017-07-19 14:29 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-07-19 14:28 [GIT PULL 00/86] perf/core improvements and fixes Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 01/86] perf annotate: Check for fused instructions Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 02/86] perf annotate: Implement visual marker for macro fusion Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 03/86] perf trace: Remove F_ from some of the fcntl command strings Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 04/86] perf trace: Beautify linux specific fcntl commands Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 05/86] tools: Update include/uapi/linux/fcntl.h copy from the kernel Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 06/86] perf trace beauty: Export the strarrays scnprintf method Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 07/86] perf trace: Only build tools/perf/trace/beauty/ when building 'perf trace' Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 08/86] perf trace beauty: Mask ignored fcntl 'arg' parameter Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 09/86] perf trace beauty: Allow accessing syscall args values in a syscall arg formatter Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 10/86] perf trace beauty: Export the "int" and "hex" syscall arg formatters Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 11/86] perf trace beauty: Introduce syscall arg beautifier for long integers Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 12/86] tools include uapi asm-generic: Grab a copy of fcntl.h Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 13/86] perf trace beauty fcntl: Basic 'arg' beautifier Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 14/86] perf trace: Beautify new write hint fcntl commands Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 15/86] perf beauty open: Detach the syscall_arg agnostic bits from the flags formatter Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 16/86] perf trace: Allow syscall_arg beautifiers to set a different return formatter Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 17/86] perf trace beauty open flags: Support O_TMPFILE and O_NOFOLLOW Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 18/86] perf trace beauty open flags: Do not depend on the system's O_LARGEFILE define Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 19/86] perf trace beauty fcntl: Beautify F_GETFL return value Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 20/86] perf trace beauty open flags: Move RDRW to the start of the output Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 21/86] perf trace beauty fcntl flags: Beautify F_SETFL arg Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 22/86] perf trace beauty fcntl: Beautify F_[GS]ETFD arg/return value Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 23/86] perf trace beauty: Give syscall return beautifier more context Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 24/86] perf trace beauty: Export the fd beautifier for use in more places Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 25/86] perf trace beauty fcntl: Augment the return of F_DUPFD(_CLOEXEC) Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 26/86] perf trace beauty: Export the pid beautifier for use in more places Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 27/86] perf trace beauty fcntl: Beautify F_GETOWN and F_SETOWN Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 28/86] perf pmu-events: Support additional POWER8+ PVR in mapfile Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 29/86] perf vendor events: Add POWER9 PMU events Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 30/86] perf vendor events: Add POWER9 PVRs to mapfile Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 31/86] tools include uapi x86: Grab a copy of unistd.h Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 32/86] tools include uapi x86: Add __NR_setns, if missing Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 33/86] tools build: Add test for setns() Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 34/86] perf symbols: Find symbols in different mount namespace Arnaldo Carvalho de Melo
2017-07-19 14:28 ` [PATCH 35/86] perf maps: Lookup maps in both intitial mountns and inner mountns Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 36/86] perf probe: Allow placing uprobes in alternate namespaces Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 37/86] perf buildid-cache: Support binary objects from other namespaces Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 38/86] perf buildid-cache: Cache debuginfo Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 39/86] perf evsel: Allow asking for max precise_ip in new_cycles() Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 40/86] perf evlist: Allow asking for max precise_ip in add_default() Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 41/86] perf record: Do not ask for precise_ip with --no-samples Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 42/86] perf test sdt: Handle realpath() failure Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 43/86] perf tests attr: Do not store failed events Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 44/86] perf tests attr: Add test_attr__ready function Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 45/86] perf tests attr: Make compare_data global Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 46/86] perf tests attr: Rename compare_data to data_equal Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 47/86] perf tests attr: Add 1s for exclude_kernel and task base bits Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 48/86] perf tests attr: Fix record dwarf test Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 49/86] perf tests attr: Fix no-delay test Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 50/86] perf tests attr: Add proper return values Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 51/86] perf tests attr: Fix cpu test disabled term setup Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 52/86] perf tests attr: Fix sample_period setup Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 53/86] perf tests attr: Fix precise_ip setup Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 54/86] perf tests attr: Fix stat sample_type setup Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 55/86] perf tests attr: Add optional term Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 56/86] perf trace beauty: Export strarray for use in per-object beautifiers Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 57/86] perf trace beauty fcntl: Beautify F_GETLEASE and F_SETLEASE arg/return Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 58/86] perf trace: Group per syscall arg formatter info into one struct Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 59/86] perf trace: Allow syscall arg formatters to request non suppression of zeros Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 60/86] perf trace beauty fcntl: Do not suppress 'cmd' when zero, should be DUPFD Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 61/86] perf trace beauty fcntl: Beautify the 'arg' for DUPFD Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 62/86] perf trace beauty: Simplify syscall return formatting Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 63/86] perf report: Enable finding kernel inline functions Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 64/86] perf header: Encapsulate read and swap Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 65/86] perf header: Add PROCESS_STR_FUN macro Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 66/86] perf header: Fail on write_padded error Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 67/86] perf util: Add const modifier to buf in "writen" function Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 68/86] perf header: Revamp do_write() Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 69/86] perf header: Add struct feat_fd for write Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 70/86] perf header: Use struct feat_fd for print Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 71/86] perf header: Use struct feat_fd to process header records Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 72/86] perf header: Don't pass struct perf_file_section to process_##_feat Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 73/86] perf header: Use struct feat_fd in read header records Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 74/86] perf header: Make write_pmu_mappings pipe-mode friendly Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 75/86] perf header: Add a buffer to struct feat_fd Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 76/86] perf header: Change FEAT_OP* macros Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 77/86] perf tool: Add show_feature_header to perf_tool Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 78/86] perf tools: Add feature header record to pipe-mode Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 79/86] perf header: Add event desc to pipe-mode header Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 80/86] perf/core: Define the common branch type classification Arnaldo Carvalho de Melo
2017-07-19 14:29 ` Arnaldo Carvalho de Melo [this message]
2017-07-19 14:29 ` [PATCH 82/86] perf record: Create a new option save_type in --branch-filter Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 83/86] perf report: Refactor the branch info printing code Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 84/86] perf util: Create branch.c/.h for common branch functions Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 85/86] perf report: Show branch type statistics for stdio mode Arnaldo Carvalho de Melo
2017-07-19 14:29 ` [PATCH 86/86] perf report: Show branch type in callchain entry Arnaldo Carvalho de Melo
2017-07-20 8:32 ` [GIT PULL 00/86] perf/core improvements and fixes Ingo Molnar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170719142950.3747-82-acme@kernel.org \
--to=acme@kernel.org \
--cc=acme@redhat.com \
--cc=ak@linux.intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=kan.liang@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=mpe@ellerman.id.au \
--cc=yao.jin@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).