Optimal edge profiling is only possible when blocks with no predecessors get an
virtual edge (BB,0) that counts the execution frequencies of this
function-exiting blocks.
This patch makes the necessary changes before actually enabling optimal edge profiling.
llvm-svn: 80667
This adds a pass to verify the current profile against the flow conditions.
This is very helpful when later on trying to perserve the profiling information
during all passes.
llvm-svn: 80666
for sanity. This didn't turn up any bugs.
Change CallGraphNode to maintain its "callsite" information in the
call edges list as a WeakVH instead of as an instruction*. This fixes
a broad class of dangling pointer bugs, and makes CallGraph have a number
of useful invariants again. This fixes the class of problem indicated
by PR4029 and PR3601.
llvm-svn: 80663
makes an eggregious hack somewhat more palatable. Bringing the LSDA forward
and making it a GV available for reference would be even better, but is
beyond the scope of what I'm looking to solve at this point.
Objective C++ code could generate function names that broke the previous
scheme. This fixes that.
llvm-svn: 80649
SCEVUnknowns, as the non-SCEVUnknown cases in the getSCEVAtScope code
can also end up repeatedly climing through the same expression trees,
which can be unusably slow when the trees are very tall.
Also, add a quick check for SCEV pointer equality to the main
SCEV comparison routine, as the full comparison code can be expensive
in the case of large expression trees.
These fix compile-time problems in some pathlogical cases.
llvm-svn: 80623
stem from the fact that we have two types of passes that need to update it:
1. callgraphscc and module passes that are explicitly aware of it
2. Functionpasses (and loop passes etc) that are interlaced with CGSCC passes
by the CGSCC Passmgr.
In the case of #1, we can reasonably expect the passes to update the call
graph just like any analysis. However, functionpasses are not and generally
should not be CG aware. This has caused us no end of problems, so this takes
a new approach. Logically, the CGSCC Pass manager can rescan every function
after it runs a function pass over it to see if the functionpass made any
updates to the IR that affect the callgraph. This allows it to catch new calls
introduced by the functionpass.
In practice, doing this would be slow. This implementation keeps track of
whether or not the current scc is dirtied by a function pass, and, if so,
delays updating the callgraph until it is actually needed again. This was
we avoid extraneous rescans, but we still have good invariants when the
callgraph is needed.
Step #2 of the "give Callgraph some sane invariants" is to change CallGraphNode
to use a CallBackVH for the callsite entry of the CallGraphNode. This way
we can immediately remove entries from the callgraph when a FunctionPass is
active instead of having dangling pointers. The current pass tries to tolerate
these dangling pointers, but it is just an evil hack.
This is related to PR3601/4835/4029. This also reverts r80541, a hack working
around the sad lack of invariants.
llvm-svn: 80566
indirect function pointer, inline it, then go to delete the body.
The problem is that the callgraph had other references to the function,
though the inliner had no way to know it, so we got a dangling pointer
and an invalid iterator out of the deal.
The fix to this is pretty simple: stop the inliner from deleting the
function by knowing that there are references to it. Do this by making
CallGraphNodes contain a refcount. This requires moving deletion of
available_externally functions to the module-level cleanup sweep where
it belongs.
llvm-svn: 80533
argpromotion and structretpromote. Basically, when replacing
a function, they used the 'changeFunction' api which changes
the entry in the function map (and steals/reuses the callgraph
node).
This has some interesting effects: first, the problem is that it doesn't
update the "callee" edges in any callees of the function in the call graph.
Second, this covers for a major problem in all the CGSCC pass stuff, which
is that it is completely broken when functions are deleted if they *don't*
reuse a CGN. (there is a cute little fixme about this though :).
This patch changes the protocol that CGSCC passes must obey: now the CGSCC
pass manager copies the SCC and preincrements its iterator to avoid passes
invalidating it. This allows CGSCC passes to mutate the current SCC. However
multiple passes may be run on that SCC, so if passes do this, they are now
required to *update* the SCC to be current when they return.
Other less interesting parts of this patch are that it makes passes update
the CG more directly, eliminates changeFunction, and requires clients of
replaceCallSite to specify the new callee CGN if they are changing it.
llvm-svn: 80527
encodings.
- Make some of the values emitted by the FDEs dependent upon the pointer
size. This is in line with how GCC does things. And it has the benefit of
working for Darwin in 64-bit mode now.
llvm-svn: 80428
- I'm still trying to figure out the cleanest way to implement this and match the assembler, currently there are some substantial differences.
llvm-svn: 80347
calls into a function and if the calls bring in arrays, try to merge
them together to reduce stack size. For example, in the testcase
we'd previously end up with 4 allocas, now we end up with 2 allocas.
As described in the comments, this is not really the ideal solution
to this problem, but it is surprisingly effective. For example, on
176.gcc, we end up eliminating 67 arrays at "gccas" time and another
24 at "llvm-ld" time.
One piece of concern that I didn't look into: at -O0 -g with
forced inlining this will almost certainly result in worse debug
info. I think this is acceptable though given that this is a case
of "debugging optimized code", and we don't want debug info to
prevent the optimizer from doing things anyway.
llvm-svn: 80215
A include/llvm/ADT/iterator.cmake
U autoconf/configure.ac
--- Reverse-merging r80161 into '.':
U cmake/config-ix.cmake
--- Reverse-merging r80171 into '.':
U Makefile
--- Reverse-merging r80173 into '.':
U configure
U include/llvm/Config/config.h.in
--- Reverse-merging r80180 into '.':
A include/llvm/ADT/iterator.h.in
Despite common miscomceptions, iterator.h is alive and well. It broke the build
bots for several hours. And yet no one bothered to look at them.
Gabor and Doug, please review your changes and make sure that they actually
build before resubmitting them.
llvm-svn: 80197
- I moved section creation back into AsmParser. I think policy decisions like
this should be pushed higher, not lower, when possible (in addition the
assembler has flags which change this behavior, for example).
llvm-svn: 80162
so get rid of this monstrosity. iterator.h.in is scheduled for deletion in my working copy,
but I wait till I see that configure gets regenerated, as it depends on it. I'll commit
then.
There are still some AC_* tests in the configure.ac dealing with iterators, those can
be zapped probably too.
llvm-svn: 80147
This is a simple AliasAnalysis implementation which works by making
ScalarEvolution queries. ScalarEvolution has a more complete understanding
of arithmetic than BasicAA's collection of ad-hoc checks, so it handles
some cases that BasicAA misses, for example p[i] and p[i+1] within the
same iteration of a loop.
This is currently experimental. It may be that the main use for this pass
will be to help find cases where BasicAA can be profitably extended, or
to help in the development of the overall AliasAnalysis infrastructure,
however it's also possible that it could grow up to become a directly
useful pass.
llvm-svn: 80098
- I haven't really tried to find the "right" way to store the fixups or apply
them, yet. This works, but isn't particularly elegant or fast.
- Still no evaluation support, so we don't actually ever not turn a fixup into
a relocation entry.
llvm-svn: 80089
code hints that it would be a good idea to inline
a function ("inline" keyword). No functional change
yet; FEs do not emit this and inliner does not use it.
llvm-svn: 80063
and introduce a new Instruction::isIdenticalTo which tests for full
identity, including the SubclassOptionalData flags. Also, fix the
Instruction::clone implementations to preserve the SubclassOptionalData
flags. Finally, teach several optimizations how to handle
SubclassOptionalData correctly, given these changes.
This fixes the counterintuitive behavior of isIdenticalTo not comparing
the full value, and clone not returning an identical clone, as well as
some subtle bugs that could be caused by these.
Thanks to Nick Lewycky for reporting this, and for an initial patch!
llvm-svn: 80038
This time there is no additional include of llvm/Config/config.h
at all. Instead I use a hard-coded preprecessor symbol:
LLVM_COMPACTIFY_SENTINELS
(should this work on the self-hosting buildbot, then
cleanups come next)
llvm-svn: 80035
should be forced to 32-bits (.long) even on 64-bit architectures. Darwin wants
these bits to be 64-bits (.quad). However, other platforms may disagree.
This is just the info right now and is part of a work-in-progress which needs
this. We'll add the actual *use* of this soon.
llvm-svn: 80024
This is conventional command-line tool behavior. -f now just means
"enable binary output on terminals".
Add a -f option to llvm-extract and llvm-link, for consistency.
Remove F_Force from raw_fd_ostream and enable overwriting and
truncating by default. Introduce an F_Excl flag to permit users to
enable a failure when the file already exists. This flag is
currently unused.
Update Makefiles and documentation accordingly.
llvm-svn: 79990
U include/llvm/BasicBlock.h
U include/llvm/ADT/ilist_node.h
U include/llvm/ADT/ilist.h
U include/llvm/CodeGen/SelectionDAG.h
U include/llvm/CodeGen/MachineFunction.h
U include/llvm/CodeGen/MachineBasicBlock.h
U include/llvm/Function.h
Revert r79938. It was causing self-hosting build failures.
llvm-svn: 79960
*) introducing new data type and export function of edge info for whole function (preparation for next patch).
*) renaming variables to make clear distinction between data and containers that contain this data.
*) updated comments and whitespaces.
*) made ProfileInfo::MissingValue a double (as it should be...).
(Discussed at http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090817/084955.html.)
llvm-svn: 79940
reduce the size of relevant "ghostly" sentinels
by a pointer.
This attempt now makes the compactification dependent
on the configure variable LLVM_COMPACT_SENTINELS
and should not cause any bootstrap failures for
llvm-gcc any more.
Please note that this is not yet the final version,
and (as settled with Chris) I shall take out the
autofoo/cmake portions in the next days.
This will also lose the assertability on sentinel
dereferencing and operator++, but that seems
an acceptable price to pay for the simplified
build logic.
llvm-svn: 79938
- The indirect table itself isn't being filled in yet.
- This isn't factored properly and is rather FIXMEd, but at the moment I'm more
focused on figuring out what it needs to do.
llvm-svn: 79910
- This is mostly complete, the main thing missing is .indirect_symbol support
(which would be straight-forward, except that the way it is implemented in
'as' makes getting an exact .o match interesting).
llvm-svn: 79899
means that raw_ostream no longer has to #include <iosfwd>. Nothing in llvm
should use raw_os_ostream.h, but llvm-gcc and some unit tests do.
llvm-svn: 79886
- I did not actually remove the configure test itself, someone who can
regenerate configure should do this (see FIXME in autoconf/configure.ac)
llvm-svn: 79881
forcing them down into various .cpp files.
This change also:
1. Renames TimeValue::toString() and Path::toString() to ::str()
for similarity with the STL.
2. Removes all stream insertion support for sys::Path, forcing
clients to call .str().
3. Removes a use of Config/alloca.h from bugpoint, using smallvector
instead.
4. Weans llvm-db off <iostream>
sys::Path really needs to be gutted, but I don't have the desire to
do it at this point.
llvm-svn: 79869
- This also shortens the Format.h implementation, and uses the print buffer
fully (it was wasting a character).
- This manifested as llvm-test failures, because one side effect was that
raw_ostream would write garbage '\x00' values into the output stream if it
happened that the string was at the end of the buffer. This meant that grep
would report 'Binary file matches', which meant the silly pattern matching
llvm-test eventually does would fail. Cute. :)
llvm-svn: 79862
MachineInstr and MachineOperand. This required eliminating a
bunch of stuff that was using DOUT, I hope that bill doesn't
mind me stealing his fun. ;-)
llvm-svn: 79813
instead of as two bools. Use this to add a F_Append flag
which has the obvious behavior.
Other unrelated changes conflated into this patch:
1. REmove EH stuff from llvm-dis and llvm-as, the try blocks
are dead.
2. Simplify the filename inference code in llvm-as/llvm-dis,
because raw_fd_ostream does the right thing with '-'.
3. Switch machine verifier to use raw_ostream instead of ostream
(Which is the thing that needed append in the first place).
llvm-svn: 79807
- Honor .globl.
- Set symbol type and section correctly ('nm' now works), and order symbols
appropriately.
- Take care to the string table so that the .o matches 'as' exactly (for ease
of testing).
llvm-svn: 79740
- The only .s syntax this honors right now is emitting labels, and some parts
of the symbol table generation are wrong or faked.
- This is enough to get nm to report such symbols... incorrectly, but still.
Also, fixed byte emission to extend the previous fragment if possible.
llvm-svn: 79739
(external was really undefined and there wasn't an explicit representation for
absolute symbols).
- This still needs some cleanup to how the absolute "pseudo" section is dealt
with, but I haven't figured out the nicest approach yet.
llvm-svn: 79733
bytes. libgcc doesn't seem to mind, but if you pass this DWARF to GDB, it
doesn't like it.
Also make the JIT memory manager to initialize it's memory to garbage in debug
mode, so that it's easier to find bugs like these in the future.
llvm-svn: 79674
- Together these form the (Mach-O) back end of the assembler.
- MCAssembler is the actual assembler backend, which is designed to have a
reasonable API. This will eventually grow to support multiple object file
implementations, but for now its Mach-O/i386 only.
- MCMachOStreamer adapts the MCStreamer "actions" API to the MCAssembler API,
e.g. converting the various directives into fragments, managing state like
the current section, and so on.
- llvm-mc will use the new backend via '-filetype=obj', which may eventually
be, but is not yet, since I hear that people like assemblers which actually
assemble.
- The only thing that works at the moment is changing sections. For the time
being I have a Python Mach-O dumping tool in test/scripts so this stuff can
be easily tested, eventually I expect to replace this with a real LLVM tool.
- More doxyments to come.
I assume that since this stuff doesn't touch any of the things which are part of
2.6 that it is ok to put this in not so long before the freeze, but if someone
objects let me know, I can pull it.
llvm-svn: 79612
conversion code do we really need?
- S.append_uint(N) can be replaced with 'raw_svector_ostream(S) << N' which is
somewhat slower due to the extra set up cost, but still plenty fast
(especially if the svector set up cost can be amortized).
llvm-svn: 79450
right.
- This class turns out to be much more convenient to use if we do this; clients
can make sure the buffer is always big enough if they care (since our current
idiom tends to be to use a SmallString<256> for the input to this we should
generally be avoiding an unnecessary malloc).
Also, add a convenience raw_svector_ostream::str method which flushes the buffer
and returns a StringRef for the vector contents.
llvm-svn: 79446
- This avoids unnecessary malloc/free overhead in the common case, and
unnecessary copying from the ostream buffer into the output vector.
llvm-svn: 79434
- These allow clients to make use of the extra elements in the vector which
have already been allocated, without requiring them to be value initialized.
llvm-svn: 79433
gcc-4.4 was optimizing away comparisons against SimpleValueType when
it was compared to a value larger than the largest value in the enum.
This patch works around it by adding one extra item to the enum so
that these tests will now be valid.
llvm-svn: 79401
- Kill off begin(), end(), and iterator. It isn't clear what these
mean. Instead provide getBufferStart(), which can be used with
GetNumBytesInBuffer to the same effect.
- Update ComputeColumn to take arguments for the buffer to scan, this
simplifies the implementation of write_impl substantially.
- This should also fix possible problems with the scanning pointer pointing
outside of the current raw_ostream buffer.
llvm-svn: 79379
- Drop the Candidates argument and fix all callers. Now that RegScavenger
tracks available registers accurately, there is no need to restict the
search.
- Make sure that no aliases of the found register are in use. This was a potential bug.
llvm-svn: 79369
TargetData is not present. It still uses TargetData when available.
This generalization also fixed some limitations in the TargetData
case; the attached testcase covers this.
llvm-svn: 79344
MCAsmStreamer. Based on this, eliminate the current section from AsmPrinter.
While I'm at it, clean up the last of the horrible "switch to null section" stuff
and add an assert. This change is in preparation for completely eliminating
asmprinter::switchtosection.
llvm-svn: 79324
more properly belong. This allows removing the front-end conditionalized
SJLJ code, and cleans up the generated IR considerably. All of the
infrastructure code (calling _Unwind_SjLj_Register/Unregister, etc) is
added by the SjLjEHPrepare pass.
llvm-svn: 79250
doing it directly. This requires const'izing a bunch of stuff that
took sections, but this seems like the right semantic thing to do:
emitting a label to a section shouldn't mutate the MCSection object
itself, for example.
llvm-svn: 79227
support unaligned mem access only for certain types. (Should it be size
instead?)
ARM v7 supports unaligned access for i16 and i32, some v6 variants support it
as well.
llvm-svn: 79127
libcall. Take advantage of this in the ARM backend to rectify broken
choice of CC when hard float is in effect. PIC16 may want to see if
it could be of use in MakePIC16Libcall, which works unchanged.
Patch by Sandeep!
llvm-svn: 79033
In order for the changes in r78424 to work properly, cast_retty<X,Y> should return an object instead of a reference, and it's not clear that this approach has real advantages.
llvm-svn: 79023
specific printer (this only works on x86, for now).
- This makes it possible to do some correctness checking of the parsing and
matching, since we can compare the results of 'as' on the original input, to
those of 'as' on the output from llvm-mc.
- In theory, we could now have an easy ATT -> Intel syntax converter. :)
llvm-svn: 78986
TargetAsmInfo. This eliminates a dependency on TargetMachine.h from
TargetRegistry.h, which technically was a layering violation.
- Clients probably can only sensibly pass in the same TargetAsmInfo as the
TargetMachine has, but there are only limited clients of this API.
llvm-svn: 78928
unbuffered. std::ostream does its own buffering, and std::string and
SmallVector both have allocation strategies intended to handle frequent
appending.
llvm-svn: 78924
x86_64-apple-darwin10.
--- Reverse-merging r78895 into '.':
U test/CodeGen/PowerPC/2008-12-12-EH.ll
U lib/Target/DarwinTargetAsmInfo.cpp
--- Reverse-merging r78892 into '.':
U include/llvm/Target/DarwinTargetAsmInfo.h
U lib/Target/X86/X86TargetAsmInfo.cpp
U lib/Target/X86/X86TargetAsmInfo.h
U lib/Target/ARM/ARMTargetAsmInfo.h
U lib/Target/ARM/ARMTargetMachine.cpp
U lib/Target/ARM/ARMTargetAsmInfo.cpp
U lib/Target/PowerPC/PPCTargetAsmInfo.cpp
U lib/Target/PowerPC/PPCTargetAsmInfo.h
U lib/Target/PowerPC/PPCTargetMachine.cpp
G lib/Target/DarwinTargetAsmInfo.cpp
llvm-svn: 78919
There have been a few times where I've wanted this but ended up leaving the
operand type unconstrained. It is easy to add this now and should help
catch errors in the future.
llvm-svn: 78849
pair instead of from a virtual method on TargetMachine. This cuts the final
ties of TargetAsmInfo to TargetMachine, meaning that MC can now use
TargetAsmInfo.
llvm-svn: 78802
the darwin version string. This should help consolidate
the variety of weird functions we have scattered around the
codebase that do stuff like this.
llvm-svn: 78792
- Used to mark fake instructions which don't correspond to an actual machine
instruction (or are duplicates of a real instruction). This is to be used for
"special cases" in the .td files, which should be ignored by things like the
assembler and disassembler. We still need a good solution to handle pervasive
duplication, like with the Int_ instructions.
- Set the bit on fake "mov 0" style instructions, which allows turning an
assembler matcher warning into a hard error.
- -2 FIXMEs.
llvm-svn: 78731
version. This allows TAI implementations to specify the directive to use
based on the mode being codegen'd for.
The real fix for this is to remove JumpTableDirective, but I don't feel
like diving into the jumptable snarl just now.
llvm-svn: 78709
The register scavenger maintains a DistanceMap that maps MI pointers to their
distance from the top of the current MBB. The DistanceMap is built
incrementally in forward() and in bulk in findFirstUse(). It is used by
scavengeRegister() to determine which candidate register has the longest
unused interval.
Unfortunately the DistanceMap contents can become outdated. The first time
scavengeRegister() is called, the DistanceMap is filled to cover the MBB. If
then instructions are inserted in the MBB (as they always are following
scavengeRegister()), the recorded distances are too short. This causes bad
behaviour in the included test case where a register use /after/ the current
position is ignored because findFirstUse() thinks is is /before/ the current
position. A "using an undefined register" assertion follows promptly.
The fix is to build a fresh DistanceMap at the top of scavengeRegister(), and
discard it after use. This means that DistanceMap is no longer needed as a
RegScavenger member variable, and forward() doesn't need to update it.
The fix then discloses issue number two in the same test case: The candidate
search in scavengeRegister() finds a CSR that has been saved in the prologue,
but is currently unused. It would be both inefficient and wrong to spill such
a register in the emergency spill slot. In the present case, the emergency
slot restore is placed immediately before the normal epilogue restore, leading
to a "Redefining a live register" assertion.
Fix number two: When scavengerRegister() stumbles upon an unused register that
is overwritten later in the MBB, return that register early. It is important
to verify that the register is defined later in the MBB, otherwise it might be
an unspilled CSR.
llvm-svn: 78650
the overloaded vector types allowed floating-point or integer vector elements.
Most of these operations actually depend on the element type, so bitcasting
was not an option.
If you include the vpadd intrinsics that I updated earlier, this gets rid
of 20 intrinsics.
llvm-svn: 78646
and short. Well, it's kinda short. Definitely nasty and brutish.
The front-end generates the register/unregister calls into the SjLj runtime,
call-site indices and landing pad dispatch. The back end fills in the LSDA
with the call-site information provided by the front end. Catch blocks are
not yet implemented.
Built on Darwin and verified no llvm-core "make check" regressions.
llvm-svn: 78625
Andreas Neustifter