------------------------------------------------------------------------
r205738 | atrick | 2014-04-07 17:29:22 -0400 (Mon, 07 Apr 2014) | 16 lines
Put a limit on ScheduleDAGSDNodes::ClusterNeighboringLoads to avoid blowing up compile time.
Fixes PR16365 - Extremely slow compilation in -O1 and -O2.
The SD scheduler has a quadratic implementation of load clustering
which absolutely blows up compile time for large blocks with constant
pool loads. The MI scheduler has a better implementation of load
clustering. However, we have not done the work yet to completely
eliminate the SD scheduler. Some benchmarks still seem to benefit from
early load clustering, although maybe by chance.
As an intermediate term fix, I just put a nice limit on the number of
DAG users to search before finding a match. With this limit there are no
binary differences in the LLVM test suite, and the PR16365 test case
does not suffer any compile time impact from this routine.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205808 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r200202 | stpworld | 2014-01-27 04:43:10 -0500 (Mon, 27 Jan 2014) | 2 lines
Additional fix for 200201: due to dependence on bitwidth test was moved to X86 directory.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205807 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r200201 | stpworld | 2014-01-27 04:18:31 -0500 (Mon, 27 Jan 2014) | 31 lines
Fix for PR18102.
Issue outcomes from DAGCombiner::MergeConsequtiveStores, more precisely from
mem-ops sequence sorting.
Consider, how MergeConsequtiveStores works for next example:
store i8 1, a[0]
store i8 2, a[1]
store i8 3, a[1] ; a[1] again.
return ; DAG starts here
1. Method will collect all the 3 stores.
2. It sorts them by distance from the base pointer (farthest with highest
index).
3. It takes first consecutive non-overlapping stores and (if possible) replaces
them with a single store instruction.
The point is, we can't determine here which 'store' instruction
would be the second after sorting ('store 2' or 'store 3').
It happens that 'store 3' would be the second, and 'store 2' would be the third.
So after merging we have the next result:
store i16 (1 | 3 << 8), base ; is a[0] but bit-casted to i16
store i8 2, a[1]
So actually we swapped 'store 3' and 'store 2' and got wrong contents in a[1].
Fix: In sort routine just also take into account mem-op sequence number.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205806 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r203725 | rafael.espindola | 2014-03-12 18:03:43 -0400 (Wed, 12 Mar 2014) | 2 lines
This test need the X86 backend, move it to the X86 sub directory.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205798 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r201104 | chandlerc | 2014-02-10 14:39:35 -0500 (Mon, 10 Feb 2014) | 26 lines
[LPM] A terribly simple fix to a terribly complex bug: PR18773.
The crux of the issue is that LCSSA doesn't preserve stateful alias
analyses. Before r200067, LICM didn't cause LCSSA to run in the LTO pass
manager, where LICM runs essentially without any of the other loop
passes. As a consequence the globalmodref-aa pass run before that loop
pass manager was able to survive the loop pass manager and be used by
DSE to eliminate stores in the function called from the loop body in
Adobe-C++/loop_unroll (and similar patterns in other benchmarks).
When LICM was taught to preserve LCSSA it had to require it as well.
This caused it to be run in the loop pass manager and because it did not
preserve AA, the stateful AA was lost. Most of LLVM's AA isn't stateful
and so this didn't manifest in most cases. Also, in most cases LCSSA was
already running, and so there was no interesting change.
The real kicker is that LCSSA by its definition (injecting PHI nodes
only) trivially preserves AA! All we need to do is mark it, and then
everything goes back to working as intended. It probably was blocking
some other weird cases of stateful AA but the only one I have is
a 1000-line IR test case from loop_unroll, so I don't really have a good
test case here.
Hopefully this fixes the regressions on performance that have been seen
since that revision.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205795 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r198863 | stpworld | 2014-01-09 07:26:12 -0500 (Thu, 09 Jan 2014) | 6 lines
Fixed old typo in ScalarEvolution, that caused wrong SCEVs zext operation.
Detailed description is here:
http://llvm.org/bugs/show_bug.cgi?id=18000#c16
For participation in bugfix process special thanks to David Wiberg.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205794 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r198744 | iain | 2014-01-08 05:22:54 -0500 (Wed, 08 Jan 2014) | 8 lines
[patch] Adjust behavior of FDE cross-section relocs for targets that don't support abs-differences.
Modern versions of OSX/Darwin's ld (ld64 > 97.17) have an optimisation present that allows the back end to omit relocations (and replace them with an absolute difference) for FDE some text section refs.
This patch allows a backend to opt-in to this behaviour by setting "DwarfFDESymbolsUseAbsDiff". At present, this is only enabled for modern x86 OSX ports.
test changes by David Fang.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205768 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197574 | rafael.espindola | 2013-12-18 10:06:25 -0500 (Wed, 18 Dec 2013) | 8 lines
Fix f64 and f128 for ppc-darwin.
This patch adds -f64:32:64 to 32 bit ppc darwin since a f64 inside a
structure are only 32 bit aligned.
The patch also drop -f128:64:128 from all ppc darwin, since f128 is
128 bit aligned.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205767 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197572 | rafael.espindola | 2013-12-18 09:35:37 -0500 (Wed, 18 Dec 2013) | 6 lines
One ppc32-darwin, a i64 inside a structure can have 32 bit alignment.
Thanks for Iain Sandoe for testing this with the original gcc.
Clang was already getting this right.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205766 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r196987 | rafael.espindola | 2013-12-10 19:09:06 -0500 (Tue, 10 Dec 2013) | 2 lines
Move PPC's getDataLayoutString out of line and document it better.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205765 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r196168 | rafael.espindola | 2013-12-02 18:04:51 -0500 (Mon, 02 Dec 2013) | 2 lines
Convert two char* that are only ever used as booleans to bool.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@205764 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r203818 | thomas.stellard | 2014-03-13 10:13:04 -0700 (Thu, 13 Mar 2014) | 7 lines
R600: LDS instructions shouldn't implicitly define OQAP
LDS instructions are pseudo instructions which model
the OQAP defs and uses within a single instruction.
This fixes a hang in the opencv MedianFilter tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204650 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r203281 | thomas.stellard | 2014-03-07 12:12:39 -0800 (Fri, 07 Mar 2014) | 4 lines
R600/SI: Using SGPRs is illegal for instructions that read carry-out
from VCC
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204649 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r201097 | thomas.stellard | 2014-02-10 08:58:30 -0800 (Mon, 10 Feb 2014) | 9 lines
R600/SI: Initialize M0 and emit S_WQM_B64 whenever DS instructions are
used
DS instructions that access local memory can only uses addresses that
are less than or equal to the value of M0. When M0 is uninitialized,
then we experience undefined behavior.
This patch also changes the behavior to emit S_WQM_B64 on pixel shaders
no matter what kind of DS instruction is used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204648 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r201096 | thomas.stellard | 2014-02-10 08:58:27 -0800 (Mon, 10 Feb 2014) | 6 lines
R600/SI: Only use S_WQM_B64 in pixel shaders
This doesn't change any functionality, since we only have two shader
types (compute and pixel) that use local memory. We're just changing
the logic to match the documentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204647 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r200776 | thomas.stellard | 2014-02-04 09:18:43 -0800 (Tue, 04 Feb 2014) | 9 lines
R600/SI: Expand i1 BR_CC
This fixes a crashes in the OpenCV test suite and also the scrypt
kernel in bfgminer.
I was unable to come up with a reduced test case for this.
https://bugs.freedesktop.org/show_bug.cgi?id=72785
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204645 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r200775 | thomas.stellard | 2014-02-04 09:18:42 -0800 (Tue, 04 Feb 2014) | 5 lines
R600/SI: Don't assume copies will be coalesced in SIFixSGPRCopies
There is no lit test for this, because it would be too big and
complicated, but it does fix a crash in the Arithm/Absdiff.* OpenCV
test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204644 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r200743 | michel.daenzer | 2014-02-03 23:12:38 -0800 (Mon, 03 Feb 2014) | 11 lines
R600/SI: Fix fneg for 0.0
V_ADD_F32 with source modifier does not produce -0.0 for this. Just
manipulate the sign bit directly instead.
Also add a pattern for (fneg (fabs ...)).
Fixes a bunch of bit encoding piglit tests with radeonsi.
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204643 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r199919 | thomas.stellard | 2014-01-23 10:49:34 -0800 (Thu, 23 Jan 2014) | 10 lines
R600: Remove successive JUMP in AnalyzeBranch when AllowModify is true
This fixes a crash in the OpenCV OpenCL test suite.
There is no lit test for this, because the test would be very large
and could easily be invalidated by changes to the scheduler
or other parts of the compiler.
Patch by: Vincent Lejeune
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204641 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r199918 | thomas.stellard | 2014-01-23 10:49:33 -0800 (Thu, 23 Jan 2014) | 8 lines
R600: Disable the BFE pattern
This pattern uses an SDNodeXForm, which isn't being emitted for some
reason. I can get it to work by attaching the PatLeaf that has the
XForm to the argument in the output pattern, but this results in an
immediate being used in a register operand, which the backend can't
handle yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204640 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r199917 | thomas.stellard | 2014-01-23 10:49:31 -0800 (Thu, 23 Jan 2014) | 6 lines
R600: Correctly handle vertex fetch clauses the precede ENDIFs
The control flow finalizer would sometimes use an ALU_POP_AFTER
instruction before the vetex fetch clause instead of using a POP
instruction after it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204639 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r202720 | thomas.stellard | 2014-03-03 07:22:00 -0800 (Mon, 03 Mar 2014) | 8 lines
Add patch level to llvm version in CMake and Autoconf
The shared library generated by autoconf will now be called
libLLVM-$(VERSION_MAJOR).$(VERSION_MINOR).$(VERSION_PATCH)$(VERSION_SUFFIX).so
and a symlink named
libLLVM-$(VERSION_MAJOR).$(VERSION_MINOR)$(VERSION_SUFFIX).so will
also be created in the install directory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@204262 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197520 | dexonsmith | 2013-12-17 12:28:21 -0800 (Tue, 17 Dec 2013) | 7
lines
Assert that the last operand is actually EFLAGS
This is another follow-up to r197503, after a post-commit review by
Andy.
<rdar://problem/15627766>
------------------------------------------------------------------------
r197505 | dexonsmith | 2013-12-17 08:20:37 -0800 (Tue, 17 Dec 2013) | 6
lines
Setting the CPU in the new vaargs test
Trying to fix buildbots after r197503 (test passes locally).
<rdar://problem/15627766>
------------------------------------------------------------------------
r197503 | dexonsmith | 2013-12-17 07:54:45 -0800 (Tue, 17 Dec 2013) | 17
lines
Revert "Revert "Mark vastart_save_xmm_regs as changing EFLAGS""
This reverts commit r197481, recommiting r197469 with an extra fix.
The vastart_save_xmm_regs pseudo-instruction expands to a test and a
branch, so it modifies EFLAGS. Mark it so, or else the scheduler might
place it in the middle of another test+branch.
This fixes a bug exposed by r192750, which changed the initial scheduler
to source-order as part of enabling the MI Scheduler for X86.
This re-commit changes the VASTART_SAVE_XMM_REGS custom inserter not to
try to save %flags, and adds a test that catches the bad behavior of
r197469.
<rdar://problem/15627766>
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@202060 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197483 | yrnkrn | 2013-12-17 00:40:11 -0800 (Tue, 17 Dec 2013) | 8 lines
There are no __register_frame and __deregister_frame functions
when using structured exception handling (SEH) on Windows 64.
http://llvm-reviews.chandlerc.com/D2378
Patch by Jonathan Liu!
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197944 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197492 | dyatkovskiy | 2013-12-17 04:07:33 -0800 (Tue, 17 Dec 2013) | 26 lines
Fix for PR18045:
http://llvm.org/bugs/show_bug.cgi?id=18045
Short issue description:
For X86 machines with sse < sse4.1 we got failures for some
particular load/store vector sequences:
$ clang-trunk -m32 -O2 test-case.c
fatal error: error in backend: Cannot select: 0x4200920: v4i32,ch = load 0x41d6ab0, 0x4205850,
0x41dcb10<LD16[getelementptr inbounds ([4 x i32]* @e, i32 0, i32 0)](align=4)> [ORD=82]
[ID=58]
0x4205850: i32 = X86ISD::Wrapper 0x41d5490 [ORD=26] [ID=43]
0x41d5490: i32 = TargetGlobalAddress<[4 x i32]* @e> 0 [ORD=26] [ID=23]
0x41dcb10: i32 = undef [ID=2]
The reason is that EltsFromConsecutiveLoads could emit such load instruction
both before and after legalize stage. Though this instruction is not legal for
machines with SSSE3 and lower.
The fix: In EltsFromConsecutiveLoads, if we have passed legalize stage, we
check whether nodes it emits are legal.
P.S.: If you get failure in time from 12:00 and till 22:00 (UTC-8),
perhaps I'll slow with response, so you better reject this commit. Thanks!
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197779 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197718 | hans | 2013-12-19 12:32:44 -0800 (Thu, 19 Dec 2013) | 10 lines
Make sys::ThreadLocal<> zero-initialized on non-thread builds (PR18205)
According to the docs, ThreadLocal<>::get() should return NULL
if no object has been set. This patch makes that the case also for non-thread
builds and adds a very basic unit test to check it.
(This was causing PR18205 because PrettyStackTraceHead didn't get zero-
initialized and we'd crash trying to read past the end of that list. We didn't
notice this so much on Linux since we'd crash after printing all the entries,
but on Mac we print into a SmallString, and would crash before printing that.)
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197778 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197449 | arnolds | 2013-12-16 17:11:01 -0800 (Mon, 16 Dec 2013) | 7 lines
LoopVectorizer: Don't if-convert constant expressions that can trap
A phi node operand or an instruction operand could be a constant expression that
can trap (division). Check that we don't vectorize such cases.
PR16729
radar://15653590
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197453 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r195411 | mgottesman | 2013-11-21 21:00:51 -0800 (Thu, 21 Nov 2013) | 1 line
[block-freq] Update data in test case to be unsigned long long to fix mingw build.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197363 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197216 | chandlerc | 2013-12-13 00:00:01 -0800 (Fri, 13 Dec 2013) | 9 lines
[inliner] Fix PR18206 by preventing inlining functions that call setjmp
through an invoke instruction.
The original patch for this was written by Mark Seaborn, but I've
reworked his test case into the existing returns_twice test case and
implemented the fix by the prior refactoring to actually run the cost
analysis over invoke instructions, and then here fixing our detection of
the returns_twice attribute to work for both calls and invokes. We never
noticed because we never saw an invoke. =[
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197352 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197215 | chandlerc | 2013-12-12 23:59:56 -0800 (Thu, 12 Dec 2013) | 24 lines
[inliner] Completely change (and fix) how the inline cost analysis
handles terminator instructions.
The inline cost analysis inheritted some pretty rough handling of
terminator insts from the original cost analysis, and then made it much,
much worse by factoring all of the important analyses into a separate
instruction visitor. That instruction visitor never visited the
terminator.
This works fine for things like conditional branches, but for many other
things we simply computed The Wrong Value. First example are
unconditional branches, which should be free but were counted as full
cost. This is most significant for conditional branches where the
condition simplifies and folds during inlining. We paid a 1 instruction
tax on every branch in a straight line specialized path. =[
Oh, we also claimed that the unreachable instruction had cost.
But it gets worse. Let's consider invoke. We never applied the call
penalty. We never accounted for the cost of the arguments. Nope. Worse
still, we didn't handle the *correctness* constraints of not inlining
recursive invokes, or exception throwing returns_twice functions. Oops.
See PR18206. Sadly, PR18206 requires yet another fix, but this
refactoring is at least a huge step in that direction.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197351 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197178 | hfinkel | 2013-12-12 12:45:24 -0800 (Thu, 12 Dec 2013) | 9 lines
Fix a use-after-free error in GlobalOpt CleanupConstantGlobalUsers
GlobalOpt's CleanupConstantGlobalUsers function uses a worklist array to manage
constant users to be visited. The pointers in this array need to be weak
handles because when we delete a constant array, we may also be holding a
pointer to one of its elements (or an element of one of its elements if we're
dealing with an array of arrays) in the worklist.
Fixes PR17347.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197322 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197228 | d0k | 2013-12-13 05:40:24 -0800 (Fri, 13 Dec 2013) | 8 lines
X86: When lowering shl_parts, don't emit shift amounts larger than the bit width.
While it's safe for the X86-specific shift nodes, dag combining will
kill generic nodes. Insert an AND to make it safe, isel will nuke it
as x86's shift instructions have an implicit AND.
Fixes PR16108, which contains a contraption to hit this case in between
constant folders.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197321 91177308-0d34-0410-b5e6-96231b3b80d8
------------------------------------------------------------------------
r197089 | hfinkel | 2013-12-11 15:12:25 -0800 (Wed, 11 Dec 2013) | 6 lines
Fix the PPC subsumes-predicate check
For one predicate to subsume another, they must both check the same condition
register. Failure to check this prerequisite was causing miscompiles.
Fixes PR18003.
------------------------------------------------------------------------
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197126 91177308-0d34-0410-b5e6-96231b3b80d8
