Mark Wielaard [Sat, 21 Jul 2018 21:40:11 +0000 (23:40 +0200)]
unstrip: Handle SHT_GROUP sections in ET_REL files.
SHT_GROUP sections are put in both the stripped and debug file.
Handle correcting the symbol table/name entry of the group only once.
The testfile was generated with the gcc annobin plugin.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sat, 21 Jul 2018 15:07:12 +0000 (17:07 +0200)]
elfcompress: Don't rewrite file if no section data needs to be updated.
If the input and output file are the same and no section needs to
be updated we really don't need to rewrite the file.
Check whether any matching section is already compressed or decompressed.
Skip the section if it doesn't need to be changed. If no section data
needs updating end with success without rewriting/updating file.
With --force the file will still always be updated/rewritten even if
no section data needs to be (de)compressed.
Acked-by: Igor Gnatenko <ignatenkobrain@fedoraproject.org>
Signed-off-by: Mark Wielaard <mark@klomp.org>
Andreas Schwab [Sat, 21 Jul 2018 21:42:07 +0000 (23:42 +0200)]
backends: add abi_cfi and set_initial_registers_tid callbacks for M68K
This fixes all testsuite failures.
Signed-off-by: Andreas Schwab <schwab@linux-m68k.org>
Mark Wielaard [Sat, 21 Jul 2018 14:10:25 +0000 (16:10 +0200)]
elfcompress: Swap fchmod and fchown calls on new file.
Calling fchmod with a suid bit on a file might silently fail or the suid
bit might be slilently cleared by a call to fchown if already set. Swap
the calls so that the owner is set first and then set the suid bit.
https://bugzilla.redhat.com/show_bug.cgi?id=1607044
Reported-and-tested-by: Igor Gnatenko <ignatenkobrain@fedoraproject.org>
Signed-off-by: Mark Wielaard <mark@klomp.org>
Andreas Schwab [Thu, 19 Jul 2018 10:41:02 +0000 (12:41 +0200)]
tests: Add core and strip tests for RISC-V
Signed-off-by: Andreas Schwab <schwab@suse.de>
Andreas Schwab [Thu, 19 Jul 2018 09:54:59 +0000 (11:54 +0200)]
Fix typo in riscv_register_info
Signed-off-by: Andreas Schwab <schwab@suse.de>
Ulf Hermann [Tue, 17 Jul 2018 11:40:26 +0000 (13:40 +0200)]
Don't include sys/uio.h if we don't need it
The PID attaching only works on linux.
Andreas Schwab [Tue, 17 Jul 2018 11:18:32 +0000 (13:18 +0200)]
backends: add core_note callback for RISC-V
Signed-off-by: Andreas Schwab <schwab@suse.de>
Ulf Hermann [Mon, 16 Jul 2018 10:01:21 +0000 (12:01 +0200)]
Remove previous test files before running the next round
strip explicitly creates the new files. This will not work on windows if
the files already exist.
Also, if strip fails to create the files for any reason, we would check
the previous run's files and succeed. However, the test should fail
then.
Signed-off-by: Ulf Hermann <ulf.hermann@qt.io>
Mark Wielaard [Tue, 17 Jul 2018 10:27:51 +0000 (12:27 +0200)]
tests: Update hello_riscv64.ko test file with debuginfo.
Also document CONFIG_DEBUG_INFO=y.
Updated test file provided by Andreas Schwab <schwab@suse.de>.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Andreas Schwab [Wed, 11 Jul 2018 10:43:06 +0000 (12:43 +0200)]
backends: add set_initial_registers_tid callback for RISC-V
This fixes the backtrace-dwarf and deleted tests, and lets
backtrace-native run a bit further.
Signed-off-by: Andreas Schwab <schwab@suse.de>
Mark Wielaard [Thu, 5 Jul 2018 14:24:57 +0000 (16:24 +0200)]
tests: Handle compressed sections in next_cfi testcase.
Some toolchains use compressed ELF sections by default.
This would make run-next-cfi-self.sh fail because it would try to
decode the compressed data. Fix by decompressing the section first.
https://sourceware.org/bugzilla/show_bug.cgi?id=23370
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 4 Jul 2018 14:08:05 +0000 (16:08 +0200)]
readelf: Don't shadow index function from string.h
On some ancient GCC versions (4.4.7 at least) -Wshadow warns about local
variables "shadowing" global function definitions.
readelf.c: In function ‘print_debug_addr_section’:
readelf.c:5265: error: declaration of ‘index’ shadows a global declaration
/usr/include/string.h:489: error: shadowed declaration is here
This is silly of course, but easy to work around.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Ross Burton [Wed, 4 Jul 2018 11:11:08 +0000 (12:11 +0100)]
Consolidate error.h inclusion in system.h
error.h isn't standard and so isn't part of the musl C library.
To easy future porting, consolidate the inclusion of error.h into system.h.
https://sourceware.org/bugzilla/show_bug.cgi?id=21008
Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Wed, 4 Jul 2018 11:11:07 +0000 (12:11 +0100)]
Require gawk in maintainer mode
gawk is required to build known_dwarf.h, so check for it in configure.ac.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Mark Wielaard [Wed, 4 Jul 2018 21:08:41 +0000 (23:08 +0200)]
config: Add upload-release.sh script.
Should be run after making a release to upload the new tar.bz2,
signature file and update the latest symlinks. Make sure to run
make distcheck first. Requires a shell account on sourceware.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 29 Jun 2018 19:02:18 +0000 (21:02 +0200)]
Prepare for 0.173
Set version to 0.173.
Mention new functionality in NEWS.
Update po/*.po files.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 28 Jun 2018 15:53:12 +0000 (17:53 +0200)]
libdw: Recognize zero terminator to end frame table in dwarf_next_cfi.
When the length is zero this is a the zero terminator that ends the
frame table. Return 1 (end of table) instead of -1 (error) in that case.
We cannot update next_off and don't want to caller to try again.
Add testcase for dwarf_next_cfi to show both .eh_frame and .debug_frame
tables and check consistency (FDEs should point to existing CIEs).
Also add a self check to make sure we can read the table from the just
build elfutils binaries.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 27 Jun 2018 12:41:22 +0000 (14:41 +0200)]
libdw: Allow .debug_frame only Dwarf.
.debug_frame is useful independent from the other .debug sections.
Add a simplified variant of the addrcfi testcase dwarfcfi.
dwarfcfi only uses dwarf_frame calls and no dwfl helpers.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 26 Jun 2018 14:37:21 +0000 (16:37 +0200)]
libdw: Remove dwarf_getscn_info from libdw.h and libdw.map.
This function was never actually implemented/provided by libdw.
And it doesn't look like something we really want to implement.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sat, 23 Jun 2018 16:07:18 +0000 (18:07 +0200)]
tests: Limit varlocs print_expr_block recursion depth.
This is only useful for bad DWARF where an expression block might have
an expression that refers to a DIE that contains the expression block
itself. But that might happen with bad DWARF generated by a fuzzer.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 22 Jun 2018 20:44:51 +0000 (22:44 +0200)]
libdw: Add dwarf_next_lines to read .debug_line tables without CUs.
It is sometimes useful to read .debug_line tables on their own without
having an associated CU DIE. DWARF5 line tables are self-contained.
Adjust dwarf_begin_elf to accept ELF files with just a .debug_line.
Add a new function dwarf_next_lines that returns the Dwarf_Files and
Dwarf_Lines while iterating over just the .debug_lines section. Since
we parse and cache the information it also will try to match the CU
a table is associated with. This is only necessary for DWARF4 line
tables (we will need at least the compilation dir from the CU) and
won't be done for DWARF5 line tables. It also isn't an error if there
is no associated CU (but will mean for DWARF4 line tables the dir list
and the file paths might not be complete).
A typical way to call this new function is:
Dwarf_Off off, next_off = 0;
Dwarf_CU *cu = NULL;
Dwarf_Files *files;
size_t nfiles;
Dwarf_Lines *lines;
size_t nlines;
int res;
while ((res = dwarf_next_lines (dbg, off = next_off, &next_off, &cu,
&files, &nfiles, &lines, &nlines)) == 0)
{
/* ... handle files and lines ... */
}
if (res < 0)
printf ("BAD dwarf_next_lines: %s\n", dwarf_errmsg (-1));
See libdw.h for the full documentation. For more examples on how to use
the function see the new testcases next-files and next-lines.
Also adjust the file paths for line tables missing a comp_dir.
They are no longer made "absolute" by prepending a slash '/' in front
of them. This really was not useful and didn't happen in any of the
testcases. They are now just kept relative.
Make eu-readelf --debug-dump=decodedline use dwarf_next_lines instead
of iterating over the CUs to show the (decoded) line tables. This allows
it to show decoded line tables even if there is no .debug_info section.
New tests have been added that mimic the get-files and get-lines tests
but use dwarf_next_lines instead of iterating over all CUs. They produce
identical output (modulo the CU information). Also add a new test file
that contains only a .debug_line section.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 20 Jun 2018 22:54:57 +0000 (00:54 +0200)]
libdw: Handle bogus CU length in dwarf_nextcu.
The length field could be so big that it would wrap around the next_offset.
We don't really care that length is bogus, but we don't want to use it to
calculate the next offset if it is.
Found by afl-fuzz.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 12 Jun 2018 10:54:35 +0000 (12:54 +0200)]
libelf: Don't return unaligned data returned from elf_getdata[_rawchunk].
For i386 and x86_64 we allow some unaligned data accesses.
We also return unaligned data from elf_getdata[_rawchunk].
But that might go wrong if we then access the ELF types inside.
When build with gcc -O3 for example the compiler might vectorize
loops accessing ELF words or types. The instructions used do require
the data is naturally aligned. If the function returnes unaligned
data the program will segfault and crash. This happens for example
with the code in dwfl_module_getdwarf.c that tries to iterate over
the hash buckets gotten through elf_getdata_rawchunk based on the
DT_[GNU]_HASH value.
This only happens when the underlying ELF file is mmapped, and it
is meant as optimization so that we don't have to copy data first
so that it is correctly aligned. In most cases the data is already
naturally aligned though. But it might not be for non-native ELF
files.
Given that it might even happen in our own code base and these
are public functions that can be used by code that might rely on
the data returned being correctly aligned for the ELF data type
requested just always return correctly aligned data.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Yonghong Song [Sat, 16 Jun 2018 20:02:43 +0000 (13:02 -0700)]
backends,bpf: add proper relocation support
Due to libdw does not have proper BPF relocation support,
the pahole cannot display filenames correctly for objects
with default llvm options. So we have to invent
a special option "llc -march=bpf -mattr=dwarfris" to
prevent llvm from generating cross-section dwarf relocation
records (https://reviews.llvm.org/rL326505).
The pahole related discussion is in linux netdev
mailing list (http://lists.openwall.net/netdev/2018/06/15/38, etc.)
We would like to add proper BPF relocation support
to libdw so eventually we could retire the special llc bpf
flag "-mattr=dwarfris".
The bpf relocations are defined in
llvm_repo:include/llvm/BinaryFormat/ELFRelocs/BPF.def:
ELF_RELOC(R_BPF_NONE, 0)
ELF_RELOC(R_BPF_64_64, 1)
ELF_RELOC(R_BPF_64_32, 10)
Removed the relocation type R_BPF_MAP_FD whoes name does not
confirm to llvm definition and replaced it with R_BPF_64_64.
The BPF object is just a relocatible object, not an executable or
a shared library, so assign ELF type to REL only in bpf_reloc.def.
Signed-off-by: Yonghong Song <yhs@fb.com>
Mark Wielaard [Thu, 21 Jun 2018 16:52:53 +0000 (18:52 +0200)]
libelf: Sync elf.h from glibc.
Add NT_PPC_PKEY, R_BPF_64_64 and R_BPF_64_32. Remove R_BPF_MAP_FD.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 17 Jun 2018 22:58:24 +0000 (00:58 +0200)]
libdw: Break dwarf_aggregate_size recursion because of type cycles.
Found by afl-fuzz. An array type (indirectly) referring to itself in the
DIE tree could blow up the stack when dwarf_aggregate_size was called.
Limit the recursion depth to MAX_DEPTH (256) entries.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 17 Jun 2018 22:23:48 +0000 (00:23 +0200)]
libdw: dwarf_peel_type break long chains/cycles.
Limit the number of chained modifiers to 64 (that is 8 chains for all
8 modifiers, most of which cannot be chained). This prevents loops in
the DWARF DIE DW_AT_type references.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 17 Jun 2018 22:23:23 +0000 (00:23 +0200)]
libdw: aggregate_size check NULL result from get_type.
aggregate_size can be called recursively with the result of get_type.
get_type can return NULL when dwarf_peel_type fails. Found by afl-fuzz.
dwarf_aggregate_size when called directly doesn't need a NULL check
because it calls and checks the result of dwarf_peel_type directly.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Andreas Schwab [Wed, 13 Jun 2018 10:28:00 +0000 (12:28 +0200)]
backends: add abi_cfi and register_info callbacks for RISC-V
From
https://github.com/riscv/riscv-isa-manual/raw/master/release/riscv-spec-v2.2.pdf
and GCC source.
Signed-off-by: Andreas Schwab <schwab@suse.de>
Luiz Angelo Daros de Luca [Sun, 17 Jun 2018 09:34:08 +0000 (11:34 +0200)]
libdw: Initialize filelist earlier in dwarf_getsrclines.c read_srclines.
I'm getting this error with 0.172:
dwarf_getsrclines.c: In function 'read_srclines':
dwarf_getsrclines.c:1074:7: error: 'filelist' may be used uninitialized in this function [-Werror=maybe-uninitialized]
free (filelist);
^~~~~~~~~~~~~~~
It seems that gcc is right here as there is "ifs" that go to "out"
(where filelist is freed) before freelist is initialized.
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Mark Wielaard [Fri, 15 Jun 2018 14:25:15 +0000 (16:25 +0200)]
libdw, readelf: Don't handle DW_FORM_data16 as expression block/location.
Also found by afl-fuzz on the varlocs testcase.
DW_FORM_data16 is constant form according to the DWARF5 spec.
But since it is 128bits it isn't really representable as Dwarf_Word.
So we treat it as block form. But we cannot treat it as an expression
block. Make sure readelf prints it as a regular block and that
dwarf_getlocation[s|_addr] doesn't treat it as location expression.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 12 Jun 2018 12:36:50 +0000 (14:36 +0200)]
readelf: Check there are at least 4 bytes available for DWARF_FORM_block4.
Found by afl-fuzz. When printing a DWARF_FORM_block4 we checked there
were only 2 bytes available (copy/paste from DW_FORM_block2 right
before). Obviously we need at least 4 bytes to read the length of a
DW_FORM_block4.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 12 Jun 2018 10:22:13 +0000 (12:22 +0200)]
readelf: Make sure print_form_data always consumes DW_FORM_strx[1234] data.
Found by afl-fuzz. When printing DW_FORM_strx[1234] data eu-readelf didn't
increase readp which meant eu-readelf would keep printing the same line
dirs or files encoded with strx[1234] names. This meant that for insane
large dir or file counts eu-readelf would just keep printing endlessly
because we never reached and of the .debug_line buffer.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 12 Jun 2018 10:18:31 +0000 (12:18 +0200)]
libdw: Make __libdw_dieabbrev more robust on failure.
Make sure to always set die->abbrev to DWARF_END_ABBREV on failure.
DWARF_END_ABBREV is also what the function reports on failure. And it
will prevent trying to lookup the abbrev ever again.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 13 Jun 2018 13:21:37 +0000 (15:21 +0200)]
readelf: While printing .debug_loc make sure that next_off doesn't overflow.
Found by the afl fuzzer. The next offset (after a locview) comes from a
DIE loclist attribute. This could be a bogus value so large it overflows
the buffer and makes us print past the end of buffer.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 13 Jun 2018 12:03:24 +0000 (14:03 +0200)]
readelf: Handle signedness of DW_FORM_implicit_const and DW_AT_const_value.
We only handles DW_FORM_sdata as a signed form, but DW_FORM_implicit_const
is also signed by default. For DW_AT_const_value we can do a little better.
GCC encodes some const_values with signed forms, even though the type
is unsigned. Lookup the (base) type of the DIE and display the const value
as their (signed) type/size (if we can determine that).
Add a new testcase run-readelf-const-values.sh that shows that.
With the new testcase the const values would come out as follows:
name (string) "i"
const_value (implicit_const)
18446744073709551615
name (string) "j"
const_value (implicit_const)
18446744073709551615
name (string) "sc"
const_value (sdata) -2
name (string) "uc"
const_value (sdata) -2
name (string) "ss"
const_value (sdata) -16
name (string) "us"
const_value (sdata) -16
name (string) "si"
const_value (sdata) -3
name (string) "ui"
const_value (sdata) -
94967296
name (string) "sl"
const_value (sdata) -1
name (string) "ul"
const_value (sdata) -1
With this patch they show up as:
name (string) "i"
const_value (implicit_const) -1
name (string) "j"
const_value (implicit_const) -1
name (string) "sc"
const_value (sdata) -2
name (string) "uc"
const_value (sdata) 254 (-2)
name (string) "ss"
const_value (sdata) -16
name (string) "us"
const_value (sdata) 65520 (-16)
name (string) "si"
const_value (sdata) -3
name (string) "ui"
const_value (sdata)
4200000000 (-
94967296)
name (string) "sl"
const_value (sdata) -1
name (string) "ul"
const_value (sdata)
18446744073709551615 (-1)
(for signed/unsigned int char, short and long)
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 11 Jun 2018 15:14:39 +0000 (17:14 +0200)]
Prepare for 0.172.
It has been only 10 days since the previous release and there are
no functional changes compared to 0.171. The speedup of eu-readelf -N
is pretty nice. And ~25 patches fix various bugs (hangs and crashes)
in dealing with bad DWARF5 data. Most have been found by running the
afl fuzzer on eu-readelf and various testcases.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 11 Jun 2018 09:22:36 +0000 (11:22 +0200)]
readelf: Return correct readp (or readendp) from print_form_data.
print_form_data returns the new readp (or readendp on error) to show how
much data was consumed. But when reading the .debug_str_offsets section
we would reuse readp and readendp. This meant the wrong readp would be
returned to the caller.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 11 Jun 2018 00:14:34 +0000 (02:14 +0200)]
readelf: Fix bounds check in print_form_data.
The afl fuzzer found that we did a wrong check in print_form_data when
comparing the remaining bytes in the buffer to an (unsigned) value read.
We were casting the value to ptrdiff_t which is a signed value and so
might turn a really big unsigned value into a negative number. Since we
know the difference between readendp and readp is zero or greater, we
should cast the pointer difference to size_t (and unsigned type) instead
before comparing with the unsigned value.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 10 Jun 2018 14:59:12 +0000 (16:59 +0200)]
libdw: Break long or circular DIE ref chains in dwarf_[has]attr_integrate.
Bad DWARF could create a very long or circular DIE ref chain by linking
DW_AT_abstract_origin or DW_AT_specification to the DIE itself. Break
the chain after seeing a large number (16) of DIEs.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 21:30:37 +0000 (23:30 +0200)]
readelf: Calculate max_entries instead of needed bytes (and overflowing).
The afl fuzzer found that we would overflow the needed bytes when
calculating how many index entries would fit in the .debug_loclists
and .debug_rnglists tables. To fix this just calculate the max number
of entries. If the offset entry count is larger than that, do emit
an error, but print up to max_entries of offsets (so the user can
more clearly see what is wrong with their table).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 19:13:25 +0000 (21:13 +0200)]
tests: Don't assert on bad DW_OP_GNU_parameter_ref target in varlocs.
If the target of a DW_OP_GNU_parameter_ref isn't a DW_TAG_formal_parameter
that is bad data (which varlocs should error on). But it isn't an internal
consistency check (for which varlocs should assert).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 18:45:48 +0000 (20:45 +0200)]
libdw: Check validity of dwarf_getabbrev arguments.
When the given Dwarf_Die was invalid we might crash and when the offset
was totally bogus we might succeed with a random abbrev.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 18:14:58 +0000 (20:14 +0200)]
libdw: dwarf_get_units should handle existing failure to open Dwarf.
The other dwarf unit/cu iterators handle a NULL Dwarf handle as an
existing error and return NULL. Don't crash.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 13:10:43 +0000 (15:10 +0200)]
libdw: Detect bad DWARF in store_implicit_value.
The afl fuzzer running against the varlocs test detected we didn't report
the value block of a DW_OP_implicit_value consistently when the DWARF was
bad. Although this doesn't cause a crash it might result in consumers
using dwarf_getlocation_implicit_value seeing an inconsistent block length
value. To fix this detect and report bad DWARF data earlier.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 13:07:16 +0000 (15:07 +0200)]
tests: Fix cfi_debug_bias assert in varlocs.
It is only a consistency issue if we actually have an cfi_debug and the
cfi_debug_bias is not zero (because they come from the same file as the
other debug data).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 12:04:40 +0000 (14:04 +0200)]
readelf, libdw: Handle too many directories or files in the line table better.
The afl fuzzer found that the way we handle "too many" directories or files
in the (DWARF5 style) line table badly. In the case of eu-readelf we would
print an endless stream of "bad directory" or "bad file". Just stop printing
when the end of data is reached. In the case of dwarf_getsrclines we would
allocate a giant amount of memory, even if there was no data to actually
read in. Sanity check that the directory and file counts seem reasonable
compared to the amount of data left (assume we need at least 1 byte of
data per form describing the dirs or files).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 8 Jun 2018 00:11:44 +0000 (02:11 +0200)]
libdw: Return an error in dwarf_getlocation_attr for missing .debug_addr.
When constructing a "fake" Dwarf_Attribute for DW_OP_GNU_const_index,
DW_OP_constx, DW_OP_GNU_addr_index or DW_OP_addrx, we would create a
fake attribute pointing to the actual data in the .debug_addr section.
We would even do that if there was no .debug_addr section assuming
dwarf_formaddr or dwarf_formudata would generate an error. But when
there is no .debug_addr there is also no fake_addr_cu, so the
dwarf_form* functions cannot check the value is correct (and crash).
Fix by returning an error early from dwarf_getlocation_attr indicating
bad DWARF data.
Found by the afl fuzzer running on the varlocs testcase.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 7 Jun 2018 21:09:41 +0000 (23:09 +0200)]
readelf: Turn format_print_dwarf into print_dwarf_addr.
We don't really need to setup a buffer, print into it and then print it
out to stdout. Simplify the code by directly printing the address (and
symbol name).
This also showed a small error in the output of DW_LLE_startx_length.
It had two unintended trailing dots.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 6 Jun 2018 23:23:05 +0000 (01:23 +0200)]
libdw: Make sure that address_size and offset_size are 4 or 8 bytes.
When interning a CU make sure that address_size and offset_size are
either 4 or 8 bytes. We really don't (want to) handle any other size.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 7 Jun 2018 14:57:23 +0000 (16:57 +0200)]
libdw: Check DIE address fall inside the CU before reading abbrev code.
The afl fuzzer found a case where we tried reading an uleb for the DIE
abbrev code without properly checking the DIE address is inside the CU.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 6 Jun 2018 21:51:12 +0000 (23:51 +0200)]
libdw: Report error in dwarf_getlocation_die for bogus opcode offset.
Found by afl fuzzer on varlocs test. varlocs sanity checks that the
given offset in the opcode corresponds to the cuoffset of the returned
DIE. In case the opcode offset was bogus this might fail because we
might wrap around and return a random DIE instead of reporting an error.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 6 Jun 2018 21:35:35 +0000 (23:35 +0200)]
tests: Use error, not assert, when trying to print a non-base type DIE.
When using the varlocs test with a fuzzer using assert for internal
sanity checks is great to find issues. But when encountering bad data
using an assert is wrong. Just use error to show we handle the data
correctly (by reporting it is bad, instead of crashing).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 21:34:46 +0000 (23:34 +0200)]
libdw: Explicitly check we could decode diridx in dwarf_getsrclines.
It is highly unlikely dwarf_formudata fails because we setup the attribute
ourselves, but better to explicitly mark diridx as bad if it does.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 21:21:05 +0000 (23:21 +0200)]
readelf: Always initialize .debug_addr unit_length, even without a header.
We would print a "fake" .debug_addr header, but didn't always setup the
unit_length (in case there was a mix of GNU DebugFission and DWARF5 tables).
Make sure to always set the unit_length (we do always calculate the next
unit offset already).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 20:50:43 +0000 (22:50 +0200)]
readelf: Set begin properly for DW_LLE_GNU_start_end_entry on addrx failure.
When printing the GNU DebugFission location entries we want to print the
start idx as begin, if we cannot find the address index.
A copy/paste error set up end instead of begin in that case causing us
to print garbage (in the unlikely event the .debug_addr table wasn't
found for this entry).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 20:41:45 +0000 (22:41 +0200)]
libdw: Don't leak arange if we cannot figure out which CU it belongs to.
In the unlikely case that __libdw_findcu fails to find the associated
CU we would leak one arange because it wasn't linked into the arangelist
list yet. Make sure to free it immediately.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 20:27:25 +0000 (22:27 +0200)]
libdw: Make sure dirarray is always properly freed in dwarf_getsrclines.
If there were more than 256 directories in the table and there was
illegal DWARF before we read them all, then we might not free the
dirarray (or the wrong one). Fix by defining the dirarray early
(before the first data sanity check) and making sure it is not
(still) equal to dirstack before freeing.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 19:52:46 +0000 (21:52 +0200)]
readelf: Don't leak lengths array when detecting an invalid hash chain.
In both handle_sysv_hash and handle_sysv_hash64 we check the has chain
isn't too long. If it is we would report an error and leak the lengths
array. Just clean up the array even in the error case.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 19:29:27 +0000 (21:29 +0200)]
libdw: Make sure id_path can contain max number of build id bytes.
The MAX_BUILD_ID_BYTES is fairly large (64), while normally build-ids
are only 20 bytes long. But if we would encounter a jumbo build-id we
should have enough room to construct the full build-id path.
We used to substract 2 bytes from the max, because 2 chars are used
as subdir. But that should be 1 (2 hex chars is just one 8 bit byte).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 5 Jun 2018 19:04:05 +0000 (21:04 +0200)]
readelf: Make room for DW_MACRO_hi_user opcode if used.
The vendor array should have room for all vendor opcode, including
DW_MACRO_hi_user if used.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 4 Jun 2018 20:23:48 +0000 (22:23 +0200)]
libdwfl: Make __libdwfl_addrsym a static function in dwfl_module_addrsym.c
__libdwfl_addrsym is only used in the dwfl_module_addrsym.c source.
There is no need to mark this as a (shared) internal function.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 7 Jun 2018 17:42:43 +0000 (19:42 +0200)]
readelf: Lookup "no" translation for no_str, not "yes".
Also use yes_str and no_str in print_debug_abbrev_section
and print_form_data.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sat, 2 Jun 2018 02:24:01 +0000 (04:24 +0200)]
readelf: Don't allocate string with asprintf, but reuse buffer with sprintf.
Since we are single threaded we can just use a static result buffer for
format_dwarf_addr as long as we make sure to print the result before
calling format_dwarf_addr again. This removes lots of malloc/free calls.
On my machine eu-readelf -N --debug-dump=info libxul.so
goes from 57 seconds to 55 seconds.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 4 Jun 2018 09:06:44 +0000 (11:06 +0200)]
readelf: Lookup gettext "yes" and "no" only once.
On my machine eu-readelf -N --debug-dump=info libxul.so > /dev/null
goes from 63 seconds to 57 seconds.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 3 Jun 2018 07:28:10 +0000 (09:28 +0200)]
readelf: Call __fsetlocking (stdout, FSETLOCKING_BYCALLER).
We only call printf on stdout from one thread, so we don't need internal
stdio locking for stdout.
On my machine eu-readelf -N --debug-dump=info libxul.so > /dev/null
goes from 65 seconds to 63 seconds.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sat, 2 Jun 2018 12:56:43 +0000 (14:56 +0200)]
tests: Drop shared libraries from self_test_files_exe.
Commit
00d89086 "tests: Split self_test_files into an exe, lib and obj
list" accidentially left the shared libraries in the self_test_files_exe.
Causing all shared libraries to be tested twice. Drop them and keep just
four ET_EXE files: addr2line, elfcmp, objdump and readelf.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 1 Jun 2018 19:16:56 +0000 (21:16 +0200)]
Prepare for 0.171.
Set version to 0.171. Update po/*.po files.
Mention DWARF5, split dwarf and GNU DebugFission support in NEWS.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 31 May 2018 14:57:09 +0000 (16:57 +0200)]
libdw: Try both the relative and absolute paths when finding a .dwo file.
We would give up if one of them failed. With this fixed a self-test with
make check succeeds when building elfutils itself with CFLAGS set to
"-gdwarf-4 -gdwarf-split -O2".
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 1 Jun 2018 01:46:10 +0000 (03:46 +0200)]
readelf: Fix .debug_types printing with implicit section_info.
Commit
314e9d7d "readelf: Handle .debug_info first if any other debug
section needs it" disabled section_info printing if it was already
handled. But section_types was an alias for section_info. So unless
section_info was explicitly printed, .debug_types wasn't.
Make section_types its own thing to print .debug_types and make
section_info imply section_types. Add a testcase to make sure
.debug_types is now printed.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 1 Jun 2018 00:23:37 +0000 (02:23 +0200)]
readelf: Deal with combined normal and split dwarf DebugFission .debug_loc.
Normal and split dwarf from GNU DebugFission look the same, but should
be treated competely separtely. When having a file with both skeletons
and real compile units only note the secoffsets into the real .debug_loc
in readelf. Otherwise or known_locslistptr will get confused.
Add a testfile that combines an normal -gdwarf-4 object with a
-gsplit-dwarf object. libdw already got this right, but add a
run-varlocs.sh test to make sure.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 31 May 2018 12:35:34 +0000 (14:35 +0200)]
tests: Split self_test_files into an exe, lib and obj list.
Introduce testrun_on_self_exe and testrun_on_self_lib.
Some tests cannot handle (unrelocated) ET_REL object files.
run-get-units-split.sh and run-unit-info.sh only handle executables
and shared libraries. This allows running the whole testsuite on an
elfutils build done with CFLAGS="-gdwarf-4 -gsplit-dwarf -O2".
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 31 May 2018 12:00:27 +0000 (14:00 +0200)]
tests: Run run-low_high_pc.sh testcase on split dwarf files.
Test that the low high pc attributes can be properly resolved also
in split dwarf setups.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 30 May 2018 13:51:12 +0000 (15:51 +0200)]
libdw: Fix overflow warning on 32bit systems with GCC8 in dwarf_getsrclines.
ndirs is read from the debug data and should be size checked before use.
https://sourceware.org/bugzilla/show_bug.cgi?id=23248
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Wed, 30 May 2018 09:54:31 +0000 (11:54 +0200)]
readelf: Use correct listptr when looking up next loc for locview attr.
We were using loclistsptr instead of locsptr in print_debug_loc_section.
https://sourceware.org/bugzilla/show_bug.cgi?id=23247
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 31 May 2018 11:01:39 +0000 (13:01 +0200)]
libdw: Don't crash on invalid die in dwarf_dieoffset.
Add explicit test in get-units-invalid for dwarf_cuoffset and
dwarf_dieoffset. Make sure dwarf_dieoffset returns (Dwarf_Off) -1
on failure.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 29 May 2018 21:49:21 +0000 (23:49 +0200)]
libdw: Handle split Dwarf Dies in dwarf_die_addr_die.
dwarf_die_addr_die can be used to turn an Dwarf_Die addr back into a
full Dwarf_Die, just given the original Dwarf debug handle. This now
also works for Dwarf_Dies which originated from a split Dwarf. Whenever
a split Dwarf_CU is found the Dwarf it originated from is registered
with the Dwarf that the skeleton Dwarf_CU came from. All registered
split Dwarfs are then searched by dwarf_die_addr_die if the addr didn't
match the main Dwarf or the alt Dwarf.
One limitation in this implementation is that only DIEs that come from
the main .debug_info in the .dwo are supported. Theoretically there could
also be DIEs in an .debug_type or from other/multiple (comdat) sections.
New tests are added for dwarf-4, dwarf-5, split-dwarf-4, split-dwarf-5
and version 4 and 5 dwo files.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 28 May 2018 22:59:28 +0000 (00:59 +0200)]
readelf, libdw: Add GNU DebugFission .debug_loc support.
GNU DebugFission .debug_loc location lists uses the .debug_loc section
in the split dwarf .dwo file. The encoding is a mix of old style DWARF
.debug_loc and new style .debug_loclists.
Add two testcases for the readelf and libdw decoders.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 29 May 2018 07:44:38 +0000 (09:44 +0200)]
readelf: Fix regression with multiple files and implicit debug_info reading.
Commit
314e9d7d "readelf: Handle .debug_info first if any other debug
section needs it" introduced a regression when handling multiple files.
The implicit and explicit printing of debuginfo weren't reset and so
the second file would not show (or scan) the .debug_info section when
needed.
Fix by resetting the implicit and explicit section printing flags.
Add a testcase that prints the .debug_loc section for two files and
check that the CUs are resolved.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 20 May 2018 19:57:39 +0000 (21:57 +0200)]
libdw: Fix memory corruption in libdw_find_split_unit.
Found by valgrind when trying to match a split unit from a .dwo file
that doesn't contain the split unit (as first) match. We would close
the split Dwarf too early, before we had inspected all units in it.
Add a testcase that simulates this. Which failed (at least under valgrind
as run by make distcheck) before the fix.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 30 Apr 2018 00:54:07 +0000 (02:54 +0200)]
readelf: Handle .debug_str_offsets.
The .debug_str_offsets tables are indirect string offsets into the
.debug_str section. For DWARF5 they can be in both the main, skeleton
and split dwarf (.dwo) files.
For DWARF4 with the GNU DebugFission extension the tables will not have
an header and they will only be in the split DWARF (.dwo) file, never in
the main (skeleton) file.
For DWARF5 the (non-split) unit DIE will have a DW_AT_str_offsets_base
attribute pointing at the actual index (after the header). The split
unit will never have this attribute (and use the table at offset zero).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 27 Apr 2018 22:10:41 +0000 (00:10 +0200)]
readelf handle .debug_addr section.
Add debug-dump=addr which will show the .debug_addr section tables.
The only tricky bit is the fact that GNU DebugFission, a DWARF4
extension, didn't produce unit table headers. So if we see a mixed
DWARF4/5 .debug_addr table we have to reconstruct the table length
from the CU DIE DW_AT_[GNU_]_addr_base offsets.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sat, 7 Apr 2018 21:48:27 +0000 (23:48 +0200)]
libdw: Handle .debug_loclists in dwarf_getlocation.
Handle all new DW_LLE opcodes in .debug_loclists in dwarf_getlocation.
__libdw_read_begin_end_pair_inc now also handles a default location
(which is simply the range [0,-1]). Since expression blocks can now
also come from the .debug_loclists section add a new fake_loclists_cu
necessary for checking bounds while parsing expression blocks.
Adapt varlocs test to handle debug-only files.
Test testfileranges5.debug and testfilesplitranges5.debug with it.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 24 Nov 2017 18:29:57 +0000 (19:29 +0100)]
readelf: Handle .debug_loclists.
The new DWARF5 .debug_loclists sections are like .debug_rnglists, but
plus locations. For Split Dwarf GCC generates the .debug_loclists fully
in the split .dwo file. Any references to addresses need to be resolved
through the skeleton .debug_addr section.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 27 May 2018 12:05:21 +0000 (14:05 +0200)]
libdwfl: Allow partial relocations also for debug files.
__libdwfl_relocate is called for get_dwarf and get_elf. We allow not all
relocations to be resolved for Elf files, but required all relocations
(in the debug sections) to be fully resoled in Dwarf files. This used to
mostly work out with .o ET_REL files when the main Elf was gotten before
the Dwarf. But with .dwo files, we (readelf) might open the .o file just
for the (skeleton) Dwarf. In this case it could happen not all relocations
in the debug sections could be resolved (.debug_info and .debug_addr might
contain referenes to undefined symbols). So allow partial relocations
also for debug files.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 8 Apr 2018 22:03:43 +0000 (00:03 +0200)]
readelf: Find skeleton units when inspecting split .dwo (--dwarf-skeleton).
To get the right context (especially addresses) when looking at a .dwo file
we really need the skeleton file. If we can find it (simply replace .dwo
with .o) then use that to get to the split DWARF units so that libdw sets
up all relevant information to resolve.
Also adds a --dwarf-skeleton option so the user can explicitly give a
skeleton file to use (for example when all .o files are linked and removed
already). Unfortunately this might not work if libdw cannot get from the
skeleton file to the .dwo file (because they have been moved around).
In that case eu-readelf "cheats", it will link up the libdw datastructures
so that the skeleton and split DWARF units are setup correctly anyway.
This does introduce a problem when trying to cleanup the Dwarf handle
ownership graph. So we will deliberately leak memory by not closing the
underlying Dwfl in that case.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 6 Apr 2018 16:53:34 +0000 (18:53 +0200)]
libdw: Handle .debug_rnglists in dwarf_ranges.
Handle all new DW_RLE opcodes in .debug_rnglists in dwarf_ranges. Extract
code for reading .debug_addr indexes from dwarf_formaddr as __libdw_addrx
to reuse in __libdw_read_begin_end_pair_inc. And add new testcase for
"plain" DWARF5 and add a new test all-dwarf-ranges to test split DWARF5.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 24 Nov 2017 10:16:06 +0000 (11:16 +0100)]
readelf: Add .debug_rnglists support.
Parse the .debug_rnglists section for DWARF5 --debug-dump=ranges.
Add testcase to show both "normal" and "split" DWARF variants are
handled for DWARF4 and DWARF5.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 25 May 2018 13:46:09 +0000 (15:46 +0200)]
__libdw_cu_addr_base: Cast offset to uintptr_t before returning as pointer.
Fixes a GCC compile issue on the elfutils-debian-i686 buildbot builder:
dwarf_getlocation_attr.c: In function ‘addr_valp’:
dwarf_getlocation_attr.c:62:12: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
return (unsigned char *) offset;
^
cc1: all warnings being treated as errors
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Tue, 22 May 2018 12:34:06 +0000 (14:34 +0200)]
libdw: Handle all address FORMs for dwarf_highpc, handle errors better.
dwarf_highpc can use any address FORM, not just DW_FORM_addr. Just try
whether the address can be resolved as address. Always set error when
attribute couldn't be found or resolved. When calculating the base
address for a CU don't try to second guess the error code, just treat
an error the same as the attribute not being there.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 20 May 2018 21:30:01 +0000 (23:30 +0200)]
libdw: Support DW_OP_addrx/constx and split DWARF addrx/constx support.
DW_OP_addrx/constx and GNU DebugFission DW_OP_GNU_addr/const_index take
as argument an index into the .debug_addr section for the associated CU.
This index gets resolved through dwarf_getlocation_attr. A new fake addr
CU is created per Dwarf for use with this new attribute. For split DWARF
files, the IDX_debug_addr gets replaced with the skeleton section and the
addr base is resolved immediately when constructing the split DWARF CU.
Move __libdw_cu_addr_base to libdwP.h to share with eu-readelf. Also
make it possible to resolve addrx[1234]/GNU_addr_index also as constant
indexes to (also) show when displaying these attributes in eu-readelf.
A new varlocs tests is added to test the resolving for both the DWARF4
and DWARF5 DW_OP variants. And now that addrx forms are resolved in
split DWARF files add the new DIEs with "single ranges" (those DIEs that
have a lowpc/highpc attribute pair) to run-all-dwarf-ranges.sh.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 19 Jan 2018 14:45:16 +0000 (15:45 +0100)]
readelf: print split CUs when given --debug-dump=info+
To show the difference between "regular" CUs and split CUs print
offsets and references between { and } instead of [ and ].
When --debug-dump=info+ is given (implied by -w) instead of
--debug-dump=info any skeleton unit will be immediately followed
by the corresponding split compile unit (from the .dwo file).
DWARF section [27] '.debug_info' at offset 0x1075:
[Offset]
Compilation unit at offset 0:
Version: 4, Abbreviation section offset: 0, Address size: 8, Offset size: 4
Unit type: skeleton (4), Unit id: 0xc152129eb4b99599
[ b] compile_unit abbrev: 1
ranges (sec_offset) range list [ 0]
low_pc (addr) +
000000000000000000
stmt_list (sec_offset) 0
GNU_dwo_name (strp) "foo.dwo"
comp_dir (strp) "/tmp"
GNU_pubnames (flag_present) yes
GNU_addr_base (sec_offset) 0
GNU_dwo_id (data8) 0xc152129eb4b99599
GNU_ranges_base (sec_offset) 0
Split compilation unit at offset 0:
Version: 4, Abbreviation section offset: 0, Address size: 8, Offset size: 4
Unit type: skeleton (4), Unit id: 0xc152129eb4b99599
{ b} compile_unit abbrev: 1
producer (GNU_str_index) "GNU C11 7.3.0 -gsplit-dwarf -g"
language (data1) C99 (12)
name (GNU_str_index) "foo.c"
comp_dir (GNU_str_index) "/tmp"
GNU_dwo_id (data8) 0xc152129eb4b99599
{ 18} subprogram abbrev: 2
external (flag_present) yes
name (GNU_str_index) "main"
decl_file (data1) foo.c (1)
decl_line (data1) 1
type (ref4) { 2b}
high_pc (data8) 3
frame_base (exprloc)
[ 0] call_frame_cfa
GNU_all_call_sites (flag_present) yes
{ 2b} base_type abbrev: 3
byte_size (data1) 4
encoding (data1) signed (5)
name (string) "int"
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sun, 20 May 2018 08:35:34 +0000 (10:35 +0200)]
libdw: Add new dwarf_cu_info function.
This allows getting a (split) subdie lazily, only when needed.
All arguments to dwarf_get_units are optional. When not given
then unit DIE and sub DIE are not looked up. This new function
allows them to be looked up when not immediately retrieved with
dwarf_get_units, or for a Dwarf_CU gotten through some other way.
Add a new testcase to make sure the results of calling dwarf_cu_info
and dwarf_get_units are consistent.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 24 May 2018 13:20:25 +0000 (15:20 +0200)]
libdw: Initialize ranges_base, add invalid DWARF test and fix expected output.
We never initialized the CU ranges_base, which meant we didn't actually
calculate it correctly. This caused bad ranges on some DIEs. The expected
output in the testcase was wrong. We also crashed on invalid dwarf.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Fri, 18 May 2018 22:46:02 +0000 (00:46 +0200)]
libdw: Handle GNU DebugFission split ranges.
GNU DebugFission split dwarf handles DW_FORM_sec_offset specially for
attributes that point to ranges. The .debug_ranges section is not in
the .dwo file, but in the main/skeleton object file. The sec_offset is
not relocated (in the ELF file), but is an offset against the skeleton
DIE DW_AT_GNU_ranges_base attribute. dwarf_formudata is changed so it
still looks like a normal offset ptr into the .debug_ranges section.
dwarf_ranges is adapted to look for the .debug_ranges in the main object
file. dwarf_highpc and dwarf_lowpc now handle the highpc and lowpc
attributes being inherited for the split unit DIE from the skeleton.
A new testcase is added to iterate over all ranges in a split GNU
DebugFission file.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 17 May 2018 22:26:56 +0000 (00:26 +0200)]
libdw: Handle split dwarf debuglines.
Split DWARF .dwo files do contain a .debug_line section, but only with
the file table, there is no actual line program. Also split DWARF CU DIEs
don't have a DW_AT_stmt_list attribute. To get at the file (and dir) table
for a split unit DIE take just the files from the .debug_line section
(at offset zero). To get the full line table use the skeleton DIE (which
does have a DW_AT_stmt_list).
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Thu, 17 May 2018 17:23:46 +0000 (19:23 +0200)]
libdw: Cache ELF directory early. Explicitly set it in dwfl.
The logic that finds alt files and dwo files relies on having an open
file descriptor. But after all needed ELF data has been read the
underlying Elf file descriptor can be closed. libdwfl in particular
closes file descriptor fairly aggressively. So capture the directory
early on. And make dwfl set it if it has recorded it. Which it will
do now before closing a file descriptor for the main Dwfl_Module file.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Mon, 14 May 2018 21:58:25 +0000 (23:58 +0200)]
libdw: Search skeleton DIE for split compile unit DIE attributes.
dwarf_attr_integrate and dwarf_hasattr_integrate should also search
for attributes from the skeleton DIE in case the given DIE is a split
compile unit DIE. Split compile unit DIEs inherit various attributes
from their skeleton unit DIE in DWARF5.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Mark Wielaard [Sat, 19 May 2018 16:35:51 +0000 (18:35 +0200)]
Mark __libdw_find_split_unit mark as internal_function.
The prototype and definition should match. On i386 this is important
because it changes the calling convention.
Signed-off-by: Mark Wielaard <mark@klomp.org>