| Message ID | 1444826502-49291-5-git-send-email-wangnan0@huawei.com |
|---|---|
| State | New |
| Headers | show |
Em Wed, Oct 14, 2015 at 12:41:15PM +0000, Wang Nan escreveu: > This patch introduces bpf__{un,}probe() functions to enable callers to > create kprobe points based on section names a BPF program. It parses > the section names in the program and creates corresponding 'struct > perf_probe_event' structures. The parse_perf_probe_command() function is > used to do the main parsing work. The resuling 'struct perf_probe_event' > is stored into program private data for further using. > > By utilizing the new probing API, this patch creates probe points during > event parsing. > > To ensure probe points be removed correctly, register an atexit hook > so even perf quit through exit() bpf__clear() is still called, so probing > points are cleared. Note that bpf_clear() should be registered before > bpf__probe() is called, so failure of bpf__probe() can still trigger > bpf__clear() to remove probe points which are already probed. > > strerror style error reporting scaffold is created by this patch. > bpf__strerror_probe() is the first error reporting function in bpf-loader.c. So, this one, for a non-root user gives me: [acme@felicio linux]$ perf record --event /tmp/foo.o sleep 1 event syntax error: '/tmp/foo.o' \___ Invalid argument (add -v to see detail) Run 'perf list' for a list of valid events Usage: perf record [<options>] [<command>] or: perf record [<options>] -- <command> [<options>] -e, --event <event> event selector. use 'perf list' to list available events [acme@felicio linux]$ -------------------- I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax error", which clearly isn't a syntax error, we need to tell the user that he or she needs special perfmissions for using sys_bpf() :-) As root: [root@felicio ~]# perf record --event /tmp/foo.o sleep event syntax error: '/tmp/foo.o' \___ Invalid argument (add -v to see detail) Run 'perf list' for a list of valid events Usage: perf record [<options>] [<command>] or: perf record [<options>] -- <command> [<options>] -e, --event <event> event selector. use 'perf list' to list available events [root@felicio ~]# ls -la /tmp/foo.o -rw-rw-r--. 1 acme acme 824 Oct 20 12:35 /tmp/foo.o [root@felicio ~]# file /tmp/foo.o /tmp/foo.o: ELF 64-bit LSB relocatable, no machine, version 1 (SYSV), not stripped Humm, its something else, this is an ancient kernel, 4.2.0, probably without eBPF support? Nope, its there: [root@felicio ~]# grep -i sys_bpf /proc/kallsyms ffffffff811829d0 T SyS_bpf ffffffff811829d0 T sys_bpf [root@felicio ~]# Its something else, we need to improve this error reporting: [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 libbpf: loading /tmp/foo.o libbpf: section .strtab, size 60, link 0, flags 0, type=3 libbpf: section .text, size 0, link 0, flags 6, type=1 libbpf: section .data, size 0, link 0, flags 3, type=1 libbpf: section .bss, size 0, link 0, flags 3, type=8 libbpf: section do_fork, size 16, link 0, flags 6, type=1 libbpf: found program do_fork libbpf: section license, size 4, link 0, flags 3, type=1 libbpf: license of /tmp/foo.o is GPL libbpf: section version, size 4, link 0, flags 3, type=1 libbpf: kernel version of /tmp/foo.o is 40100 libbpf: section .symtab, size 96, link 1, flags 0, type=2 bpf: config program 'do_fork' symbol:do_fork file:(null) line:0 offset:0 return:0 lazy:(null) bpf: 'do_fork': event name is missing event syntax error: '/tmp/foo.o' \___ Invalid argument (add -v to see detail) Run 'perf list' for a list of valid events Usage: perf record [<options>] [<command>] or: perf record [<options>] -- <command> [<options>] -e, --event <event> event selector. use 'perf list' to list available events [root@felicio ~]# [root@felicio ~]# grep do_fork /proc/kallsyms ffffffff81099ab0 T _do_fork ffffffff81ccc800 d do_fork_test [root@felicio ~]# $ echo '__attribute__((section("_do_fork"), used)) int fork(void *ctx) {return 0;} char _license[] __attribute__((section("license"), used)) = "GPL";int _version __attribute__((section("version"), used)) = 0x40100;' | clang -D__KERNEL__ $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR -c - -target bpf -O2 -o /tmp/foo.o [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 libbpf: loading /tmp/foo.o libbpf: section .strtab, size 61, link 0, flags 0, type=3 libbpf: section .text, size 0, link 0, flags 6, type=1 libbpf: section .data, size 0, link 0, flags 3, type=1 libbpf: section .bss, size 0, link 0, flags 3, type=8 libbpf: section _do_fork, size 16, link 0, flags 6, type=1 libbpf: found program _do_fork libbpf: section license, size 4, link 0, flags 3, type=1 libbpf: license of /tmp/foo.o is GPL libbpf: section version, size 4, link 0, flags 3, type=1 libbpf: kernel version of /tmp/foo.o is 40100 libbpf: section .symtab, size 96, link 1, flags 0, type=2 bpf: config program '_do_fork' symbol:_do_fork file:(null) line:0 offset:0 return:0 lazy:(null) bpf: '_do_fork': event name is missing event syntax error: '/tmp/foo.o' \___ Invalid argument (add -v to see detail) Run 'perf list' for a list of valid events Usage: perf record [<options>] [<command>] or: perf record [<options>] -- <command> [<options>] -e, --event <event> event selector. use 'perf list' to list available events [root@felicio ~]# So it still doesn't work, doesn't look like it is trying to find a vmlinux, will look at another patch IIRC is in this patchkit allowing us to tell 'perf record' where to find it... But it can as well use kallsyms... - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
On 10/20/15 1:12 PM, Arnaldo Carvalho de Melo wrote: > I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax > error", which clearly isn't a syntax error, we need to tell the user that he or she > needs special perfmissions for using sys_bpf():-) perfmissions? Are you coining new words -- permissions for perf? ;-) -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Em Tue, Oct 20, 2015 at 01:16:04PM -0600, David Ahern escreveu: > On 10/20/15 1:12 PM, Arnaldo Carvalho de Melo wrote: > >I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax > >error", which clearly isn't a syntax error, we need to tell the user that he or she > >needs special perfmissions for using sys_bpf():-) > > perfmissions? Are you coining new words -- permissions for perf? ;-) LOL, totally unintended 8-) - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Em Tue, Oct 20, 2015 at 04:12:14PM -0300, Arnaldo Carvalho de Melo escreveu: > Em Wed, Oct 14, 2015 at 12:41:15PM +0000, Wang Nan escreveu: > > This patch introduces bpf__{un,}probe() functions to enable callers to > > create kprobe points based on section names a BPF program. It parses > > the section names in the program and creates corresponding 'struct > > perf_probe_event' structures. The parse_perf_probe_command() function is > > used to do the main parsing work. The resuling 'struct perf_probe_event' > > is stored into program private data for further using. > > > > By utilizing the new probing API, this patch creates probe points during > > event parsing. > > > > To ensure probe points be removed correctly, register an atexit hook > > so even perf quit through exit() bpf__clear() is still called, so probing > > points are cleared. Note that bpf_clear() should be registered before > > bpf__probe() is called, so failure of bpf__probe() can still trigger > > bpf__clear() to remove probe points which are already probed. > > > > strerror style error reporting scaffold is created by this patch. > > bpf__strerror_probe() is the first error reporting function in bpf-loader.c. > > So, this one, for a non-root user gives me: > > [acme@felicio linux]$ perf record --event /tmp/foo.o sleep 1 > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [acme@felicio linux]$ > > -------------------- > > I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax > error", which clearly isn't a syntax error, we need to tell the user that he or she > needs special perfmissions for using sys_bpf() :-) > > As root: > > [root@felicio ~]# perf record --event /tmp/foo.o sleep > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# ls -la /tmp/foo.o > -rw-rw-r--. 1 acme acme 824 Oct 20 12:35 /tmp/foo.o > [root@felicio ~]# file /tmp/foo.o > /tmp/foo.o: ELF 64-bit LSB relocatable, no machine, version 1 (SYSV), not stripped > > > Humm, its something else, this is an ancient kernel, 4.2.0, probably without > eBPF support? Nope, its there: > > [root@felicio ~]# grep -i sys_bpf /proc/kallsyms > ffffffff811829d0 T SyS_bpf > ffffffff811829d0 T sys_bpf > [root@felicio ~]# > > Its something else, we need to improve this error reporting: > > [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > libbpf: loading /tmp/foo.o > libbpf: section .strtab, size 60, link 0, flags 0, type=3 > libbpf: section .text, size 0, link 0, flags 6, type=1 > libbpf: section .data, size 0, link 0, flags 3, type=1 > libbpf: section .bss, size 0, link 0, flags 3, type=8 > libbpf: section do_fork, size 16, link 0, flags 6, type=1 > libbpf: found program do_fork > libbpf: section license, size 4, link 0, flags 3, type=1 > libbpf: license of /tmp/foo.o is GPL > libbpf: section version, size 4, link 0, flags 3, type=1 > libbpf: kernel version of /tmp/foo.o is 40100 > libbpf: section .symtab, size 96, link 1, flags 0, type=2 > bpf: config program 'do_fork' > symbol:do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > bpf: 'do_fork': event name is missing > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# > > [root@felicio ~]# grep do_fork /proc/kallsyms > ffffffff81099ab0 T _do_fork > ffffffff81ccc800 d do_fork_test > [root@felicio ~]# > > $ echo '__attribute__((section("_do_fork"), used)) int fork(void *ctx) {return 0;} char _license[] __attribute__((section("license"), used)) = "GPL";int _version __attribute__((section("version"), used)) = 0x40100;' | clang -D__KERNEL__ $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR -c - -target bpf -O2 -o /tmp/foo.o > > [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > libbpf: loading /tmp/foo.o > libbpf: section .strtab, size 61, link 0, flags 0, type=3 > libbpf: section .text, size 0, link 0, flags 6, type=1 > libbpf: section .data, size 0, link 0, flags 3, type=1 > libbpf: section .bss, size 0, link 0, flags 3, type=8 > libbpf: section _do_fork, size 16, link 0, flags 6, type=1 > libbpf: found program _do_fork > libbpf: section license, size 4, link 0, flags 3, type=1 > libbpf: license of /tmp/foo.o is GPL > libbpf: section version, size 4, link 0, flags 3, type=1 > libbpf: kernel version of /tmp/foo.o is 40100 > libbpf: section .symtab, size 96, link 1, flags 0, type=2 > bpf: config program '_do_fork' > symbol:_do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > bpf: '_do_fork': event name is missing > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# > > So it still doesn't work, doesn't look like it is trying to find a vmlinux, > will look at another patch IIRC is in this patchkit allowing us to tell > 'perf record' where to find it... But it can as well use kallsyms... That doesn't seem to be the problem, the last check after parse_perf_probe_command() is failing: + pr_debug("bpf: config program '%s'\n", config_str); + err = parse_perf_probe_command(config_str, pev); + if (err < 0) { + pr_debug("bpf: '%s' is not a valid config string\n", + config_str); + err = -EINVAL; + goto errout; + } + + if (pev->group && strcmp(pev->group, PERF_BPF_PROBE_GROUP)) { + pr_debug("bpf: '%s': group for event is set and not '%s'.\n", + config_str, PERF_BPF_PROBE_GROUP); + err = -EINVAL; + goto errout; + } else if (!pev->group) + pev->group = strdup(PERF_BPF_PROBE_GROUP); + + if (!pev->group) { + pr_debug("bpf: strdup failed\n"); + err = -ENOMEM; + goto errout; + } + + if (!pev->event) { + pr_debug("bpf: '%s': event name is missing\n", + config_str); + err = -EINVAL; + goto errout; + } + pr_debug("bpf: config '%s' is ok\n", config_str); I.e. pev->event is NULL, is this expected at this point? I'll continue investigating later. What I have is at the perf/ebpf branch in my tree. - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
On 2015/10/21 3:12, Arnaldo Carvalho de Melo wrote: > Em Wed, Oct 14, 2015 at 12:41:15PM +0000, Wang Nan escreveu: >> This patch introduces bpf__{un,}probe() functions to enable callers to >> create kprobe points based on section names a BPF program. It parses >> the section names in the program and creates corresponding 'struct >> perf_probe_event' structures. The parse_perf_probe_command() function is >> used to do the main parsing work. The resuling 'struct perf_probe_event' >> is stored into program private data for further using. >> >> By utilizing the new probing API, this patch creates probe points during >> event parsing. >> >> To ensure probe points be removed correctly, register an atexit hook >> so even perf quit through exit() bpf__clear() is still called, so probing >> points are cleared. Note that bpf_clear() should be registered before >> bpf__probe() is called, so failure of bpf__probe() can still trigger >> bpf__clear() to remove probe points which are already probed. >> >> strerror style error reporting scaffold is created by this patch. >> bpf__strerror_probe() is the first error reporting function in bpf-loader.c. > So, this one, for a non-root user gives me: > > [acme@felicio linux]$ perf record --event /tmp/foo.o sleep 1 > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [acme@felicio linux]$ It should be: # perf record -e ./test_config_map.o ls Failed to init vmlinux path. event syntax error: './test_config_map.o' \___ You need to be root, and /proc/sys/kernel/kptr_restrict should be 0 (add -v to see detail) Run 'perf list' for a list of valid events usage: perf record [<options>] [<command>] or: perf record [<options>] -- <command> [<options>] -e, --event <event> event selector. use 'perf list' to list available events So this is another problem related to probing. I'll have a look at your lot. Thank you. > -------------------- > > I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax > error", which clearly isn't a syntax error, we need to tell the user that he or she > needs special perfmissions for using sys_bpf() :-) > > As root: > > [root@felicio ~]# perf record --event /tmp/foo.o sleep > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# ls -la /tmp/foo.o > -rw-rw-r--. 1 acme acme 824 Oct 20 12:35 /tmp/foo.o > [root@felicio ~]# file /tmp/foo.o > /tmp/foo.o: ELF 64-bit LSB relocatable, no machine, version 1 (SYSV), not stripped > > > Humm, its something else, this is an ancient kernel, 4.2.0, probably without > eBPF support? Nope, its there: > > [root@felicio ~]# grep -i sys_bpf /proc/kallsyms > ffffffff811829d0 T SyS_bpf > ffffffff811829d0 T sys_bpf > [root@felicio ~]# > > Its something else, we need to improve this error reporting: > > [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > libbpf: loading /tmp/foo.o > libbpf: section .strtab, size 60, link 0, flags 0, type=3 > libbpf: section .text, size 0, link 0, flags 6, type=1 > libbpf: section .data, size 0, link 0, flags 3, type=1 > libbpf: section .bss, size 0, link 0, flags 3, type=8 > libbpf: section do_fork, size 16, link 0, flags 6, type=1 > libbpf: found program do_fork > libbpf: section license, size 4, link 0, flags 3, type=1 > libbpf: license of /tmp/foo.o is GPL > libbpf: section version, size 4, link 0, flags 3, type=1 > libbpf: kernel version of /tmp/foo.o is 40100 > libbpf: section .symtab, size 96, link 1, flags 0, type=2 > bpf: config program 'do_fork' > symbol:do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > bpf: 'do_fork': event name is missing > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# > > [root@felicio ~]# grep do_fork /proc/kallsyms > ffffffff81099ab0 T _do_fork > ffffffff81ccc800 d do_fork_test > [root@felicio ~]# > > $ echo '__attribute__((section("_do_fork"), used)) int fork(void *ctx) {return 0;} char _license[] __attribute__((section("license"), used)) = "GPL";int _version __attribute__((section("version"), used)) = 0x40100;' | clang -D__KERNEL__ $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR -c - -target bpf -O2 -o /tmp/foo.o > > [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > libbpf: loading /tmp/foo.o > libbpf: section .strtab, size 61, link 0, flags 0, type=3 > libbpf: section .text, size 0, link 0, flags 6, type=1 > libbpf: section .data, size 0, link 0, flags 3, type=1 > libbpf: section .bss, size 0, link 0, flags 3, type=8 > libbpf: section _do_fork, size 16, link 0, flags 6, type=1 > libbpf: found program _do_fork > libbpf: section license, size 4, link 0, flags 3, type=1 > libbpf: license of /tmp/foo.o is GPL > libbpf: section version, size 4, link 0, flags 3, type=1 > libbpf: kernel version of /tmp/foo.o is 40100 > libbpf: section .symtab, size 96, link 1, flags 0, type=2 > bpf: config program '_do_fork' > symbol:_do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > bpf: '_do_fork': event name is missing > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# > > So it still doesn't work, doesn't look like it is trying to find a vmlinux, > will look at another patch IIRC is in this patchkit allowing us to tell > 'perf record' where to find it... But it can as well use kallsyms... > > - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
On 2015/10/21 3:12, Arnaldo Carvalho de Melo wrote: > Em Wed, Oct 14, 2015 at 12:41:15PM +0000, Wang Nan escreveu: >> This patch introduces bpf__{un,}probe() functions to enable callers to >> create kprobe points based on section names a BPF program. It parses >> the section names in the program and creates corresponding 'struct >> perf_probe_event' structures. The parse_perf_probe_command() function is >> used to do the main parsing work. The resuling 'struct perf_probe_event' >> is stored into program private data for further using. >> >> By utilizing the new probing API, this patch creates probe points during >> event parsing. >> >> To ensure probe points be removed correctly, register an atexit hook >> so even perf quit through exit() bpf__clear() is still called, so probing >> points are cleared. Note that bpf_clear() should be registered before >> bpf__probe() is called, so failure of bpf__probe() can still trigger >> bpf__clear() to remove probe points which are already probed. >> >> strerror style error reporting scaffold is created by this patch. >> bpf__strerror_probe() is the first error reporting function in bpf-loader.c. > So, this one, for a non-root user gives me: > > [acme@felicio linux]$ perf record --event /tmp/foo.o sleep 1 > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [acme@felicio linux]$ > > -------------------- > > I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax > error", which clearly isn't a syntax error, we need to tell the user that he or she > needs special perfmissions for using sys_bpf() :-) > > As root: > > [root@felicio ~]# perf record --event /tmp/foo.o sleep > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# ls -la /tmp/foo.o > -rw-rw-r--. 1 acme acme 824 Oct 20 12:35 /tmp/foo.o > [root@felicio ~]# file /tmp/foo.o > /tmp/foo.o: ELF 64-bit LSB relocatable, no machine, version 1 (SYSV), not stripped > > > Humm, its something else, this is an ancient kernel, 4.2.0, probably without > eBPF support? Nope, its there: > > [root@felicio ~]# grep -i sys_bpf /proc/kallsyms > ffffffff811829d0 T SyS_bpf > ffffffff811829d0 T sys_bpf > [root@felicio ~]# > > Its something else, we need to improve this error reporting: > > [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > libbpf: loading /tmp/foo.o > libbpf: section .strtab, size 60, link 0, flags 0, type=3 > libbpf: section .text, size 0, link 0, flags 6, type=1 > libbpf: section .data, size 0, link 0, flags 3, type=1 > libbpf: section .bss, size 0, link 0, flags 3, type=8 > libbpf: section do_fork, size 16, link 0, flags 6, type=1 > libbpf: found program do_fork > libbpf: section license, size 4, link 0, flags 3, type=1 > libbpf: license of /tmp/foo.o is GPL > libbpf: section version, size 4, link 0, flags 3, type=1 > libbpf: kernel version of /tmp/foo.o is 40100 > libbpf: section .symtab, size 96, link 1, flags 0, type=2 > bpf: config program 'do_fork' > symbol:do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > bpf: 'do_fork': event name is missing BPF report the problem, but it is a little bit hard to understand... > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# > > [root@felicio ~]# grep do_fork /proc/kallsyms > ffffffff81099ab0 T _do_fork > ffffffff81ccc800 d do_fork_test > [root@felicio ~]# > > $ echo '__attribute__((section("_do_fork"), used)) int fork(void *ctx) {return 0;} char _license[] __attribute__((section("license"), used)) = "GPL";int _version __attribute__((section("version"), used)) = 0x40100;' | clang -D__KERNEL__ $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR -c - -target bpf -O2 -o /tmp/foo.o In your program you only provide "do_fork", but we need "key=value" syntax. "key" will become the name of created kprobe. Please try "__attribute__((section("func=do_fork"), used)) " instead. I think when event name is missing we'd better construct one name for it like perf probe, but then we need to deal with perf probe code again. It may require another patch. For this patch, I think we can assign a new errorno so bpf__strerror_probe() can give more information to let user know whether the problem is reside in bpf program or perf configuration. Do you think ENOEXEC is a good choice? Thank you. > > [root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > libbpf: loading /tmp/foo.o > libbpf: section .strtab, size 61, link 0, flags 0, type=3 > libbpf: section .text, size 0, link 0, flags 6, type=1 > libbpf: section .data, size 0, link 0, flags 3, type=1 > libbpf: section .bss, size 0, link 0, flags 3, type=8 > libbpf: section _do_fork, size 16, link 0, flags 6, type=1 > libbpf: found program _do_fork > libbpf: section license, size 4, link 0, flags 3, type=1 > libbpf: license of /tmp/foo.o is GPL > libbpf: section version, size 4, link 0, flags 3, type=1 > libbpf: kernel version of /tmp/foo.o is 40100 > libbpf: section .symtab, size 96, link 1, flags 0, type=2 > bpf: config program '_do_fork' > symbol:_do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > bpf: '_do_fork': event name is missing > event syntax error: '/tmp/foo.o' > \___ Invalid argument > > (add -v to see detail) > Run 'perf list' for a list of valid events > > Usage: perf record [<options>] [<command>] > or: perf record [<options>] -- <command> [<options>] > > -e, --event <event> event selector. use 'perf list' to list available events > [root@felicio ~]# > > So it still doesn't work, doesn't look like it is trying to find a vmlinux, > will look at another patch IIRC is in this patchkit allowing us to tell > 'perf record' where to find it... But it can as well use kallsyms... > > - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Em Wed, Oct 21, 2015 at 11:31:57AM +0800, Wangnan (F) escreveu: > > > On 2015/10/21 3:12, Arnaldo Carvalho de Melo wrote: > >Em Wed, Oct 14, 2015 at 12:41:15PM +0000, Wang Nan escreveu: > >>This patch introduces bpf__{un,}probe() functions to enable callers to > >>create kprobe points based on section names a BPF program. It parses > >>the section names in the program and creates corresponding 'struct > >>perf_probe_event' structures. The parse_perf_probe_command() function is > >>used to do the main parsing work. The resuling 'struct perf_probe_event' > >>is stored into program private data for further using. > >> > >>By utilizing the new probing API, this patch creates probe points during > >>event parsing. > >> > >>To ensure probe points be removed correctly, register an atexit hook > >>so even perf quit through exit() bpf__clear() is still called, so probing > >>points are cleared. Note that bpf_clear() should be registered before > >>bpf__probe() is called, so failure of bpf__probe() can still trigger > >>bpf__clear() to remove probe points which are already probed. > >> > >>strerror style error reporting scaffold is created by this patch. > >>bpf__strerror_probe() is the first error reporting function in bpf-loader.c. > >So, this one, for a non-root user gives me: > > > >[acme@felicio linux]$ perf record --event /tmp/foo.o sleep 1 > >event syntax error: '/tmp/foo.o' > > \___ Invalid argument > > > >(add -v to see detail) > >Run 'perf list' for a list of valid events > > > > Usage: perf record [<options>] [<command>] > > or: perf record [<options>] -- <command> [<options>] > > > > -e, --event <event> event selector. use 'perf list' to list available events > >[acme@felicio linux]$ > > > >-------------------- > > > >I.e. no libbpf error (good!) but then, just an -EINVAL as the "event syntax > >error", which clearly isn't a syntax error, we need to tell the user that he or she > >needs special perfmissions for using sys_bpf() :-) > > > >As root: > > > >[root@felicio ~]# perf record --event /tmp/foo.o sleep > >event syntax error: '/tmp/foo.o' > > \___ Invalid argument > > > >(add -v to see detail) > >Run 'perf list' for a list of valid events > > > > Usage: perf record [<options>] [<command>] > > or: perf record [<options>] -- <command> [<options>] > > > > -e, --event <event> event selector. use 'perf list' to list available events > >[root@felicio ~]# ls -la /tmp/foo.o > >-rw-rw-r--. 1 acme acme 824 Oct 20 12:35 /tmp/foo.o > >[root@felicio ~]# file /tmp/foo.o > >/tmp/foo.o: ELF 64-bit LSB relocatable, no machine, version 1 (SYSV), not stripped > > > > > >Humm, its something else, this is an ancient kernel, 4.2.0, probably without > >eBPF support? Nope, its there: > > > >[root@felicio ~]# grep -i sys_bpf /proc/kallsyms > >ffffffff811829d0 T SyS_bpf > >ffffffff811829d0 T sys_bpf > >[root@felicio ~]# > > > >Its something else, we need to improve this error reporting: > > > >[root@felicio ~]# perf record -v --event /tmp/foo.o sleep 1 > >libbpf: loading /tmp/foo.o > >libbpf: section .strtab, size 60, link 0, flags 0, type=3 > >libbpf: section .text, size 0, link 0, flags 6, type=1 > >libbpf: section .data, size 0, link 0, flags 3, type=1 > >libbpf: section .bss, size 0, link 0, flags 3, type=8 > >libbpf: section do_fork, size 16, link 0, flags 6, type=1 > >libbpf: found program do_fork > >libbpf: section license, size 4, link 0, flags 3, type=1 > >libbpf: license of /tmp/foo.o is GPL > >libbpf: section version, size 4, link 0, flags 3, type=1 > >libbpf: kernel version of /tmp/foo.o is 40100 > >libbpf: section .symtab, size 96, link 1, flags 0, type=2 > >bpf: config program 'do_fork' > >symbol:do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > >bpf: 'do_fork': event name is missing > > BPF report the problem, but it is a little bit hard to understand... > > >event syntax error: '/tmp/foo.o' > > \___ Invalid argument > > > >(add -v to see detail) > >Run 'perf list' for a list of valid events > > > > Usage: perf record [<options>] [<command>] > > or: perf record [<options>] -- <command> [<options>] > > > > -e, --event <event> event selector. use 'perf list' to list available events > >[root@felicio ~]# > > > >[root@felicio ~]# grep do_fork /proc/kallsyms > >ffffffff81099ab0 T _do_fork > >ffffffff81ccc800 d do_fork_test > >[root@felicio ~]# > > > >$ echo '__attribute__((section("_do_fork"), used)) int fork(void *ctx) {return 0;} char _license[] __attribute__((section("license"), used)) = "GPL";int _version __attribute__((section("version"), used)) = 0x40100;' | clang -D__KERNEL__ $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR -c - -target bpf -O2 -o /tmp/foo.o > In your program you only provide "do_fork", but we need "key=value" > syntax. "key" will become the name of created kprobe. Please try > "__attribute__((section("func=do_fork"), used)) " instead. > I think when event name is missing we'd better construct one name for > it like perf probe, but then we need to deal with perf probe code > again. It may require another patch. Nah, lets go with what we have, i.e. I'll take that into account and test with the expected form. > For this patch, I think we can assign a new errorno so > bpf__strerror_probe() can give more information to let user know > whether the problem is reside in bpf program or perf configuration. Do > you think ENOEXEC is a good choice? Unsure if we should use existing errno codes in cases like this, probably better to use a BPF_ERRNO__ENOALIAS or somesuch. - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Em Wed, Oct 21, 2015 at 10:28:03AM -0300, Arnaldo Carvalho de Melo escreveu: > Em Wed, Oct 21, 2015 at 11:31:57AM +0800, Wangnan (F) escreveu: > > >bpf: config program 'do_fork' > > >symbol:do_fork file:(null) line:0 offset:0 return:0 lazy:(null) > > >bpf: 'do_fork': event name is missing > > > > BPF report the problem, but it is a little bit hard to understand... > > > > >event syntax error: '/tmp/foo.o' > > > \___ Invalid argument > > > > > >(add -v to see detail) > > >Run 'perf list' for a list of valid events > > > > > > Usage: perf record [<options>] [<command>] > > > or: perf record [<options>] -- <command> [<options>] > > > > > > -e, --event <event> event selector. use 'perf list' to list available events > > >[root@felicio ~]# > > > > > >[root@felicio ~]# grep do_fork /proc/kallsyms > > >ffffffff81099ab0 T _do_fork > > >ffffffff81ccc800 d do_fork_test > > >[root@felicio ~]# > > > > > >$ echo '__attribute__((section("_do_fork"), used)) int fork(void *ctx) {return 0;} char _license[] __attribute__((section("license"), used)) = "GPL";int _version __attribute__((section("version"), used)) = 0x40100;' | clang -D__KERNEL__ $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR -c - -target bpf -O2 -o /tmp/foo.o > > > In your program you only provide "do_fork", but we need "key=value" > > syntax. "key" will become the name of created kprobe. Please try > > "__attribute__((section("func=do_fork"), used)) " instead. > > > I think when event name is missing we'd better construct one name for > > it like perf probe, but then we need to deal with perf probe code > > again. It may require another patch. > > Nah, lets go with what we have, i.e. I'll take that into account and > test with the expected form. So, pushed this cset with my committer testing notes: https://git.kernel.org/cgit/linux/kernel/git/acme/linux.git/commit/?h=perf/ebpf&id=bd3c477e57233819580f5824dea3a6922a860a24 continuing... - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
diff --git a/tools/perf/util/bpf-loader.c b/tools/perf/util/bpf-loader.c index ab56073..56f6fe8 100644 --- a/tools/perf/util/bpf-loader.c +++ b/tools/perf/util/bpf-loader.c @@ -10,6 +10,8 @@ #include "perf.h" #include "debug.h" #include "bpf-loader.h" +#include "probe-event.h" +#include "probe-finder.h" // for MAX_PROBES #define DEFINE_PRINT_FN(name, level) \ static int libbpf_##name(const char *fmt, ...) \ @@ -27,6 +29,10 @@ DEFINE_PRINT_FN(warning, 0) DEFINE_PRINT_FN(info, 0) DEFINE_PRINT_FN(debug, 1) +struct bpf_prog_priv { + struct perf_probe_event pev; +}; + struct bpf_object *bpf__prepare_load(const char *filename) { struct bpf_object *obj; @@ -52,6 +58,220 @@ void bpf__clear(void) { struct bpf_object *obj, *tmp; - bpf_object__for_each_safe(obj, tmp) + bpf_object__for_each_safe(obj, tmp) { + bpf__unprobe(obj); bpf_object__close(obj); + } +} + +static void +bpf_prog_priv__clear(struct bpf_program *prog __maybe_unused, + void *_priv) +{ + struct bpf_prog_priv *priv = _priv; + + cleanup_perf_probe_events(&priv->pev, 1); + free(priv); +} + +static int +config_bpf_program(struct bpf_program *prog) +{ + struct perf_probe_event *pev = NULL; + struct bpf_prog_priv *priv = NULL; + const char *config_str; + int err; + + config_str = bpf_program__title(prog, false); + if (!config_str) { + pr_debug("bpf: unable to get title for program\n"); + return -EINVAL; + } + + priv = calloc(sizeof(*priv), 1); + if (!priv) { + pr_debug("bpf: failed to alloc priv\n"); + return -ENOMEM; + } + pev = &priv->pev; + + pr_debug("bpf: config program '%s'\n", config_str); + err = parse_perf_probe_command(config_str, pev); + if (err < 0) { + pr_debug("bpf: '%s' is not a valid config string\n", + config_str); + err = -EINVAL; + goto errout; + } + + if (pev->group && strcmp(pev->group, PERF_BPF_PROBE_GROUP)) { + pr_debug("bpf: '%s': group for event is set and not '%s'.\n", + config_str, PERF_BPF_PROBE_GROUP); + err = -EINVAL; + goto errout; + } else if (!pev->group) + pev->group = strdup(PERF_BPF_PROBE_GROUP); + + if (!pev->group) { + pr_debug("bpf: strdup failed\n"); + err = -ENOMEM; + goto errout; + } + + if (!pev->event) { + pr_debug("bpf: '%s': event name is missing\n", + config_str); + err = -EINVAL; + goto errout; + } + pr_debug("bpf: config '%s' is ok\n", config_str); + + err = bpf_program__set_private(prog, priv, bpf_prog_priv__clear); + if (err) { + pr_debug("Failed to set priv for program '%s'\n", config_str); + goto errout; + } + + return 0; + +errout: + if (pev) + clear_perf_probe_event(pev); + free(priv); + return err; +} + +static int bpf__prepare_probe(void) +{ + static int err = 0; + static bool initialized = false; + + /* + * Make err static, so if init failed the first, bpf__prepare_probe() + * fails each time without calling init_probe_symbol_maps multiple + * times. + */ + if (initialized) + return err; + + initialized = true; + err = init_probe_symbol_maps(false); + if (err < 0) + pr_debug("Failed to init_probe_symbol_maps\n"); + probe_conf.max_probes = MAX_PROBES; + return err; +} + +int bpf__probe(struct bpf_object *obj) +{ + int err = 0; + struct bpf_program *prog; + struct bpf_prog_priv *priv; + struct perf_probe_event *pev; + + err = bpf__prepare_probe(); + if (err) { + pr_debug("bpf__prepare_probe failed\n"); + return err; + } + + bpf_object__for_each_program(prog, obj) { + err = config_bpf_program(prog); + if (err) + goto out; + + err = bpf_program__get_private(prog, (void **)&priv); + if (err || !priv) + goto out; + pev = &priv->pev; + + err = convert_perf_probe_events(pev, 1); + if (err < 0) { + pr_debug("bpf_probe: failed to convert perf probe events"); + goto out; + } + + err = apply_perf_probe_events(pev, 1); + if (err < 0) { + pr_debug("bpf_probe: failed to apply perf probe events"); + goto out; + } + } +out: + return err < 0 ? err : 0; +} + +#define EVENTS_WRITE_BUFSIZE 4096 +int bpf__unprobe(struct bpf_object *obj) +{ + int err, ret = 0; + struct bpf_program *prog; + struct bpf_prog_priv *priv; + + bpf_object__for_each_program(prog, obj) { + int i; + + err = bpf_program__get_private(prog, (void **)&priv); + if (err || !priv) + continue; + + for (i = 0; i < priv->pev.ntevs; i++) { + struct probe_trace_event *tev = &priv->pev.tevs[i]; + char name_buf[EVENTS_WRITE_BUFSIZE]; + struct strfilter *delfilter; + + snprintf(name_buf, EVENTS_WRITE_BUFSIZE, + "%s:%s", tev->group, tev->event); + name_buf[EVENTS_WRITE_BUFSIZE - 1] = '\0'; + + delfilter = strfilter__new(name_buf, NULL); + if (!delfilter) { + pr_debug("Failed to create filter for unprobing\n"); + ret = -ENOMEM; + continue; + } + + err = del_perf_probe_events(delfilter); + strfilter__delete(delfilter); + if (err) { + pr_debug("Failed to delete %s\n", name_buf); + ret = err; + continue; + } + } + } + return ret; +} + +#define bpf__strerror_head(err, buf, size) \ + char sbuf[STRERR_BUFSIZE], *emsg;\ + if (!size)\ + return 0;\ + if (err < 0)\ + err = -err;\ + emsg = strerror_r(err, sbuf, sizeof(sbuf));\ + switch (err) {\ + default:\ + scnprintf(buf, size, "%s", emsg);\ + break; + +#define bpf__strerror_entry(val, fmt...)\ + case val: {\ + scnprintf(buf, size, fmt);\ + break;\ + } + +#define bpf__strerror_end(buf, size)\ + }\ + buf[size - 1] = '\0'; + +int bpf__strerror_probe(struct bpf_object *obj __maybe_unused, + int err, char *buf, size_t size) +{ + bpf__strerror_head(err, buf, size); + bpf__strerror_entry(EEXIST, "Probe point exist. Try use 'perf probe -d \"*\"'"); + bpf__strerror_entry(EPERM, "You need to be root, and /proc/sys/kernel/kptr_restrict should be 0\n"); + bpf__strerror_entry(ENOENT, "You need to check probing points in BPF file\n"); + bpf__strerror_end(buf, size); + return 0; } diff --git a/tools/perf/util/bpf-loader.h b/tools/perf/util/bpf-loader.h index f402d7c..b819622 100644 --- a/tools/perf/util/bpf-loader.h +++ b/tools/perf/util/bpf-loader.h @@ -11,11 +11,18 @@ #include "debug.h" struct bpf_object; +#define PERF_BPF_PROBE_GROUP "perf_bpf_probe" #ifdef HAVE_LIBBPF_SUPPORT struct bpf_object *bpf__prepare_load(const char *filename); void bpf__clear(void); + +int bpf__probe(struct bpf_object *obj); +int bpf__unprobe(struct bpf_object *obj); +int bpf__strerror_probe(struct bpf_object *obj, int err, + char *buf, size_t size); + #else static inline struct bpf_object * bpf__prepare_load(const char *filename __maybe_unused) @@ -25,5 +32,28 @@ bpf__prepare_load(const char *filename __maybe_unused) } static inline void bpf__clear(void) { } + +static inline int bpf__probe(struct bpf_object *obj __maybe_unused) { return 0;} +static inline int bpf__unprobe(struct bpf_object *obj __maybe_unused) { return 0;} + +static inline int +__bpf_strerror(char *buf, size_t size) +{ + if (!size) + return 0; + strncpy(buf, + "ERROR: eBPF object loading is disabled during compiling.\n", + size); + buf[size - 1] = '\0'; + return 0; +} + +static inline int +bpf__strerror_probe(struct bpf_object *obj __maybe_unused, + int err __maybe_unused, + char *buf, size_t size) +{ + return __bpf_strerror(buf, size); +} #endif #endif diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index a02abd3..06ff8d6 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -536,6 +536,7 @@ int parse_events_load_bpf_obj(struct parse_events_evlist *data, { int err; char errbuf[BUFSIZ]; + static bool registered_unprobe_atexit = false; if (IS_ERR(obj) || !obj) { snprintf(errbuf, sizeof(errbuf), @@ -545,6 +546,22 @@ int parse_events_load_bpf_obj(struct parse_events_evlist *data, } /* + * Register atexit handler before calling bpf__probe() so + * bpf__probe() don't need to unprobe probe points its already + * created when failure. + */ + if (!registered_unprobe_atexit) { + atexit(bpf__clear); + registered_unprobe_atexit = true; + } + + err = bpf__probe(obj); + if (err) { + bpf__strerror_probe(obj, err, errbuf, sizeof(errbuf)); + goto errout; + } + + /* * Temporary add a dummy event here so we can check whether * basic bpf loader works. Following patches will replace * dummy event by useful evsels.
This patch introduces bpf__{un,}probe() functions to enable callers to create kprobe points based on section names a BPF program. It parses the section names in the program and creates corresponding 'struct perf_probe_event' structures. The parse_perf_probe_command() function is used to do the main parsing work. The resuling 'struct perf_probe_event' is stored into program private data for further using. By utilizing the new probing API, this patch creates probe points during event parsing. To ensure probe points be removed correctly, register an atexit hook so even perf quit through exit() bpf__clear() is still called, so probing points are cleared. Note that bpf_clear() should be registered before bpf__probe() is called, so failure of bpf__probe() can still trigger bpf__clear() to remove probe points which are already probed. strerror style error reporting scaffold is created by this patch. bpf__strerror_probe() is the first error reporting function in bpf-loader.c. Signed-off-by: Wang Nan <wangnan0@huawei.com> Cc: Alexei Starovoitov <ast@plumgrid.com> Cc: Brendan Gregg <brendan.d.gregg@gmail.com> Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: David Ahern <dsahern@gmail.com> Cc: He Kuang <hekuang@huawei.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kaixu Xia <xiakaixu@huawei.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Zefan Li <lizefan@huawei.com> Cc: pi3orama@163.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Link: http://lkml.kernel.org/n/ebpf-6yw9eg0ej3l4jnqhinngkw86@git.kernel.org --- tools/perf/util/bpf-loader.c | 222 ++++++++++++++++++++++++++++++++++++++++- tools/perf/util/bpf-loader.h | 30 ++++++ tools/perf/util/parse-events.c | 17 ++++ 3 files changed, 268 insertions(+), 1 deletion(-)