1
0
Fork 0
mirror of https://github.com/tobast/libunwind-eh_elf.git synced 2024-11-01 07:39:26 +01:00
Commit graph

3024 commits

Author SHA1 Message Date
Kostik Belousov
1e6959a769 Shut down compiler warning about format mismatch. 2012-03-13 13:17:07 +02:00
Kostik Belousov
e8eb583450 Handle byteswap.h and endian.h inclusion. 2012-03-13 13:09:25 +02:00
Arun Sharma
7e5d2791f9 Update namespace checks for dwarf_find_unwind_tables 2012-03-12 20:11:49 -07:00
Arun Sharma
6da4dde70e Compile fix for platforms other than linux-i386 2012-03-12 20:11:49 -07:00
Arun Sharma
f0bb806ecb Move the test case to tests dir. 2012-03-12 20:11:49 -07:00
Arun Sharma
0f17b52171 Fixup regnames on x64 2012-03-12 20:11:49 -07:00
Arun Sharma
b08ae72d6c Add support for unwinding coredump images
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
2012-03-12 20:11:08 -07:00
Arun Sharma
d276b7a961 Introduce get_proc_name_in_image()
This is similar to get_proc_name() but takes an already mapped
image.

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
2012-03-12 20:09:18 -07:00
Arun Sharma
6525714c27 Move some dwarf functionality from ptrace to dwarf
Move ptrace-independent code from src/ptrace/_UPT_find_proc_info.c
to src/dwarf/{G,L}find_unwind_table.c. Name this moved function
dwarf_find_unwind_table().

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
2012-03-12 20:06:37 -07:00
Arun Sharma
25ee9f8172 Introduce struct elf_dyn_info
This is a common part of struct UPT_info
and struct UCD_info (to be introduced later).
Make _UPTi_find_unwind_table function operate only on this part
of struct UPT_info.

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
2012-03-12 20:04:19 -07:00
Arun Sharma
3d08506936 Fix undefined symbol: _Uelf64_valid_object
Without this change, we end up building libunwind-${arch}.a with
missing symbols.

Signed-off-by: Jan Kratochvil <jan.kratochvil@redhat.com>
2012-02-12 19:53:38 -08:00
Konstantin Belousov
59328832f7 Provide the FreeBSD implementation for _UI_siglongjmp_cont on x86.
On FreeBSD, _NSIG is at least 128, so just pass a pointer to the
signal mask in the register.
Add my copyright.
2012-02-05 11:47:47 +02:00
Konstantin Belousov
1774384098 FreeBSD also supports PT_GNU_STACK. 2012-02-05 11:42:02 +02:00
Konstantin Belousov
29dae2171e Style. 2012-02-05 11:40:17 +02:00
Konstantin Belousov
29b32cf142 Fix silly mistakes in the definitions related to the FreeBSD offsets in
jmpbuf on x86.
2012-02-05 11:39:19 +02:00
Tommi Rantala
51445f61d3 ARM: fix unwind table search when IP less than first table entry
Properly return error when searching for unwind table entry for an IP
that is less than the first available table entry.
2012-02-01 07:57:15 -08:00
Tommi Rantala
ec4b23bf70 ARM: fix off-by-one in exidx VFP pop
The range of VFP registers we need to pop is inclusive.
2012-01-24 20:43:55 -08:00
Tommi Rantala
458c661be7 Fix libunwind-generic.so symbolic link never being created in make install
Commit 7d43108f9c ("No
libunwind-generic.so if configured with --disable-shared") introduced a
check to avoid creating a broken libunwind-generic.so link, but the
result of the commit is that libunwind-generic.so is never created (at
least when installing to a clean directory).

We need to check for the installed libunwind-$(arch).so file,
libunwind-generic.so will be the symbolic link name.
2012-01-23 07:38:59 -08:00
Arun Sharma
52ca68c770 Fix a race condition
There is a window of time between the munmap and the tls_cache being
marked as destroyed, where there could be a bad access to memory that
has been unmapped/freed. Reorder the code a bit to close the window.

Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
2011-12-16 10:45:51 -08:00
Arun Sharma
cf2f3d3b75 Correct the inverted check
Typo when manually applying the previous commit.
2011-12-02 08:35:00 -08:00
Arun Sharma
2f328202ee Fix a compiler warning
Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
2011-12-01 22:08:08 -08:00
Konstantin Belousov
2b606faa21 Fix compilation on non-glibc machines.
Commit 297d9cd07d (Fix for failing test-setjmp)
breaks non glibc systems, since __GLIBC_PREREQ is not defined there.
As a consequence, preprocessor aborts with an error.

Trying to hide __GLIBC_PREREQ under #ifdef __GLIBC would require
either code duplication, or moving the longjmp implementation into
the separate file, which is included twice. In fact, I am not sure
in any use of the __GLIBC_PREREQ at the compile time, because the
compiled code can be run on the later version of glibc.

Below is the patch, tested on FreeBSD x86/x86_64 and Scientific Linux 6.1
x86_64. I compile the code always, but keep it in under unused static
symbol. In principle, the code could be optimized out by linker.

[ Minor formatting edits: asharma@fb.com ]
2011-11-30 21:59:09 -08:00
Arun Sharma
1010880548 Address x86_64 crashes when using sigaltstack
The crashes were tracked down to f->rpb_cfa_offset being incorrect.

The problem is that {rsp,rbp}_cfa_offset only have 15 bits, but for
SIGRETURN frame they are filled with:

// src/x86_64/Gstash_frame.c

   f->cfa_reg_offset = d->cfa - c->sigcontext_addr;
   f->rbp_cfa_offset = DWARF_GET_LOC(d->loc[RBP]) - d->cfa;
   f->rsp_cfa_offset = DWARF_GET_LOC(d->loc[RSP]) - d->cfa;

The problem is that the delta here can be arbitrarily large when
sigaltstack is used, and can easily overflow the 15 and 30-bit fields.

When signal handler starts running, the stack layout is:

 ... higher addresses ...
        ucontext
 CFA->
        __restore_rt (== pretcode in rt_sigframe from
                      linux-2.6/arch/x86/include/asm/sigframe.h)
 SP ->
       ... sighandler runs on this stack.

 ... lower addresses ...

This makes it very convenient to find ucontext from the CFA.

Attached patch re-tested on Linux/x86_64, no new failures.

Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
Reviwed-by: Lassi Tuura <lat@cern.ch>
2011-11-27 18:34:38 -08:00
Paul Pluzhnikov
297d9cd07d Fix for failing test-setjmp
On Mon, Nov 14, 2011 at 5:02 PM, Paul Pluzhnikov <ppluzhnikov@google.com> wrote:

> P.S. test-setjmp is failing for me (before or after the patch).
> When I enable assertions (to confirm my new assertions are correct), I see:
>
>  lt-test-setjmp: ../../src/dwarf/Gparser.c:754: apply_reg_state: \
>    Assertion `rs->reg[17].where == DWARF_WHERE_EXPR' failed.
>
> which likely explains that failure.

The problem is actually two-fold:

First, the loops in {sig,}longjmp.c are "do { ... } while (unw_step() >= 0);"

But unw_step() returns 0 on reaching the end of the chain (_start),
and the loop should stop there.

The second problem is that with this commit:
  c67da0b50e
glibc obfuscates value of SP in jmp_buf, so we might as well just give up.

Patch attached.

Thanks,
--
Paul Pluzhnikov
2011-11-19 15:37:35 -08:00
Ken Werner
d84e5d5d24 Call dwarf_step only if CONFIG_DEBUG_FRAME is defined
The ARM EABI does not use the .eh_frame and .eh_frame_hdr sections for unwinding. Therefore it doesn't make sense to call dwarf_step if CONFIG_DEBUG_FRAME is not defined.
2011-11-19 15:37:23 -08:00
Arun Sharma
786882ff8c Remove unused variable segbase 2011-11-12 10:02:09 -08:00
Ken Werner
1e644990c9 Only declare the start/end variables in case CONFIG_DEBUG_FRAME is defined.
Prevents compiler warnings about unused variables.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-11-12 09:47:12 -08:00
Ken Werner
25aeae9adf Fix breakage when host != target
Define the sigreturn syscall numbers instead of including on <sys/syscall.h>.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-11-03 08:24:18 -07:00
Arun Sharma
d20df8b318 Fix up the breakage when host != target 2011-10-31 22:33:47 -07:00
Arun Sharma
851f14281c Fix some compiler warnings 2011-10-29 17:18:49 -07:00
Ken Werner
3977724e82 Add missing <signal.h> include directive
The testcase tests/Gtest-dyn1.c uses the signal() function and should
therefore include the corresponding header file.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-29 17:12:43 -07:00
Ken Werner
91494b732b Define GNU and processor specific values for the Phdr p_type field
Define GNU and processor specific values for the Phdr p_type field in case
they aren't defined by <elf.h> already.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-29 17:12:43 -07:00
Ken Werner
0eba2169fb [ARM] Add support for systems that don't support ucontext.h
Define unw_tdep_context rather than using ucontext_t in order to support
systems that lack ucontext.h. Note that POSIX.1-2008 removed getcontext,
makecontext and swapcontext from its specification.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-29 17:12:43 -07:00
Ken Werner
059676cb00 Prevent the use of struct dl_phdr_info outside of dl_iterate_phdr
Since the dl_iterate_phdr is required for local unwinding only the use of
struct dl_phdr_info can be eliminated in case libunwind gets compiled for
remote unwinding. This enhances libunwinds portability to targets that
don't provide any dl_iterate_phdr functionality.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-29 17:12:43 -07:00
Ken Werner
0fed502a81 UNW_REMOTE_ONLY fixes
Exclude <link.h> because it is only required for local unwinding when
iterating over the program headers.

Have the following DWARF related functions available in case of
UNW_REMOTE_ONLY because they are used by libunwind-ptrace:
  dwarf_find_debug_frame
  locate_debug_info
  find_binary_for_address
  load_debug_frame
  debug_frame_tab_new
  debug_frame_tab_append
  debug_frame_tab_shrink
  debug_frame_tab_compare

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-29 17:12:43 -07:00
Ken Werner
2c865b6ed3 Include <sys/mman.h> to have MAP_ANONYMOUS defined
This change prevents libunwind_i.h from using a self-defined MAP_ANONYMOUS and
therefore avoids collisions in case the system header gets pulled in later.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-29 17:12:42 -07:00
Arun Sharma
0a26727ea2 Fix TLS destructor ordering problems
Glibc calls thread-specific dtors in the order in which the keys were added,
so the first dtor is the trace_cache_free() one. Then thread-specific
data for some other key is free()d, which calls into unw_backtrace(),
which uses dangling cache and munmapped cache->frames.

[ Minor rename + compiler warning fix: asharma@fb.com ]
Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
2011-10-29 17:12:36 -07:00
Arun Sharma
08077a4962 pthread_once() workaround for FreeBSD and Solaris
On FreeBSD, as well as on the Solaris < 10, weak pthread_once stub is
always exported from libc. But it does nothing, which means that if
threaded library is not loaded, then pthread_once() call do not actually
call the initializer finction. The construct
  if (likely (pthread_once != 0))
  {
    pthread_once(&trace_cache_once, &trace_cache_init_once);
then fails to initialize the trace cache on x86_64.

Work around by checking that the initializer was indeed called.
Note that this can break if libthr is loaded dynamically, but my belief
is that there is no platforms which allow dynamic loading of the threading
library.
2011-10-29 16:53:30 -07:00
Ken Werner
9a311bf48b [ARM] fix for the check-namespace test case
Add ARM specific symbols and ignore some symbols generated by the ARM Linux
default linker script.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-15 23:10:28 -07:00
Ken Werner
7f41e701e3 Have tdep_get_elf_image available in case of UNW_REMOTE_ONLY
Because it gets used by libunwind-ptrace.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-15 23:10:28 -07:00
Ken Werner
da4743f52a Exclude arm_find_proc_info and arm_put_unwind_info in case of UNW_REMOTE_ONLY
Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-15 23:10:28 -07:00
Ken Werner
c01ea3f6d6 Have dwarf.h include pthread.h
This is required for the pthread_mutex_t type that is beeing used in the
dwarf_rs_cache structure.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-15 23:10:28 -07:00
Ken Werner
197571915c Change _UPTi_find_unwind_table to return an integer.
This keeps the definition on IA64 and all the other architectures in sync with
the declaration of _UPTi_find_unwind_table. This also mimics the behaviour of
what's done for local unwinding and allows the function to provide more than
one way to undwind.

Signed-off-by: Ken Werner <ken.werner@linaro.org>
2011-10-15 23:10:27 -07:00
Arun Sharma
c4133e879b undef FORTIFY_SOURCE
It's a bit annoying to undefine at configure or compile time.

Signed-off-by: Arun Sharma <asharma@fb.com>
2011-10-15 23:10:00 -07:00
Arun Sharma
42b67080eb Fix linker errors with recent binutils
Without this, tests that called unw_getcontext() got a linker error

Signed-off-by: Arun Sharma <asharma@fb.com>
2011-10-15 23:03:46 -07:00
Arun
f89fb17695 Fix incorrect calls to memset.
Found when compiling libunwind with clang.

Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
2011-10-02 22:43:28 -07:00
Arun Sharma
962366dace Fixup compile errors on ia64.
Suggested-by: Harald Servat <harald.servat@bsc.es>
2011-09-23 11:11:34 -07:00
Arun Sharma
ac0acd06f3 Bump up the library version 2011-09-10 19:57:59 -07:00
Arun Sharma
1d8e2c1ae5 Explicitly specify the libraries
Without this linking fails with gold and newer versions of ld.
2011-09-10 19:57:10 -07:00
Arun Sharma
3bf76e8d19 Installed include files can't include config.h
config.h is not copied to $DESTDIR after make install. ms-abi users
will need to manually define it for now.
2011-09-05 10:47:13 -07:00