Summary:
When a call instruction is the last instruction in a function, the
backtrace PC will point past the end of the function. We already had
special code to handle that, but we did not handle the case where the PC
ends up outside of the bounds of the module containing the function,
which is a situation that occured in TestNoreturnUnwind on android for
some arch/compiler combinations.
I fix this by adding an argument to Address resolution code which states
that we are ok with addresses pointing to the end of a module/section to
resolve to that module/section.
I create a reproducible test case for this situation by hand-crafting an
executable which has a noreturn function at the end of a module.
Reviewers: jasonmolenda, jingham
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D32022
llvm-svn: 304976
This renames the LLDB error class to Status, as discussed
on the lldb-dev mailing list.
A change of this magnitude cannot easily be done without
find and replace, but that has potential to catch unwanted
occurrences of common strings such as "Error". Every effort
was made to find all the obvious things such as the word "Error"
appearing in a string, etc, but it's possible there are still
some lingering occurences left around. Hopefully nothing too
serious.
llvm-svn: 302872
All references to Host and Core have been removed, so this
class can now safely be lowered into Utility.
Differential Revision: https://reviews.llvm.org/D30559
llvm-svn: 296909
Also found/fixed one bug identified by this warning in
RenderScriptx86ABIFixups.cpp where a string literal was being used in an
effort to provide a name for an instruction/register, but was instead
being passed as the bool 'isVolatile' parameter.
llvm-svn: 291198
This is a large API change that removes the two functions from
StreamString that return a std::string& and a const std::string&,
and instead provide one function which returns a StringRef.
Direct access to the underlying buffer violates the concept of
a "stream" which is intended to provide forward only access,
and makes porting to llvm::raw_ostream more difficult in the
future.
Differential Revision: https://reviews.llvm.org/D26698
llvm-svn: 287152
*** to conform to clang-format’s LLVM style. This kind of mass change has
*** two obvious implications:
Firstly, merging this particular commit into a downstream fork may be a huge
effort. Alternatively, it may be worth merging all changes up to this commit,
performing the same reformatting operation locally, and then discarding the
merge for this particular commit. The commands used to accomplish this
reformatting were as follows (with current working directory as the root of
the repository):
find . \( -iname "*.c" -or -iname "*.cpp" -or -iname "*.h" -or -iname "*.mm" \) -exec clang-format -i {} +
find . -iname "*.py" -exec autopep8 --in-place --aggressive --aggressive {} + ;
The version of clang-format used was 3.9.0, and autopep8 was 1.2.4.
Secondly, “blame” style tools will generally point to this commit instead of
a meaningful prior commit. There are alternatives available that will attempt
to look through this change and find the appropriate prior commit. YMMV.
llvm-svn: 280751
for TestNamespaceLookup.py; didn't see anything obviously wrong so I'll
need to look at this more closely before re-committing. (passed OK on
macOS ;)
llvm-svn: 273531
There's uses of "macosx" that will be more tricky to
change, like in triples (e.g. "x86_64-apple-macosx10.11") -
for now I'm just updating source comments and strings printed
for humans.
llvm-svn: 273524
values for the pc or return address register.
On ios with arm64 and a binary that has multiple functions without
individual symbol boundaries, we end up with an assembly profile
unwind plan that says lr=<same> - that is, the link register contents
are unmodified from the caller's value. This gets the unwinder in
a loop.
When we're off the 0th frame, we never want to look to a caller for
a pc or return-address register value.
Add checks to ReadGPRValue and ReadRegister to prevent both the pc
and ra register values from recursing.
If this causes problems with backtraces on android, let me know or
back it out and I'll look into it -- but I think these are
straightforward and don't expect problems.
<rdar://problem/24610365>
llvm-svn: 270162
RegisterContextLLDB::InitializeNonZerothFrame already has code to attempt
to detect and handle the case where the PC points beyond the end of a
function, but there are certain cases where this doesn't work correctly.
In fact, there are *two* different places where this detection is attempted,
and the failure is in fact a result of an unfortunate interaction between
those two separate attempts.
First, the ResolveSymbolContextForAddress routine is called with the
resolve_tail_call_address flag set to true. This causes the routine
to internally accept a PC pointing beyond the end of a function, and
still resolving the PC to that function symbol.
Second, the InitializeNonZerothFrame routine itself maintains a
"decr_pc_and_recompute_addr_range" flag and, if that turns out to
be true, itself decrements the PC by one and searches again for
a symbol at that new PC value.
Both approaches correctly identify the symbol associated with the PC.
However, the problem is now that later on, we also need to find the
DWARF CFI record associated with the PC. This is done in the
RegisterContextLLDB::GetFullUnwindPlanForFrame routine, and uses
the "m_current_offset_backed_up_one" member variable.
However, that variable only actually contains the PC "backed up by
one" if the *second* approach above was taken. If the function was
already identified via the first approach above, that member variable
is *not* backed up by one but simply points to the original PC.
This in turn causes GetEHFrameUnwindPlan to not correctly identify
the DWARF CFI record associated with the PC.
Now, in many cases, if the first method had to back up the PC by one,
we *still* use the second method too, because of this piece of code:
// Or if we're in the middle of the stack (and not "above" an asynchronous event like sigtramp),
// and our "current" pc is the start of a function...
if (m_sym_ctx_valid
&& GetNextFrame()->m_frame_type != eTrapHandlerFrame
&& GetNextFrame()->m_frame_type != eDebuggerFrame
&& addr_range.GetBaseAddress().IsValid()
&& addr_range.GetBaseAddress().GetSection() == m_current_pc.GetSection()
&& addr_range.GetBaseAddress().GetOffset() == m_current_pc.GetOffset())
{
decr_pc_and_recompute_addr_range = true;
}
In many cases, when the PC is one beyond the end of the current function,
it will indeed then be exactly at the start of the next function. But this
is not always the case, e.g. if there happens to be alignment padding
between the end of one function and the start of the next.
In those cases, we may sucessfully look up the function symbol via
ResolveSymbolContextForAddress, but *not* set decr_pc_and_recompute_addr_range,
and therefore fail to find the correct DWARF CFI record.
A very simple fix for this problem is to just never use the first method.
Call ResolveSymbolContextForAddress with resolve_tail_call_address set
to false, which will cause it to fail if the PC is beyond the end of
the current function; or else, identify the next function if the PC
is also at the start of the next function. In either case, we will
then set the decr_pc_and_recompute_addr_range variable and back up the
PC anyway, but this time also find the correct DWARF CFI.
A related problem is that the ResolveSymbolContextForAddress sometimes
returns a "symbol" with empty name. This turns out to be an ELF section
symbol. Now, usually those get type eSymbolTypeInvalid. However, there
is code in ObjectFileELF::ParseSymbols that tries to change the type of
invalid symbols to eSymbolTypeCode or eSymbolTypeData if the symbol
lies within the code or data section.
Unfortunately, this check also hits the symbol for the code section
itself, which is then marked as eSymbolTypeCode. While the size of
the section symbol is 0 according to the ELF file, LLDB considers
this size invalid and attempts to figure out the "correct" size.
Depending on how this goes, we may end up with a symbol that overlays
part of the code section, even outside areas covered by real function
symbols.
Therefore, if we call ResolveSymbolContextForAddress with PC pointing
beyond the end of a function, we may get this bogus section symbol.
This again means InitializeNonZerothFrame thinks we have a valid PC,
but then we don't find any unwind info for it.
The fix for this problem is me to simply always leave ELF section
symbols as type eSymbolTypeInvalid.
Differential Revision: http://reviews.llvm.org/D18975
llvm-svn: 267363
If the UnwindPlan did not identify how to unwind the stack pointer
register, LLDB currently assumes it can determine to caller's SP
from the current frame's CFA. This is true on most platforms
where CFA is by definition equal to the incoming SP at function
entry.
However, on the s390x target, we instead define the CFA to equal
the incoming SP plus an offset of 160 bytes. This is because
our ABI defines that the caller has to provide a register save
area of size 160 bytes. This area is allocated by the caller,
but is considered part of the callee's stack frame, and therefore
the CFA is defined as pointing to the top of this area.
In order to make this work on s390x, this patch introduces a new
ABI callback GetFallbackRegisterLocation that provides platform-
specific fallback register locations for unwinding. The existing
code to handle SP unwinding as well as volatile registers is moved
into the default implementation of that ABI callback, to allow
targets where that implementation is incorrect to override it.
This patch in itself is a no-op for all existing platforms.
But it is a pre-requisite for adding s390x support.
Differential Revision: http://reviews.llvm.org/D18977
llvm-svn: 266307
Additionally fix the type of some dwarf expression where we had a
confusion between scalar and load address types after a dereference.
Differential revision: http://reviews.llvm.org/D17604
llvm-svn: 262014
Summary:
The testcase TestNoreturnUnwind.py was failing
because the unwind from the vdso library was not
successful for clang compiler while it was passing
for gcc. It was passing for gcc since the unwind plan
used was the assembly plan and the ebp register was
set by the main function in case of gcc and was not
used by the functions in the call flow to the vdso, whereas
clang did not emit assembly prologue for main and so
the assembly unwind was failing. Normally in case of
failure of assembly unwind, lldb switches to EH CFI frame
based unwinding, but this was not happening for
the first frame. This patch tries to fix this behaviour by
falling to EH CFI frame based unwinding in case of assembly
unwind failure even for the first frame.
The test is still marked as XFAIL since it relys on the fix
of another bug.
Reviewers: lldb-commits, jingham, zturner, tberghammer, jasonmolenda
Subscribers: jasonmolenda
Differential Revision: http://reviews.llvm.org/D15046
llvm-svn: 257465
* Use .ARM.exidx as a fallback unwind plan for non-call site when the
instruction emulation based unwind failed.
* Work around an old compiler issue where the compiler isn't sort the
entries in .ARM.exidx based on their address.
* Fix unwind info parsing when the virtual file address >= 0x80000000
* Fix bug in unwind info parsing when neither lr nor pc is explicitly
restored.
Differential revision: http://reviews.llvm.org/D13380
llvm-svn: 249119
.ARM.exidx/.ARM.extab sections contain unwind information used on ARM
architecture from unwinding from an exception.
Differential revision: http://reviews.llvm.org/D13245
llvm-svn: 248903
In some special case (e.g. signal handlers, hand written assembly) it is
valid to have 2 stack frame with the same CFA value. This CL change the
looping stack detection code to report a loop only if at least 3
consecutive frames have the same CFA.
Differential revision: http://reviews.llvm.org/D12699
llvm-svn: 247133
Previously if the instruction emulation based unwind plan failed then
we fall back to the arch default unwind plan. Change it to fall back
to the eh_frame based one even on non call sites if we have eh_frame
as that one tend to be more reliable.
Differential revision: http://reviews.llvm.org/D10902
llvm-svn: 241334
* Add and fix the emulation of several instruction.
* Disable frame pointer usage on Android.
* Specify return address register for the unwind plan instead of explict
tracking the value of RA.
* Replace prologue detection heuristics (unreliable in several cases)
with a logic to follow the branch instructions and restore the CFI
value based on them. The target address for a branch should have the
same CFI as the source address (if they are in the same function).
* Handle symbols in ELF files where the symbol size is not specified
with calcualting their size based on the next symbol (already done
in MachO files).
* Fix architecture in FuncUnwinders with filling up the inforamtion
missing from the object file with the architecture of the target.
* Add code to read register wehn the value is set to "IsSame" as it
meanse the value of a register in the parent frame is the same as the
value in the current frame.
Differential revision: http://reviews.llvm.org/D10447
llvm-svn: 240533
Summary:
This patch enables evaluation of DWARF expressions setting the CFA during stack unwinding.
This makes TestSigtrampUnwind "almost" pass on linux. I am not enabling the test yet since the
symbol name for the signal trampoline does not get resolved properly due to a different bug, but
apart from that, the backtrace is sane.
I am unsure how this change affects Mac. I think it makes the unwinder prefer the DWARF unwind
plan instead of some custom platform-dependant plan. However, it does not affect the end result
- the stack unwinding works as expected.
Reviewers: jasonmolenda
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D7792
llvm-svn: 230211
Summary:
This change refactors UnwindPlan::Row to be able to store the fact that the CFA is value is set
by evaluating a dwarf expression (DW_CFA_def_cfa_expression). This is achieved by creating a new
class CFAValue and moving all CFA setting/getting code there. Note that code using the new
CFAValue::isDWARFExpression is not yet present and will be added in a follow-up patch. Therefore,
this patch should not change the functionality in any way.
Test Plan: Ran tests on Mac and Linux. No regressions detected.
Reviewers: jasonmolenda, clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D7755
llvm-svn: 230210
introduced subtle bugs in two places in
RegisterContextLLDB::GetFullUnwindPlanForFrame where
it specifically wanted to get an eh_frame unwind plan
and was using "Get CallSite UnwindPlan" as synonymous
with that. But now we have two different types of
unwind plan that can be returned in that case, and
compact unwind won't behaves as needed.
<rdar://problem/19528559>
llvm-svn: 226631
so that we will use the UnwindPlan's rule for providing the stack
pointer BEFORE we use the trick of using the callee's CFA address
as the stack pointer. When we're in a _sigtramp frame, the CFA of
the _sigtramp stack frame is not the same as the stack pointer value
when the async interrupt occurred -- we need to use the eh_frame
rules for retrieving the correct value.
<rdar://problem/18913548>
llvm-svn: 225427
a register value that is live in the stack frame 0 register context.
Fixes a problem where retrieving a register value on stack frame #n
would involved O(n!) stack frame checks. This could be very slow on
a deep stack when retrieving register values that had not been
modified/saved by any of the stack frames. Not common, but annoying
when it was hit.
<rdar://problem/19010211>
llvm-svn: 223843
section for x86_64 and i386 targets on Darwin systems. Currently only the
compact unwind encoding for normal frame-using functions is supported but it
will be easy handle frameless functions when I have a bit more free time to
test it. The LSDA and personality routines for functions are also retrieved
correctly for functions from the compact unwind section.
This new code is very fresh -- it passes the lldb testsuite and I've done
by-hand inspection of many functions and am getting correct behavior for all
of them. There may need to be some bug fixing over the next couple weeks as
I exercise and test it further. But I think it's fine right now so I'm
committing it.
<rdar://problem/13220837>
llvm-svn: 223625
UnwindLLDB::AddOneMoreFrame to try the fallback unwind plan on
that same stack frame before it tries the fallback unwind plan
on the "next" or callee frame.
In RegisterContextLLDB::TryFallbackUnwindPlan, when we're
trying the fallback unwind plan to see if it is valid, make
sure we change all of the object ivars that might be used in
the process of fetching the CFA & caller's saved pc value
and restore those if we decide not to use the fallback
unwindplan.
<rdar://problem/19035079>
llvm-svn: 222601
where it is retrieving the Return Address register contents
on a target where that's a thing. If we fail to get a valid
RA, we force a switch to the fallback unwind plan. This patch
adds a sanity check for that fallback unwind plan -- it must
get a valid CFA for this frame in addition to being able to
retrieve the caller's PC -- and it correctly marks the unwind
rules as failing if the fallback unwind plan fails.
<rdar://problem/19010211>
llvm-svn: 222301
RegisterContextLLDB. I have core files of half a dozen tricky
unwind situations on x86/arm and they're all working pretty much
correctly at this point, but we'll need to keep an eye out for
unwinder regressions for a little while; it's tricky to get these
heuristics completely correct in all unwind situations.
<rdar://problem/18937193>
llvm-svn: 221866
After r221575 TestCallStopAndContinue and TestCallThatRestarts started
crashing on FreeBSD with a null temporary_module_sp in
RegisterContextLLDB::InitializeNonZerothFrame().
llvm-svn: 221805
The addition of RegisterNumber introduced a bug where if the PC is stored in a
return address register, such as on ARM and PowerPC, this register number is
retrieved and used, but never checked in the row if it's saved. Correct this by
setting the variable that's used to the new register number.
Patch by Jason Molenda.
llvm-svn: 221790
Summary:
PowerPC handles the stack chain with the current stack pointer being a pointer
to the backchain (CFA). LLDB currently has no way of handling this, so this
adds a "CFA is dereferenced from a register" type.
Discussed with Jason Molenda, who also provided the initial patch for this.
Reviewers: jasonmolenda
Reviewed By: jasonmolenda
Subscribers: emaste, lldb-commits
Differential Revision: http://reviews.llvm.org/D6182
llvm-svn: 221788
it in RegisterContext.cpp.
There's a lot of bookkeeping code in RegisterContextLLDB where it has
to convert between different register numbering schemes and it makes
some methods like SavedLocationForRegister very hard to read or
maintain. Abstract all of the details about different register numbering
systems for a given register into this new class to make it easier
to understand what the method is doing.
Also add register name printing to all of the logging -- that's easy to
get now that I've got an object to represent the register numbers.
There were some gnarly corner cases of this method that I believe
I've translated correctly - initial testing looks good but it's
possible I missed a corner case, especially with architectures which
uses a link-register aka return address register like arm32/arm64.
Basic behavior is correct but there are a lot of corner casese that are
handled in this method ...
llvm-svn: 221577
If a noreturn function was the last function in a section,
we wouldn't correctly back up the saved-pc value into the
correct section leading to us showing the wrong function in
the backtrace.
Also add a backtrace test with an attempt to elicit this
particular layout. It happens to work out with clang -Os
but other compilers may not quite get the same layout I'm
getting at that opt setting. We'll still be exercising the
basic noreturn handling in the unwinder even if we don't get
one function at the very end of a section.
<rdar://problem/16051613>
llvm-svn: 221575
let's let lldb try the arch default unwind every time but not destructively --
it doesn't permanently replace the main unwind method for that function from
now on.
This fix is for <rdar://problem/18683658>.
I tested it against Ryan Brown's go program test case and also a
collection of core files of tricky unwind scenarios
<rdar://problem/15664282> <rdar://problem/15835846>
<rdar://problem/15982682> <rdar://problem/16099440>
<rdar://problem/17364005> <rdar://problem/18556719>
that I've fixed over the last 6-9 months.
llvm-svn: 221238
is "invalid" -- it is past the end of the stack trace. Add a new
method IsCompletedStackWalk() so we can tell if an invalid stack
frame is from a complete backtrace or if it might be worth re-trying
the last unwind with a different method.
This fixes the unwinder problems Ryan Brown was having with go
programs. The unwinder can (under the right circumstances) still
destructively replace unwind plans permanently - I'll work on
that in a different patch.
<rdar://problem/18683658>
llvm-svn: 221229
in GetFullUnwindPlanForFrame() - the code was mostly checking
that we had an active Process and ABI but not always.
clang static analyzer fixit.
llvm-svn: 219772
the backtrace, try falling back to the architecture default
unwind plan and see if we can backtrace a little further.
<rdar://problem/18556719>
llvm-svn: 219247
detct unwind loops but there was a code path through there (using
architecture default unwind plans) that didn't do the check, and
could end up with an infinite loop unwind. Move that code into a
separate method and call it from both places where it is needed.
Also remove the use of ABI::FunctionCallsChangeCFA in that check.
I thought about it a lot and none of the architecutres that we're
supporting today can have a looping CFA.
Since the unwinder isn't using ABI::FunctionCallsChangeCFA() and
ABI::StackUsesFrames(), and the unwinder was the only reason
those methods exists, I removed them from the ABI and all its
plugins.
<rdar://problem/17364005>
llvm-svn: 216992
We decided to use assmbly profiler instead of eh_frame for frame 0 because for compiler generated code, eh_frame is usually synchronous(a.k.a. only valid at call site); and we have no way to tell if it's asynchronous or not.
But for x86 & x86_64 compiler generated code:
1. clang & GCC describes all prologue instructions in eh_frame;
2. mid-function stack pointer altering instructions can be easily detected.
So we can grab eh_frame, and use assembly profiler to augment it into asynchronous unwind table.
This change also benefits hand-written assembly; eh_frame for hand-written assembly is often asynchronous,so we have a much better chance to successfully unwind through them.
Change by Tong Shen.
llvm-svn: 216406
(lldb) b puts
(lldb) expr -g -i0 -- (int)puts("hello")
First we will stop at the entry point of the expression before it runs, then we can step over a few times and hit the breakpoint in "puts", then we can continue and finishing stepping and fininsh the expression.
Main features:
- New ObjectFileJIT class that can be easily created for JIT functions
- debug info can now be enabled when parsing expressions
- source for any function that is run throught the JIT is now saved in LLDB process specific temp directory and cleaned up on exit
- "expr -g --" allows you to single step through your expression function with source code
<rdar://problem/16382881>
llvm-svn: 204682
Jason Molenda