diff options
| author | Linus Torvalds <torvalds@linux-foundation.org> | 2025-12-05 09:37:41 -0800 |
|---|---|---|
| committer | Linus Torvalds <torvalds@linux-foundation.org> | 2025-12-05 09:37:41 -0800 |
| commit | 36492b7141b9abc967e92c991af32c670351dc16 (patch) | |
| tree | 7e1ea41e1bc327a1c4b028174b118050a5945ea3 /scripts/elf-parse.c | |
| parent | 5779de8d36ac5a0c929f276096a499b03ae0afa7 (diff) | |
| parent | b21f90e2e4503847ffeb00a9ef4d6d390291f902 (diff) | |
Merge tag 'tracepoints-v6.19' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull unused tracepoints update from Steven Rostedt:
"Detect unused tracepoints.
If a tracepoint is defined but never used (TRACE_EVENT() created but
no trace_<tracepoint>() called), it can take up to or more than 5K of
memory each. This can add up as there are around a hundred unused
tracepoints with various configs. That is 500K of wasted memory.
Add a make build parameter of "UT=1" to have the build warn if an
unused tracepoint is detected in the build. This allows detection of
unused tracepoints to be upstream so that outreachy and the mentoring
project can have new developers look for fixing them, without having
these warnings suddenly show up when someone upgrades their kernel.
When all known unused tracepoints are removed, then the "UT=1" build
parameter can be removed and unused tracepoints will always warn. This
will catch new unused tracepoints after the current ones have been
removed.
Summary:
- Separate out elf functions from sorttable.c
Move out the ELF parsing functions from sorttable.c so that the
tracing tooling can use it.
- Add a tracepoint verifier tool to the build process
If "UT=1" is added to the kernel command line, any unused
tracepoints will trigger a warning at build time.
- Do not warn about unused tracepoints for tracepoints that are
exported
There are sever cases where a tracepoint is created by the kernel
and used by modules. Since there's no easy way to detect if these
are truly unused since the users are in modules, if a tracepoint is
exported, assume it will eventually be used by a module. Note,
there's not many exported tracepoints so this should not be a
problem to ignore them.
- Have building of modules also detect unused tracepoints
Do not only check the main vmlinux for unused tracepoints, also
check modules. If a module is defining a tracepoint it should be
using it.
- Add the tracepoint-update program to the ignore file
The new tracepoint-update program needs to be ignored by git"
* tag 'tracepoints-v6.19' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
scripts: add tracepoint-update to the list of ignores files
tracing: Add warnings for unused tracepoints for modules
tracing: Allow tracepoint-update.c to work with modules
tracepoint: Do not warn for unused event that is exported
tracing: Add a tracepoint verification check at build time
sorttable: Move ELF parsing into scripts/elf-parse.[ch]
Diffstat (limited to 'scripts/elf-parse.c')
| -rw-r--r-- | scripts/elf-parse.c | 198 |
1 files changed, 198 insertions, 0 deletions
diff --git a/scripts/elf-parse.c b/scripts/elf-parse.c new file mode 100644 index 000000000000..99869ff91a8c --- /dev/null +++ b/scripts/elf-parse.c @@ -0,0 +1,198 @@ +#include <sys/types.h> +#include <sys/mman.h> +#include <sys/stat.h> +#include <fcntl.h> +#include <stdio.h> +#include <stdlib.h> +#include <stdbool.h> +#include <string.h> +#include <unistd.h> +#include <errno.h> + +#include "elf-parse.h" + +struct elf_funcs elf_parser; + +/* + * Get the whole file as a programming convenience in order to avoid + * malloc+lseek+read+free of many pieces. If successful, then mmap + * avoids copying unused pieces; else just read the whole file. + * Open for both read and write. + */ +static void *map_file(char const *fname, size_t *size) +{ + int fd; + struct stat sb; + void *addr = NULL; + + fd = open(fname, O_RDWR); + if (fd < 0) { + perror(fname); + return NULL; + } + if (fstat(fd, &sb) < 0) { + perror(fname); + goto out; + } + if (!S_ISREG(sb.st_mode)) { + fprintf(stderr, "not a regular file: %s\n", fname); + goto out; + } + + addr = mmap(0, sb.st_size, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0); + if (addr == MAP_FAILED) { + fprintf(stderr, "Could not mmap file: %s\n", fname); + goto out; + } + + *size = sb.st_size; + +out: + close(fd); + return addr; +} + +static int elf_parse(const char *fname, void *addr, uint32_t types) +{ + Elf_Ehdr *ehdr = addr; + uint16_t type; + + switch (ehdr->e32.e_ident[EI_DATA]) { + case ELFDATA2LSB: + elf_parser.r = rle; + elf_parser.r2 = r2le; + elf_parser.r8 = r8le; + elf_parser.w = wle; + elf_parser.w8 = w8le; + break; + case ELFDATA2MSB: + elf_parser.r = rbe; + elf_parser.r2 = r2be; + elf_parser.r8 = r8be; + elf_parser.w = wbe; + elf_parser.w8 = w8be; + break; + default: + fprintf(stderr, "unrecognized ELF data encoding %d: %s\n", + ehdr->e32.e_ident[EI_DATA], fname); + return -1; + } + + if (memcmp(ELFMAG, ehdr->e32.e_ident, SELFMAG) != 0 || + ehdr->e32.e_ident[EI_VERSION] != EV_CURRENT) { + fprintf(stderr, "unrecognized ELF file %s\n", fname); + return -1; + } + + type = elf_parser.r2(&ehdr->e32.e_type); + if (!((1 << type) & types)) { + fprintf(stderr, "Invalid ELF type file %s\n", fname); + return -1; + } + + switch (ehdr->e32.e_ident[EI_CLASS]) { + case ELFCLASS32: { + elf_parser.ehdr_shoff = ehdr32_shoff; + elf_parser.ehdr_shentsize = ehdr32_shentsize; + elf_parser.ehdr_shstrndx = ehdr32_shstrndx; + elf_parser.ehdr_shnum = ehdr32_shnum; + elf_parser.shdr_addr = shdr32_addr; + elf_parser.shdr_offset = shdr32_offset; + elf_parser.shdr_link = shdr32_link; + elf_parser.shdr_size = shdr32_size; + elf_parser.shdr_name = shdr32_name; + elf_parser.shdr_type = shdr32_type; + elf_parser.shdr_entsize = shdr32_entsize; + elf_parser.sym_type = sym32_type; + elf_parser.sym_name = sym32_name; + elf_parser.sym_value = sym32_value; + elf_parser.sym_shndx = sym32_shndx; + elf_parser.rela_offset = rela32_offset; + elf_parser.rela_info = rela32_info; + elf_parser.rela_addend = rela32_addend; + elf_parser.rela_write_addend = rela32_write_addend; + + if (elf_parser.r2(&ehdr->e32.e_ehsize) != sizeof(Elf32_Ehdr) || + elf_parser.r2(&ehdr->e32.e_shentsize) != sizeof(Elf32_Shdr)) { + fprintf(stderr, + "unrecognized ET_EXEC/ET_DYN file: %s\n", fname); + return -1; + } + + } + break; + case ELFCLASS64: { + elf_parser.ehdr_shoff = ehdr64_shoff; + elf_parser.ehdr_shentsize = ehdr64_shentsize; + elf_parser.ehdr_shstrndx = ehdr64_shstrndx; + elf_parser.ehdr_shnum = ehdr64_shnum; + elf_parser.shdr_addr = shdr64_addr; + elf_parser.shdr_offset = shdr64_offset; + elf_parser.shdr_link = shdr64_link; + elf_parser.shdr_size = shdr64_size; + elf_parser.shdr_name = shdr64_name; + elf_parser.shdr_type = shdr64_type; + elf_parser.shdr_entsize = shdr64_entsize; + elf_parser.sym_type = sym64_type; + elf_parser.sym_name = sym64_name; + elf_parser.sym_value = sym64_value; + elf_parser.sym_shndx = sym64_shndx; + elf_parser.rela_offset = rela64_offset; + elf_parser.rela_info = rela64_info; + elf_parser.rela_addend = rela64_addend; + elf_parser.rela_write_addend = rela64_write_addend; + + if (elf_parser.r2(&ehdr->e64.e_ehsize) != sizeof(Elf64_Ehdr) || + elf_parser.r2(&ehdr->e64.e_shentsize) != sizeof(Elf64_Shdr)) { + fprintf(stderr, + "unrecognized ET_EXEC/ET_DYN file: %s\n", + fname); + return -1; + } + + } + break; + default: + fprintf(stderr, "unrecognized ELF class %d %s\n", + ehdr->e32.e_ident[EI_CLASS], fname); + return -1; + } + return 0; +} + +int elf_map_machine(void *addr) +{ + Elf_Ehdr *ehdr = addr; + + return elf_parser.r2(&ehdr->e32.e_machine); +} + +int elf_map_long_size(void *addr) +{ + Elf_Ehdr *ehdr = addr; + + return ehdr->e32.e_ident[EI_CLASS] == ELFCLASS32 ? 4 : 8; +} + +void *elf_map(char const *fname, size_t *size, uint32_t types) +{ + void *addr; + int ret; + + addr = map_file(fname, size); + if (!addr) + return NULL; + + ret = elf_parse(fname, addr, types); + if (ret < 0) { + elf_unmap(addr, *size); + return NULL; + } + + return addr; +} + +void elf_unmap(void *addr, size_t size) +{ + munmap(addr, size); +} |