Re: [PATCH v5 2/3] perf: add helper map__fprintf_dsoname_dsoff

From: Changbin Du
Date: Sun Apr 23 2023 - 00:33:10 EST

Next message: Xu,Rongbo: "Re: [PATCH] [PATCH] fuse: no interrupt for lookup request"
Previous message: Zhou Furong: "Re: [PATCH v7 7/7] drm/i915: track gt pm wakerefs"
In reply to: Adrian Hunter: "Re: [PATCH v5 2/3] perf: add helper map__fprintf_dsoname_dsoff"
Next in thread: Changbin Du: "[PATCH v5 3/3] perf: script: add new output field 'dsoff' to print dso offset"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Fri, Apr 21, 2023 at 10:30:45AM +0300, Adrian Hunter wrote:
> On 21/04/23 08:04, Changbin Du wrote:
> >> What do you get if you try below diff on top of
> >> your patches:
> >>
> >> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> >> index c7bf1ac14e90..df0d21141185 100644
> >> --- a/tools/perf/builtin-script.c
> >> +++ b/tools/perf/builtin-script.c
> >> @@ -576,8 +576,11 @@ static void set_print_ip_opts(struct perf_event_attr *attr)
> >> if (PRINT_FIELD(DSO))
> >> output[type].print_ip_opts |= EVSEL__PRINT_DSO;
> >>
> >> - if (PRINT_FIELD(DSOFF))
> >> + if (PRINT_FIELD(DSOFF)) {
> >> output[type].print_ip_opts |= EVSEL__PRINT_DSOFF;
> >> + /* DSO offset is relative to dso->longname */
> >> + symbol_conf.show_kernel_path = true;
> >> + }
> >>
> >> if (PRINT_FIELD(SYMOFFSET))
> >> output[type].print_ip_opts |= EVSEL__PRINT_SYMOFFSET;
> >> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> >> index a86614599269..19ebfd3468cc 100644
> >> --- a/tools/perf/util/dso.c
> >> +++ b/tools/perf/util/dso.c
> >> @@ -67,6 +67,42 @@ char dso__symtab_origin(const struct dso *dso)
> >> return origin[dso->symtab_type];
> >> }
> >>
> >> +bool dso__is_file(const struct dso *dso)
> >> +{
> >> + switch (dso->binary_type) {
> >> + case DSO_BINARY_TYPE__KALLSYMS:
> >> + case DSO_BINARY_TYPE__GUEST_KALLSYMS:
> >> + return false;
> >> + case DSO_BINARY_TYPE__VMLINUX:
> >> + case DSO_BINARY_TYPE__GUEST_VMLINUX:
> >> + return true;
> >> + case DSO_BINARY_TYPE__JAVA_JIT:
> >> + return false;
> >> + case DSO_BINARY_TYPE__DEBUGLINK:
> >> + case DSO_BINARY_TYPE__BUILD_ID_CACHE:
> >> + case DSO_BINARY_TYPE__BUILD_ID_CACHE_DEBUGINFO:
> >> + case DSO_BINARY_TYPE__FEDORA_DEBUGINFO:
> >> + case DSO_BINARY_TYPE__UBUNTU_DEBUGINFO:
> >> + case DSO_BINARY_TYPE__MIXEDUP_UBUNTU_DEBUGINFO:
> >> + case DSO_BINARY_TYPE__BUILDID_DEBUGINFO:
> >> + case DSO_BINARY_TYPE__SYSTEM_PATH_DSO:
> >> + case DSO_BINARY_TYPE__GUEST_KMODULE:
> >> + case DSO_BINARY_TYPE__GUEST_KMODULE_COMP:
> >> + case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE:
> >> + case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP:
> >> + case DSO_BINARY_TYPE__KCORE:
> >> + case DSO_BINARY_TYPE__GUEST_KCORE:
> >> + case DSO_BINARY_TYPE__OPENEMBEDDED_DEBUGINFO:
> >> + return true;
> >> + case DSO_BINARY_TYPE__BPF_PROG_INFO:
> >> + case DSO_BINARY_TYPE__BPF_IMAGE:
> >> + case DSO_BINARY_TYPE__OOL:
> >> + case DSO_BINARY_TYPE__NOT_FOUND:
> >> + default:
> >> + return false;
> >> + }
> >> +}
> >> +
> >> int dso__read_binary_type_filename(const struct dso *dso,
> >> enum dso_binary_type type,
> >> char *root_dir, char *filename, size_t size)
> >> diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
> >> index 0b7c7633b9f6..fb33f5224fb6 100644
> >> --- a/tools/perf/util/dso.h
> >> +++ b/tools/perf/util/dso.h
> >> @@ -396,6 +396,8 @@ static inline bool dso__is_kallsyms(struct dso *dso)
> >> return dso->kernel && dso->long_name[0] != '/';
> >> }
> >>
> >> +bool dso__is_file(const struct dso *dso);
> >> +
> >> void dso__free_a2l(struct dso *dso);
> >>
> >> enum dso_type dso__type(struct dso *dso, struct machine *machine);
> >> diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
> >> index 7da96b41100f..9b79f88d371c 100644
> >> --- a/tools/perf/util/map.c
> >> +++ b/tools/perf/util/map.c
> >> @@ -447,11 +447,12 @@ size_t map__fprintf_dsoname(struct map *map, FILE *fp)
> >>
> >> size_t map__fprintf_dsoname_dsoff(struct map *map, bool print_off, u64 addr, FILE *fp)
> >> {
> >> + const struct dso *dso = map ? map__dso(map) : NULL;
> >> int printed = 0;
> >>
> >> printed += fprintf(fp, " (");
> >> printed += map__fprintf_dsoname(map, fp);
> >> - if (print_off && map && map__dso(map) && !map__dso(map)->kernel)
> >> + if (print_off && dso && dso__is_file(dso))
> >> printed += fprintf(fp, "+0x%" PRIx64, addr);
> >> printed += fprintf(fp, ")");
> >>
> >>
> >
> > Here are the outputs with above change.
> >
> > For elf in build-id cache, it works as expected.
> > $ sudo perf script -F +dsoff
> > perf-exec 12768 135.648023: 1 cycles: ffffffff96c8ee44 native_write_msr+0x4 (/home/changbin/.debug/.build-id/5e/2fa721660d663f38b6e1aa98d6fa3776974b54/elf+0x28ee44)
> > perf-exec 12768 135.648028: 1 cycles: ffffffff96c8ee44 native_write_msr+0x4 (/home/changbin/.debug/.build-id/5e/2fa721660d663f38b6e1aa98d6fa3776974b54/elf+0x28ee44)
> > perf-exec 12768 135.648030: 11 cycles: ffffffff96c8ee44 native_write_msr+0x4 (/home/changbin/.debug/.build-id/5e/2fa721660d663f38b6e1aa98d6fa3776974b54/elf+0x28ee44)
> > perf-exec 12768 135.648031: 295 cycles: ffffffff96c8ee46 native_write_msr+0x6 (/home/changbin/.debug/.build-id/5e/2fa721660d663f38b6e1aa98d6fa3776974b54/elf+0x28ee46)
> > perf-exec 12768 135.648032: 8850 cycles: ffffffff96c4c686 native_sched_clock+0x66 (/home/changbin/.debug/.build-id/5e/2fa721660d663f38b6e1aa98d6fa3776974b54/elf+0x24c686)
>
> A bit messy though. User can use option --show-kernel-path
> so let's not force that after all.
>
> > ls 27521 501.120978: 4309123 cycles: 7f31cb51c591 _dl_sort_maps+0x301 (/usr/lib/x86_64-linux-gnu/ld-2.31.so)
> >
> > But when I specify my vmlinux: (the binary_type is DSO_BINARY_TYPE__SYSTEM_PATH_DSO)
> > $ sudo perf script -k linux/vmlinux -F +dsoff
> > perf-exec 12768 135.648023: 1 cycles: ffffffff96c8ee44 [unknown] (/lib/modules/6.2.12/build/vmlinux)
> > perf-exec 12768 135.648028: 1 cycles: ffffffff96c8ee44 [unknown] (/lib/modules/6.2.12/build/vmlinux+0xffffffff96c8ee44)
> > perf-exec 12768 135.648030: 11 cycles: ffffffff96c8ee44 [unknown] (/lib/modules/6.2.12/build/vmlinux+0xffffffff96c8ee44)
> > perf-exec 12768 135.648031: 295 cycles: ffffffff96c8ee46 [unknown] (/lib/modules/6.2.12/build/vmlinux+0xffffffff96c8ee46)
> >
> > This is for kcore file:
> > $ sudo perf script --kallsyms /proc/kallsyms -F +dsoff
> > perf-exec 18922 267.284368: 1 cycles: ffffffff96c8ee44 native_write_msr+0x4 (/proc/kcore+0x7fff96c91e44)
> > perf-exec 18922 267.284372: 1 cycles: ffffffff96c8ee44 native_write_msr+0x4 (/proc/kcore+0x7fff96c91e44)
> > perf-exec 18922 267.284374: 11 cycles: ffffffff96c8ee44 native_write_msr+0x4 (/proc/kcore+0x7fff96c91e44)
> >
> > For ko, it's wierd that the binary_type of first one is DSO_BINARY_TYPE__NOT_FOUND.
> > $ sudo perf script -F +dsoff | grep '.ko'
> > gedit 37410 769.927194: 199304 cycles: ffffffffc0a3d050 ipt_do_table+0x0 (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko)
> > gedit 37410 770.459667: 271035 cycles: ffffffffc0a3d050 ipt_do_table+0x0 (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0xf0)
> > gedit 37410 770.649838: 271878 cycles: ffffffffc0a3d050 ipt_do_table+0x0 (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0xf0)
> > gedit 37410 771.239221: 216084 cycles: ffffffffc0a3d13b ipt_do_table+0xeb (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0x1db)
> > gedit 37410 771.257816: 219469 cycles: ffffffffc0a3d165 ipt_do_table+0x115 (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0x205)
> > gedit 37410 771.531158: 288970 cycles: ffffffffc0a3d151 ipt_do_table+0x101 (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0x1f1)
> > gedit 37410 771.816916: 321215 cycles: ffffffffc0a3d151 ipt_do_table+0x101 (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0x1f1)
> > gedit 37410 773.624786: 332528 cycles: ffffffffc0a3d2ea ipt_do_table+0x29a (/lib/modules/6.2.12/kernel/net/ipv4/netfilter/ip_tables.ko+0x38a)
> >
> > Maybe we can just filter by name? e.g. '[kernel.kallsyms]', '[guest.kernel.kallsyms]'.
>
> Using the dso->binary_type deals with other cases also.
> We can just change case DSO_BINARY_TYPE__NOT_FOUND to return
> true.
>
> How about this:
>
> commit abaf1cbf5be5d50b1b3682511b92794394b72178
> Author: Adrian Hunter <adrian.hunter@xxxxxxxxx>
> Date: Fri Apr 21 10:15:14 2023 +0300
>
> perf script: Refine printing of dso offset
>
> Print dso offset only for object files, and in those cases force using the
> dso->long_name if the dso->name starts with '[' or the dso is kcore, in
> order to avoid special names such as [vdso], or mixing up kcore with
> vmlinux.
>
> Signed-off-by: Adrian Hunter <adrian.hunter@xxxxxxxxx>
>
> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> index a86614599269..046fbfcfdaab 100644
> --- a/tools/perf/util/dso.c
> +++ b/tools/perf/util/dso.c
> @@ -67,6 +67,39 @@ char dso__symtab_origin(const struct dso *dso)
> return origin[dso->symtab_type];
> }
>
> +bool dso__is_object_file(const struct dso *dso)
> +{
> + switch (dso->binary_type) {
> + case DSO_BINARY_TYPE__KALLSYMS:
> + case DSO_BINARY_TYPE__GUEST_KALLSYMS:
> + case DSO_BINARY_TYPE__JAVA_JIT:
> + case DSO_BINARY_TYPE__BPF_PROG_INFO:
> + case DSO_BINARY_TYPE__BPF_IMAGE:
> + case DSO_BINARY_TYPE__OOL:
> + return false;
> + case DSO_BINARY_TYPE__VMLINUX:
> + case DSO_BINARY_TYPE__GUEST_VMLINUX:
> + case DSO_BINARY_TYPE__DEBUGLINK:
> + case DSO_BINARY_TYPE__BUILD_ID_CACHE:
> + case DSO_BINARY_TYPE__BUILD_ID_CACHE_DEBUGINFO:
> + case DSO_BINARY_TYPE__FEDORA_DEBUGINFO:
> + case DSO_BINARY_TYPE__UBUNTU_DEBUGINFO:
> + case DSO_BINARY_TYPE__MIXEDUP_UBUNTU_DEBUGINFO:
> + case DSO_BINARY_TYPE__BUILDID_DEBUGINFO:
> + case DSO_BINARY_TYPE__SYSTEM_PATH_DSO:
> + case DSO_BINARY_TYPE__GUEST_KMODULE:
> + case DSO_BINARY_TYPE__GUEST_KMODULE_COMP:
> + case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE:
> + case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP:
> + case DSO_BINARY_TYPE__KCORE:
> + case DSO_BINARY_TYPE__GUEST_KCORE:
> + case DSO_BINARY_TYPE__OPENEMBEDDED_DEBUGINFO:
> + case DSO_BINARY_TYPE__NOT_FOUND:
> + default:
> + return true;
> + }
> +}
> +
> int dso__read_binary_type_filename(const struct dso *dso,
> enum dso_binary_type type,
> char *root_dir, char *filename, size_t size)
> diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
> index 0b7c7633b9f6..b23a157c914d 100644
> --- a/tools/perf/util/dso.h
> +++ b/tools/perf/util/dso.h
> @@ -379,23 +379,25 @@ void dso__reset_find_symbol_cache(struct dso *dso);
> size_t dso__fprintf_symbols_by_name(struct dso *dso, FILE *fp);
> size_t dso__fprintf(struct dso *dso, FILE *fp);
>
> -static inline bool dso__is_vmlinux(struct dso *dso)
> +static inline bool dso__is_vmlinux(const struct dso *dso)
> {
> return dso->binary_type == DSO_BINARY_TYPE__VMLINUX ||
> dso->binary_type == DSO_BINARY_TYPE__GUEST_VMLINUX;
> }
>
> -static inline bool dso__is_kcore(struct dso *dso)
> +static inline bool dso__is_kcore(const struct dso *dso)
> {
> return dso->binary_type == DSO_BINARY_TYPE__KCORE ||
> dso->binary_type == DSO_BINARY_TYPE__GUEST_KCORE;
> }
>
> -static inline bool dso__is_kallsyms(struct dso *dso)
> +static inline bool dso__is_kallsyms(const struct dso *dso)
> {
> return dso->kernel && dso->long_name[0] != '/';
> }
>
> +bool dso__is_object_file(const struct dso *dso);
> +
> void dso__free_a2l(struct dso *dso);
>
> enum dso_type dso__type(struct dso *dso, struct machine *machine);
> diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
> index 7da96b41100f..5e7808b0bc87 100644
> --- a/tools/perf/util/map.c
> +++ b/tools/perf/util/map.c
> @@ -424,14 +424,21 @@ size_t map__fprintf(struct map *map, FILE *fp)
> map__start(map), map__end(map), map__pgoff(map), dso->name);
> }
>
> -size_t map__fprintf_dsoname(struct map *map, FILE *fp)
> +static bool prefer_dso_long_name(const struct dso *dso, bool print_off)
> +{
> + return dso->long_name &&
> + (symbol_conf.show_kernel_path ||
> + (print_off && (dso->name[0] == '[' || dso__is_kcore(dso))));
> +}
> +
> +static size_t __map__fprintf_dsoname(struct map *map, bool print_off, FILE *fp)
> {
> char buf[symbol_conf.pad_output_len_dso + 1];
> const char *dsoname = "[unknown]";
> const struct dso *dso = map ? map__dso(map) : NULL;
>
> if (dso) {
> - if (symbol_conf.show_kernel_path && dso->long_name)
> + if (prefer_dso_long_name(dso, print_off))
> dsoname = dso->long_name;
> else
> dsoname = dso->name;
> @@ -445,13 +452,21 @@ size_t map__fprintf_dsoname(struct map *map, FILE *fp)
> return fprintf(fp, "%s", dsoname);
> }
>
> +size_t map__fprintf_dsoname(struct map *map, FILE *fp)
> +{
> + return __map__fprintf_dsoname(map, false, fp);
> +}
> +
> size_t map__fprintf_dsoname_dsoff(struct map *map, bool print_off, u64 addr, FILE *fp)
> {
> + const struct dso *dso = map ? map__dso(map) : NULL;
> int printed = 0;
>
> + if (print_off && (!dso || !dso__is_object_file(dso)))
> + print_off = false;
> printed += fprintf(fp, " (");
> - printed += map__fprintf_dsoname(map, fp);
> - if (print_off && map && map__dso(map) && !map__dso(map)->kernel)
> + printed += __map__fprintf_dsoname(map, print_off, fp);
> + if (print_off)
> printed += fprintf(fp, "+0x%" PRIx64, addr);
> printed += fprintf(fp, ")");
>
>
>
I am fine with this polish patch. So will send this standalone or append to this
thread?

--
Cheers,
Changbin Du

Next message: Xu,Rongbo: "Re: [PATCH] [PATCH] fuse: no interrupt for lookup request"
Previous message: Zhou Furong: "Re: [PATCH v7 7/7] drm/i915: track gt pm wakerefs"
In reply to: Adrian Hunter: "Re: [PATCH v5 2/3] perf: add helper map__fprintf_dsoname_dsoff"
Next in thread: Changbin Du: "[PATCH v5 3/3] perf: script: add new output field 'dsoff' to print dso offset"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]