Summary:
Windows unwinding is weird. The unwind rules do not (always) describe
the precise layout of the stack, but rather expect the debugger to scan
the stack for something which looks like a plausible return address, and
the unwind based on that. The reason this works somewhat reliably is
because the the unwinder also has access to the frame sizes of the
functions on the stack. This allows it (in most cases) to skip function
pointers in local variables or function arguments, which could otherwise
be mistaken for return addresses.
Implementing this kind of unwind mechanism in lldb was a bit challenging
because we expect to be able to statically describe (in the UnwindPlan)
structure, the layout of the stack for any given instruction. Giving a
precise desription of this is not possible, because it requires
correlating information from two functions -- the pushed arguments to a
function are considered a part of the callers stack frame, and their
size needs to be considered when unwinding the caller, but they are only
present in the unwind entry of the callee. The callee may end up being
in a completely different module, or it may not even be possible to
determine it statically (indirect calls).
This patch implements this functionality by introducing a couple of new
APIs:
SymbolFile::GetParameterStackSize - return the amount of stack space
taken up by parameters of this function.
SymbolFile::GetOwnFrameSize - the size of this function's frame. This
excludes the parameters, but includes stuff like local variables and
spilled registers.
These functions are then used by the unwinder to compute the estimated
location of the return address. This address is not always exact,
because the stack may contain some additional values -- for instance, if
we're getting ready to call a function then the stack will also contain
partially set up arguments, but we will not know their size because we
haven't called the function yet. For this reason the unwinder will crawl
up the stack from the return address position, and look for something
that looks like a possible return address. Currently, we assume that
something is a valid return address if it ends up pointing to an
executable section.
All of this logic kicks in when the UnwindPlan sets the value of CFA as
"isHeuristicallyDetected", which is also the final new API here. Right
now, only SymbolFileBreakpad implements these APIs, but in the future
SymbolFilePDB will use them too.
Differential Revision: https://reviews.llvm.org/D66638
llvm-svn: 373072
Summary:
This patch makes it possible to unwind via breakpad STACK WIN records.
It is "basic" because two important features are missing:
- support for the .raSearch keyword
- support for multiple STACK WIN records within a single function
Right now, we just reject the .raSearch records, and always pick the
first record for the whole function
SymbolFileBreakpad, and so I think it can serve as a good example of
what is needed of the symbol file and unwinding machinery to make this
work.
However, it is already useful for unwinding in some situations, and it
sets up the general framework for the parsing of these kinds of records,
which reduces the size of the followup patches implementing the two
other components.
Reviewers: amccarth, rnk, markmentovai
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D67067
llvm-svn: 371017
Summary:
Previous patch (r360409) introduced the "symbol file unwind plan"
concept, but that plan wasn't used for unwinding yet. With this patch,
we start to consider the new plan as a possible strategy for both
synchronous and asynchronous unwinding. I also add a test that asserts
that unwinding via breakpad STACK CFI info works end-to-end.
Reviewers: jasonmolenda, clayborg
Subscribers: lldb-commits, amccarth, markmentovai
Differential Revision: https://reviews.llvm.org/D61853
llvm-svn: 361618
Summary:
This patch implements the GetUnwindPlan interface (added in the previous
patch) for SymbolFileBreakpad, and uses it to generate unwind plans from
STACK CFI records in breakpad files.
We first perform a light-weight parse of the breakpad in order to build
up a map of regions covered by the unwind info so that we can later jump
to the right record when we need to unwind a specific function.
The actual parsing is relatively straight-forward, as the STACK CFI records
are just another (text) form of the eh_frame unwind instructions, and
the same goes for lldb's UnwindPlans. The newly-introduced
PostfixExpression API is used to convert the breakpad postfix
expressions into DWARF. The generated dwarf expressions are stored in a
BumpPtrAllocator, as the UnwindPlan does not take ownership of the
expression data it references (usually this is static data in an object
file, so special ownership is needed).
At this moment the generated unwind plans aren't used in the actual
unwind machinery (only in the image show-unwind command), but that is
coming in a separate patch.
Reviewers: amccarth, clayborg, markmentovai
Subscribers: aprantl, jasonmolenda, lldb-commits
Differential Revision: https://reviews.llvm.org/D61733
llvm-svn: 360574
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
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 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
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
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:
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
Pavel Labath