Lit has the ability to set a timeout for individual tests. This patch
enables that functionality with a default of 10 minutes.
Currently we rely on the bots to kill the whole test suite. However this
doesn't tell us which test caused the timeout. Furthermore, when running
the test suite during development, I have to manually kill the tests
that time out to get the lit output at then end. This fixes both
inconveniences.
llvm-svn: 357555
This re-commits r354263, which was because it uncovered with handling of
modules with empty (zero) UUIDs. This would cause us to treat two
modules as intentical even though they were not. This caused an assert
in PlaceholderObjectFile::SetLoadAddress to fire, because we were trying
to load the module twice even though it was designed to be only loaded
at a specific address. (The same problem also existed with the previous
implementation, but it had no asserts to warn us about this.) These
issues have now been fixed in r356896.
windows bot. The issue there was that ObjectFilePECOFF vended its base
address through the incorrect interface. SymbolFilePDB depended on that,
which lead to assertion failures when SymbolFilePDB was attempting to
use the placeholder object files as a base. This has been fixed in
r354258
The original commit message was:
The reason this wasn't working was that ProcessMinidump was creating odd
object-file-less modules, and SymbolFileBreakpad required the module to
have an associated object file because it needed to get its base
address.
This fixes that by introducing a PlaceholderObjectFile to serve as a
dummy object file. The general idea for this is taken from D55142, but
I've reworked it a bit to avoid the need for the PlaceholderModule
class. Now that we have an object file, our modules are sufficiently
similar to regular modules that we can use the regular Module class
almost out of the box -- the only thing I needed to tweak was the
Module::CreateModuleFromObjectFile functon to set the module's FileSpec
in addition to it's architecture. This wasn't needed for ObjectFileJIT
(the other user of CreateModuleFromObjectFile), but it shouldn't hurt it
either, and the change seems like a straightforward extension of this
function.
Reviewers: clayborg, lemo, amccarth
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D57751
llvm-svn: 357060
At the moment when --repl is passed to lldb it silently ignores any
commands passed via the options below:
--one-line-before-file <command>
Tells the debugger to execute this one-line lldb command before any file provided on the command line has been loaded.
--one-line <command>
Tells the debugger to execute this one-line lldb command after any file provided on the command line has been loaded.
--source-before-file <file>
Tells the debugger to read in and execute the lldb commands in the given file, before any file has been loaded.
--source <file>
Tells the debugger to read in and execute the lldb commands in the given file, after any file has been loaded.
-O <value> Alias for --one-line-before-file
-o <value> Alias for --one-line
-S <value> Alias for --source-before-file
-s <value> Alias for --source
The -O and -S options are quite useful when writing tests for the REPL
though, e.g. to change settings prior to entering REPL mode. This
patch updates the driver to still respect the commands supplied via -O
and -S when passing --repl instead of silently ignoring them. As -s
and -o don't really make sense in REPL mode, commands supplied via
those options are still ignored, but the driver now emits a warning to
make that clear to the user.
Patch by Nathan Hawes!
Differential Revision: https://reviews.llvm.org/D59681
llvm-svn: 356911
This patch begins the process of migrating the "minidump" plugin to the
minidump parser in llvm. The llvm parser is not fully finished yet, but
even now, a lot of things can be switched over. The gradual migration
process will allow us to easier detect if things break than doing a big
one-step migration. Doing it early will allow us to make sure that the
llvm parser fits the use case that we need in lldb.
In this patch I start with the various minidump constants, which have
their llvm equivalent. It doesn't contain any functional changes. The
diff just reflects the different naming of things in llvm.
llvm-svn: 356898
On Linux, a QEnvironment packet is sent for every environment variable.
This breaks replay when the number of environment variables is different
then during capture. The solution is to always reply with OK.
llvm-svn: 356643
Summary:
This is a preparatory step to enable adding of unwind plans by symbol
file plugins.
Although at the surface it seems that currently symbol files have
nothing to do with unwinding, this isn't entirely correct even now. The
mere act of adding a symbol file can have the effect of making more
sections (typically .debug_frame) available to the unwinding machinery,
so that it can have more unwind strategies to choose from.
Up until now, we've had a bug, which went largely unnoticed, where
unwind info in the manually added symbols files (target symbols add) was
being ignored during unwinding. Reinitializing the UnwindTable fixes
that bug too.
Reviewers: clayborg, jasonmolenda, alexshap
Subscribers: jdoerfert, lldb-commits
Differential Revision: https://reviews.llvm.org/D58347
llvm-svn: 356361
Yesterday I noticed a reproducer test failing after making a local
change. Removing the reproducer directory solved the issue. Add a test
case that detects this.
llvm-svn: 355941
Tablegen doesn't support options that are both flags and take values as
an argument. I noticed this when doing the tablegen rewrite, but forgot
that that affected the reproducer --capture flag.
This patch makes --capture a flag and adds --capture-path to specify a
path for the reproducer. In reality I expect this to be mostly used for
testing, but it could be useful nonetheless.
Differential revision: https://reviews.llvm.org/D59238
llvm-svn: 355936
Summary:
Within .lldbinit, regex commands can be structured as a list of substitutions over
multiple lines. It's possible that this is uninentional, but it works and has
benefits.
For example:
command regex <command-name>
s/pat1/repl1/
s/pat2/repl2/
...
I use this form of `command regex` in my `~/.lldbinit`, because it makes it
clearer to write and read compared to a single line definition, because
multiline substitutions don't need to be quoted, and are broken up one per line.
However, multiline definitions result in usage instructions being printed for
each use. The result is that every time I run `lldb`, I get a dozen or more
lines of noise. With this change, the instructions are only printed when
`command regex` is invoked interactively, or from a terminal, neither of which
are true when lldb is sourcing `~/.lldbinit`.
Reviewers: clayborg, jingham
Reviewed By: clayborg
Subscribers: jdoerfert, kastiglione, xiaobai, keith, lldb-commits
Differential Revision: https://reviews.llvm.org/D48752
llvm-svn: 355793
Adjust the XFAIL-ing tests to match consistent results from buildbot.
I'm going to work on differences between them and my local results
following this.
llvm-svn: 355774
Summary: DW_OP_GNU_addr_index has been renamed as DW_OP_addrx in the standard. clang produces DW_OP_addrx tags and with this change lldb starts to process them.
Reviewers: aprantl, jingham, davide, clayborg, serge-sans-paille
Reviewed By: aprantl
Subscribers: jdoerfert, dblaikie, labath, shafik, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D59004
llvm-svn: 355629
After D55626 I see a failure in my Fedora buildbot.
There is uninitialized variable as the Foo constructor has not been run and foo
is an autovariable.
(lldb) breakpoint set -f foo.cpp -l 11
Breakpoint 1: where = TestDataFormatter.test.tmp.out`main + 30 at foo.cpp:11:7, address = 0x000000000040112e
(lldb) run
Process 801065 stopped
* thread #1, name = 'TestDataFormatt', stop reason = breakpoint 1.1
frame #0: 0x000000000040112e TestDataFormatter.test.tmp.out`main(argc=1, argv=0x00007fffffffcc48) at foo.cpp:11:7
8 };
9
10 int main(int argc, char **argv) {
-> 11 Foo foo(1, 2.22);
12 return 0;
13 }
Process 801065 launched: '.../tools/lldb/lit/Reproducer/Functionalities/Output/TestDataFormatter.test.tmp.out' (x86_64)
(lldb) frame var
(int) argc = 1
(char **) argv = 0x00007fffffffcc48
(Foo) foo = (m_i = 4198432, m_d = 0)
While the testcase expects m_i will be 0.
Differential Revision: https://reviews.llvm.org/D59088
llvm-svn: 355611
lldb/cmake/modules/LLDBConfig.cmake does not build lldb-server on Windows:
if (CMAKE_SYSTEM_NAME MATCHES "Android|Darwin|FreeBSD|Linux|NetBSD")
set(LLDB_CAN_USE_LLDB_SERVER 1)
Also do not append 'platform' parameter twice - although that was quietly
ignored.
llvm-svn: 355579
This patch adds test that check that functionality in lldb continues to
work when replaying a reproducer.
- Entries in image list are identical.
- That stepping behaves the same.
- That the data formatters behave the same.
Differential revision: https://reviews.llvm.org/D55626
llvm-svn: 355570
In mail
[lldb-dev] Remote debugging a docker process
https://lists.llvm.org/pipermail/lldb-dev/2019-March/014795.html
user was confused by --min-gdbserver-port and --max-gdbserver-port options
being ignored. I think there is even a bug that --max-gdbserver-port is upper
exclusive limit (and not upper inclusive limit appropriate for max).
At least this patch should catch such mistake by an error message. The question
is whether --max-gdbserver-port should not be changed to really be max and not
max+1 but that would break backward compatibility.
Now the mail example does produce:
error: --min-gdbserver-port (5001) is not lower than --max-gdbserver-port (5001)
Differential Revision: https://reviews.llvm.org/D58962
llvm-svn: 355554
Pass appropriate -L and -Wl,-rpath flags pointing out to the LLVM
library directory on NetBSD. This is necessary since clang on NetBSD
requires libc++ but it is not installed as part of the system
by default. For the purpose of running buildbot, we want LLDB to use
just-built libc++.
Differential Revision: https://reviews.llvm.org/D58630
llvm-svn: 355502
Now that the LLDB instrumentation macros are in place, we should use
that to test reproducer replay.
Differential revision: https://reviews.llvm.org/D58565
llvm-svn: 355470
Summary:
clang-cl tries to match cl's interface, and treats /U as "Removes a
predefined macro" as cl does. When you feed clang-cl a file that begins with
'/U' (e.g. /Users/xiaobai/foo.c), clang-cl will emit a warning and in some cases
an error, like so:
clang-9: warning: '/Users/xiaobai/foo.c' treated as the '/U' option [-Wslash-u-filename]
clang-9: note: Use '--' to treat subsequent arguments as filenames
clang-9: error: no input files
If you're using clang-cl, make sure '--' is passed before the source file.
Differential Revision: https://reviews.llvm.org/D58860
llvm-svn: 355341
Add a convenience 'expectedFailureNetBSD' decorator and mark all tests
currently failing on NetBSD with it. Also skip a few tests that hang
the test suite. This should establish a baseline for the test suite
and get us closer to enabling tests on buildbot. This will help us
catch regressions while we still have a lot of work to do to get tests
working.
It seems that there are also some flaky tests. I am going to address
them later on.
Differential Revision: https://reviews.llvm.org/D58527
llvm-svn: 355320
This stanza was removed in r355213, but it seems that patch did not
fully fix the problem, as the test still fails sporadically
(particularly under heavy load) on linux.
llvm-svn: 355276
When the debugger is run in sync mode, you need to
be able to tell whether a hijacked resume is for some
special purpose (like waiting for the SIGSTOP on attach)
or just to perform a synchronous resume. Target::Launch was doing
that wrong, and that caused stop-hooks on process launch
in source files to behave incorrectly.
<rdar://problem/48115661>
Differential Revision: https://reviews.llvm.org/D58727
llvm-svn: 355213
Set LD_LIBRARY_PATH or local platform's equivalent of it when running
the 'Suite' tests. This is necessary when running tests inside build
tree with BUILD_SHARED_LIBS enabled, in order to make the LLDB modules
load freshly built LLVM libraries.
The code is copied from clang (test/Unit/lit.cfg). SHLIBDIR
substitution is added to site-config (already present in top-level LLDB
site-config) to future-proof this into supporting stand-alone builds
with shared LLDB libraries.
Differential Revision: https://reviews.llvm.org/D58610
llvm-svn: 354920
They aren't designed to nest recursively, so this will prevent that.
Also add a --auto-continue flag, putting "continue" in the stop hook makes
the stop hooks fight one another in multi-threaded programs.
Also allow more than one -o options so you can make more complex stop hooks w/o
having to go into the editor.
<rdar://problem/48115661>
Differential Revision: https://reviews.llvm.org/D58394
llvm-svn: 354706
Split the recognition into NetBSD executables & shared libraries
and core(5) files.
Introduce new owner type: "NetBSD-CORE", as core(5) files are not tagged
in the same way as regular NetBSD executables.
Stop using incorrectly ABI_TAG and ABI_SIZE. Introduce IDENT_TAG,
IDENT_DECSZ, IDENT_NAMESZ and PROCINFO.
The new values detect correctly the NetBSD images.
The patch has been originally written by Kamil Rytarowski. I've added
tests and applied minor code changes per review. The work has been
sponsored by the NetBSD Foundation.
Differential Revision: https://reviews.llvm.org/D42870
llvm-svn: 354466
Currently we'd always print the LLDB_REGISTER macro, even if the
LLDB_RECORD macro was already present. This patches changes that to make
it easier to incrementally update the macros.
Note that it's still possible for the RECORD and REGISTER macros to get
out of sync.
llvm-svn: 354400
Facebook creates minidump files that contain specific information about why things crash. Adding ways to dump these allows tools to be made that can auto download symbols based on the information that is contained in the minidump files.
Differential Revision: https://reviews.llvm.org/D58398
llvm-svn: 354385
This reverts r354263, because it uncovered a problem in handling of the
minidumps with conflicting UUIDs. If a minidump contains two files with
the same UUID, we will not create to placeholder modules for them, but
instead reuse the first one for the second instance. This creates a
problem because these modules have their load address hardcoded in them
(and I've added an assert to verify that).
Technically this is not a problem with this patch, as the same issue
existed in the previous implementation, but it did not have the assert
which would diagnose that. Nonetheless, I am reverting this until I
figure out what's the best course of action in this situation.
llvm-svn: 354324
Importing cxx modules doesn't seem to work on Windows:
error: a.out :: Class 'tagARRAYDESC' has a member 'tdescElem' of type
'tagTYPEDESC' which does not have a complete definition.
error: a.out :: Class 'tagPARAMDESCEX' has a member 'varDefaultValue' of type
'tagVARIANT' which does not have a complete definition.
llvm-svn: 354300
The test had an implicit constructor for the Foo struct. Also, as the
instrumentation doesn't have to be reproducer specific, I moved the
tests into the lit/tools directory.
llvm-svn: 354294
In r353906 we hooked up clang and lldb's reproducer infrastructure to
capture files used by clang. This patch adds the necessary logic to have
clang reuse the files from lldb's reproducer during replay.
Differential revision: https://reviews.llvm.org/D58309
llvm-svn: 354283
This re-commits r353677, which was reverted due to test failures on the
windows bot. The issue there was that ObjectFilePECOFF vended its base
address through the incorrect interface. SymbolFilePDB depended on that,
which lead to assertion failures when SymbolFilePDB was attempting to
use the placeholder object files as a base. This has been fixed in
r354258
It also fixes one small problem in the original patch. The issue was that the
Module class would attempt to overwrite the object file we created in
CreateModuleFromObjectFile if the file corresponding to the placeholder object
file happened to exist (but we have already disqualified it due to UUID
mismatch. The fix is simple -- we set the m_did_load_objfile flag to properly
record the fact that we have already created an object file for the module.
The original commit message was:
The reason this wasn't working was that ProcessMinidump was creating odd
object-file-less modules, and SymbolFileBreakpad required the module to
have an associated object file because it needed to get its base
address.
This fixes that by introducing a PlaceholderObjectFile to serve as a
dummy object file. The general idea for this is taken from D55142, but
I've reworked it a bit to avoid the need for the PlaceholderModule
class. Now that we have an object file, our modules are sufficiently
similar to regular modules that we can use the regular Module class
almost out of the box -- the only thing I needed to tweak was the
Module::CreateModuleFromObjectFile functon to set the module's FileSpec
in addition to it's architecture. This wasn't needed for ObjectFileJIT
(the other user of CreateModuleFromObjectFile), but it shouldn't hurt it
either, and the change seems like a straightforward extension of this
function.
Reviewers: clayborg, lemo, amccarth
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D57751
llvm-svn: 354263
COFF files are modelled in lldb as having one big container section
spanning the entire module image, with the actual sections being
subsections of that. In this model, the base address is simply the
address of the first byte of that section.
This also removes the hack where ObjectFilePECOFF was using the
m_file_offset field to communicate this information. Using file offset
for this purpose is completely wrong, as that is supposed to indicate
where is this ObjectFile located in the file on disk. This field is only
meaningful for fat binaries, and should normally be 0.
Both PDB plugins have been updated to use GetBaseAddress instead of
GetFileOffset.
llvm-svn: 354258
ExecControl/StopHook/stop-hook-threads.test is flaky on Linux (it's
consistently failing on my machine, but doesn't fail on a co-worker's).
I'm seeing the following assertion failure:
```
CommandObject.cpp:145: bool lldb_private::CommandObject::CheckRequirements(lldb_private::CommandReturnObject&): Assertion `m_exe_ctx.GetTargetPtr() == NULL' failed.
```
Interestingly, this doesn't happen when typing the same commands in
interactive mode. The cause seems to be that, in synchronous execution
mode continue waits until the process stops again, and that includes
running any stop-hooks for that later stop, so we end with a stack trace
like this (lots of frames omitted for clarity):
```
abort()
CommandObject::CheckRequirements() <-- this is again the same instance of CommandObjectProcessContinue, fails assertion because the previous continue command hasn't finished.
Target::RunStopHooks()
CommandObjectProcessContinue::DoExecute()
Target::RunStopHooks()
```
In general, it seems like using process control commands inside
stop-hooks does not have very well defined semantics. You don't even
need multiple threads to make that assertion fail, you can build
```
int main() {
printf("1\n"); // break1
printf("2\n"); // break2
}
```
and then on lldb
```
target stop-hook add -o continue
break set -f stop-hook-simple.cpp -p "break1"
break set -f stop-hook-simple.cpp -p "break2"
run
```
In this case it's even worse because the presence of multiple threads
makes it prone to race conditions. In some tests I ran with a simpler
version of this test case, I was hitting either the previous assertion
failure or the following issue:
1. Two threads reach a breakpoint
2. First stop-hook does a process continue
3. Threads end
4. Second stop-hook runs, complains about process not existing.
This change disables the test on Linux. It's already marked as XFAIL on
Windows, so maybe we should just delete it until it's clear what should
be the expected behavior in these cases. Or maybe try to come up with a
way to write a similar multithreaded test without calling continue from
a stop hook, I don't know.
Differential Revision: https://reviews.llvm.org/D58257
llvm-svn: 354149
Summary:
As suggested by Pavel, we shouldn't let our tests parse the local .lldbinit to prevent random test failures
due to changed settings.
Fixes Minidump/Windows/Sigsegv/sigsegv.test (and probably others too).
Reviewers: labath, serge-sans-paille
Reviewed By: labath
Subscribers: abidh, lldb-commits, zturner
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D58235
llvm-svn: 354038
Summary:
This is coming from the discussion in D55356 (the most interesting part
happened on the mailing list, so it isn't reflected on the review page).
In short the issue is that lldb assumes that all bytes of a module image
in memory will be backed by a "section". This isn't the case for PECOFF
files because the initial bytes of the module image will contain the
file header, which does not correspond to any normal section in the
file. In particular, this means it is not possible to implement
GetBaseAddress function for PECOFF files, because that's supposed point
to the first byte of that header.
If my (limited) understanding of how PECOFF files work is correct, then
the OS is expecded to load the entire module into one continuous chunk
of memory. The address of that chunk (+/- ASLR) is given by the "image
base" field in the COFF header, and it's size by "image size". All of
the COFF sections are then loaded into this range.
If that's true, then we can model this behavior in lldb by creating a
"container" section to represent the entire module image, and then place
other sections inside that. This would make be consistent with how MachO
and ELF files are modelled (except that those can have multiple
top-level containers as they can be loaded into multiple discontinuous
chunks of memory).
This change required a small number of fixups in the PDB plugins, which
assumed a certain order of sections within the object file (which
obivously changes now). I fix this by changing the lookup code to use
section IDs (which are unchanged) instead of indexes. This has the nice
benefit of removing spurious -1s in the plugins as the section IDs in
the pdbs match the 1-based section IDs in the COFF plugin.
Besides making the implementation of GetBaseAddress possible, this also
improves the lookup of addresses in the gaps between the object file
sections, which will now be correctly resolved as belonging to the
object file.
Reviewers: zturner, amccarth, stella.stamenova, clayborg, lemo
Reviewed By: clayborg, lemo
Subscribers: JDevlieghere, abidh, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D56537
llvm-svn: 353916
This patch hooks up clang and lldb's reproducers functionality. It
ensures that when capturing a reproducer, headers and modules imported
through the expression parser are collected.
Differential revision: https://reviews.llvm.org/D58076
llvm-svn: 353906
Summary:
This patch makes virtual bases to be added in the correct order to the bases
list. It is important because `VTableContext` (`MicrosoftVTableContext` in our
case) uses then the order of virtual bases in the list to restore the virtual
table indexes. These indexes are used then to resolve the layout of the virtual
bases.
We haven't enough information about offsets of virtual bases regarding to the
object (moreover, in a common case we can't rely on such information, see the
example here: https://reviews.llvm.org/D53506#1272306 ), but there should be
enough information to restore the layout of the virtual bases from the indexes
in runtime. After D53506 this information is used whenever possible, so there
should be no problems with virtual bases' fields reading.
Reviewers: zturner, rnk, stella.stamenova
Subscribers: abidh, teemperor, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D56904
llvm-svn: 353806
Summary:
`clang-cl` can't compile tests containing `char16_t` and `char32_t` types
without the MSVC compatibility option passed. This patch adds the option to the
`clang-cl` call in the `build.py` script by default.
Reviewers: zturner, labath, stella.stamenova, serge-sans-paille
Reviewed By: labath
Subscribers: lldb-commits, leonid.mashinskiy
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D57809
llvm-svn: 353709
Summary:
This adds support for auto-detection of path style to SymbolFileBreakpad
(similar to how r351328 did the same for DWARF). We guess each file
entry separately, as we have no idea which file came from which compile
units (and different compile units can have different path styles). The
breakpad generates should have already converted the paths to absolute
ones, so this guess should be reasonable accurate, but as always with
these kinds of things, it is hard to give guarantees about anything.
In an attempt to bring some unity to the path guessing logic, I move the
guessing logic from inside SymbolFileDWARF into the FileSpec class and
have both symbol files use it to implent their desired behavior.
Reviewers: clayborg, lemo, JDevlieghere
Subscribers: aprantl, markmentovai, lldb-commits
Differential Revision: https://reviews.llvm.org/D57895
llvm-svn: 353702
Fix the build helper to find lld-link via PATH lookup, rather than
making a fragile assumption that it will be present in the 'compiler
directory'. This fixes tests on Gentoo where clang and lld
are installed in different directories.
Differential Revision: https://reviews.llvm.org/D58001
llvm-svn: 353701
Skip running lldb-mi tests when Python support is disabled. This causes
lldb-mi to unconditionally fail, and therefore all the relevant tests
fail as well.
Differential Revision: https://reviews.llvm.org/D58000
llvm-svn: 353700
Summary:
The reason this wasn't working was that ProcessMinidump was creating odd
object-file-less modules, and SymbolFileBreakpad required the module to
have an associated object file because it needed to get its base
address.
This fixes that by introducing a PlaceholderObjectFile to serve as a
dummy object file. The general idea for this is taken from D55142, but
I've reworked it a bit to avoid the need for the PlaceholderModule
class. Now that we have an object file, our modules are sufficiently
similar to regular modules that we can use the regular Module class
almost out of the box -- the only thing I needed to tweak was the
Module::CreateModuleFromObjectFile functon to set the module's FileSpec
in addition to it's architecture. This wasn't needed for ObjectFileJIT
(the other user of CreateModuleFromObjectFile), but it shouldn't hurt it
either, and the change seems like a straightforward extension of this
function.
Reviewers: clayborg, lemo, amccarth
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D57751
llvm-svn: 353677
The tests are failing on windows because the paths in the symbol file
are parsed using the host path style. I'm working on a patch to have
SymbolFileBreakpad auto-detect the correct path style (similar to dwarf
r351328).
I originally wanted to make this a part of the initial line-table patch,
but then I simply forgot.
llvm-svn: 353410
Summary:
This patch teaches SymbolFileBreakpad to parse the line information in
breakpad files and present it to lldb.
The trickiest question here was what kind of "compile units" to present
to lldb, as there really isn't enough information in breakpad files to
correctly reconstruct those.
A couple of options were considered
- have the entire file be one compile unit
- have one compile unit for each FILE record
- have one compile unit for each FUNC record
The main drawback of the first approach is that all of the files would
be considered "headers" by lldb, and so they wouldn't be searched if
target.inline-breakpoint-strategy=never. The single compile unit would
also be huge, and there isn't a good way to name it.
The second approach will create mostly correct compile units for cpp
files, but it will still be wrong for headers. However, the biggest
drawback here seemed to be the fact that this can cause a compile unit
to change mid-function (for example when a function from another file is
inlined or another file is #included into a function). While I don't
know of any specific thing that would break in this case, it does sound
like a thing that we should avoid.
In the end, we chose the third option, as it didn't seem to have any
major disadvantages, though it was not ideal either. One disadvantage
here is that this generates a large number of compile units, and there
is still a question on how to name it. We chose to simply name it after
the first line record in that function. This should be correct 99.99% of
the time, though it can produce somewhat strange results if the very
first line record comes from an #included file.
Reviewers: clayborg, zturner, lemo, markmentovai
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D56595
llvm-svn: 353404
This patch introduces a new tool called 'lldb-instr'. It automates the
workflow of inserting LLDB_RECORD and LLDB_REGSITER macros for
instrumentation.
Because the tool won't be part of the build process, I didn't want to
over-complicate it. SB_RECORD macros are inserted in place, while
SB_REGISTER macros are printed to stdout, and have to be manually copied
into the Registry's constructor. Additionally, the utility makes no
attempt to properly format the inserted macros. Please use clang-format
to format the changes after running the tool.
Differential revision: https://reviews.llvm.org/D56822
llvm-svn: 353271
stored relative to VFRAME
Summary:
This patch makes LLDB able to retrieve proper values for function arguments and
local variables stored in PDB relative to VFRAME register.
Patch contains retrieval of corresponding FPO table entries from PDB and a
generic translator from FPO programs to DWARF expressions to get correct VFRAME
value.
Patch also improves variables-locations.test and makes this test passable on
x86.
Patch By: leonid.mashinsky
Reviewers: zturner, asmith, stella.stamenova, aleksandr.urakov
Reviewed By: zturner
Subscribers: arphaman, labath, mgorny, aprantl, JDevlieghere, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D55122
llvm-svn: 352845
This patch adds the file provider which is responsible for capturing
files used by LLDB.
When capturing a reproducer, we use a file collector that is very
similar to the one used in clang. For every file that we touch, we add
an entry with a mapping from its virtual to its real path. When we
decide to generate a reproducer we copy over the files and their
permission into to reproducer folder.
When replaying a reproducer, we load the VFS mapping and instantiate a
RedirectingFileSystem. The latter will transparently use the files
available in the reproducer.
I've tested this on two macOS machines with an artificial example.
Still, it is very likely that I missed some places where we (still) use
native file system calls. I'm hoping to flesh those out while testing
with more advanced examples. However, I will fix those things in
separate patches.
Differential revision: https://reviews.llvm.org/D54617
llvm-svn: 352538
Summary:
This patch adds the basic support of methods reconstruction by native PDB
plugin. It contains only most obvious changes (it processes LF_ONEMETHOD and
LF_METHOD records), some things still remain unsolved:
- mangled names retrieving;
- support of template methods.
Reviewers: zturner, labath, lemo, stella.stamenova
Reviewed by: zturner
Differential Revision: https://reviews.llvm.org/D56126
llvm-svn: 352464
This patch extends SymbolFileBreakpad::AddSymbols to include the symbols
from the FUNC records too. These symbols come from the debug info and
have a size associated with them, so they are given preference in case
there is a PUBLIC record for the same address.
To achieve this, I first pre-process the symbols into a temporary
DenseMap, and then insert the uniqued symbols into the module's symtab.
Reviewers: clayborg, lemo, zturner
Reviewed By: clayborg
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D56590
llvm-svn: 351781
The new LLVM header is one line shorter than the old one, which lead to
some test failures. Ideally tests should rely on line numbers for
breakpoints or output, but that's a different discussion. Hopefully this
turns the bots green again.
llvm-svn: 351779
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
This reapplies commit r351330, which was reverted due to a failing test on
macos. The failure was because the SymbolVendor used on MacOS was stricter than
the default (or ELF) symbol vendor, and rejected the symbol file because it's
UUID did not match the object file.
This version of the patch adds a uuid load command to the test macho file to
make sure the UUIDs match.
llvm-svn: 351447
This reapplies r350802, which was reverted because of issues with
parsing posix-style paths on windows hosts (and vice-versa). These have
since been fixed in r351328, and lldb should now recognise the path
style used in a dwarf compile unit correctly.
llvm-svn: 351435
This patch changes the behavior when printing C++ function references:
where we previously would get a <could not determine size>, there is
now a <no summary available>. It's not clear to me whether this is a
bug or an omission, but it's one step further than LLDB previously
got.
Differential Revision: https://reviews.llvm.org/D56798
llvm-svn: 351376
In this test we have deliberately removed all information which may hint
at the correct path style, so we cannot assert that lldb uses a
particular style. Instead, we should just check that it does something
vaguely reasonable.
llvm-svn: 351359
Summary:
Adding a breakpad symbol file to an existing MachO module with "target symbols
add" currently works only if one's host platform is a mac. This is
because SymbolVendorMacOSX (which is the one responsible for loading
symbols for MachO files) is conditionally compiled for the mac platform.
While we will sooner or later have a special symbol vendor for breakpad
files (to enable more advanced searching), and so this flow could be
made to work through that, it's not clear to me whether this should be a
requirement for the "target symbols add" flow to work. After all, since
the user has explicitly specified the symbol file to use, the symbol
vendor plugin's job is pretty much done.
This patch teaches the default symbol vendor to respect module's symbol
file spec, and load the symbol from that file if it is specified (and no
plugin requests any special handling).
Reviewers: clayborg, zturner, lemo
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D56589
llvm-svn: 351330
Summary:
If we opened a file which was produced on system with different path
syntax, we would parse the paths from the debug info incorrectly.
The reason for that is that we would parse the paths as they were
native. For example this meant that on linux we would treat the entire
windows path as a single file name with no directory component, and then
we would concatenate that with the single directory component from the
DW_AT_comp_dir attribute. When parsing posix paths on windows, we would
at least get the directory separators right, but we still would treat
the posix paths as relative, and concatenate them where we shouldn't.
This patch attempts to remedy this by guessing the path syntax used in
each compile unit. (Unfortunately, there is no info in DWARF which would
give the definitive path style used by the produces, so guessing is all
we can do.) Currently, this guessing is based on the DW_AT_comp_dir
attribute of the compile unit, but this can be refined later if needed
(for example, the DW_AT_name of the compile unit may also contain some
useful info). This style is then used when parsing the line table of
that compile unit.
This patch is sufficient to make the line tables come out right, and
enable breakpoint setting by file name work correctly. Setting a
breakpoint by full path still has some kinks (specifically, using a
windows-style full path will not work on linux because the path will be
parsed as a linux path), but this will require larger changes in how
breakpoint setting works.
Reviewers: clayborg, zturner, JDevlieghere
Subscribers: aprantl, lldb-commits
Differential Revision: https://reviews.llvm.org/D56543
llvm-svn: 351328
Summary:
This commit adds the glue code necessary to integrate the
SymbolFileBreakpad into the plugin system. Most of the methods are
stubbed out. The only method implemented method is AddSymbols, which
parses the PUBLIC "section" of the breakpad "object file", and fills out
the Module's symtab.
To enable testing this, I've made two additional changes:
- dump Symtab from the SymbolVendor class. The symtab was already being
dumped as a part of the object file dump, but that happened before
symbol vendor kicked in, so it did not reflect any symbols added
there.
- add ability to explicitly specify the external symbol file in
lldb-test (so that the object file could be linked with the breakpad
symbol file). To make things simpler, I've changed lldb-test from
consuming multiple inputs (and dumping their symbols) to having it
just process a single file per invocation. This was not a problem
since everyone was using it that way already.
Reviewers: clayborg, zturner, lemo, markmentovai, amccarth
Subscribers: mgorny, lldb-commits
Differential Revision: https://reviews.llvm.org/D56173
llvm-svn: 350924
The code was assuming that the elf file will have a PT_LOAD segment
starting from the first byte of the file. While this is true for files
generated by most linkers (it's a way of saving space), it is not a
requirement. And files not satisfying this constraint can still be
perfectly executable. yaml2obj is one of the tools which produces files
like this.
This patch relaxes the check in ObjectFileELF to take the address of the
first PT_LOAD segment as the base address of the object (instead of the
one with the offset 0). Since the PT_LOAD segments are supposed to be
sorted according to the VM address, this entry will also be the one with
the lowest VM address.
If we ever run into files which don't have the PT_LOAD segments sorted,
we can easily change this code to return the lowest VM address as the
base address (if that is the correct thing to do for these files).
llvm-svn: 350923
Typedefs are represented as S_UDT records in the globals stream. This
creates a strange situation where "types" are actually represented as
"symbols", so they need special handling.
In order to test this, we don't just use lldb and print out some
variables causing the AST to get created, because variables whose type
is a typedef will have debug info referencing the original type, not the
typedef. So we use lldb-test instead which will parse all debug info in
the entire file. This exposed some problems with lldb-test and the
native reader, mainly that certain types of obscure symbols which we can
find when iterating every single record would trigger crashes. These
have been fixed as well so that lldb-test can be used to test this
functionality.
Differential Revision: https://reviews.llvm.org/D56461
llvm-svn: 350888
This reverts commit r350802 because the test fails on windows. This
happens because we treat the paths as windows paths even though they
have linux path separators in the asm file. That results in wrong paths
being computed (\tmp\tmp\a.c instead of /tmp/a.c).
Reverting until I can figure out what to do with this.
llvm-svn: 350810
If a section name is exactly 8 bytes long (or has been truncated to 8
bytes), it will not contain the terminating nul character. This means
reading the name as a c string will pick up random data following the
name field (which happens to be the section vm size).
This fixes the name computation to avoid out-of-bounds access and adds a
test.
Reviewers: zturner, stella.stamenova
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D56124
llvm-svn: 350809
Summary:
The concept of a base address was already present in the implementation
(it's needed for computing section load addresses properly), but it was
never exposed through this function. This fixes that.
llvm-svn: 350804
Summary:
The motivation for this is being able to write tests for the upcoming
breakpad line table parser, but this could be useful for testing the
low-level workings of any line table format. Or simply for viewing the
line table information with more detail (the brief format doesn't
include any of the flags for end_of_prologue and similar).
I've also removed the load_addresses argument from the
DumpCompileUnitLineTable function, as it wasn't being used anywhere.
Reviewers: clayborg, zturner
Subscribers: JDevlieghere, lldb-commits
Differential Revision: https://reviews.llvm.org/D56315
llvm-svn: 350802
CHECK-DAG can't really be mixed with CHECK-NEXT statements because
each non DAG check sets a new search-origin for following CHECK-DAG
statements. This was passing by coincidence before, but a benign
change in the way we process symbols caused the order of the output
to be different, which triggered this test to fail.
This change makes the test resilient against ordering problems by
running a separate invocation of FileCheck for each function that
we want to test.
Note that with the Native PDB reader, we have full control over
the ordering that symbols are processed in, so we don't have
to worry about different machines returning things in different
orders due to different DIA SDK versions.
llvm-svn: 350773
ParseDeclsForContext was originally created to serve the very specific
case where the context is a function block. It was never intended to be
used for arbitrary DeclContexts, however due to the generic name, the
DWARF and PDB plugins implemented it in this way "just in case". Then,
lldb-test came along and decided to use it in that way.
Related to this, there are a set of functions in the SymbolFile class
interface whose requirements and expectations are not documented. For
example, if you call ParseCompileUnitFunctions, there's an inherent
requirement that you create entries in the underlying clang AST for
these functions as well as their signature types, because in order to
create an lldb_private::Function object, you have to pass it a
CompilerType for the parameter representing the signature.
On the other hand, there is no similar requirement (either inherent or
documented) if one were to call ParseDeclsForContext. Specifically, if
one calls ParseDeclsForContext, and some variable declarations, types,
and other things are added to the clang AST, is it necessary to create
lldb::Variable, lldb::Type, etc objects representing them? Nobody knows.
There is, however, an accidental requirement, because since all of the
plugins implemented this just in case, lldb-test came along and used
ParsedDeclsForContext, and then wrote check lines that depended on this.
When I went to try and implemented the NativePDB reader, I did not
adhere to this (in fact, from a layering perspective I went out of my
way to avoid it), and as a result the existing DIA PDB tests don't work
when the native PDB reader is enabled, because they expect that calling
ParseDeclsForContext will modify the *module's* view of symbols, and not
just the internal AST.
All of this confusion, however, can be avoided if we simply stick to
using ParseDeclsForContext for its original intended use case (blocks),
and use a different function (ParseAllDebugSymbols) for its intended use
case which is, unsuprisingly, to parse all the debug symbols (which is
all lldb-test really wanted to do anyway).
In the future, I would like to change ParseDeclsForContext to
ParseDeclsForFunctionBlock, then delete all of the dead code inside that
handles other types of DeclContexts (and probably even assert if the
DeclContext is anything other than a block).
A few PDB tests needed to be fixed up as a result of this, and this also
exposed a couple of bugs in the DIA PDB reader (doesn't matter much
since it should be going away soon, but worth mentioning) where the
appropriate AST entries weren't being created always.
Differential Revision: https://reviews.llvm.org/D56418
llvm-svn: 350764
Summary:
This is the result of the discussion in D55356, where it was suggested
as a solution to representing the addresses that logically belong to a
module in memory, but are not a part of any of its sections.
The ELF PT_LOAD segments are similar to the MachO "load commands",
except that the relationship between them and the object file sections
is a bit weaker. While in the MachO case, the sections belonging to a
specific segment are placed directly inside it in the object file
logical structur, in the ELF case, the sections and segments form two
separate hierarchies. This means that it is in theory possible to create
an elf file where only a part of a section would belong to some segment
(and another part to a different one). However, I am not aware of any
tool which would produce such a file (and most tools will have problems
ingesting them), so this means it is still possible to follow the MachO
model and make sections children of the PT_LOAD segments.
In case we run into (corrupt?) files with overlapping sections, I have
added code (and tests) which adjusts the sizes and/or drops the offending
sections in order to present a reasonable image to the upper layers of
LLDB. This is mostly done for completeness, as I don't anticipate
running into this situation in the real world. However, if we do run
into it, and the current behavior is not suitable for some reason, we
can implement this logic differently.
Reviewers: clayborg, jankratochvil, krytarowski, joerg, espindola
Subscribers: emaste, arichardson, lldb-commits
Differential Revision: https://reviews.llvm.org/D55998
llvm-svn: 350742
Summary: In standalone builds `LLVM_BINARY_DIR` was equal to `LLDB_BINARY_DIR` so far. This is counterintuitive and invalidated the values of `LLDB_DEFAULT_TEST_DSYMUTIL/FILECHECK/COMPILER` etc.
Reviewers: zturner, labath, clayborg, JDevlieghere, stella.stamenova, serge-sans-paille
Reviewed By: labath
Subscribers: mgorny, friss, lldb-commits, #lldb
Differential Revision: https://reviews.llvm.org/D56443
llvm-svn: 350738
Summary:
In standalone builds of LLDB we currently have no target `count` that tests can depend on. This may be fixed in the future by exporting the target from LLVM similar to targets llvm-config, dsymutil and others.
In-tree build with this patch:
```
$ ninja -t query tools/lldb/lit/check-lldb-lit
tools/lldb/lit/check-lldb-lit:
input: phony
tools/lldb/lit/CMakeFiles/check-lldb-lit
bin/FileCheck
bin/clang
bin/count
bin/darwin-debug
bin/debugserver
bin/dsymutil
bin/llc
bin/lldb
bin/lldb-mi
bin/lldb-server
bin/lldb-test
bin/llvm-config
bin/llvm-mc
bin/llvm-objcopy
bin/not
bin/yaml2obj
lib/liblldb.dylib
projects/libcxx/lib/cxx
tools/lldb/unittests/LLDBUnitTests
outputs:
tools/lldb/test/check-lldb
check-lldb-lit
```
Standalone build with this patch:
```
$ ninja -t query lit/check-lldb-lit
lit/check-lldb-lit:
input: phony
lit/CMakeFiles/check-lldb-lit
bin/darwin-debug
bin/debugserver
bin/lldb
bin/lldb-mi
bin/lldb-server
bin/lldb-test
lib/liblldb.dylib
unittests/LLDBUnitTests
outputs:
test/check-lldb
check-lldb-lit
```
Reviewers: davide, aprantl, JDevlieghere, alexshap
Reviewed By: davide
Subscribers: mgorny, lldb-commits, #lldb
Differential Revision: https://reviews.llvm.org/D56389
llvm-svn: 350538
Summary:
This patch allows ObjectFileBreakpad to parse the contents of Breakpad
files into sections. This sounds slightly odd at first, but in essence
its not too different from how other object files handle things. For
example in elf files, the symtab section consists of a number of
"records", where each record represents a single symbol. The same is
true for breakpad's PUBLIC section, except in this case, the records will be
textual instead of binary.
To keep sections contiguous, I create a new section every time record
type changes. Normally, the breakpad processor will group all records of
the same type in one block, but the format allows them to be intermixed,
so in general, the "object file" may contain multiple sections with the
same record type.
Reviewers: clayborg, zturner, lemo, markmentovai, amccarth
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D55434
llvm-svn: 350511
This is a first step towards getting lldb-test symbols working
with the native plugin. There is a remaining issue, which is
that the plugin expects that ParseDeclsForContext will also
create lldb symbols rather than just the decls, but the native
pdb plugin doesn't currently do this. This will be addressed
in a followup patch.
llvm-svn: 350243
A Previous patch added support for creating VarDecls for global
variables. This patch updates this test to be more strict and
actually check these, not just the types.
llvm-svn: 350242
There were several problems preventing this from working. The
first is that when the PDB had an absolute path to the main
source file, we would construct an invalid path by prepending the
compilation directory to it anyway. So we needed to check if the
path is already absolute first.
Second, LLDB assumes that the zero'th item in the support file list
is the main compilation unit. We were respecting this requirement,
but LLDB *also* requires that file to appear somewhere in the list
starting from index 1 as well. So the main compilation file should
appear in the support file list twice. And when parsing a line
table, it expects the LineEntry records to be constructed using
the 1-based index. With these two fixes we can now set breakpoints
by file and line using the native PDB reader.
llvm-svn: 350240
Summary:
r346165 introduced a bug, where we would fail to parse the size of an
array if that size happened to match an existing die offset.
The logic was:
if (DWARFDIE count = die.GetReferencedDie(DW_AT_count))
num_elements = compute_vla_size(count);
else
num_elements = die.GetUsigned(DW_AT_count); // a fixed-size array
The problem with this logic was that GetReferencedDie did not take the
form class of the attribute into account, and would happily return a die
reference for any form, if its value happened to match some die.
As this behavior is inconsistent with how llvm's DWARFFormValue class
operates, I chose to fix the problem by making our version of this class
match the llvm behavior. For this to work, I had to add an explicit form
class check to the .apple_XXX tables parsing code, because they do
(incorrectly?) use data forms as die references.
Reviewers: aprantl, clayborg
Subscribers: JDevlieghere, lldb-commits
Differential Revision: https://reviews.llvm.org/D55991
llvm-svn: 350086
Previously we would create these for local variables but not for
global variables.
Also updated existing tests which created global variables to check
for them in the resulting AST.
llvm-svn: 349854
Unexpected successes should be considered failures because they can hide
regressions when not addressed. When a test is fixed and not re-enabled,
it can easily regress without us noticing.
I couldn't find a good way to make this change other than changing it in
the unittest2 framework. I know this is less than optimal but since we
have the dependency checked in and the change is pretty fundamental to
the framework I think it's not unreasonable.
Differential revision: https://reviews.llvm.org/D55835
llvm-svn: 349818
This test is now marked as unsupported on Windows - it is not technically "unsupported" on Windows, but it fails because "expr ptr" does not evaluate correctly. However, the error message contains the expected string, so the test "passes" despite the fact that the commands failed
The following bug has been opened for it: llvm.org/pr40119
llvm-svn: 349784
We had a use after free where we were assigning the result of a function
that returned a string to a StringRef. After fixing this use after
free, one of the DIA PDB tests now passes with the native PDB reader,
so we enable the test under native mode as well. The goal is to
eventually make all the tests pass under both, at which point we can
disable them all under DIA mode.
llvm-svn: 349673
The build.py script always runs the compiler in C++ mode, regardless of
the file extension. This results in mangled names presented to the
linker which in turn cannot find the printf symbol.
While we figure out how to solve this issue I've turned the source file
into a cpp file and added extern c. This should unbreak the bots.
llvm-svn: 349642
We reconstruct the AST hierarchy by trying to hack up a mangled
name for the parent type using the child type's mangled name.
This was failing for enums because their tag type is represented
with two letters ("W4") instead of one letter ("T", "U", etc) as
it is with classes, structs, and unions. After accounting for
this we can now correctly determine when an enum is nested
inside of a namespace or a class.
llvm-svn: 349565
Summary:
The first section header does not define a real section. Instead it is
used for various elf extensions. This patch skips creation of a section
for index 0.
This has one furtunate side-effect, in that it allows us to use the section
header index as the Section ID (where 0 is also invalid). This way, we
can get rid of a lot of spurious +1s in the ObjectFileELF code.
Reviewers: clayborg, krytarowski, joerg, espindola
Subscribers: emaste, lldb-commits, arichardson
Differential Revision: https://reviews.llvm.org/D55757
llvm-svn: 349498
Summary:
This should enable the compiler to find the system linker for the link
step.
Reviewers: stella.stamenova, zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D55736
llvm-svn: 349461
Each process plug-in can create its own custom commands. I figured it would be nice to be able to dump things from the minidump file from the lldb command line, so I added the start of the some custom commands.
Currently you can dump:
minidump stream directory
all linux specifc streams, most of which are strings
each linux stream individually if desired, or all with --linux
The idea is we can expand the command set to dump more things, search for data in the core file, and much more. This patch gets us started.
Differential Revision: https://reviews.llvm.org/D55727
llvm-svn: 349429
Previously the code that parsed debug info to create lldb's Symbol
objects such as Variable, Type, Function, etc was tightly coupled
to the AST reconstruction code. This made it difficult / impossible
to implement functions such as ParseDeclsForContext() that were only
supposed to be operating on clang AST's. By splitting these apart,
the logic becomes much cleaner and we have a clear separation of
responsibilities.
llvm-svn: 349383
This is a little dangerous since the crashlog files aren't 100%
unambiguous, but the risk is mitigated by using a non-greedy +?
pattern.
rdar://problem/38478511
Differential Revision: https://reviews.llvm.org/D55608
llvm-svn: 349367
Summary:
Previously lldb-test's LinePrinter would output the indentation spaces
even on completely empty lines. This is not nice, as trailing spaces get
flagged as errors in some tools/editors, and it prevents FileCheck's
CHECK-EMPTY from working.
Equally annoying was the fact that the LinePrinter did not terminate
it's output with a newline (instead it would leave the unterminated hanging
indent from the last NewLine() command), which meant that the shell prompt
following the lldb-test command came out wrong.
This fixes both issues by changing how newlines are handled. NewLine(),
which was ending the previous line ('\n') *and* begging the next line by
printing the indent, is now "demoted" to just printing literal "\n".
Instead, lines are now delimited via a helper Line object, which makes
sure the line is indented and terminated in an RAII fashion. The typical
usage would be:
Printer.line() << "This text will be indented and terminated";
If one needs to do more work than it will fit into a single statement,
one can also assign the result of the line() function to a local
variable. The line will then be terminated when that object goes out of
scope.
Reviewers: zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D55597
llvm-svn: 349269
Summary:
This patch attempts to move as much code as possible out of the
CreateSections function to make room for future improvements there. Some
of this may be slightly over-engineered (VMAddressProvider), but I
wanted to keep the logic of this function very simple, because once I
start taking segment headers into acount (as discussed in D55356), the
function is going to grow significantly.
While in there, I also added tests for various bits of functionality.
This should be NFC, except that I changed the order of hac^H^Heuristicks
for determining section type slightly. Previously, name-based deduction
(.symtab -> symtab) would take precedence over type-based (SHT_SYMTAB ->
symtab) one. In fact we would assert if we ran into a .text section with
type SHT_SYMTAB. Though unlikely to matter in practice, this order
seemed wrong to me, so I have inverted it.
Reviewers: clayborg, krytarowski, espindola
Subscribers: emaste, arichardson, lldb-commits
Differential Revision: https://reviews.llvm.org/D55706
llvm-svn: 349268
Since we're actually running an executable on the host now, different
versions of Windows could load different system libraries, so we need
to regex out the number of loaded modules.
llvm-svn: 349175
This patch adds support for parsing and evaluating local variables.
using the native pdb plugin.
Differential Revision: https://reviews.llvm.org/D55575
llvm-svn: 349067
We've recently developed a convention where the tests are placed into
subfolders according to the object file type. This applies that
convention to existing tests too.
llvm-svn: 349027
Move code into a separate function, and replace the if-else chain with
llvm::StringSwitch.
A slight behavioral change is that now I use the section flags
(SHF_TLS) instead of the section name to set the thread-specific
property. There is no explanation in the original commit introducing
this (r153537) as to why that was done this way, but the new behavior
should be more correct.
llvm-svn: 348936
Previously, lldb-test would only print top-level sections. However, in
lldb, sections can contain other sections. This teaches lldb-test to
print nested sections too.
llvm-svn: 348924
Summary:
This implements the gcc builder in build.py script to allow it to
compile host executables when running on a non-windows host. Where it
made sense, I tried to share code with the msvc builder by moving stuff
to the base class.
Reviewers: zturner
Subscribers: mehdi_amini, dexonsmith, lldb-commits
Differential Revision: https://reviews.llvm.org/D55430
llvm-svn: 348918
This re-commits r348592, which was reverted due to a failing test on
macos.
The issue was that I was passing a null pointer for the
"CreateMemoryInstance" callback when registering ObjectFileBreakpad,
which caused crashes when attemping to load modules from memory. The
correct thing to do is to pass a callback which always returns a null
pointer (as breakpad files are never loaded in inferior memory).
It turns out that there is only one test which exercises this code path,
and it's mac-only, so I've create a new test which should run everywhere
(except windows, as one cannot delete an executable which is being run).
Unfortunately, this test still fails on linux for other reasons, but at
least it gives us something to aim for.
The original commit message was:
This patch adds the scaffolding necessary for lldb to recognise symbol
files generated by breakpad. These (textual) files contain just enough
information to be able to produce a backtrace from a crash
dump. This information includes:
- UUID, architecture and name of the module
- line tables
- list of symbols
- unwind information
A minimal breakpad file could look like this:
MODULE Linux x86_64 0000000024B5D199F0F766FFFFFF5DC30 a.out
INFO CODE_ID 00000000B52499D1F0F766FFFFFF5DC3
FILE 0 /tmp/a.c
FUNC 1010 10 0 _start
1010 4 4 0
1014 5 5 0
1019 5 6 0
101e 2 7 0
PUBLIC 1010 0 _start
STACK CFI INIT 1010 10 .cfa: $rsp 8 + .ra: .cfa -8 + ^
STACK CFI 1011 $rbp: .cfa -16 + ^ .cfa: $rsp 16 +
STACK CFI 1014 .cfa: $rbp 16 +
Even though this data would normally be considered "symbol" information,
in the current lldb infrastructure it is assumed every SymbolFile object
is backed by an ObjectFile instance. So, in order to better interoperate
with the rest of the code (particularly symbol vendors).
In this patch I just parse the breakpad header, which is enough to
populate the UUID and architecture fields of the ObjectFile interface.
The rough plan for followup patches is to expose the individual parts of
the breakpad file as ObjectFile "sections", which can then be used by
other parts of the codebase (SymbolFileBreakpad ?) to vend the necessary
information.
Reviewers: clayborg, zturner, lemo, amccarth
Subscribers: mgorny, fedor.sergeev, markmentovai, lldb-commits
Differential Revision: https://reviews.llvm.org/D55214
llvm-svn: 348773
The test still only passes when not run from VS because the previous patch did not remove the original build commands.... This also simplifies the build command by removing some defaults
llvm-svn: 348664
Previously we would create an lldb::Function object for each function
parsed, but we would not add these to the clang AST. This is a first
step towards getting local variable support working, as we first need an
AST decl so that when we create local variable entries, they have the
proper DeclContext.
Differential Revision: https://reviews.llvm.org/D55384
llvm-svn: 348631
Summary:
This patch adds the scaffolding necessary for lldb to recognise symbol
files generated by breakpad. These (textual) files contain just enough
information to be able to produce a backtrace from a crash
dump. This information includes:
- UUID, architecture and name of the module
- line tables
- list of symbols
- unwind information
A minimal breakpad file could look like this:
MODULE Linux x86_64 0000000024B5D199F0F766FFFFFF5DC30 a.out
INFO CODE_ID 00000000B52499D1F0F766FFFFFF5DC3
FILE 0 /tmp/a.c
FUNC 1010 10 0 _start
1010 4 4 0
1014 5 5 0
1019 5 6 0
101e 2 7 0
PUBLIC 1010 0 _start
STACK CFI INIT 1010 10 .cfa: $rsp 8 + .ra: .cfa -8 + ^
STACK CFI 1011 $rbp: .cfa -16 + ^ .cfa: $rsp 16 +
STACK CFI 1014 .cfa: $rbp 16 +
Even though this data would normally be considered "symbol" information,
in the current lldb infrastructure it is assumed every SymbolFile object
is backed by an ObjectFile instance. So, in order to better interoperate
with the rest of the code (particularly symbol vendors).
In this patch I just parse the breakpad header, which is enough to
populate the UUID and architecture fields of the ObjectFile interface.
The rough plan for followup patches is to expose the individual parts of
the breakpad file as ObjectFile "sections", which can then be used by
other parts of the codebase (SymbolFileBreakpad ?) to vend the necessary
information.
Reviewers: clayborg, zturner, lemo, amccarth
Subscribers: mgorny, fedor.sergeev, markmentovai, lldb-commits
Differential Revision: https://reviews.llvm.org/D55214
llvm-svn: 348592
Summary:
This parses entries in pecoff import tables for imported DLLs and
is intended as the first step to allow LLDB to load a PE's shared
modules when creating a target on the LLDB console.
Reviewers: rnk, zturner, aleksandr.urakov, lldb-commits, labath, asmith
Reviewed By: labath, asmith
Subscribers: labath, lemo, clayborg, Hui, mgorny, mgrang, teemperor
Differential Revision: https://reviews.llvm.org/D53094
llvm-svn: 348527
In compile-and-link mode, the user doesn't specify the name of the
object files to generate, because there could be multiple inputs
on a single command line and this would be hard to specify. So the
script just tries to be smart and figure out the best object file
names. However, if two build scripts are running in parallel and
using the same source files as input, they would previously race
to write the same object files, since the computed name only considered
the source file names when computing the object file names.
With this patch, we also consider the final executable name. In a
way, this "namespaces" the generated object files so that as long
as the final executable file names don't clash, the intermediate
object file names won't clash either.
llvm-svn: 348511
George Rimar