From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Ingo Molnar <mingo@kernel.org>, Thomas Gleixner <tglx@linutronix.de>
Cc: "Jiri Olsa" <jolsa@kernel.org>,
"Namhyung Kim" <namhyung@kernel.org>,
"Clark Williams" <williams@redhat.com>,
linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
"Arnaldo Carvalho de Melo" <acme@redhat.com>,
"Adrian Hunter" <adrian.hunter@intel.com>,
"Luis Cláudio Gonçalves" <lclaudio@redhat.com>
Subject: [PATCH 09/37] perf trace: Put the per-syscall entry/exit prog_array BPF map infrastructure in place
Date: Mon, 22 Jul 2019 14:38:11 -0300 [thread overview]
Message-ID: <20190722173839.22898-10-acme@kernel.org> (raw)
In-Reply-To: <20190722173839.22898-1-acme@kernel.org>
From: Arnaldo Carvalho de Melo <acme@redhat.com>
I.e. look for "syscalls_sys_enter" and "syscalls_sys_exit" BPF maps of
type PROG_ARRAY and populate it with the handlers as specified per
syscall, for now only 'open' is wiring it to something, in time all
syscalls that need to copy arguments entering a syscall or returning
from one will set these to the right handlers, reusing when possible
pre-existing ones.
Next step is to use bpf_tail_call() into that.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Luis Cláudio Gonçalves <lclaudio@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lkml.kernel.org/n/tip-t0p4u43i9vbpzs1xtowna3gb@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/builtin-trace.c | 76 ++++++++++++++++++-
.../examples/bpf/augmented_raw_syscalls.c | 14 ++++
2 files changed, 86 insertions(+), 4 deletions(-)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 6cc696edf24a..fb8b8e78d7b5 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -1,4 +1,3 @@
-// SPDX-License-Identifier: GPL-2.0-only
/*
* builtin-trace.c
*
@@ -83,6 +82,10 @@ struct trace {
int max;
struct syscall *table;
struct bpf_map *map;
+ struct { // per syscall BPF_MAP_TYPE_PROG_ARRAY
+ struct bpf_map *sys_enter,
+ *sys_exit;
+ } prog_array;
struct {
struct perf_evsel *sys_enter,
*sys_exit,
@@ -1619,6 +1622,22 @@ static int trace__validate_ev_qualifier(struct trace *trace)
goto out;
}
+static __maybe_unused bool trace__syscall_enabled(struct trace *trace, int id)
+{
+ bool in_ev_qualifier;
+
+ if (trace->ev_qualifier_ids.nr == 0)
+ return true;
+
+ in_ev_qualifier = bsearch(&id, trace->ev_qualifier_ids.entries,
+ trace->ev_qualifier_ids.nr, sizeof(int), intcmp) != NULL;
+
+ if (in_ev_qualifier)
+ return !trace->not_ev_qualifier;
+
+ return trace->not_ev_qualifier;
+}
+
/*
* args is to be interpreted as a series of longs but we need to handle
* 8-byte unaligned accesses. args points to raw_data within the event
@@ -2784,6 +2803,18 @@ static void trace__init_syscall_bpf_progs(struct trace *trace, int id)
}
}
+static int trace__bpf_prog_sys_enter_fd(struct trace *trace, int id)
+{
+ struct syscall *sc = trace__syscall_info(trace, NULL, id);
+ return sc ? bpf_program__fd(sc->bpf_prog.sys_enter) : bpf_program__fd(trace->syscalls.unaugmented_prog);
+}
+
+static int trace__bpf_prog_sys_exit_fd(struct trace *trace, int id)
+{
+ struct syscall *sc = trace__syscall_info(trace, NULL, id);
+ return sc ? bpf_program__fd(sc->bpf_prog.sys_exit) : bpf_program__fd(trace->syscalls.unaugmented_prog);
+}
+
static void trace__init_bpf_map_syscall_args(struct trace *trace, int id, struct bpf_map_syscall_entry *entry)
{
struct syscall *sc = trace__syscall_info(trace, NULL, id);
@@ -2837,10 +2868,8 @@ static int __trace__init_syscalls_bpf_map(struct trace *trace, bool enabled)
int err = 0, key;
for (key = 0; key < trace->sctbl->syscalls.nr_entries; ++key) {
- if (enabled) {
+ if (enabled)
trace__init_bpf_map_syscall_args(trace, key, &value);
- trace__init_syscall_bpf_progs(trace, key);
- }
err = bpf_map_update_elem(fd, &key, &value, BPF_ANY);
if (err)
@@ -2859,6 +2888,34 @@ static int trace__init_syscalls_bpf_map(struct trace *trace)
return __trace__init_syscalls_bpf_map(trace, enabled);
}
+
+static int trace__init_syscalls_bpf_prog_array_maps(struct trace *trace)
+{
+ int map_enter_fd = bpf_map__fd(trace->syscalls.prog_array.sys_enter),
+ map_exit_fd = bpf_map__fd(trace->syscalls.prog_array.sys_exit);
+ int err = 0, key;
+
+ for (key = 0; key < trace->sctbl->syscalls.nr_entries; ++key) {
+ int prog_fd;
+
+ if (!trace__syscall_enabled(trace, key))
+ continue;
+
+ trace__init_syscall_bpf_progs(trace, key);
+
+ // It'll get at least the "!raw_syscalls:unaugmented"
+ prog_fd = trace__bpf_prog_sys_enter_fd(trace, key);
+ err = bpf_map_update_elem(map_enter_fd, &key, &prog_fd, BPF_ANY);
+ if (err)
+ break;
+ prog_fd = trace__bpf_prog_sys_exit_fd(trace, key);
+ err = bpf_map_update_elem(map_exit_fd, &key, &prog_fd, BPF_ANY);
+ if (err)
+ break;
+ }
+
+ return err;
+}
#else
static int trace__set_ev_qualifier_bpf_filter(struct trace *trace __maybe_unused)
{
@@ -2875,6 +2932,11 @@ static struct bpf_program *trace__find_bpf_program_by_title(struct trace *trace
{
return NULL;
}
+
+static int trace__init_syscalls_bpf_prog_array_maps(struct trace *trace __maybe_unused)
+{
+ return 0;
+}
#endif // HAVE_LIBBPF_SUPPORT
static int trace__set_ev_qualifier_filter(struct trace *trace)
@@ -3129,6 +3191,10 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
if (trace->syscalls.map)
trace__init_syscalls_bpf_map(trace);
+ if (trace->syscalls.prog_array.sys_enter)
+ trace__init_syscalls_bpf_prog_array_maps(trace);
+
+
if (trace->ev_qualifier_ids.nr > 0) {
err = trace__set_ev_qualifier_filter(trace);
if (err < 0)
@@ -3754,6 +3820,8 @@ static void trace__set_bpf_map_filtered_pids(struct trace *trace)
static void trace__set_bpf_map_syscalls(struct trace *trace)
{
trace->syscalls.map = trace__find_bpf_map_by_name(trace, "syscalls");
+ trace->syscalls.prog_array.sys_enter = trace__find_bpf_map_by_name(trace, "syscalls_sys_enter");
+ trace->syscalls.prog_array.sys_exit = trace__find_bpf_map_by_name(trace, "syscalls_sys_exit");
}
static int trace__config(const char *var, const char *value, void *arg)
diff --git a/tools/perf/examples/bpf/augmented_raw_syscalls.c b/tools/perf/examples/bpf/augmented_raw_syscalls.c
index 66b33b299349..c66474a6ccf4 100644
--- a/tools/perf/examples/bpf/augmented_raw_syscalls.c
+++ b/tools/perf/examples/bpf/augmented_raw_syscalls.c
@@ -33,6 +33,20 @@ struct syscall {
bpf_map(syscalls, ARRAY, int, struct syscall, 512);
+/*
+ * What to augment at entry?
+ *
+ * Pointer arg payloads (filenames, etc) passed from userspace to the kernel
+ */
+bpf_map(syscalls_sys_enter, PROG_ARRAY, u32, u32, 512);
+
+/*
+ * What to augment at exit?
+ *
+ * Pointer arg payloads returned from the kernel (struct stat, etc) to userspace.
+ */
+bpf_map(syscalls_sys_exit, PROG_ARRAY, u32, u32, 512);
+
struct syscall_enter_args {
unsigned long long common_tp_fields;
long syscall_nr;
--
2.21.0
next prev parent reply other threads:[~2019-07-22 17:40 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-22 17:38 [GIT PULL] perf/core improvements and fixes Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 01/37] perf include bpf: Add bpf_tail_call() prototype Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 02/37] perf bpf: Do not attach a BPF prog to a tracepoint if its name starts with ! Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 03/37] perf evsel: Store backpointer to attached bpf_object Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 04/37] perf trace: Add pointer to BPF object containing __augmented_syscalls__ Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 05/37] perf trace: Look up maps just on the __augmented_syscalls__ BPF object Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 06/37] perf trace: Order -e syscalls table Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 07/37] perf trace: Add BPF handler for unaugmented syscalls Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 08/37] perf trace: Allow specifying the bpf prog to augment specific syscalls Arnaldo Carvalho de Melo
2019-07-22 17:38 ` Arnaldo Carvalho de Melo [this message]
2019-07-22 17:38 ` [PATCH 10/37] perf trace: Handle raw_syscalls:sys_enter just like the BPF_OUTPUT augmented event Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 11/37] perf augmented_raw_syscalls: Add handler for "openat" Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 12/37] perf augmented_raw_syscalls: Switch to using BPF_MAP_TYPE_PROG_ARRAY Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 13/37] perf augmented_raw_syscalls: Support copying two string syscall args Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 14/37] perf trace: Look for default name for entries in the syscalls prog array Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 15/37] perf augmented_raw_syscalls: Rename augmented_args_filename to augmented_args_payload Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 16/37] perf script: Fix --max-blocks man page description Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 17/37] perf script: Improve man page description of metrics Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 18/37] perf script: Fix off by one in brstackinsn IPC computation Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 19/37] perf tools: Fix proper buffer size for feature processing Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 20/37] perf stat: Fix segfault for event group in repeat mode Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 21/37] perf augmented_raw_syscalls: Augment sockaddr arg in 'connect' Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 22/37] perf trace beauty: Make connect's addrlen be printed as an int, not hex Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 23/37] perf trace beauty: Disable fd->pathname when close() not enabled Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 24/37] perf trace beauty: Do not try to use the fd->pathname beautifier for bind/connect fd arg Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 25/37] perf trace beauty: Beautify 'sendto's sockaddr arg Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 26/37] perf trace beauty: Beautify bind's " Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 27/37] perf stat: Always separate stalled cycles per insn Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 28/37] perf session: Fix loading of compressed data split across adjacent records Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 29/37] perf trace beauty: Add BPF augmenter for the 'rename' syscall Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 30/37] perf trace: Forward error codes when trying to read syscall info Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 31/37] perf trace: Mark syscall ids that are not allocated to avoid unnecessary error messages Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 32/37] perf trace: Preallocate the syscall table Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 33/37] perf trace: Reuse BPF augmenters from syscalls with similar args signature Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 34/37] perf trace: Add "sendfile64" alias to the "sendfile" syscall Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 35/37] perf probe: Set pev->nargs to zero after freeing pev->args entries Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 36/37] perf probe: Avoid calling freeing routine multiple times for same pointer Arnaldo Carvalho de Melo
2019-07-22 17:38 ` [PATCH 37/37] perf build: Do not use -Wshadow on gcc < 4.8 Arnaldo Carvalho de Melo
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=20190722173839.22898-10-acme@kernel.org \
--to=acme@kernel.org \
--cc=acme@redhat.com \
--cc=adrian.hunter@intel.com \
--cc=jolsa@kernel.org \
--cc=lclaudio@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=namhyung@kernel.org \
--cc=tglx@linutronix.de \
--cc=williams@redhat.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).