libperf: Add sampling interface
Jiri Olsa
reviewerv2 cover libperf: Add sampling interface
hi,
sending changes for exporting basic sampling interface in libperf. It's now possible to use following code in applications via libperf:
--- (example is without error checks for simplicity)
struct perf_event_attr attr = {
.type = PERF_TYPE_TRACEPOINT,
.sample_period = 1,
.wakeup_watermark = 1,
.disabled = 1,
};
/* ... setup attr */
cpus = perf_cpu_map__new(NULL);
evlist = perf_evlist__new();
evsel = perf_evsel__new(&attr);
perf_evlist__add(evlist, evsel);
perf_evlist__set_maps(evlist, cpus, NULL);
err = perf_evlist__open(evlist);
err = perf_evlist__mmap(evlist, 4);
err = perf_evlist__enable(evlist);/* ... monitored area, plus all the other cpus */
err = perf_evlist__disable(evlist);
perf_evlist__for_each_mmap(evlist, map) {
if (perf_mmap__read_init(map) < 0)
continue; while ((event = perf_mmap__read_event(map)) != NULL) {
perf_mmap__consume(map);
}
perf_mmap__read_done(map);}
perf_evlist__delete(evlist); perf_cpu_map__put(cpus);
--- (end)
Nothing is carved in stone so far, the interface is exported as is available in perf now and we can change it as we want.
New tests are added in test-evlist.c to do thread and cpu based sampling.
All the functionality should not change, however there's considerable mmap code rewrite.
Now we have perf_evlist__mmap_ops called by both perf and libperf mmaps functions with specific 'struct perf_evlist_mmap_ops' callbacks:
- get - to get mmap object, both libperf and perf use different objects, because perf needs to carry more data for aio, compression and auxtrace
- mmap - to actually mmap the object, it's simple mmap for libperf, but more work for perf wrt aio, compression and auxtrace
- idx - callback to get current IDs, used only in perf for auxtrace setup
It would be great if guys could run your usual workloads to see if all is fine.. so far so good in my tests ;-)
It's also available in here:
git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
perf/lib
v2 changes:
- rebased to latest perf/core
- portion of patches already taken
- explained mmap refcnt management in following patch changelog:
libperf: Centralize map refcnt setting
thanks,
jirka
---
Jiri Olsa (36):
libperf: Add perf_mmap__init() function
libperf: Add 'struct perf_mmap_param'
libperf: Add perf_mmap__mmap_len() function
libperf: Add perf_mmap__mmap() function
libperf: Add perf_mmap__get() function
libperf: Add perf_mmap__unmap() function
libperf: Add perf_mmap__put() function
perf tools: Use perf_mmap way to detect aux mmap
libperf: Add perf_mmap__consume() function
libperf: Add perf_mmap__read_init() function
libperf: Add perf_mmap__read_done() function
libperf: Add perf_mmap__read_event() function
libperf: Add perf_evlist__mmap()/munmap() functions
libperf: Add perf_evlist__mmap_ops function
libperf: Add perf_evlist_mmap_ops::idx callback
libperf: Add perf_evlist_mmap_ops::get callback
libperf: Add perf_evlist_mmap_ops::mmap callback perf tools: Add perf_evlist__mmap_cb_idx function perf tools: Add perf_evlist__mmap_cb_get function perf tools: Add perf_evlist__mmap_cb_mmap function perf tools: Switch to libperf mmap interface
libperf: Centralize map refcnt setting
libperf: Move pollfd allocation to libperf
libperf: Add perf_evlist__exit function
libperf: Add perf_evlist__purge function
libperf: Add perf_evlist__filter_pollfd function
libperf: Add perf_evlist__for_each_mmap function
libperf: Move mmap allocation to perf_evlist__mmap_ops::get
libperf: Move mask setup to perf_evlist__mmap_ops function
libperf: Link static tests with libapi.a
libperf: Add _GNU_SOURCE define to compilation
libperf: Add tests_mmap_thread test
libperf: Add tests_mmap_cpus test
libperf: Keep count of failed tests
libperf: Do not export perf_evsel__init/perf_evlist__init
libperf: Add pr_err macro
tools/perf/arch/x86/tests/perf-time-to-tsc.c | 9 ++-- tools/perf/builtin-kvm.c | 11 ++--- tools/perf/builtin-record.c | 10 ++--- tools/perf/builtin-top.c | 9 ++-- tools/perf/builtin-trace.c | 9 ++-- tools/perf/lib/Build | 1 + tools/perf/lib/Makefile | 7 +++- tools/perf/lib/evlist.c | 357 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tools/perf/lib/include/internal/evlist.h | 43 ++++++++++++++++++++ tools/perf/lib/include/internal/evsel.h | 1 + tools/perf/lib/include/internal/mmap.h | 46 ++++++++++++++++----- tools/perf/lib/include/internal/tests.h | 20 +++++++-- tools/perf/lib/include/perf/core.h | 3 ++ tools/perf/lib/include/perf/evlist.h | 15 ++++++- tools/perf/lib/include/perf/evsel.h | 2 - tools/perf/lib/include/perf/mmap.h | 14 +++++++ tools/perf/lib/internal.h | 5 +++ tools/perf/lib/libperf.map | 10 ++++- tools/perf/lib/mmap.c | 274 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tools/perf/lib/tests/Makefile | 8 ++-- tools/perf/lib/tests/test-cpumap.c | 2 +- tools/perf/lib/tests/test-evlist.c | 218 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- tools/perf/lib/tests/test-evsel.c | 2 +- tools/perf/lib/tests/test-threadmap.c | 2 +- tools/perf/tests/backward-ring-buffer.c | 7 ++-- tools/perf/tests/bpf.c | 7 ++-- tools/perf/tests/code-reading.c | 9 ++-- tools/perf/tests/keep-tracking.c | 9 ++-- tools/perf/tests/mmap-basic.c | 9 ++-- tools/perf/tests/openat-syscall-tp-fields.c | 9 ++-- tools/perf/tests/perf-record.c | 9 ++-- tools/perf/tests/sw-clock.c | 9 ++-- tools/perf/tests/switch-tracking.c | 9 ++-- tools/perf/tests/task-exit.c | 9 ++-- tools/perf/util/evlist.c | 256 +++++++++++++++++++++++++++--------------------------------------------------------------------------------------- tools/perf/util/mmap.c | 260 +++++++------------------------------------------------------------------------------------------------------------- tools/perf/util/mmap.h | 28 +++---------- tools/perf/util/python.c | 7 ++-- 38 files changed, 1161 insertions(+), 554 deletions(-) create mode 100644 tools/perf/lib/include/perf/mmap.h create mode 100644 tools/perf/lib/mmap.c
Jiri Olsa
reviewerPATCH v2 1/36 libperf: Add perf_mmap__init() function
Add perf_mmap__init() function to initialize 'struct perf_mmap' object.
Add it to a new mmap.c object, that will carry all the mmap related functions.
--- tools/perf/lib/Build | 1 + tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/mmap.c | 9 +++++++++ tools/perf/util/evlist.c | 5 ++--- 4 files changed, 14 insertions(+), 3 deletions(-) create mode 100644 tools/perf/lib/mmap.c
tools/perf/lib/Build +1 -0
@@ -3,6 +3,7 @@ libperf-y += cpumap.o libperf-y += threadmap.o libperf-y += evsel.o libperf-y += evlist.o +libperf-y += mmap.o libperf-y += zalloc.o libperf-y += xyarray.o libperf-y += lib.o
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -29,4 +29,6 @@ struct perf_mmap { char event_copy[PERF_SAMPLE_MAX_SIZE] __aligned(8); }; +void perf_mmap__init (struct perf_mmap *map,bool overwrite); +#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c new file +9 -0
@@ -0,0 +1,9 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <internal/mmap.h> + +void perf_mmap__init (struct perf_mmap *map,bool overwrite) +{ + map->fd =-1 ; + map->overwrite = overwrite; +refcount_set (&map->refcnt,0 ); +}
tools/perf/util/evlist.c +2 -3
@@ -605,8 +605,6 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return NULL ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { - map[i].core.fd =-1 ; - map[i].core.overwrite = overwrite;/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last
@@ -616,8 +614,9 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -refcount_set (&map[i].core.refcnt,0 ); +perf_mmap__init (&map[i].core, overwrite); } +return map; }
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 2/36 libperf: Add 'struct perf_mmap_param'
In reply to Jiri OlsaAdd libperf's version of mmap params 'struct perf_mmap_param' object with the basics: 'prot' and 'mask'. Encapsulate it in the current 'struct mmap_params' object.
--- tools/perf/lib/include/internal/mmap.h | 5 +++++ tools/perf/util/evlist.c | 14 +++++++++----- tools/perf/util/mmap.c | 4 ++-- tools/perf/util/mmap.h | 3 ++- 4 files changed, 18 insertions(+), 8 deletions(-)
tools/perf/lib/include/internal/mmap.h +5 -0
@@ -29,6 +29,11 @@ struct perf_mmap { char event_copy[PERF_SAMPLE_MAX_SIZE] __aligned(8); }; +struct perf_mmap_param { +int prot; +int mask; +}; +void perf_mmap__init (struct perf_mmap *map,bool overwrite);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/util/evlist.c +9 -5
@@ -643,7 +643,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,int fd;int cpu; - mp->prot =PROT_READ |PROT_WRITE ; + mp->core.prot =PROT_READ |PROT_WRITE ;if (evsel->core.attr.write_backward) { output = _output_overwrite; maps = evlist->overwrite_mmap;
@@ -656,7 +656,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY )perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); } - mp->prot &= ~PROT_WRITE ; + mp->core.prot &= ~PROT_WRITE ; }if (evsel->core.system_wide && thread)
@@ -897,8 +897,12 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages,* Its value is decided by evsel's write_backward. * So &mp should not be passed through const pointer. */ -struct mmap_params mp = { .nr_cblocks = nr_cblocks, .affinity = affinity, .flush = flush, - .comp_level = comp_level }; +struct mmap_params mp = { + .nr_cblocks = nr_cblocks, + .affinity = affinity, + .flush = flush, + .comp_level = comp_level + };if (!evlist->mmap) evlist->mmap =evlist__alloc_mmap (evlist, false);
@@ -910,7 +914,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); - mp.mask = evlist->core.mmap_len - page_size -1 ; + mp.core.mask = evlist->core.mmap_len - page_size -1 ;auxtrace_mmap_params__init (&mp.auxtrace_mp, evlist->core.mmap_len, auxtrace_pages, auxtrace_overwrite);
tools/perf/util/mmap.c +2 -2
@@ -370,8 +370,8 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu)*/ refcount_set (&map->core.refcnt,2 ); map->core.prev =0 ; - map->core.mask = mp->mask; - map->core.base =mmap (NULL ,perf_mmap__mmap_len (map), mp->prot, + map->core.mask = mp->core.mask; + map->core.base =mmap (NULL ,perf_mmap__mmap_len (map), mp->core.prot,MAP_SHARED , fd,0 );if (map->core.base ==MAP_FAILED ) { pr_debug2("failed to mmap perf event ring buffer, error %d\n",
tools/perf/util/mmap.h +2 -1
@@ -37,7 +37,8 @@ struct mmap { };struct mmap_params { -int prot, mask, nr_cblocks, affinity, flush, comp_level; +struct perf_mmap_param core; +int nr_cblocks, affinity, flush, comp_level;struct auxtrace_mmap_params auxtrace_mp; };
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 3/36 libperf: Add perf_mmap__mmap_len() function
In reply to Jiri OlsaMove perf_mmap__mmap_len() from tools/perf wto libperf, it will be used in the following patches. And rename the existing perf's function to mmap__mmap_len().
--- tools/perf/builtin-record.c | 4 ++-- tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/mmap.c | 6 ++++++ tools/perf/util/mmap.c | 20 ++++++++++---------- tools/perf/util/mmap.h | 2 +- 5 files changed, 21 insertions(+), 13 deletions(-)
tools/perf/builtin-record.c +2 -2
@@ -276,7 +276,7 @@ static int record__aio_pushfn(struct mmap *map, void *to, void *buf, size_t sizeif (record__comp_enabled (aio->rec)) { size =zstd_compress (aio->rec->session, aio->data + aio->size, -perf_mmap__mmap_len (map) - aio->size, +mmap__mmap_len (map) - aio->size, buf, size); }else {memcpy (aio->data + aio->size, buf, size);
@@ -488,7 +488,7 @@ static int record__pushfn(struct mmap *map, void *to, void *bf, size_t size)struct record *rec = to;if (record__comp_enabled (rec)) { - size =zstd_compress (rec->session, map->data,perf_mmap__mmap_len (map), bf, size); + size =zstd_compress (rec->session, map->data,mmap__mmap_len (map), bf, size); bf = map->data; }
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -34,6 +34,8 @@ struct perf_mmap_param { int mask; }; +size_t perf_mmap__mmap_len (struct perf_mmap *map); +void perf_mmap__init (struct perf_mmap *map,bool overwrite);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +6 -0
@@ -1,5 +1,6 @@// SPDX-License-Identifier: GPL-2.0 #include <internal/mmap.h> +#include <internal/lib.h> void perf_mmap__init (struct perf_mmap *map,bool overwrite) {
@@ -7,3 +8,8 @@ void perf_mmap__init(struct perf_mmap *map, bool overwrite) map->overwrite = overwrite;refcount_set (&map->refcnt,0 ); } + +size_t perf_mmap__mmap_len (struct perf_mmap *map) +{ +return map->mask +1 + page_size; +}
tools/perf/util/mmap.c +10 -10
@@ -23,9 +23,9 @@#include "../perf.h" #include <internal/lib.h> /* page_size */ -size_t perf_mmap__mmap_len (struct mmap *map) +size_t mmap__mmap_len (struct mmap *map) { -return map->core.mask +1 + page_size; +return perf_mmap__mmap_len (&map->core); }/* When check_messup is true, 'end' must points to a good entry */
@@ -170,7 +170,7 @@ static int perf_mmap__aio_enabled(struct mmap *map)#ifdef HAVE_LIBNUMA_SUPPORT static int perf_mmap__aio_alloc (struct mmap *map,int idx) { - map->aio.data[idx] =mmap (NULL ,perf_mmap__mmap_len (map),PROT_READ |PROT_WRITE , + map->aio.data[idx] =mmap (NULL ,mmap__mmap_len (map),PROT_READ |PROT_WRITE ,MAP_PRIVATE |MAP_ANONYMOUS ,0 ,0 );if (map->aio.data[idx] ==MAP_FAILED ) { map->aio.data[idx] =NULL ;
@@ -183,7 +183,7 @@ static int perf_mmap__aio_alloc(struct mmap *map, int idx)static void perf_mmap__aio_free (struct mmap *map,int idx) {if (map->aio.data[idx]) { -munmap (map->aio.data[idx],perf_mmap__mmap_len (map)); +munmap (map->aio.data[idx],mmap__mmap_len (map)); map->aio.data[idx] =NULL ; } }
@@ -196,7 +196,7 @@ static int perf_mmap__aio_bind(struct mmap *map, int idx, int cpu, int affinity)if (affinity !=PERF_AFFINITY_SYS &&cpu__max_node () >1 ) { data = map->aio.data[idx]; - mmap_len =perf_mmap__mmap_len (map); + mmap_len =mmap__mmap_len (map); node_mask =1UL <<cpu__get_node (cpu);if (mbind (data, mmap_len,MPOL_BIND , &node_mask,1 ,0 )) { pr_err("Failed to bind [%p-%p] AIO buffer to node %d: error %m\n",
@@ -210,7 +210,7 @@ static int perf_mmap__aio_bind(struct mmap *map, int idx, int cpu, int affinity)#else /* !HAVE_LIBNUMA_SUPPORT */ static int perf_mmap__aio_alloc (struct mmap *map,int idx) { - map->aio.data[idx] =malloc (perf_mmap__mmap_len (map)); + map->aio.data[idx] =malloc (mmap__mmap_len (map));if (map->aio.data[idx] ==NULL )return -1 ;
@@ -315,11 +315,11 @@ void perf_mmap__munmap(struct mmap *map) {perf_mmap__aio_munmap (map);if (map->data !=NULL ) { -munmap (map->data,perf_mmap__mmap_len (map)); +munmap (map->data,mmap__mmap_len (map)); map->data =NULL ; }if (map->core.base !=NULL ) { -munmap (map->core.base,perf_mmap__mmap_len (map)); +munmap (map->core.base,mmap__mmap_len (map)); map->core.base =NULL ; map->core.fd =-1 ;refcount_set (&map->core.refcnt,0 );
@@ -371,7 +371,7 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu)refcount_set (&map->core.refcnt,2 ); map->core.prev =0 ; map->core.mask = mp->core.mask; - map->core.base =mmap (NULL ,perf_mmap__mmap_len (map), mp->core.prot, + map->core.base =mmap (NULL ,mmap__mmap_len (map), mp->core.prot,MAP_SHARED , fd,0 );if (map->core.base ==MAP_FAILED ) { pr_debug2("failed to mmap perf event ring buffer, error %d\n",
@@ -389,7 +389,7 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu) map->comp_level = mp->comp_level;if (map->comp_level && !perf_mmap__aio_enabled (map)) { - map->data =mmap (NULL ,perf_mmap__mmap_len (map),PROT_READ |PROT_WRITE , + map->data =mmap (NULL ,mmap__mmap_len (map),PROT_READ |PROT_WRITE ,MAP_PRIVATE |MAP_ANONYMOUS ,0 ,0 );if (map->data ==MAP_FAILED ) { pr_debug2("failed to mmap data buffer, error %d\n",
tools/perf/util/mmap.h +1 -1
@@ -67,7 +67,7 @@ union perf_event *perf_mmap__read_event(struct mmap *map);int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size)); -size_t perf_mmap__mmap_len (struct mmap *map); +size_t mmap__mmap_len (struct mmap *map);int perf_mmap__read_init (struct mmap *md);void perf_mmap__read_done (struct mmap *map);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 5/36 libperf: Add perf_mmap__get() function
In reply to Jiri OlsaMove perf_mmap__get() from tools/perf to libperf in the internal header internal/mmap.h.
--- tools/perf/builtin-record.c | 2 +- tools/perf/lib/include/internal/mmap.h | 1 + tools/perf/lib/mmap.c | 5 +++++ tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 5 ----- tools/perf/util/mmap.h | 1 - 6 files changed, 8 insertions(+), 8 deletions(-)
tools/perf/builtin-record.c +1 -1
@@ -293,7 +293,7 @@ static int record__aio_pushfn(struct mmap *map, void *to, void *buf, size_t size* after started aio request completion or at record__aio_push() * if the request failed to start. */ -perf_mmap__get (map); +perf_mmap__get (&map->core); } aio->size += size;
tools/perf/lib/include/internal/mmap.h +1 -0
@@ -39,5 +39,6 @@ size_t perf_mmap__mmap_len(struct perf_mmap *map);void perf_mmap__init (struct perf_mmap *map,bool overwrite);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu); +void perf_mmap__get (struct perf_mmap *map);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +5 -0
@@ -31,3 +31,8 @@ int perf_mmap__mmap(struct perf_mmap *map, struct perf_mmap_param *mp, map->cpu = cpu;return 0 ; } + +void perf_mmap__get (struct perf_mmap *map) +{ +refcount_inc (&map->refcnt); +}
tools/perf/util/evlist.c +1 -1
@@ -677,7 +677,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )return -1 ; -perf_mmap__get (&maps[idx]); +perf_mmap__get (&maps[idx].core); } revent =perf_evlist__should_poll (evlist, evsel) ?POLLIN :0 ;
tools/perf/util/mmap.c +0 -5
@@ -110,11 +110,6 @@ static bool perf_mmap__empty(struct mmap *map)return perf_mmap__read_head (map) == map->core.prev && !map->auxtrace_mmap.base; } -void perf_mmap__get (struct mmap *map) -{ -refcount_inc (&map->core.refcnt); -} -void perf_mmap__put (struct mmap *map) {BUG_ON (map->core.base &&refcount_read (&map->core.refcnt) ==0 );
tools/perf/util/mmap.h +0 -1
@@ -45,7 +45,6 @@ struct mmap_params {int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void perf_mmap__munmap (struct mmap *map); -void perf_mmap__get (struct mmap *map);void perf_mmap__put (struct mmap *map);void perf_mmap__consume (struct mmap *map);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 7/36 libperf: Add perf_mmap__put() function
In reply to Jiri OlsaMove perf_mmap__put() from tools/perf to libperf.
Once perf_mmap__put() is moved, we need a way to call application related unmap code (AIO and aux related code for eprf), when the map goes away.
Add the perf_mmap::unmap callback to do that.
The unmap path from perf is:
perf_mmap__put (libperf)
perf_mmap__munmap (libperf)
map->unmap_cb -> perf_mmap__unmap_cb (perf)
mmap__munmap (perf)--- tools/perf/builtin-record.c | 4 ++-- tools/perf/lib/include/internal/mmap.h | 31 ++++++++++++++++---------- tools/perf/lib/mmap.c | 14 +++++++++++- tools/perf/util/evlist.c | 17 +++++++++----- tools/perf/util/mmap.c | 11 +-------- tools/perf/util/mmap.h | 2 -- 6 files changed, 47 insertions(+), 32 deletions(-)
tools/perf/builtin-record.c +2 -2
@@ -197,7 +197,7 @@ static int record__aio_complete(struct mmap *md, struct aiocb *cblock)* every aio write request started in record__aio_push() so * decrement it because the request is now complete. */ -perf_mmap__put (md); +perf_mmap__put (&md->core); rc =1 ; }else {/*
@@ -332,7 +332,7 @@ static int record__aio_push(struct record *rec, struct mmap *map, off_t *off)* map->refcount is decremented in record__aio_complete() after * aio write operation finishes successfully. */ -perf_mmap__put (map); +perf_mmap__put (&map->core); }return ret;
tools/perf/lib/include/internal/mmap.h +19 -12
@@ -10,23 +10,28 @@/* perf sample has 16 bits size limit */ #define PERF_SAMPLE_MAX_SIZE (1 << 16) +struct perf_mmap ; + +typedef void (*libperf_unmap_cb_t )(struct perf_mmap *map); +/** * struct perf_mmap - perf's ring buffer mmap details * * @refcnt - e.g. code using PERF_EVENT_IOC_SET_OUTPUT to share this */ struct perf_mmap { -void *base; -int mask; -int fd; -int cpu; -refcount_t refcnt; -u64 prev; -u64 start; -u64 end; -bool overwrite; -u64 flush; -char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; +void *base; +int mask; +int fd; +int cpu; +refcount_t refcnt; +u64 prev; +u64 start; +u64 end; +bool overwrite; +u64 flush; +libperf_unmap_cb_t unmap_cb; +char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; };struct perf_mmap_param {
@@ -36,10 +41,12 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map); -void perf_mmap__init (struct perf_mmap *map,bool overwrite); +void perf_mmap__init (struct perf_mmap *map,bool overwrite, +libperf_unmap_cb_t unmap_cb);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu);void perf_mmap__munmap (struct perf_mmap *map);void perf_mmap__get (struct perf_mmap *map); +void perf_mmap__put (struct perf_mmap *map);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +13 -1
@@ -3,10 +3,12 @@#include <internal/mmap.h> #include <internal/lib.h> -void perf_mmap__init (struct perf_mmap *map,bool overwrite) +void perf_mmap__init (struct perf_mmap *map,bool overwrite, +libperf_unmap_cb_t unmap_cb) { map->fd =-1 ; map->overwrite = overwrite; + map->unmap_cb = unmap_cb;refcount_set (&map->refcnt,0 ); }
@@ -40,9 +42,19 @@ void perf_mmap__munmap(struct perf_mmap *map) map->fd =-1 ;refcount_set (&map->refcnt,0 ); } +if (map && map->unmap_cb) + map->unmap_cb (map); }void perf_mmap__get (struct perf_mmap *map) {refcount_inc (&map->refcnt); } + +void perf_mmap__put (struct perf_mmap *map) +{ +BUG_ON (map->base &&refcount_read (&map->refcnt) ==0 ); + +if (refcount_dec_and_test (&map->refcnt)) +perf_mmap__munmap (map); +}
tools/perf/util/evlist.c +12 -5
@@ -409,7 +409,7 @@ static void perf_evlist__munmap_filtered(struct fdarray *fda, int fd,struct mmap *map = fda->priv[fd].ptr;if (map) -perf_mmap__put (map); +perf_mmap__put (&map->core); }int evlist__filter_pollfd (struct evlist *evlist,short revents_and_mask)
@@ -577,11 +577,11 @@ static void evlist__munmap_nofree(struct evlist *evlist)if (evlist->mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -mmap__munmap (&evlist->mmap[i]); +perf_mmap__munmap (&evlist->mmap[i].core);if (evlist->overwrite_mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -mmap__munmap (&evlist->overwrite_mmap[i]); +perf_mmap__munmap (&evlist->overwrite_mmap[i].core); }void evlist__munmap (struct evlist *evlist)
@@ -591,6 +591,13 @@ void evlist__munmap(struct evlist *evlist)zfree (&evlist->overwrite_mmap); } +static void perf_mmap__unmap_cb (struct perf_mmap *map) +{ + struct mmap *m = container_of(map, struct mmap, core); + +mmap__munmap (m); +} +static struct mmap *evlist__alloc_mmap (struct evlist *evlist,bool overwrite) {
@@ -614,7 +621,7 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i].core, overwrite); +perf_mmap__init (&map[i].core, overwrite, perf_mmap__unmap_cb); }return map;
@@ -691,7 +698,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,*/ if (!evsel->core.system_wide &&perf_evlist__add_pollfd (&evlist->core, fd, &maps[idx], revent) <0 ) { -perf_mmap__put (&maps[idx]); +perf_mmap__put (&maps[idx].core);return -1 ; }
tools/perf/util/mmap.c +1 -10
@@ -110,14 +110,6 @@ static bool perf_mmap__empty(struct mmap *map)return perf_mmap__read_head (map) == map->core.prev && !map->auxtrace_mmap.base; } -void perf_mmap__put (struct mmap *map) -{ -BUG_ON (map->core.base &&refcount_read (&map->core.refcnt) ==0 ); - -if (refcount_dec_and_test (&map->core.refcnt)) -mmap__munmap (map); -} -void perf_mmap__consume (struct mmap *map) {if (!map->core.overwrite) {
@@ -127,7 +119,7 @@ void perf_mmap__consume(struct mmap *map) }if (refcount_read (&map->core.refcnt) ==1 &&perf_mmap__empty (map)) -perf_mmap__put (map); +perf_mmap__put (&map->core); } int __weak auxtrace_mmap__mmap(struct auxtrace_mmap *mm __maybe_unused,
@@ -308,7 +300,6 @@ static void perf_mmap__aio_munmap(struct mmap *map __maybe_unused)void mmap__munmap (struct mmap *map) { -perf_mmap__munmap (&map->core);perf_mmap__aio_munmap (map);if (map->data !=NULL ) {munmap (map->data,mmap__mmap_len (map));
tools/perf/util/mmap.h +0 -2
@@ -45,8 +45,6 @@ struct mmap_params {int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void mmap__munmap (struct mmap *map); -void perf_mmap__put (struct mmap *map); -void perf_mmap__consume (struct mmap *map);static inline u64 perf_mmap__read_head (struct mmap *mm)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 10/36 libperf: Add perf_mmap__read_init() function
In reply to Jiri OlsaMove perf_mmap__read_init() from tools/perf to libperf and export it in perf/mmap.h header.
And add pr_debug2()/pr_debug3() macros support, because the code is using them.
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/perf/core.h | 2 + tools/perf/lib/include/perf/mmap.h | 1 + tools/perf/lib/internal.h | 2 + tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 84 ++++++++++++++++++++ tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 82 +------------------ tools/perf/util/mmap.h | 1 - tools/perf/util/python.c | 2 +- 23 files changed, 107 insertions(+), 98 deletions(-)
tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 -1
@@ -118,7 +118,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybefor (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/builtin-kvm.c +1 -1
@@ -760,7 +760,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx, *mmap_time =ULLONG_MAX ; md = &evlist->mmap[idx]; - err =perf_mmap__read_init (md); + err =perf_mmap__read_init (&md->core);if (err <0 )return (err == -EAGAIN ) ?0 :-1 ;
tools/perf/builtin-top.c +1 -1
@@ -870,7 +870,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)union perf_event *event; md = opts->overwrite ? &evlist->overwrite_mmap[idx] : &evlist->mmap[idx]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )return ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/builtin-trace.c +1 -1
@@ -3451,7 +3451,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)struct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/lib/include/perf/core.h +2 -0
@@ -12,6 +12,8 @@ enum libperf_print_level {LIBPERF_WARN ,LIBPERF_INFO ,LIBPERF_DEBUG , +LIBPERF_DEBUG2 , + LIBPERF_DEBUG3, }; typedef int (*libperf_print_fn_t)(enum libperf_print_level level,
tools/perf/lib/include/perf/mmap.h +1 -0
@@ -7,5 +7,6 @@struct perf_mmap ; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); +LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/internal.h +2 -0
@@ -14,5 +14,7 @@ do { \#define pr_warning (fmt, ...) __pr(LIBPERF_WARN, fmt, ##__VA_ARGS__)#define pr_info (fmt, ...) __pr(LIBPERF_INFO, fmt, ##__VA_ARGS__)#define pr_debug (fmt, ...) __pr(LIBPERF_DEBUG, fmt, ##__VA_ARGS__) +#define pr_debug2 (fmt, ...) __pr(LIBPERF_DEBUG2, fmt, ##__VA_ARGS__) +#define pr_debug3 (fmt, ...) __pr(LIBPERF_DEBUG3, fmt, ##__VA_ARGS__)#endif /* __LIBPERF_INTERNAL_H */
tools/perf/lib/libperf.map +1 -0
@@ -41,6 +41,7 @@ LIBPERF_0.0.1 { perf_evlist__set_maps; perf_evlist__poll; perf_mmap__consume; + perf_mmap__read_init; local: *; };
tools/perf/lib/mmap.c +84 -0
97 lines not on this page · raw message
tools/perf/tests/backward-ring-buffer.c +1 -1
@@ -38,7 +38,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,struct mmap *map = &evlist->overwrite_mmap[i];union perf_event *event; -perf_mmap__read_init (map); +perf_mmap__read_init (&map->core);while ((event =perf_mmap__read_event (map)) !=NULL ) {const u32 type = event->header.type;
tools/perf/tests/bpf.c +1 -1
@@ -185,7 +185,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),struct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/code-reading.c +1 -1
@@ -426,7 +426,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,for (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/keep-tracking.c +1 -1
@@ -39,7 +39,7 @@ static int find_comm(struct evlist *evlist, const char *comm) found =0 ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) { if (event->header.type == PERF_RECORD_COMM &&
tools/perf/tests/mmap-basic.c +1 -1
@@ -114,7 +114,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuse } md = &evlist->mmap[0 ]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto out_init;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/openat-syscall-tp-fields.c +1 -1
@@ -93,7 +93,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subteststruct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/perf-record.c +1 -1
@@ -171,7 +171,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusstruct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/sw-clock.c +1 -1
@@ -100,7 +100,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)evlist__disable (evlist); md = &evlist->mmap[0 ]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto out_init;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/switch-tracking.c +1 -1
@@ -270,7 +270,7 @@ static int process_events(struct evlist *evlist,for (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/task-exit.c +1 -1
@@ -118,7 +118,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused retry: md = &evlist->mmap[0 ]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto out_init;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/util/evlist.c +1 -1
@@ -1740,7 +1740,7 @@ static void *perf_evlist__poll_thread(void *arg)struct mmap *map = &evlist->mmap[i];union perf_event *event; -if (perf_mmap__read_init (map)) +if (perf_mmap__read_init (&map->core))continue ;while ((event =perf_mmap__read_event (map)) !=NULL ) {struct evsel *evsel =perf_evlist__event2evsel (evlist, event);
tools/perf/util/mmap.c +1 -81
94 lines not on this page · raw message
tools/perf/util/mmap.h +0 -1
@@ -54,6 +54,5 @@ int perf_mmap__push(struct mmap *md, void *to,size_t mmap__mmap_len (struct mmap *map); -int perf_mmap__read_init (struct mmap *md);void perf_mmap__read_done (struct mmap *map);#endif /*__PERF_MMAP_H */
tools/perf/util/python.c +1 -1
@@ -1017,7 +1017,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist,if (!md)return NULL ; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto end; event =perf_mmap__read_event (md);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 11/36 libperf: Add perf_mmap__read_done() function
In reply to Jiri OlsaMove perf_mmap__read_init() from tools/perf to libperf and export it in the perf/mmap.h header.
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/perf/mmap.h | 1 + tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 17 +++++++++++++++++ tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 17 ----------------- tools/perf/util/mmap.h | 1 - 20 files changed, 34 insertions(+), 33 deletions(-)
tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 -1
@@ -142,7 +142,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybe next_event:perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (!comm1_time || !comm2_time)
tools/perf/builtin-kvm.c +1 -1
@@ -794,7 +794,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,break ; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core);return n; }
tools/perf/builtin-top.c +1 -1
@@ -894,7 +894,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx) } } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }static void perf_top__mmap_read (struct perf_top *top)
tools/perf/builtin-trace.c +1 -1
@@ -3471,7 +3471,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv) draining = true; } } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (trace->nr_events == before) {
tools/perf/lib/include/perf/mmap.h +1 -0
@@ -8,5 +8,6 @@ struct perf_mmap; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map); +LIBPERF_API void perf_mmap__read_done(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/libperf.map +1 -0
@@ -42,6 +42,7 @@ LIBPERF_0.0.1 { perf_evlist__poll; perf_mmap__consume; perf_mmap__read_init; + perf_mmap__read_done; local: *; };
tools/perf/lib/mmap.c +17 -0
@@ -174,3 +174,20 @@ int perf_mmap__read_init(struct perf_mmap *map)return __perf_mmap__read_init (map); } + +/* +* Mandatory for overwrite mode +* The direction of overwrite mode is backward. +* The last perf_mmap__read() will set tail to map->core.prev. +* Need to correct the map->core.prev to head which is the end of next read. +*/ +void perf_mmap__read_done (struct perf_mmap *map) +{ +/* +* Check if event was unmapped due to a POLLHUP/POLLERR. +*/ +if (!refcount_read (&map->refcnt)) +return ; + + map->prev =perf_mmap__read_head (map); +}
tools/perf/tests/backward-ring-buffer.c +1 -1
@@ -54,7 +54,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,return TEST_FAIL ; } } -perf_mmap__read_done (map); +perf_mmap__read_done (&map->core); }return TEST_OK ; }
tools/perf/tests/bpf.c +1 -1
@@ -194,7 +194,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),if (type ==PERF_RECORD_SAMPLE ) count ++; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (count != expect) {
tools/perf/tests/code-reading.c +1 -1
@@ -435,7 +435,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,if (ret <0 )return ret; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }return 0 ; }
tools/perf/tests/keep-tracking.c +1 -1
@@ -49,7 +49,7 @@ static int find_comm(struct evlist *evlist, const char *comm) found +=1 ;perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }return found; }
tools/perf/tests/mmap-basic.c +1 -1
@@ -142,7 +142,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuse nr_events[evsel->idx]++;perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); out_init: err =0 ;
tools/perf/tests/openat-syscall-tp-fields.c +1 -1
@@ -124,7 +124,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtest goto out_ok; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (nr_events == before)
tools/perf/tests/perf-record.c +1 -1
@@ -279,7 +279,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusperf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }/*
tools/perf/tests/sw-clock.c +1 -1
@@ -120,7 +120,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id) next_event:perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); out_init:if ((u64 ) nr_samples == total_periods) {
tools/perf/tests/switch-tracking.c +1 -1
@@ -280,7 +280,7 @@ static int process_events(struct evlist *evlist,if (ret <0 ) goto out_free_nodes; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); } events_array =calloc (cnt,sizeof (struct event_node ));
tools/perf/tests/task-exit.c +1 -1
@@ -127,7 +127,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unusedperf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); out_init:if (!exited || !nr_exit) {
tools/perf/util/evlist.c +1 -1
@@ -1753,7 +1753,7 @@ static void *perf_evlist__poll_thread(void *arg)perf_mmap__consume (&map->core); got_data = true; } -perf_mmap__read_done (map); +perf_mmap__read_done (&map->core); }if (draining && !got_data)
tools/perf/util/mmap.c +0 -17
@@ -405,20 +405,3 @@ int perf_mmap__push(struct mmap *md, void *to, out:return rc; } - -/* -* Mandatory for overwrite mode -* The direction of overwrite mode is backward. -* The last perf_mmap__read() will set tail to map->core.prev. -* Need to correct the map->core.prev to head which is the end of next read. -*/ -void perf_mmap__read_done (struct mmap *map) -{ -/* -* Check if event was unmapped due to a POLLHUP/POLLERR. -*/ -if (!refcount_read (&map->core.refcnt)) -return ; - - map->core.prev =perf_mmap__read_head (&map->core); -}
tools/perf/util/mmap.h +0 -1
@@ -54,5 +54,4 @@ int perf_mmap__push(struct mmap *md, void *to,size_t mmap__mmap_len (struct mmap *map); -void perf_mmap__read_done (struct mmap *map);#endif /*__PERF_MMAP_H */
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 12/36 libperf: Add perf_mmap__read_event() function
In reply to Jiri OlsaMove perf_mmap__read_event() from tools/perf to libperf and export it in the perf/mmap.h header.
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/internal/mmap.h | 1 + tools/perf/lib/include/perf/mmap.h | 1 + tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 79 ++++++++++++++++++++ tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 77 ------------------- tools/perf/util/mmap.h | 2 - tools/perf/util/python.c | 2 +- 22 files changed, 98 insertions(+), 95 deletions(-)
tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 -1
@@ -121,7 +121,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybeif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample; if (event->header.type != PERF_RECORD_COMM ||
tools/perf/builtin-kvm.c +1 -1
@@ -764,7 +764,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,if (err <0 )return (err == -EAGAIN ) ?0 :-1 ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { err =perf_evlist__parse_sample_timestamp (evlist, event, ×tamp);if (err) {perf_mmap__consume (&md->core);
tools/perf/builtin-top.c +1 -1
@@ -873,7 +873,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)if (perf_mmap__read_init (&md->core) <0 )return ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {int ret; ret =perf_evlist__parse_sample_timestamp (evlist, event, &last_timestamp);
tools/perf/builtin-trace.c +1 -1
@@ -3454,7 +3454,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { ++trace->nr_events; err =trace__deliver_event (trace, event);
tools/perf/lib/include/internal/mmap.h +1 -0
@@ -11,6 +11,7 @@#define PERF_SAMPLE_MAX_SIZE (1 << 16)struct perf_mmap ; +union perf_event ;typedef void (*libperf_unmap_cb_t )(struct perf_mmap *map);
tools/perf/lib/include/perf/mmap.h +1 -0
@@ -9,5 +9,6 @@ struct perf_mmap; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map); LIBPERF_API void perf_mmap__read_done(struct perf_mmap *map); +LIBPERF_API union perf_event *perf_mmap__read_event(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/libperf.map +1 -0
@@ -43,6 +43,7 @@ LIBPERF_0.0.1 { perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done; + perf_mmap__read_event; local: *; };
tools/perf/lib/mmap.c +79 -0
91 lines not on this page · raw message
tools/perf/tests/backward-ring-buffer.c +1 -1
@@ -39,7 +39,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,union perf_event *event;perf_mmap__read_init (&map->core); -while ((event =perf_mmap__read_event (map)) !=NULL ) { +while ((event =perf_mmap__read_event (&map->core)) !=NULL ) {const u32 type = event->header.type;switch (type) {
tools/perf/tests/bpf.c +1 -1
@@ -188,7 +188,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;if (type ==PERF_RECORD_SAMPLE )
tools/perf/tests/code-reading.c +1 -1
@@ -429,7 +429,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { ret =process_event (machine, evlist, event, state);perf_mmap__consume (&md->core);if (ret <0 )
tools/perf/tests/keep-tracking.c +1 -1
@@ -41,7 +41,7 @@ static int find_comm(struct evlist *evlist, const char *comm) md = &evlist->mmap[i];if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { if (event->header.type == PERF_RECORD_COMM && (pid_t)event->comm.pid == getpid() && (pid_t)event->comm.tid == getpid() &&
tools/perf/tests/mmap-basic.c +1 -1
@@ -117,7 +117,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuseif (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample;if (event->header.type !=PERF_RECORD_SAMPLE ) {
tools/perf/tests/openat-syscall-tp-fields.c +1 -1
@@ -96,7 +96,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtestif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;int tp_flags;struct perf_sample sample;
tools/perf/tests/perf-record.c +1 -1
@@ -174,7 +174,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;const char *name =perf_event__name (type);
tools/perf/tests/sw-clock.c +1 -1
@@ -103,7 +103,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)if (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample;if (event->header.type !=PERF_RECORD_SAMPLE )
tools/perf/tests/switch-tracking.c +1 -1
@@ -273,7 +273,7 @@ static int process_events(struct evlist *evlist,if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { cnt +=1 ; ret =add_event (evlist, &events, event);perf_mmap__consume (&md->core);
tools/perf/tests/task-exit.c +1 -1
@@ -121,7 +121,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unusedif (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {if (event->header.type ==PERF_RECORD_EXIT ) nr_exit++;
tools/perf/util/evlist.c +1 -1
@@ -1742,7 +1742,7 @@ static void *perf_evlist__poll_thread(void *arg)if (perf_mmap__read_init (&map->core))continue ; -while ((event =perf_mmap__read_event (map)) !=NULL ) { +while ((event =perf_mmap__read_event (&map->core)) !=NULL ) {struct evsel *evsel =perf_evlist__event2evsel (evlist, event);if (evsel && evsel->side_band.cb)
tools/perf/util/mmap.c +0 -77
83 lines not on this page · raw message
tools/perf/util/mmap.h +0 -2
@@ -47,8 +47,6 @@ void mmap__munmap(struct mmap *map);union perf_event *perf_mmap__read_forward (struct mmap *map); -union perf_event *perf_mmap__read_event (struct mmap *map); -int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size));
tools/perf/util/python.c +1 -1
@@ -1020,7 +1020,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist,if (perf_mmap__read_init (&md->core) <0 ) goto end; - event =perf_mmap__read_event (md); + event =perf_mmap__read_event (&md->core);if (event !=NULL ) {PyObject *pyevent =pyrf_event__new (event);struct pyrf_event *pevent = (struct pyrf_event *)pyevent;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 13/36 libperf: Add perf_evlist__mmap()/munmap() functions
In reply to Jiri OlsaAdd libperf's version of perf_evlist__mmap()/munmap() functions and exporting them in the perf/evlist.h header.
It's the backbone of what we have in perf code. The following changes will add needed callbacks and then we'll finally switch the perf code to use libperf's version.
Add mmap/mmap_ovw 'struct perf_mmap' object arrays to hold maps for libperf's evlist.
--- tools/perf/lib/evlist.c | 236 +++++++++++++++++++++++ tools/perf/lib/include/internal/evlist.h | 2 + tools/perf/lib/include/perf/evlist.h | 3 + tools/perf/lib/libperf.map | 2 + 4 files changed, 243 insertions(+)
tools/perf/lib/evlist.c +236 -0
@@ -8,13 +8,20 @@#include <internal/evlist.h> #include <internal/evsel.h> #include <internal/xyarray.h> +#include <internal/mmap.h> +#include <internal/cpumap.h> +#include <internal/threadmap.h> +#include <internal/xyarray.h> +#include <internal/lib.h> #include <linux/zalloc.h> +#include <sys/ioctl.h> #include <stdlib.h> #include <errno.h> #include <unistd.h> #include <fcntl.h> #include <signal.h> #include <poll.h> +#include <sys/mman.h> #include <perf/cpumap.h> #include <perf/threadmap.h> #include <api/fd/array.h>
@@ -103,6 +110,10 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)void perf_evlist__delete (struct perf_evlist *evlist) { +if (evlist ==NULL ) +return ; + +perf_evlist__munmap (evlist);free (evlist); }
@@ -281,3 +292,228 @@ int perf_evlist__poll(struct perf_evlist *evlist, int timeout) {return fdarray__poll (&evlist->pollfd, timeout); } + +static struct perf_mmap *perf_evlist__alloc_mmap (struct perf_evlist *evlist,bool overwrite) +{ +int i; +struct perf_mmap *map; + + evlist->nr_mmaps =perf_cpu_map__nr (evlist->cpus); +if (perf_cpu_map__empty (evlist->cpus)) + evlist->nr_mmaps =perf_thread_map__nr (evlist->threads); + + map =zalloc (evlist->nr_mmaps *sizeof (struct perf_mmap )); +if (!map) +return NULL ; + +for (i =0 ; i < evlist->nr_mmaps; i++) { +/* +* When the perf_mmap() call is made we grab one refcount, plus +* one extra to let perf_mmap__consume() get the last +* events after all real references (perf_mmap__get()) are +* dropped. +* +* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and +* thus does perf_mmap__get() on it. +*/ +perf_mmap__init (&map[i], overwrite,NULL ); + } + +return map; +} + +static void perf_evlist__set_sid_idx (struct perf_evlist *evlist, +struct perf_evsel *evsel,int idx,int cpu, +int thread) +{ +struct perf_sample_id *sid =SID (evsel, cpu, thread); + + sid->idx = idx; +if (evlist->cpus && cpu >=0 ) + sid->cpu = evlist->cpus->map[cpu]; +else + sid->cpu =-1 ; +if (!evsel->system_wide && evlist->threads && thread >=0 ) + sid->tid =perf_thread_map__pid (evlist->threads, thread); +else + sid->tid =-1 ; +} + +static struct perf_mmap * +perf_evlist__map_get (struct perf_evlist *evlist,bool overwrite,int idx) +{ +struct perf_mmap *map = &evlist->mmap[idx]; + +if (overwrite) { +if (!evlist->mmap_ovw) { + evlist->mmap_ovw =perf_evlist__alloc_mmap (evlist, true); +if (!evlist->mmap_ovw) +return NULL ; + } + map = &evlist->mmap_ovw[idx]; + } + +return map; +} + +#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y)) + +static int +mmap_per_evsel (struct perf_evlist *evlist,int idx, +struct perf_mmap_param *mp,int cpu_idx, +int thread,int *_output,int *_output_overwrite) +{ +int evlist_cpu =perf_cpu_map__cpu (evlist->cpus, cpu_idx); +struct perf_evsel *evsel; +int revent; + +perf_evlist__for_each_entry (evlist, evsel) { +bool overwrite = evsel->attr.write_backward; +struct perf_mmap *map; +int *output, fd, cpu; + +if (evsel->system_wide && thread) +continue ; + + cpu =perf_cpu_map__idx (evsel->cpus, evlist_cpu); +if (cpu ==-1 ) +continue ; + + map =perf_evlist__map_get (evlist, overwrite, idx); +if (map ==NULL ) +return -ENOMEM ; + +if (overwrite) { + mp->prot =PROT_READ ; + output = _output_overwrite; + }else { + mp->prot =PROT_READ |PROT_WRITE ; + output = _output; + } + + fd =FD (evsel, cpu, thread); + +if (*output ==-1 ) { + *output = fd; + +if (perf_mmap__mmap (map, mp, *output, evlist_cpu) <0 ) +return -1 ; + }else { +if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 ) +return -1 ; + +perf_mmap__get (map); + } + + revent = !overwrite ?POLLIN :0 ; + +if (!evsel->system_wide && +perf_evlist__add_pollfd (evlist, fd, map, revent) <0 ) { +perf_mmap__put (map); +return -1 ; + } + +if (evsel->attr.read_format &PERF_FORMAT_ID ) { +if (perf_evlist__id_add_fd (evlist, evsel, cpu, thread, + fd) <0 ) +return -1 ; +perf_evlist__set_sid_idx (evlist, evsel, idx, cpu, + thread); + } + } + +return 0 ; +} + +static int +mmap_per_thread (struct perf_evlist *evlist,struct perf_mmap_param *mp) +{ +int thread; +int nr_threads =perf_thread_map__nr (evlist->threads); + +for (thread =0 ; thread < nr_threads; thread++) { +int output =-1 ; +int output_overwrite =-1 ; + +if (mmap_per_evsel (evlist, thread, mp,0 , thread, + &output, &output_overwrite)) + goto out_unmap; + } + +return 0 ; + +out_unmap: +perf_evlist__munmap (evlist); +return -1 ; +} + +static int +mmap_per_cpu (struct perf_evlist *evlist,struct perf_mmap_param *mp) +{ +int nr_threads =perf_thread_map__nr (evlist->threads); +int nr_cpus =perf_cpu_map__nr (evlist->cpus); +int cpu, thread; + +for (cpu =0 ; cpu < nr_cpus; cpu++) { +int output =-1 ; +int output_overwrite =-1 ; + +for (thread =0 ; thread < nr_threads; thread++) { +if (mmap_per_evsel (evlist, cpu, mp, cpu, + thread, &output, &output_overwrite)) + goto out_unmap; + } + } + +return 0 ; + +out_unmap: +perf_evlist__munmap (evlist); +return -1 ; +} + +int perf_evlist__mmap (struct perf_evlist *evlist,int pages) +{ +struct perf_evsel *evsel; +const struct perf_cpu_map *cpus = evlist->cpus; +const struct perf_thread_map *threads = evlist->threads; +struct perf_mmap_param mp; + +if (!evlist->mmap) + evlist->mmap =perf_evlist__alloc_mmap (evlist, false); +if (!evlist->mmap) +return -ENOMEM ; + +perf_evlist__for_each_entry (evlist, evsel) { +if ((evsel->attr.read_format &PERF_FORMAT_ID ) && + evsel->sample_id ==NULL && +perf_evsel__alloc_id (evsel,perf_cpu_map__nr (cpus), threads->nr) <0 ) +return -ENOMEM ; + } + + evlist->mmap_len = (pages +1 ) * page_size; + mp.mask = evlist->mmap_len - page_size -1 ; + +if (perf_cpu_map__empty (cpus)) +return mmap_per_thread (evlist, &mp); + +return mmap_per_cpu (evlist, &mp); +} + +void perf_evlist__munmap (struct perf_evlist *evlist) +{ +int i; + +if (evlist->mmap) { +for (i =0 ; i < evlist->nr_mmaps; i++) +perf_mmap__munmap (&evlist->mmap[i]); + } + +if (evlist->mmap_ovw) { +for (i =0 ; i < evlist->nr_mmaps; i++) +perf_mmap__munmap (&evlist->mmap_ovw[i]); + } + +zfree (&evlist->mmap); +zfree (&evlist->mmap_ovw); +}
tools/perf/lib/include/internal/evlist.h +2 -0
@@ -22,6 +22,8 @@ struct perf_evlist {size_t mmap_len;struct fdarray pollfd;struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ]; +struct perf_mmap *mmap; +struct perf_mmap *mmap_ovw; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
tools/perf/lib/include/perf/evlist.h +3 -0
@@ -33,4 +33,7 @@ LIBPERF_API void perf_evlist__set_maps(struct perf_evlist *evlist, struct perf_thread_map *threads); LIBPERF_API int perf_evlist__poll(struct perf_evlist *evlist, int timeout); +LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); +LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist); +#endif /* __LIBPERF_EVLIST_H */
tools/perf/lib/libperf.map +2 -0
@@ -40,6 +40,8 @@ LIBPERF_0.0.1 { perf_evlist__next; perf_evlist__set_maps; perf_evlist__poll; + perf_evlist__mmap; + perf_evlist__munmap; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 14/36 libperf: Add perf_evlist__mmap_ops function
In reply to Jiri OlsaTo be able to pass specific callbacks to evlist's mmap. There will be specific call to this function from perf's evlist__mmap() and libperf's perf_evlist__mmap() functions in following changes.
--- tools/perf/lib/evlist.c | 24 ++++++++++++++++++------ tools/perf/lib/include/internal/evlist.h | 8 ++++++++ 2 files changed, 26 insertions(+), 6 deletions(-)
tools/perf/lib/evlist.c +18 -6
@@ -472,12 +472,16 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_mmap_param *mp)return -1 ; } -int perf_evlist__mmap (struct perf_evlist *evlist,int pages) +int perf_evlist__mmap_ops (struct perf_evlist *evlist, +struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp) {struct perf_evsel *evsel;const struct perf_cpu_map *cpus = evlist->cpus;const struct perf_thread_map *threads = evlist->threads; -struct perf_mmap_param mp; + +if (!ops) +return -EINVAL ;if (!evlist->mmap) evlist->mmap =perf_evlist__alloc_mmap (evlist, false);
@@ -491,13 +495,21 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages)return -ENOMEM ; } +if (perf_cpu_map__empty (cpus)) +return mmap_per_thread (evlist, mp); + +return mmap_per_cpu (evlist, mp); +} + +int perf_evlist__mmap (struct perf_evlist *evlist,int pages) +{ +struct perf_mmap_param mp; +struct perf_evlist_mmap_ops ops; + evlist->mmap_len = (pages +1 ) * page_size; mp.mask = evlist->mmap_len - page_size -1 ; -if (perf_cpu_map__empty (cpus)) -return mmap_per_thread (evlist, &mp); - -return mmap_per_cpu (evlist, &mp); +return perf_evlist__mmap_ops (evlist, &ops, &mp); }void perf_evlist__munmap (struct perf_evlist *evlist)
tools/perf/lib/include/internal/evlist.h +8 -0
@@ -11,6 +11,7 @@struct perf_cpu_map ;struct perf_thread_map ; +struct perf_mmap_param ;struct perf_evlist {struct list_head entries;
@@ -26,10 +27,17 @@ struct perf_evlist {struct perf_mmap *mmap_ovw; }; +struct perf_evlist_mmap_ops { +}; +int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);int perf_evlist__add_pollfd (struct perf_evlist *evlist,int fd,void *ptr,short revent); +int perf_evlist__mmap_ops (struct perf_evlist *evlist, +struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp); +/** * __perf_evlist__for_each_entry - iterate thru all the evsels * @list: list_head instance to iterate
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 16/36 libperf: Add perf_evlist_mmap_ops::get callback
In reply to Jiri OlsaAdd the perf_evlist_mmap_ops::get callback to be called in mmap_per_evsel to get/allocate the 'struct perf_mmap' object.
Add the libperf's perf_evlist__mmap_cb_get() function as libperf's get callback.
--- tools/perf/lib/evlist.c | 18 ++++++++++-------- tools/perf/lib/include/internal/evlist.h | 3 +++ 2 files changed, 13 insertions(+), 8 deletions(-)
tools/perf/lib/evlist.c +10 -8
@@ -340,7 +340,7 @@ static void perf_evlist__set_sid_idx(struct perf_evlist *evlist, }static struct perf_mmap * -perf_evlist__map_get (struct perf_evlist *evlist,bool overwrite,int idx) +perf_evlist__mmap_cb_get (struct perf_evlist *evlist,bool overwrite,int idx) {struct perf_mmap *map = &evlist->mmap[idx];
@@ -359,8 +359,8 @@ perf_evlist__map_get(struct perf_evlist *evlist, bool overwrite, int idx)#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y))static int -mmap_per_evsel (struct perf_evlist *evlist,int idx, -struct perf_mmap_param *mp,int cpu_idx, +mmap_per_evsel (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, +int idx,struct perf_mmap_param *mp,int cpu_idx,int thread,int *_output,int *_output_overwrite) {int evlist_cpu =perf_cpu_map__cpu (evlist->cpus, cpu_idx);
@@ -379,7 +379,7 @@ mmap_per_evsel(struct perf_evlist *evlist, int idx,if (cpu ==-1 )continue ; - map =perf_evlist__map_get (evlist, overwrite, idx); + map = ops->get (evlist, overwrite, idx);if (map ==NULL )return -ENOMEM ;
@@ -439,7 +439,7 @@ mmap_per_thread(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (ops->idx) ops->idx (evlist, mp, thread, false); -if (mmap_per_evsel (evlist, thread, mp,0 , thread, +if (mmap_per_evsel (evlist, ops, thread, mp,0 , thread, &output, &output_overwrite)) goto out_unmap; }
@@ -467,7 +467,7 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops, ops->idx (evlist, mp, cpu, true);for (thread =0 ; thread < nr_threads; thread++) { -if (mmap_per_evsel (evlist, cpu, mp, cpu, +if (mmap_per_evsel (evlist, ops, cpu, mp, cpu, thread, &output, &output_overwrite)) goto out_unmap; }
@@ -488,7 +488,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,const struct perf_cpu_map *cpus = evlist->cpus;const struct perf_thread_map *threads = evlist->threads; -if (!ops) +if (!ops || !ops->get)return -EINVAL ;if (!evlist->mmap)
@@ -512,7 +512,9 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,int perf_evlist__mmap (struct perf_evlist *evlist,int pages) {struct perf_mmap_param mp; -struct perf_evlist_mmap_ops ops = {0 }; +struct perf_evlist_mmap_ops ops = { + .get = perf_evlist__mmap_cb_get, + }; evlist->mmap_len = (pages +1 ) * page_size; mp.mask = evlist->mmap_len - page_size -1 ;
tools/perf/lib/include/internal/evlist.h +3 -0
@@ -29,9 +29,12 @@ struct perf_evlist {typedef void (*perf_evlist_mmap__cb_idx_t )(struct perf_evlist *,struct perf_mmap_param *,int ,bool ); +typedef struct perf_mmap * +(*perf_evlist_mmap__cb_get_t )(struct perf_evlist *,bool ,int );struct perf_evlist_mmap_ops {perf_evlist_mmap__cb_idx_t idx; +perf_evlist_mmap__cb_get_t get; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 17/36 libperf: Add perf_evlist_mmap_ops::mmap callback
In reply to Jiri OlsaAdding perf_evlist_mmap_ops::mmap callback to be called in mmap_per_evsel to actually mmap the map.
Adding libperf's perf_evlist__mmap_cb_mmap function as libperf's mmap callback.
New mmaped map gets refcount set to 2 in mmap__mmap, we follow that in mmap callback. We will move this to common place after we switch to perf_evlist__mmap.
--- tools/perf/lib/evlist.c | 29 +++++++++++++++++++++--- tools/perf/lib/include/internal/evlist.h | 3 +++ 2 files changed, 29 insertions(+), 3 deletions(-)
tools/perf/lib/evlist.c +26 -3
@@ -358,6 +358,28 @@ perf_evlist__mmap_cb_get(struct perf_evlist *evlist, bool overwrite, int idx)#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y)) +static int +perf_evlist__mmap_cb_mmap (struct perf_mmap *map,struct perf_mmap_param *mp, +int output,int cpu) +{ +/* +* The last one will be done at perf_mmap__consume(), so that we +* make sure we don't prevent tools from consuming every last event in +* the ring buffer. +* +* I.e. we can get the POLLHUP meaning that the fd doesn't exist +* anymore, but the last events for it are still in the ring buffer, +* waiting to be consumed. +* +* Tools can chose to ignore this at their own discretion, but the +* evlist layer can't just drop it when filtering events in +* perf_evlist__filter_pollfd(). +*/ +refcount_set (&map->refcnt,2 ); + +return perf_mmap__mmap (map, mp, output, cpu); +} + static int mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops, int idx, struct perf_mmap_param *mp, int cpu_idx,
@@ -396,7 +418,7 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (*output ==-1 ) { *output = fd; -if (perf_mmap__mmap (map, mp, *output, evlist_cpu) <0 ) +if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )
@@ -488,7 +510,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,const struct perf_cpu_map *cpus = evlist->cpus;const struct perf_thread_map *threads = evlist->threads; -if (!ops || !ops->get) +if (!ops || !ops->get || !ops->mmap)return -EINVAL ;if (!evlist->mmap)
@@ -513,7 +535,8 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages) {struct perf_mmap_param mp;struct perf_evlist_mmap_ops ops = { - .get = perf_evlist__mmap_cb_get, + .get = perf_evlist__mmap_cb_get, + .mmap = perf_evlist__mmap_cb_mmap, }; evlist->mmap_len = (pages +1 ) * page_size;
tools/perf/lib/include/internal/evlist.h +3 -0
@@ -31,10 +31,13 @@ typedef void (*perf_evlist_mmap__cb_idx_t)(struct perf_evlist*, struct perf_mmap_param*, int, bool);typedef struct perf_mmap * (*perf_evlist_mmap__cb_get_t )(struct perf_evlist *,bool ,int ); +typedef int +(*perf_evlist_mmap__cb_mmap_t )(struct perf_mmap *,struct perf_mmap_param *,int ,int );struct perf_evlist_mmap_ops {perf_evlist_mmap__cb_idx_t idx;perf_evlist_mmap__cb_get_t get; +perf_evlist_mmap__cb_mmap_t mmap; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 19/36 perf tools: Add perf_evlist__mmap_cb_get function
In reply to Jiri OlsaAdding perf_evlist__mmap_cb_get function to return 'struct perf_mmap' object during perf_evlist__mmap_ops call.
The array of 'struct mmap' is allocated via evlist__alloc_mmap, in this callback we simply returns pointer to the base object.
--- tools/perf/util/evlist.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+)
tools/perf/util/evlist.c +24 -0
@@ -726,6 +726,29 @@ perf_evlist__mmap_cb_idx(struct perf_evlist *_evlist,auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, idx, per_cpu); } +static struct perf_mmap * +perf_evlist__mmap_cb_get (struct perf_evlist *_evlist,bool overwrite,int idx) +{ + struct evlist *evlist = container_of(_evlist, struct evlist, core); +struct mmap *maps = evlist->mmap; + +if (overwrite) { + maps = evlist->overwrite_mmap; + +if (!maps) { + maps =evlist__alloc_mmap (evlist, true); +if (!maps) +return NULL ; + + evlist->overwrite_mmap = maps; +if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY ) +perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); + } + } + +return &maps[idx].core; +} +static int evlist__mmap_per_cpu (struct evlist *evlist,struct mmap_params *mp) {
@@ -924,6 +947,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, };struct perf_evlist_mmap_ops ops__maybe_unused = { .idx = perf_evlist__mmap_cb_idx, + .get = perf_evlist__mmap_cb_get, };if (!evlist->mmap)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 20/36 perf tools: Add perf_evlist__mmap_cb_mmap function
In reply to Jiri OlsaAdding perf_evlist__mmap_cb_mmap function to call perf specific mmap__mmap function during perf_evlist__mmap_ops call.
--- tools/perf/util/evlist.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)
tools/perf/util/evlist.c +13 -2
@@ -749,6 +749,16 @@ perf_evlist__mmap_cb_get(struct perf_evlist *_evlist, bool overwrite, int idx)return &maps[idx].core; } +static int +perf_evlist__mmap_cb_mmap (struct perf_mmap *_map,struct perf_mmap_param *_mp, +int output,int cpu) +{ + struct mmap *map = container_of(_map, struct mmap, core); + struct mmap_params *mp = container_of(_mp, struct mmap_params, core); + +return mmap__mmap (map, mp, output, cpu); +} +static int evlist__mmap_per_cpu (struct evlist *evlist,struct mmap_params *mp) {
@@ -946,8 +956,9 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .comp_level = comp_level };struct perf_evlist_mmap_ops ops__maybe_unused = { - .idx = perf_evlist__mmap_cb_idx, - .get = perf_evlist__mmap_cb_get, + .idx = perf_evlist__mmap_cb_idx, + .get = perf_evlist__mmap_cb_get, + .mmap = perf_evlist__mmap_cb_mmap, };if (!evlist->mmap)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 21/36 perf tools: Switch to libperf mmap interface
In reply to Jiri OlsaSwitching to libperf mmap interface by calling directly perf_evlist__mmap_ops and removing perf's evlist__mmap_per_* functions.
By switching libperf perf_evlist__mmap we need to operate over 'struct perf_mmap' in evlist__add_pollfd, so making related changes there.
--- tools/perf/util/evlist.c | 179 +-------------------------------------- 1 file changed, 4 insertions(+), 175 deletions(-)
tools/perf/util/evlist.c +4 -175
@@ -407,10 +407,10 @@ int evlist__add_pollfd(struct evlist *evlist, int fd)static void perf_evlist__munmap_filtered (struct fdarray *fda,int fd,void *arg__maybe_unused ) { -struct mmap *map = fda->priv[fd].ptr; +struct perf_mmap *map = fda->priv[fd].ptr;if (map) -perf_mmap__put (&map->core); +perf_mmap__put (map); }int evlist__filter_pollfd (struct evlist *evlist,short revents_and_mask)
@@ -424,22 +424,6 @@ int evlist__poll(struct evlist *evlist, int timeout)return perf_evlist__poll (&evlist->core, timeout); } -static void perf_evlist__set_sid_idx (struct evlist *evlist, -struct evsel *evsel,int idx,int cpu, -int thread) -{ -struct perf_sample_id *sid =SID (evsel, cpu, thread); - sid->idx = idx; -if (evlist->core.cpus && cpu >=0 ) - sid->cpu = evlist->core.cpus->map[cpu]; -else - sid->cpu =-1 ; -if (!evsel->core.system_wide && evlist->core.threads && thread >=0 ) - sid->tid =perf_thread_map__pid (evlist->core.threads, thread); -else - sid->tid =-1 ; -} -struct perf_sample_id *perf_evlist__id2sid (struct evlist *evlist,u64 id) {struct hlist_head *head;
@@ -628,93 +612,6 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return map; } -static bool -perf_evlist__should_poll (struct evlist *evlist__maybe_unused , -struct evsel *evsel) -{ -if (evsel->core.attr.write_backward) -return false; -return true; -} - -static int evlist__mmap_per_evsel (struct evlist *evlist,int idx, -struct mmap_params *mp,int cpu_idx, -int thread,int *_output,int *_output_overwrite) -{ -struct evsel *evsel; -int revent; -int evlist_cpu =cpu_map__cpu (evlist->core.cpus, cpu_idx); - -evlist__for_each_entry (evlist, evsel) { -struct mmap *maps = evlist->mmap; -int *output = _output; -int fd; -int cpu; - - mp->core.prot =PROT_READ |PROT_WRITE ; -if (evsel->core.attr.write_backward) { - output = _output_overwrite; - maps = evlist->overwrite_mmap; - -if (!maps) { - maps =evlist__alloc_mmap (evlist, true); -if (!maps) -return -1 ; - evlist->overwrite_mmap = maps; -if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY ) -perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); - } - mp->core.prot &= ~PROT_WRITE ; - } - -if (evsel->core.system_wide && thread) -continue ; - - cpu =perf_cpu_map__idx (evsel->core.cpus, evlist_cpu); -if (cpu ==-1 ) -continue ; - - fd =FD (evsel, cpu, thread); - -if (*output ==-1 ) { - *output = fd; - -if (mmap__mmap (&maps[idx], mp, *output, evlist_cpu) <0 ) -return -1 ; - }else { -if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 ) -return -1 ; - -perf_mmap__get (&maps[idx].core); - } - - revent =perf_evlist__should_poll (evlist, evsel) ?POLLIN :0 ; - -/* -* The system_wide flag causes a selected event to be opened -* always without a pid. Consequently it will never get a -* POLLHUP, but it is used for tracking in combination with -* other events, so it should not need to be polled anyway. -* Therefore don't add it for polling. -*/ -if (!evsel->core.system_wide && -perf_evlist__add_pollfd (&evlist->core, fd, &maps[idx], revent) <0 ) { -perf_mmap__put (&maps[idx].core); -return -1 ; - } - -if (evsel->core.attr.read_format &PERF_FORMAT_ID ) { -if (perf_evlist__id_add_fd (&evlist->core, &evsel->core, cpu, thread, - fd) <0 ) -return -1 ; -perf_evlist__set_sid_idx (evlist, evsel, idx, cpu, - thread); - } - } - -return 0 ; -} - static void perf_evlist__mmap_cb_idx(struct perf_evlist *_evlist, struct perf_mmap_param *_mp,
@@ -759,61 +656,6 @@ perf_evlist__mmap_cb_mmap(struct perf_mmap *_map, struct perf_mmap_param *_mp,return mmap__mmap (map, mp, output, cpu); } -static int evlist__mmap_per_cpu (struct evlist *evlist, -struct mmap_params *mp) -{ -int cpu, thread; -int nr_cpus =perf_cpu_map__nr (evlist->core.cpus); -int nr_threads =perf_thread_map__nr (evlist->core.threads); - -pr_debug2 ("perf event ring buffer mmapped per cpu\n" ); -for (cpu =0 ; cpu < nr_cpus; cpu++) { -int output =-1 ; -int output_overwrite =-1 ; - -auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, cpu, - true); - -for (thread =0 ; thread < nr_threads; thread++) { -if (evlist__mmap_per_evsel (evlist, cpu, mp, cpu, - thread, &output, &output_overwrite)) - goto out_unmap; - } - } - -return 0 ; - -out_unmap: -evlist__munmap_nofree (evlist); -return -1 ; -} - -static int evlist__mmap_per_thread (struct evlist *evlist, -struct mmap_params *mp) -{ -int thread; -int nr_threads =perf_thread_map__nr (evlist->core.threads); - -pr_debug2 ("perf event ring buffer mmapped per thread\n" ); -for (thread =0 ; thread < nr_threads; thread++) { -int output =-1 ; -int output_overwrite =-1 ; - -auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, thread, - false); - -if (evlist__mmap_per_evsel (evlist, thread, mp,0 , thread, - &output, &output_overwrite)) - goto out_unmap; - } - -return 0 ; - -out_unmap: -evlist__munmap_nofree (evlist); -return -1 ; -} -unsigned long perf_event_mlock_kb_in_pages (void ) {unsigned long pages;
@@ -941,9 +783,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, bool auxtrace_overwrite, int nr_cblocks, int affinity, int flush, int comp_level) { - struct evsel *evsel; -const struct perf_cpu_map *cpus = evlist->core.cpus; -const struct perf_thread_map *threads = evlist->core.threads;/* * Delay setting mp.prot: set it before calling perf_mmap__mmap. * Its value is decided by evsel's write_backward.
@@ -955,7 +794,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .flush = flush, .comp_level = comp_level }; -struct perf_evlist_mmap_ops ops__maybe_unused = { +struct perf_evlist_mmap_ops ops = { .idx = perf_evlist__mmap_cb_idx, .get = perf_evlist__mmap_cb_get, .mmap = perf_evlist__mmap_cb_mmap,
@@ -976,17 +815,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages,auxtrace_mmap_params__init (&mp.auxtrace_mp, evlist->core.mmap_len, auxtrace_pages, auxtrace_overwrite); -evlist__for_each_entry (evlist, evsel) { -if ((evsel->core.attr.read_format &PERF_FORMAT_ID ) && - evsel->core.sample_id ==NULL && -perf_evsel__alloc_id (&evsel->core,perf_cpu_map__nr (cpus), threads->nr) <0 ) -return -ENOMEM ; - } - -if (perf_cpu_map__empty (cpus)) -return evlist__mmap_per_thread (evlist, &mp); - -return evlist__mmap_per_cpu (evlist, &mp); +return perf_evlist__mmap_ops (&evlist->core, &ops, &mp.core); }int evlist__mmap (struct evlist *evlist,unsigned int pages)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 22/36 libperf: Centralize map refcnt setting
In reply to Jiri OlsaCurrently when new map is mmaps we set its refcnt
to 2 in perf_evlist_mmap_ops::mmap callback.
Every mmap is given refcnt set to 2 when it's first mmaped:
- 1 for current user, which will be taken out by call to perf_evlist__munmap_filtered, where we find out there's no more data comming from kernel to this mmap
- 1 for drain code where in perf_mmap__consume the mmap is released if it is empty
Moving this common setup into libperf generic code before mmap callback is called.
--- tools/perf/lib/evlist.c | 30 +++++++++++++++--------------- tools/perf/util/mmap.c | 15 --------------- 2 files changed, 15 insertions(+), 30 deletions(-)
tools/perf/lib/evlist.c +15 -15
@@ -362,21 +362,6 @@ static int perf_evlist__mmap_cb_mmap(struct perf_mmap *map, struct perf_mmap_param *mp, int output, int cpu) { -/* -* The last one will be done at perf_mmap__consume(), so that we -* make sure we don't prevent tools from consuming every last event in -* the ring buffer. -* -* I.e. we can get the POLLHUP meaning that the fd doesn't exist -* anymore, but the last events for it are still in the ring buffer, -* waiting to be consumed. -* -* Tools can chose to ignore this at their own discretion, but the -* evlist layer can't just drop it when filtering events in -* perf_evlist__filter_pollfd(). -*/ - refcount_set(&map->refcnt, 2); - return perf_mmap__mmap(map, mp, output, cpu); }
@@ -418,6 +403,21 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (*output ==-1 ) { *output = fd; +/* +* The last one will be done at perf_mmap__consume(), so that we +* make sure we don't prevent tools from consuming every last event in +* the ring buffer. +* +* I.e. we can get the POLLHUP meaning that the fd doesn't exist +* anymore, but the last events for it are still in the ring buffer, +* waiting to be consumed. +* +* Tools can chose to ignore this at their own discretion, but the +* evlist layer can't just drop it when filtering events in +* perf_evlist__filter_pollfd(). +*/ +refcount_set (&map->refcnt,2 ); +if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; }else {
tools/perf/util/mmap.c +0 -15
@@ -243,21 +243,6 @@ static void perf_mmap__setup_affinity_mask(struct mmap *map, struct mmap_paramsint mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu) { -/* -* The last one will be done at perf_mmap__consume(), so that we -* make sure we don't prevent tools from consuming every last event in -* the ring buffer. -* -* I.e. we can get the POLLHUP meaning that the fd doesn't exist -* anymore, but the last events for it are still in the ring buffer, -* waiting to be consumed. -* -* Tools can chose to ignore this at their own discretion, but the -* evlist layer can't just drop it when filtering events in -* perf_evlist__filter_pollfd(). -*/ -refcount_set (&map->core.refcnt,2 ); -if (perf_mmap__mmap (&map->core, &mp->core, fd, cpu)) {pr_debug2 ("failed to mmap perf event ring buffer, error %d\n" , errno);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 23/36 libperf: Move pollfd allocation to libperf
In reply to Jiri OlsaIt's needed in libperf only, moving it toperf_evlist__mmap_ops function.
--- tools/perf/lib/evlist.c | 5 +++++ tools/perf/util/evlist.c | 4 ---- 2 files changed, 5 insertions(+), 4 deletions(-)
tools/perf/lib/evlist.c +5 -0
@@ -34,6 +34,7 @@ void perf_evlist__init(struct perf_evlist *evlist)INIT_HLIST_HEAD (&evlist->heads[i]);INIT_LIST_HEAD (&evlist->entries); evlist->nr_entries =0 ; +fdarray__init (&evlist->pollfd,64 ); } static void __perf_evlist__propagate_maps(struct perf_evlist *evlist,
@@ -114,6 +115,7 @@ void perf_evlist__delete(struct perf_evlist *evlist)return ;perf_evlist__munmap (evlist); +fdarray__exit (&evlist->pollfd);free (evlist); }
@@ -525,6 +527,9 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,return -ENOMEM ; } +if (evlist->pollfd.entries ==NULL &&perf_evlist__alloc_pollfd (evlist) <0 ) +return -ENOMEM ; +if (perf_cpu_map__empty (cpus))return mmap_per_thread (evlist, ops, mp);
tools/perf/util/evlist.c +0 -4
@@ -58,7 +58,6 @@ void evlist__init(struct evlist *evlist, struct perf_cpu_map *cpus, {perf_evlist__init (&evlist->core);perf_evlist__set_maps (&evlist->core, cpus, threads); -fdarray__init (&evlist->core.pollfd,64 ); evlist->workload.pid =-1 ; evlist->bkw_mmap_state =BKW_MMAP_NOTREADY ; }
@@ -805,9 +804,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages,if (!evlist->mmap)return -ENOMEM ; -if (evlist->core.pollfd.entries ==NULL &&perf_evlist__alloc_pollfd (&evlist->core) <0 ) -return -ENOMEM ; - evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); mp.core.mask = evlist->core.mmap_len - page_size -1 ;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 25/36 libperf: Add perf_evlist__purge function
In reply to Jiri OlsaAdding static perf_evlist__purge function to purge and delete evsels from evlist.
Adding also perf_evlist__for_each_entry_safe which is used by perf_evlist__purge.
--- tools/perf/lib/evlist.c | 13 +++++++++++++ tools/perf/lib/include/internal/evlist.h | 18 ++++++++++++++++++ 2 files changed, 31 insertions(+)
tools/perf/lib/evlist.c +13 -0
@@ -109,6 +109,18 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)return next; } +static void perf_evlist__purge (struct perf_evlist *evlist) +{ +struct perf_evsel *pos, *n; + +perf_evlist__for_each_entry_safe (evlist, n, pos) { +list_del_init (&pos->node); +perf_evsel__delete (pos); + } + + evlist->nr_entries =0 ; +} +void perf_evlist__exit (struct perf_evlist *evlist) {perf_cpu_map__put (evlist->cpus);
@@ -125,6 +137,7 @@ void perf_evlist__delete(struct perf_evlist *evlist)perf_evlist__munmap (evlist);perf_evlist__close (evlist); +perf_evlist__purge (evlist);perf_evlist__exit (evlist);free (evlist); }
tools/perf/lib/include/internal/evlist.h +18 -0
@@ -82,6 +82,24 @@ void perf_evlist__exit(struct perf_evlist *evlist);#define perf_evlist__for_each_entry_reverse (evlist, evsel) \ __perf_evlist__for_each_entry_reverse(&(evlist)->entries, evsel) +/** +* __perf_evlist__for_each_entry_safe - safely iterate thru all the evsels +* @list: list_head instance to iterate +* @tmp: struct evsel temp iterator +* @evsel: struct evsel iterator +*/ +#define __perf_evlist__for_each_entry_safe (list, tmp, evsel) \ + list_for_each_entry_safe(evsel, tmp, list, node) + +/** +* perf_evlist__for_each_entry_safe - safely iterate thru all the evsels +* @evlist: evlist instance to iterate +* @evsel: struct evsel iterator +* @tmp: struct evsel temp iterator +*/ +#define perf_evlist__for_each_entry_safe (evlist, tmp, evsel) \ + __perf_evlist__for_each_entry_safe(&(evlist)->entries, tmp, evsel) +static inline struct perf_evsel *perf_evlist__first (struct perf_evlist *evlist) { return list_entry(evlist->entries.next, struct perf_evsel, node);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 26/36 libperf: Add perf_evlist__filter_pollfd function
In reply to Jiri OlsaAdding perf_evlist__filter_pollfd function and exporting it in perf/evlist.h header, so libperf users can check if descriptor is still alive.
--- tools/perf/lib/evlist.c | 15 +++++++++++++++ tools/perf/lib/include/perf/evlist.h | 2 ++ tools/perf/lib/libperf.map | 1 + tools/perf/util/evlist.c | 12 +----------- 4 files changed, 19 insertions(+), 11 deletions(-)
tools/perf/lib/evlist.c +15 -0
@@ -313,6 +313,21 @@ int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd,return pos; } +static void perf_evlist__munmap_filtered (struct fdarray *fda,int fd, +void *arg__maybe_unused ) +{ +struct perf_mmap *map = fda->priv[fd].ptr; + +if (map) +perf_mmap__put (map); +} + +int perf_evlist__filter_pollfd (struct perf_evlist *evlist,short revents_and_mask) +{ +return fdarray__filter (&evlist->pollfd, revents_and_mask, + perf_evlist__munmap_filtered,NULL ); +} +int perf_evlist__poll (struct perf_evlist *evlist,int timeout) {return fdarray__poll (&evlist->pollfd, timeout);
tools/perf/lib/include/perf/evlist.h +2 -0
@@ -32,6 +32,8 @@ LIBPERF_API void perf_evlist__set_maps(struct perf_evlist *evlist,struct perf_cpu_map *cpus, struct perf_thread_map *threads); LIBPERF_API int perf_evlist__poll(struct perf_evlist *evlist, int timeout); +LIBPERF_API int perf_evlist__filter_pollfd(struct perf_evlist *evlist, +short revents_and_mask); LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist);
tools/perf/lib/libperf.map +1 -0
@@ -42,6 +42,7 @@ LIBPERF_0.0.1 { perf_evlist__poll; perf_evlist__mmap; perf_evlist__munmap; + perf_evlist__filter_pollfd; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;
tools/perf/util/evlist.c +1 -11
@@ -399,19 +399,9 @@ int evlist__add_pollfd(struct evlist *evlist, int fd)return perf_evlist__add_pollfd (&evlist->core, fd,NULL ,POLLIN ); } -static void perf_evlist__munmap_filtered (struct fdarray *fda,int fd, -void *arg__maybe_unused ) -{ -struct perf_mmap *map = fda->priv[fd].ptr; - -if (map) -perf_mmap__put (map); -} -int evlist__filter_pollfd (struct evlist *evlist,short revents_and_mask) { -return fdarray__filter (&evlist->core.pollfd, revents_and_mask, - perf_evlist__munmap_filtered,NULL ); +return perf_evlist__filter_pollfd (&evlist->core, revents_and_mask); }int evlist__poll (struct evlist *evlist,int timeout)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 27/36 libperf: Add perf_evlist__for_each_mmap function
In reply to Jiri OlsaAdding perf_evlist__for_each_mmap function and exporting it in perf/evlist.h header, so user can iterate through 'struct perf_mmap' objects.
Adding internal perf_mmap__link function to do the actual linking.
--- tools/perf/lib/evlist.c | 26 +++++++++++++++++++++++- tools/perf/lib/include/internal/evlist.h | 2 ++ tools/perf/lib/include/internal/mmap.h | 5 +++-- tools/perf/lib/include/perf/evlist.h | 9 ++++++++ tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 6 ++++-- tools/perf/util/evlist.c | 4 +++- 7 files changed, 47 insertions(+), 6 deletions(-)
tools/perf/lib/evlist.c +25 -1
@@ -347,6 +347,8 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, booreturn NULL ;for (i =0 ; i < evlist->nr_mmaps; i++) { +struct perf_mmap *prev = i ? &map[i -1 ] :NULL ; +/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last
@@ -356,7 +358,7 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, boo* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i], overwrite,NULL ); +perf_mmap__init (&map[i], prev, overwrite,NULL ); }return map;
@@ -405,6 +407,15 @@ perf_evlist__mmap_cb_mmap(struct perf_mmap *map, struct perf_mmap_param *mp,return perf_mmap__mmap (map, mp, output, cpu); } +static void perf_evlist__set_mmap0 (struct perf_evlist *evlist,struct perf_mmap *map, +bool overwrite) +{ +if (overwrite) + evlist->mmap_ovw0 = map; +else + evlist->mmap0 = map; +} + static int mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops, int idx, struct perf_mmap_param *mp, int cpu_idx,
@@ -460,6 +471,9 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; + +if (!idx) +perf_evlist__set_mmap0 (evlist, map, overwrite); }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )return -1 ;
@@ -605,3 +619,13 @@ void perf_evlist__munmap(struct perf_evlist *evlist)zfree (&evlist->mmap);zfree (&evlist->mmap_ovw); } + +struct perf_mmap * +perf_evlist__next_mmap (struct perf_evlist *evlist,struct perf_mmap *map, +bool overwrite) +{ +if (map) +return map->next; + +return overwrite ? evlist->mmap_ovw0 : evlist->mmap0; +}
tools/perf/lib/include/internal/evlist.h +2 -0
@@ -25,6 +25,8 @@ struct perf_evlist {struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ];struct perf_mmap *mmap;struct perf_mmap *mmap_ovw; +struct perf_mmap *mmap0; +struct perf_mmap *mmap_ovw0; }; typedef void
tools/perf/lib/include/internal/mmap.h +3 -2
@@ -33,6 +33,7 @@ struct perf_mmap {u64 flush;libperf_unmap_cb_t unmap_cb;char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; +struct perf_mmap *next; };struct perf_mmap_param {
@@ -42,8 +43,8 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map); -void perf_mmap__init (struct perf_mmap *map,bool overwrite, -libperf_unmap_cb_t unmap_cb); +void perf_mmap__init (struct perf_mmap *map,struct perf_mmap *prev, +bool overwrite,libperf_unmap_cb_t unmap_cb);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu);void perf_mmap__munmap (struct perf_mmap *map);
tools/perf/lib/include/perf/evlist.h +9 -0
@@ -3,6 +3,7 @@#define __LIBPERF_EVLIST_H#include <perf/core.h> +#include <stdbool.h> struct perf_evlist ;struct perf_evsel ;
@@ -38,4 +39,12 @@ LIBPERF_API int perf_evlist__filter_pollfd(struct perf_evlist *evlist, LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist); +LIBPERF_API struct perf_mmap *perf_evlist__next_mmap(struct perf_evlist *evlist, +struct perf_mmap *map, +bool overwrite); +#define perf_evlist__for_each_mmap (evlist, pos, overwrite) \ + for ((pos) = perf_evlist__next_mmap((evlist), NULL, overwrite); \ + (pos) != NULL; \ + (pos) = perf_evlist__next_mmap((evlist), (pos), overwrite)) +#endif /* __LIBPERF_EVLIST_H */
tools/perf/lib/libperf.map +1 -0
@@ -43,6 +43,7 @@ LIBPERF_0.0.1 { perf_evlist__mmap; perf_evlist__munmap; perf_evlist__filter_pollfd; + perf_evlist__next_mmap; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;
tools/perf/lib/mmap.c +4 -2
@@ -12,13 +12,15 @@#include <internal/lib.h> #include "internal.h" -void perf_mmap__init (struct perf_mmap *map,bool overwrite, -libperf_unmap_cb_t unmap_cb) +void perf_mmap__init (struct perf_mmap *map,struct perf_mmap *prev, +bool overwrite,libperf_unmap_cb_t unmap_cb) { map->fd =-1 ; map->overwrite = overwrite; map->unmap_cb = unmap_cb;refcount_set (&map->refcnt,0 ); +if (prev) + prev->next = map; }size_t perf_mmap__mmap_len (struct perf_mmap *map)
tools/perf/util/evlist.c +3 -1
@@ -582,6 +582,8 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return NULL ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { +struct perf_mmap *prev = i ? &map[i -1 ].core :NULL ; +/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last
@@ -591,7 +593,7 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i].core, overwrite, perf_mmap__unmap_cb); +perf_mmap__init (&map[i].core, prev, overwrite, perf_mmap__unmap_cb); }return map;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 28/36 libperf: Move mmap allocation to perf_evlist__mmap_ops::get
In reply to Jiri OlsaMoving allocation of mmap array into perf_evlist__mmap_ops::get, so we centralize the mmap allocation.
Moving also nr_mmap setup to perf_evlist__mmap_ops so it's centralized and shared by both perf and libperf mmap code.
--- tools/perf/lib/evlist.c | 42 ++++++++++++++++++++++++---------------- tools/perf/util/evlist.c | 24 +++++++++-------------- 2 files changed, 34 insertions(+), 32 deletions(-)
tools/perf/lib/evlist.c +25 -17
@@ -338,10 +338,6 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, booint i;struct perf_mmap *map; - evlist->nr_mmaps =perf_cpu_map__nr (evlist->cpus); -if (perf_cpu_map__empty (evlist->cpus)) - evlist->nr_mmaps =perf_thread_map__nr (evlist->threads); - map =zalloc (evlist->nr_mmaps *sizeof (struct perf_mmap ));if (!map)return NULL ;
@@ -384,18 +380,22 @@ static void perf_evlist__set_sid_idx(struct perf_evlist *evlist,static struct perf_mmap *perf_evlist__mmap_cb_get (struct perf_evlist *evlist,bool overwrite,int idx) { -struct perf_mmap *map = &evlist->mmap[idx]; +struct perf_mmap *maps; -if (overwrite) { -if (!evlist->mmap_ovw) { - evlist->mmap_ovw =perf_evlist__alloc_mmap (evlist, true); -if (!evlist->mmap_ovw) -return NULL ; - } - map = &evlist->mmap_ovw[idx]; + maps = overwrite ? evlist->mmap_ovw : evlist->mmap; + +if (!maps) { + maps =perf_evlist__alloc_mmap (evlist, overwrite); +if (!maps) +return NULL ; + +if (overwrite) + evlist->mmap_ovw = maps; +else + evlist->mmap = maps; } -return map; +return &maps[idx]; }#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y))
@@ -556,6 +556,17 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,return -1 ; } +static int perf_evlist__nr_mmaps (struct perf_evlist *evlist) +{ +int nr_mmaps; + + nr_mmaps =perf_cpu_map__nr (evlist->cpus); +if (perf_cpu_map__empty (evlist->cpus)) + nr_mmaps =perf_thread_map__nr (evlist->threads); + +return nr_mmaps; +} +int perf_evlist__mmap_ops (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops,struct perf_mmap_param *mp)
@@ -567,10 +578,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,if (!ops || !ops->get || !ops->mmap)return -EINVAL ; -if (!evlist->mmap) - evlist->mmap =perf_evlist__alloc_mmap (evlist, false); -if (!evlist->mmap) -return -ENOMEM ; + evlist->nr_mmaps =perf_evlist__nr_mmaps (evlist);perf_evlist__for_each_entry (evlist, evsel) { if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
tools/perf/util/evlist.c +9 -15
@@ -574,9 +574,6 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,int i;struct mmap *map; - evlist->core.nr_mmaps =perf_cpu_map__nr (evlist->core.cpus); -if (perf_cpu_map__empty (evlist->core.cpus)) - evlist->core.nr_mmaps =perf_thread_map__nr (evlist->core.threads); map =zalloc (evlist->core.nr_mmaps *sizeof (struct mmap ));if (!map)return NULL ;
@@ -614,19 +611,21 @@ static struct perf_mmap* perf_evlist__mmap_cb_get(struct perf_evlist *_evlist, bool overwrite, int idx) { struct evlist *evlist = container_of(_evlist, struct evlist, core); -struct mmap *maps = evlist->mmap; +struct mmap *maps; -if (overwrite) { - maps = evlist->overwrite_mmap; + maps = overwrite ? evlist->overwrite_mmap : evlist->mmap; -if (!maps) { - maps =evlist__alloc_mmap (evlist, true); -if (!maps) -return NULL ; +if (!maps) { + maps =evlist__alloc_mmap (evlist, overwrite); +if (!maps) +return NULL ; +if (overwrite) { evlist->overwrite_mmap = maps;if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY )perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); + }else { + evlist->mmap = maps; } }
@@ -787,11 +786,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .mmap = perf_evlist__mmap_cb_mmap, }; -if (!evlist->mmap) - evlist->mmap =evlist__alloc_mmap (evlist, false); -if (!evlist->mmap) -return -ENOMEM ; - evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); mp.core.mask = evlist->core.mmap_len - page_size -1 ;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 31/36 libperf: Add _GNU_SOURCE define to compilation
In reply to Jiri Olsa--- tools/perf/lib/Makefile | 1 + tools/perf/lib/tests/Makefile | 2 ++ 2 files changed, 3 insertions(+)
tools/perf/lib/Makefile +1 -0
@@ -73,6 +73,7 @@ override CFLAGS += -Werror -Wall override CFLAGS += -fPIC override CFLAGS += $(INCLUDES) override CFLAGS += -fvisibility=hidden +override CFLAGS += -D_GNU_SOURCE all:
tools/perf/lib/tests/Makefile +2 -0
@@ -12,6 +12,8 @@ else CFLAGS := -g -Wall endif +CFLAGS += -D_GNU_SOURCE + all: include $(srctree)/tools/scripts/Makefile.include
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 32/36 libperf: Add tests_mmap_thread test
In reply to Jiri OlsaAdding mmaping tests that generates 100 prctl calls in monitored child process and validates it gets 100 events in ring buffer.
--- tools/perf/lib/tests/test-evlist.c | 119 +++++++++++++++++++++++++++++ 1 file changed, 119 insertions(+)
tools/perf/lib/tests/test-evlist.c +119 -0
@@ -1,12 +1,21 @@// SPDX-License-Identifier: GPL-2.0 #include <stdio.h> #include <stdarg.h> +#include <unistd.h> +#include <stdlib.h> #include <linux/perf_event.h> +#include <linux/limits.h> +#include <sys/types.h> +#include <sys/wait.h> +#include <sys/prctl.h> #include <perf/cpumap.h> #include <perf/threadmap.h> #include <perf/evlist.h> #include <perf/evsel.h> +#include <perf/mmap.h> +#include <perf/event.h> #include <internal/tests.h> +#include <api/fs/fs.h> static int libperf_print (enum libperf_print_level level,const char *fmt,va_list ap)
@@ -181,6 +190,115 @@ static int test_stat_thread_enable(void)return 0 ; } +static int test_mmap_thread (void ) +{ +struct perf_evlist *evlist; +struct perf_evsel *evsel; +struct perf_mmap *map; +struct perf_cpu_map *cpus; +struct perf_thread_map *threads; +struct perf_event_attr attr = { + .type =PERF_TYPE_TRACEPOINT , + .sample_period =1 , + .wakeup_watermark =1 , + .disabled =1 , + }; +char path[PATH_MAX ]; +int id, err, pid, go_pipe[2 ]; +union perf_event *event; +char bf; +int count =0 ; + +snprintf (path,PATH_MAX ,"%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id" , +sysfs__mountpoint ()); + +if (filename__read_int (path, &id)) { +fprintf (stderr,"error: failed to get tracepoint id: %s\n" , path); +return -1 ; + } + + attr.config = id; + + err =pipe (go_pipe); +__T ("failed to create pipe" , err ==0 ); + +fflush (NULL ); + + pid =fork (); +if (!pid) { +int i; + +read (go_pipe[0 ], &bf,1 ); + +/* Generate 100 prctl calls. */ +for (i =0 ; i <100 ; i++) +prctl (0 ,0 ,0 ,0 ,0 ); + +exit (0 ); + } + + threads =perf_thread_map__new_dummy (); +__T ("failed to create threads" , threads); + + cpus =perf_cpu_map__dummy_new (); +__T ("failed to create cpus" , cpus); + +perf_thread_map__set_pid (threads,0 , pid); + + evlist =perf_evlist__new (); +__T ("failed to create evlist" , evlist); + + evsel =perf_evsel__new (&attr); +__T ("failed to create evsel1" , evsel); + +perf_evlist__add (evlist, evsel); + +perf_evlist__set_maps (evlist, cpus, threads); + + err =perf_evlist__open (evlist); +__T ("failed to open evlist" , err ==0 ); + + err =perf_evlist__mmap (evlist,4 ); +__T ("failed to mmap evlist" , err ==0 ); + +perf_evlist__enable (evlist); + +/* kick the child and wait for it to finish */ +write (go_pipe[1 ], &bf,1 ); +waitpid (pid,NULL ,0 ); + +/* +* There's no need to call perf_evlist__disable, +* monitored process is dead now. +*/ + +perf_evlist__for_each_mmap (evlist, map, false) { +if (perf_mmap__read_init (map) <0 ) +continue ; + +while ((event =perf_mmap__read_event (map)) !=NULL ) { + count++; +perf_mmap__consume (map); + } + +perf_mmap__read_done (map); + } + +/* calls perf_evlist__munmap/perf_evlist__close */ +perf_evlist__delete (evlist); + +perf_thread_map__put (threads); +perf_cpu_map__put (cpus); + +/* +* The generated prctl calls should match the +* number of events in the buffer. +*/ +__T ("failed count" , count ==100 ); + +return 0 ; +} +int main (int argc,char **argv) { __T_START;
@@ -190,6 +308,7 @@ int main(int argc, char **argv)test_stat_cpu ();test_stat_thread ();test_stat_thread_enable (); +test_mmap_thread (); __T_OK;return 0 ;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 34/36 libperf: Keep count of failed tests
In reply to Jiri OlsaKeeping the count of failed tests, so we
get better output with failures, like:
# make tests
...
running static:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...FAILED test-evlist.c:53 failed to create evsel2
FAILED test-evlist.c:163 failed to create evsel2
FAILED test-evlist.c:287 failed count
FAILED (3)
- running test-evsel.c...OK
running dynamic:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...FAILED test-evlist.c:53 failed to create evsel2
FAILED test-evlist.c:163 failed to create evsel2
FAILED test-evlist.c:287 failed count
FAILED (3)
- running test-evsel.c...OK
...--- tools/perf/lib/include/internal/tests.h | 20 +++++++++++++++++--- tools/perf/lib/tests/test-cpumap.c | 2 +- tools/perf/lib/tests/test-evlist.c | 2 +- tools/perf/lib/tests/test-evsel.c | 2 +- tools/perf/lib/tests/test-threadmap.c | 2 +- 5 files changed, 21 insertions(+), 7 deletions(-)
tools/perf/lib/include/internal/tests.h +17 -3
@@ -4,14 +4,28 @@#include <stdio.h> -#define __T_START fprintf(stdout, "- running %s...", __FILE__) -#define __T_OK fprintf(stdout, "OK\n") -#define __T_FAIL fprintf(stdout, "FAIL\n") +int tests_failed; + +#define __T_START \ +do { \ + fprintf(stdout, "- running %s...", __FILE__); \ + fflush(NULL); \ + tests_failed = 0; \ +} while (0) + +#define __T_END \ +do { \ + if (tests_failed) \ + fprintf(stdout, " FAILED (%d)\n", tests_failed); \ + else \ + fprintf(stdout, "OK\n"); \ +} while (0)#define __T (text, cond) \ do { \ if (!(cond)) { \ fprintf(stderr, "FAILED %s:%d %s\n", __FILE__, __LINE__, text); \ + tests_failed++; \ return -1; \ } \ } while (0)
tools/perf/lib/tests/test-cpumap.c +1 -1
@@ -26,6 +26,6 @@ int main(int argc, char **argv)perf_cpu_map__put (cpus);perf_cpu_map__put (cpus); - __T_OK; + __T_END;return 0 ; }
tools/perf/lib/tests/test-evlist.c +1 -1
@@ -407,6 +407,6 @@ int main(int argc, char **argv)test_mmap_thread ();test_mmap_cpus (); - __T_OK; + __T_END;return 0 ; }
tools/perf/lib/tests/test-evsel.c +1 -1
@@ -130,6 +130,6 @@ int main(int argc, char **argv)test_stat_thread ();test_stat_thread_enable (); - __T_OK; + __T_END;return 0 ; }
tools/perf/lib/tests/test-threadmap.c +1 -1
@@ -26,6 +26,6 @@ int main(int argc, char **argv)perf_thread_map__put (threads);perf_thread_map__put (threads); - __T_OK; + __T_END;return 0 ; }
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 35/36 libperf: Do not export perf_evsel__init/perf_evlist__init
In reply to Jiri OlsaThere's no point to export perf_evsel__init/perf_evlist__init, it's called from perf_evsel__new/perf_evlist__new respectively.
It's used only from perf where perf_evsel/perf_evlist is embedded perf's evsel/evlist.
--- tools/perf/lib/include/internal/evlist.h | 1 + tools/perf/lib/include/internal/evsel.h | 1 + tools/perf/lib/include/perf/evlist.h | 1 - tools/perf/lib/include/perf/evsel.h | 2 -- tools/perf/lib/libperf.map | 2 -- 5 files changed, 2 insertions(+), 5 deletions(-)
tools/perf/lib/include/internal/evlist.h +1 -0
@@ -50,6 +50,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, struct perf_mmap_param *mp); +void perf_evlist__init (struct perf_evlist *evlist);void perf_evlist__exit (struct perf_evlist *evlist);/**
tools/perf/lib/include/internal/evsel.h +1 -0
@@ -50,6 +50,7 @@ struct perf_evsel {bool system_wide; }; +void perf_evsel__init (struct perf_evsel *evsel,struct perf_event_attr *attr);int perf_evsel__alloc_fd (struct perf_evsel *evsel,int ncpus,int nthreads);void perf_evsel__close_fd (struct perf_evsel *evsel);void perf_evsel__free_fd (struct perf_evsel *evsel);
tools/perf/lib/include/perf/evlist.h +0 -1
@@ -10,7 +10,6 @@ struct perf_evsel;struct perf_cpu_map ;struct perf_thread_map ; -LIBPERF_API void perf_evlist__init(struct perf_evlist *evlist); LIBPERF_API void perf_evlist__add(struct perf_evlist *evlist,struct perf_evsel *evsel); LIBPERF_API void perf_evlist__remove(struct perf_evlist *evlist,
tools/perf/lib/include/perf/evsel.h +0 -2
@@ -21,8 +21,6 @@ struct perf_counts_values { }; }; -LIBPERF_API void perf_evsel__init(struct perf_evsel *evsel, -struct perf_event_attr *attr); LIBPERF_API struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr); LIBPERF_API void perf_evsel__delete(struct perf_evsel *evsel); LIBPERF_API int perf_evsel__open(struct perf_evsel *evsel, struct perf_cpu_map *cpus,
tools/perf/lib/libperf.map +0 -2
@@ -21,7 +21,6 @@ LIBPERF_0.0.1 { perf_evsel__delete; perf_evsel__enable; perf_evsel__disable; - perf_evsel__init; perf_evsel__open; perf_evsel__close; perf_evsel__read;
@@ -34,7 +33,6 @@ LIBPERF_0.0.1 { perf_evlist__close; perf_evlist__enable; perf_evlist__disable; - perf_evlist__init; perf_evlist__add; perf_evlist__remove; perf_evlist__next;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 36/36 libperf: Add pr_err macro
In reply to Jiri OlsaAnd missing include for "perf/core.h" header, which provides LIBPERF_* debug levels and adding missing pr_err support.
--- tools/perf/lib/include/perf/core.h | 1 + tools/perf/lib/internal.h | 3 +++ 2 files changed, 4 insertions(+)
tools/perf/lib/include/perf/core.h +1 -0
@@ -9,6 +9,7 @@#endif enum libperf_print_level { +LIBPERF_ERR ,LIBPERF_WARN ,LIBPERF_INFO ,LIBPERF_DEBUG ,
tools/perf/lib/internal.h +3 -0
@@ -2,6 +2,8 @@#ifndef __LIBPERF_INTERNAL_H#define __LIBPERF_INTERNAL_H +#include <perf/core.h> +void libperf_print (enum libperf_print_level level,const char *format, ...) __attribute__((format (printf,2 ,3 )));
@@ -11,6 +13,7 @@ do { \ libperf_print(level, "libperf: " fmt, ##__VA_ARGS__); \ } while (0) +#define pr_err (fmt, ...) __pr(LIBPERF_ERR, fmt, ##__VA_ARGS__)#define pr_warning (fmt, ...) __pr(LIBPERF_WARN, fmt, ##__VA_ARGS__)#define pr_info (fmt, ...) __pr(LIBPERF_INFO, fmt, ##__VA_ARGS__)#define pr_debug (fmt, ...) __pr(LIBPERF_DEBUG, fmt, ##__VA_ARGS__)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 33/36 libperf: Add tests_mmap_cpus test
In reply to Jiri OlsaAdding mmaping tests that generates prctl call on every cpu validates it gets all the related events in ring buffer.
--- tools/perf/lib/tests/test-evlist.c | 97 ++++++++++++++++++++++++++++++ 1 file changed, 97 insertions(+)
tools/perf/lib/tests/test-evlist.c +97 -0
@@ -1,5 +1,6 @@// SPDX-License-Identifier: GPL-2.0 #include <stdio.h> +#include <sched.h> #include <stdarg.h> #include <unistd.h> #include <stdlib.h>
@@ -299,6 +300,101 @@ static int test_mmap_thread(void)return 0 ; } +static int test_mmap_cpus (void ) +{ +struct perf_evlist *evlist; +struct perf_evsel *evsel; +struct perf_mmap *map; +struct perf_cpu_map *cpus; +struct perf_event_attr attr = { + .type =PERF_TYPE_TRACEPOINT , + .sample_period =1 , + .wakeup_watermark =1 , + .disabled =1 , + }; +cpu_set_t saved_mask; +char path[PATH_MAX ]; +int id, err, cpu, tmp; +union perf_event *event; +int count =0 ; + +snprintf (path,PATH_MAX ,"%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id" , +sysfs__mountpoint ()); + +if (filename__read_int (path, &id)) { +fprintf (stderr,"error: failed to get tracepoint id: %s\n" , path); +return -1 ; + } + + attr.config = id; + + cpus =perf_cpu_map__new (NULL ); +__T ("failed to create cpus" , cpus); + + evlist =perf_evlist__new (); +__T ("failed to create evlist" , evlist); + + evsel =perf_evsel__new (&attr); +__T ("failed to create evsel1" , evsel); + +perf_evlist__add (evlist, evsel); + +perf_evlist__set_maps (evlist, cpus,NULL ); + + err =perf_evlist__open (evlist); +__T ("failed to open evlist" , err ==0 ); + + err =perf_evlist__mmap (evlist,4 ); +__T ("failed to mmap evlist" , err ==0 ); + +perf_evlist__enable (evlist); + + err =sched_getaffinity (0 ,sizeof (saved_mask), &saved_mask); +__T ("sched_getaffinity failed" , err ==0 ); + +perf_cpu_map__for_each_cpu (cpu, tmp, cpus) { +cpu_set_t mask; + +CPU_ZERO (&mask); +CPU_SET (cpu, &mask); + + err =sched_setaffinity (0 ,sizeof (mask), &mask); +__T ("sched_setaffinity failed" , err ==0 ); + +prctl (0 ,0 ,0 ,0 ,0 ); + } + + err =sched_setaffinity (0 ,sizeof (saved_mask), &saved_mask); +__T ("sched_setaffinity failed" , err ==0 ); + +perf_evlist__disable (evlist); + +perf_evlist__for_each_mmap (evlist, map, false) { +if (perf_mmap__read_init (map) <0 ) +continue ; + +while ((event =perf_mmap__read_event (map)) !=NULL ) { + count++; +perf_mmap__consume (map); + } + +perf_mmap__read_done (map); + } + +/* calls perf_evlist__munmap/perf_evlist__close */ +perf_evlist__delete (evlist); + +/* +* The generated prctl events should match the +* number of cpus or be bigger (we are system-wide). +*/ +__T ("failed count" , count >=perf_cpu_map__nr (cpus)); + +perf_cpu_map__put (cpus); + +return 0 ; +} +int main (int argc,char **argv) { __T_START;
@@ -309,6 +405,7 @@ int main(int argc, char **argv)test_stat_thread ();test_stat_thread_enable ();test_mmap_thread (); +test_mmap_cpus (); __T_OK;return 0 ;
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 29/36 libperf: Move mask setup to perf_evlist__mmap_ops function
In reply to Jiri OlsaMoving mask setup to perf_evlist__mmap_ops function, because it's same on both perf and libperf path.
--- tools/perf/lib/evlist.c | 3 ++- tools/perf/util/evlist.c | 1 - 2 files changed, 2 insertions(+), 2 deletions(-)
tools/perf/lib/evlist.c +2 -1
@@ -578,6 +578,8 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,if (!ops || !ops->get || !ops->mmap)return -EINVAL ; + mp->mask = evlist->mmap_len - page_size -1 ; + evlist->nr_mmaps =perf_evlist__nr_mmaps (evlist);perf_evlist__for_each_entry (evlist, evsel) {
@@ -605,7 +607,6 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages) }; evlist->mmap_len = (pages +1 ) * page_size; - mp.mask = evlist->mmap_len - page_size -1 ;return perf_evlist__mmap_ops (evlist, &ops, &mp); }
tools/perf/util/evlist.c +0 -1
@@ -788,7 +788,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); - mp.core.mask = evlist->core.mmap_len - page_size -1 ;auxtrace_mmap_params__init (&mp.auxtrace_mp, evlist->core.mmap_len, auxtrace_pages, auxtrace_overwrite);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 30/36 libperf: Link static tests with libapi.a
In reply to Jiri OlsaBoth static and dynamic tests needs to link libapi.a, because it's using its functions. Adding also include path for libapi includes.
--- tools/perf/lib/Makefile | 1 + tools/perf/lib/tests/Makefile | 6 +++--- 2 files changed, 4 insertions(+), 3 deletions(-)
tools/perf/lib/Makefile +1 -0
@@ -107,6 +107,7 @@ else endif LIBAPI = $(API_PATH)libapi.a +export LIBAPI $(LIBAPI): FORCE $(Q)$(MAKE) -C $(LIB_DIR) O=$(OUTPUT) $(OUTPUT)libapi.a
tools/perf/lib/tests/Makefile +3 -3
@@ -16,13 +16,13 @@ all: include $(srctree)/tools/scripts/Makefile.include -INCLUDE = -I$(srctree)/tools/perf/lib/include -I$(srctree)/tools/include +INCLUDE = -I$(srctree)/tools/perf/lib/include -I$(srctree)/tools/include -I$(srctree)/tools/lib $(TESTS_A): FORCE - $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -o $@ $(subst -a,.c,$@) ../libperf.a + $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -o $@ $(subst -a,.c,$@) ../libperf.a $(LIBAPI) $(TESTS_SO): FORCE - $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -L.. -o $@ $(subst -so,.c,$@) -lperf + $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -L.. -o $@ $(subst -so,.c,$@) $(LIBAPI) -lperf all: $(TESTS_A) $(TESTS_SO)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 24/36 libperf: Add perf_evlist__exit function
In reply to Jiri OlsaAdding perf_evlist__exit function, so far it's not exported and added only for internal use for perf and libperf.
Using it to release cpus/threads and pollfd array.
--- tools/perf/lib/evlist.c | 12 +++++++++++- tools/perf/lib/include/internal/evlist.h | 2 ++ tools/perf/util/evlist.c | 6 +----- 3 files changed, 14 insertions(+), 6 deletions(-)
tools/perf/lib/evlist.c +11 -1
@@ -109,13 +109,23 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)return next; } +void perf_evlist__exit (struct perf_evlist *evlist) +{ +perf_cpu_map__put (evlist->cpus); +perf_thread_map__put (evlist->threads); + evlist->cpus =NULL ; + evlist->threads =NULL ; +fdarray__exit (&evlist->pollfd); +} +void perf_evlist__delete (struct perf_evlist *evlist) {if (evlist ==NULL )return ;perf_evlist__munmap (evlist); -fdarray__exit (&evlist->pollfd); +perf_evlist__close (evlist); +perf_evlist__exit (evlist);free (evlist); }
tools/perf/lib/include/internal/evlist.h +2 -0
@@ -48,6 +48,8 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, struct perf_mmap_param *mp); +void perf_evlist__exit (struct perf_evlist *evlist); +/** * __perf_evlist__for_each_entry - iterate thru all the evsels * @list: list_head instance to iterate
tools/perf/util/evlist.c +1 -5
@@ -138,7 +138,7 @@ void evlist__exit(struct evlist *evlist) {zfree (&evlist->mmap);zfree (&evlist->overwrite_mmap); -fdarray__exit (&evlist->core.pollfd); +perf_evlist__exit (&evlist->core); }void evlist__delete (struct evlist *evlist)
@@ -148,10 +148,6 @@ void evlist__delete(struct evlist *evlist)evlist__munmap (evlist);evlist__close (evlist); -perf_cpu_map__put (evlist->core.cpus); -perf_thread_map__put (evlist->core.threads); - evlist->core.cpus =NULL ; - evlist->core.threads =NULL ;evlist__purge (evlist);evlist__exit (evlist);free (evlist);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 18/36 perf tools: Add perf_evlist__mmap_cb_idx function
In reply to Jiri OlsaAdding perf_evlist__mmap_cb_idx function to callauxtrace_mmap_params__set_idx on each new index
during perf_evlist__mmap_ops call.
--- tools/perf/util/evlist.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
tools/perf/util/evlist.c +14 -0
@@ -715,6 +715,17 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,return 0 ; } +static void +perf_evlist__mmap_cb_idx (struct perf_evlist *_evlist, +struct perf_mmap_param *_mp, +int idx,bool per_cpu) +{ + struct evlist *evlist = container_of(_evlist, struct evlist, core); + struct mmap_params *mp = container_of(_mp, struct mmap_params, core); + +auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, idx, per_cpu); +} +static int evlist__mmap_per_cpu (struct evlist *evlist,struct mmap_params *mp) {
@@ -911,6 +922,9 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .flush = flush, .comp_level = comp_level }; +struct perf_evlist_mmap_ops ops__maybe_unused = { + .idx = perf_evlist__mmap_cb_idx, + };if (!evlist->mmap) evlist->mmap =evlist__alloc_mmap (evlist, false);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 15/36 libperf: Add perf_evlist_mmap_ops::idx callback
In reply to Jiri OlsaAdd the perf_evlist_mmap_ops::idx callback to be called in mmap_per_cpu and mmap_per_thread with current cpu and thread indexes.
It's used by current aux code, so perf will use this callback to set the aux index.
--- tools/perf/lib/evlist.c | 18 +++++++++++++----- tools/perf/lib/include/internal/evlist.h | 4 ++++ 2 files changed, 17 insertions(+), 5 deletions(-)
tools/perf/lib/evlist.c +13 -5
@@ -426,7 +426,8 @@ mmap_per_evsel(struct perf_evlist *evlist, int idx, }static int -mmap_per_thread (struct perf_evlist *evlist,struct perf_mmap_param *mp) +mmap_per_thread (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp) {int thread;int nr_threads =perf_thread_map__nr (evlist->threads);
@@ -435,6 +436,9 @@ mmap_per_thread(struct perf_evlist *evlist, struct perf_mmap_param *mp)int output =-1 ;int output_overwrite =-1 ; +if (ops->idx) + ops->idx (evlist, mp, thread, false); +if (mmap_per_evsel (evlist, thread, mp,0 , thread, &output, &output_overwrite)) goto out_unmap;
@@ -448,7 +452,8 @@ mmap_per_thread(struct perf_evlist *evlist, struct perf_mmap_param *mp) }static int -mmap_per_cpu (struct perf_evlist *evlist,struct perf_mmap_param *mp) +mmap_per_cpu (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp) {int nr_threads =perf_thread_map__nr (evlist->threads);int nr_cpus =perf_cpu_map__nr (evlist->cpus);
@@ -458,6 +463,9 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_mmap_param *mp)int output =-1 ;int output_overwrite =-1 ; +if (ops->idx) + ops->idx (evlist, mp, cpu, true); +for (thread =0 ; thread < nr_threads; thread++) {if (mmap_per_evsel (evlist, cpu, mp, cpu, thread, &output, &output_overwrite))
@@ -496,15 +504,15 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist, }if (perf_cpu_map__empty (cpus)) -return mmap_per_thread (evlist, mp); +return mmap_per_thread (evlist, ops, mp); -return mmap_per_cpu (evlist, mp); +return mmap_per_cpu (evlist, ops, mp); }int perf_evlist__mmap (struct perf_evlist *evlist,int pages) {struct perf_mmap_param mp; -struct perf_evlist_mmap_ops ops; +struct perf_evlist_mmap_ops ops = {0 }; evlist->mmap_len = (pages +1 ) * page_size; mp.mask = evlist->mmap_len - page_size -1 ;
tools/perf/lib/include/internal/evlist.h +4 -0
@@ -27,7 +27,11 @@ struct perf_evlist {struct perf_mmap *mmap_ovw; }; +typedef void +(*perf_evlist_mmap__cb_idx_t )(struct perf_evlist *,struct perf_mmap_param *,int ,bool ); +struct perf_evlist_mmap_ops { +perf_evlist_mmap__cb_idx_t idx; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 9/36 libperf: Add perf_mmap__consume() function
In reply to Jiri OlsaMove perf_mmap__consume() vrom tools/perf to libperf and export it in the perf/mmap.h header.
Move also the needed helpers perf_mmap__write_tail(), perf_mmap__read_head() and perf_mmap__empty().
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 3 +- tools/perf/builtin-kvm.c | 5 +-- tools/perf/builtin-top.c | 3 +- tools/perf/builtin-trace.c | 3 +- tools/perf/lib/Makefile | 5 +-- tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/include/perf/mmap.h | 11 +++++++ tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 32 ++++++++++++++++++++ tools/perf/tests/backward-ring-buffer.c | 1 + tools/perf/tests/bpf.c | 1 + tools/perf/tests/code-reading.c | 3 +- tools/perf/tests/keep-tracking.c | 3 +- tools/perf/tests/mmap-basic.c | 3 +- tools/perf/tests/openat-syscall-tp-fields.c | 3 +- tools/perf/tests/perf-record.c | 3 +- tools/perf/tests/sw-clock.c | 3 +- tools/perf/tests/switch-tracking.c | 3 +- tools/perf/tests/task-exit.c | 3 +- tools/perf/util/evlist.c | 3 +- tools/perf/util/mmap.c | 32 +++++--------------- tools/perf/util/mmap.h | 12 -------- tools/perf/util/python.c | 3 +- 23 files changed, 87 insertions(+), 54 deletions(-) create mode 100644 tools/perf/lib/include/perf/mmap.h
tools/perf/arch/x86/tests/perf-time-to-tsc.c +2 -1
@@ -9,6 +9,7 @@#include <sys/prctl.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "parse-events.h"
@@ -139,7 +140,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybe comm2_time = sample.time; } next_event: -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md); }
tools/perf/builtin-kvm.c +3 -2
@@ -46,6 +46,7 @@#include <semaphore.h> #include <signal.h> #include <math.h> +#include <perf/mmap.h> static const char *get_filename_for_perf_kvm (void ) {
@@ -766,7 +767,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,while ((event =perf_mmap__read_event (md)) !=NULL ) { err =perf_evlist__parse_sample_timestamp (evlist, event, ×tamp);if (err) { -perf_mmap__consume (md); +perf_mmap__consume (&md->core);pr_err ("Failed to parse sample\n" );return -1 ; }
@@ -776,7 +777,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,* FIXME: Here we can't consume the event, as perf_session__queue_event will * point to it, and it'll get possibly overwritten by the kernel. */ -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (err) {pr_err ("Failed to enqueue sample: %d\n" , err);
tools/perf/builtin-top.c +2 -1
@@ -82,6 +82,7 @@#include <linux/err.h> #include <linux/ctype.h> +#include <perf/mmap.h> static volatile int done;static volatile int resize;
@@ -883,7 +884,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)if (ret)break ; -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (top->qe.rotate) {pthread_mutex_lock (&top->qe.mutex);
tools/perf/builtin-trace.c +2 -1
@@ -77,6 +77,7 @@#include <sys/sysmacros.h> #include <linux/ctype.h> +#include <perf/mmap.h> #ifndef O_CLOEXEC # define O_CLOEXEC 02000000
@@ -3460,7 +3461,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)if (err) goto out_disable; -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (interrupted) goto out_disable;
tools/perf/lib/Makefile +3 -2
@@ -172,8 +172,9 @@ install_headers: $(call do_install,include/perf/cpumap.h,$(prefix)/include/perf,644); \ $(call do_install,include/perf/threadmap.h,$(prefix)/include/perf,644); \ $(call do_install,include/perf/evlist.h,$(prefix)/include/perf,644); \ - $(call do_install,include/perf/evsel.h,$(prefix)/include/perf,644); - $(call do_install,include/perf/event.h,$(prefix)/include/perf,644); + $(call do_install,include/perf/evsel.h,$(prefix)/include/perf,644); \ + $(call do_install,include/perf/event.h,$(prefix)/include/perf,644); \ + $(call do_install,include/perf/mmap.h,$(prefix)/include/perf,644); install_pkgconfig: $(LIBPERF_PC) $(call QUIET_INSTALL, $(LIBPERF_PC)) \
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -49,4 +49,6 @@ void perf_mmap__munmap(struct perf_mmap *map);void perf_mmap__get (struct perf_mmap *map);void perf_mmap__put (struct perf_mmap *map); +u64 perf_mmap__read_head (struct perf_mmap *map); +#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/include/perf/mmap.h new file +11 -0
@@ -0,0 +1,11 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __LIBPERF_MMAP_H +#define __LIBPERF_MMAP_H + +#include <perf/core.h> + +struct perf_mmap ; + +LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); + +#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/libperf.map +1 -0
@@ -40,6 +40,7 @@ LIBPERF_0.0.1 { perf_evlist__next; perf_evlist__set_maps; perf_evlist__poll; + perf_mmap__consume; local: *; };
tools/perf/lib/mmap.c +32 -0
@@ -1,5 +1,8 @@// SPDX-License-Identifier: GPL-2.0 #include <sys/mman.h> +#include <linux/ring_buffer.h> +#include <linux/perf_event.h> +#include <perf/mmap.h> #include <internal/mmap.h> #include <internal/lib.h>
@@ -58,3 +61,32 @@ void perf_mmap__put(struct perf_mmap *map)if (refcount_dec_and_test (&map->refcnt))perf_mmap__munmap (map); } + +static inline void perf_mmap__write_tail (struct perf_mmap *md,u64 tail) +{ +ring_buffer_write_tail (md->base, tail); +} + +u64 perf_mmap__read_head (struct perf_mmap *map) +{ +return ring_buffer_read_head (map->base); +} + +static bool perf_mmap__empty (struct perf_mmap *map) +{ +struct perf_event_mmap_page *pc = map->base; + +return perf_mmap__read_head (map) == map->prev && !pc->aux_size; +} + +void perf_mmap__consume (struct perf_mmap *map) +{ +if (!map->overwrite) { +u64 old = map->prev; + +perf_mmap__write_tail (map, old); + } + +if (refcount_read (&map->refcnt) ==1 &&perf_mmap__empty (map)) +perf_mmap__put (map); +}
tools/perf/tests/backward-ring-buffer.c +1 -0
@@ -13,6 +13,7 @@#include "util/mmap.h" #include <errno.h> #include <linux/string.h> +#include <perf/mmap.h> #define NR_ITERS 111
tools/perf/tests/bpf.c +1 -0
@@ -15,6 +15,7 @@#include <linux/string.h> #include <api/fs/fs.h> #include <bpf/bpf.h> +#include <perf/mmap.h> #include "tests.h" #include "llvm.h" #include "debug.h"
tools/perf/tests/code-reading.c +2 -1
@@ -10,6 +10,7 @@#include <sys/param.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "dso.h"
@@ -430,7 +431,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,while ((event =perf_mmap__read_event (md)) !=NULL ) { ret =process_event (machine, evlist, event, state); -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (ret <0 )return ret; }
tools/perf/tests/keep-tracking.c +2 -1
@@ -5,6 +5,7 @@#include <sys/prctl.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "parse-events.h"
@@ -46,7 +47,7 @@ static int find_comm(struct evlist *evlist, const char *comm) (pid_t)event->comm.tid == getpid() && strcmp(event->comm.comm, comm) == 0) found +=1 ; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md); }
tools/perf/tests/mmap-basic.c +2 -1
@@ -16,6 +16,7 @@#include <linux/kernel.h> #include <linux/string.h> #include <perf/evlist.h> +#include <perf/mmap.h> /* * This test will generate random numbers of calls to some getpid syscalls,
@@ -139,7 +140,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuse goto out_delete_evlist; } nr_events[evsel->idx]++; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md);
tools/perf/tests/openat-syscall-tp-fields.c +2 -1
@@ -13,6 +13,7 @@#include "debug.h" #include "util/mmap.h" #include <errno.h> +#include <perf/mmap.h> #ifndef O_DIRECTORY #define O_DIRECTORY 00200000
@@ -103,7 +104,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtest ++nr_events;if (type !=PERF_RECORD_SAMPLE ) { -perf_mmap__consume (md); +perf_mmap__consume (&md->core);continue ; }
tools/perf/tests/perf-record.c +2 -1
@@ -6,6 +6,7 @@#include <pthread.h> #include <sched.h> +#include <perf/mmap.h> #include "evlist.h" #include "evsel.h" #include "debug.h"
@@ -276,7 +277,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unus ++errs; } -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md); }
tools/perf/tests/sw-clock.c +2 -1
@@ -15,6 +15,7 @@#include "util/mmap.h" #include "util/thread_map.h" #include <perf/evlist.h> +#include <perf/mmap.h> #define NR_LOOPS 10000000
@@ -117,7 +118,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id) total_periods += sample.period; nr_samples++; next_event: -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md);
tools/perf/tests/switch-tracking.c +2 -1
@@ -8,6 +8,7 @@#include <linux/zalloc.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "parse-events.h"
@@ -275,7 +276,7 @@ static int process_events(struct evlist *evlist,while ((event =perf_mmap__read_event (md)) !=NULL ) { cnt +=1 ; ret =add_event (evlist, &events, event); -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (ret <0 ) goto out_free_nodes; }
tools/perf/tests/task-exit.c +2 -1
@@ -12,6 +12,7 @@#include <linux/string.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> static int exited;static int nr_exit;
@@ -124,7 +125,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unusedif (event->header.type ==PERF_RECORD_EXIT ) nr_exit++; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md);
tools/perf/util/evlist.c +2 -1
@@ -42,6 +42,7 @@#include <perf/evlist.h> #include <perf/evsel.h> #include <perf/cpumap.h> +#include <perf/mmap.h> #include <internal/xyarray.h>
@@ -1749,7 +1750,7 @@ static void *perf_evlist__poll_thread(void *arg)else pr_warning ("cannot locate proper evsel for the side band event\n" ); -perf_mmap__consume (map); +perf_mmap__consume (&map->core); got_data = true; }perf_mmap__read_done (map);
tools/perf/util/mmap.c +7 -25
@@ -13,6 +13,7 @@#include <stdlib.h> #include <string.h> #include <unistd.h> // sysconf() +#include <perf/mmap.h> #ifdef HAVE_LIBNUMA_SUPPORT #include <numaif.h> #endif
@@ -95,7 +96,7 @@ union perf_event *perf_mmap__read_event(struct mmap *map)/* non-overwirte doesn't pause the ringbuffer */ if (!map->core.overwrite) - map->core.end =perf_mmap__read_head (map); + map->core.end =perf_mmap__read_head (&map->core); event =perf_mmap__read (map, &map->core.start, map->core.end);
@@ -105,25 +106,6 @@ union perf_event *perf_mmap__read_event(struct mmap *map)return event; } -static bool perf_mmap__empty (struct mmap *map) -{ -struct perf_event_mmap_page *pc = map->core.base; - -return perf_mmap__read_head (map) == map->core.prev && !pc->aux_size; -} - -void perf_mmap__consume (struct mmap *map) -{ -if (!map->core.overwrite) { -u64 old = map->core.prev; - -perf_mmap__write_tail (map, old); - } - -if (refcount_read (&map->core.refcnt) ==1 &&perf_mmap__empty (map)) -perf_mmap__put (&map->core); -} - int __weak auxtrace_mmap__mmap(struct auxtrace_mmap *mm __maybe_unused, struct auxtrace_mmap_params *mp __maybe_unused, void *userpg __maybe_unused,
@@ -420,7 +402,7 @@ static int overwrite_rb_find_range(void *buf, int mask, u64 *start, u64 *end)*/ static int __perf_mmap__read_init (struct mmap *md) { -u64 head =perf_mmap__read_head (md); +u64 head =perf_mmap__read_head (&md->core);u64 old = md->core.prev;unsigned char *data = md->core.base + page_size;unsigned long size;
@@ -437,7 +419,7 @@ static int __perf_mmap__read_init(struct mmap *md)WARN_ONCE (1 ,"failed to keep up with mmap data. (warn only once)\n" ); md->core.prev = head; -perf_mmap__consume (md); +perf_mmap__consume (&md->core);return -EAGAIN ; }
@@ -466,7 +448,7 @@ int perf_mmap__read_init(struct mmap *map)int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size)) { -u64 head =perf_mmap__read_head (md); +u64 head =perf_mmap__read_head (&md->core);unsigned char *data = md->core.base + page_size;unsigned long size;void *buf;
@@ -499,7 +481,7 @@ int perf_mmap__push(struct mmap *md, void *to, } md->core.prev = head; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); out:return rc; }
@@ -518,5 +500,5 @@ void perf_mmap__read_done(struct mmap *map)if (!refcount_read (&map->core.refcnt))return ; - map->core.prev =perf_mmap__read_head (map); + map->core.prev =perf_mmap__read_head (&map->core); }
tools/perf/util/mmap.h +0 -12
@@ -45,18 +45,6 @@ struct mmap_params {int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void mmap__munmap (struct mmap *map); -void perf_mmap__consume (struct mmap *map); - -static inline u64 perf_mmap__read_head (struct mmap *mm) -{ -return ring_buffer_read_head (mm->core.base); -} - -static inline void perf_mmap__write_tail (struct mmap *md,u64 tail) -{ -ring_buffer_write_tail (md->core.base, tail); -} -union perf_event *perf_mmap__read_forward (struct mmap *map);union perf_event *perf_mmap__read_event (struct mmap *map);
tools/perf/util/python.c +2 -1
@@ -6,6 +6,7 @@#include <linux/err.h> #include <perf/cpumap.h> #include <traceevent/event-parse.h> +#include <perf/mmap.h> #include "evlist.h" #include "callchain.h" #include "evsel.h"
@@ -1039,7 +1040,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist, err =perf_evsel__parse_sample (evsel, event, &pevent->sample);/* Consume the even only after we parsed it out. */ -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (err) return PyErr_Format(PyExc_OSError,
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 8/36 perf tools: Use perf_mmap way to detect aux mmap
In reply to Jiri OlsaWe will move this code to libperf shortly, so we need to free it of 'struct auxtrace_mmap' usage, because it won't be available in libperf (for now).
The perf_event_mmap_page::aux_size is set when the aux mmap is mapped, so the check is equivalent.
--- tools/perf/util/mmap.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tools/perf/util/mmap.c +3 -1
@@ -107,7 +107,9 @@ union perf_event *perf_mmap__read_event(struct mmap *map)static bool perf_mmap__empty (struct mmap *map) { -return perf_mmap__read_head (map) == map->core.prev && !map->auxtrace_mmap.base; +struct perf_event_mmap_page *pc = map->core.base; + +return perf_mmap__read_head (map) == map->core.prev && !pc->aux_size; }void perf_mmap__consume (struct mmap *map)
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 6/36 libperf: Add perf_mmap__unmap() function
In reply to Jiri OlsaMove perf_mmap__unmap() from tools/perf to libperf, to internal header internal/mmap.h. It will be used in the following patches. And rename the existing perf's function to mmap__munmap().
--- tools/perf/lib/include/internal/mmap.h | 1 + tools/perf/lib/mmap.c | 10 ++++++++++ tools/perf/util/evlist.c | 4 ++-- tools/perf/util/mmap.c | 11 +++-------- tools/perf/util/mmap.h | 2 +- 5 files changed, 17 insertions(+), 11 deletions(-)
tools/perf/lib/include/internal/mmap.h +1 -0
@@ -39,6 +39,7 @@ size_t perf_mmap__mmap_len(struct perf_mmap *map);void perf_mmap__init (struct perf_mmap *map,bool overwrite);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu); +void perf_mmap__munmap (struct perf_mmap *map);void perf_mmap__get (struct perf_mmap *map);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +10 -0
@@ -32,6 +32,16 @@ int perf_mmap__mmap(struct perf_mmap *map, struct perf_mmap_param *mp,return 0 ; } +void perf_mmap__munmap (struct perf_mmap *map) +{ +if (map && map->base !=NULL ) { +munmap (map->base,perf_mmap__mmap_len (map)); + map->base =NULL ; + map->fd =-1 ; +refcount_set (&map->refcnt,0 ); + } +} +void perf_mmap__get (struct perf_mmap *map) {refcount_inc (&map->refcnt);
tools/perf/util/evlist.c +2 -2
@@ -577,11 +577,11 @@ static void evlist__munmap_nofree(struct evlist *evlist)if (evlist->mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -perf_mmap__munmap (&evlist->mmap[i]); +mmap__munmap (&evlist->mmap[i]);if (evlist->overwrite_mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -perf_mmap__munmap (&evlist->overwrite_mmap[i]); +mmap__munmap (&evlist->overwrite_mmap[i]); }void evlist__munmap (struct evlist *evlist)
tools/perf/util/mmap.c +3 -8
@@ -115,7 +115,7 @@ void perf_mmap__put(struct mmap *map)BUG_ON (map->core.base &&refcount_read (&map->core.refcnt) ==0 );if (refcount_dec_and_test (&map->core.refcnt)) -perf_mmap__munmap (map); +mmap__munmap (map); }void perf_mmap__consume (struct mmap *map)
@@ -306,19 +306,14 @@ static void perf_mmap__aio_munmap(struct mmap *map __maybe_unused) }#endif -void perf_mmap__munmap (struct mmap *map) +void mmap__munmap (struct mmap *map) { +perf_mmap__munmap (&map->core);perf_mmap__aio_munmap (map);if (map->data !=NULL ) {munmap (map->data,mmap__mmap_len (map)); map->data =NULL ; } -if (map->core.base !=NULL ) { -munmap (map->core.base,mmap__mmap_len (map)); - map->core.base =NULL ; - map->core.fd =-1 ; -refcount_set (&map->core.refcnt,0 ); - }auxtrace_mmap__munmap (&map->auxtrace_mmap); }
tools/perf/util/mmap.h +1 -1
@@ -43,7 +43,7 @@ struct mmap_params { };int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu); -void perf_mmap__munmap (struct mmap *map); +void mmap__munmap (struct mmap *map);void perf_mmap__put (struct mmap *map);
Signature · 1 line
2.21.0
Jiri Olsa
reviewerPATCH v2 4/36 libperf: Add perf_mmap__mmap() function
In reply to Jiri OlsaMove perf_mmap__mmap() from tools/perf to libperf, it will be used in the following patches. And rename the existing perf's function to mmap__mmap().
--- tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/mmap.c | 18 ++++++++++++++++++ tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 12 +++--------- tools/perf/util/mmap.h | 2 +- 5 files changed, 25 insertions(+), 11 deletions(-)
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -37,5 +37,7 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map);void perf_mmap__init (struct perf_mmap *map,bool overwrite); +int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp, +int fd,int cpu);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +18 -0
@@ -1,4 +1,5 @@// SPDX-License-Identifier: GPL-2.0 +#include <sys/mman.h> #include <internal/mmap.h> #include <internal/lib.h>
@@ -13,3 +14,20 @@ size_t perf_mmap__mmap_len(struct perf_mmap *map) {return map->mask +1 + page_size; } + +int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp, +int fd,int cpu) +{ + map->prev =0 ; + map->mask = mp->mask; + map->base =mmap (NULL ,perf_mmap__mmap_len (map), mp->prot, +MAP_SHARED , fd,0 ); +if (map->base ==MAP_FAILED ) { + map->base =NULL ; +return -1 ; + } + + map->fd = fd; + map->cpu = cpu; +return 0 ; +}
tools/perf/util/evlist.c +1 -1
@@ -671,7 +671,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,if (*output ==-1 ) { *output = fd; -if (perf_mmap__mmap (&maps[idx], mp, *output, evlist_cpu) <0 ) +if (mmap__mmap (&maps[idx], mp, *output, evlist_cpu) <0 )return -1 ; }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )
tools/perf/util/mmap.c +3 -9
@@ -353,7 +353,7 @@ static void perf_mmap__setup_affinity_mask(struct mmap *map, struct mmap_paramsCPU_SET (map->core.cpu, &map->affinity_mask); } -int perf_mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu) +int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu) {/* * The last one will be done at perf_mmap__consume(), so that we
@@ -369,18 +369,12 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu)* perf_evlist__filter_pollfd(). */ refcount_set (&map->core.refcnt,2 ); - map->core.prev =0 ; - map->core.mask = mp->core.mask; - map->core.base =mmap (NULL ,mmap__mmap_len (map), mp->core.prot, -MAP_SHARED , fd,0 ); -if (map->core.base ==MAP_FAILED ) { + +if (perf_mmap__mmap (&map->core, &mp->core, fd, cpu)) {pr_debug2 ("failed to mmap perf event ring buffer, error %d\n" , errno); - map->core.base =NULL ;return -1 ; } - map->core.fd = fd; - map->core.cpu = cpu;perf_mmap__setup_affinity_mask (map, mp);
tools/perf/util/mmap.h +1 -1
@@ -42,7 +42,7 @@ struct mmap_params {struct auxtrace_mmap_params auxtrace_mp; }; -int perf_mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu); +int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void perf_mmap__munmap (struct mmap *map);void perf_mmap__get (struct mmap *map);
Signature · 1 line
2.21.0
Re: [PATCH 07/36] libperf: Add perf_mmap__put() function
In reply to Jiri OlsaEm Mon, Oct 07, 2019 at 02:53:15PM +0200, Jiri Olsa escreveu:
Quoted 21 lines
+++
b/tools/perf/lib/mmap.ctools/perf/lib/mmap.c +5
-1@@ -3,10 +3,12 @@#include <internal/mmap.h> #include <internal/lib.h> -void perf_mmap__init (struct perf_mmap *map,bool overwrite) +void perf_mmap__init (struct perf_mmap *map,bool overwrite, +libperf_unmap_cb_t unmap_cb) { map->fd =-1 ; map->overwrite = overwrite; + map->unmap_cb = unmap_cb;refcount_set (&map->refcnt,0 ); }@@ -40,9 +42,19 @@ void perf_mmap__munmap(struct perf_mmap *map) map->fd =-1 ;refcount_set (&map->refcnt,0 ); } +if (map && map->unmap_cb) + map->unmap_cb (map);
tools/perf/lib/mmap.c void perf_mmap__munmap(struct perf_mmap *map) · line 56
Quoted 10 lines
tools/perf/lib/mmap.c +4
-0}void perf_mmap__get (struct perf_mmap *map) {refcount_inc (&map->refcnt); } + +void perf_mmap__put (struct perf_mmap *map) +{ +BUG_ON (map->base &&refcount_read (&map->refcnt) ==0 );
Added missing linux/kernel.h to this file to pick up BUG_ON().
- Arnaldo
Re: [PATCH 12/36] libperf: Add perf_mmap__read_event() function
In reply to Jiri OlsaEm Mon, Oct 07, 2019 at 02:53:20PM +0200, Jiri Olsa escreveu:
Quoted 92 lines
Move
perf_mmap__read_event()fromtools/perfto libperf and export it in theperf/mmap.hheader.Signed-off-by Jiri Olsa <[email protected]>Cc Alexander Shishkin <[email protected]>Cc Michael Petlan <[email protected]>Cc Namhyung Kim <[email protected]>Cc Peter Zijlstra <[email protected]>Signed-off-by Arnaldo Carvalho de Melo <[email protected]>--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/internal/mmap.h | 1 + tools/perf/lib/include/perf/mmap.h | 1 + tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 79 ++++++++++++++++++++ tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 77 ------------------- tools/perf/util/mmap.h | 2 - tools/perf/util/python.c | 2 +- 22 files changed, 98 insertions(+), 95 deletions(-)tools/perf/arch/x86/tests/perf-time-to-tsc.c +1
-1@@ -121,7 +121,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybeif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample; if (event->header.type != PERF_RECORD_COMM ||tools/perf/builtin-kvm.c +1
-1@@ -764,7 +764,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,if (err <0 )return (err == -EAGAIN ) ?0 :-1 ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { err =perf_evlist__parse_sample_timestamp (evlist, event, ×tamp);if (err) {perf_mmap__consume (&md->core);tools/perf/builtin-top.c +1
-1@@ -873,7 +873,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)if (perf_mmap__read_init (&md->core) <0 )return ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {int ret; ret =perf_evlist__parse_sample_timestamp (evlist, event, &last_timestamp);tools/perf/builtin-trace.c +1
-1@@ -3454,7 +3454,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { ++trace->nr_events; err =trace__deliver_event (trace, event);tools/perf/lib/include/internal/mmap.h +0
-0
tools/perf/lib/include/internal/mmap.h line 14
Quoted 4 lines
tools/perf/lib/include/internal/mmap.h +1
-0@@ -11,6 +11,7 @@#define PERF_SAMPLE_MAX_SIZE (1 << 16)struct perf_mmap ; +union perf_event ;
Why are you adding this here?
Quoted 381 lines
tools/perf/lib/include/internal/mmap.h +0
-0typedef void (*libperf_unmap_cb_t )(struct perf_mmap *map);tools/perf/lib/include/perf/mmap.h +1
-0@@ -9,5 +9,6 @@ struct perf_mmap; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map); LIBPERF_API void perf_mmap__read_done(struct perf_mmap *map); +LIBPERF_API union perf_event *perf_mmap__read_event(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */ tools/perf/lib/libperf.map +1
-0@@ -43,6 +43,7 @@ LIBPERF_0.0.1 { perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done; + perf_mmap__read_event; local: *; };tools/perf/lib/mmap.c +79
-0@@ -3,9 +3,11 @@#include <inttypes.h> #include <asm/bug.h> #include <errno.h> +#include <string.h> #include <linux/ring_buffer.h> #include <linux/perf_event.h> #include <perf/mmap.h> +#include <perf/event.h> #include <internal/mmap.h> #include <internal/lib.h> #include "internal.h" @@ -191,3 +193,80 @@ void perf_mmap__read_done(struct perf_mmap *map) map->prev =perf_mmap__read_head (map); } + +/* When check_messup is true, 'end' must points to a good entry */ +static union perf_event *perf_mmap__read (struct perf_mmap *map, +u64 *startp,u64 end) +{ +unsigned char *data = map->base + page_size; +union perf_event *event =NULL ; +int diff = end - *startp; + +if (diff >= (int )sizeof (event->header)) { +size_t size; + + event = (union perf_event *)&data[*startp & map->mask]; + size = event->header.size; + +if (size <sizeof (event->header) || diff < (int )size) +return NULL ; + +/* +* Event straddles the mmap boundary -- header should always +* be inside due to u64 alignment of output. +*/ +if ((*startp & map->mask) + size != ((*startp + size) & map->mask)) { +unsigned int offset = *startp; +unsigned int len =min (sizeof (*event), size), cpy; +void *dst = map->event_copy; + +do { + cpy =min (map->mask +1 - (offset & map->mask), len); +memcpy (dst, &data[offset & map->mask], cpy); + offset += cpy; + dst += cpy; + len -= cpy; + }while (len); + + event = (union perf_event *)map->event_copy; + } + + *startp += size; + } + +return event; +} + +/* +* Read event from ring buffer one by one. +* Return one event for each call. +* +* Usage: +* perf_mmap__read_init() +* while(event = perf_mmap__read_event()) { +* //process the event +* perf_mmap__consume() +* } +* perf_mmap__read_done() +*/ +union perf_event *perf_mmap__read_event (struct perf_mmap *map) +{ +union perf_event *event; + +/* +* Check if event was unmapped due to a POLLHUP/POLLERR. +*/ +if (!refcount_read (&map->refcnt)) +return NULL ; + +/* non-overwirte doesn't pause the ringbuffer */ +if (!map->overwrite) + map->end =perf_mmap__read_head (map); + + event =perf_mmap__read (map, &map->start, map->end); + +if (!map->overwrite) + map->prev = map->start; + +return event; +}tools/perf/tests/backward-ring-buffer.c +1
-1@@ -39,7 +39,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,union perf_event *event;perf_mmap__read_init (&map->core); -while ((event =perf_mmap__read_event (map)) !=NULL ) { +while ((event =perf_mmap__read_event (&map->core)) !=NULL ) {const u32 type = event->header.type;switch (type) {tools/perf/tests/bpf.c +1
-1@@ -188,7 +188,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;if (type ==PERF_RECORD_SAMPLE )tools/perf/tests/code-reading.c +1
-1@@ -429,7 +429,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { ret =process_event (machine, evlist, event, state);perf_mmap__consume (&md->core);if (ret <0 )tools/perf/tests/keep-tracking.c +1
-1@@ -41,7 +41,7 @@ static int find_comm(struct evlist *evlist, const char *comm) md = &evlist->mmap[i];if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { if (event->header.type == PERF_RECORD_COMM && (pid_t)event->comm.pid == getpid() && (pid_t)event->comm.tid == getpid() &&tools/perf/tests/mmap-basic.c +1
-1@@ -117,7 +117,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuseif (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample;if (event->header.type !=PERF_RECORD_SAMPLE ) {tools/perf/tests/openat-syscall-tp-fields.c +1
-1@@ -96,7 +96,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtestif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;int tp_flags;struct perf_sample sample;tools/perf/tests/perf-record.c +1
-1@@ -174,7 +174,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;const char *name =perf_event__name (type);tools/perf/tests/sw-clock.c +1
-1@@ -103,7 +103,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)if (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample;if (event->header.type !=PERF_RECORD_SAMPLE )tools/perf/tests/switch-tracking.c +1
-1@@ -273,7 +273,7 @@ static int process_events(struct evlist *evlist,if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { cnt +=1 ; ret =add_event (evlist, &events, event);perf_mmap__consume (&md->core);tools/perf/tests/task-exit.c +1
-1@@ -121,7 +121,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unusedif (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {if (event->header.type ==PERF_RECORD_EXIT ) nr_exit++;tools/perf/util/evlist.c +1
-1@@ -1742,7 +1742,7 @@ static void *perf_evlist__poll_thread(void *arg)if (perf_mmap__read_init (&map->core))continue ; -while ((event =perf_mmap__read_event (map)) !=NULL ) { +while ((event =perf_mmap__read_event (&map->core)) !=NULL ) {struct evsel *evsel =perf_evlist__event2evsel (evlist, event);if (evsel && evsel->side_band.cb)tools/perf/util/mmap.c +0
-77@@ -29,83 +29,6 @@ size_t mmap__mmap_len(struct mmap *map)return perf_mmap__mmap_len (&map->core); } -/* When check_messup is true, 'end' must points to a good entry */ -static union perf_event *perf_mmap__read (struct mmap *map, -u64 *startp,u64 end) -{ -unsigned char *data = map->core.base + page_size; -union perf_event *event =NULL ; -int diff = end - *startp; - -if (diff >= (int )sizeof (event->header)) { -size_t size; - - event = (union perf_event *)&data[*startp & map->core.mask]; - size = event->header.size; - -if (size <sizeof (event->header) || diff < (int )size) -return NULL ; - -/* -* Event straddles the mmap boundary -- header should always -* be inside due to u64 alignment of output. -*/ -if ((*startp & map->core.mask) + size != ((*startp + size) & map->core.mask)) { -unsigned int offset = *startp; -unsigned int len =min (sizeof (*event), size), cpy; -void *dst = map->core.event_copy; - -do { - cpy =min (map->core.mask +1 - (offset & map->core.mask), len); -memcpy (dst, &data[offset & map->core.mask], cpy); - offset += cpy; - dst += cpy; - len -= cpy; - }while (len); - - event = (union perf_event *)map->core.event_copy; - } - - *startp += size; - } - -return event; -} - -/* -* Read event from ring buffer one by one. -* Return one event for each call. -* -* Usage: -* perf_mmap__read_init() -* while(event = perf_mmap__read_event()) { -* //process the event -* perf_mmap__consume() -* } -* perf_mmap__read_done() -*/ -union perf_event *perf_mmap__read_event (struct mmap *map) -{ -union perf_event *event; - -/* -* Check if event was unmapped due to a POLLHUP/POLLERR. -*/ -if (!refcount_read (&map->core.refcnt)) -return NULL ; - -/* non-overwirte doesn't pause the ringbuffer */ -if (!map->core.overwrite) - map->core.end =perf_mmap__read_head (&map->core); - - event =perf_mmap__read (map, &map->core.start, map->core.end); - -if (!map->core.overwrite) - map->core.prev = map->core.start; - -return event; -} - int __weak auxtrace_mmap__mmap(struct auxtrace_mmap *mm __maybe_unused, struct auxtrace_mmap_params *mp __maybe_unused, void *userpg __maybe_unused,tools/perf/util/mmap.h +0
-2@@ -47,8 +47,6 @@ void mmap__munmap(struct mmap *map);union perf_event *perf_mmap__read_forward (struct mmap *map); -union perf_event *perf_mmap__read_event (struct mmap *map); -int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size));tools/perf/util/python.c +1
-1@@ -1020,7 +1020,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist,if (perf_mmap__read_init (&md->core) <0 ) goto end; - event =perf_mmap__read_event (md); + event =perf_mmap__read_event (&md->core);if (event !=NULL ) {PyObject *pyevent =pyrf_event__new (event);struct pyrf_event *pevent = (struct pyrf_event *)pyevent;Signature · 1 line
2.21.0
Signature · 1 line
- Arnaldo
Re: [PATCH 12/36] libperf: Add perf_mmap__read_event() function
On Thu, Oct 10, 2019 at 11:47:00AM -0300, Arnaldo Carvalho de Melo wrote:
SNIP
Quoted 11 lines
tools/perf/lib/include/internal/mmap.h line 14
Quoted 9 lines
tools/perf/lib/include/internal/mmap.h +1
-0@@ -11,6 +11,7 @@#define PERF_SAMPLE_MAX_SIZE (1 << 16)struct perf_mmap ; +union perf_event ;Why are you adding this here?
oops, it should be in lib/include/perf/mmap.h ;-)
plz let me know if you want me to repost
thanks,
jirka
Re: [PATCH 27/36] libperf: Add perf_evlist__for_each_mmap function
In reply to Jiri OlsaEm Mon, Oct 07, 2019 at 02:53:35PM +0200, Jiri Olsa escreveu:
Quoted 86 lines
Adding
perf_evlist__for_each_mmapfunction and exporting it inperf/evlist.hheader, so user can iterate through 'structperf_mmap' objects.Adding internal
perf_mmap__linkfunction to do the actual linking.Signed-off-by Jiri Olsa <[email protected]>--- tools/perf/lib/evlist.c | 26 +++++++++++++++++++++++- tools/perf/lib/include/internal/evlist.h | 2 ++ tools/perf/lib/include/internal/mmap.h | 5 +++-- tools/perf/lib/include/perf/evlist.h | 9 ++++++++ tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 6 ++++-- tools/perf/util/evlist.c | 4 +++- 7 files changed, 47 insertions(+), 6 deletions(-)tools/perf/lib/evlist.c +25
-1@@ -347,6 +347,8 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, booreturn NULL ;for (i =0 ; i < evlist->nr_mmaps; i++) { +struct perf_mmap *prev = i ? &map[i -1 ] :NULL ; +/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last @@ -356,7 +358,7 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, boo* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i], overwrite,NULL ); +perf_mmap__init (&map[i], prev, overwrite,NULL ); }return map;@@ -405,6 +407,15 @@ perf_evlist__mmap_cb_mmap(struct perf_mmap *map, struct perf_mmap_param *mp,return perf_mmap__mmap (map, mp, output, cpu); } +static void perf_evlist__set_mmap0 (struct perf_evlist *evlist,struct perf_mmap *map, +bool overwrite) +{ +if (overwrite) + evlist->mmap_ovw0 = map; +else + evlist->mmap0 = map; +} + static int mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops, int idx, struct perf_mmap_param *mp, int cpu_idx,@@ -460,6 +471,9 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; + +if (!idx) +perf_evlist__set_mmap0 (evlist, map, overwrite); }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )return -1 ;@@ -605,3 +619,13 @@ void perf_evlist__munmap(struct perf_evlist *evlist)zfree (&evlist->mmap);zfree (&evlist->mmap_ovw); } + +struct perf_mmap * +perf_evlist__next_mmap (struct perf_evlist *evlist,struct perf_mmap *map, +bool overwrite) +{ +if (map) +return map->next; + +return overwrite ? evlist->mmap_ovw0 : evlist->mmap0; +}tools/perf/lib/include/internal/evlist.h +0
-0
tools/perf/lib/include/internal/evlist.h struct perf_evlist { · line 29
Quoted 5 lines
tools/perf/lib/include/internal/evlist.h +2
-0@@ -25,6 +25,8 @@ struct perf_evlist {struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ];struct perf_mmap *mmap;struct perf_mmap *mmap_ovw; +struct perf_mmap *mmap0; +struct perf_mmap *mmap_ovw0;
Why do you use this funny name? Using this "0" suffix? Please pick something more informative and add a comment on the struct perf_evlist definiton.
I've processed up to the previous patch, pushed to acme/tmp.perf/core.
- Arnaldo
Quoted 109 lines
tools/perf/lib/include/internal/evlist.h +0
-0}; typedef voidtools/perf/lib/include/internal/mmap.h +3
-2@@ -33,6 +33,7 @@ struct perf_mmap {u64 flush;libperf_unmap_cb_t unmap_cb;char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; +struct perf_mmap *next; };struct perf_mmap_param {@@ -42,8 +43,8 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map); -void perf_mmap__init (struct perf_mmap *map,bool overwrite, -libperf_unmap_cb_t unmap_cb); +void perf_mmap__init (struct perf_mmap *map,struct perf_mmap *prev, +bool overwrite,libperf_unmap_cb_t unmap_cb);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu);void perf_mmap__munmap (struct perf_mmap *map);tools/perf/lib/include/perf/evlist.h +9
-0@@ -3,6 +3,7 @@#define __LIBPERF_EVLIST_H#include <perf/core.h> +#include <stdbool.h> struct perf_evlist ;struct perf_evsel ;@@ -38,4 +39,12 @@ LIBPERF_API int perf_evlist__filter_pollfd(struct perf_evlist *evlist, LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist); +LIBPERF_API struct perf_mmap *perf_evlist__next_mmap(struct perf_evlist *evlist, +struct perf_mmap *map, +bool overwrite); +#define perf_evlist__for_each_mmap (evlist, pos, overwrite) \ + for ((pos) = perf_evlist__next_mmap((evlist), NULL, overwrite); \ + (pos) != NULL; \ + (pos) = perf_evlist__next_mmap((evlist), (pos), overwrite)) +#endif /* __LIBPERF_EVLIST_H */ tools/perf/lib/libperf.map +1
-0@@ -43,6 +43,7 @@ LIBPERF_0.0.1 { perf_evlist__mmap; perf_evlist__munmap; perf_evlist__filter_pollfd; + perf_evlist__next_mmap; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;tools/perf/lib/mmap.c +4
-2@@ -12,13 +12,15 @@#include <internal/lib.h> #include "internal.h" -void perf_mmap__init (struct perf_mmap *map,bool overwrite, -libperf_unmap_cb_t unmap_cb) +void perf_mmap__init (struct perf_mmap *map,struct perf_mmap *prev, +bool overwrite,libperf_unmap_cb_t unmap_cb) { map->fd =-1 ; map->overwrite = overwrite; map->unmap_cb = unmap_cb;refcount_set (&map->refcnt,0 ); +if (prev) + prev->next = map; }size_t perf_mmap__mmap_len (struct perf_mmap *map)tools/perf/util/evlist.c +3
-1@@ -582,6 +582,8 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return NULL ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { +struct perf_mmap *prev = i ? &map[i -1 ].core :NULL ; +/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last @@ -591,7 +593,7 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i].core, overwrite, perf_mmap__unmap_cb); +perf_mmap__init (&map[i].core, prev, overwrite, perf_mmap__unmap_cb); }return map;Signature · 1 line
2.21.0
Signature · 1 line
- Arnaldo
Re: [PATCH 12/36] libperf: Add perf_mmap__read_event() function
In reply to Jiri OlsaEm Thu, Oct 10, 2019 at 05:32:32PM +0200, Jiri Olsa escreveu:
Quoted 11 lines · Arnaldo Carvalho de Melo
SNIP
Quoted 7 lines
Quoted 7 lines
tools/perf/lib/include/internal/mmap.h +0
-0@@ -11,6 +11,7 @@#define PERF_SAMPLE_MAX_SIZE (1 << 16)
Quoted 7 lines
Quoted 4 lines
tools/perf/lib/include/internal/mmap.h line 14
Quoted 2 lines
tools/perf/lib/include/internal/mmap.h +1
-0struct perf_mmap ; +union perf_event ;Why are you adding this here?
oops, it should be in
lib/include/perf/mmap.h;-)
plz let me know if you want me to repost
I fixed it already, its in acme/tmp.perf/core, will continue later, possibly will push to Ingo what I have there if it passes all the tests, its getting too big a perf/core. :-)
- Arnaldo
Re: [PATCH 27/36] libperf: Add perf_evlist__for_each_mmap function
In reply to Arnaldo Carvalho de MeloOn Thu, Oct 10, 2019 at 01:06:22PM -0300, Arnaldo Carvalho de Melo wrote:
SNIP
Quoted 6 lines
Quoted 6 lines
tools/perf/lib/include/internal/evlist.h +0
-0@@ -25,6 +25,8 @@ struct perf_evlist {struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ];
Quoted 8 lines
tools/perf/lib/include/internal/evlist.h struct perf_evlist { · line 29
Quoted 4 lines
tools/perf/lib/include/internal/evlist.h +2
-0struct perf_mmap *mmap;struct perf_mmap *mmap_ovw; +struct perf_mmap *mmap0; +struct perf_mmap *mmap_ovw0;Why do you use this funny name? Using this "0" suffix? Please pick something more informative and add a comment on the struct
perf_evlistdefiniton.
it's the first mmap.. I can spell it ;-)
Quoted 2 lines
I've processed up to the previous patch, pushed to acme/tmp.perf/core.
great, I'll rebase and continue from there
thanks,
jirka
[tip: perf/core] libperf: Centralize map refcnt setting
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 285aaeac8c5d537b56b70169e21ac29ae5caa8e1 Gitweb: https://git.kernel.org/tip/285aaeac8c5d537b56b70169e21ac29ae5caa8e1 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:30 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:52:41 -03:00
libperf: Centralize map refcnt setting
Currently when a new map is mmapped we set its refcnt to 2 in the perf_evlist_mmap_ops::mmap callback.
Every mmap gets its refcnt set to 2 when it's first mmaped:
- 1 for the current user, which will be taken out by a call to perf_evlist__munmap_filtered(), where we find out there's no more data comming from kernel to this mmap.
- 1 for the drain code where in perf_mmap__consume() the mmap is released if it is empty.
Move this common setup into libperf's generic code before the mmap callback is called.
--- tools/perf/lib/evlist.c | 30 +++++++++++++++--------------- tools/perf/util/mmap.c | 15 --------------- 2 files changed, 15 insertions(+), 30 deletions(-)
tools/perf/lib/evlist.c +15 -15
@@ -362,21 +362,6 @@ static int perf_evlist__mmap_cb_mmap(struct perf_mmap *map, struct perf_mmap_param *mp, int output, int cpu) { -/* -* The last one will be done at perf_mmap__consume(), so that we -* make sure we don't prevent tools from consuming every last event in -* the ring buffer. -* -* I.e. we can get the POLLHUP meaning that the fd doesn't exist -* anymore, but the last events for it are still in the ring buffer, -* waiting to be consumed. -* -* Tools can chose to ignore this at their own discretion, but the -* evlist layer can't just drop it when filtering events in -* perf_evlist__filter_pollfd(). -*/ - refcount_set(&map->refcnt, 2); - return perf_mmap__mmap(map, mp, output, cpu); }
@@ -418,6 +403,21 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (*output ==-1 ) { *output = fd; +/* +* The last one will be done at perf_mmap__consume(), so that we +* make sure we don't prevent tools from consuming every last event in +* the ring buffer. +* +* I.e. we can get the POLLHUP meaning that the fd doesn't exist +* anymore, but the last events for it are still in the ring buffer, +* waiting to be consumed. +* +* Tools can chose to ignore this at their own discretion, but the +* evlist layer can't just drop it when filtering events in +* perf_evlist__filter_pollfd(). +*/ +refcount_set (&map->refcnt,2 ); +if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; }else {
tools/perf/util/mmap.c +0 -15
@@ -243,21 +243,6 @@ static void perf_mmap__setup_affinity_mask(struct mmap *map, struct mmap_paramsint mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu) { -/* -* The last one will be done at perf_mmap__consume(), so that we -* make sure we don't prevent tools from consuming every last event in -* the ring buffer. -* -* I.e. we can get the POLLHUP meaning that the fd doesn't exist -* anymore, but the last events for it are still in the ring buffer, -* waiting to be consumed. -* -* Tools can chose to ignore this at their own discretion, but the -* evlist layer can't just drop it when filtering events in -* perf_evlist__filter_pollfd(). -*/ -refcount_set (&map->core.refcnt,2 ); -if (perf_mmap__mmap (&map->core, &mp->core, fd, cpu)) {pr_debug2 ("failed to mmap perf event ring buffer, error %d\n" , errno);
[tip: perf/core] libperf: Introduce perf_evlist_mmap_ops::mmap callback
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: b5911e7ac28cb34f21b7380915ce98518078f114 Gitweb: https://git.kernel.org/tip/b5911e7ac28cb34f21b7380915ce98518078f114 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:25 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:22:21 -03:00
libperf: Introduce perf_evlist_mmap_ops::mmap callback
Add the perf_evlist_mmap_ops::mmap callback to be called in mmap_per_evsel() to actually mmap the map.
Add libperf's perf_evlist__mmap_cb_mmap() function as libperf's mmap callback.
New mmaped map gets refcount set to 2 in mmap__mmap(), we follow that in mmap callback. We will move this to common place after we switch to perf_evlist__mmap().
--- tools/perf/lib/evlist.c | 29 ++++++++++++++++++++--- tools/perf/lib/include/internal/evlist.h | 3 ++- 2 files changed, 29 insertions(+), 3 deletions(-)
tools/perf/lib/evlist.c +26 -3
@@ -359,6 +359,28 @@ perf_evlist__mmap_cb_get(struct perf_evlist *evlist, bool overwrite, int idx)#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y))static int +perf_evlist__mmap_cb_mmap (struct perf_mmap *map,struct perf_mmap_param *mp, +int output,int cpu) +{ +/* +* The last one will be done at perf_mmap__consume(), so that we +* make sure we don't prevent tools from consuming every last event in +* the ring buffer. +* +* I.e. we can get the POLLHUP meaning that the fd doesn't exist +* anymore, but the last events for it are still in the ring buffer, +* waiting to be consumed. +* +* Tools can chose to ignore this at their own discretion, but the +* evlist layer can't just drop it when filtering events in +* perf_evlist__filter_pollfd(). +*/ +refcount_set (&map->refcnt,2 ); + +return perf_mmap__mmap (map, mp, output, cpu); +} + +static int mmap_per_evsel (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops,int idx,struct perf_mmap_param *mp,int cpu_idx,int thread,int *_output,int *_output_overwrite)
@@ -396,7 +418,7 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (*output ==-1 ) { *output = fd; -if (perf_mmap__mmap (map, mp, *output, evlist_cpu) <0 ) +if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )
@@ -488,7 +510,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,const struct perf_cpu_map *cpus = evlist->cpus;const struct perf_thread_map *threads = evlist->threads; -if (!ops || !ops->get) +if (!ops || !ops->get || !ops->mmap)return -EINVAL ;if (!evlist->mmap)
@@ -513,7 +535,8 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages) {struct perf_mmap_param mp;struct perf_evlist_mmap_ops ops = { - .get = perf_evlist__mmap_cb_get, + .get = perf_evlist__mmap_cb_get, + .mmap = perf_evlist__mmap_cb_mmap, }; evlist->mmap_len = (pages +1 ) * page_size;
tools/perf/lib/include/internal/evlist.h +3 -0
@@ -31,10 +31,13 @@ typedef void (*perf_evlist_mmap__cb_idx_t)(struct perf_evlist*, struct perf_mmap_param*, int, bool);typedef struct perf_mmap * (*perf_evlist_mmap__cb_get_t )(struct perf_evlist *,bool ,int ); +typedef int +(*perf_evlist_mmap__cb_mmap_t )(struct perf_mmap *,struct perf_mmap_param *,int ,int );struct perf_evlist_mmap_ops {perf_evlist_mmap__cb_idx_t idx;perf_evlist_mmap__cb_get_t get; +perf_evlist_mmap__cb_mmap_t mmap; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
[tip: perf/core] perf tools: Introduce perf_evlist__mmap_cb_idx()
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 9abd2ab2377613425e1c362694f85b110f5bace2 Gitweb: https://git.kernel.org/tip/9abd2ab2377613425e1c362694f85b110f5bace2 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:26 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:23:52 -03:00
perf tools: Introduce perf_evlist__mmap_cb_idx()
Add perf_evlist__mmap_cb_idx function to call auxtrace_mmap_params__set_idx() on each new index during perf_evlist__mmap_ops call.
--- tools/perf/util/evlist.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
tools/perf/util/evlist.c +14 -0
@@ -739,6 +739,17 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,return 0 ; } +static void +perf_evlist__mmap_cb_idx (struct perf_evlist *_evlist, +struct perf_mmap_param *_mp, +int idx,bool per_cpu) +{ + struct evlist *evlist = container_of(_evlist, struct evlist, core); + struct mmap_params *mp = container_of(_mp, struct mmap_params, core); + +auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, idx, per_cpu); +} +static int evlist__mmap_per_cpu (struct evlist *evlist,struct mmap_params *mp) {
@@ -935,6 +946,9 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .flush = flush, .comp_level = comp_level }; +struct perf_evlist_mmap_ops ops__maybe_unused = { + .idx = perf_evlist__mmap_cb_idx, + };if (!evlist->mmap) evlist->mmap =evlist__alloc_mmap (evlist, false);
[tip: perf/core] libperf: Adopt perf_mmap__put() function from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 80e53d1148231d7d4fdc4cd89e5393616b33bf82 Gitweb: https://git.kernel.org/tip/80e53d1148231d7d4fdc4cd89e5393616b33bf82 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:15 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 10:09:25 -03:00
libperf: Adopt perf_mmap__put() function from tools/perf
Move perf_mmap__put() from tools/perf to libperf.
Once perf_mmap__put() is moved, we need a way to call application related unmap code (AIO and aux related code for eprf), when the map goes away.
Add the perf_mmap::unmap callback to do that.
The unmap path from perf is:
perf_mmap__put (libperf)
perf_mmap__munmap (libperf)
map->unmap_cb -> perf_mmap__unmap_cb (perf)
mmap__munmap (perf)Committer notes:
Add missing linux/kernel.h to tools/perf/lib/mmap.c to get the BUG_ON definition.
--- tools/perf/builtin-record.c | 4 +-- tools/perf/lib/include/internal/mmap.h | 31 +++++++++++++++---------- tools/perf/lib/mmap.c | 15 +++++++++++- tools/perf/util/evlist.c | 17 +++++++++----- tools/perf/util/mmap.c | 11 +--------- tools/perf/util/mmap.h | 2 +-- 6 files changed, 48 insertions(+), 32 deletions(-)
tools/perf/builtin-record.c +2 -2
@@ -197,7 +197,7 @@ static int record__aio_complete(struct mmap *md, struct aiocb *cblock)* every aio write request started in record__aio_push() so * decrement it because the request is now complete. */ -perf_mmap__put (md); +perf_mmap__put (&md->core); rc =1 ; }else {/*
@@ -332,7 +332,7 @@ static int record__aio_push(struct record *rec, struct mmap *map, off_t *off)* map->refcount is decremented in record__aio_complete() after * aio write operation finishes successfully. */ -perf_mmap__put (map); +perf_mmap__put (&map->core); }return ret;
tools/perf/lib/include/internal/mmap.h +19 -12
@@ -10,23 +10,28 @@/* perf sample has 16 bits size limit */ #define PERF_SAMPLE_MAX_SIZE (1 << 16) +struct perf_mmap ; + +typedef void (*libperf_unmap_cb_t )(struct perf_mmap *map); +/** * struct perf_mmap - perf's ring buffer mmap details * * @refcnt - e.g. code using PERF_EVENT_IOC_SET_OUTPUT to share this */ struct perf_mmap { -void *base; -int mask; -int fd; -int cpu; -refcount_t refcnt; -u64 prev; -u64 start; -u64 end; -bool overwrite; -u64 flush; -char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; +void *base; +int mask; +int fd; +int cpu; +refcount_t refcnt; +u64 prev; +u64 start; +u64 end; +bool overwrite; +u64 flush; +libperf_unmap_cb_t unmap_cb; +char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; };struct perf_mmap_param {
@@ -36,10 +41,12 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map); -void perf_mmap__init (struct perf_mmap *map,bool overwrite); +void perf_mmap__init (struct perf_mmap *map,bool overwrite, +libperf_unmap_cb_t unmap_cb);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu);void perf_mmap__munmap (struct perf_mmap *map);void perf_mmap__get (struct perf_mmap *map); +void perf_mmap__put (struct perf_mmap *map);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +14 -1
@@ -2,11 +2,14 @@#include <sys/mman.h> #include <internal/mmap.h> #include <internal/lib.h> +#include <linux/kernel.h> -void perf_mmap__init (struct perf_mmap *map,bool overwrite) +void perf_mmap__init (struct perf_mmap *map,bool overwrite, +libperf_unmap_cb_t unmap_cb) { map->fd =-1 ; map->overwrite = overwrite; + map->unmap_cb = unmap_cb;refcount_set (&map->refcnt,0 ); }
@@ -40,9 +43,19 @@ void perf_mmap__munmap(struct perf_mmap *map) map->fd =-1 ;refcount_set (&map->refcnt,0 ); } +if (map && map->unmap_cb) + map->unmap_cb (map); }void perf_mmap__get (struct perf_mmap *map) {refcount_inc (&map->refcnt); } + +void perf_mmap__put (struct perf_mmap *map) +{ +BUG_ON (map->base &&refcount_read (&map->refcnt) ==0 ); + +if (refcount_dec_and_test (&map->refcnt)) +perf_mmap__munmap (map); +}
tools/perf/util/evlist.c +12 -5
@@ -433,7 +433,7 @@ static void perf_evlist__munmap_filtered(struct fdarray *fda, int fd,struct mmap *map = fda->priv[fd].ptr;if (map) -perf_mmap__put (map); +perf_mmap__put (&map->core); }int evlist__filter_pollfd (struct evlist *evlist,short revents_and_mask)
@@ -601,11 +601,11 @@ static void evlist__munmap_nofree(struct evlist *evlist)if (evlist->mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -mmap__munmap (&evlist->mmap[i]); +perf_mmap__munmap (&evlist->mmap[i].core);if (evlist->overwrite_mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -mmap__munmap (&evlist->overwrite_mmap[i]); +perf_mmap__munmap (&evlist->overwrite_mmap[i].core); }void evlist__munmap (struct evlist *evlist)
@@ -615,6 +615,13 @@ void evlist__munmap(struct evlist *evlist)zfree (&evlist->overwrite_mmap); } +static void perf_mmap__unmap_cb (struct perf_mmap *map) +{ + struct mmap *m = container_of(map, struct mmap, core); + +mmap__munmap (m); +} +static struct mmap *evlist__alloc_mmap (struct evlist *evlist,bool overwrite) {
@@ -638,7 +645,7 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i].core, overwrite); +perf_mmap__init (&map[i].core, overwrite, perf_mmap__unmap_cb); }return map;
@@ -715,7 +722,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,*/ if (!evsel->core.system_wide &&perf_evlist__add_pollfd (&evlist->core, fd, &maps[idx], revent) <0 ) { -perf_mmap__put (&maps[idx]); +perf_mmap__put (&maps[idx].core);return -1 ; }
tools/perf/util/mmap.c +1 -10
@@ -110,14 +110,6 @@ static bool perf_mmap__empty(struct mmap *map)return perf_mmap__read_head (map) == map->core.prev && !map->auxtrace_mmap.base; } -void perf_mmap__put (struct mmap *map) -{ -BUG_ON (map->core.base &&refcount_read (&map->core.refcnt) ==0 ); - -if (refcount_dec_and_test (&map->core.refcnt)) -mmap__munmap (map); -} -void perf_mmap__consume (struct mmap *map) {if (!map->core.overwrite) {
@@ -127,7 +119,7 @@ void perf_mmap__consume(struct mmap *map) }if (refcount_read (&map->core.refcnt) ==1 &&perf_mmap__empty (map)) -perf_mmap__put (map); +perf_mmap__put (&map->core); } int __weak auxtrace_mmap__mmap(struct auxtrace_mmap *mm __maybe_unused,
@@ -308,7 +300,6 @@ static void perf_mmap__aio_munmap(struct mmap *map __maybe_unused)void mmap__munmap (struct mmap *map) { -perf_mmap__munmap (&map->core);perf_mmap__aio_munmap (map);if (map->data !=NULL ) {munmap (map->data,mmap__mmap_len (map));
tools/perf/util/mmap.h +0 -2
@@ -45,8 +45,6 @@ struct mmap_params {int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void mmap__munmap (struct mmap *map); -void perf_mmap__put (struct mmap *map); -void perf_mmap__consume (struct mmap *map);static inline u64 perf_mmap__read_head (struct mmap *mm)
[tip: perf/core] libperf: Add 'struct perf_mmap_param'
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: e440979faf6ac8048e1792af383df6af78dd1cb0 Gitweb: https://git.kernel.org/tip/e440979faf6ac8048e1792af383df6af78dd1cb0 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:10 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 09:40:00 -03:00
libperf: Add 'struct perf_mmap_param'
Add libperf's version of mmap params 'struct perf_mmap_param' object with the basics: 'prot' and 'mask'. Encapsulate it in the current 'struct mmap_params' object.
--- tools/perf/lib/include/internal/mmap.h | 5 +++++ tools/perf/util/evlist.c | 14 +++++++++----- tools/perf/util/mmap.c | 4 ++-- tools/perf/util/mmap.h | 3 ++- 4 files changed, 18 insertions(+), 8 deletions(-)
tools/perf/lib/include/internal/mmap.h +5 -0
@@ -29,6 +29,11 @@ struct perf_mmap { char event_copy[PERF_SAMPLE_MAX_SIZE] __aligned(8); }; +struct perf_mmap_param { +int prot; +int mask; +}; +void perf_mmap__init (struct perf_mmap *map,bool overwrite);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/util/evlist.c +9 -5
@@ -667,7 +667,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,int fd;int cpu; - mp->prot =PROT_READ |PROT_WRITE ; + mp->core.prot =PROT_READ |PROT_WRITE ;if (evsel->core.attr.write_backward) { output = _output_overwrite; maps = evlist->overwrite_mmap;
@@ -680,7 +680,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY )perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); } - mp->prot &= ~PROT_WRITE ; + mp->core.prot &= ~PROT_WRITE ; }if (evsel->core.system_wide && thread)
@@ -921,8 +921,12 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages,* Its value is decided by evsel's write_backward. * So &mp should not be passed through const pointer. */ -struct mmap_params mp = { .nr_cblocks = nr_cblocks, .affinity = affinity, .flush = flush, - .comp_level = comp_level }; +struct mmap_params mp = { + .nr_cblocks = nr_cblocks, + .affinity = affinity, + .flush = flush, + .comp_level = comp_level + };if (!evlist->mmap) evlist->mmap =evlist__alloc_mmap (evlist, false);
@@ -934,7 +938,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); - mp.mask = evlist->core.mmap_len - page_size -1 ; + mp.core.mask = evlist->core.mmap_len - page_size -1 ;auxtrace_mmap_params__init (&mp.auxtrace_mp, evlist->core.mmap_len, auxtrace_pages, auxtrace_overwrite);
tools/perf/util/mmap.c +2 -2
@@ -370,8 +370,8 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu)*/ refcount_set (&map->core.refcnt,2 ); map->core.prev =0 ; - map->core.mask = mp->mask; - map->core.base =mmap (NULL ,perf_mmap__mmap_len (map), mp->prot, + map->core.mask = mp->core.mask; + map->core.base =mmap (NULL ,perf_mmap__mmap_len (map), mp->core.prot,MAP_SHARED , fd,0 );if (map->core.base ==MAP_FAILED ) { pr_debug2("failed to mmap perf event ring buffer, error %d\n",
tools/perf/util/mmap.h +2 -1
@@ -37,7 +37,8 @@ struct mmap { };struct mmap_params { -int prot, mask, nr_cblocks, affinity, flush, comp_level; +struct perf_mmap_param core; +int nr_cblocks, affinity, flush, comp_level;struct auxtrace_mmap_params auxtrace_mp; };
[tip: perf/core] libperf: Adopt perf_mmap__read_done() from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 32fdc2ca7e2ae8ae5d0ff660ca7783acd8ee6396 Gitweb: https://git.kernel.org/tip/32fdc2ca7e2ae8ae5d0ff660ca7783acd8ee6396 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:19 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 11:45:32 -03:00
libperf: Adopt perf_mmap__read_done() from tools/perf
Move perf_mmap__read_init() from tools/perf to libperf and export it in the perf/mmap.h header.
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/perf/mmap.h | 1 + tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 17 +++++++++++++++++ tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 17 ----------------- tools/perf/util/mmap.h | 1 - 20 files changed, 34 insertions(+), 33 deletions(-)
tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 -1
@@ -142,7 +142,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybe next_event:perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (!comm1_time || !comm2_time)
tools/perf/builtin-kvm.c +1 -1
@@ -794,7 +794,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,break ; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core);return n; }
tools/perf/builtin-top.c +1 -1
@@ -894,7 +894,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx) } } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }static void perf_top__mmap_read (struct perf_top *top)
tools/perf/builtin-trace.c +1 -1
@@ -3821,7 +3821,7 @@ again: draining = true; } } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (trace->nr_events == before) {
tools/perf/lib/include/perf/mmap.h +1 -0
@@ -8,5 +8,6 @@ struct perf_mmap; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map); +LIBPERF_API void perf_mmap__read_done(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/libperf.map +1 -0
@@ -42,6 +42,7 @@ LIBPERF_0.0.1 { perf_evlist__poll; perf_mmap__consume; perf_mmap__read_init; + perf_mmap__read_done; local: *; };
tools/perf/lib/mmap.c +17 -0
@@ -175,3 +175,20 @@ int perf_mmap__read_init(struct perf_mmap *map)return __perf_mmap__read_init (map); } + +/* +* Mandatory for overwrite mode +* The direction of overwrite mode is backward. +* The last perf_mmap__read() will set tail to map->core.prev. +* Need to correct the map->core.prev to head which is the end of next read. +*/ +void perf_mmap__read_done (struct perf_mmap *map) +{ +/* +* Check if event was unmapped due to a POLLHUP/POLLERR. +*/ +if (!refcount_read (&map->refcnt)) +return ; + + map->prev =perf_mmap__read_head (map); +}
tools/perf/tests/backward-ring-buffer.c +1 -1
@@ -54,7 +54,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,return TEST_FAIL ; } } -perf_mmap__read_done (map); +perf_mmap__read_done (&map->core); }return TEST_OK ; }
tools/perf/tests/bpf.c +1 -1
@@ -194,7 +194,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),if (type ==PERF_RECORD_SAMPLE ) count ++; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (count != expect) {
tools/perf/tests/code-reading.c +1 -1
@@ -435,7 +435,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,if (ret <0 )return ret; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }return 0 ; }
tools/perf/tests/keep-tracking.c +1 -1
@@ -49,7 +49,7 @@ static int find_comm(struct evlist *evlist, const char *comm) found +=1 ;perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }return found; }
tools/perf/tests/mmap-basic.c +1 -1
@@ -142,7 +142,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuse nr_events[evsel->idx]++;perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); out_init: err =0 ;
tools/perf/tests/openat-syscall-tp-fields.c +1 -1
@@ -124,7 +124,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtest goto out_ok; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }if (nr_events == before)
tools/perf/tests/perf-record.c +1 -1
@@ -279,7 +279,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusperf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); }/*
tools/perf/tests/sw-clock.c +1 -1
@@ -120,7 +120,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id) next_event:perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); out_init:if ((u64 ) nr_samples == total_periods) {
tools/perf/tests/switch-tracking.c +1 -1
@@ -280,7 +280,7 @@ static int process_events(struct evlist *evlist,if (ret <0 ) goto out_free_nodes; } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); } events_array =calloc (cnt,sizeof (struct event_node ));
tools/perf/tests/task-exit.c +1 -1
@@ -127,7 +127,7 @@ retry:perf_mmap__consume (&md->core); } -perf_mmap__read_done (md); +perf_mmap__read_done (&md->core); out_init:if (!exited || !nr_exit) {
tools/perf/util/evlist.c +1 -1
@@ -1822,7 +1822,7 @@ static void *perf_evlist__poll_thread(void *arg)perf_mmap__consume (&map->core); got_data = true; } -perf_mmap__read_done (map); +perf_mmap__read_done (&map->core); }if (draining && !got_data)
tools/perf/util/mmap.c +0 -17
@@ -405,20 +405,3 @@ int perf_mmap__push(struct mmap *md, void *to, out:return rc; } - -/* -* Mandatory for overwrite mode -* The direction of overwrite mode is backward. -* The last perf_mmap__read() will set tail to map->core.prev. -* Need to correct the map->core.prev to head which is the end of next read. -*/ -void perf_mmap__read_done (struct mmap *map) -{ -/* -* Check if event was unmapped due to a POLLHUP/POLLERR. -*/ -if (!refcount_read (&map->core.refcnt)) -return ; - - map->core.prev =perf_mmap__read_head (&map->core); -}
tools/perf/util/mmap.h +0 -1
@@ -54,5 +54,4 @@ int perf_mmap__push(struct mmap *md, void *to,size_t mmap__mmap_len (struct mmap *map); -void perf_mmap__read_done (struct mmap *map);#endif /*__PERF_MMAP_H */
[tip: perf/core] libperf: Adopt perf_evlist__filter_pollfd() from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 84227cb11ff4d9815b9b1daf0c1a2bd7e9274c58 Gitweb: https://git.kernel.org/tip/84227cb11ff4d9815b9b1daf0c1a2bd7e9274c58 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:34 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:58:45 -03:00
libperf: Adopt perf_evlist__filter_pollfd() from tools/perf
Introduce the perf_evlist__filter_pollfd function and export it in the perf/evlist.h header, so that libperf users can check if the descriptor is still alive.
--- tools/perf/lib/evlist.c | 15 +++++++++++++++ tools/perf/lib/include/perf/evlist.h | 2 ++ tools/perf/lib/libperf.map | 1 + tools/perf/util/evlist.c | 12 +----------- 4 files changed, 19 insertions(+), 11 deletions(-)
tools/perf/lib/evlist.c +15 -0
@@ -313,6 +313,21 @@ int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd,return pos; } +static void perf_evlist__munmap_filtered (struct fdarray *fda,int fd, +void *arg__maybe_unused ) +{ +struct perf_mmap *map = fda->priv[fd].ptr; + +if (map) +perf_mmap__put (map); +} + +int perf_evlist__filter_pollfd (struct perf_evlist *evlist,short revents_and_mask) +{ +return fdarray__filter (&evlist->pollfd, revents_and_mask, + perf_evlist__munmap_filtered,NULL ); +} +int perf_evlist__poll (struct perf_evlist *evlist,int timeout) {return fdarray__poll (&evlist->pollfd, timeout);
tools/perf/lib/include/perf/evlist.h +2 -0
@@ -32,6 +32,8 @@ LIBPERF_API void perf_evlist__set_maps(struct perf_evlist *evlist,struct perf_cpu_map *cpus, struct perf_thread_map *threads); LIBPERF_API int perf_evlist__poll(struct perf_evlist *evlist, int timeout); +LIBPERF_API int perf_evlist__filter_pollfd(struct perf_evlist *evlist, +short revents_and_mask); LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist);
tools/perf/lib/libperf.map +1 -0
@@ -42,6 +42,7 @@ LIBPERF_0.0.1 { perf_evlist__poll; perf_evlist__mmap; perf_evlist__munmap; + perf_evlist__filter_pollfd; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;
tools/perf/util/evlist.c +1 -11
@@ -423,19 +423,9 @@ int evlist__add_pollfd(struct evlist *evlist, int fd)return perf_evlist__add_pollfd (&evlist->core, fd,NULL ,POLLIN ); } -static void perf_evlist__munmap_filtered (struct fdarray *fda,int fd, -void *arg__maybe_unused ) -{ -struct perf_mmap *map = fda->priv[fd].ptr; - -if (map) -perf_mmap__put (map); -} -int evlist__filter_pollfd (struct evlist *evlist,short revents_and_mask) { -return fdarray__filter (&evlist->core.pollfd, revents_and_mask, - perf_evlist__munmap_filtered,NULL ); +return perf_evlist__filter_pollfd (&evlist->core, revents_and_mask); }int evlist__poll (struct evlist *evlist,int timeout)
[tip: perf/core] libperf: Move the pollfd allocation from tools/perf to libperf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 230662e15ed6cc63ecf72ed1bffa3cadef486850 Gitweb: https://git.kernel.org/tip/230662e15ed6cc63ecf72ed1bffa3cadef486850 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:31 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:54:35 -03:00
libperf: Move the pollfd allocation from tools/perf to libperf
It's needed in libperf only, so move it to the perf_evlist__mmap_ops() function.
--- tools/perf/lib/evlist.c | 5 +++++ tools/perf/util/evlist.c | 4 ---- 2 files changed, 5 insertions(+), 4 deletions(-)
tools/perf/lib/evlist.c +5 -0
@@ -34,6 +34,7 @@ void perf_evlist__init(struct perf_evlist *evlist)INIT_HLIST_HEAD (&evlist->heads[i]);INIT_LIST_HEAD (&evlist->entries); evlist->nr_entries =0 ; +fdarray__init (&evlist->pollfd,64 ); } static void __perf_evlist__propagate_maps(struct perf_evlist *evlist,
@@ -114,6 +115,7 @@ void perf_evlist__delete(struct perf_evlist *evlist)return ;perf_evlist__munmap (evlist); +fdarray__exit (&evlist->pollfd);free (evlist); }
@@ -525,6 +527,9 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,return -ENOMEM ; } +if (evlist->pollfd.entries ==NULL &&perf_evlist__alloc_pollfd (evlist) <0 ) +return -ENOMEM ; +if (perf_cpu_map__empty (cpus))return mmap_per_thread (evlist, ops, mp);
tools/perf/util/evlist.c +0 -4
@@ -58,7 +58,6 @@ void evlist__init(struct evlist *evlist, struct perf_cpu_map *cpus, {perf_evlist__init (&evlist->core);perf_evlist__set_maps (&evlist->core, cpus, threads); -fdarray__init (&evlist->core.pollfd,64 ); evlist->workload.pid =-1 ; evlist->bkw_mmap_state =BKW_MMAP_NOTREADY ; }
@@ -829,9 +828,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages,if (!evlist->mmap)return -ENOMEM ; -if (evlist->core.pollfd.entries ==NULL &&perf_evlist__alloc_pollfd (&evlist->core) <0 ) -return -ENOMEM ; - evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); mp.core.mask = evlist->core.mmap_len - page_size -1 ;
[tip: perf/core] libperf: Adopt perf_mmap__mmap_len() function from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: bf59b3053e63783520c2810fc3f676553bc7eedd Gitweb: https://git.kernel.org/tip/bf59b3053e63783520c2810fc3f676553bc7eedd Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:11 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 09:41:38 -03:00
libperf: Adopt perf_mmap__mmap_len() function from tools/perf
Move perf_mmap__mmap_len() from tools/perf wto libperf, it will be used in the following patches. And rename the existing perf's function to mmap__mmap_len().
--- tools/perf/builtin-record.c | 4 ++-- tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/mmap.c | 6 ++++++ tools/perf/util/mmap.c | 20 ++++++++++---------- tools/perf/util/mmap.h | 2 +- 5 files changed, 21 insertions(+), 13 deletions(-)
tools/perf/builtin-record.c +2 -2
@@ -276,7 +276,7 @@ static int record__aio_pushfn(struct mmap *map, void *to, void *buf, size_t sizeif (record__comp_enabled (aio->rec)) { size =zstd_compress (aio->rec->session, aio->data + aio->size, -perf_mmap__mmap_len (map) - aio->size, +mmap__mmap_len (map) - aio->size, buf, size); }else {memcpy (aio->data + aio->size, buf, size);
@@ -488,7 +488,7 @@ static int record__pushfn(struct mmap *map, void *to, void *bf, size_t size)struct record *rec = to;if (record__comp_enabled (rec)) { - size =zstd_compress (rec->session, map->data,perf_mmap__mmap_len (map), bf, size); + size =zstd_compress (rec->session, map->data,mmap__mmap_len (map), bf, size); bf = map->data; }
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -34,6 +34,8 @@ struct perf_mmap_param { int mask; }; +size_t perf_mmap__mmap_len (struct perf_mmap *map); +void perf_mmap__init (struct perf_mmap *map,bool overwrite);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +6 -0
@@ -1,5 +1,6 @@// SPDX-License-Identifier: GPL-2.0 #include <internal/mmap.h> +#include <internal/lib.h> void perf_mmap__init (struct perf_mmap *map,bool overwrite) {
@@ -7,3 +8,8 @@ void perf_mmap__init(struct perf_mmap *map, bool overwrite) map->overwrite = overwrite;refcount_set (&map->refcnt,0 ); } + +size_t perf_mmap__mmap_len (struct perf_mmap *map) +{ +return map->mask +1 + page_size; +}
tools/perf/util/mmap.c +10 -10
@@ -23,9 +23,9 @@#include "../perf.h" #include <internal/lib.h> /* page_size */ -size_t perf_mmap__mmap_len (struct mmap *map) +size_t mmap__mmap_len (struct mmap *map) { -return map->core.mask +1 + page_size; +return perf_mmap__mmap_len (&map->core); }/* When check_messup is true, 'end' must points to a good entry */
@@ -170,7 +170,7 @@ static int perf_mmap__aio_enabled(struct mmap *map)#ifdef HAVE_LIBNUMA_SUPPORT static int perf_mmap__aio_alloc (struct mmap *map,int idx) { - map->aio.data[idx] =mmap (NULL ,perf_mmap__mmap_len (map),PROT_READ |PROT_WRITE , + map->aio.data[idx] =mmap (NULL ,mmap__mmap_len (map),PROT_READ |PROT_WRITE ,MAP_PRIVATE |MAP_ANONYMOUS ,0 ,0 );if (map->aio.data[idx] ==MAP_FAILED ) { map->aio.data[idx] =NULL ;
@@ -183,7 +183,7 @@ static int perf_mmap__aio_alloc(struct mmap *map, int idx)static void perf_mmap__aio_free (struct mmap *map,int idx) {if (map->aio.data[idx]) { -munmap (map->aio.data[idx],perf_mmap__mmap_len (map)); +munmap (map->aio.data[idx],mmap__mmap_len (map)); map->aio.data[idx] =NULL ; } }
@@ -196,7 +196,7 @@ static int perf_mmap__aio_bind(struct mmap *map, int idx, int cpu, int affinity)if (affinity !=PERF_AFFINITY_SYS &&cpu__max_node () >1 ) { data = map->aio.data[idx]; - mmap_len =perf_mmap__mmap_len (map); + mmap_len =mmap__mmap_len (map); node_mask =1UL <<cpu__get_node (cpu);if (mbind (data, mmap_len,MPOL_BIND , &node_mask,1 ,0 )) { pr_err("Failed to bind [%p-%p] AIO buffer to node %d: error %m\n",
@@ -210,7 +210,7 @@ static int perf_mmap__aio_bind(struct mmap *map, int idx, int cpu, int affinity)#else /* !HAVE_LIBNUMA_SUPPORT */ static int perf_mmap__aio_alloc (struct mmap *map,int idx) { - map->aio.data[idx] =malloc (perf_mmap__mmap_len (map)); + map->aio.data[idx] =malloc (mmap__mmap_len (map));if (map->aio.data[idx] ==NULL )return -1 ;
@@ -315,11 +315,11 @@ void perf_mmap__munmap(struct mmap *map) {perf_mmap__aio_munmap (map);if (map->data !=NULL ) { -munmap (map->data,perf_mmap__mmap_len (map)); +munmap (map->data,mmap__mmap_len (map)); map->data =NULL ; }if (map->core.base !=NULL ) { -munmap (map->core.base,perf_mmap__mmap_len (map)); +munmap (map->core.base,mmap__mmap_len (map)); map->core.base =NULL ; map->core.fd =-1 ;refcount_set (&map->core.refcnt,0 );
@@ -371,7 +371,7 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu)refcount_set (&map->core.refcnt,2 ); map->core.prev =0 ; map->core.mask = mp->core.mask; - map->core.base =mmap (NULL ,perf_mmap__mmap_len (map), mp->core.prot, + map->core.base =mmap (NULL ,mmap__mmap_len (map), mp->core.prot,MAP_SHARED , fd,0 );if (map->core.base ==MAP_FAILED ) { pr_debug2("failed to mmap perf event ring buffer, error %d\n",
@@ -389,7 +389,7 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu) map->comp_level = mp->comp_level;if (map->comp_level && !perf_mmap__aio_enabled (map)) { - map->data =mmap (NULL ,perf_mmap__mmap_len (map),PROT_READ |PROT_WRITE , + map->data =mmap (NULL ,mmap__mmap_len (map),PROT_READ |PROT_WRITE ,MAP_PRIVATE |MAP_ANONYMOUS ,0 ,0 );if (map->data ==MAP_FAILED ) { pr_debug2("failed to mmap data buffer, error %d\n",
tools/perf/util/mmap.h +1 -1
@@ -67,7 +67,7 @@ union perf_event *perf_mmap__read_event(struct mmap *map);int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size)); -size_t perf_mmap__mmap_len (struct mmap *map); +size_t mmap__mmap_len (struct mmap *map);int perf_mmap__read_init (struct mmap *md);void perf_mmap__read_done (struct mmap *map);
[tip: perf/core] libperf: Adopt perf_mmap__unmap() function from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 59d7ea620b58fa7d107834a81528e3098f1c27b0 Gitweb: https://git.kernel.org/tip/59d7ea620b58fa7d107834a81528e3098f1c27b0 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:14 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 10:05:57 -03:00
libperf: Adopt perf_mmap__unmap() function from tools/perf
Move perf_mmap__unmap() from tools/perf to libperf, to internal header internal/mmap.h. It will be used in the following patches. And rename the existing perf's function to mmap__munmap().
--- tools/perf/lib/include/internal/mmap.h | 1 + tools/perf/lib/mmap.c | 10 ++++++++++ tools/perf/util/evlist.c | 4 ++-- tools/perf/util/mmap.c | 11 +++-------- tools/perf/util/mmap.h | 2 +- 5 files changed, 17 insertions(+), 11 deletions(-)
tools/perf/lib/include/internal/mmap.h +1 -0
@@ -39,6 +39,7 @@ size_t perf_mmap__mmap_len(struct perf_mmap *map);void perf_mmap__init (struct perf_mmap *map,bool overwrite);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu); +void perf_mmap__munmap (struct perf_mmap *map);void perf_mmap__get (struct perf_mmap *map);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +10 -0
@@ -32,6 +32,16 @@ int perf_mmap__mmap(struct perf_mmap *map, struct perf_mmap_param *mp,return 0 ; } +void perf_mmap__munmap (struct perf_mmap *map) +{ +if (map && map->base !=NULL ) { +munmap (map->base,perf_mmap__mmap_len (map)); + map->base =NULL ; + map->fd =-1 ; +refcount_set (&map->refcnt,0 ); + } +} +void perf_mmap__get (struct perf_mmap *map) {refcount_inc (&map->refcnt);
tools/perf/util/evlist.c +2 -2
@@ -601,11 +601,11 @@ static void evlist__munmap_nofree(struct evlist *evlist)if (evlist->mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -perf_mmap__munmap (&evlist->mmap[i]); +mmap__munmap (&evlist->mmap[i]);if (evlist->overwrite_mmap)for (i =0 ; i < evlist->core.nr_mmaps; i++) -perf_mmap__munmap (&evlist->overwrite_mmap[i]); +mmap__munmap (&evlist->overwrite_mmap[i]); }void evlist__munmap (struct evlist *evlist)
tools/perf/util/mmap.c +3 -8
@@ -115,7 +115,7 @@ void perf_mmap__put(struct mmap *map)BUG_ON (map->core.base &&refcount_read (&map->core.refcnt) ==0 );if (refcount_dec_and_test (&map->core.refcnt)) -perf_mmap__munmap (map); +mmap__munmap (map); }void perf_mmap__consume (struct mmap *map)
@@ -306,19 +306,14 @@ static void perf_mmap__aio_munmap(struct mmap *map __maybe_unused) }#endif -void perf_mmap__munmap (struct mmap *map) +void mmap__munmap (struct mmap *map) { +perf_mmap__munmap (&map->core);perf_mmap__aio_munmap (map);if (map->data !=NULL ) {munmap (map->data,mmap__mmap_len (map)); map->data =NULL ; } -if (map->core.base !=NULL ) { -munmap (map->core.base,mmap__mmap_len (map)); - map->core.base =NULL ; - map->core.fd =-1 ; -refcount_set (&map->core.refcnt,0 ); - }auxtrace_mmap__munmap (&map->auxtrace_mmap); }
tools/perf/util/mmap.h +1 -1
@@ -43,7 +43,7 @@ struct mmap_params { };int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu); -void perf_mmap__munmap (struct mmap *map); +void mmap__munmap (struct mmap *map);void perf_mmap__put (struct mmap *map);
[tip: perf/core] libperf: Adopt perf_mmap__mmap() function from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 32c261c070c222858148c2171698d2954242ddd9 Gitweb: https://git.kernel.org/tip/32c261c070c222858148c2171698d2954242ddd9 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:12 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 09:42:59 -03:00
libperf: Adopt perf_mmap__mmap() function from tools/perf
Move perf_mmap__mmap() from tools/perf to libperf, it will be used in the following patches. And rename the existing perf's function to mmap__mmap().
--- tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/mmap.c | 18 ++++++++++++++++++ tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 12 +++--------- tools/perf/util/mmap.h | 2 +- 5 files changed, 25 insertions(+), 11 deletions(-)
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -37,5 +37,7 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map);void perf_mmap__init (struct perf_mmap *map,bool overwrite); +int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp, +int fd,int cpu);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +18 -0
@@ -1,4 +1,5 @@// SPDX-License-Identifier: GPL-2.0 +#include <sys/mman.h> #include <internal/mmap.h> #include <internal/lib.h>
@@ -13,3 +14,20 @@ size_t perf_mmap__mmap_len(struct perf_mmap *map) {return map->mask +1 + page_size; } + +int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp, +int fd,int cpu) +{ + map->prev =0 ; + map->mask = mp->mask; + map->base =mmap (NULL ,perf_mmap__mmap_len (map), mp->prot, +MAP_SHARED , fd,0 ); +if (map->base ==MAP_FAILED ) { + map->base =NULL ; +return -1 ; + } + + map->fd = fd; + map->cpu = cpu; +return 0 ; +}
tools/perf/util/evlist.c +1 -1
@@ -695,7 +695,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,if (*output ==-1 ) { *output = fd; -if (perf_mmap__mmap (&maps[idx], mp, *output, evlist_cpu) <0 ) +if (mmap__mmap (&maps[idx], mp, *output, evlist_cpu) <0 )return -1 ; }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )
tools/perf/util/mmap.c +3 -9
@@ -353,7 +353,7 @@ static void perf_mmap__setup_affinity_mask(struct mmap *map, struct mmap_paramsCPU_SET (map->core.cpu, &map->affinity_mask); } -int perf_mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu) +int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu) {/* * The last one will be done at perf_mmap__consume(), so that we
@@ -369,18 +369,12 @@ int perf_mmap__mmap(struct mmap *map, struct mmap_params *mp, int fd, int cpu)* perf_evlist__filter_pollfd(). */ refcount_set (&map->core.refcnt,2 ); - map->core.prev =0 ; - map->core.mask = mp->core.mask; - map->core.base =mmap (NULL ,mmap__mmap_len (map), mp->core.prot, -MAP_SHARED , fd,0 ); -if (map->core.base ==MAP_FAILED ) { + +if (perf_mmap__mmap (&map->core, &mp->core, fd, cpu)) {pr_debug2 ("failed to mmap perf event ring buffer, error %d\n" , errno); - map->core.base =NULL ;return -1 ; } - map->core.fd = fd; - map->core.cpu = cpu;perf_mmap__setup_affinity_mask (map, mp);
tools/perf/util/mmap.h +1 -1
@@ -42,7 +42,7 @@ struct mmap_params {struct auxtrace_mmap_params auxtrace_mp; }; -int perf_mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu); +int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void perf_mmap__munmap (struct mmap *map);void perf_mmap__get (struct mmap *map);
[tip: perf/core] libperf: Adopt perf_mmap__read_event() from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 151ed5d70da87720022e4171227733a008b3c719 Gitweb: https://git.kernel.org/tip/151ed5d70da87720022e4171227733a008b3c719 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:20 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 11:49:46 -03:00
libperf: Adopt perf_mmap__read_event() from tools/perf
Move perf_mmap__read_event() from tools/perf to libperf and export it in the perf/mmap.h header.
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/perf/mmap.h | 2 +- tools/perf/lib/libperf.map | 1 +- tools/perf/lib/mmap.c | 79 +++++++++++++++++++- tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 77 +------------------- tools/perf/util/mmap.h | 2 +- tools/perf/util/python.c | 2 +- 21 files changed, 98 insertions(+), 95 deletions(-)
tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 -1
@@ -121,7 +121,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybeif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample; if (event->header.type != PERF_RECORD_COMM ||
tools/perf/builtin-kvm.c +1 -1
@@ -764,7 +764,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,if (err <0 )return (err == -EAGAIN ) ?0 :-1 ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { err =perf_evlist__parse_sample_timestamp (evlist, event, ×tamp);if (err) {perf_mmap__consume (&md->core);
tools/perf/builtin-top.c +1 -1
@@ -873,7 +873,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)if (perf_mmap__read_init (&md->core) <0 )return ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {int ret; ret =perf_evlist__parse_sample_timestamp (evlist, event, &last_timestamp);
tools/perf/builtin-trace.c +1 -1
@@ -3804,7 +3804,7 @@ again:if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { ++trace->nr_events; err =trace__deliver_event (trace, event);
tools/perf/lib/include/perf/mmap.h +2 -0
@@ -5,9 +5,11 @@#include <perf/core.h> struct perf_mmap ; +union perf_event ; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map); LIBPERF_API void perf_mmap__read_done(struct perf_mmap *map); +LIBPERF_API union perf_event *perf_mmap__read_event(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/libperf.map +1 -0
@@ -43,6 +43,7 @@ LIBPERF_0.0.1 { perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done; + perf_mmap__read_event; local: *; };
tools/perf/lib/mmap.c +79 -0
91 lines not on this page · raw message
tools/perf/tests/backward-ring-buffer.c +1 -1
@@ -39,7 +39,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,union perf_event *event;perf_mmap__read_init (&map->core); -while ((event =perf_mmap__read_event (map)) !=NULL ) { +while ((event =perf_mmap__read_event (&map->core)) !=NULL ) {const u32 type = event->header.type;switch (type) {
tools/perf/tests/bpf.c +1 -1
@@ -188,7 +188,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;if (type ==PERF_RECORD_SAMPLE )
tools/perf/tests/code-reading.c +1 -1
@@ -429,7 +429,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { ret =process_event (machine, evlist, event, state);perf_mmap__consume (&md->core);if (ret <0 )
tools/perf/tests/keep-tracking.c +1 -1
@@ -41,7 +41,7 @@ static int find_comm(struct evlist *evlist, const char *comm) md = &evlist->mmap[i];if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { if (event->header.type == PERF_RECORD_COMM && (pid_t)event->comm.pid == getpid() && (pid_t)event->comm.tid == getpid() &&
tools/perf/tests/mmap-basic.c +1 -1
@@ -117,7 +117,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuseif (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample;if (event->header.type !=PERF_RECORD_SAMPLE ) {
tools/perf/tests/openat-syscall-tp-fields.c +1 -1
@@ -96,7 +96,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtestif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;int tp_flags;struct perf_sample sample;
tools/perf/tests/perf-record.c +1 -1
@@ -174,7 +174,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusif (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {const u32 type = event->header.type;const char *name =perf_event__name (type);
tools/perf/tests/sw-clock.c +1 -1
@@ -103,7 +103,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)if (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {struct perf_sample sample;if (event->header.type !=PERF_RECORD_SAMPLE )
tools/perf/tests/switch-tracking.c +1 -1
@@ -273,7 +273,7 @@ static int process_events(struct evlist *evlist,if (perf_mmap__read_init (&md->core) <0 )continue ; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) { cnt +=1 ; ret =add_event (evlist, &events, event);perf_mmap__consume (&md->core);
tools/perf/tests/task-exit.c +1 -1
@@ -121,7 +121,7 @@ retry:if (perf_mmap__read_init (&md->core) <0 ) goto out_init; -while ((event =perf_mmap__read_event (md)) !=NULL ) { +while ((event =perf_mmap__read_event (&md->core)) !=NULL ) {if (event->header.type ==PERF_RECORD_EXIT ) nr_exit++;
tools/perf/util/evlist.c +1 -1
@@ -1811,7 +1811,7 @@ static void *perf_evlist__poll_thread(void *arg)if (perf_mmap__read_init (&map->core))continue ; -while ((event =perf_mmap__read_event (map)) !=NULL ) { +while ((event =perf_mmap__read_event (&map->core)) !=NULL ) {struct evsel *evsel =perf_evlist__event2evsel (evlist, event);if (evsel && evsel->side_band.cb)
tools/perf/util/mmap.c +0 -77
83 lines not on this page · raw message
tools/perf/util/mmap.h +0 -2
@@ -47,8 +47,6 @@ void mmap__munmap(struct mmap *map);union perf_event *perf_mmap__read_forward (struct mmap *map); -union perf_event *perf_mmap__read_event (struct mmap *map); -int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size));
tools/perf/util/python.c +1 -1
@@ -1026,7 +1026,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist,if (perf_mmap__read_init (&md->core) <0 ) goto end; - event =perf_mmap__read_event (md); + event =perf_mmap__read_event (&md->core);if (event !=NULL ) {PyObject *pyevent =pyrf_event__new (event);struct pyrf_event *pevent = (struct pyrf_event *)pyevent;
[tip: perf/core] libperf: Add perf_mmap__init() function
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 353120b48d4f61288e4745b0c8a191784b11c0f4 Gitweb: https://git.kernel.org/tip/353120b48d4f61288e4745b0c8a191784b11c0f4 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:09 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 09:37:25 -03:00
libperf: Add perf_mmap__init() function
Add perf_mmap__init() function to initialize 'struct perf_mmap' objects.
Add it to a new mmap.c source file, that will carry all the mmap related functions.
--- tools/perf/lib/Build | 1 + tools/perf/lib/include/internal/mmap.h | 2 ++ tools/perf/lib/mmap.c | 9 +++++++++ tools/perf/util/evlist.c | 5 ++--- 4 files changed, 14 insertions(+), 3 deletions(-) create mode 100644 tools/perf/lib/mmap.c
tools/perf/lib/Build +1 -0
@@ -3,6 +3,7 @@ libperf-y += cpumap.o libperf-y += threadmap.o libperf-y += evsel.o libperf-y += evlist.o +libperf-y += mmap.o libperf-y += zalloc.o libperf-y += xyarray.o libperf-y += lib.o
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -29,4 +29,6 @@ struct perf_mmap { char event_copy[PERF_SAMPLE_MAX_SIZE] __aligned(8); }; +void perf_mmap__init (struct perf_mmap *map,bool overwrite); +#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c new file +9 -0
@@ -0,0 +1,9 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <internal/mmap.h> + +void perf_mmap__init (struct perf_mmap *map,bool overwrite) +{ + map->fd =-1 ; + map->overwrite = overwrite; +refcount_set (&map->refcnt,0 ); +}
tools/perf/util/evlist.c +2 -3
@@ -629,8 +629,6 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return NULL ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { - map[i].core.fd =-1 ; - map[i].core.overwrite = overwrite;/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last
@@ -640,8 +638,9 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -refcount_set (&map[i].core.refcnt,0 ); +perf_mmap__init (&map[i].core, overwrite); } +return map; }
[tip: perf/core] perf evlist: Switch to libperf's mmap interface
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 923d0f1868cb331d660fb569ecd00c39889905f6 Gitweb: https://git.kernel.org/tip/923d0f1868cb331d660fb569ecd00c39889905f6 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:29 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:46:04 -03:00
perf evlist: Switch to libperf's mmap interface
Switch to the libperf mmap interface by calling directly perf_evlist__mmap_ops() and removing perf's evlist__mmap_per_* functions.
By switching to libperf perf_evlist__mmap() we need to operate over 'struct perf_mmap' in evlist__add_pollfd, so make the related changes there.
--- tools/perf/util/evlist.c | 179 +-------------------------------------- 1 file changed, 4 insertions(+), 175 deletions(-)
tools/perf/util/evlist.c +4 -175
@@ -431,10 +431,10 @@ int evlist__add_pollfd(struct evlist *evlist, int fd)static void perf_evlist__munmap_filtered (struct fdarray *fda,int fd,void *arg__maybe_unused ) { -struct mmap *map = fda->priv[fd].ptr; +struct perf_mmap *map = fda->priv[fd].ptr;if (map) -perf_mmap__put (&map->core); +perf_mmap__put (map); }int evlist__filter_pollfd (struct evlist *evlist,short revents_and_mask)
@@ -448,22 +448,6 @@ int evlist__poll(struct evlist *evlist, int timeout)return perf_evlist__poll (&evlist->core, timeout); } -static void perf_evlist__set_sid_idx (struct evlist *evlist, -struct evsel *evsel,int idx,int cpu, -int thread) -{ -struct perf_sample_id *sid =SID (evsel, cpu, thread); - sid->idx = idx; -if (evlist->core.cpus && cpu >=0 ) - sid->cpu = evlist->core.cpus->map[cpu]; -else - sid->cpu =-1 ; -if (!evsel->core.system_wide && evlist->core.threads && thread >=0 ) - sid->tid =perf_thread_map__pid (evlist->core.threads, thread); -else - sid->tid =-1 ; -} -struct perf_sample_id *perf_evlist__id2sid (struct evlist *evlist,u64 id) {struct hlist_head *head;
@@ -652,93 +636,6 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return map; } -static bool -perf_evlist__should_poll (struct evlist *evlist__maybe_unused , -struct evsel *evsel) -{ -if (evsel->core.attr.write_backward) -return false; -return true; -} - -static int evlist__mmap_per_evsel (struct evlist *evlist,int idx, -struct mmap_params *mp,int cpu_idx, -int thread,int *_output,int *_output_overwrite) -{ -struct evsel *evsel; -int revent; -int evlist_cpu =cpu_map__cpu (evlist->core.cpus, cpu_idx); - -evlist__for_each_entry (evlist, evsel) { -struct mmap *maps = evlist->mmap; -int *output = _output; -int fd; -int cpu; - - mp->core.prot =PROT_READ |PROT_WRITE ; -if (evsel->core.attr.write_backward) { - output = _output_overwrite; - maps = evlist->overwrite_mmap; - -if (!maps) { - maps =evlist__alloc_mmap (evlist, true); -if (!maps) -return -1 ; - evlist->overwrite_mmap = maps; -if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY ) -perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); - } - mp->core.prot &= ~PROT_WRITE ; - } - -if (evsel->core.system_wide && thread) -continue ; - - cpu =perf_cpu_map__idx (evsel->core.cpus, evlist_cpu); -if (cpu ==-1 ) -continue ; - - fd =FD (evsel, cpu, thread); - -if (*output ==-1 ) { - *output = fd; - -if (mmap__mmap (&maps[idx], mp, *output, evlist_cpu) <0 ) -return -1 ; - }else { -if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 ) -return -1 ; - -perf_mmap__get (&maps[idx].core); - } - - revent =perf_evlist__should_poll (evlist, evsel) ?POLLIN :0 ; - -/* -* The system_wide flag causes a selected event to be opened -* always without a pid. Consequently it will never get a -* POLLHUP, but it is used for tracking in combination with -* other events, so it should not need to be polled anyway. -* Therefore don't add it for polling. -*/ -if (!evsel->core.system_wide && -perf_evlist__add_pollfd (&evlist->core, fd, &maps[idx], revent) <0 ) { -perf_mmap__put (&maps[idx].core); -return -1 ; - } - -if (evsel->core.attr.read_format &PERF_FORMAT_ID ) { -if (perf_evlist__id_add_fd (&evlist->core, &evsel->core, cpu, thread, - fd) <0 ) -return -1 ; -perf_evlist__set_sid_idx (evlist, evsel, idx, cpu, - thread); - } - } - -return 0 ; -} - static void perf_evlist__mmap_cb_idx(struct perf_evlist *_evlist, struct perf_mmap_param *_mp,
@@ -783,61 +680,6 @@ perf_evlist__mmap_cb_mmap(struct perf_mmap *_map, struct perf_mmap_param *_mp,return mmap__mmap (map, mp, output, cpu); } -static int evlist__mmap_per_cpu (struct evlist *evlist, -struct mmap_params *mp) -{ -int cpu, thread; -int nr_cpus =perf_cpu_map__nr (evlist->core.cpus); -int nr_threads =perf_thread_map__nr (evlist->core.threads); - -pr_debug2 ("perf event ring buffer mmapped per cpu\n" ); -for (cpu =0 ; cpu < nr_cpus; cpu++) { -int output =-1 ; -int output_overwrite =-1 ; - -auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, cpu, - true); - -for (thread =0 ; thread < nr_threads; thread++) { -if (evlist__mmap_per_evsel (evlist, cpu, mp, cpu, - thread, &output, &output_overwrite)) - goto out_unmap; - } - } - -return 0 ; - -out_unmap: -evlist__munmap_nofree (evlist); -return -1 ; -} - -static int evlist__mmap_per_thread (struct evlist *evlist, -struct mmap_params *mp) -{ -int thread; -int nr_threads =perf_thread_map__nr (evlist->core.threads); - -pr_debug2 ("perf event ring buffer mmapped per thread\n" ); -for (thread =0 ; thread < nr_threads; thread++) { -int output =-1 ; -int output_overwrite =-1 ; - -auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, thread, - false); - -if (evlist__mmap_per_evsel (evlist, thread, mp,0 , thread, - &output, &output_overwrite)) - goto out_unmap; - } - -return 0 ; - -out_unmap: -evlist__munmap_nofree (evlist); -return -1 ; -} -unsigned long perf_event_mlock_kb_in_pages (void ) {unsigned long pages;
@@ -965,9 +807,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, bool auxtrace_overwrite, int nr_cblocks, int affinity, int flush, int comp_level) { - struct evsel *evsel; -const struct perf_cpu_map *cpus = evlist->core.cpus; -const struct perf_thread_map *threads = evlist->core.threads;/* * Delay setting mp.prot: set it before calling perf_mmap__mmap. * Its value is decided by evsel's write_backward.
@@ -979,7 +818,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .flush = flush, .comp_level = comp_level }; -struct perf_evlist_mmap_ops ops__maybe_unused = { +struct perf_evlist_mmap_ops ops = { .idx = perf_evlist__mmap_cb_idx, .get = perf_evlist__mmap_cb_get, .mmap = perf_evlist__mmap_cb_mmap,
@@ -1000,17 +839,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages,auxtrace_mmap_params__init (&mp.auxtrace_mp, evlist->core.mmap_len, auxtrace_pages, auxtrace_overwrite); -evlist__for_each_entry (evlist, evsel) { -if ((evsel->core.attr.read_format &PERF_FORMAT_ID ) && - evsel->core.sample_id ==NULL && -perf_evsel__alloc_id (&evsel->core,perf_cpu_map__nr (cpus), threads->nr) <0 ) -return -ENOMEM ; - } - -if (perf_cpu_map__empty (cpus)) -return evlist__mmap_per_thread (evlist, &mp); - -return evlist__mmap_per_cpu (evlist, &mp); +return perf_evlist__mmap_ops (&evlist->core, &ops, &mp.core); }int evlist__mmap (struct evlist *evlist,unsigned int pages)
[tip: perf/core] libperf: Adopt perf_mmap__read_init() from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 7c4d41824f9afc659ba425a41018546531cffd72 Gitweb: https://git.kernel.org/tip/7c4d41824f9afc659ba425a41018546531cffd72 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:18 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 11:45:21 -03:00
libperf: Adopt perf_mmap__read_init() from tools/perf
Move perf_mmap__read_init() from tools/perf to libperf and export it in perf/mmap.h header.
And add pr_debug2()/pr_debug3() macros support, because the code is using them.
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 2 +- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/lib/include/perf/core.h | 2 +- tools/perf/lib/include/perf/mmap.h | 1 +- tools/perf/lib/internal.h | 2 +- tools/perf/lib/libperf.map | 1 +- tools/perf/lib/mmap.c | 84 +++++++++++++++++++- tools/perf/tests/backward-ring-buffer.c | 2 +- tools/perf/tests/bpf.c | 2 +- tools/perf/tests/code-reading.c | 2 +- tools/perf/tests/keep-tracking.c | 2 +- tools/perf/tests/mmap-basic.c | 2 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/perf-record.c | 2 +- tools/perf/tests/sw-clock.c | 2 +- tools/perf/tests/switch-tracking.c | 2 +- tools/perf/tests/task-exit.c | 2 +- tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 82 +------------------- tools/perf/util/mmap.h | 1 +- tools/perf/util/python.c | 2 +- 23 files changed, 107 insertions(+), 98 deletions(-)
tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 -1
@@ -118,7 +118,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybefor (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/builtin-kvm.c +1 -1
@@ -760,7 +760,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx, *mmap_time =ULLONG_MAX ; md = &evlist->mmap[idx]; - err =perf_mmap__read_init (md); + err =perf_mmap__read_init (&md->core);if (err <0 )return (err == -EAGAIN ) ?0 :-1 ;
tools/perf/builtin-top.c +1 -1
@@ -870,7 +870,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)union perf_event *event; md = opts->overwrite ? &evlist->overwrite_mmap[idx] : &evlist->mmap[idx]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )return ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/builtin-trace.c +1 -1
@@ -3801,7 +3801,7 @@ again:struct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/lib/include/perf/core.h +2 -0
@@ -12,6 +12,8 @@ enum libperf_print_level {LIBPERF_WARN ,LIBPERF_INFO ,LIBPERF_DEBUG , +LIBPERF_DEBUG2 , + LIBPERF_DEBUG3, }; typedef int (*libperf_print_fn_t)(enum libperf_print_level level,
tools/perf/lib/include/perf/mmap.h +1 -0
@@ -7,5 +7,6 @@struct perf_mmap ; LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); +LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map);#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/internal.h +2 -0
@@ -14,5 +14,7 @@ do { \#define pr_warning (fmt, ...) __pr(LIBPERF_WARN, fmt, ##__VA_ARGS__)#define pr_info (fmt, ...) __pr(LIBPERF_INFO, fmt, ##__VA_ARGS__)#define pr_debug (fmt, ...) __pr(LIBPERF_DEBUG, fmt, ##__VA_ARGS__) +#define pr_debug2 (fmt, ...) __pr(LIBPERF_DEBUG2, fmt, ##__VA_ARGS__) +#define pr_debug3 (fmt, ...) __pr(LIBPERF_DEBUG3, fmt, ##__VA_ARGS__)#endif /* __LIBPERF_INTERNAL_H */
tools/perf/lib/libperf.map +1 -0
@@ -41,6 +41,7 @@ LIBPERF_0.0.1 { perf_evlist__set_maps; perf_evlist__poll; perf_mmap__consume; + perf_mmap__read_init; local: *; };
tools/perf/lib/mmap.c +84 -0
98 lines not on this page · raw message
tools/perf/tests/backward-ring-buffer.c +1 -1
@@ -38,7 +38,7 @@ static int count_samples(struct evlist *evlist, int *sample_count,struct mmap *map = &evlist->overwrite_mmap[i];union perf_event *event; -perf_mmap__read_init (map); +perf_mmap__read_init (&map->core);while ((event =perf_mmap__read_event (map)) !=NULL ) {const u32 type = event->header.type;
tools/perf/tests/bpf.c +1 -1
@@ -185,7 +185,7 @@ static int do_test(struct bpf_object *obj, int (*func)(void),struct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/code-reading.c +1 -1
@@ -426,7 +426,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,for (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/keep-tracking.c +1 -1
@@ -39,7 +39,7 @@ static int find_comm(struct evlist *evlist, const char *comm) found =0 ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) { if (event->header.type == PERF_RECORD_COMM &&
tools/perf/tests/mmap-basic.c +1 -1
@@ -114,7 +114,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuse } md = &evlist->mmap[0 ]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto out_init;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/openat-syscall-tp-fields.c +1 -1
@@ -93,7 +93,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subteststruct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/perf-record.c +1 -1
@@ -171,7 +171,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unusstruct mmap *md; md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/sw-clock.c +1 -1
@@ -100,7 +100,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)evlist__disable (evlist); md = &evlist->mmap[0 ]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto out_init;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/switch-tracking.c +1 -1
@@ -270,7 +270,7 @@ static int process_events(struct evlist *evlist,for (i =0 ; i < evlist->core.nr_mmaps; i++) { md = &evlist->mmap[i]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 )continue ;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/tests/task-exit.c +1 -1
@@ -118,7 +118,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused retry: md = &evlist->mmap[0 ]; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto out_init;while ((event =perf_mmap__read_event (md)) !=NULL ) {
tools/perf/util/evlist.c +1 -1
@@ -1809,7 +1809,7 @@ static void *perf_evlist__poll_thread(void *arg)struct mmap *map = &evlist->mmap[i];union perf_event *event; -if (perf_mmap__read_init (map)) +if (perf_mmap__read_init (&map->core))continue ;while ((event =perf_mmap__read_event (map)) !=NULL ) {struct evsel *evsel =perf_evlist__event2evsel (evlist, event);
tools/perf/util/mmap.c +1 -81
94 lines not on this page · raw message
tools/perf/util/mmap.h +0 -1
@@ -54,6 +54,5 @@ int perf_mmap__push(struct mmap *md, void *to,size_t mmap__mmap_len (struct mmap *map); -int perf_mmap__read_init (struct mmap *md);void perf_mmap__read_done (struct mmap *map);#endif /*__PERF_MMAP_H */
tools/perf/util/python.c +1 -1
@@ -1023,7 +1023,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist,if (!md)return NULL ; -if (perf_mmap__read_init (md) <0 ) +if (perf_mmap__read_init (&md->core) <0 ) goto end; event =perf_mmap__read_event (md);
[tip: perf/core] libperf: Adopt perf_mmap__consume() function from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 7728fa0cfaeb7d25b12c8865c733359cc8e5fb13 Gitweb: https://git.kernel.org/tip/7728fa0cfaeb7d25b12c8865c733359cc8e5fb13 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:17 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 11:43:49 -03:00
libperf: Adopt perf_mmap__consume() function from tools/perf
Move perf_mmap__consume() vrom tools/perf to libperf and export it in the perf/mmap.h header.
Move also the needed helpers perf_mmap__write_tail(), perf_mmap__read_head() and perf_mmap__empty().
--- tools/perf/arch/x86/tests/perf-time-to-tsc.c | 3 +- tools/perf/builtin-kvm.c | 5 +-- tools/perf/builtin-top.c | 3 +- tools/perf/builtin-trace.c | 3 +- tools/perf/lib/Makefile | 5 +-- tools/perf/lib/include/internal/mmap.h | 2 +- tools/perf/lib/include/perf/mmap.h | 11 +++++++- tools/perf/lib/libperf.map | 1 +- tools/perf/lib/mmap.c | 32 +++++++++++++++++++- tools/perf/tests/backward-ring-buffer.c | 1 +- tools/perf/tests/bpf.c | 1 +- tools/perf/tests/code-reading.c | 3 +- tools/perf/tests/keep-tracking.c | 3 +- tools/perf/tests/mmap-basic.c | 3 +- tools/perf/tests/openat-syscall-tp-fields.c | 3 +- tools/perf/tests/perf-record.c | 3 +- tools/perf/tests/sw-clock.c | 3 +- tools/perf/tests/switch-tracking.c | 3 +- tools/perf/tests/task-exit.c | 3 +- tools/perf/util/evlist.c | 3 +- tools/perf/util/mmap.c | 32 ++++--------------- tools/perf/util/mmap.h | 12 +------- tools/perf/util/python.c | 3 +- 23 files changed, 87 insertions(+), 54 deletions(-) create mode 100644 tools/perf/lib/include/perf/mmap.h
tools/perf/arch/x86/tests/perf-time-to-tsc.c +2 -1
@@ -9,6 +9,7 @@#include <sys/prctl.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "parse-events.h"
@@ -139,7 +140,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybe comm2_time = sample.time; } next_event: -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md); }
tools/perf/builtin-kvm.c +3 -2
@@ -46,6 +46,7 @@#include <semaphore.h> #include <signal.h> #include <math.h> +#include <perf/mmap.h> static const char *get_filename_for_perf_kvm (void ) {
@@ -766,7 +767,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,while ((event =perf_mmap__read_event (md)) !=NULL ) { err =perf_evlist__parse_sample_timestamp (evlist, event, ×tamp);if (err) { -perf_mmap__consume (md); +perf_mmap__consume (&md->core);pr_err ("Failed to parse sample\n" );return -1 ; }
@@ -776,7 +777,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,* FIXME: Here we can't consume the event, as perf_session__queue_event will * point to it, and it'll get possibly overwritten by the kernel. */ -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (err) {pr_err ("Failed to enqueue sample: %d\n" , err);
tools/perf/builtin-top.c +2 -1
@@ -82,6 +82,7 @@#include <linux/err.h> #include <linux/ctype.h> +#include <perf/mmap.h> static volatile int done;static volatile int resize;
@@ -883,7 +884,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)if (ret)break ; -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (top->qe.rotate) {pthread_mutex_lock (&top->qe.mutex);
tools/perf/builtin-trace.c +2 -1
@@ -77,6 +77,7 @@#include <sys/sysmacros.h> #include <linux/ctype.h> +#include <perf/mmap.h> #ifndef O_CLOEXEC # define O_CLOEXEC 02000000
@@ -3810,7 +3811,7 @@ again:if (err) goto out_disable; -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (interrupted) goto out_disable;
tools/perf/lib/Makefile +3 -2
@@ -172,8 +172,9 @@ install_headers: $(call do_install,include/perf/cpumap.h,$(prefix)/include/perf,644); \ $(call do_install,include/perf/threadmap.h,$(prefix)/include/perf,644); \ $(call do_install,include/perf/evlist.h,$(prefix)/include/perf,644); \ - $(call do_install,include/perf/evsel.h,$(prefix)/include/perf,644); - $(call do_install,include/perf/event.h,$(prefix)/include/perf,644); + $(call do_install,include/perf/evsel.h,$(prefix)/include/perf,644); \ + $(call do_install,include/perf/event.h,$(prefix)/include/perf,644); \ + $(call do_install,include/perf/mmap.h,$(prefix)/include/perf,644); install_pkgconfig: $(LIBPERF_PC) $(call QUIET_INSTALL, $(LIBPERF_PC)) \
tools/perf/lib/include/internal/mmap.h +2 -0
@@ -49,4 +49,6 @@ void perf_mmap__munmap(struct perf_mmap *map);void perf_mmap__get (struct perf_mmap *map);void perf_mmap__put (struct perf_mmap *map); +u64 perf_mmap__read_head (struct perf_mmap *map); +#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/include/perf/mmap.h new file +11 -0
@@ -0,0 +1,11 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __LIBPERF_MMAP_H +#define __LIBPERF_MMAP_H + +#include <perf/core.h> + +struct perf_mmap ; + +LIBPERF_API void perf_mmap__consume(struct perf_mmap *map); + +#endif /* __LIBPERF_MMAP_H */
tools/perf/lib/libperf.map +1 -0
@@ -40,6 +40,7 @@ LIBPERF_0.0.1 { perf_evlist__next; perf_evlist__set_maps; perf_evlist__poll; + perf_mmap__consume; local: *; };
tools/perf/lib/mmap.c +32 -0
@@ -1,5 +1,8 @@// SPDX-License-Identifier: GPL-2.0 #include <sys/mman.h> +#include <linux/ring_buffer.h> +#include <linux/perf_event.h> +#include <perf/mmap.h> #include <internal/mmap.h> #include <internal/lib.h> #include <linux/kernel.h>
@@ -59,3 +62,32 @@ void perf_mmap__put(struct perf_mmap *map)if (refcount_dec_and_test (&map->refcnt))perf_mmap__munmap (map); } + +static inline void perf_mmap__write_tail (struct perf_mmap *md,u64 tail) +{ +ring_buffer_write_tail (md->base, tail); +} + +u64 perf_mmap__read_head (struct perf_mmap *map) +{ +return ring_buffer_read_head (map->base); +} + +static bool perf_mmap__empty (struct perf_mmap *map) +{ +struct perf_event_mmap_page *pc = map->base; + +return perf_mmap__read_head (map) == map->prev && !pc->aux_size; +} + +void perf_mmap__consume (struct perf_mmap *map) +{ +if (!map->overwrite) { +u64 old = map->prev; + +perf_mmap__write_tail (map, old); + } + +if (refcount_read (&map->refcnt) ==1 &&perf_mmap__empty (map)) +perf_mmap__put (map); +}
tools/perf/tests/backward-ring-buffer.c +1 -0
@@ -13,6 +13,7 @@#include "util/mmap.h" #include <errno.h> #include <linux/string.h> +#include <perf/mmap.h> #define NR_ITERS 111
tools/perf/tests/bpf.c +1 -0
@@ -15,6 +15,7 @@#include <linux/string.h> #include <api/fs/fs.h> #include <bpf/bpf.h> +#include <perf/mmap.h> #include "tests.h" #include "llvm.h" #include "debug.h"
tools/perf/tests/code-reading.c +2 -1
@@ -10,6 +10,7 @@#include <sys/param.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "dso.h"
@@ -430,7 +431,7 @@ static int process_events(struct machine *machine, struct evlist *evlist,while ((event =perf_mmap__read_event (md)) !=NULL ) { ret =process_event (machine, evlist, event, state); -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (ret <0 )return ret; }
tools/perf/tests/keep-tracking.c +2 -1
@@ -5,6 +5,7 @@#include <sys/prctl.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "parse-events.h"
@@ -46,7 +47,7 @@ static int find_comm(struct evlist *evlist, const char *comm) (pid_t)event->comm.tid == getpid() && strcmp(event->comm.comm, comm) == 0) found +=1 ; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md); }
tools/perf/tests/mmap-basic.c +2 -1
@@ -16,6 +16,7 @@#include <linux/kernel.h> #include <linux/string.h> #include <perf/evlist.h> +#include <perf/mmap.h> /* * This test will generate random numbers of calls to some getpid syscalls,
@@ -139,7 +140,7 @@ int test__basic_mmap(struct test *test __maybe_unused, int subtest __maybe_unuse goto out_delete_evlist; } nr_events[evsel->idx]++; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md);
tools/perf/tests/openat-syscall-tp-fields.c +2 -1
@@ -13,6 +13,7 @@#include "debug.h" #include "util/mmap.h" #include <errno.h> +#include <perf/mmap.h> #ifndef O_DIRECTORY #define O_DIRECTORY 00200000
@@ -103,7 +104,7 @@ int test__syscall_openat_tp_fields(struct test *test __maybe_unused, int subtest ++nr_events;if (type !=PERF_RECORD_SAMPLE ) { -perf_mmap__consume (md); +perf_mmap__consume (&md->core);continue ; }
tools/perf/tests/perf-record.c +2 -1
@@ -6,6 +6,7 @@#include <pthread.h> #include <sched.h> +#include <perf/mmap.h> #include "evlist.h" #include "evsel.h" #include "debug.h"
@@ -276,7 +277,7 @@ int test__PERF_RECORD(struct test *test __maybe_unused, int subtest __maybe_unus ++errs; } -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md); }
tools/perf/tests/sw-clock.c +2 -1
@@ -15,6 +15,7 @@#include "util/mmap.h" #include "util/thread_map.h" #include <perf/evlist.h> +#include <perf/mmap.h> #define NR_LOOPS 10000000
@@ -117,7 +118,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id) total_periods += sample.period; nr_samples++; next_event: -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md);
tools/perf/tests/switch-tracking.c +2 -1
@@ -8,6 +8,7 @@#include <linux/zalloc.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> #include "debug.h" #include "parse-events.h"
@@ -275,7 +276,7 @@ static int process_events(struct evlist *evlist,while ((event =perf_mmap__read_event (md)) !=NULL ) { cnt +=1 ; ret =add_event (evlist, &events, event); -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (ret <0 ) goto out_free_nodes; }
tools/perf/tests/task-exit.c +2 -1
@@ -12,6 +12,7 @@#include <linux/string.h> #include <perf/cpumap.h> #include <perf/evlist.h> +#include <perf/mmap.h> static int exited;static int nr_exit;
@@ -124,7 +125,7 @@ retry:if (event->header.type ==PERF_RECORD_EXIT ) nr_exit++; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); }perf_mmap__read_done (md);
tools/perf/util/evlist.c +2 -1
@@ -42,6 +42,7 @@#include <perf/evlist.h> #include <perf/evsel.h> #include <perf/cpumap.h> +#include <perf/mmap.h> #include <internal/xyarray.h>
@@ -1818,7 +1819,7 @@ static void *perf_evlist__poll_thread(void *arg)else pr_warning ("cannot locate proper evsel for the side band event\n" ); -perf_mmap__consume (map); +perf_mmap__consume (&map->core); got_data = true; }perf_mmap__read_done (map);
tools/perf/util/mmap.c +7 -25
@@ -13,6 +13,7 @@#include <stdlib.h> #include <string.h> #include <unistd.h> // sysconf() +#include <perf/mmap.h> #ifdef HAVE_LIBNUMA_SUPPORT #include <numaif.h> #endif
@@ -95,7 +96,7 @@ union perf_event *perf_mmap__read_event(struct mmap *map)/* non-overwirte doesn't pause the ringbuffer */ if (!map->core.overwrite) - map->core.end =perf_mmap__read_head (map); + map->core.end =perf_mmap__read_head (&map->core); event =perf_mmap__read (map, &map->core.start, map->core.end);
@@ -105,25 +106,6 @@ union perf_event *perf_mmap__read_event(struct mmap *map)return event; } -static bool perf_mmap__empty (struct mmap *map) -{ -struct perf_event_mmap_page *pc = map->core.base; - -return perf_mmap__read_head (map) == map->core.prev && !pc->aux_size; -} - -void perf_mmap__consume (struct mmap *map) -{ -if (!map->core.overwrite) { -u64 old = map->core.prev; - -perf_mmap__write_tail (map, old); - } - -if (refcount_read (&map->core.refcnt) ==1 &&perf_mmap__empty (map)) -perf_mmap__put (&map->core); -} - int __weak auxtrace_mmap__mmap(struct auxtrace_mmap *mm __maybe_unused, struct auxtrace_mmap_params *mp __maybe_unused, void *userpg __maybe_unused,
@@ -420,7 +402,7 @@ static int overwrite_rb_find_range(void *buf, int mask, u64 *start, u64 *end)*/ static int __perf_mmap__read_init (struct mmap *md) { -u64 head =perf_mmap__read_head (md); +u64 head =perf_mmap__read_head (&md->core);u64 old = md->core.prev;unsigned char *data = md->core.base + page_size;unsigned long size;
@@ -437,7 +419,7 @@ static int __perf_mmap__read_init(struct mmap *md)WARN_ONCE (1 ,"failed to keep up with mmap data. (warn only once)\n" ); md->core.prev = head; -perf_mmap__consume (md); +perf_mmap__consume (&md->core);return -EAGAIN ; }
@@ -466,7 +448,7 @@ int perf_mmap__read_init(struct mmap *map)int perf_mmap__push (struct mmap *md,void *to,int push (struct mmap *map,void *to,void *buf,size_t size)) { -u64 head =perf_mmap__read_head (md); +u64 head =perf_mmap__read_head (&md->core);unsigned char *data = md->core.base + page_size;unsigned long size;void *buf;
@@ -499,7 +481,7 @@ int perf_mmap__push(struct mmap *md, void *to, } md->core.prev = head; -perf_mmap__consume (md); +perf_mmap__consume (&md->core); out:return rc; }
@@ -518,5 +500,5 @@ void perf_mmap__read_done(struct mmap *map)if (!refcount_read (&map->core.refcnt))return ; - map->core.prev =perf_mmap__read_head (map); + map->core.prev =perf_mmap__read_head (&map->core); }
tools/perf/util/mmap.h +0 -12
@@ -45,18 +45,6 @@ struct mmap_params {int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void mmap__munmap (struct mmap *map); -void perf_mmap__consume (struct mmap *map); - -static inline u64 perf_mmap__read_head (struct mmap *mm) -{ -return ring_buffer_read_head (mm->core.base); -} - -static inline void perf_mmap__write_tail (struct mmap *md,u64 tail) -{ -ring_buffer_write_tail (md->core.base, tail); -} -union perf_event *perf_mmap__read_forward (struct mmap *map);union perf_event *perf_mmap__read_event (struct mmap *map);
tools/perf/util/python.c +2 -1
@@ -6,6 +6,7 @@#include <linux/err.h> #include <perf/cpumap.h> #include <traceevent/event-parse.h> +#include <perf/mmap.h> #include "evlist.h" #include "callchain.h" #include "evsel.h"
@@ -1045,7 +1046,7 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist, err =perf_evsel__parse_sample (evsel, event, &pevent->sample);/* Consume the even only after we parsed it out. */ -perf_mmap__consume (md); +perf_mmap__consume (&md->core);if (err) return PyErr_Format(PyExc_OSError,
[tip: perf/core] libperf: Introduce perf_evlist__purge()
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 696f27c994ed056fd176ce9dc51c7988e148e4b0 Gitweb: https://git.kernel.org/tip/696f27c994ed056fd176ce9dc51c7988e148e4b0 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:33 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:57:22 -03:00
libperf: Introduce perf_evlist__purge()
Add a static perf_evlist__purge() function to purge evsels from a evlist.
Add also perf_evlist__for_each_entry_safe() which is used by perf_evlist__purge().
--- tools/perf/lib/evlist.c | 13 +++++++++++++ tools/perf/lib/include/internal/evlist.h | 18 ++++++++++++++++++ 2 files changed, 31 insertions(+)
tools/perf/lib/evlist.c +13 -0
@@ -109,6 +109,18 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)return next; } +static void perf_evlist__purge (struct perf_evlist *evlist) +{ +struct perf_evsel *pos, *n; + +perf_evlist__for_each_entry_safe (evlist, n, pos) { +list_del_init (&pos->node); +perf_evsel__delete (pos); + } + + evlist->nr_entries =0 ; +} +void perf_evlist__exit (struct perf_evlist *evlist) {perf_cpu_map__put (evlist->cpus);
@@ -125,6 +137,7 @@ void perf_evlist__delete(struct perf_evlist *evlist)perf_evlist__munmap (evlist);perf_evlist__close (evlist); +perf_evlist__purge (evlist);perf_evlist__exit (evlist);free (evlist); }
tools/perf/lib/include/internal/evlist.h +18 -0
@@ -82,6 +82,24 @@ void perf_evlist__exit(struct perf_evlist *evlist);#define perf_evlist__for_each_entry_reverse (evlist, evsel) \ __perf_evlist__for_each_entry_reverse(&(evlist)->entries, evsel) +/** +* __perf_evlist__for_each_entry_safe - safely iterate thru all the evsels +* @list: list_head instance to iterate +* @tmp: struct evsel temp iterator +* @evsel: struct evsel iterator +*/ +#define __perf_evlist__for_each_entry_safe (list, tmp, evsel) \ + list_for_each_entry_safe(evsel, tmp, list, node) + +/** +* perf_evlist__for_each_entry_safe - safely iterate thru all the evsels +* @evlist: evlist instance to iterate +* @evsel: struct evsel iterator +* @tmp: struct evsel temp iterator +*/ +#define perf_evlist__for_each_entry_safe (evlist, tmp, evsel) \ + __perf_evlist__for_each_entry_safe(&(evlist)->entries, tmp, evsel) +static inline struct perf_evsel *perf_evlist__first (struct perf_evlist *evlist) { return list_entry(evlist->entries.next, struct perf_evsel, node);
[tip: perf/core] perf tools: Use perf_mmap way to detect aux mmap
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 1d40ae4e1784bfa1646fd153ca022db21511284f Gitweb: https://git.kernel.org/tip/1d40ae4e1784bfa1646fd153ca022db21511284f Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:16 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 10:11:54 -03:00
perf tools: Use perf_mmap way to detect aux mmap
We will move this code to libperf shortly, so we need to free it of 'struct auxtrace_mmap' usage, because it won't be available in libperf (for now).
The perf_event_mmap_page::aux_size is set when the aux mmap is mapped, so the check is equivalent.
--- tools/perf/util/mmap.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tools/perf/util/mmap.c +3 -1
@@ -107,7 +107,9 @@ union perf_event *perf_mmap__read_event(struct mmap *map)static bool perf_mmap__empty (struct mmap *map) { -return perf_mmap__read_head (map) == map->core.prev && !map->auxtrace_mmap.base; +struct perf_event_mmap_page *pc = map->core.base; + +return perf_mmap__read_head (map) == map->core.prev && !pc->aux_size; }void perf_mmap__consume (struct mmap *map)
[tip: perf/core] libperf: Adopt perf_evlist__mmap()/munmap() from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: d1a177595b3a824c72dacb0f9d1a4e5906eaef0e Gitweb: https://git.kernel.org/tip/d1a177595b3a824c72dacb0f9d1a4e5906eaef0e Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:21 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:15:58 -03:00
libperf: Adopt perf_evlist__mmap()/munmap() from tools/perf
Add libperf's version of perf_evlist__mmap()/munmap() functions and exporting them in the perf/evlist.h header.
It's the backbone of what we have in perf code. The following changes will add needed callbacks and then we'll finally switch the perf code to use libperf's version.
Add mmap/mmap_ovw 'struct perf_mmap' object arrays to hold maps for libperf's evlist.
--- tools/perf/lib/evlist.c | 236 ++++++++++++++++++++++- tools/perf/lib/include/internal/evlist.h | 2 +- tools/perf/lib/include/perf/evlist.h | 3 +- tools/perf/lib/libperf.map | 2 +- 4 files changed, 243 insertions(+)
tools/perf/lib/evlist.c +236 -0
@@ -8,13 +8,20 @@#include <internal/evlist.h> #include <internal/evsel.h> #include <internal/xyarray.h> +#include <internal/mmap.h> +#include <internal/cpumap.h> +#include <internal/threadmap.h> +#include <internal/xyarray.h> +#include <internal/lib.h> #include <linux/zalloc.h> +#include <sys/ioctl.h> #include <stdlib.h> #include <errno.h> #include <unistd.h> #include <fcntl.h> #include <signal.h> #include <poll.h> +#include <sys/mman.h> #include <perf/cpumap.h> #include <perf/threadmap.h> #include <api/fd/array.h>
@@ -103,6 +110,10 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)void perf_evlist__delete (struct perf_evlist *evlist) { +if (evlist ==NULL ) +return ; + +perf_evlist__munmap (evlist);free (evlist); }
@@ -281,3 +292,228 @@ int perf_evlist__poll(struct perf_evlist *evlist, int timeout) {return fdarray__poll (&evlist->pollfd, timeout); } + +static struct perf_mmap *perf_evlist__alloc_mmap (struct perf_evlist *evlist,bool overwrite) +{ +int i; +struct perf_mmap *map; + + evlist->nr_mmaps =perf_cpu_map__nr (evlist->cpus); +if (perf_cpu_map__empty (evlist->cpus)) + evlist->nr_mmaps =perf_thread_map__nr (evlist->threads); + + map =zalloc (evlist->nr_mmaps *sizeof (struct perf_mmap )); +if (!map) +return NULL ; + +for (i =0 ; i < evlist->nr_mmaps; i++) { +/* +* When the perf_mmap() call is made we grab one refcount, plus +* one extra to let perf_mmap__consume() get the last +* events after all real references (perf_mmap__get()) are +* dropped. +* +* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and +* thus does perf_mmap__get() on it. +*/ +perf_mmap__init (&map[i], overwrite,NULL ); + } + +return map; +} + +static void perf_evlist__set_sid_idx (struct perf_evlist *evlist, +struct perf_evsel *evsel,int idx,int cpu, +int thread) +{ +struct perf_sample_id *sid =SID (evsel, cpu, thread); + + sid->idx = idx; +if (evlist->cpus && cpu >=0 ) + sid->cpu = evlist->cpus->map[cpu]; +else + sid->cpu =-1 ; +if (!evsel->system_wide && evlist->threads && thread >=0 ) + sid->tid =perf_thread_map__pid (evlist->threads, thread); +else + sid->tid =-1 ; +} + +static struct perf_mmap * +perf_evlist__map_get (struct perf_evlist *evlist,bool overwrite,int idx) +{ +struct perf_mmap *map = &evlist->mmap[idx]; + +if (overwrite) { +if (!evlist->mmap_ovw) { + evlist->mmap_ovw =perf_evlist__alloc_mmap (evlist, true); +if (!evlist->mmap_ovw) +return NULL ; + } + map = &evlist->mmap_ovw[idx]; + } + +return map; +} + +#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y)) + +static int +mmap_per_evsel (struct perf_evlist *evlist,int idx, +struct perf_mmap_param *mp,int cpu_idx, +int thread,int *_output,int *_output_overwrite) +{ +int evlist_cpu =perf_cpu_map__cpu (evlist->cpus, cpu_idx); +struct perf_evsel *evsel; +int revent; + +perf_evlist__for_each_entry (evlist, evsel) { +bool overwrite = evsel->attr.write_backward; +struct perf_mmap *map; +int *output, fd, cpu; + +if (evsel->system_wide && thread) +continue ; + + cpu =perf_cpu_map__idx (evsel->cpus, evlist_cpu); +if (cpu ==-1 ) +continue ; + + map =perf_evlist__map_get (evlist, overwrite, idx); +if (map ==NULL ) +return -ENOMEM ; + +if (overwrite) { + mp->prot =PROT_READ ; + output = _output_overwrite; + }else { + mp->prot =PROT_READ |PROT_WRITE ; + output = _output; + } + + fd =FD (evsel, cpu, thread); + +if (*output ==-1 ) { + *output = fd; + +if (perf_mmap__mmap (map, mp, *output, evlist_cpu) <0 ) +return -1 ; + }else { +if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 ) +return -1 ; + +perf_mmap__get (map); + } + + revent = !overwrite ?POLLIN :0 ; + +if (!evsel->system_wide && +perf_evlist__add_pollfd (evlist, fd, map, revent) <0 ) { +perf_mmap__put (map); +return -1 ; + } + +if (evsel->attr.read_format &PERF_FORMAT_ID ) { +if (perf_evlist__id_add_fd (evlist, evsel, cpu, thread, + fd) <0 ) +return -1 ; +perf_evlist__set_sid_idx (evlist, evsel, idx, cpu, + thread); + } + } + +return 0 ; +} + +static int +mmap_per_thread (struct perf_evlist *evlist,struct perf_mmap_param *mp) +{ +int thread; +int nr_threads =perf_thread_map__nr (evlist->threads); + +for (thread =0 ; thread < nr_threads; thread++) { +int output =-1 ; +int output_overwrite =-1 ; + +if (mmap_per_evsel (evlist, thread, mp,0 , thread, + &output, &output_overwrite)) + goto out_unmap; + } + +return 0 ; + +out_unmap: +perf_evlist__munmap (evlist); +return -1 ; +} + +static int +mmap_per_cpu (struct perf_evlist *evlist,struct perf_mmap_param *mp) +{ +int nr_threads =perf_thread_map__nr (evlist->threads); +int nr_cpus =perf_cpu_map__nr (evlist->cpus); +int cpu, thread; + +for (cpu =0 ; cpu < nr_cpus; cpu++) { +int output =-1 ; +int output_overwrite =-1 ; + +for (thread =0 ; thread < nr_threads; thread++) { +if (mmap_per_evsel (evlist, cpu, mp, cpu, + thread, &output, &output_overwrite)) + goto out_unmap; + } + } + +return 0 ; + +out_unmap: +perf_evlist__munmap (evlist); +return -1 ; +} + +int perf_evlist__mmap (struct perf_evlist *evlist,int pages) +{ +struct perf_evsel *evsel; +const struct perf_cpu_map *cpus = evlist->cpus; +const struct perf_thread_map *threads = evlist->threads; +struct perf_mmap_param mp; + +if (!evlist->mmap) + evlist->mmap =perf_evlist__alloc_mmap (evlist, false); +if (!evlist->mmap) +return -ENOMEM ; + +perf_evlist__for_each_entry (evlist, evsel) { +if ((evsel->attr.read_format &PERF_FORMAT_ID ) && + evsel->sample_id ==NULL && +perf_evsel__alloc_id (evsel,perf_cpu_map__nr (cpus), threads->nr) <0 ) +return -ENOMEM ; + } + + evlist->mmap_len = (pages +1 ) * page_size; + mp.mask = evlist->mmap_len - page_size -1 ; + +if (perf_cpu_map__empty (cpus)) +return mmap_per_thread (evlist, &mp); + +return mmap_per_cpu (evlist, &mp); +} + +void perf_evlist__munmap (struct perf_evlist *evlist) +{ +int i; + +if (evlist->mmap) { +for (i =0 ; i < evlist->nr_mmaps; i++) +perf_mmap__munmap (&evlist->mmap[i]); + } + +if (evlist->mmap_ovw) { +for (i =0 ; i < evlist->nr_mmaps; i++) +perf_mmap__munmap (&evlist->mmap_ovw[i]); + } + +zfree (&evlist->mmap); +zfree (&evlist->mmap_ovw); +}
tools/perf/lib/include/internal/evlist.h +2 -0
@@ -22,6 +22,8 @@ struct perf_evlist {size_t mmap_len;struct fdarray pollfd;struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ]; +struct perf_mmap *mmap; +struct perf_mmap *mmap_ovw; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
tools/perf/lib/include/perf/evlist.h +3 -0
@@ -33,4 +33,7 @@ LIBPERF_API void perf_evlist__set_maps(struct perf_evlist *evlist, struct perf_thread_map *threads); LIBPERF_API int perf_evlist__poll(struct perf_evlist *evlist, int timeout); +LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); +LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist); +#endif /* __LIBPERF_EVLIST_H */
tools/perf/lib/libperf.map +2 -0
@@ -40,6 +40,8 @@ LIBPERF_0.0.1 { perf_evlist__next; perf_evlist__set_maps; perf_evlist__poll; + perf_evlist__mmap; + perf_evlist__munmap; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;
[tip: perf/core] libperf: Adopt perf_mmap__get() function from tools/perf
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: e75710f063e29ae7715c57b45eb27c2d504b32ca Gitweb: https://git.kernel.org/tip/e75710f063e29ae7715c57b45eb27c2d504b32ca Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:13 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 09:53:27 -03:00
libperf: Adopt perf_mmap__get() function from tools/perf
Move perf_mmap__get() from tools/perf to libperf in the internal header internal/mmap.h.
--- tools/perf/builtin-record.c | 2 +- tools/perf/lib/include/internal/mmap.h | 1 + tools/perf/lib/mmap.c | 5 +++++ tools/perf/util/evlist.c | 2 +- tools/perf/util/mmap.c | 5 ----- tools/perf/util/mmap.h | 1 - 6 files changed, 8 insertions(+), 8 deletions(-)
tools/perf/builtin-record.c +1 -1
@@ -293,7 +293,7 @@ static int record__aio_pushfn(struct mmap *map, void *to, void *buf, size_t size* after started aio request completion or at record__aio_push() * if the request failed to start. */ -perf_mmap__get (map); +perf_mmap__get (&map->core); } aio->size += size;
tools/perf/lib/include/internal/mmap.h +1 -0
@@ -39,5 +39,6 @@ size_t perf_mmap__mmap_len(struct perf_mmap *map);void perf_mmap__init (struct perf_mmap *map,bool overwrite);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu); +void perf_mmap__get (struct perf_mmap *map);#endif /* __LIBPERF_INTERNAL_MMAP_H */
tools/perf/lib/mmap.c +5 -0
@@ -31,3 +31,8 @@ int perf_mmap__mmap(struct perf_mmap *map, struct perf_mmap_param *mp, map->cpu = cpu;return 0 ; } + +void perf_mmap__get (struct perf_mmap *map) +{ +refcount_inc (&map->refcnt); +}
tools/perf/util/evlist.c +1 -1
@@ -701,7 +701,7 @@ static int evlist__mmap_per_evsel(struct evlist *evlist, int idx,if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )return -1 ; -perf_mmap__get (&maps[idx]); +perf_mmap__get (&maps[idx].core); } revent =perf_evlist__should_poll (evlist, evsel) ?POLLIN :0 ;
tools/perf/util/mmap.c +0 -5
@@ -110,11 +110,6 @@ static bool perf_mmap__empty(struct mmap *map)return perf_mmap__read_head (map) == map->core.prev && !map->auxtrace_mmap.base; } -void perf_mmap__get (struct mmap *map) -{ -refcount_inc (&map->core.refcnt); -} -void perf_mmap__put (struct mmap *map) {BUG_ON (map->core.base &&refcount_read (&map->core.refcnt) ==0 );
tools/perf/util/mmap.h +0 -1
@@ -45,7 +45,6 @@ struct mmap_params {int mmap__mmap (struct mmap *map,struct mmap_params *mp,int fd,int cpu);void perf_mmap__munmap (struct mmap *map); -void perf_mmap__get (struct mmap *map);void perf_mmap__put (struct mmap *map);void perf_mmap__consume (struct mmap *map);
[tip: perf/core] perf evlist: Introduce perf_evlist__mmap_cb_get()
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: bb1b1885e2f22afb8bc7981cd865fe4b0e3d975b Gitweb: https://git.kernel.org/tip/bb1b1885e2f22afb8bc7981cd865fe4b0e3d975b Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:27 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:30:21 -03:00
perf evlist: Introduce perf_evlist__mmap_cb_get()
Add the perf_evlist__mmap_cb_get() function to return 'struct perf_mmap' object during perf_evlist__mmap_ops() call.
The array of 'struct mmap' is allocated via evlist__alloc_mmap(), in this callback we simply returns pointer to the base object.
--- tools/perf/util/evlist.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+)
tools/perf/util/evlist.c +24 -0
@@ -750,6 +750,29 @@ perf_evlist__mmap_cb_idx(struct perf_evlist *_evlist,auxtrace_mmap_params__set_idx (&mp->auxtrace_mp, evlist, idx, per_cpu); } +static struct perf_mmap * +perf_evlist__mmap_cb_get (struct perf_evlist *_evlist,bool overwrite,int idx) +{ + struct evlist *evlist = container_of(_evlist, struct evlist, core); +struct mmap *maps = evlist->mmap; + +if (overwrite) { + maps = evlist->overwrite_mmap; + +if (!maps) { + maps =evlist__alloc_mmap (evlist, true); +if (!maps) +return NULL ; + + evlist->overwrite_mmap = maps; +if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY ) +perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); + } + } + +return &maps[idx].core; +} +static int evlist__mmap_per_cpu (struct evlist *evlist,struct mmap_params *mp) {
@@ -948,6 +971,7 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, };struct perf_evlist_mmap_ops ops__maybe_unused = { .idx = perf_evlist__mmap_cb_idx, + .get = perf_evlist__mmap_cb_get, };if (!evlist->mmap)
[tip: perf/core] libperf: Introduce perf_evlist_mmap_ops::idx callback
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 1fcbb75cc574072ab457dbbaa74fc7064b691e86 Gitweb: https://git.kernel.org/tip/1fcbb75cc574072ab457dbbaa74fc7064b691e86 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:23 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:20:08 -03:00
libperf: Introduce perf_evlist_mmap_ops::idx callback
Add the perf_evlist_mmap_ops::idx callback to be called in mmap_per_cpu() and mmap_per_thread() with current cpu and thread indexes.
It's used by current aux code, so perf will use this callback to set the aux index.
--- tools/perf/lib/evlist.c | 18 +++++++++++++----- tools/perf/lib/include/internal/evlist.h | 4 ++++ 2 files changed, 17 insertions(+), 5 deletions(-)
tools/perf/lib/evlist.c +13 -5
@@ -426,7 +426,8 @@ mmap_per_evsel(struct perf_evlist *evlist, int idx, }static int -mmap_per_thread (struct perf_evlist *evlist,struct perf_mmap_param *mp) +mmap_per_thread (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp) {int thread;int nr_threads =perf_thread_map__nr (evlist->threads);
@@ -435,6 +436,9 @@ mmap_per_thread(struct perf_evlist *evlist, struct perf_mmap_param *mp)int output =-1 ;int output_overwrite =-1 ; +if (ops->idx) + ops->idx (evlist, mp, thread, false); +if (mmap_per_evsel (evlist, thread, mp,0 , thread, &output, &output_overwrite)) goto out_unmap;
@@ -448,7 +452,8 @@ out_unmap: }static int -mmap_per_cpu (struct perf_evlist *evlist,struct perf_mmap_param *mp) +mmap_per_cpu (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp) {int nr_threads =perf_thread_map__nr (evlist->threads);int nr_cpus =perf_cpu_map__nr (evlist->cpus);
@@ -458,6 +463,9 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_mmap_param *mp)int output =-1 ;int output_overwrite =-1 ; +if (ops->idx) + ops->idx (evlist, mp, cpu, true); +for (thread =0 ; thread < nr_threads; thread++) {if (mmap_per_evsel (evlist, cpu, mp, cpu, thread, &output, &output_overwrite))
@@ -496,15 +504,15 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist, }if (perf_cpu_map__empty (cpus)) -return mmap_per_thread (evlist, mp); +return mmap_per_thread (evlist, ops, mp); -return mmap_per_cpu (evlist, mp); +return mmap_per_cpu (evlist, ops, mp); }int perf_evlist__mmap (struct perf_evlist *evlist,int pages) {struct perf_mmap_param mp; -struct perf_evlist_mmap_ops ops; +struct perf_evlist_mmap_ops ops = {0 }; evlist->mmap_len = (pages +1 ) * page_size; mp.mask = evlist->mmap_len - page_size -1 ;
tools/perf/lib/include/internal/evlist.h +4 -0
@@ -27,7 +27,11 @@ struct perf_evlist {struct perf_mmap *mmap_ovw; }; +typedef void +(*perf_evlist_mmap__cb_idx_t )(struct perf_evlist *,struct perf_mmap_param *,int ,bool ); +struct perf_evlist_mmap_ops { +perf_evlist_mmap__cb_idx_t idx; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
[tip: perf/core] libperf: Add perf_evlist_mmap_ops::get callback
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 3a8bb58121987a8405d6f96cd8815025e564605d Gitweb: https://git.kernel.org/tip/3a8bb58121987a8405d6f96cd8815025e564605d Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:24 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:21:11 -03:00
libperf: Add perf_evlist_mmap_ops::get callback
Add the perf_evlist_mmap_ops::get callback to be called in mmap_per_evsel() to get/allocate the 'struct perf_mmap' object.
Add the libperf's perf_evlist__mmap_cb_get() function as libperf's get callback.
--- tools/perf/lib/evlist.c | 18 ++++++++++-------- tools/perf/lib/include/internal/evlist.h | 3 +++ 2 files changed, 13 insertions(+), 8 deletions(-)
tools/perf/lib/evlist.c +10 -8
@@ -340,7 +340,7 @@ static void perf_evlist__set_sid_idx(struct perf_evlist *evlist, }static struct perf_mmap * -perf_evlist__map_get (struct perf_evlist *evlist,bool overwrite,int idx) +perf_evlist__mmap_cb_get (struct perf_evlist *evlist,bool overwrite,int idx) {struct perf_mmap *map = &evlist->mmap[idx];
@@ -359,8 +359,8 @@ perf_evlist__map_get(struct perf_evlist *evlist, bool overwrite, int idx)#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y))static int -mmap_per_evsel (struct perf_evlist *evlist,int idx, -struct perf_mmap_param *mp,int cpu_idx, +mmap_per_evsel (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, +int idx,struct perf_mmap_param *mp,int cpu_idx,int thread,int *_output,int *_output_overwrite) {int evlist_cpu =perf_cpu_map__cpu (evlist->cpus, cpu_idx);
@@ -379,7 +379,7 @@ mmap_per_evsel(struct perf_evlist *evlist, int idx,if (cpu ==-1 )continue ; - map =perf_evlist__map_get (evlist, overwrite, idx); + map = ops->get (evlist, overwrite, idx);if (map ==NULL )return -ENOMEM ;
@@ -439,7 +439,7 @@ mmap_per_thread(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (ops->idx) ops->idx (evlist, mp, thread, false); -if (mmap_per_evsel (evlist, thread, mp,0 , thread, +if (mmap_per_evsel (evlist, ops, thread, mp,0 , thread, &output, &output_overwrite)) goto out_unmap; }
@@ -467,7 +467,7 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops, ops->idx (evlist, mp, cpu, true);for (thread =0 ; thread < nr_threads; thread++) { -if (mmap_per_evsel (evlist, cpu, mp, cpu, +if (mmap_per_evsel (evlist, ops, cpu, mp, cpu, thread, &output, &output_overwrite)) goto out_unmap; }
@@ -488,7 +488,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,const struct perf_cpu_map *cpus = evlist->cpus;const struct perf_thread_map *threads = evlist->threads; -if (!ops) +if (!ops || !ops->get)return -EINVAL ;if (!evlist->mmap)
@@ -512,7 +512,9 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,int perf_evlist__mmap (struct perf_evlist *evlist,int pages) {struct perf_mmap_param mp; -struct perf_evlist_mmap_ops ops = {0 }; +struct perf_evlist_mmap_ops ops = { + .get = perf_evlist__mmap_cb_get, + }; evlist->mmap_len = (pages +1 ) * page_size; mp.mask = evlist->mmap_len - page_size -1 ;
tools/perf/lib/include/internal/evlist.h +3 -0
@@ -29,9 +29,12 @@ struct perf_evlist {typedef void (*perf_evlist_mmap__cb_idx_t )(struct perf_evlist *,struct perf_mmap_param *,int ,bool ); +typedef struct perf_mmap * +(*perf_evlist_mmap__cb_get_t )(struct perf_evlist *,bool ,int );struct perf_evlist_mmap_ops {perf_evlist_mmap__cb_idx_t idx; +perf_evlist_mmap__cb_get_t get; };int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);
[tip: perf/core] libperf: Introduce perf_evlist__mmap_ops()
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 0b5ea10d4c312f5b17af9d09187efb9418517bec Gitweb: https://git.kernel.org/tip/0b5ea10d4c312f5b17af9d09187efb9418517bec Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:22 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:18:00 -03:00
libperf: Introduce perf_evlist__mmap_ops()
To be able to pass specific callbacks to evlist's mmap.
There will be a specific call to this function from perf's evlist__mmap() and libperf's perf_evlist__mmap() functions in following changes.
--- tools/perf/lib/evlist.c | 24 +++++++++++++++++------ tools/perf/lib/include/internal/evlist.h | 8 ++++++++- 2 files changed, 26 insertions(+), 6 deletions(-)
tools/perf/lib/evlist.c +18 -6
@@ -472,12 +472,16 @@ out_unmap:return -1 ; } -int perf_evlist__mmap (struct perf_evlist *evlist,int pages) +int perf_evlist__mmap_ops (struct perf_evlist *evlist, +struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp) {struct perf_evsel *evsel;const struct perf_cpu_map *cpus = evlist->cpus;const struct perf_thread_map *threads = evlist->threads; -struct perf_mmap_param mp; + +if (!ops) +return -EINVAL ;if (!evlist->mmap) evlist->mmap =perf_evlist__alloc_mmap (evlist, false);
@@ -491,13 +495,21 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages)return -ENOMEM ; } +if (perf_cpu_map__empty (cpus)) +return mmap_per_thread (evlist, mp); + +return mmap_per_cpu (evlist, mp); +} + +int perf_evlist__mmap (struct perf_evlist *evlist,int pages) +{ +struct perf_mmap_param mp; +struct perf_evlist_mmap_ops ops; + evlist->mmap_len = (pages +1 ) * page_size; mp.mask = evlist->mmap_len - page_size -1 ; -if (perf_cpu_map__empty (cpus)) -return mmap_per_thread (evlist, &mp); - -return mmap_per_cpu (evlist, &mp); +return perf_evlist__mmap_ops (evlist, &ops, &mp); }void perf_evlist__munmap (struct perf_evlist *evlist)
tools/perf/lib/include/internal/evlist.h +8 -0
@@ -11,6 +11,7 @@struct perf_cpu_map ;struct perf_thread_map ; +struct perf_mmap_param ;struct perf_evlist {struct list_head entries;
@@ -26,10 +27,17 @@ struct perf_evlist {struct perf_mmap *mmap_ovw; }; +struct perf_evlist_mmap_ops { +}; +int perf_evlist__alloc_pollfd (struct perf_evlist *evlist);int perf_evlist__add_pollfd (struct perf_evlist *evlist,int fd,void *ptr,short revent); +int perf_evlist__mmap_ops (struct perf_evlist *evlist, +struct perf_evlist_mmap_ops *ops, +struct perf_mmap_param *mp); +/** * __perf_evlist__for_each_entry - iterate thru all the evsels * @list: list_head instance to iterate
[tip: perf/core] perf evlist: Introduce perf_evlist__mmap_cb_mmap()
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: b80132b12a78ec71de2b3320cc49d4a0b2cd7c46 Gitweb: https://git.kernel.org/tip/b80132b12a78ec71de2b3320cc49d4a0b2cd7c46 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:28 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:44:59 -03:00
perf evlist: Introduce perf_evlist__mmap_cb_mmap()
Add the perf_evlist__mmap_cb_mmap() function to call perf specific mmap__mmap() function during perf_evlist__mmap_ops() call.
--- tools/perf/util/evlist.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)
tools/perf/util/evlist.c +13 -2
@@ -773,6 +773,16 @@ perf_evlist__mmap_cb_get(struct perf_evlist *_evlist, bool overwrite, int idx)return &maps[idx].core; } +static int +perf_evlist__mmap_cb_mmap (struct perf_mmap *_map,struct perf_mmap_param *_mp, +int output,int cpu) +{ + struct mmap *map = container_of(_map, struct mmap, core); + struct mmap_params *mp = container_of(_mp, struct mmap_params, core); + +return mmap__mmap (map, mp, output, cpu); +} +static int evlist__mmap_per_cpu (struct evlist *evlist,struct mmap_params *mp) {
@@ -970,8 +980,9 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .comp_level = comp_level };struct perf_evlist_mmap_ops ops__maybe_unused = { - .idx = perf_evlist__mmap_cb_idx, - .get = perf_evlist__mmap_cb_get, + .idx = perf_evlist__mmap_cb_idx, + .get = perf_evlist__mmap_cb_get, + .mmap = perf_evlist__mmap_cb_mmap, };if (!evlist->mmap)
[tip: perf/core] libperf: Introduce perf_evlist__exit()
In reply to Jiri OlsaThe following commit has been merged into the perf/core branch of tip:
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 93dd6e2831ff399f7685aa2157b997b6392efac8 Gitweb: https://git.kernel.org/tip/93dd6e2831ff399f7685aa2157b997b6392efac8 Author: Jiri Olsa <[email protected]> AuthorDate: Mon, 07 Oct 2019 14:53:32 +02:00 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitterDate: Thu, 10 Oct 2019 12:56:01 -03:00
libperf: Introduce perf_evlist__exit()
Add the perf_evlist__exit() function, so far it's not exported and added only for internal use for perf and libperf.
USe it to release cpus/threads and pollfd array.
--- tools/perf/lib/evlist.c | 12 +++++++++++- tools/perf/lib/include/internal/evlist.h | 2 ++ tools/perf/util/evlist.c | 6 +----- 3 files changed, 14 insertions(+), 6 deletions(-)
tools/perf/lib/evlist.c +11 -1
@@ -109,13 +109,23 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)return next; } +void perf_evlist__exit (struct perf_evlist *evlist) +{ +perf_cpu_map__put (evlist->cpus); +perf_thread_map__put (evlist->threads); + evlist->cpus =NULL ; + evlist->threads =NULL ; +fdarray__exit (&evlist->pollfd); +} +void perf_evlist__delete (struct perf_evlist *evlist) {if (evlist ==NULL )return ;perf_evlist__munmap (evlist); -fdarray__exit (&evlist->pollfd); +perf_evlist__close (evlist); +perf_evlist__exit (evlist);free (evlist); }
tools/perf/lib/include/internal/evlist.h +2 -0
@@ -48,6 +48,8 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, struct perf_mmap_param *mp); +void perf_evlist__exit (struct perf_evlist *evlist); +/** * __perf_evlist__for_each_entry - iterate thru all the evsels * @list: list_head instance to iterate
tools/perf/util/evlist.c +1 -5
@@ -138,7 +138,7 @@ void evlist__exit(struct evlist *evlist) {zfree (&evlist->mmap);zfree (&evlist->overwrite_mmap); -fdarray__exit (&evlist->core.pollfd); +perf_evlist__exit (&evlist->core); }void evlist__delete (struct evlist *evlist)
@@ -148,10 +148,6 @@ void evlist__delete(struct evlist *evlist)evlist__munmap (evlist);evlist__close (evlist); -perf_cpu_map__put (evlist->core.cpus); -perf_thread_map__put (evlist->core.threads); - evlist->core.cpus =NULL ; - evlist->core.threads =NULL ;evlist__purge (evlist);evlist__exit (evlist);free (evlist);