From 2492c465ad3ae6860ebfff1c9032865017835e70 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 4 Jul 2016 19:35:47 -0300 Subject: perf build: Add feature detection for libelf's elf_getshdrstrndx() That appeared after 0.140, and will be used in the SDT code, so, to avoid bisection break on older systems, add a feature detection and provide a stub with a pr_debug() to keep it building. Cc: Ananth N Mavinakayanahalli Cc: Brendan Gregg Cc: Hemant Kumar Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-80y0eldgweorqnwha9rvfxjr@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'tools/perf/util/symbol-elf.c') diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 87a297dd8901..b222552c7159 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -54,6 +54,14 @@ static int elf_getphdrnum(Elf *elf, size_t *dst) } #endif +#ifndef HAVE_ELF_GETSHDRSTRNDX_SUPPORT +static int elf_getshdrstrndx(Elf *elf __maybe_unused, size_t *dst __maybe_unused) +{ + pr_err("%s: update your libelf to > 0.140, this one lacks elf_getshdrstrndx().\n", __func__); + return -1; +} +#endif + #ifndef NT_GNU_BUILD_ID #define NT_GNU_BUILD_ID 3 #endif -- cgit v1.2.3 From 060fa0c7a3e0bb4f1426ee79dfd38e2a4c80067a Mon Sep 17 00:00:00 2001 From: Hemant Kumar Date: Fri, 1 Jul 2016 17:03:46 +0900 Subject: perf sdt: ELF support for SDT This patch serves the initial support to identify and list SDT events in binaries. When programs containing SDT markers are compiled, gcc with the help of assembler directives identifies them and places them in the section ".note.stapsdt". To find these markers from the binaries, one needs to traverse through this section and parse the relevant details like the name, type and location of the marker. Also, the original location could be skewed due to the effect of prelinking. If that is the case, the locations need to be adjusted. The functions in this patch open a given ELF, find out the SDT section, parse the relevant details, adjust the location (if necessary) and populate them in a list. A typical note entry in ".note.stapsdt" section is as follows : |--nhdr.n_namesz--| ------------------------------------ | nhdr | "stapsdt" | ----- |----------------------------------| | | | | | | nhdr.n_descsize | "provider_name" "note_name" | | | | ----- |----------------------------------| | nhdr | "stapsdt" | |... The above shows an excerpt from the section ".note.stapsdt". 'nhdr' is a structure which has the note name size (n_namesz), note description size (n_desc_sz) and note type (n_type). So, in order to parse the note note info, we need nhdr to tell us where to start from. As can be seen from , the name of the SDT notes given is "stapsdt". But this is not the identifier of the note. After that, we go to description of the note to find out its location, the address of the ".stapsdt.base" section and the semaphore address. Then, we find the provider name and the SDT marker name and then follow the arguments. Signed-off-by: Hemant Kumar Reviewed-by: Masami Hiramatsu Acked-by: Namhyung Kim Cc: Ananth N Mavinakayanahalli Cc: Brendan Gregg Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/146736022628.27797.1201368329092908163.stgit@devbox Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 252 +++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/symbol.h | 22 ++++ 2 files changed, 274 insertions(+) (limited to 'tools/perf/util/symbol-elf.c') diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index b222552c7159..6f15b92cbf70 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -1789,6 +1789,258 @@ void kcore_extract__delete(struct kcore_extract *kce) unlink(kce->extract_filename); } +/** + * populate_sdt_note : Parse raw data and identify SDT note + * @elf: elf of the opened file + * @data: raw data of a section with description offset applied + * @len: note description size + * @type: type of the note + * @sdt_notes: List to add the SDT note + * + * Responsible for parsing the @data in section .note.stapsdt in @elf and + * if its an SDT note, it appends to @sdt_notes list. + */ +static int populate_sdt_note(Elf **elf, const char *data, size_t len, + struct list_head *sdt_notes) +{ + const char *provider, *name; + struct sdt_note *tmp = NULL; + GElf_Ehdr ehdr; + GElf_Addr base_off = 0; + GElf_Shdr shdr; + int ret = -EINVAL; + + union { + Elf64_Addr a64[NR_ADDR]; + Elf32_Addr a32[NR_ADDR]; + } buf; + + Elf_Data dst = { + .d_buf = &buf, .d_type = ELF_T_ADDR, .d_version = EV_CURRENT, + .d_size = gelf_fsize((*elf), ELF_T_ADDR, NR_ADDR, EV_CURRENT), + .d_off = 0, .d_align = 0 + }; + Elf_Data src = { + .d_buf = (void *) data, .d_type = ELF_T_ADDR, + .d_version = EV_CURRENT, .d_size = dst.d_size, .d_off = 0, + .d_align = 0 + }; + + tmp = (struct sdt_note *)calloc(1, sizeof(struct sdt_note)); + if (!tmp) { + ret = -ENOMEM; + goto out_err; + } + + INIT_LIST_HEAD(&tmp->note_list); + + if (len < dst.d_size + 3) + goto out_free_note; + + /* Translation from file representation to memory representation */ + if (gelf_xlatetom(*elf, &dst, &src, + elf_getident(*elf, NULL)[EI_DATA]) == NULL) { + pr_err("gelf_xlatetom : %s\n", elf_errmsg(-1)); + goto out_free_note; + } + + /* Populate the fields of sdt_note */ + provider = data + dst.d_size; + + name = (const char *)memchr(provider, '\0', data + len - provider); + if (name++ == NULL) + goto out_free_note; + + tmp->provider = strdup(provider); + if (!tmp->provider) { + ret = -ENOMEM; + goto out_free_note; + } + tmp->name = strdup(name); + if (!tmp->name) { + ret = -ENOMEM; + goto out_free_prov; + } + + if (gelf_getclass(*elf) == ELFCLASS32) { + memcpy(&tmp->addr, &buf, 3 * sizeof(Elf32_Addr)); + tmp->bit32 = true; + } else { + memcpy(&tmp->addr, &buf, 3 * sizeof(Elf64_Addr)); + tmp->bit32 = false; + } + + if (!gelf_getehdr(*elf, &ehdr)) { + pr_debug("%s : cannot get elf header.\n", __func__); + ret = -EBADF; + goto out_free_name; + } + + /* Adjust the prelink effect : + * Find out the .stapsdt.base section. + * This scn will help us to handle prelinking (if present). + * Compare the retrieved file offset of the base section with the + * base address in the description of the SDT note. If its different, + * then accordingly, adjust the note location. + */ + if (elf_section_by_name(*elf, &ehdr, &shdr, SDT_BASE_SCN, NULL)) { + base_off = shdr.sh_offset; + if (base_off) { + if (tmp->bit32) + tmp->addr.a32[0] = tmp->addr.a32[0] + base_off - + tmp->addr.a32[1]; + else + tmp->addr.a64[0] = tmp->addr.a64[0] + base_off - + tmp->addr.a64[1]; + } + } + + list_add_tail(&tmp->note_list, sdt_notes); + return 0; + +out_free_name: + free(tmp->name); +out_free_prov: + free(tmp->provider); +out_free_note: + free(tmp); +out_err: + return ret; +} + +/** + * construct_sdt_notes_list : constructs a list of SDT notes + * @elf : elf to look into + * @sdt_notes : empty list_head + * + * Scans the sections in 'elf' for the section + * .note.stapsdt. It, then calls populate_sdt_note to find + * out the SDT events and populates the 'sdt_notes'. + */ +static int construct_sdt_notes_list(Elf *elf, struct list_head *sdt_notes) +{ + GElf_Ehdr ehdr; + Elf_Scn *scn = NULL; + Elf_Data *data; + GElf_Shdr shdr; + size_t shstrndx, next; + GElf_Nhdr nhdr; + size_t name_off, desc_off, offset; + int ret = 0; + + if (gelf_getehdr(elf, &ehdr) == NULL) { + ret = -EBADF; + goto out_ret; + } + if (elf_getshdrstrndx(elf, &shstrndx) != 0) { + ret = -EBADF; + goto out_ret; + } + + /* Look for the required section */ + scn = elf_section_by_name(elf, &ehdr, &shdr, SDT_NOTE_SCN, NULL); + if (!scn) { + ret = -ENOENT; + goto out_ret; + } + + if ((shdr.sh_type != SHT_NOTE) || (shdr.sh_flags & SHF_ALLOC)) { + ret = -ENOENT; + goto out_ret; + } + + data = elf_getdata(scn, NULL); + + /* Get the SDT notes */ + for (offset = 0; (next = gelf_getnote(data, offset, &nhdr, &name_off, + &desc_off)) > 0; offset = next) { + if (nhdr.n_namesz == sizeof(SDT_NOTE_NAME) && + !memcmp(data->d_buf + name_off, SDT_NOTE_NAME, + sizeof(SDT_NOTE_NAME))) { + /* Check the type of the note */ + if (nhdr.n_type != SDT_NOTE_TYPE) + goto out_ret; + + ret = populate_sdt_note(&elf, ((data->d_buf) + desc_off), + nhdr.n_descsz, sdt_notes); + if (ret < 0) + goto out_ret; + } + } + if (list_empty(sdt_notes)) + ret = -ENOENT; + +out_ret: + return ret; +} + +/** + * get_sdt_note_list : Wrapper to construct a list of sdt notes + * @head : empty list_head + * @target : file to find SDT notes from + * + * This opens the file, initializes + * the ELF and then calls construct_sdt_notes_list. + */ +int get_sdt_note_list(struct list_head *head, const char *target) +{ + Elf *elf; + int fd, ret; + + fd = open(target, O_RDONLY); + if (fd < 0) + return -EBADF; + + elf = elf_begin(fd, PERF_ELF_C_READ_MMAP, NULL); + if (!elf) { + ret = -EBADF; + goto out_close; + } + ret = construct_sdt_notes_list(elf, head); + elf_end(elf); +out_close: + close(fd); + return ret; +} + +/** + * cleanup_sdt_note_list : free the sdt notes' list + * @sdt_notes: sdt notes' list + * + * Free up the SDT notes in @sdt_notes. + * Returns the number of SDT notes free'd. + */ +int cleanup_sdt_note_list(struct list_head *sdt_notes) +{ + struct sdt_note *tmp, *pos; + int nr_free = 0; + + list_for_each_entry_safe(pos, tmp, sdt_notes, note_list) { + list_del(&pos->note_list); + free(pos->name); + free(pos->provider); + free(pos); + nr_free++; + } + return nr_free; +} + +/** + * sdt_notes__get_count: Counts the number of sdt events + * @start: list_head to sdt_notes list + * + * Returns the number of SDT notes in a list + */ +int sdt_notes__get_count(struct list_head *start) +{ + struct sdt_note *sdt_ptr; + int count = 0; + + list_for_each_entry(sdt_ptr, start, note_list) + count++; + return count; +} + void symbol__elf_init(void) { elf_version(EV_CURRENT); diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index b10d558a8803..699f7cbcfe72 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -342,4 +342,26 @@ void arch__sym_update(struct symbol *s, GElf_Sym *sym); int arch__choose_best_symbol(struct symbol *syma, struct symbol *symb); +/* structure containing an SDT note's info */ +struct sdt_note { + char *name; /* name of the note*/ + char *provider; /* provider name */ + bool bit32; /* whether the location is 32 bits? */ + union { /* location, base and semaphore addrs */ + Elf64_Addr a64[3]; + Elf32_Addr a32[3]; + } addr; + struct list_head note_list; /* SDT notes' list */ +}; + +int get_sdt_note_list(struct list_head *head, const char *target); +int cleanup_sdt_note_list(struct list_head *sdt_notes); +int sdt_notes__get_count(struct list_head *start); + +#define SDT_BASE_SCN ".stapsdt.base" +#define SDT_NOTE_SCN ".note.stapsdt" +#define SDT_NOTE_TYPE 3 +#define SDT_NOTE_NAME "stapsdt" +#define NR_ADDR 3 + #endif /* __PERF_SYMBOL */ -- cgit v1.2.3 From cc31078cf13f67c489ad6a5c48dea657f5f88d11 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 12 Jul 2016 11:04:13 -0300 Subject: perf symbols: Provide a GElf_Nhdr typedef This one can be safely defined to be Elf64_Nhdr, as it is in elfutils's libelf, but not on musl libc, as both Elf64_Nhdr and Elf32_Nhdr have the same layout. Cc: Adrian Hunter Cc: David Ahern Cc: Jiri Olsa Cc: Namhyung Kim Cc: Wang Nan Link: http://lkml.kernel.org/n/tip-w8z8614l03lc8bip4ijbywbt@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools/perf/util/symbol-elf.c') diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 6f15b92cbf70..79a6a190f41d 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -16,6 +16,7 @@ #define EM_AARCH64 183 /* ARM 64 bit */ #endif +typedef Elf64_Nhdr GElf_Nhdr; #ifdef HAVE_CPLUS_DEMANGLE_SUPPORT extern char *cplus_demangle(const char *, int); -- cgit v1.2.3 From 1c1a3a4729aae712c55e001e151ef008d030d4a7 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 12 Jul 2016 12:19:09 -0300 Subject: perf tools: Add feature detection for gelf_getnote() That is not present on some libelf implementations, such as the one used in Alpine Linux: libelf-0.8.13. This ends up disabling the SDT code, that relies on this function. One alternative would be to provide an weak fallback implementation or the open coded variant used by the buildid sysfs notes reading code. Cc: Adrian Hunter Cc: Ananth N Mavinakayanahalli Cc: Brendan Gregg Cc: David Ahern Cc: Hemant Kumar Cc: Jiri Olsa Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Wang Nan Link: http://lkml.kernel.org/n/tip-82lh22ybedy9b9lych8xj12g@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/build/Makefile.feature | 1 + tools/build/feature/Makefile | 4 ++++ tools/build/feature/test-all.c | 5 +++++ tools/build/feature/test-libelf-gelf_getnote.c | 7 +++++++ tools/perf/config/Makefile | 6 ++++++ tools/perf/util/build-id.c | 2 +- tools/perf/util/probe-file.c | 2 ++ tools/perf/util/symbol-elf.c | 2 ++ 8 files changed, 28 insertions(+), 1 deletion(-) create mode 100644 tools/build/feature/test-libelf-gelf_getnote.c (limited to 'tools/perf/util/symbol-elf.c') diff --git a/tools/build/Makefile.feature b/tools/build/Makefile.feature index 3dd529bb0604..fe12bee57418 100644 --- a/tools/build/Makefile.feature +++ b/tools/build/Makefile.feature @@ -40,6 +40,7 @@ FEATURE_TESTS_BASIC := \ libbfd \ libelf \ libelf-getphdrnum \ + libelf-gelf_getnote \ libelf-getshdrstrndx \ libelf-mmap \ libnuma \ diff --git a/tools/build/feature/Makefile b/tools/build/feature/Makefile index 674711629ef0..d6017c1abdb0 100644 --- a/tools/build/feature/Makefile +++ b/tools/build/feature/Makefile @@ -17,6 +17,7 @@ FILES= \ test-cplus-demangle.bin \ test-libelf.bin \ test-libelf-getphdrnum.bin \ + test-libelf-gelf_getnote.bin \ test-libelf-getshdrstrndx.bin \ test-libelf-mmap.bin \ test-libnuma.bin \ @@ -99,6 +100,9 @@ $(OUTPUT)test-libelf-mmap.bin: $(OUTPUT)test-libelf-getphdrnum.bin: $(BUILD) -lelf +$(OUTPUT)test-libelf-gelf_getnote.bin: + $(BUILD) -lelf + $(OUTPUT)test-libelf-getshdrstrndx.bin: $(BUILD) -lelf diff --git a/tools/build/feature/test-all.c b/tools/build/feature/test-all.c index 7433cca33306..843aed024a3a 100644 --- a/tools/build/feature/test-all.c +++ b/tools/build/feature/test-all.c @@ -49,6 +49,10 @@ # include "test-libelf-getphdrnum.c" #undef main +#define main main_test_libelf_gelf_getnote +# include "test-libelf-gelf_getnote.c" +#undef main + #define main main_test_libelf_getshdrstrndx # include "test-libelf-getshdrstrndx.c" #undef main @@ -153,6 +157,7 @@ int main(int argc, char *argv[]) main_test_dwarf(); main_test_dwarf_getlocations(); main_test_libelf_getphdrnum(); + main_test_libelf_gelf_getnote(); main_test_libelf_getshdrstrndx(); main_test_libunwind(); main_test_libaudit(); diff --git a/tools/build/feature/test-libelf-gelf_getnote.c b/tools/build/feature/test-libelf-gelf_getnote.c new file mode 100644 index 000000000000..d78cf4d5271f --- /dev/null +++ b/tools/build/feature/test-libelf-gelf_getnote.c @@ -0,0 +1,7 @@ +#include +#include + +int main(void) +{ + return gelf_getnote(NULL, 0, NULL, NULL, NULL); +} diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile index 97d47140a637..5ac428060779 100644 --- a/tools/perf/config/Makefile +++ b/tools/perf/config/Makefile @@ -309,6 +309,12 @@ ifndef NO_LIBELF CFLAGS += -DHAVE_ELF_GETPHDRNUM_SUPPORT endif + ifeq ($(feature-libelf-gelf_getnote), 1) + CFLAGS += -DHAVE_GELF_GETNOTE_SUPPORT + else + msg := $(warning gelf_getnote() not found on libelf, SDT support disabled); + endif + ifeq ($(feature-libelf-getshdrstrndx), 1) CFLAGS += -DHAVE_ELF_GETSHDRSTRNDX_SUPPORT endif diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index e1a16408da9c..1e504e40dac8 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -533,7 +533,7 @@ int build_id_cache__list_build_ids(const char *pathname, return ret; } -#ifdef HAVE_LIBELF_SUPPORT +#if defined(HAVE_LIBELF_SUPPORT) && defined(HAVE_GELF_GETNOTE_SUPPORT) static int build_id_cache__add_sdt_cache(const char *sbuild_id, const char *realname) { diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c index 98398b55a03f..e705a742ee1e 100644 --- a/tools/perf/util/probe-file.c +++ b/tools/perf/util/probe-file.c @@ -624,6 +624,7 @@ out_err: return ret; } +#ifdef HAVE_GELF_GETNOTE_SUPPORT static unsigned long long sdt_note__get_addr(struct sdt_note *note) { return note->bit32 ? (unsigned long long)note->addr.a32[0] @@ -682,6 +683,7 @@ int probe_cache__scan_sdt(struct probe_cache *pcache, const char *pathname) cleanup_sdt_note_list(&sdtlist); return ret; } +#endif static int probe_cache_entry__write(struct probe_cache_entry *entry, int fd) { diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 79a6a190f41d..cebf98ec27bc 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -1790,6 +1790,7 @@ void kcore_extract__delete(struct kcore_extract *kce) unlink(kce->extract_filename); } +#ifdef HAVE_GELF_GETNOTE_SUPPORT /** * populate_sdt_note : Parse raw data and identify SDT note * @elf: elf of the opened file @@ -2041,6 +2042,7 @@ int sdt_notes__get_count(struct list_head *start) count++; return count; } +#endif void symbol__elf_init(void) { -- cgit v1.2.3 From cae15db74999edb96dd9f5bbd4d55849391dd92b Mon Sep 17 00:00:00 2001 From: David Tolnay Date: Sat, 9 Jul 2016 00:20:00 -0700 Subject: perf symbols: Add Rust demangling Rust demangling is another step after bfd demangling. Add a diagnosis to identify mangled Rust symbols based on the hash that the Rust mangler appends as the last path component, as well as other characteristics. Add a demangler to reconstruct the original symbol. Committer notes: How I tested it: Enabled COPR on Fedora 24 and then installed the 'rust-binary' package, with it: $ cat src/main.rs fn main() { println!("Hello, world!"); } $ cat Cargo.toml [package] name = "hello_world" version = "0.0.1" authors = [ "Arnaldo Carvalho de Melo " ] $ perf record cargo bench Compiling hello_world v0.0.1 (file:///home/acme/projects/hello_world) Running target/release/hello_world-d4b9dab4b2a47d75 running 0 tests test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.096 MB perf.data (1457 samples) ] $ Before this patch: $ perf report --stdio --dsos librbml-e8edd0fd.so # dso: librbml-e8edd0fd.so # # Total Lost Samples: 0 # # Samples: 1K of event 'cycles:u' # Event count (approx.): 979599126 # # Overhead Command Symbol # ........ ....... ............................................................................................................. # 1.78% rustc [.] rbml::reader::maybe_get_doc::hb9d387df6024b15b 1.50% rustc [.] _$LT$reader..DocsIterator$LT$$u27$a$GT$$u20$as$u20$std..iter..Iterator$GT$::next::hd9af9e60d79a35c8 1.20% rustc [.] rbml::reader::doc_at::hc88107fba445af31 0.46% rustc [.] _$LT$reader..TaggedDocsIterator$LT$$u27$a$GT$$u20$as$u20$std..iter..Iterator$GT$::next::h0cb40e696e4bb489 0.35% rustc [.] rbml::reader::Decoder::_next_int::h66eef7825a398bc3 0.29% rustc [.] rbml::reader::Decoder::_next_sub::h8e5266005580b836 0.15% rustc [.] rbml::reader::get_doc::h094521c645459139 0.14% rustc [.] _$LT$reader..Decoder$LT$$u27$doc$GT$$u20$as$u20$serialize..Decoder$GT$::read_u32::h0acea2fff9669327 0.07% rustc [.] rbml::reader::Decoder::next_doc::h6714d469c9dfaf91 0.07% rustc [.] _ZN4rbml6reader10doc_as_u6417h930b740aa94f1d3aE@plt 0.06% rustc [.] _fini $ After: $ perf report --stdio --dsos librbml-e8edd0fd.so # dso: librbml-e8edd0fd.so # # Total Lost Samples: 0 # # Samples: 1K of event 'cycles:u' # Event count (approx.): 979599126 # # Overhead Command Symbol # ........ ....... ................................................................. # 1.78% rustc [.] rbml::reader::maybe_get_doc 1.50% rustc [.] as std::iter::Iterator>::next 1.20% rustc [.] rbml::reader::doc_at 0.46% rustc [.] as std::iter::Iterator>::next 0.35% rustc [.] rbml::reader::Decoder::_next_int 0.29% rustc [.] rbml::reader::Decoder::_next_sub 0.15% rustc [.] rbml::reader::get_doc 0.14% rustc [.] as serialize::Decoder>::read_u32 0.07% rustc [.] rbml::reader::Decoder::next_doc 0.07% rustc [.] _ZN4rbml6reader10doc_as_u6417h930b740aa94f1d3aE@plt 0.06% rustc [.] _fini $ Signed-off-by: David Tolnay Tested-by: Arnaldo Carvalho de Melo Cc: Alexander Shishkin Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/5780B7FA.3030602@gmail.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/Build | 1 + tools/perf/util/demangle-rust.c | 269 ++++++++++++++++++++++++++++++++++++++++ tools/perf/util/demangle-rust.h | 7 ++ tools/perf/util/symbol-elf.c | 8 ++ 4 files changed, 285 insertions(+) create mode 100644 tools/perf/util/demangle-rust.c create mode 100644 tools/perf/util/demangle-rust.h (limited to 'tools/perf/util/symbol-elf.c') diff --git a/tools/perf/util/Build b/tools/perf/util/Build index eda68f582884..2fa7d8b69873 100644 --- a/tools/perf/util/Build +++ b/tools/perf/util/Build @@ -113,6 +113,7 @@ libperf-y += scripting-engines/ libperf-$(CONFIG_ZLIB) += zlib.o libperf-$(CONFIG_LZMA) += lzma.o libperf-y += demangle-java.o +libperf-y += demangle-rust.o ifdef CONFIG_JITDUMP libperf-$(CONFIG_LIBELF) += jitdump.o diff --git a/tools/perf/util/demangle-rust.c b/tools/perf/util/demangle-rust.c new file mode 100644 index 000000000000..f9dafa888c06 --- /dev/null +++ b/tools/perf/util/demangle-rust.c @@ -0,0 +1,269 @@ +#include +#include "util.h" +#include "debug.h" + +#include "demangle-rust.h" + +/* + * Mangled Rust symbols look like this: + * + * _$LT$std..sys..fd..FileDesc$u20$as$u20$core..ops..Drop$GT$::drop::hc68340e1baa4987a + * + * The original symbol is: + * + * ::drop + * + * The last component of the path is a 64-bit hash in lowercase hex, prefixed + * with "h". Rust does not have a global namespace between crates, an illusion + * which Rust maintains by using the hash to distinguish things that would + * otherwise have the same symbol. + * + * Any path component not starting with a XID_Start character is prefixed with + * "_". + * + * The following escape sequences are used: + * + * "," => $C$ + * "@" => $SP$ + * "*" => $BP$ + * "&" => $RF$ + * "<" => $LT$ + * ">" => $GT$ + * "(" => $LP$ + * ")" => $RP$ + * " " => $u20$ + * "'" => $u27$ + * "[" => $u5b$ + * "]" => $u5d$ + * "~" => $u7e$ + * + * A double ".." means "::" and a single "." means "-". + * + * The only characters allowed in the mangled symbol are a-zA-Z0-9 and _.:$ + */ + +static const char *hash_prefix = "::h"; +static const size_t hash_prefix_len = 3; +static const size_t hash_len = 16; + +static bool is_prefixed_hash(const char *start); +static bool looks_like_rust(const char *sym, size_t len); +static bool unescape(const char **in, char **out, const char *seq, char value); + +/* + * INPUT: + * sym: symbol that has been through BFD-demangling + * + * This function looks for the following indicators: + * + * 1. The hash must consist of "h" followed by 16 lowercase hex digits. + * + * 2. As a sanity check, the hash must use between 5 and 15 of the 16 possible + * hex digits. This is true of 99.9998% of hashes so once in your life you + * may see a false negative. The point is to notice path components that + * could be Rust hashes but are probably not, like "haaaaaaaaaaaaaaaa". In + * this case a false positive (non-Rust symbol has an important path + * component removed because it looks like a Rust hash) is worse than a + * false negative (the rare Rust symbol is not demangled) so this sets the + * balance in favor of false negatives. + * + * 3. There must be no characters other than a-zA-Z0-9 and _.:$ + * + * 4. There must be no unrecognized $-sign sequences. + * + * 5. There must be no sequence of three or more dots in a row ("..."). + */ +bool +rust_is_mangled(const char *sym) +{ + size_t len, len_without_hash; + + if (!sym) + return false; + + len = strlen(sym); + if (len <= hash_prefix_len + hash_len) + /* Not long enough to contain "::h" + hash + something else */ + return false; + + len_without_hash = len - (hash_prefix_len + hash_len); + if (!is_prefixed_hash(sym + len_without_hash)) + return false; + + return looks_like_rust(sym, len_without_hash); +} + +/* + * A hash is the prefix "::h" followed by 16 lowercase hex digits. The hex + * digits must comprise between 5 and 15 (inclusive) distinct digits. + */ +static bool is_prefixed_hash(const char *str) +{ + const char *end; + bool seen[16]; + size_t i; + int count; + + if (strncmp(str, hash_prefix, hash_prefix_len)) + return false; + str += hash_prefix_len; + + memset(seen, false, sizeof(seen)); + for (end = str + hash_len; str < end; str++) + if (*str >= '0' && *str <= '9') + seen[*str - '0'] = true; + else if (*str >= 'a' && *str <= 'f') + seen[*str - 'a' + 10] = true; + else + return false; + + /* Count how many distinct digits seen */ + count = 0; + for (i = 0; i < 16; i++) + if (seen[i]) + count++; + + return count >= 5 && count <= 15; +} + +static bool looks_like_rust(const char *str, size_t len) +{ + const char *end = str + len; + + while (str < end) + switch (*str) { + case '$': + if (!strncmp(str, "$C$", 3)) + str += 3; + else if (!strncmp(str, "$SP$", 4) + || !strncmp(str, "$BP$", 4) + || !strncmp(str, "$RF$", 4) + || !strncmp(str, "$LT$", 4) + || !strncmp(str, "$GT$", 4) + || !strncmp(str, "$LP$", 4) + || !strncmp(str, "$RP$", 4)) + str += 4; + else if (!strncmp(str, "$u20$", 5) + || !strncmp(str, "$u27$", 5) + || !strncmp(str, "$u5b$", 5) + || !strncmp(str, "$u5d$", 5) + || !strncmp(str, "$u7e$", 5)) + str += 5; + else + return false; + break; + case '.': + /* Do not allow three or more consecutive dots */ + if (!strncmp(str, "...", 3)) + return false; + /* Fall through */ + case 'a' ... 'z': + case 'A' ... 'Z': + case '0' ... '9': + case '_': + case ':': + str++; + break; + default: + return false; + } + + return true; +} + +/* + * INPUT: + * sym: symbol for which rust_is_mangled(sym) returns true + * + * The input is demangled in-place because the mangled name is always longer + * than the demangled one. + */ +void +rust_demangle_sym(char *sym) +{ + const char *in; + char *out; + const char *end; + + if (!sym) + return; + + in = sym; + out = sym; + end = sym + strlen(sym) - (hash_prefix_len + hash_len); + + while (in < end) + switch (*in) { + case '$': + if (!(unescape(&in, &out, "$C$", ',') + || unescape(&in, &out, "$SP$", '@') + || unescape(&in, &out, "$BP$", '*') + || unescape(&in, &out, "$RF$", '&') + || unescape(&in, &out, "$LT$", '<') + || unescape(&in, &out, "$GT$", '>') + || unescape(&in, &out, "$LP$", '(') + || unescape(&in, &out, "$RP$", ')') + || unescape(&in, &out, "$u20$", ' ') + || unescape(&in, &out, "$u27$", '\'') + || unescape(&in, &out, "$u5b$", '[') + || unescape(&in, &out, "$u5d$", ']') + || unescape(&in, &out, "$u7e$", '~'))) { + pr_err("demangle-rust: unexpected escape sequence"); + goto done; + } + break; + case '_': + /* + * If this is the start of a path component and the next + * character is an escape sequence, ignore the + * underscore. The mangler inserts an underscore to make + * sure the path component begins with a XID_Start + * character. + */ + if ((in == sym || in[-1] == ':') && in[1] == '$') + in++; + else + *out++ = *in++; + break; + case '.': + if (in[1] == '.') { + /* ".." becomes "::" */ + *out++ = ':'; + *out++ = ':'; + in += 2; + } else { + /* "." becomes "-" */ + *out++ = '-'; + in++; + } + break; + case 'a' ... 'z': + case 'A' ... 'Z': + case '0' ... '9': + case ':': + *out++ = *in++; + break; + default: + pr_err("demangle-rust: unexpected character '%c' in symbol\n", + *in); + goto done; + } + +done: + *out = '\0'; +} + +static bool unescape(const char **in, char **out, const char *seq, char value) +{ + size_t len = strlen(seq); + + if (strncmp(*in, seq, len)) + return false; + + **out = value; + + *in += len; + *out += 1; + + return true; +} diff --git a/tools/perf/util/demangle-rust.h b/tools/perf/util/demangle-rust.h new file mode 100644 index 000000000000..7b41ead7e0dd --- /dev/null +++ b/tools/perf/util/demangle-rust.h @@ -0,0 +1,7 @@ +#ifndef __PERF_DEMANGLE_RUST +#define __PERF_DEMANGLE_RUST 1 + +bool rust_is_mangled(const char *str); +void rust_demangle_sym(char *str); + +#endif /* __PERF_DEMANGLE_RUST */ diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index cebf98ec27bc..a34321e9b44d 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -7,6 +7,7 @@ #include "symbol.h" #include "demangle-java.h" +#include "demangle-rust.h" #include "machine.h" #include "vdso.h" #include @@ -1081,6 +1082,13 @@ new_symbol: demangled = bfd_demangle(NULL, elf_name, demangle_flags); if (demangled == NULL) demangled = java_demangle_sym(elf_name, JAVA_DEMANGLE_NORET); + else if (rust_is_mangled(demangled)) + /* + * Input to Rust demangling is the BFD-demangled + * name which it Rust-demangles in place. + */ + rust_demangle_sym(demangled); + if (demangled != NULL) elf_name = demangled; } -- cgit v1.2.3