Use __attribute__ ((fallthrough)) to indicate switch case fall through
instead of a comment. This ensures that the fallthrough warning is not
triggered even if the file is pre-processed (hence stripping the
comments) before it is compiled.
The actual fallback implementation is hidden behind a FALLBACK macro in
case the compiler doesn't support it.
Finally, the -Wimplict-fallthrough warning was upgraded to only allow
the attribute to satisfy it; a comment alone is no longer sufficient.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
We won't use the e_shoff value in that case because we will set
elf->state.elf[64|32].scns.cnt to zero to indicate not to read
any section header data from the file.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Add ELF_E_INVALID_ELF which is set when the ELF file data is bad.
This is different from ELF_E_INVALID_FILE which is set when the file
could not be read.
Signed-off-by: Mark Wielaard <mark@klomp.org>
gcc defaults to using struct layouts that follow the native conventions,
even if __attribute__((packed)) is given. In order to get the layout we
expect, we need to tell gcc to always use the gcc struct layout, at
least for packed structs. To do this, we can use the gcc_struct
attribute.
This is important, not only for porting to windows, but also potentially
for other platforms, as the bugs resulting from struct layout
differences are rather subtle and hard to find.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
If so, define attribute_hidden to be empty. Also, use attribute_hidden
in all places where we hide symbols. If this attribute is missing, it
simply means that we cannot hide private symbols in the binary using
attributes. This disables some optimizations and may increase the risk
of symbol name clashes with other libraries, but is not fatal.
However, we still employ linker version scripts to explicitly define
the exported symbols. This serves much of the same purpose. Also, as
all our symbols are prefixed with the library name, and "__" for
private ones, the chance of clashes is low anyway.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
Those flags are not available on all platforms, and omitting them when
not available will not cause any harm. In particular:
-z,defs disallows undefined symbols in object files. This option is
unsupported if the target binary format enforces the same condition
already. Furthermore it is only a compile time sanity check. When it is
omitted, the same binary is produced.
-z,relro instructs the loader to mark sections read-only after loading
the library, where possible. This is a hardening mechanism. If it is
unavailable, the functionality of the code is not affected in any way.
-fPIC instructs the compiler to produce position independent code. While
this is preferable to relocatable code, relocatable code also works and
may even be faster. Relocatable code might just be loaded into memory
multiple times for different processes.
-fPIE is the same thing as -fPIC for executables rather than shared
libraries.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
Since POWER8, PowerPC 64 supports Hardware Transactional Memory, which has
three special purpose registers associated to it: tfhar, tfiar, and texasr.
This commit add HTM SPRs set as known note type so it's possible to use
'readelf --notes' to inspect the HTM SPRs in a coredump file generated in
such a machines.
Signed-off-by: Gustavo Romero <gromero@linux.vnet.ibm.com>
Link them all with -z,defs,-z,relro,--no-undefined, provide complete
dependencies for the link steps, and add libeu.a to each one. libeu.a
contains useful library functionality that each of them might use. The
linker will strip unneeded symbols, so linking it in won't hurt even if
none of the functions are used.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
config.h doesn't have include guards, so including it twice is bad. We
deal with this by checking for PACKAGE_NAME, but only in some places.
Once we start using gnulib, we will need to include config.h before any
gnulib-generated headers. This is problematic if we include it
transitively through our own private headers.
In order to set a clear rule about inclusion of config.h, it is now
included in every .c file as first header, but not in any header. This
will definitely avoid double-inclusion and satisfy the condition that it
has to be included before gnulib headers. It comes at the price of
adding some redundancy, but there is no clean way to avoid this.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
__attribute__ is a GNU extension. If we want to link against the
libraries using a different compiler, it needs to be disabled. It was
already disabled in libdw.h, and this patch extends this to the other
headers. We move the defines to libelf.h as that is included in all
the others.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
When building with gcc -Os it seems we can inline read_number_entries
but if that function fails then n will not be initialized. GCC seems not
to realize that in that case n won't be used at all. Explicitly initialize
n to zero to prevent a spurious error: 'n' may be used uninitialized in
this function [-Werror=maybe-uninitialized] in that case.
https://sourceware.org/bugzilla/show_bug.cgi?id=21011
Signed-off-by: Mark Wielaard <mark@klomp.org>
When ELF section data was used, but not updated or marked as dirty and
there also existed non-dirty sections and some padding was needed between
the sections (possibly because of alignment) then elf_update might write
"fill" over some of the existing data. This happened because in that case
the last_position was not updated correctly.
Includes a new testcase fillfile that fails before this patch by showing
fill instead of the expected data in some section data. It succeeds with
this patch.
https://sourceware.org/bugzilla/show_bug.cgi?id=21199
Signed-off-by: Mark Wielaard <mark@klomp.org>
unsigned long int is not always capable to have pointer in some cases
(LLP64, for example). Return a void pointer instead. Other libelf
implementations will also make this change (or already have).
Also update the documentation to state what is created and that NULL
is returned on error (don't document that the returned value is a
pointer to the actual header created).
Signed-off-by: Akihiko Odaki <akihiko.odaki.4i@stu.hosei.ac.jp>
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Bad sh_off or sh_size could trigger a bad malloc or read. Sanity check
the header values first before trying to malloc a huge buffer or reading
any data that will certainly fail.
https://bugzilla.redhat.com/show_bug.cgi?id=1387584
Signed-off-by: Mark Wielaard <mjw@redhat.com>
There are various sanity checks that depend on knowing the file size
of the underlying ELF file which we only used when mmapping the ELF file.
Although we probably won't crash if we use pread to try to read from
the file, we still might return completely bogus data structures. This
could cause us to malloc insane amounts of memory.
Always try to get the maxsize when unknown in elf_begin.c (read_file).
https://bugzilla.redhat.com/show_bug.cgi?id=1388057
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Define/open code memrchr, rawmemchr, powerof2 and TEMP_FAILURE_RETRY if
not available through system headers.
Signed-off-by: Akihiko Odaki <akihiko.odaki.4i@stu.hosei.ac.jp>
Signed-off-by: Mark Wielaard <mjw@redhat.com>
This change also creates a new header file libeu.h to provide the
prototypes for the function of libeu. That hides the definition of function
crc32, which can conflict with zlib, from libelf. It also prevents mistakes
to refer those functions from a component which doesn't link with libeu,
such as libelf.
Signed-off-by: Akihiko Odaki <akihiko.odaki.4i@stu.hosei.ac.jp>
The testcase added to run-strip-reloc.sh for strip-compressed.o showed
a memory leak when ran under valgrind (configure --enable-valgrind).
For a mmapped ELF file when existing section data was compressed
elf_end would fail to release the new compressed data buffer assigned
to rawdata_base. For non-mapped files rawdata_base is always freed.
For decompressed data rawdata_base is released together with zdata_base.
Use the Elf_Scn flag ELF_T_MALLOCED to track whether rawdata_base
points to malloced memory and free it in elf_end even for mmapped
ELF files.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Nobody has hacked on eu-ld in a very long time. It didn't really work.
And we didn't install it by default in the spec file. Remove sources,
the build rules and any (now) unused code.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
elf[32|64]_updatenull would sanity check the e_type before allowing to
update the phdrs. This prevents creating an ET_REL file with phdrs. It
also prevents creating any vendor specific ELF file having phdrs. We
only check this when updating/writing out the file. But we would just
read such files. Don't prevent people from creating unexpected ELF files.
elflint will warn for such files.
While writing a new testcase for this another bug was found that
prevented updating a just created phdr because elf_getphdrnum would
sanity check the phdr offset in the file (which doesn't exist yet).
Fix that by only doing such a sanity check if the phdrs haven't been
read in or created yet.
This second bug should have been found by the existing elfshphehdr
test, but that test contained a typo checking elf_getphdrnum.
It tested that the called failed when there were no phdrs, but then
elf_getphdrnum should simply succeed and return zero.
https://bugzilla.redhat.com/show_bug.cgi?id=1352232
Signed-off-by: Mark Wielaard <mjw@redhat.com>
When getting section headers it is assumed that the first section
is on the first section list. However, it is possible that the
first section list only contains the zeroth section, in which
case either illegal memory access occurs or elf_nextscn()
erroneously returns NULL.
With this patch, checks are added to avoid the illegal memory
access and (if available) the second section list is looked at
to find the first section.
A new test emptyfile is added that tests adding a section to
and "empty" ELF file 32/64 class with ELF_C_RDWR[_MMAP].
Signed-off-by: John Ogness <john.ogness@linutronix.de>
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Update elf.h from glibc and recognize R_386_GOT32X, R_X86_64_GOTPCRELX
and R_x86_64_REX_GOTPCRELX as non-dynamic relocations.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
In commit c0748e "libelf: More checking of valid sh_addralign values." we
adjusted bogus alignment of data buffers if they were greater than the
offset of the data in the file. This works OK, except when there is no
data in the file. So make sure to not adjust any NOBITS sections.
Also adds a test that shows the issue and makes sure elflint is called
with --gnu in run-strip-test.sh.
https://bugzilla.redhat.com/show_bug.cgi?id=1303845
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Older glibc elf.h might not define the new ELF compression defines and
types. If not just define them in libelf.h directly to make the libelf
headers work on older glibc systems.
Also include a testcase to check the libelf headers build against the
system elf.h.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=810885
Signed-off-by: Mark Wielaard <mjw@redhat.com>
elf_strptr indexes into the section data. This is defined as index into
the uncompressed data of the section. If the section is compressed make
sure the uncompressed data is available, but don't really decompress the
section header (elf_getdata will still return compressed data).
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Elf_Data of a compressed section has type ELF_T_CHDR. This type can be
xlated to the file or memory representation. This will make sure the Chdr
is in the correct endianess. The compressed data following the Chdr isn't
translated.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
* Move nested functions to file scope
in libelf/elf_begin.c and elf32_updatefile.c
to compile with clang.
Signed-off-by: Chih-Hung Hsieh <chh@google.com>
Some systems don't have loff_t, like FreeBSD where off_t always supports
large files. We need a standardized 64-bit signed type for the public
header, without depending on configuration... OK, just use int64_t.
Signed-off-by: Josh Stone <jistone@redhat.com>
Don't use posix_fallocate when not using mmap. It is enough to ftruncate
the file to the right size. pwrite will report an error if there is no
disk space left. And on file systems that don't support fallocate it
might duplicate writes in that case. When using posix_fallocate do ignore
most errors. Other libc implementations don't guarantee the call actually
works always and even with glibc there might be an unexpected error from
the fallback code when the file system doesn't support fallocate. That is
pretty nasty since we might get a SIGBUS in that case when writing to the
mmapped memory. But the chance of that happening is very small. And will
normally never happen with glibc. So only report an error when
posix_fallocate reports ENOSPC.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
AC_SYS_LARGEFILE defines _FILE_OFFSET_BITS in config.h if needed for
LFS, and this automatically maps things like open to open64. But quite
a few places used explicit 64-bit names, which won't work on platforms
like FreeBSD where off_t is always 64-bit and there are no foo64 names.
It's better to just trust that AC_SYS_LARGEFILE is doing it correctly.
But we can verify this too, as some file could easily forget to include
config.h. The new tests/run-lfs-symbols.sh checks all build targets
against lfs-symbols (taken from lintian) to make sure everything was
implicitly mapped to 64-bit variants when _FILE_OFFSET_BITS is set.
Signed-off-by: Josh Stone <jistone@redhat.com>
Note, elfutils does not explicitly enable AM_SILENT_RULES. It's only
available starting from automake 1.11, but starting from automake 1.13
silent rules are always generated, defaulting to verbose. $(AM_V_foo)
additions should be no-ops on systems that don't support silent rules.
To be silent, use "./configure --enable-silent-rules" or "make V=0".
Signed-off-by: Josh Stone <jistone@redhat.com>
This avoids relocation overflows in sparc/sparc64 targets while
linking, where the reachable data using -fpic is only 4kb.
Signed-off-by: Jose E. Marchesi <jose.marchesi@oracle.com>
When e_version is EV_NONE we should set it to EV_CURRENT like we do for
the EI_VERSION and like we set EI_DATA to the correct byte order when set
to ELFDATANONE. Likewise we should always set e_shentsize like we do for
e_phentsize, not just when ELF_F_LAYOUT isn't set.
Add a new elfshphehdr testcase to check the above.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Since we banned old style function definitions GCC is able to diagnose
function definitions that don't match the function declaration:
elf32_getehdr.c:78: error: conflicting types for ‘__elf64_getehdr_wrlock’
libelfP.h:498: note: previous declaration of ‘__elf64_getehdr_wrlock’
This happens on i386 because there internal functions are marked with:
# define internal_function __attribute__ ((regparm (3), stdcall))
Make sure all internal function declarations and definitions are marked
with internal_function.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
Andreas Schwab