Reapply r286007, modified to compile with pre-C++11 compilers:

Pull in r219009 from upstream llvm trunk (by Adam Nemet):

[ISel] Keep matching state consistent when folding during X86 address match

In the X86 backend, matching an address is initiated by the 'addr' complex
pattern and its friends. During this process we may reassociate and-of-shift
into shift-of-and (FoldMaskedShiftToScaledMask) to allow folding of the
shift into the scale of the address.

However as demonstrated by the testcase, this can trigger CSE of not only the
shift and the AND which the code is prepared for but also the underlying load
node. In the testcase this node is sitting in the RecordedNode and MatchScope
data structures of the matcher and becomes a deleted node upon CSE. Returning
from the complex pattern function, we try to access it again hitting an assert
because the node is no longer a load even though this was checked before.

Now obviously changing the DAG this late is bending the rules but I think it
makes sense somewhat. Outside of addresses we prefer and-of-shift because it
may lead to smaller immediates (FoldMaskAndShiftToScale is an even better
example because it create a non-canonical node). We currently don't recognize
addresses during DAGCombiner where arguably this canonicalization should be
performed. On the other hand, having this in the matcher allows us to cover
all the cases where an address can be used in an instruction.

I've also talked a little bit to Dan Gohman on llvm-dev who added the RAUW for
the new shift node in FoldMaskedShiftToScaledMask. This RAUW is responsible
for initiating the recursive CSE on users
(http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-September/076903.html) but it
is not strictly necessary since the shift is hooked into the visited user. Of
course it's safer to keep the DAG consistent at all times (e.g. for accurate
number of uses, etc.).

So rather than changing the fundamentals, I've decided to continue along the
previous patches and detect the CSE. This patch installs a very targeted
DAGUpdateListener for the duration of a complex-pattern match and updates the
matching state accordingly. (Previous patches used HandleSDNode to detect the
CSE but that's not practical here). The listener is only installed on X86.

I tested that there is no measurable overhead due to this while running
through the spec2k BC files with llc. The only thing we pay for is the
creation of the listener. The callback never ever triggers in spec2k since
this is a corner case.

Fixes rdar://problem/18206171

This fixes a possible crash in x86 code generation when compiling recent
llvm/clang trunk sources.

Direct commit to stable/10, since head already has llvm/clang 3.6.1,
which includes this fix.

Reported by: jonathan, theraven
Upstream PR: https://llvm.org/bugs/show_bug.cgi?id=24249

Revert r286007-r286009 for now, until I can figure out how to make the
fix compile with older gcc and libstdc++.

Pull in r219009 from upstream llvm trunk (by Adam Nemet):

[ISel] Keep matching state consistent when folding during X86 address match

In the X86 backend, matching an address is initiated by the 'addr' complex
pattern and its friends. During this process we may reassociate and-of-shift
into shift-of-and (FoldMaskedShiftToScaledMask) to allow folding of the
shift into the scale of the address.

However as demonstrated by the testcase, this can trigger CSE of not only the
shift and the AND which the code is prepared for but also the underlying load
node. In the testcase this node is sitting in the RecordedNode and MatchScope
data structures of the matcher and becomes a deleted node upon CSE. Returning
from the complex pattern function, we try to access it again hitting an assert
because the node is no longer a load even though this was checked before.

Now obviously changing the DAG this late is bending the rules but I think it
makes sense somewhat. Outside of addresses we prefer and-of-shift because it
may lead to smaller immediates (FoldMaskAndShiftToScale is an even better
example because it create a non-canonical node). We currently don't recognize
addresses during DAGCombiner where arguably this canonicalization should be
performed. On the other hand, having this in the matcher allows us to cover
all the cases where an address can be used in an instruction.

I've also talked a little bit to Dan Gohman on llvm-dev who added the RAUW for
the new shift node in FoldMaskedShiftToScaledMask. This RAUW is responsible
for initiating the recursive CSE on users
(http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-September/076903.html) but it
is not strictly necessary since the shift is hooked into the visited user. Of
course it's safer to keep the DAG consistent at all times (e.g. for accurate
number of uses, etc.).

So rather than changing the fundamentals, I've decided to continue along the
previous patches and detect the CSE. This patch installs a very targeted
DAGUpdateListener for the duration of a complex-pattern match and updates the
matching state accordingly. (Previous patches used HandleSDNode to detect the
CSE but that's not practical here). The listener is only installed on X86.

I tested that there is no measurable overhead due to this while running
through the spec2k BC files with llc. The only thing we pay for is the
creation of the listener. The callback never ever triggers in spec2k since
this is a corner case.

Fixes rdar://problem/18206171

This fixes a possible crash in x86 code generation when compiling recent
llvm/clang trunk sources.

Direct commit to stable/10, since head already has llvm/clang 3.6.1,
which includes this fix.

Reported by: jonathan, theraven
Upstream PR: https://llvm.org/bugs/show_bug.cgi?id=24249

MFC r274442:

Pull in r221709 from upstream llvm trunk (by Frédéric Riss):

Totally forget deallocated SDNodes in SDDbgInfo.

What would happen before that commit is that the SDDbgValues associated with
a deallocated SDNode would be marked Invalidated, but SDDbgInfo would keep
a map entry keyed by the SDNode pointer pointing to this list of invalidated
SDDbgNodes. As the memory gets reused, the list might get wrongly associated
with another new SDNode. As the SDDbgValues are cloned when they are transfered,
this can lead to an exponential number of SDDbgValues being produced during
DAGCombine like in http://llvm.org/bugs/show_bug.cgi?id=20893

Note that the previous behavior wasn't really buggy as the invalidation made
sure that the SDDbgValues won't be used. This commit can be considered a
memory optimization and as such is really hard to validate in a unit-test.

This should fix abnormally large memory usage and resulting OOM crashes
when compiling certain ports with debug information.

Reported by: Dmitry Marakasov <amdmi3@amdmi3.ru>
Upstream PRs: http://llvm.org/PR19031 http://llvm.org/PR20893

MFC r274483:

The fix imported into llvm in r274442 contains some C++11 constructs,
which gcc in base cannot handle. Replace these with C++98 equivalents.

While here, add the patch for the adapted fix.

Reported by: bz, kib
Pointy hat to: dim

MFC debug info for variadic functions

r264826: Merge LLVM r202188:

Debug info: Support variadic functions.
Variadic functions have an unspecified parameter tag after the last
argument. In IR this is represented as an unspecified parameter in the
subroutine type.

Paired commit with CFE r202185.

rdar://problem/13690847

This re-applies r202184 + a bugfix in DwarfDebug's argument handling.

This merge includes a change to use the LLVM 3.4 API in
lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:

DwarfUnit -> CompileUnit

r264827: Merge Clang r202185:

Debug info: Generate debug info for variadic functions.
Paired commit with LLVM.

rdar://problem/13690847

This merege includes changes to use the Clang 3.4 API (revisions
199686 and 200082) in lib/CodeGen/CGDebugInfo.cpp:

getParamType -> getArgType
getNumParams -> getNumArgs
getReturnType -> getResultType

r264828: Add patches corresponding to r264826 and r264827

Sponsored by: DARPA, AFRL

MFC r265925:

Upgrade our copy of llvm/clang to 3.4.1 release. This release contains
mostly fixes, for the following upstream bugs:

http://llvm.org/PR16365 http://llvm.org/PR17473 http://llvm.org/PR18000
http://llvm.org/PR18068 http://llvm.org/PR18102 http://llvm.org/PR18165
http://llvm.org/PR18260 http://llvm.org/PR18290 http://llvm.org/PR18316
http://llvm.org/PR18460 http://llvm.org/PR18473 http://llvm.org/PR18515
http://llvm.org/PR18526 http://llvm.org/PR18600 http://llvm.org/PR18762
http://llvm.org/PR18773 http://llvm.org/PR18860 http://llvm.org/PR18994
http://llvm.org/PR19007 http://llvm.org/PR19010 http://llvm.org/PR19033
http://llvm.org/PR19059 http://llvm.org/PR19144 http://llvm.org/PR19326

MFC r263312:

Pull in r196939 from upstream llvm trunk (by Reid Kleckner):

Reland "Fix miscompile of MS inline assembly with stack realignment"

This re-lands commit r196876, which was reverted in r196879.

The tests have been fixed to pass on platforms with a stack alignment
larger than 4.

Update to clang side tests will land shortly.

Pull in r196986 from upstream llvm trunk (by Reid Kleckner):

Revert the backend fatal error from r196939

The combination of inline asm, stack realignment, and dynamic allocas
turns out to be too common to reject out of hand.

ASan inserts empy inline asm fragments and uses aligned allocas.
Compiling any trivial function containing a dynamic alloca with ASan is
enough to trigger the check.

XFAIL the test cases that would be miscompiled and add one that uses the
relevant functionality.

Pull in r202930 from upstream llvm trunk (by Hans Wennborg):

Check for dynamic allocas and inline asm that clobbers sp before building
selection dag (PR19012)

In X86SelectionDagInfo::EmitTargetCodeForMemcpy we check with MachineFrameInfo
to make sure that ESI isn't used as a base pointer register before we choose to
emit rep movs (which clobbers esi).

The problem is that MachineFrameInfo wouldn't know about dynamic allocas or
inline asm that clobbers the stack pointer until SelectionDAGBuilder has
encountered them.

This patch fixes the problem by checking for such things when building the
FunctionLoweringInfo.

Differential Revision: http://llvm-reviews.chandlerc.com/D2954

Together, these commits fix the problem encountered in the devel/emacs
port on the i386 architecture, where a combination of stack realignment,
alloca() and memcpy() could incidentally clobber the %esi register,
leading to segfaults in the temacs build-time utility.

See also: http://llvm.org/PR18171 and http://llvm.org/PR19012

Reported by: ashish
PR: ports/183064

MFC r263313:

Pull in r203311 from upstream llvm trunk (by Arnold Schwaighofer):

ISel: Make VSELECT selection terminate in cases where the condition type has to
be split and the result type widened.

When the condition of a vselect has to be split it makes no sense widening the
vselect and thereby widening the condition. We end up in an endless loop of
widening (vselect result type) and splitting (condition mask type) doing this.
Instead, split both the condition and the vselect and widen the result.

I ran this over the test suite with i686 and mattr=+sse and saw no regressions.

Fixes PR18036.

With this fix the original problem case from the graphics/rawtherapee
port (posted in http://llvm.org/PR18036 ) now compiles within ~97MB RSS.

Reported by: mandree

MFC r263320:

Add separate patch files for all the customizations we have currently
applied to our copy of llvm/clang. These can be applied in alphabetical
order to a pristine llvm/clang 3.4 release source tree, to result in the
same version used in FreeBSD.

This is intended to clearly document all the changes until now, which
mostly consist of cherry pickings from the respective upstream trunks,
plus a number of hand-written FreeBSD-specific ones. Hopefully those
can eventually be cleaned up and sent upstream too.

MFC r262613:

Merge the projects/clang-sparc64 branch back to head. This brings in
several updates from the llvm and clang trunks to make the sparc64
backend fully functional.

Apart from one patch to sys/sparc64/include/pcpu.h which is still under
discussion, this makes it possible to let clang fully build world and
kernel for sparc64.

Any assistance with testing this on actual sparc64 hardware is greatly
appreciated, as there will unavoidably be bugs left.

Many thanks go to Roman Divacky for his upstream work on getting the
sparc64 backend into shape.

MFC r262985:

Repair a few minor mismerges from r262261 in the clang-sparc64 project
branch. This is also to minimize differences with upstream.

MFC 261991:

Upgrade our copy of llvm/clang to 3.4 release. This version supports
all of the features in the current working draft of the upcoming C++
standard, provisionally named C++1y.

The code generator's performance is greatly increased, and the loop
auto-vectorizer is now enabled at -Os and -O2 in addition to -O3. The
PowerPC backend has made several major improvements to code generation
quality and compile time, and the X86, SPARC, ARM32, Aarch64 and SystemZ
backends have all seen major feature work.

Release notes for llvm and clang can be found here:
<http://llvm.org/releases/3.4/docs/ReleaseNotes.html>
<http://llvm.org/releases/3.4/tools/clang/docs/ReleaseNotes.html>

MFC 262121 (by emaste):

Update lldb for clang/llvm 3.4 import

This commit largely restores the lldb source to the upstream r196259
snapshot with the addition of threaded inferior support and a few bug
fixes.

Specific upstream lldb revisions restored include:
SVN git
181387 779e6ac
181703 7bef4e2
182099 b31044e
182650 f2dcf35
182683 0d91b80
183862 15c1774
183929 99447a6
184177 0b2934b
184948 4dc3761
184954 007e7bc
186990 eebd175

Sponsored by: DARPA, AFRL

MFC 262186 (by emaste):

Fix mismerge in r262121

A break statement was lost in the merge. The error had no functional
impact, but restore it to reduce the diff against upstream.

MFC 262303:

Pull in r197521 from upstream clang trunk (by rdivacky):

Use the integrated assembler by default on FreeBSD/ppc and ppc64.

Requested by: jhibbits

MFC 262611:

Pull in r196874 from upstream llvm trunk:

Fix a crash that occurs when PWD is invalid.

MCJIT needs to be able to run in hostile environments, even when PWD
is invalid. There's no need to crash MCJIT in this case.

The obvious fix is to simply leave MCContext's CompilationDir empty
when PWD can't be determined. This way, MCJIT clients,
and other clients that link with LLVM don't need a valid working directory.

If we do want to guarantee valid CompilationDir, that should be done
only for clients of getCompilationDir(). This is as simple as checking
for an empty string.

The only current use of getCompilationDir is EmitGenDwarfInfo, which
won't conceivably run with an invalid working dir. However, in the
purely hypothetically and untestable case that this happens, the
AT_comp_dir will be omitted from the compilation_unit DIE.

This should help fix assertions occurring with ports-mgmt/tinderbox,
when it is using jails, and sometimes invalidates clang's current
working directory.

Reported by: decke

MFC 262809:

Pull in r203007 from upstream clang trunk:

Don't produce an alias between destructors with different calling conventions.

Fixes pr19007.

(Please note that is an LLVM PR identifier, not a FreeBSD one.)

This should fix Firefox and/or libxul crashes (due to problems with
regparm/stdcall calling conventions) on i386.

Reported by: multiple users on freebsd-current
PR: bin/187103

MFC 263048:

Repair recognition of "CC" as an alias for the C++ compiler, since it
was silently broken by upstream for a Windows-specific use-case.

Apparently some versions of CMake still rely on this archaic feature...

Reported by: rakuco

MFC 263049:

Garbage collect the old way of adding the libstdc++ include directories
in clang's InitHeaderSearch.cpp. This has been superseded by David
Chisnall's commit in r255321.

Moreover, if libc++ is used, the libstdc++ include directories should
not be in the search path at all. These directories are now only used
if you pass -stdlib=libstdc++.

MFC r258005: Merge upstream LLVM r192118:

Formally added an explicit enum for DWARF TLS support. No functionality
change.

Sponsored by: DARPA, AFRL

MFC r258003: Merge upstream LLVM r182803:

[Mips] Add Mips specific dynamic table entry tags.

Sponsored by: DARPA, AFRL

Copy head (r256279) to stable/10 as part of the 10.0-RELEASE cycle.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

Pull in r189644 from upstream llvm trunk:

Add ms_abi and sysv_abi attribute handling.

Based on a patch by Benno Rice!

This will help to develop EFI support.

Approved by: re (kib)
Verified by: benno
MFC after: 1 week

Import llvm r187614 (git 44c8e34), for lldb's use:

Author: Daniel Malea <daniel.malea@intel.com>
Date: Thu Aug 1 21:18:16 2013 +0000

Fixed the Intel-syntax X86 disassembler to respect the (existing)
option for hexadecimal immediates, to match AT&T syntax. This also
brings a new option for C-vs-MASM-style hex.

Patch by Richard Mitton
Reviewed: http://llvm-reviews.chandlerc.com/D1243

Pull in r185594 from llvm trunk:

Add MachineBasicBlock::addLiveIn().

This function adds a live-in physical register to an MBB and ensures
that it is copied to a virtual register immediately.

Pull in r185615 from llvm trunk:

Live-in copies go *after* EH_LABELs.

This will soon be tested by exception handling working at all.

Pull in r185617 from llvm trunk:

Simplify landing pad lowering.

Stop using the ISD::EXCEPTIONADDR and ISD::EHSELECTION when lowering
landing pad arguments. These nodes were previously legalized into
CopyFromReg nodes, but that never worked properly because the
CopyFromReg node weren't guaranteed to be scheduled at the top of the
basic block.

This meant the exception pointer and selector registers could be
clobbered before being copied to a virtual register.

This patch copies the two physical registers to virtual registers at
the beginning of the basic block, and lowers the landingpad instruction
directly to two CopyFromReg nodes reading the *virtual* registers. This
is safe because virtual registers don't get clobbered.

A future patch will remove the ISD::EXCEPTIONADDR and ISD::EHSELECTION
nodes.

Together, these changes fix llvm PR 16038 ('qt4 webcore file results in
"Bad machine code: Using an undefined physical register"'), and should
make it possible again to compile the www/qt4-webkit port again on the
i386 arch, without using a CPUTYPE=i686 or higher setting.

Upgrade our copy of llvm/clang to 3.3 release.

Release notes are still in the works, these will follow soon.

MFC after: 1 month

Upgrade our copy of llvm/clang to trunk r178860, in preparation of the
upcoming 3.3 release (branching and freezing expected in a few weeks).

Preliminary release notes can be found at the usual location:
<http://llvm.org/docs/ReleaseNotes.html>

An MFC is planned once the actual 3.3 release is finished.

Upgrade our copy of llvm/clang to 3.2 release.

Release notes for llvm:
http://llvm.org/releases/3.2/docs/ReleaseNotes.html

Release notes for clang:
http://llvm.org/releases/3.2/tools/clang/docs/ReleaseNotes.html

MFC after: 2 weeks

Upgrade our copy of llvm/clang to r168974, from upstream's release_32
branch. This is effectively llvm/clang 3.2 RC2; the 3.2 release is
coming soon.

Pull in r164132 from upstream llvm trunk:

When creating MCAsmBackend pass the CPU string as well. In X86AsmBackend
store this and use it to not emit long nops when the CPU is geode which
doesnt support them.

Fixes PR11212.

Pull in r164133 from upstream clang trunk:

Follow up on llvm r164132.

This should prevent illegal instructions when building world on Geode
CPUs (e.g. Soekris).

MFC after: 3 days

Upgrade our copy of llvm/clang to trunk r162107. With thanks to
Benjamin Kramer and Joerg Sonnenberger for their input and fixes.

Pull in r156591 from upstream llvm trunk:

Allow unique_file to take a mode for file permissions, but default
to user only read/write.

and r156592 from upstream clang trunk:

For final output files create them with mode 0664 to match other
compilers and expected defaults.

This should fix clang creating files with mode 0600.

Reported by: James <james@hicag.org>
MFC after: 3 days

Upgrade our copy of llvm/clang to r155985, from upstream's release_31
branch. This brings us very close to the 3.1 release, which is planned
for May 14th.

MFC after: 2 weeks

Upgrade our copy of llvm/clang to trunk r154661, in preparation of the
upcoming 3.1 release (expected in a few weeks). Preliminary release
notes can be found at: <http://llvm.org/docs/ReleaseNotes.html>

MFC after: 2 weeks

Upgrade our copy of llvm/clang to r142614, from upstream's release_30
branch. This brings us very close to the 3.0 release, which is expected
in a week or two.

MFC after: 1 week

Upgrade our copy of llvm/clang to r135360, from upstream's trunk.

Upgrade our copy of llvm/clang to r132879, from upstream's trunk.

Upgrade our copy of llvm/clang to r130700, from upstream's trunk.

Update llvm/clang to trunk r126547.

There are several bugfixes in this update, but the most important one is
to ensure __start_ and __stop_ symbols for linker sets and kernel module
metadata are always emitted in object files:

http://llvm.org/bugs/show_bug.cgi?id=9292

Before this fix, if you compiled kernel modules with clang, they would
not be properly processed by kldxref, and if they had any dependencies,
the kernel would fail to load those. Another problem occurred when
attempting to mount a tmpfs filesystem, which would result in 'operation
not supported by device'.

Upgrade our copy of llvm/clang to r126079, from upstream's trunk.

This contains many improvements, primarily better C++ support, an
integrated assembler for x86 and support for -pg.

Remove more unneeded files and directories from contrib/llvm. This
still allows us to build tblgen and clang, and further reduces the
footprint in the tree.

Approved by: rpaulo (mentor)

Upgrade Clang and LLVM to the 2.8 release. See here for release notes:
http://llvm.org/releases/2.8/docs/ReleaseNotes.html

Approved by: rpaulo (mentor)

Upgrade our Clang in base to r114020, from upstream's release_28 branch.

Approved-by: rpaulo (mentor)

Upgrade our Clang in base to r108428.

This commit merges the latest LLVM sources from the vendor space. It
also updates the build glue to match the new sources. Clang's version
number is changed to match LLVM's, which means /usr/include/clang/2.0
has been renamed to /usr/include/clang/2.8.

Obtained from: projects/clangbsd

Import LLVM/clang from vendor stripped of docs/ test/ website/ www/ examples/
in llvm/ and/or llvm/contrib/clang/ respectively.

Approved by: ed (mentor)
Approved by: core

Update LLVM to r104832.

Update LLVM to r103004.

Update LLVM to r100520.

Update LLVM to r100285.

Update LLVM to r100181.

Update LLVM to r99115.

Update LLVM to r98631.

Update LLVM to r98164.

Update LLVM to r97873.

Update LLVM to 97654.

Update LLVM to r96341.

Update LLVM to r94309.

Update LLVM to 93512.

Update LLVM to 92395.

Update LLVM to 91430.

Update LLVM to r90226.

Remove includes that were removed upstream.

Update LLVM to r89337.

Update LLVM to r89205.

Update LLVM to r86140.

Update LLVM to r86025.

These files does not belong here.

Update LLVM to r84949.

Delete all stale files.

Update llvm to r84175.

Update llvm to r84119.

Import LLVM 74788.

Import LLVM r74383.

Import LLVM r73984.

It seems I keep importing sources at very unlucky moments. Let's see
what this revision of LLVM does.

Import LLVM r73954.

Update LLVM sources to r73879.

Import LLVM r73340.

Import LLVM r73070.

Now I'm going to stop importing code for a while. I spent a lot of time
the last couple of days figuring out which LLVM commit caused g++ to
miscompile, which in its turn caused Clang to miscompile the next build.
I had to run `make buildworld' twice each time I bisected a revision.

Import LLVM r73021.

Import LLVM, at r72995.

We should now have support for #pragma weak.

Import LLVM, at r72805, which fixes PR4315 and PR4316.

Normally I'm not updating sources this often, but I want to get rid of
this breakage, because right now I can't offer a proper source snapshot
yet.

Import LLVM, at r72770.

This should fix LLVM PR4225.

Import LLVM, at r72732.