OSDN Git Service
Dale Johannesen [Tue, 3 Feb 2009 19:33:06 +0000 (19:33 +0000)]
DebugLoc propagation. 2/3 through file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63650
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 19:30:11 +0000 (19:30 +0000)]
make scalar conversion handle stores of first class
aggregate values. loads are not yet handled (coming
soon to an sroa near you).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63649
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 18:57:45 +0000 (18:57 +0000)]
Change the post-RA scheduler to iterate through the
basic-block segments bottom-up instead of top down. This
is the first step in a general restructuring of the way
register liveness is tracked in the post-RA scheduler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63643
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 18:53:21 +0000 (18:53 +0000)]
Tevert part of the x86 subtarget logic changes: when -march=x86-64
is given, override the subtarget settings and enable 64-bit support.
This restores the earlier behavior, and fixes regressions on
Non-64-bit-capable x86-32 hosts.
This isn't necessarily the best approach, but the most obvious
alternative is to require -mcpu=x86-64 or -mattr=+64bit to be used
with -march=x86-64 when the host doesn't have 64-bit support. This
makes things little more consistent, but it's less convenient, and
it has the practical drawback of requiring lots of test changes, so
I opted for the above approach for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63642
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Tue, 3 Feb 2009 18:46:32 +0000 (18:46 +0000)]
First initialize DAG otherwise dwarf writer is used uninitialized.
Duncan spotted this. Thanks!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63641
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 18:15:05 +0000 (18:15 +0000)]
Make SROA produce a vector only when the alloca is actually
accessed at least once as a vector. This prevents it from
compiling the example in not-a-vector into:
define double @test(double %A, double %B) {
%tmp4 = insertelement <7 x double> undef, double %A, i32 0
%tmp = insertelement <7 x double> %tmp4, double %B, i32 4
%tmp2 = extractelement <7 x double> %tmp, i32 4
ret double %tmp2
}
instead, producing the integer code. Producing vectors when they
aren't otherwise in the program is dangerous because a lot of other
code treats them carefully and doesn't want to break them down.
OTOH, many things want to break down tasty i448's.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63638
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 18:13:00 +0000 (18:13 +0000)]
this produces an undefined result, just check that the alloca is gone
and that sroa doesn't crash.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63637
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 3 Feb 2009 10:23:33 +0000 (10:23 +0000)]
Fix PR3411. When replacing values, nodes are analyzed
in any old order. Since analyzing a node analyzes its
operands also, this can mean that when we pop a node
off the list of nodes to be analyzed, it may already
have been analyzed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63632
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 3 Feb 2009 10:05:09 +0000 (10:05 +0000)]
APInt'fy SimplifyDemandedVectorElts so it can analyze vectors with more than 64 elements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63631
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 07:39:50 +0000 (07:39 +0000)]
add a method to BumpPtrAllocator that allows allocating elements
with a specified alignment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63629
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 3 Feb 2009 07:13:24 +0000 (07:13 +0000)]
Add LLVM plugin for gold.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63623
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 3 Feb 2009 07:10:30 +0000 (07:10 +0000)]
Regenerate configure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63622
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 3 Feb 2009 07:10:08 +0000 (07:10 +0000)]
Add the ability to pass the path to binutils source to configure. This is
needed to build the LLVM gold plugin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63621
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 07:08:57 +0000 (07:08 +0000)]
add another case of undefined behavior without crashing, PR3466.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63620
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 3 Feb 2009 06:30:37 +0000 (06:30 +0000)]
Revert r63600. It didn't fix the bug, it just moved it a bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63618
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 3 Feb 2009 06:27:22 +0000 (06:27 +0000)]
Teach AliasAnalysis that a bunch of the atomic intrinsics only dereference their arguments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63616
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 3 Feb 2009 04:34:40 +0000 (04:34 +0000)]
Update the callgraph when replacing InvokeInst with CallInst when inlining.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63600
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 02:29:34 +0000 (02:29 +0000)]
Explicitly pass in debug location information to BuildMI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63599
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 02:20:52 +0000 (02:20 +0000)]
Pass in something sensible for the debug location information when creating the
initial PHI nodes of the machine function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63598
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 02:11:28 +0000 (02:11 +0000)]
fix a bitcode reader bug where it can't handle extractelement correctly:
the index of the value being extracted is always an i32. This fixes PR3465
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63597
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 02:01:43 +0000 (02:01 +0000)]
Teach ConvertUsesToScalar to handle memset, allowing it to handle
crazy cases like:
struct f { int A, B, C, D, E, F; };
short test4() {
struct f A;
A.A = 1;
memset(&A.B, 2, 12);
return A.C;
}
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63596
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Tue, 3 Feb 2009 01:55:44 +0000 (01:55 +0000)]
Fill in some missing DL propagation in getNode()s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63595
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 01:55:42 +0000 (01:55 +0000)]
Use SDL->getCurDebugLoc() instead of unknown loc for landing pads.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63594
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 01:33:28 +0000 (01:33 +0000)]
Explicitly pass in the "unknown" debug location. This is probably not
correct. We need more infrastructure before we can get the DebugLoc info for
these instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63593
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 01:33:26 +0000 (01:33 +0000)]
Delete these two tests. They are specific to x86-64, and there's no
reliable way to do this with the current dejagnu infrastructure.
If someone can figure out how to fix these tests so that they test
what they are intended to test without spuriously failing on any
popular platforms, they are invited to reinstate them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63592
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 01:32:22 +0000 (01:32 +0000)]
Alphabetize includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63591
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 01:30:09 +0000 (01:30 +0000)]
rearrange how SRoA handles promotion of allocas to vectors.
With the new world order, it can handle cases where the first
store into the alloca is an element of the vector, instead of
requiring the first analyzed store to have the vector type
itself. This allows us to un-xfail
test/CodeGen/X86/vec_ins_extract.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63590
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 01:28:32 +0000 (01:28 +0000)]
Move isIdentifiedObject and isNoAliasCall into AliasAnalysis.cpp since
they are useful to analyses other than BasicAliasAnalysis.cpp. Include
the full comment for isIdentifiedObject in the header file. Thanks to
Chris for suggeseting this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63589
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 3 Feb 2009 01:13:52 +0000 (01:13 +0000)]
this test produces an undefined value, we don't care
what it is, but we do want the alloca promoted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63587
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 01:02:39 +0000 (01:02 +0000)]
Propagate debug loc info during SDNode -> machine instr creation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63585
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 00:55:04 +0000 (00:55 +0000)]
Create DebugLoc information in FastISel. Several temporary methods were
created. Specifically, those BuildMIs which use
"DebugLoc::getUnknownLoc()". I'll remove them soon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63584
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Tue, 3 Feb 2009 00:47:48 +0000 (00:47 +0000)]
Propagation in TargetLowering. Includes passing a DL
into SimplifySetCC which gets called elsewhere.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63583
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 00:35:11 +0000 (00:35 +0000)]
It fails on Linux. XFAIL that machine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63582
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Feb 2009 00:27:09 +0000 (00:27 +0000)]
This is passing for us. Should it have been reenabled?
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63580
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 00:20:22 +0000 (00:20 +0000)]
Add explicit -march=x86 to these tests so that they don't
default to -march=x86-64 on 64-bit hosts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63579
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 00:08:45 +0000 (00:08 +0000)]
Use the SubclassData field to hold ExtType, isTrunc, and MemIndexedMode
information. This eliminates the need for the Flags field in MemSDNode,
so this makes LoadSDNode and StoreSDNode smaller. Also, it makes
FoldingSetNodeIDs for loads and stores two AddIntegers smaller.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63577
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 3 Feb 2009 00:04:43 +0000 (00:04 +0000)]
Change Feature64Bit to not imply FeatureSSE2. All x86-64 hardware has
SSE2, however it's possible to disable SSE2, and the subtarget support
code thinks that if 64-bit implies SSE2 and SSE2 is disabled then
64-bit should also be disabled. Instead, just mark all the 64-bit
subtargets as explicitly supporting SSE2.
Also, move the code that makes -march=x86-64 enable 64-bit support by
default to only apply when there is no explicit subtarget. If you
need to specify a subtarget and you want 64-bit code, you'll need to
select a subtarget that supports 64-bit code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63575
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 2 Feb 2009 23:46:53 +0000 (23:46 +0000)]
Last DebugLoc propagation for this file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63574
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 2 Feb 2009 23:43:59 +0000 (23:43 +0000)]
Fix another test to not use -mcpu=yonah with 64-bit code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63572
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 2 Feb 2009 22:50:08 +0000 (22:50 +0000)]
Yonah does not support x86-64. Change the -mcpu value to one that does.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63561
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 2 Feb 2009 22:49:46 +0000 (22:49 +0000)]
More DebugLoc propagation. This should be everything
except LegalizeOp itself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63560
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Stump [Mon, 2 Feb 2009 22:48:49 +0000 (22:48 +0000)]
Improve -fno-opt style option processing to not require an extra
option to make the -fno- form on the option. We also document the new
form in the CommandLine documentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63559
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Mon, 2 Feb 2009 22:42:01 +0000 (22:42 +0000)]
MergeValueInto is too smart: it might choose to do the merge the opposite direction.
Live interval reconstruction needs to account for this, and scour its maps to
prevent dangling references.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63558
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 2 Feb 2009 22:12:50 +0000 (22:12 +0000)]
DebugLoc propagation. ExpandOp and PromoteOp,
among others.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63555
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Mon, 2 Feb 2009 21:57:34 +0000 (21:57 +0000)]
Only force SSE level if it is not correct.
Add an assert to check HasX86_64 status.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63552
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Mon, 2 Feb 2009 21:09:36 +0000 (21:09 +0000)]
Run dsymutil on darwin, when it is expected, before running gdb test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63548
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 2 Feb 2009 20:47:48 +0000 (20:47 +0000)]
Commit missing files.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63545
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 20:44:45 +0000 (20:44 +0000)]
inline SROA::ConvertToScalar, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63544
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 2 Feb 2009 20:41:04 +0000 (20:41 +0000)]
More DebugLoc propagation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63543
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Mon, 2 Feb 2009 20:23:02 +0000 (20:23 +0000)]
remove #if 0 code on Bill's request.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63542
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 2 Feb 2009 19:46:41 +0000 (19:46 +0000)]
Remove trailing spaces.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63540
91177308-0d34-0410-b5e6-
96231b3b80d8
Steve Naroff [Mon, 2 Feb 2009 19:22:02 +0000 (19:22 +0000)]
Update hand-crafted VC++ build.
- Add PACKAGE_STRING definition to Windows config.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63539
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 2 Feb 2009 19:03:57 +0000 (19:03 +0000)]
DebugLoc propagation for int<->fp conversions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63537
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 2 Feb 2009 18:33:18 +0000 (18:33 +0000)]
Refactor PerformPHIConstruction, no functionality changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63536
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 18:15:33 +0000 (18:15 +0000)]
xfail this for now, will fix shortly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63533
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 18:12:58 +0000 (18:12 +0000)]
update test
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63532
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 2 Feb 2009 18:04:24 +0000 (18:04 +0000)]
Reposition the DebugLoc field to reduce padding space
on LP64 hosts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63530
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 18:02:59 +0000 (18:02 +0000)]
Fix a bug which caused us to miscompile a couple of Ada
tests. Thanks for the beautiful reduced testcase Duncan!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63529
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Mon, 2 Feb 2009 17:51:41 +0000 (17:51 +0000)]
Do not add redundant arguments in a method definition DIE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63527
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Mon, 2 Feb 2009 17:50:43 +0000 (17:50 +0000)]
Make this test case smaller.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63526
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjiv Gupta [Mon, 2 Feb 2009 17:19:39 +0000 (17:19 +0000)]
Duncan's patch. Further to 64382. Takes care of illegal types for shift amount.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63523
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjiv Gupta [Mon, 2 Feb 2009 16:53:06 +0000 (16:53 +0000)]
Made the common case of default address space directive as non-virtual for performance reasons. Provide a single virtual interface for directives of all sizes in non-default address spaces.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63521
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 2 Feb 2009 10:06:20 +0000 (10:06 +0000)]
Fix a comment (bytes -> bits), reformat a comment
and remove trailing whitespace. No functionality
change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63511
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 2 Feb 2009 09:53:14 +0000 (09:53 +0000)]
Fix an obvious thinko.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63510
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 2 Feb 2009 09:15:04 +0000 (09:15 +0000)]
ADD / SUB / SMUL / UMUL with overflow second result top bits must be zero.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63509
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 2 Feb 2009 09:10:57 +0000 (09:10 +0000)]
This passes on x86-32 linux at least.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63508
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 2 Feb 2009 09:07:13 +0000 (09:07 +0000)]
Make the XFAIL line actually match x86-32 targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63507
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 2 Feb 2009 08:19:07 +0000 (08:19 +0000)]
Add comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63506
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 2 Feb 2009 08:07:36 +0000 (08:07 +0000)]
Teach LowerBRCOND to recognize (xor (setcc x), 1). The xor inverts the condition. It's normally transformed by the dag combiner, unless the condition is set by a arithmetic op with overflow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63505
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 07:40:17 +0000 (07:40 +0000)]
reject things like: zext <4 x i32> %tmp to i256
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63504
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 07:33:15 +0000 (07:33 +0000)]
validation error fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63503
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 07:32:36 +0000 (07:32 +0000)]
Document type upreferences (PR3380), patch by Stein Roger Skafløtten
with enhancements and corrections by me.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63502
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 07:24:28 +0000 (07:24 +0000)]
Fix PR3372
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63501
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 07:15:30 +0000 (07:15 +0000)]
reduce indentation, (~XorCST->getValue()).isSignBit() -> isMaxSignedValue()
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63500
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 2 Feb 2009 06:55:45 +0000 (06:55 +0000)]
reduce testcase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63499
91177308-0d34-0410-b5e6-
96231b3b80d8
Mon P Wang [Mon, 2 Feb 2009 06:37:55 +0000 (06:37 +0000)]
Preserve more SourceValue information.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63498
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sun, 1 Feb 2009 18:24:20 +0000 (18:24 +0000)]
add 2 more testcases for -mattr=-sse (r63495).
--This line, and those below, will be ignaored--
A test/CodeGen/X86/nosse-error1.ll
A test/CodeGen/X86/nosse-error2.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63496
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sun, 1 Feb 2009 18:15:56 +0000 (18:15 +0000)]
Implement -mno-sse: if SSE is disabled on x86-64, don't store XMM on stack for
var-args, and don't allow FP return values
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63495
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sun, 1 Feb 2009 18:06:53 +0000 (18:06 +0000)]
Fix PR3453 and probably a bunch of other potential
crashes or wrong code with codegen of large integers:
eliminate the legacy getIntegerVTBitMask and
getIntegerVTSignBit methods, which returned their
value as a uint64_t, so couldn't handle huge types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63494
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 1 Feb 2009 11:19:36 +0000 (11:19 +0000)]
Forgot some more DebugLoc propagations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63493
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Sun, 1 Feb 2009 08:41:54 +0000 (08:41 +0000)]
Fix test failures causes by my previous commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63492
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Sun, 1 Feb 2009 07:06:00 +0000 (07:06 +0000)]
Fix an issue in PHI construction that was exposed by GCC 4.2 producing a different set iteration order for the reg_iterator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63490
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sun, 1 Feb 2009 06:42:27 +0000 (06:42 +0000)]
Fix PR3423: Link llvm on ARM EABI machines. Patch by Robert Schuster.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63489
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Sat, 31 Jan 2009 22:04:51 +0000 (22:04 +0000)]
DebugLoc propagation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63488
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sat, 31 Jan 2009 21:30:05 +0000 (21:30 +0000)]
Reinstate this optimization to fold icmp of xor when possible. Don't try to
turn icmp eq a+x, b+x into icmp eq a, b if a+x or b+x has other uses. This
may have been increasing register pressure leading to the bzip2 slowdown.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63487
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Sat, 31 Jan 2009 21:04:24 +0000 (21:04 +0000)]
DebugLoc propagation. Done with file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63486
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Sat, 31 Jan 2009 20:01:02 +0000 (20:01 +0000)]
DebugLoc propagation. Done with file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63485
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 19:05:27 +0000 (19:05 +0000)]
Fix PR3452 (an infinite loop bootstrapping) by disabling the recent
improvements to the EvaluateInDifferentType code. This code works
by just inserted a bunch of new code and then seeing if it is
useful. Instcombine is not allowed to do this: it can only insert
new code if it is useful, and only when it is converging to a more
canonical fixed point. Now that we iterate when DCE makes progress,
this causes an infinite loop when the code ends up not being used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63483
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sat, 31 Jan 2009 15:50:11 +0000 (15:50 +0000)]
Fix PR3401: when using large integers, the type
returned by getShiftAmountTy may be too small
to hold shift values (it is an i8 on x86-32).
Before and during type legalization, use a large
but legal type for shift amounts: getPointerTy;
afterwards use getShiftAmountTy, fixing up any
shift amounts with a big type during operation
legalization. Thanks to Dan for writing the
original patch (which I shamelessly pillaged).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63482
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 08:40:03 +0000 (08:40 +0000)]
now that all the pieces are in place, teach instcombine's
simplifydemandedbits to simplify instructions with *multiple
uses* in contexts where it can get away with it. This allows
it to simplify the code in multi-use-or.ll into a single 'add
double'.
This change is particularly interesting because it will cover
up for some common codegen bugs with large integers created due
to the recent SROA patch. When working on fixing those bugs,
this should be disabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63481
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 08:24:16 +0000 (08:24 +0000)]
simplify/clarify control flow and improve comments, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63480
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 08:15:18 +0000 (08:15 +0000)]
make some fairly meaty internal changes to how SimplifyDemandedBits works.
Now, if it detects that "V" is the same as some other value,
SimplifyDemandedBits returns the new value instead of RAUW'ing it immediately.
This has two benefits:
1) simpler code in the recursive SimplifyDemandedBits routine.
2) it allows future fun stuff in instcombine where an operation has multiple
uses and can be simplified in one context, but not all.
#2 isn't implemented yet, this patch should have no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63479
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 07:34:19 +0000 (07:34 +0000)]
add accessors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63478
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 07:26:06 +0000 (07:26 +0000)]
minor cleanups
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63477
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 07:04:22 +0000 (07:04 +0000)]
make sure to set Changed=true when instcombine hacks on the code,
not doing so prevents it from properly iterating and prevents it
from deleting the entire body of dce-iterate.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63476
91177308-0d34-0410-b5e6-
96231b3b80d8
Mon P Wang [Sat, 31 Jan 2009 06:50:54 +0000 (06:50 +0000)]
Used "-enable-unsafe-fp-math" to allow this transformation - (a * b -c) = c - a *b.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63475
91177308-0d34-0410-b5e6-
96231b3b80d8
Mon P Wang [Sat, 31 Jan 2009 06:07:45 +0000 (06:07 +0000)]
If unsafe FP optimization is not set, don't allow -(A-B) => B-A because
when A==B, -0.0 != +0.0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63474
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 31 Jan 2009 03:12:48 +0000 (03:12 +0000)]
Don't use DebugLoc::getUnknownLoc(). Default to something hopefully sensible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63473
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 31 Jan 2009 02:28:54 +0000 (02:28 +0000)]
Simplify and generalize the SROA "convert to scalar" transformation to
be able to handle *ANY* alloca that is poked by loads and stores of
bitcasts and GEPs with constant offsets. Before the code had a number
of annoying limitations and caused it to miss cases such as storing into
holes in structs and complex casts (as in bitfield-sroa) where we had
unions of bitfields etc. This also handles a number of important cases
that are exposed due to the ABI lowering stuff we do to pass stuff by
value.
One case that is pretty great is that we compile
2006-11-07-InvalidArrayPromote.ll into:
define i32 @func(<4 x float> %v0, <4 x float> %v1) nounwind {
%tmp10 = call <4 x i32> @llvm.x86.sse2.cvttps2dq(<4 x float> %v1)
%tmp105 = bitcast <4 x i32> %tmp10 to i128
%tmp1056 = zext i128 %tmp105 to i256
%tmp.upgrd.43 = lshr i256 %tmp1056, 96
%tmp.upgrd.44 = trunc i256 %tmp.upgrd.43 to i32
ret i32 %tmp.upgrd.44
}
which turns into:
_func:
subl $28, %esp
cvttps2dq %xmm1, %xmm0
movaps %xmm0, (%esp)
movl 12(%esp), %eax
addl $28, %esp
ret
Which is pretty good code all things considering :).
One effect of this is that SROA will start generating arbitrary bitwidth
integers that are a multiple of 8 bits. In the case above, we got a
256 bit integer, but the codegen guys assure me that it can handle the
simple and/or/shift/zext stuff that we're doing on these operations.
This addresses rdar://
6532315
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63469
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Sat, 31 Jan 2009 02:22:37 +0000 (02:22 +0000)]
Move CurDebugLoc into SelectionDAGLowering.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63468
91177308-0d34-0410-b5e6-
96231b3b80d8