* Commit: readelf: Add --display-section option
@ 2024-04-10 11:23 Nick Clifton
0 siblings, 0 replies; only message in thread
From: Nick Clifton @ 2024-04-10 11:23 UTC (permalink / raw)
To: binutils
[-- Attachment #1: Type: text/plain, Size: 954 bytes --]
Hi Guys,
Sometimes when using readelf you only want to see information on one
particular section, even if there are multiple sections that would
match the dumping criteria. (For example when examining the RELR
relocations in a file, you may not be interested in the REL or RELA
relocations).
Objdump and objcopy have their -j options to restrict their operations
to just a single specified section and I decided that it would be nice
to have something similar for readelf. So I have created the attached
patch which implements -j/--display-section=<name|number>. Unlike
objcopy and objdump however you do not have to tell readelf what to do
with the specified section. Instead readelf will look at the
section's type and display the information accordingly -
relocations for SHT_REL sections for example or notes for SHT_NOTE
sections.
Tested with a wide variety of different binutils configurations.
Cheers
Nick
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: readelf-j.patch --]
[-- Type: text/x-patch, Size: 21116 bytes --]
diff --git a/binutils/NEWS b/binutils/NEWS
index 9c7c8f1f603..be744e3b2c4 100644
--- a/binutils/NEWS
+++ b/binutils/NEWS
@@ -1,5 +1,9 @@
-*- text -*-
+* Readelf now has a -j/--display-section option which takes the name or index
+ of a section and displays its contents according to its type. The option can
+ be used multiple times on the command line to display multiple sections.
+
* Base register 0 is now printed as "0" instead of "%r0" in s390 disassembly.
* When objdump or readelf are used to display the contents of a .eh_frame
diff --git a/binutils/doc/binutils.texi b/binutils/doc/binutils.texi
index 79461559a56..4e67b6e0da9 100644
--- a/binutils/doc/binutils.texi
+++ b/binutils/doc/binutils.texi
@@ -5012,6 +5012,7 @@ readelf [@option{-a}|@option{--all}]
[@option{-x} <number or name>|@option{--hex-dump=}<number or name>]
[@option{-p} <number or name>|@option{--string-dump=}<number or name>]
[@option{-R} <number or name>|@option{--relocated-dump=}<number or name>]
+ [@option{-j} <number or name>|@option{--display-section=}<number or name>]
[@option{-z}|@option{--decompress}]
[@option{-c}|@option{--archive-index}]
[@option{-w[lLiaprmfFsoORtUuTgAck]}|
@@ -5270,6 +5271,8 @@ displayed.
Displays the contents of the indicated section as a hexadecimal bytes.
A number identifies a particular section by index in the section table;
any other string identifies all sections with that name in the object file.
+This option can be repeated multiple times on the command line in
+order to request multiple hex dumps.
@item -R <number or name>
@itemx --relocated-dump=<number or name>
@@ -5278,12 +5281,31 @@ bytes. A number identifies a particular section by index in the
section table; any other string identifies all sections with that name
in the object file. The contents of the section will be relocated
before they are displayed.
+This option can be repeated multiple times on the command line in
+order to request multiple relocated dumps.
@item -p <number or name>
@itemx --string-dump=<number or name>
Displays the contents of the indicated section as printable strings.
A number identifies a particular section by index in the section table;
any other string identifies all sections with that name in the object file.
+This option can be repeated multiple times on the command line in
+order to request multiple string dumps.
+
+@item -j <number or name>
+@itemx --display-section
+Displays the contents of the indicated section according to its
+section header type. Sections containing relocations will be
+displayed as if the @option{--relocations} option had been used,
+sections contains symbols will be displayed as if the @option{--syms}
+option had been used and so on.
+
+A number identifies a particular section by index in the section
+table; any other string identifies all sections with that name in the
+input file(s).
+
+This option can be repeated multiple times on the command line in
+order to request multiple section dumps.
@item -z
@itemx --decompress
diff --git a/binutils/elfcomm.c b/binutils/elfcomm.c
index ad7c5744f20..b8289b77161 100644
--- a/binutils/elfcomm.c
+++ b/binutils/elfcomm.c
@@ -63,6 +63,20 @@ warn (const char *message, ...)
va_end (args);
}
+void
+inform (const char *message, ...)
+{
+ va_list args;
+
+ /* Try to keep info messages in sync with the program's normal output. */
+ fflush (stdout);
+
+ va_start (args, message);
+ fprintf (stderr, _("%s: Info: "), program_name);
+ vfprintf (stderr, message, args);
+ va_end (args);
+}
+
void (*byte_put) (unsigned char *, uint64_t, unsigned int);
void
diff --git a/binutils/elfcomm.h b/binutils/elfcomm.h
index 4cda2a7b6a8..9728d6166d9 100644
--- a/binutils/elfcomm.h
+++ b/binutils/elfcomm.h
@@ -28,6 +28,7 @@
extern void error (const char *, ...) ATTRIBUTE_PRINTF_1;
extern void warn (const char *, ...) ATTRIBUTE_PRINTF_1;
+extern void inform (const char *, ...) ATTRIBUTE_PRINTF_1;
extern void (*byte_put) (unsigned char *, uint64_t, unsigned int);
extern void byte_put_little_endian (unsigned char *, uint64_t, unsigned int);
diff --git a/binutils/readelf.c b/binutils/readelf.c
index c1006480b7b..fa0de3a7e0d 100644
--- a/binutils/readelf.c
+++ b/binutils/readelf.c
@@ -188,12 +188,15 @@ typedef struct elf_section_list
/* Flag bits indicating particular types of dump. */
#define HEX_DUMP (1 << 0) /* The -x command line switch. */
+#ifdef SUPPORT_DISASSEMBLY
#define DISASS_DUMP (1 << 1) /* The -i command line switch. */
+#endif
#define DEBUG_DUMP (1 << 2) /* The -w command line switch. */
#define STRING_DUMP (1 << 3) /* The -p command line switch. */
#define RELOC_DUMP (1 << 4) /* The -R command line switch. */
#define CTF_DUMP (1 << 5) /* The --ctf command line switch. */
#define SFRAME_DUMP (1 << 6) /* The --sframe command line switch. */
+#define AUTO_DUMP (1 << 7) /* The -j command line switch. */
typedef unsigned char dump_type;
@@ -402,6 +405,9 @@ static const char * get_symbol_version_string
(Filedata *, bool, const char *, size_t, unsigned,
Elf_Internal_Sym *, enum versioned_symbol_info *, unsigned short *);
+static bool process_notes_at
+ (Filedata *, Elf_Internal_Shdr *, uint64_t, uint64_t, uint64_t);
+
#define UNKNOWN -1
static inline const char *
@@ -5414,7 +5420,6 @@ get_section_type_name (Filedata * filedata, unsigned int sh_type)
case SHT_SYMTAB: return "SYMTAB";
case SHT_STRTAB: return "STRTAB";
case SHT_RELA: return "RELA";
- case SHT_RELR: return "RELR";
case SHT_HASH: return "HASH";
case SHT_DYNAMIC: return "DYNAMIC";
case SHT_NOTE: return "NOTE";
@@ -5422,20 +5427,27 @@ get_section_type_name (Filedata * filedata, unsigned int sh_type)
case SHT_REL: return "REL";
case SHT_SHLIB: return "SHLIB";
case SHT_DYNSYM: return "DYNSYM";
+ /* 12 and 13 are not defined. */
case SHT_INIT_ARRAY: return "INIT_ARRAY";
case SHT_FINI_ARRAY: return "FINI_ARRAY";
case SHT_PREINIT_ARRAY: return "PREINIT_ARRAY";
- case SHT_GNU_HASH: return "GNU_HASH";
case SHT_GROUP: return "GROUP";
case SHT_SYMTAB_SHNDX: return "SYMTAB SECTION INDICES";
+ case SHT_RELR: return "RELR";
+ /* End of generic section types. */
+
+ /* OS specific section types: */
case SHT_GNU_verdef: return "VERDEF";
case SHT_GNU_verneed: return "VERNEED";
case SHT_GNU_versym: return "VERSYM";
+ case SHT_GNU_INCREMENTAL_INPUTS: return "GNU_INCREMENTAL_INPUTS";
case 0x6ffffff0: return "VERSYM";
+ case SHT_GNU_ATTRIBUTES: return "GNU_ATTRIBUTES";
+ case SHT_GNU_HASH: return "GNU_HASH";
+ case SHT_GNU_LIBLIST: return "GNU_LIBLIST";
case 0x6ffffffc: return "VERDEF";
case 0x7ffffffd: return "AUXILIARY";
case 0x7fffffff: return "FILTER";
- case SHT_GNU_LIBLIST: return "GNU_LIBLIST";
default:
if ((sh_type >= SHT_LOPROC) && (sh_type <= SHT_HIPROC))
@@ -5595,6 +5607,7 @@ static struct option options[] =
{"help", no_argument, 0, 'H'},
{"file-header", no_argument, 0, 'h'},
{"histogram", no_argument, 0, 'I'},
+ {"display-section", required_argument, 0, 'j'},
{"lint", no_argument, 0, 'L'},
{"enable-checks", no_argument, 0, 'L'},
{"program-headers", no_argument, 0, 'l'},
@@ -5729,6 +5742,9 @@ usage (FILE * stream)
Dump the relocated contents of section <number|name>\n"));
fprintf (stream, _("\
-z --decompress Decompress section before dumping it\n"));
+ fprintf (stream, _("\n\
+ -j --display-section=<name|number>\n\
+ Display the contents of the indicated section. Can be repeated\n"));
fprintf (stream, _("\
-w --debug-dump[a/=abbrev, A/=addr, r/=aranges, c/=cu_index, L/=decodedline,\n\
f/=frames, F/=frames-interp, g/=gdb_index, i/=info, o/=loc,\n\
@@ -5893,7 +5909,7 @@ parse_args (struct dump_data *dumpdata, int argc, char ** argv)
usage (stderr);
while ((c = getopt_long
- (argc, argv, "ACDHILNPR:STU:VWXacdeghi:lnp:rstuvw::x:z", options, NULL)) != EOF)
+ (argc, argv, "ACDHILNPR:STU:VWXacdeghi:j:lnp:rstuvw::x:z", options, NULL)) != EOF)
{
switch (c)
{
@@ -5976,6 +5992,9 @@ parse_args (struct dump_data *dumpdata, int argc, char ** argv)
do_follow_links = true;
dump_any_debugging = true;
break;
+ case 'j':
+ request_dump (dumpdata, AUTO_DUMP);
+ break;
case 'x':
request_dump (dumpdata, HEX_DUMP);
break;
@@ -8877,6 +8896,86 @@ static struct
{ "PLT", DT_JMPREL, DT_PLTRELSZ, reltype_unknown }
};
+static relocation_type
+rel_type_from_sh_type (unsigned int sh_type)
+{
+ switch (sh_type)
+ {
+ case SHT_RELA: return reltype_rela;
+ case SHT_REL: return reltype_rel;
+ case SHT_RELR: return reltype_relr;
+ default: return reltype_unknown;
+ }
+}
+
+static bool
+display_relocations (Elf_Internal_Shdr * section,
+ Filedata * filedata)
+{
+ if (section->sh_type != SHT_RELA
+ && section->sh_type != SHT_REL
+ && section->sh_type != SHT_RELR)
+ return false;
+
+ uint64_t rel_size = section->sh_size;
+
+ if (rel_size == 0)
+ return false;
+
+ if (filedata->is_separate)
+ printf (_("\nIn linked file '%s' relocation section "),
+ filedata->file_name);
+ else
+ printf (_("\nRelocation section "));
+
+ if (filedata->string_table == NULL)
+ printf ("%d", section->sh_name);
+ else
+ printf ("'%s'", printable_section_name (filedata, section));
+
+ uint64_t num_rela = rel_size / section->sh_entsize;
+ uint64_t rel_offset = section->sh_offset;
+
+ printf (ngettext (" at offset %#" PRIx64
+ " contains %" PRIu64 " entry:\n",
+ " at offset %#" PRIx64
+ " contains %" PRId64 " entries:\n",
+ num_rela),
+ rel_offset, num_rela);
+
+ relocation_type rel_type = rel_type_from_sh_type (section->sh_type);
+
+ if (section->sh_link == 0
+ || section->sh_link >= filedata->file_header.e_shnum)
+ /* Symbol data not available. */
+ return dump_relocations (filedata, rel_offset, rel_size,
+ NULL, 0, NULL, 0, rel_type,
+ false /* is_dynamic */);
+
+ Elf_Internal_Shdr * symsec = filedata->section_headers + section->sh_link;
+
+ if (symsec->sh_type != SHT_SYMTAB
+ && symsec->sh_type != SHT_DYNSYM)
+ return false;
+
+ Elf_Internal_Sym * symtab;
+ uint64_t nsyms;
+ uint64_t strtablen = 0;
+ char * strtab = NULL;
+
+ if (!get_symtab (filedata, symsec, &symtab, &nsyms, &strtab, &strtablen))
+ return false;
+
+ bool res = dump_relocations (filedata, rel_offset, rel_size,
+ symtab, nsyms, strtab, strtablen,
+ rel_type,
+ symsec->sh_type == SHT_DYNSYM);
+ free (strtab);
+ free (symtab);
+
+ return res;
+}
+
/* Process the reloc section. */
static bool
@@ -8968,72 +9067,8 @@ process_relocs (Filedata * filedata)
i < filedata->file_header.e_shnum;
i++, section++)
{
- if ( section->sh_type != SHT_RELA
- && section->sh_type != SHT_REL
- && section->sh_type != SHT_RELR)
- continue;
-
- rel_offset = section->sh_offset;
- rel_size = section->sh_size;
-
- if (rel_size)
- {
- relocation_type rel_type;
- uint64_t num_rela;
-
- if (filedata->is_separate)
- printf (_("\nIn linked file '%s' relocation section "),
- filedata->file_name);
- else
- printf (_("\nRelocation section "));
-
- if (filedata->string_table == NULL)
- printf ("%d", section->sh_name);
- else
- printf ("'%s'", printable_section_name (filedata, section));
-
- num_rela = rel_size / section->sh_entsize;
- printf (ngettext (" at offset %#" PRIx64
- " contains %" PRIu64 " entry:\n",
- " at offset %#" PRIx64
- " contains %" PRId64 " entries:\n",
- num_rela),
- rel_offset, num_rela);
-
- rel_type = section->sh_type == SHT_RELA ? reltype_rela :
- section->sh_type == SHT_REL ? reltype_rel : reltype_relr;
-
- if (section->sh_link != 0
- && section->sh_link < filedata->file_header.e_shnum)
- {
- Elf_Internal_Shdr *symsec;
- Elf_Internal_Sym *symtab;
- uint64_t nsyms;
- uint64_t strtablen = 0;
- char *strtab = NULL;
-
- symsec = filedata->section_headers + section->sh_link;
- if (symsec->sh_type != SHT_SYMTAB
- && symsec->sh_type != SHT_DYNSYM)
- continue;
-
- if (!get_symtab (filedata, symsec,
- &symtab, &nsyms, &strtab, &strtablen))
- continue;
-
- dump_relocations (filedata, rel_offset, rel_size,
- symtab, nsyms, strtab, strtablen,
- rel_type,
- symsec->sh_type == SHT_DYNSYM);
- free (strtab);
- free (symtab);
- }
- else
- dump_relocations (filedata, rel_offset, rel_size,
- NULL, 0, NULL, 0, rel_type, false /* is_dynamic */);
-
- found = true;
- }
+ if (display_relocations (section, filedata))
+ found = true;
}
if (! found)
@@ -14098,6 +14133,77 @@ print_symbol_table_heading (void)
}
}
+static bool
+dump_symbol_section (Elf_Internal_Shdr * section,
+ Filedata * filedata)
+{
+ if (section->sh_entsize == 0)
+ {
+ printf (_("\nSymbol table '%s' has a sh_entsize of zero!\n"),
+ printable_section_name (filedata, section));
+ return false;
+ }
+
+ uint64_t num_syms = section->sh_size / section->sh_entsize;
+
+ if (filedata->is_separate)
+ printf (ngettext ("\nIn linked file '%s' symbol section '%s'"
+ " contains %" PRIu64 " entry:\n",
+ "\nIn linked file '%s' symbol section '%s'"
+ " contains %" PRIu64 " entries:\n",
+ num_syms),
+ filedata->file_name,
+ printable_section_name (filedata, section),
+ num_syms);
+ else
+ printf (ngettext ("\nSymbol table '%s' contains %" PRIu64
+ " entry:\n",
+ "\nSymbol table '%s' contains %" PRIu64
+ " entries:\n",
+ num_syms),
+ printable_section_name (filedata, section),
+ num_syms);
+
+ print_symbol_table_heading ();
+
+ Elf_Internal_Sym * symtab = get_elf_symbols (filedata, section, & num_syms);
+ if (symtab == NULL)
+ /* An error message will have already been displayed. */
+ return false;
+
+ char * strtab = NULL;
+ uint64_t strtab_size = 0;
+
+ if (section->sh_link == filedata->file_header.e_shstrndx)
+ {
+ strtab = filedata->string_table;
+ strtab_size = filedata->string_table_length;
+ }
+ else if (section->sh_link < filedata->file_header.e_shnum)
+ {
+ Elf_Internal_Shdr * string_sec;
+
+ string_sec = filedata->section_headers + section->sh_link;
+
+ strtab = (char *) get_data (NULL, filedata, string_sec->sh_offset,
+ 1, string_sec->sh_size,
+ _("string table"));
+ strtab_size = strtab != NULL ? string_sec->sh_size : 0;
+ }
+
+ uint64_t si;
+
+ for (si = 0; si < num_syms; si++)
+ print_symbol (filedata, si, symtab, section, strtab, strtab_size);
+
+ free (symtab);
+
+ if (strtab != filedata->string_table)
+ free (strtab);
+
+ return true;
+}
+
/* Dump the symbol table. */
static bool
@@ -14152,74 +14258,13 @@ process_symbol_table (Filedata * filedata)
i < filedata->file_header.e_shnum;
i++, section++)
{
- char * strtab = NULL;
- uint64_t strtab_size = 0;
- Elf_Internal_Sym * symtab;
- uint64_t si, num_syms;
-
if ((section->sh_type != SHT_SYMTAB
&& section->sh_type != SHT_DYNSYM)
|| (!do_syms
&& section->sh_type == SHT_SYMTAB))
continue;
- if (section->sh_entsize == 0)
- {
- printf (_("\nSymbol table '%s' has a sh_entsize of zero!\n"),
- printable_section_name (filedata, section));
- continue;
- }
-
- num_syms = section->sh_size / section->sh_entsize;
-
- if (filedata->is_separate)
- printf (ngettext ("\nIn linked file '%s' symbol section '%s'"
- " contains %" PRIu64 " entry:\n",
- "\nIn linked file '%s' symbol section '%s'"
- " contains %" PRIu64 " entries:\n",
- num_syms),
- filedata->file_name,
- printable_section_name (filedata, section),
- num_syms);
- else
- printf (ngettext ("\nSymbol table '%s' contains %" PRIu64
- " entry:\n",
- "\nSymbol table '%s' contains %" PRIu64
- " entries:\n",
- num_syms),
- printable_section_name (filedata, section),
- num_syms);
-
- print_symbol_table_heading ();
-
- symtab = get_elf_symbols (filedata, section, & num_syms);
- if (symtab == NULL)
- continue;
-
- if (section->sh_link == filedata->file_header.e_shstrndx)
- {
- strtab = filedata->string_table;
- strtab_size = filedata->string_table_length;
- }
- else if (section->sh_link < filedata->file_header.e_shnum)
- {
- Elf_Internal_Shdr * string_sec;
-
- string_sec = filedata->section_headers + section->sh_link;
-
- strtab = (char *) get_data (NULL, filedata, string_sec->sh_offset,
- 1, string_sec->sh_size,
- _("string table"));
- strtab_size = strtab != NULL ? string_sec->sh_size : 0;
- }
-
- for (si = 0; si < num_syms; si++)
- print_symbol (filedata, si, symtab, section,
- strtab, strtab_size);
-
- free (symtab);
- if (strtab != filedata->string_table)
- free (strtab);
+ dump_symbol_section (section, filedata);
}
}
else if (do_syms)
@@ -17016,6 +17061,65 @@ process_section_contents (Filedata * filedata)
if (filedata->is_separate && ! process_links)
dump &= DEBUG_DUMP;
+ if (dump & AUTO_DUMP)
+ {
+ switch (section->sh_type)
+ {
+ case SHT_PROGBITS:
+ /* FIXME: There are lots of different type of section that have
+ SHT_PROGBITS set in their header - code, debug info, etc. So
+ we should check the section's name and interpret its contents
+ that way, rather than just defaulting to a byte dump. */
+#ifdef SUPPORT_DISASSEMBLY
+ res &= disassemble_section (section, filedata);
+#else
+ res &= dump_section_as_bytes (section, filedata, false);
+#endif
+ break;
+
+ case SHT_DYNSYM:
+ case SHT_SYMTAB:
+ res &= dump_symbol_section (section, filedata);
+ break;
+
+ case SHT_STRTAB:
+ res &= dump_section_as_strings (section, filedata);
+ break;
+
+ case SHT_RELA:
+ case SHT_REL:
+ case SHT_RELR:
+ res &= display_relocations (section, filedata);
+ break;
+
+ case SHT_NOTE:
+ res &= process_notes_at (filedata, section, section->sh_offset,
+ section->sh_size, section->sh_addralign);
+ break;
+
+ case SHT_NULL:
+ inform (_("Unable to display section %d - it has a NULL type\n"), i);
+ break;
+
+ case SHT_NOBITS:
+ inform (_("Unable to display section %d - it has no contents\n"), i);
+ break;
+
+ case SHT_HASH:
+ case SHT_DYNAMIC:
+ case SHT_GROUP:
+ case SHT_GNU_ATTRIBUTES:
+ /* FIXME: Implement these. */
+ /* Fall through. */
+ default:
+ /* FIXME: Add Proc and OS specific section types ? */
+ warn (_("Unable to determine how to dump section %d (type %#x)\n"),
+ i, section->sh_type);
+ res = false;
+ break;
+ }
+ }
+
#ifdef SUPPORT_DISASSEMBLY
if (dump & DISASS_DUMP)
{
@@ -17622,7 +17726,8 @@ display_arm_attribute (unsigned char * p,
static unsigned char *
display_gnu_attribute (unsigned char * p,
- unsigned char * (* display_proc_gnu_attribute) (unsigned char *, unsigned int, const unsigned char * const),
+ unsigned char * (* display_proc_gnu_attribute)
+ (unsigned char *, unsigned int, const unsigned char * const),
const unsigned char * const end)
{
unsigned int tag;
diff --git a/binutils/testsuite/binutils-all/readelf.exp b/binutils/testsuite/binutils-all/readelf.exp
index 399b3c7f276..b91134b4326 100644
--- a/binutils/testsuite/binutils-all/readelf.exp
+++ b/binutils/testsuite/binutils-all/readelf.exp
@@ -634,3 +634,11 @@ readelf_find_size $tempfile 2
# Make sure that readelf can decode the contents.
readelf_test -wi $tempfile dw5-op.W
}
+
+# Test the -j/--display-section option.
+# Check that multiple options accumulate.
+# Check that both numbers and names can be used.
+readelf_test {-j .rela.debug_info --display-section=.rel.debug_info} $tempfile display-section.r
+readelf_test --display-section=0 $tempfile display-section.0
+
+
--- /dev/null 2024-04-10 08:22:57.640408832 +0100
+++ binutils/testsuite/binutils-all/display-section.r 2024-04-10 10:46:48.465031933 +0100
@@ -0,0 +1,4 @@
+.*: Warning: Section '.*' was not dumped because it does not exist
+
+Relocation section '.*' at offset 0x.* contains .* entries:
+#...
--- /dev/null 2024-04-10 08:22:57.640408832 +0100
+++ binutils/testsuite/binutils-all/display-section.0 2024-04-10 10:32:05.871683369 +0100
@@ -0,0 +1 @@
+.*: Info: Unable to display section 0 - it has a NULL type
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2024-04-10 11:23 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-04-10 11:23 Commit: readelf: Add --display-section option Nick Clifton
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).