1
0
Fork 0
mirror of https://github.com/tobast/libunwind-eh_elf.git synced 2024-11-08 02:18:11 +01:00
Commit graph

24 commits

Author SHA1 Message Date
Martin Milata
d93d96ad83 Fix compilation on IA64
- Add tdep macro for {dwarf,ia64}_find_unwind_table so that ia64
   doesn't try to use dwarf code.
 - Fix extraneous #if.
 - Fix mistyped filename in Makefile.am.
 - Link ia64-specific tests with correct libraries.

Signed-off-by: Martin Milata <mmilata@redhat.com>
2012-05-23 20:52:28 -07:00
Arun Sharma
e09f9701ff Handle register nums > 16 on x86_64
gcc generates them when using ms-abi. Support disabled by
default since it increases the cache footprint of the library.
2011-06-18 20:48:49 -07:00
Lassi Tuura
ae5c1f2adf Performance optimisations for fast trace.
Insert static branch prediction predicates in useful places and avoid
unnecessary code in the hottest paths. Bypass unnecessary indirect
calls, in particular to access_mem(), when known to be safe.
2011-04-17 20:34:38 -07:00
Arun Sharma
e2962af9d3 Implement a cheaper getcontext()
Since the fast unwinding code path doesn't need the full context,
a faster target dependent getcontext is implemented.

Signed-off-by: Lassi Tuura <lat@cern.ch>
2011-04-05 22:07:05 -07:00
Lassi Tuura
5f38f35d5d Drop a call frame in tdep_trace and avoid a call to unw_step.
Dropping the extra frame for unw_backtrace itself using unw_step is
approximately 15% slower than skipping the frame in tdep_trace.  So
drop the frame in the latter, and make the function a private
implementation detail for libunwind, not an exported interface.

Also moves unw_getcontext call back into unw_backtrace to avoid an
extra call frame in case slow_backtrace does not get inlined into
unw_backtrace.
2011-04-01 00:00:39 -07:00
Lassi Tuura
9e98f15e9a Fast back-trace for x86_64 for only collecting the call stack.
Adds new function to perform a pure stack walk without unwinding,
functionally similar to backtrace() but accelerated by an address
attribute cache the caller maintains across calls.
2011-03-24 22:33:17 -07:00
Lassi Tuura
28f33c8ce0 Auto-detect whether to use msync() or mincore() for address validation. 2011-03-24 21:02:28 -07:00
Arun Sharma
1787a2fd28 Add path and pathlen arguments to tdep_get_elf_image()
Signed-off-by: Andris Zeila <andris.zeila@accenture.com>
2010-05-15 12:14:09 -07:00
Lassi Tuura
dac2d001af Identify signal frames by augmentation attribute. 2010-04-24 17:18:26 -07:00
Konstantin Belousov
71c4161247 Remove extra empty line 2010-03-13 18:33:22 +02:00
Konstantin Belousov
71cbdc697f Provide real values for jmpbuf offsets for FreeBSD 2010-03-08 21:49:38 +02:00
Konstantin Belousov
752ce15c4f Preliminary version of freebsd syscall unwinder 2010-03-08 19:45:54 +02:00
Konstantin Belousov
c64723835c Support walk over freebsd sigframes 2010-03-07 21:53:01 +02:00
Arun Sharma
649f1fb344 [X86-64] For local unwinding, we have a defence mechanism against
bad/missing unwind information, which could result in libunwind
dereferencing bad pointers. This mechanism is based on msync(2) system
call and significantly reduces the chances of a bad pointer
dereference in libunwind.

The original idea was to turn this mechanism on only when necessary
i.e. libunwind didn't find proper unwind information for a IP.

There are a couple of problems in the current implementation.

* The flag is global and is modified without locking
* The flag isn't reset when starting a new unwind

The attached patch makes ->validate a per-thread setting by moving it
into struct cursor from unw_local_addr_space and resets it to false
when starting a new unwind. As a result, cursor->as_arg points to the
cursor itself instead of the ucontext (for the local case).

This was found to reduce the number of msync() system calls from an
application using libunwind significantly.

Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
Signed-off-by: Arun Sharma <arun.sharma@google.com>
2008-06-16 14:35:53 -06:00
Daniel Jacobowitz
3842dac733 Add initial ARM and MIPS support. To support this, also enable the
reading of .debug_frame sections (used in lieu of .eh_frame sections
when they're not available).
2008-02-04 17:16:37 -07:00
Arun Sharma (अरुण)
795529eee3 [x86-64] Add a known good address cache.
This improves the performance of Lperf-simple from 500ns to 195ns
on the avg.

Signed-off-by: Arun Sharma <arun.sharma@google.com>
2006-08-10 21:30:01 -06:00
David Mosberger-Tang
1fdb08e5fc Various fixes to get x86 to compile and work again.
include/dwarf.h: Declare dwarf_reg_state_pool and dwarf_cie_info_pool.
include/dwarf_i.h: Include libunwind_i.h instead of tdep.h.
	Make dwarf_to_unw_regnum() a macro so it doesn't get compiled
	into an object file merely because it include dwarf_i.h (important
	when optimization is turned off).
	(dwarf_read_encoded_pointer_inlined): New function.
include/tdep-x86/libunwind_i.h: Add include of "mempool.h".
include/tdep-x86_64/libunwind_i.h: Add include of "mempool.h".
2006-07-26 22:36:03 -06:00
Arun Sharma
8cdd8bfed8 [x86-64] Make even UNW_LOCAL_ONLY accessors go through validation.
Signed-off-by: Arun Sharma <arun.sharma@google.com>
2006-07-26 21:39:46 -06:00
Arun Sharma
00db7f752a Implement a ip -> dwarf_reg_state cache.
Signed-off-by: Arun Sharma <arun.sharma@google.com>
2006-07-26 21:18:49 -06:00
David Mosberger-Tang
37be21c01e Restore more files wiped out by Bitkeeper->CVS->GIT transition. 2006-07-26 15:09:39 -06:00
hp.com!davidm
e3ac62c5f1 Rename: include/x86_64/jmpbuf.h -> include/tdep-x86_64/jmpbuf.h
(Logical change 1.294)
2005-05-20 09:48:08 +00:00
hp.com!davidm
180a1ececa Initial revision 2005-05-20 09:48:08 +00:00
hp.com!davidm
bea50a5d41 Adjust for renaming of file.
(tdep_put_unwind_info): Adjust for change of arguments.

2005/05/19 07:42:45-07:00 hp.com!davidm
Rename: include/tdep-x86_64.h -> include/tdep-x86_64/libunwind_i.h

(Logical change 1.294)
2005-05-20 09:48:08 +00:00
hp.com!davidm
e5944788ad Adjust comment.
2005/05/19 08:13:00-07:00 hp.com!davidm
Rename: include/x86_64/dwarf-config.h -> include/tdep-x86_64/dwarf-config.h

(Logical change 1.294)
2005-05-20 09:48:08 +00:00