While it doesn't make a *ton* of sense for POSIX paths to be
in PDBs, it's possible to occur in real scenarios involving
cross compilation.
The tools need to be able to handle this, because certain types
of debugging scenarios are possible without a running process
and so don't necessarily require you to be on a Windows system.
These include post-mortem debugging and binary forensics (e.g.
using a debugger to disassemble functions and examine symbols
without running the process).
There's changes in clang, LLD, and lldb in this patch. After
this the cross-platform disassembly and source-list tests pass
on Linux.
Furthermore, the behavior of LLD can now be summarized by a much
simpler rule than before: Unless you specify /pdbsourcepath and
/pdbaltpath, the PDB ends up with paths that are valid within
the context of the machine that the link is performed on.
Differential Revision: https://reviews.llvm.org/D53149
llvm-svn: 344269
LC_BUILD_VERSION load command handling - this
commit is a combination of patches by Adrian
Prantl and myself. llvm::Triple::BridgeOS
isn't defined yet, so all references to that
are currently commented out.
Also update Xcode project file to build the
NativePDB etc plugins.
<rdar://problem/43353615>
llvm-svn: 344209
Summary:
If the process exits before any initial stop then notify the debugger
of the error otherwise WaitForDebuggerConnection() will be blocked.
An example of this issue is when a process fails to load a dependent DLL.
In addition to the fix, remove a duplicate call to FreeProcessHandles() in
DebuggerThread::HandleExitProcessEvent() and use decimal format
for all thread IDs.
Reviewers: rnk, zturner, aleksandr.urakov
Reviewed By: zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D53090
llvm-svn: 344168
The existing SymbolFilePDB only works on Windows, as it is written
against a closed-source Microsoft SDK that ships with their debugging
tools.
There are several reasons we want to bypass this and go straight to the
bits of the PDB, but just to list a few:
More room for optimization. We can't see inside the implementation of
the Microsoft SDK, so we don't always know if we're doing things in the
most efficient way possible. For example, setting a breakpoint on main
of a big program currently takes several seconds. With the
implementation here, the time is unnoticeable.
We want to be able to symbolize Windows minidumps even if not on
Windows. Someone should be able to debug Windows minidumps as if they
were on Windows, given that no running process is necessary.
This patch is a very crude first attempt at filling out some of the
basic pieces.
I've implemented FindFunctions, ParseCompileUnitLineTable, and
ResolveSymbolContext for a limited subset of possible parameter values,
which is just enough to get it to display something nice for the
breakpoint location.
I've added several tests exercising this functionality which are limited
enough to work on all platforms but still exercise this functionality.
I'll try to add as many tests of this nature as I can, but at some
point we'll need a live process.
For now, this plugin is enabled always on non-Windows, and by setting
the environment variable LLDB_USE_NATIVE_PDB_READER=1 on Windows.
Eventually, once it's at parity with the Windows implementation, we'll
delete the Windows DIA-based implementation.
Differential Revision: https://reviews.llvm.org/D53002
llvm-svn: 344154
There are several places that call `FindRanges`,
all of them use `Slide` to adjust the ranges found
by the base address.
All except one, which does the same manually in a loop.
Patch updates it to use `Slide` for consistency.
llvm-svn: 344122
This adds a basic support of the .debug_rnglists section.
Only the DW_RLE_start_length and DW_RLE_end_of_list entries are supported.
Differential revision: https://reviews.llvm.org/D52981
llvm-svn: 344119
This patch teaches lldb to detect when there are missing frames in a
backtrace due to a sequence of tail calls, and to fill in the backtrace
with artificial tail call frames when this happens. This is only done
when the execution history can be determined from the call graph and
from the return PC addresses of calls on the stack. Ambiguous sequences
of tail calls (e.g anything involving tail calls and recursion) are
detected and ignored.
Depends on D49887.
Differential Revision: https://reviews.llvm.org/D50478
llvm-svn: 343900
Summary:
Add settings to control command echoing:
```
(lldb) settings set interpreter.echo-commands true
(lldb) settings set interpreter.echo-comment-commands true
```
Both settings default to true, which keeps LLDB's existing behavior in non-interactive mode (echo all command inputs to the output).
So far the only way to change this behavior was the `--source-quietly` flag, which disables all output including evaluation results.
Now `echo-commands` allows to turn off echoing for commands, while evaluation results are still printed. No effect if `--source-quietly` was present.
`echo-comment-commands` allows to turn off echoing for commands in case they are pure comment lines. No effect if `echo-commands` is false.
Note that the behavior does not change immediately! The new settings take effect only with the next command source.
LLDB lit test are the main motivation for this feature. So far incoming `#CHECK` line have always been echoed to the output and so they could never fail. Now we can disable it in lit-lldb-init.
Todos: Finish test for this feature. Add to lit-lldb-init. Check for failing lit tests.
Reviewers: aprantl, jasonmolenda, JDevlieghere
Subscribers: friss, lldb-commits
Differential Revision: https://reviews.llvm.org/D52788
llvm-svn: 343859
Summary:
This function existed (with identical code) in both NativeProcessLinux
and NativeProcessNetBSD, and it is likely that it would be useful to any
future implementation of NativeProcessProtocol.
Therefore I move it to the base class.
Reviewers: krytarowski
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D52719
llvm-svn: 343683
This is a follow-up to https://reviews.llvm.org/D46362.
When evaluating a complex expression in DWARFExpression::Evaluate,
file addresses must be resolved to load addresses before we can
perform operations such as DW_OP_deref on them.
For this the address goes through three steps
1. Read the file address as stored in the DWARF
2. Link/relocate the file address (when reading from a .dSYM, this is a no-op)
3. Convert the file address to a load address.
D46362 implemented step (3) by resolving the file address using the
Module that the original DWARF came from. In the case of a dSYM that
is correct, but when reading from .o files, we need to look up
relocated/linked addresses, so the right place to look them up is the
current frame's module. This patch fixes that by setting the
expression's Module to point to the linked debugmap object.
A word a bout the unorthodox testcase: The motivating testcase for
this fix is in Swift, but I managed to hand-modify LLVM-IR for a
trivial C program to exhibit the same problem, so we can fix this in
llvm.org.
rdar://problem/44689915
Differential Revision: https://reviews.llvm.org/D52678
llvm-svn: 343612
Summary:
If there is no newline the "lldb" prompt could be on the wrong line. To reproduce the missing newline you can do 'image dump smytab' on any binary.
Previously
Symtab, file = D:\upstream\build\Debug\bin\clang-diff.exe, num_symbols = 0(lldb)
Now
Symtab, file = D:\upstream\build\Debug\bin\clang-diff.exe, num_symbols = 0
(lldb)
Reviewers: zturner, aleksandr.urakov, lldb-commits
Subscribers: abidh
Differential Revision: https://reviews.llvm.org/D52627
llvm-svn: 343497
Currently we reject our own default disassembly-format string because it
contains two backticks which causes everything in between to be
interpreter as an expression by the command interpreter. This patch
fixes that by escaping backticks when dumping format strings.
llvm-svn: 343471
Summary:
This function encodes the knowledge of whether the PC points to the
breakpoint instruction of the one following it after the breakpoint is
"hit". This behavior mainly(*) depends on the architecture and not on the
OS, so it makes sense for it to be implemented in the base class, where
it can be shared between different implementations (Linux and NetBSD
atm).
(*) It is possible for an OS to expose a different API, perhaps by doing
some fixups in the kernel. In this case, the implementation can override
this function to implement custom behavior.
Reviewers: krytarowski, zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D52532
llvm-svn: 343409
Summary:
The `ClangUserExpression::GetLanguageForExpr` method is currently a big
source of sadness, as it's name implies that it's an accessor method, but it actually
is also initializing some variables that we need for parsing. This caused that we
currently call this getter just for it's side effects while ignoring it's return value,
which is confusing for the reader.
This patch renames it to `UpdateLanguageForExpr` and merges all calls to the
method into a single call in `ClangUserExpression::PrepareForParsing` (as calling
this method is anyway mandatory for parsing to succeed)
While looking at the code, I also found that we actually have two language
variables in this class hierarchy. The normal `Language` from the UserExpression
class and the `LanguageForExpr` that we implemented in this subclass. Both
don't seem to actually contain the same value, so we probably should look at this
next.
Reviewers: xbolva00
Reviewed By: xbolva00
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D52561
llvm-svn: 343191
Summary:
This patch implements restoring of the calling convention from PDB.
It is necessary for expressions evaluation, if we want to call a function
of the debuggee process with a calling convention other than ccall.
Reviewers: clayborg, zturner, labath, asmith
Reviewed By: clayborg
Subscribers: teemperor, lldb-commits, stella.stamenova
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D52501
llvm-svn: 343084
NativeProcessProtocol::ReadMemoryWithoutTrap had a bug, where it failed
to properly remove inserted breakpoint opcodes if the memory read
partially overlapped the trap opcode. This could not happen on x86
because it has a one-byte breakpoint instruction, but it could happen on
arm, which has a 4-byte breakpoint instruction (in arm mode).
Since triggerring this condition would only be possible on an arm
machine (and even then it would be a bit tricky). I test this using a
NativeProcessProtocol unit test.
llvm-svn: 343076
max number of stack frames to backtrace, make it a setting,
target.process.thread.max-backtrace-depth.
Add a test case for the setting.
<rdar://problem/28759559>
llvm-svn: 343029
AbsPosToLineColumnPos is the only reader of m_user_expression_start_pos
and actually treats it like a size_t. Also the value we store in
m_user_expression_start_pos is originally a size_t, so it makes sense
to change the type of this variable to size_t.
llvm-svn: 342804
When creating a target, lldb loads all dependent files (i.e. libs in
LC_LOAD_DYLIB for Mach-O). This can be confusing, especially when two
versions of the same library end up in the shared cache. It's possible
to change this behavior, by specifying target create -d <target> these
dependents are not loaded.
This patch changes the default behavior to only load dependent files
only when the target is an executable. When creating a target for a
library, it is now no longer necessary to pass -d. The user can still
override this behavior by specifying the -d option to change this
behavior.
rdar://problem/43721382
Differential revision: https://reviews.llvm.org/D51934
llvm-svn: 342634
This is an NFC commit to refactor the "load dependent files" parameter
from a boolean to an enum value. We want to be able to specify a
default, in which case we decide whether or not to load the dependent
files based on whether the target is an executable or not (i.e. a
dylib).
This is a dependency for D51934.
Differential revision: https://reviews.llvm.org/D51859
llvm-svn: 342633
Summary:
Completing inside the expression command now uses the new description API
to also provide additional information to the user. For now this information
are the types of variables/fields and the signatures of completed function calls.
Reviewers: #lldb, JDevlieghere
Reviewed By: JDevlieghere
Subscribers: JDevlieghere, lldb-commits
Differential Revision: https://reviews.llvm.org/D52103
llvm-svn: 342385
They're not that common, and falling back is definitely
better than throwing an error instead of the result. If we
feel motivated, we might end up implementing support for these,
but it's unclear whether it's worth the effort/complexity.
Fixes PR38925.
<rdar://problem/44436068>
llvm-svn: 342262
Summary:
This patch adds some symbol tag checks before using the `IPDBRawSymbol`
interface to improve safety and readability.
Reviewers: zturner
Reviewed By: zturner
Subscribers: lldb-commits, stella.stamenova
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D51967
llvm-svn: 342208
I started from a clean slate to do the checkin, but forgot to svn add the new files.
Do that now.
Also add the one new source file to CMakeLists.txt
llvm-svn: 342190
This change allows you to write a new breakpoint type where the
logic for setting breakpoints is determined by a Python callback
written using the SB API's.
Differential Revision: https://reviews.llvm.org/D51830
llvm-svn: 342185
Summary:
This patch adds a framework for adding descriptions to the command completions we provide.
It also adds descriptions for completed top-level commands so that we can test this code.
Completions are in general supposed to be displayed alongside the completion itself. The descriptions
can be used to provide additional information about the completion to the user. Examples for descriptions
are function signatures when completing function calls in the expression command or the binary name
when providing completion for a symbol.
There is still some boilerplate code from the old completion API left in LLDB (mostly because the respective
APIs are reused for non-completion related purposes, so the CompletionRequest doesn't make sense to be
used), so that's why I still had to change some function signatures. Also, as the old API only passes around a
list of matches, and the descriptions are for these functions just another list, I had to add some code that
essentially just ensures that both lists are always the same side (e.g. all the manual calls to
`descriptions->AddString(X)` below a `matches->AddString(Y)` call).
The initial command descriptions that come with this patch are just reusing the existing
short help that is already added in LLDB.
An example completion with descriptions looks like this:
```
(lldb) pl
Available completions:
platform -- Commands to manage and create platforms.
plugin -- Commands for managing LLDB plugins.
```
Reviewers: #lldb, jingham
Reviewed By: #lldb, jingham
Subscribers: jingham, JDevlieghere, lldb-commits
Differential Revision: https://reviews.llvm.org/D51175
llvm-svn: 342181
The two existing implementations have the function implemented
identically, and there's no reason to believe that this would be
different for other implementations.
llvm-svn: 342167
This patch improves the support of DWARF5.
Particularly the reporting of source code locations.
Differential revision: https://reviews.llvm.org/D51935
llvm-svn: 342153
After landing r341457, we started seeing a failure on the swift-lldb
bots. The change was correct and pretty straightforward, a DW_OP_constu
was replaced with DW_OP_lit23, the value remaining identical.
0x000000f4: DW_TAG_variable
DW_AT_location (0x00000000
[0x0000000100000a51, 0x0000000100000d47): DW_OP_lit23, DW_OP_stack_value)
DW_AT_name ("number")
However, this broke LLDB.
(Int) number = <extracting data from value failed>
The value was read correctly, but apparently the value's type was different.
When reading a constu it was reading a uint64 (m_type = e_ulonglong) while for
the literal, it got a signed int (m_type = e_sint). This change makes sure we
read the value as an unsigned.
Differential revision: https://reviews.llvm.org/D51730
llvm-svn: 342142
Summary:
Similar to what we did in D50681, we now stop manually byte counting here
in the SourceManager.
Reviewers: #lldb, JDevlieghere
Reviewed By: #lldb, JDevlieghere
Subscribers: JDevlieghere, abidh, lldb-commits
Differential Revision: https://reviews.llvm.org/D50809
llvm-svn: 342121
Summary:
Building LLDB with xcodebuild sets the compatibility version of liblldb
in LLDB.framework. Building the framework with cmake does not set the
compatibility version, and so it defaults to 0.0.0. This is a discrepency in the
difference between the xcode build and the cmake build.
I tested this change by building without this patch. From the build tree I ran
`otool -L Library/Frameworks/LLDB.framework/Versions/A/LLDB` and got this:
```
@rpath/LLDB.framework/Versions/A/LLDB (compatibility version 0.0.0, current version 8.0.0)
```
Did the same with this patch and the output contained this:
```
@rpath/LLDB.framework/Versions/A/LLDB (compatibility version 1.0.0, current version 8.0.0)
```
Reviewers: clayborg, labath
Subscribers: mgorny, lldb-commits
Differential Revision: https://reviews.llvm.org/D51959
llvm-svn: 342066
- gcc warning about using binary or for or-ing two comparisons (a == b | a == c)
- llvm style prefers static functions to functions in an anonymous namespace
llvm-svn: 342051
Summary:
One of the conclusions of the discussion on D49740 was that SafeMachO is better
off in the Host module (as that's the only place which should include
mach/machine.h, which is what this header is working around). Also, Utility,
which is the only module which cannot include Host, should not be doing
anything with object file formats.
This patch implements that move, and also removes any unneded includes of that
file.
I've verified that MacOS still compiles after this.
Reviewers: jingham, zturner, teemperor
Subscribers: fedor.sergeev, lldb-commits
Differential Revision: https://reviews.llvm.org/D50383
llvm-svn: 342050
Summary:
This patch removes the manual byte counting in all internal Stream methods.
This is now done by the automatic byte counting provided by calling `GetWrittenBytes()`
before and after writing the data (which is automatically done for us by the `ByteDelta`
utility class).
Reviewers: #lldb, JDevlieghere
Reviewed By: JDevlieghere
Subscribers: JDevlieghere, labath, lldb-commits
Differential Revision: https://reviews.llvm.org/D50681
llvm-svn: 342044
The warning is about heap-allocating a struct with bigger alignment
requirements than the standard heap allocator provides.
AFAICT, all uses of the XSAVE struct are already heap-allocated, so this
high alignment does not actually have any effect and removing it should
be NFC.
I have also done some digging in the commit history. This alignment
requirement was since the XSAVE struct was introduced in r180572 when
adding AVX register support for linux. It does not mention the alignment
specifically, so I am guessing this was just put there because the
corresponging XSAVE cpu instruction requires its buffer to be 64-byte
aligned. However, LLDB will not be normally reading this struct via the
XSAVE instruction directly. Instead we will ask the kernel to copy the
buffer saved when suspeding the inferior. This should not require such
strict alignment (in fact, linux kernel will happily do this for any
alignment).
llvm-svn: 342029
Summary:
This commit fixes following problems after rL341782:
- Broken SymbolFilePDBTests
- Warning on comparison of integers of different signs
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D51162
llvm-svn: 341942
Summary:
The check is inverted here: If we have error messages, we should print those instead
of our default error message. But currently we print the default message when we
actually have a sensible error to print.
Fixes https://bugs.llvm.org/show_bug.cgi?id=38383
Thanks Nat for the patch!
Reviewers: #lldb, JDevlieghere
Reviewed By: JDevlieghere
Subscribers: JDevlieghere, lldb-commits
Differential Revision: https://reviews.llvm.org/D51602
llvm-svn: 341940
Summary: An address breakpoint of the form "b 0x1000" won't resolve if it's created while the process isn't running. This patch deletes Address::SectionWasDeleted, renames Address::SectionWasDeletedPrivate to SectionWasDeleted (and makes it public), and changes the section check in Breakpoint::ModulesChanged back to its original form
Reviewers: jingham, #lldb
Reviewed By: jingham
Subscribers: davide, lldb-commits
Differential Revision: https://reviews.llvm.org/D51816
llvm-svn: 341849
Summary:
This patch adds an implementation of retrieving of declarations and declaration
contexts based on PDB symbols.
PDB has different type symbols for const-qualified types, and this
implementation ensures that only one declaration was created for both const
and non-const types, but creates different compiler types for them.
The implementation also processes the case when there are two symbols
corresponding to a variable. It's possible e.g. for class static variables,
they has one global symbol and one symbol belonging to a class.
PDB has no info about namespaces, so this implementation parses the full symbol
name and tries to figure out if the symbol belongs to namespace or not,
and then creates nested namespaces if necessary.
Reviewers: asmith, zturner, labath
Reviewed By: asmith
Subscribers: aleksandr.urakov, teemperor, lldb-commits, stella.stamenova
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D51162
llvm-svn: 341782
This recommits r341487, which was reverted due to failing tests with
clang. It turned out I had incorrectly expected that the literal arrays
passed to ArrayRef constructor will have static (permanent) storage.
This was only the case with gcc, while clang was constructing them on
stack, leading to dangling pointers when the function returns.
The fix is to explicitly assign static storage duration to the opcode
arrays.
llvm-svn: 341758
Summary:
Previously we SetUseColor(true) wrongly when output was not a terminal so it broken some (not public) bots.
Thanks for issue report, @stella.stamenova
Reviewers: stella.stamenova, zturner
Reviewed By: stella.stamenova
Subscribers: abidh, lldb-commits, stella.stamenova
Differential Revision: https://reviews.llvm.org/D51772
llvm-svn: 341746
In a subsequent commit, I will need to expose the search depth
to the SB API's, so I'm moving this define into lldb-enumerations
where it will get added to the lldb module.
llvm-svn: 341690
This showed up in an Ubsan build of lldb (inside the CFAbsoluteTime
data formatter). As we only care about the bit pattern, we just
round to the nearest double, and truncate to a size that fits
in ulonglong_t.
<rdar://problem/44229924>
llvm-svn: 341682
code. This will enable disassembly of the optional subset of
neon that some Cortex cores support. Add a unit test to check
that a few of these instructions disassemble as expected.
<rdar://problem/26674303>
llvm-svn: 341623
The GetLanguageForExpr has side effects, so we can't remove this
call without breaking the completion mechanism. However, we can
keep the change that gets rid of this unnecessary variable.
llvm-svn: 341535
reads an ObjectFileMachO's string table in one chunk. Originally
this was commented out because binaries in the system's shared cache
all share a mega-string table and so reading the entire mega-strtab
for each binary was a performance problem.
In the reinstated code, I add a check that the binary we're reading
from memory is not in the shared cache (there isn't a constant in
<mach-o/loader.h> for this bit yet; we hardcode the value in one
other place in ObjectFileMachO alread). For binaries that we're
reading out of memory that are NOT in the shared cache, reading
the string table in one chunk is a big performance improvement.
Also have debugserver send up the flags value for binaries in its
response to the jGetLoadedDynamicLibrariesInfos request.
NFC.
<rdar://problem/33604496>
llvm-svn: 341511
This patch allows LLDB to print column info in backtraces et al. if
available, which is useful when the backtrace contains a frame like
the following:
f(can_crash(0), can_crash(1));
Differential Revision: https://reviews.llvm.org/D51661
llvm-svn: 341506
return the opcode as a Expected<ArrayRef> instead of a
Status+pointer+size combo.
I also move the linux implementation to the base class, as the trap
opcodes are likely to be the same for all/most implementations of the
class (except the arm one, where linux chooses a different opcode than
what the arm spec recommends, which I keep linux-specific).
llvm-svn: 341487
Jason Molenda
