https://crrev.com/c/3327644 introduced the ability for dump_syms to
operate on in memory data, which has the consequence of not going
through the same input validation as the dump_syms cli tool. In certain
cases, it is possible that architecture info can't be reliably
determined, e.g. new architectures that breakpad is unware of. In that
case, dump_syms should avoid crashing when calling ReadSymbolData and
return false instead.
Change-Id: Ie9acdf811300084f1d5916f4778754f8abca10e0
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3572251
Reviewed-by: Ivan Penkov <ivanpe@chromium.org>
1. Visit DW_TAG_class_type when it's inside DW_TAG_subprogram.
2. Only warn when we can't get the name for the DIE and it has DW_AT_specification that is not in the specification map.
Bug: 1078932
Change-Id: Id3126aec305658f8f65c01675a8e9e3ea03f3651
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3579855
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
- don't do iter decrement when the map empty.
- add dummy file with id equals to 0 to represent unknown file.
Change-Id: I3fe55a459c9fa835bbe0c4272e4ac12b1150c034
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3425732
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
In trying to create a backend service that can process both ELF and
Mach-O binaries, I found that symbol collisions occur when trying to
link different implementations of FileID. This change puts the
different implementations into separate namespaces to avoid the
collision.
Change-Id: I15aabb222803f2ffbda15ed13e66793bae32ddce
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3421417
Reviewed-by: Mark Mentovai <mark@chromium.org>
dump_syms assumes it is operating on a file and can access a compliant
file system. This change allows dump_syms to operate on the contents of
a file that has already been read into memory. This is useful in a
server context where the file does not exist on the local file system.
Change-Id: Id63f115c2df287083d548dadd5ac487f97bde057
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3327644
Reviewed-by: Mark Mentovai <mark@chromium.org>
- Make handle_inline default to false in PDBSourceLineWriter constructor.
- Add --i flag for symupload to generate inline information.
Change-Id: I3149173ee635a503b9508a12ef572f8b6e5c5dfe
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3364804
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
This adds INLINE and INLINE_ORIGIN records on Window dump_syms. It also
adds more LINE records that represents the inner most callsite line info
inside a function.
Bug: chromium:1190878
Change-Id: I15c2044709f8ca831b03a453910d036f749452c6
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3133606
Reviewed-by: Lei Zhang <thestig@chromium.org>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Reviewed-by: Ivan Penkov <ivanpe@chromium.org>
- Ignore DW_TAG_inlined_subroutine with empty range.
- Don't stop parsing after parsing malformed INLINE/INLINE_ORIGIN
records, because reports can still be generated without them but won't
have inlined frames.
Bug: 1190878
Change-Id: I445105ad06b9146268f7d064e85b0d162c3f2a39
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3321166
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
The rust-demangle-capi crate hasn't been updated since 2016 and
out-of-date. Instead, Breakpad needs to use C API offered by the
rustc-demangle to demangle Rust symbols.
*** TESTING ***
1) Set up rustc-demangle
> git clone https://github.com/rust-lang/rustc-demangle.git
> cd rustc-demangle
> cargo build -p rustc-demangle-capi --release
2) Breakpad
> ./configure --with-rustc-demangle=<path to rustc-demangle>
> make check src/common/dward_cu_to_module
Change-Id: Ib68b62ef329f1397bc379a1d04c632781e4b2069
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3273324
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
This reverts commit 54d878abcb.
54d878abcb changed the dump_syms format incompatibly. This must be
redone in a multi-step process: the processor must be made to understand
the old and new formats simultaneously and the processor service must be
rebuilt and run with that update before dump_syms output can change to
use the new format.
Bug: chromium:1263390
Change-Id: I5b6f8aff8ea2916b2c07ac6a74b569fa27db51b9
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3244775
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Processor shows incorrect source file name if a frame have an inlined
frame and their source files are different.
Consider this example:
FILE 0 /tmp/a.h
FILE 1 /tmp/a.cpp
INLINE_ORIGIN 0 0 foo()
FUNC 1110 a 0 main
INLINE 0 22 0 1110 7
1110 7 3 0
1117 3 23 1
When querying the address 0x1110, we know this line 0x1110 corresponds
to /tmp/a.h line 3 and it's inside a inlined function foo() which is
defined at /tmp/a.h and called at line 22. But we don't know at which
file it's being called at line 22. So, we will get stacks like this:
void foo() /tmp/a.h:3
int main() /tmp/a.h:22
The correct stacks should be this:
void foo() /tmp/a.h:3
int main() /tmp/a.cpp:22
In this change:
1. Remove file_id field for INLINE_ORIGIN record.
2. Add call_site_file_id for INLINE record to represents the file where
this call being inlined.
After adding call_site_file_id to it (as third field), it looks like
this:
FILE 0 /tmp/a.h
FILE 1 /tmp/a.cpp
INLINE_ORIGIN 0 foo()
FUNC 1110 a 0 main
INLINE 0 22 1 0 1110 7
1110 7 3 0
1117 3 23 1
Bug: 1190878
Change-Id: Ibbb697d2f7e1b6ac3208cac6fae4353c8743198d
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3232838
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
After ff5892c5da added the new StringView,
building fails with GCC 6 due to it apparently failing to properly find
the type for nullptr_t resulting in the following error:
In file included from ../src/common/module.h:49:0,
from ../src/common/dwarf_cfi_to_module.h:49,
from ../src/common/linux/dump_symbols.cc:59:
../src/common/string_view.h:55:27: error: field 'nullptr_t' has incomplete type 'google_breakpad::StringView'
StringView(nullptr_t) = delete;
^~~~~~
../src/common/string_view.h:42:7: note: definition of 'class google_breakpad::StringView' is not complete until the closing brace
class StringView {
^~~~~~~~~~
This can be fixed by adding the std:: namespace to nullptr_t.
Change-Id: I00a090d307ebe21d1143eac4a605ff319ce27048
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3201997
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
The context arguments are of type DWORD_PTR which is actually a
integer type, not a pointer, so using NULL here causes a type
missmatch warning:
error: passing NULL to non-pointer argument 8 [...]
Change-Id: Ia52f51fd0cd33af3b139f0427dec6c59c2455d0a
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3168663
Reviewed-by: Primiano Tucci <primiano@chromium.org>
- Added StringView which is used as a reference to a string, but
doesn't own the string.
- Removed the old string pool in DwarfCUToModule::FilePrivate, since
it's doing string copy.
- Added a string pool in Module to store functions/inline origins'
names (mangled and demangled).
- The peak memory usage drops from 20.6 GB to 12.5 GB when disabling
inline records and drops from 36 GB to 20.3 GB when enabling inline records.
Bug: chromium:1246974, chromium:1250351
Change-Id: Ie7e9740ea10c1930a0fc58c6becaae2d718b83b8
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3189410
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
It moves InlineOriginMap to module.h. Let Module keeps the global InlineOriginMap to easily get all referenced InlineOrigin when emitting. And release allocated memory inside its destructor.
Verified that the symbol file with inline records for chrome is the same before and after this change.
Change-Id: I7541aa05d3d2df0b9d52d670cab58241baecf20d
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3171638
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
This change makes sure dump_syms process DW_TAG_inlined_subroutine only when -d flag is given, which save memory and time when -d is not given. Before this, it always processes DW_TAG_inlined_subroutine and -d determines whether or not to emit INLINE records.
Bug: chromium:1250351, chromium:1246974
Change-Id: I54725ba1e513cafe17268ca389ff8acc9c11b25e
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3166674
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Temporarily works around an issue on Mac where the system version of
NXGetLocalArchInfo is returning x86 information on x86_64 devices,
which results in dump_syms failing on said devices. Instead, the
Breakpad implementation of NXGetLocalArchInfo, which is meant for
dump_syms_mac on Linux, will be used until the system version is fixed.
Bug: 1242776
Change-Id: Id398338e580eb9c67c61f9f01670d2e7dbe86bea
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3143524
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Dwarf generated by Clang -g1 will not have DW_AT_inline attribute for some DW_TAG_subprograms even if they are inlined. This warning recently increased a lot (~ 3 million) due to DW_TAG_inlined_subroutine also complains about unknown abstract origin. It caused infra failure in building bots.
Bug: 1241579
Change-Id: I9b5135925b71aa915760c140bcf73fc603bb77d3
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3111782
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Building fails for some people because configure requires c++11 but make_unique is a c++14 feature.
Change-Id: I23ce689fc92e9e90a95e7643ff29602f6b32ccbb
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3107784
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
The size of symbol file for chrome binary increased from 577 MB to
1205 MB. There are 7,453,748 INLINE records and 1,268,493 INLINE_ORIGIN
records.
Bug: 1190878
Change-Id: I802ec1b4574c14f74ff80d0f69daf3c81085778a
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2915828
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Section 7.26 of the DWARF 5 spec describes a header of
either 8 or 16 bytes before the offsets begin.
Bug: b/187205051
Change-Id: I1ba01008dcd7a533f59d3865762ca09b9d43032b
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/3016609
Reviewed-by: Sterling Augustine <saugustine@google.com>
iOS closes an app’s network connections when the app is backgrounded. This can cause an in-progress upload request to fail. We can mitigate this by requesting additional background execution time using the `UIApplication` background task APIs.
BUG=b:130302235
Change-Id: Ifd8e14ca82c736ad7dd60dcdd0d4bbcabb76f5ad
Signed-off-by: Darren Mo <darrenmo@google.com>
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2251020
Reviewed-by: Mark Mentovai <mark@chromium.org>
PUBLIC records.
For chrome, the symbol file size dropped from 661 MB to 577 MB.
The number of PUBLIC records dropped from 559416 to 91930.
Change-Id: I7f703d619f2acc7c83f002e6f588f6a6569e5c87
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2916731
Reviewed-by: Lei Zhang <thestig@chromium.org>
macho_reader assumes that every section in a segment exists, or none of the
sections exist in the file (for dSYM files).
https://reviews.llvm.org/D94460?id=315965 added __eh_frame section to the
__TEXT segments in dSYM files. All the other sections are removed, but still
have non-zero size in the header. macho_reader only looks at `fileoff` and
`filesize` fields to determine the size of the segment, but it looks at `addr`
and `size` to determine the size of the section, therefore it determines that
the sections would not fit in the segment and refused to parse the file.
In this case the removed sections all have offset == 0. Ignore such cases.
Change-Id: Ife771f7b302c1bc81c673b1103492c41321b5e3e
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2911204
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Before this change, forward_ref_die_to_func is always empty.
Because forward reference DIEs could cross compilation units,
forward_ref_die_to_func is moved into FilePrivate.
By using "grep "name omitted" sym | wc -l", the number of functions
with no name is reduced from 26951 to 203 if forward reference only is
allowed within CU. It's reduced from 203 to 6 if crossing compilation
units is allowed.
Change-Id: Ie2a457abfc0c4d8e68fe0fa595a27ea4abf33a76
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2904093
Reviewed-by: Mike Frysinger <vapier@chromium.org>
If there is an address index reference to .debug_addr section before
DW_AT_addr_base, addr_base_ will be 0. It will retrieve wrong address.
An example could be DW_AT_loc_pc occurs before DW_AT_addr_base.
Change-Id: Id2b337f5235470cc9beaf05a62efebbde797dacf
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2900806
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
std::allocator<T>::pointer and related are removed in C++20, so moving to
std::allocator_traits which is available since C++11.
Change-Id: Ie67b5c24b27e59edf5595a3575f2794748c3817f
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/2894004
Reviewed-by: Mike Frysinger <vapier@chromium.org>
