summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* Port libunwind to FreeBSD/powerpc64.Piotr Kubaj2020-09-216-117/+248
|
* Fix unwinding of pre-linked librariesMikhail Durnev2020-09-214-8/+38
| | | | | | | | | | | Prelinker updates section .eh_frame but does not change section .debug_frame. Libunwind can work with prelinked .eh_frame, but if fails to find call frame info in unmodified .debug_frame because it does not add the load offset. ELF load offset from PT_LOAD p_vaddr has to be used to correctly interpret addresses in the .debug_frame section. Signed-off-by: Mikhail Durnev <mikhail_durnev@mentor.com>
* Use __linux__ define everywhere instead of __linuxMilian Wolff2020-09-0813-21/+21
| | | | | | | | | | | | | | | The libunwind code base uses __linux__ already in many locations, but some others relied on __linux instead. Apparently on some older toolchain configurations, such as Ubuntu Trusty that's still used for AppImage generation sometimes, __linux is never defined - only __linux__ is. Fixes compiler warning on such platforms: ``` coredump/_UPT_get_dyn_info_list_addr.c: In function 'get_list_addr': coredump/_UPT_get_dyn_info_list_addr.c:86:3: warning: #warning Implement get_list_addr(), please. [-Wcpp] # warning Implement get_list_addr(), please. ^ ```
* ia64: Fix dyn_info_list.S with newer binutilsJessica Clarke2020-08-041-1/+1
| | | | | | | | Using binutils 2.34 or 2.35, libunwind fails to link on ia64 with: /usr/bin/ld: .IA_64.unwind has both ordered [`.IA_64.unwind' in unwind/.libs/GetIPInfo.o] and unordered [`.IA_64.unwind' in ia64/.libs/dyn_info_list.o] sections Fix this by using the right flags and type for the section, taken from binutils's gas/testsuite/gas/ia64/unwind.s.
* Link Ltest-cxx-exceptions with libunwindAdrian Bunk2020-07-151-0/+1
| | | | | | | Bernhard Übelacker noticed that Ltest-cxx-exceptions was actually testing the gcc implementation due to not linking with libunwind. Fixes: #129
* linux: Use /proc/[pid]/rootMikhail Durnev2020-07-151-2/+32
| | | | | | | | When we are backtracing processes running in a container, we cannot rely on ELF paths in /proc/[pid]/maps. Those paths are valid inside the container. We have to add /proc/[pid]/root to each path to access the ELF files from outside of the container. Signed-off-by: Mikhail Durnev <mikhail_durnev@mentor.com>
* Undo unnecessary debug-related changeStephen Webb2020-07-021-21/+18
|
* Use NT_FILE to get backing file names if availableStephen Webb2020-07-028-45/+196
| | | | | A Linux core dump passes backing file names for LOAD segments. The coredump library should take advantage of them if they are present.
* Refactor Linux-specific codedump codeStephen Webb2020-07-025-78/+293
| | | | | | Moved Linux-specific core file code into its own separate source file. Moved some generic ELF-note handling code into its own separate source file, too, to replace some macros in anticipation of re-using them for non-Linux core files.
* Make SHF_COMPRESSED use contingent on its existenceStephen Webb2020-07-021-0/+4
| | | | | | | | | | A recent change broke builds for many non-recent and non-Linux targets by adding code dependent on a non-portable preprocessor macro SHF_COMPRESSED found in some more recent GLIBC elf.h headers. This fixes the firt part of #190. Signed-off-by: Stephen Webb <swebb@blackberry.com>
* Fix a typo caught by CIStephen Webb2020-06-181-1/+1
| | | | Fixed a missing conjunction on the mips target.
* Fix typo in a previous changeStephen Webb2020-06-181-1/+1
| | | | | Somehow a typo crept in in a previous change. You try and try but still can't win.
* Fix problems found in aarch64 buildStephen Webb2020-06-181-1/+5
| | | | Make an additional Linuxism dependent on building for Linux.
* Autodetect <sys/syscall.h>Stephen Webb2020-06-184-8/+11
| | | | | Use the system header <sys/syscall.h> only on those OS targets for which it is available.
* Handle <elf.h> that uses enums instead of definesStephen Webb2020-06-189-13/+35
| | | | | | | | | | Handle the case in which the system-supplied elf.h file uses C enums instead of macros preprocessor defines for some constants. Made the inclusion of the discovered elf.h file more consistent for portability to non-Linux targets. Put some libunwind-specific ELF macros into the UNW_ namespace.
* Rationalize endianness macrosStephen Webb2020-06-1821-101/+87
| | | | | | | | | Extend the endianness macros to support systems that define them in <sys/param.h> and replace the use of names that violate ISO/IEC 9899. Rename endianness macros used in code to use properly-namespaced macros for portability and cross-hosting compatibility. Add inline functions to perform common operations with the macros.
* Consistently use POSIX <ucontext.h>Stephen Webb2020-06-186-6/+6
| | | | | Use the (former) POSIX-standard <ucontext.h> everywhere instead of sometimes using the Linux-specxific <sys/ucontext.h>.
* Fix typosSteve MacLean2020-06-183-3/+3
|
* Fix placement of alignas() specifierSteve MacLean2020-06-183-3/+3
| | | | | | | | | | | | The alignment specifier must appear before the typename i.e. clang9 refused to compile uint8_t alignas(MAX_ALIGN) __reserved[128]; but accepted alignas(MAX_ALIGN) uint8_t __reserved[128];
* Added --disable-unwind-header optionPedro Navarro2020-06-182-1/+12
| | | | | | When set, unwind.h won't be copied to the installation directory. Some platforms provide their own unwind.h that might conflict with the one supplied with libunwind
* Fix Windows abstraction SIZEOF_*Steve MacLean2020-06-043-6/+29
| | | | Use actual sizes from current kernels for the Windows abstraction layer
* Fix Travis job matrixSteve MacLean2020-06-041-4/+6
|
* Add Windows build supportSteve MacLean2020-06-045-0/+420
| | | | | Add CMake file for Windows build Enable Windows builds in Travis CI
* Add simple Windows platform abstraction layerSteve MacLean2020-06-0413-0/+369
| | | | | | | | Add required unix style headers as a platform abstraction layer This will allow compilation on Windows. Simple single threaded implementation of the platform abstraction layer for compilation on windows
* Add freebsd-elf* headers for WIndows compileSteve MacLean2020-06-043-0/+1362
|
* Remove aux directorySteve MacLean2020-06-044-7871/+1
| | | | | | | | | | | | | aux is a reserved filename on Windows. Presence of the directory in the repo will prevent a git checkout on Windows. The aux directory contained GPL code related to autotools, it appears to not be directly referenced or necessary. Assume it was a commit error. Remove it. Add aux to the list of autotool releated ignores in .gitignore
* aarch64: Fix __sigset build issue on muslCDave Watson2020-05-271-1/+1
| | | | | | Described in issue #176. Taken from Alpine Linux project. Change __sigset_t to sigset_t
* Define UNW_EMPTY_STRUCT for non-GNUCSteve MacLean2020-05-212-2/+10
|
* Fix warningsSteve MacLean2020-05-211-2/+2
|
* Use C11 atomicsSteve MacLean2020-05-2161-312/+128
|
* Use C11 alignas(x)Steve MacLean2020-05-214-17/+7
|
* Use C11 _Thread_localSteve MacLean2020-05-2110-49/+17
|
* Only use one set of macros in Gexpr.cSteve MacLean2020-05-211-35/+1
| | | | Fix Gexpr macro warnings
* Remove #include compiler.h in public headerSteve MacLean2020-05-211-2/+12
|
* Fix MSVC UNW_REMOTE_ONLY compilation errorsSteve MacLean2020-05-2015-24/+124
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix errors observed while compiling libunwind for UNW_REMOTE_ONLY, library src, amd64 & aarch64 with an amd64 MSVC compile, and arm with an x86 MSVC compiler Use standard compliant variadic macros GNUC offer non-compliant variadic macros Switch to standrds compiant variadic macros Replace non-standard statement expression GNUC statement expression is non-ISO C compliant Where a simple inline function will be equivalent, use that instead Gexpr.c support compilers w/o statement expressions For this case a simple inline function would change semantics Make a close approximation with a separate set of macros for non GNUC Fix UNW_REMOTE_ONLY placement Exclude function call not required for remote only builds Fix __attribute__((packed)) On MSVC use pragma pack() Add THREAD_LOCAL macro Avoid bare use of __thread Add MSVC fetch_and_add() support Use macro ALIGNED(x) in place of __attribute__((aligned(x)) Rename local ltoa to avoid name collision The Windows x86 SDK headers define ltoa. Rename this local function to avoid a name collision. ISO C doesn't allow empty structures Add padding to allow empty arm/aarch64 structures to compile on MSVC
* Fix a memory leak issueAdeel2020-05-181-9/+8
| | | | | | | | | | | | ```sh $ cppcheck --enable=all . 2>&1 | grep -i -B5 -A5 leak ... Checking src/tilegx/Gcreate_addr_space.c ... src/tilegx/Gcreate_addr_space.c:52:5: error: Memory leak: as [memleak] return NULL; ^ ... ```
* src/arm/Gex_tables.c: resolve possible null pointer dereferenceIlya Shipitsin2020-05-181-3/+3
| | | | | | | | found by cppcheck [src/arm/Gex_tables.c:159] -> [src/arm/Gex_tables.c:155]: (warning) Either the condition 'buf!=NULL' is redundant or there is pointer arithmetic with NULL pointer. Upstreamed from dotnet/coreclr#26869
* Fix error in aarch64 unw_sigcontextSteve MacLean2020-05-181-1/+1
| | | | | | | | Introduced by #71 __reseverved needs to be big enough to store a unw_fpsimd_context_t Which includes 32 128-bit registers, stored as 64 64-bit half registers. Fix off by 2x issue
* Fix off-by-one error in x86_64 stack framesJon Haslam2020-04-291-1/+1
|
* Reflect current status of Solaris x64 in the READMEAdeel2020-04-281-16/+2
|
* Update README to reflect Solaris x86-64 status w/ this PRJason King2020-04-281-1/+1
|
* Stop stepping on Solaris and illumos after reaching an IP of 0Jason King2020-04-281-0/+7
|
* Fix Gtest-trace output for illumos/SolarisJason King2020-04-281-1/+1
|
* Fix signal stacks on illumos/SolarisJason King2020-04-282-73/+51
|
* Add additional symbols to ignore on illumos and SolarisJason King2020-04-281-0/+5
|
* Treat PT_SUNW_UNWIND sections on illumos and Solaris as equivalent to ↵Jason King2020-04-282-0/+7
| | | | PT_GNU_EH_FRAME
* Update Gtest-bt for illumos and SolarisJason King2020-04-281-3/+3
|
* coredump: Unmap elf image before mapping another (#169)Hans-Christian Egtvedt2020-04-021-0/+4
| | | | | | | | Make sure to unmap the currently mapped elf image before mapping another image in CD_get_proc_name. If not, the reference to that image is lost and the memory is leaked. Signed-off-by: Sturle Mastberg <smastber@cisco.com> Signed-off-by: Hans-Christian Noren Egtvedt <hegtvedt@cisco.com>
* Bump version to 1.5-rc1v1.5-rc1Dave Watson2020-03-311-1/+1
|
* Fix compilation with -fno-common.Yichao Yu2020-03-3115-80/+89
| | | | | | | Making all other archs consistent with IA64 which should not have this problem. Also move the FIXME to the correct place. Also add some minimum comments about this...
View Lorry Controller Status