1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-28 06:22:51 +01:00
Commit Graph

44377 Commits

Author SHA1 Message Date
Chris Lattner
60dcb2b5c2 reduce indentation and improve comments, no functionality change.
llvm-svn: 122389
2010-12-22 07:36:50 +00:00
Wesley Peck
2759b7bc98 Don't generate carry bit when loading immediate values on the Microblaze.
llvm-svn: 122385
2010-12-22 01:29:32 +00:00
Wesley Peck
0d64db9772 Add support for some of the LLVM atomic operations to the MBlaze backend.
llvm-svn: 122384
2010-12-22 01:15:01 +00:00
Wesley Peck
32c95e5ef4 Modeling the carry bit in the MSR register of the MicroBlaze.
llvm-svn: 122381
2010-12-22 00:53:07 +00:00
Wesley Peck
fc577ca406 Fix a regression introduced into the MBlaze delay slot filler.
llvm-svn: 122379
2010-12-22 00:22:59 +00:00
Owen Anderson
b4f1511864 Give GVN back the ability to perform simple conditional propagation on conditional branch values.
I still think that LVI should be handling this, but that capability is some ways off in the future,
and this matters for some significant benchmarks.

llvm-svn: 122378
2010-12-21 23:54:34 +00:00
Matt Beaumont-Gay
b03146b50a GCC objects to the two sides of a conditional expression having different enum
types, but they're just getting converted to unsigned anyway, so cast first
(and ask questions later).

llvm-svn: 122377
2010-12-21 23:43:23 +00:00
Owen Anderson
1ee218de0e Remove dead code.
llvm-svn: 122371
2010-12-21 22:31:24 +00:00
Andrew Trick
afec190a28 In DelayForLiveRegsBottomUp, handle instructions that read and write
the same physical register. Simplifies the fix from the previous
checkin r122211.

llvm-svn: 122370
2010-12-21 22:27:44 +00:00
Andrew Trick
1e3ad9f721 whitespace
llvm-svn: 122368
2010-12-21 22:25:04 +00:00
Dale Johannesen
e0fb87c3d7 Reapply 122353-122355 with fixes. 122354 was wrong;
the shift type was needed one place, the shift count
type another.  The transform in 123555 had the same
problem.

llvm-svn: 122366
2010-12-21 21:55:50 +00:00
Benjamin Kramer
369872edfc Add some x86 specific dagcombines for conditional increments.
(add Y, (sete  X, 0)) -> cmp X, 1; adc  0, Y
(add Y, (setne X, 0)) -> cmp X, 1; sbb -1, Y
(sub (sete  X, 0), Y) -> cmp X, 1; sbb  0, Y
(sub (setne X, 0), Y) -> cmp X, 1; adc -1, Y

for
  unsigned foo(unsigned a, unsigned b) {
    if (a == 0) b++;
    return b;
  }
we now get:
  foo:
    cmpl  $1, %edi
    movl  %esi, %eax
    adcl  $0, %eax
    ret
instead of:
  foo:
    testl %edi, %edi
    sete  %al
    movzbl  %al, %eax
    addl  %esi, %eax
    ret

llvm-svn: 122364
2010-12-21 21:41:44 +00:00
Benjamin Kramer
fd119319ea GVN's Expression is not POD-like (it contains a SmallVector). Simplify code while at it.
llvm-svn: 122362
2010-12-21 21:30:19 +00:00
Dale Johannesen
972aba543a Revert 122353-122355 for the moment, they broke stuff.
llvm-svn: 122360
2010-12-21 21:22:27 +00:00
Rafael Espindola
a1091c31ba Simplify EvaluateAsAbsolute now that EvaluateAsRelocatableImpl does all
the folding it can.

llvm-svn: 122359
2010-12-21 20:51:42 +00:00
Rafael Espindola
3d9afa3c44 Don't relax org or align. They change size as the relaxation happens, but they
are not actually relaxed. For example, a section with only alignments will never
needs relaxation.

llvm-svn: 122356
2010-12-21 20:35:18 +00:00
Dale Johannesen
39186cfb0b Add a new transform to DAGCombiner.
llvm-svn: 122355
2010-12-21 20:10:51 +00:00
Dale Johannesen
5f3e7b08f6 Get the type of a shift from the shift, not from its shift
count operand.  These should be the same but apparently are
not always, and this is cleaner anyway.  This improves the
code in an existing test.

llvm-svn: 122354
2010-12-21 20:06:19 +00:00
Dale Johannesen
bad19334ee Shift by the word size is invalid IR; don't create it.
llvm-svn: 122353
2010-12-21 20:00:06 +00:00
Chris Lattner
8a3058137a fix some typos
llvm-svn: 122349
2010-12-21 18:05:22 +00:00
Stuart Hastings
fedc21e594 Fix indentation, add comment.
llvm-svn: 122345
2010-12-21 17:16:58 +00:00
Duncan Sands
c7868f60bc Visit instructions deterministically. Use a FIFO so as to approximately
visit instructions before their uses, since InstructionSimplify does a
better job in that case.  All this prompted by Frits van Bommel.

llvm-svn: 122343
2010-12-21 17:08:55 +00:00
Stuart Hastings
a1f786efa9 Missing logic for nested CALLSEQ_START/END.
llvm-svn: 122342
2010-12-21 17:07:24 +00:00
Duncan Sands
6e2fd37144 If an instruction simplifies, try again to simplify any uses of it. This is
not very important since the pass is only used for testing, but it does make
it more realistic.  Suggested by Frits van Bommel.

llvm-svn: 122336
2010-12-21 16:12:03 +00:00
Daniel Dunbar
1c98bbd915 MC/Mach-O: Shuffle enums a bit to make it harder to inadvertently use the wrong
type.

llvm-svn: 122334
2010-12-21 15:26:45 +00:00
Duncan Sands
b4497c7e0f While I don't think any later transforms can fire, it seems cleaner to
not assume this (for example in case more transforms get added below
it).  Suggested by Frits van Bommel.

llvm-svn: 122332
2010-12-21 15:03:43 +00:00
Duncan Sands
7e94cad96b Fix inverted condition noticed by Frits van Bommel.
llvm-svn: 122331
2010-12-21 14:48:48 +00:00
Duncan Sands
6b0ed93a81 Pull a few more simplifications out of instcombine (there are still
plenty left though!), in particular for multiplication.

llvm-svn: 122330
2010-12-21 14:00:22 +00:00
Duncan Sands
0bd25425b6 Teach InstructionSimplify about distributive laws. These transforms fire
quite often, but don't make much difference in practice presumably because
instcombine also knows them and more.

llvm-svn: 122328
2010-12-21 13:32:22 +00:00
Duncan Sands
2ca8f72575 Move checking of the recursion limit into the various Thread methods.
No functionality change.

llvm-svn: 122327
2010-12-21 09:09:15 +00:00
Duncan Sands
5880f299da Add generic simplification of associative operations, generalizing
a couple of existing transforms.  This fires surprisingly often, for
example when compiling gcc "(X+(-1))+1->X" fires quite a lot as well
as various "and" simplifications (usually with a phi node operand).
Most of the time this doesn't make a real difference since the same
thing would have been done elsewhere anyway, eg: by instcombine, but
there are a few places where this results in simplifications that we
were not doing before.

llvm-svn: 122326
2010-12-21 08:49:00 +00:00
Cameron Zwarich
0243f1d21e Incremental progress towards a new implementation of StrongPHIElimination. Most
of the problems with my last attempt were in the updating of LiveIntervals
rather than the coalescing itself. Therefore, I decided to get that right first
by essentially reimplementing the existing PHIElimination using LiveIntervals.

It works correctly, with only a few tests failing (which may not be legitimate
failures) and no new verifier failures (at least as far as I can tell, I didn't
count the number per file).

llvm-svn: 122321
2010-12-21 06:54:43 +00:00
Bob Wilson
01593c55a2 Add ARM-specific DAG combining to cast i64 vector element load/stores to f64.
Type legalization splits up i64 values into pairs of i32 values, which leads
to poor quality code when inserting or extracting i64 vector elements.
If the vector element is loaded or stored, it can be treated as an f64 value
and loaded or stored directly from a VPR register.  Use the pre-legalization
DAG combiner to cast those vector elements to f64 types so that the type
legalizer won't mess them up.  Radar 8755338.

llvm-svn: 122319
2010-12-21 06:43:19 +00:00
Rafael Espindola
e0e966982a Layout one section until no relaxations are done and then move to the next
section.

This helps because in practice sections form a dag with debug sections pointing
to text sections. Finishing up the text sections first makes the debug section
relaxation trivial.

llvm-svn: 122314
2010-12-21 04:22:09 +00:00
Eric Christopher
72ceef0a74 Arm and thumb call instructions are also in different orders.
Fixes rdar://8782223

llvm-svn: 122313
2010-12-21 03:50:43 +00:00
Chris Lattner
65c5243bd6 rename MVT::Flag to MVT::Glue. "Flag" is a terrible name for
something that just glues two nodes together, even if it is
sometimes used for flags.

llvm-svn: 122310
2010-12-21 02:38:05 +00:00
Eric Christopher
81ae56b33c If we're not using reg+reg offset we're using reg+imm, set the opcode
to be the one we want to use. bugpoint reduced testcase is a little large,
I'll see if I can simplify it down more.

Fixes part of rdar://8782207

llvm-svn: 122307
2010-12-21 02:12:07 +00:00
Chris Lattner
b37e697277 improve "cannot yet select" errors a trivial amount: now
they are just as useless, but at least a bit more gramatical

llvm-svn: 122305
2010-12-21 02:07:03 +00:00
Bill Wendling
856080c8a1 Fix a copy-pasto. When the tBR_JTr instruction was converted to using the
tPseudoInst class, its size was changed from "special" to "2 bytes". This is
incorrect because the jump table will no longer be taken into account when
calculating branch offsets.
<rdar://problem/8782216>

llvm-svn: 122303
2010-12-21 01:57:15 +00:00
Bill Wendling
8c7f90099b Comment cleanups.
llvm-svn: 122302
2010-12-21 01:54:40 +00:00
Jakob Stoklund Olesen
e9eb1be4dd Add EdgeBundles to SplitKit.
Edge bundles is an annotation on the CFG that turns it into a bipartite directed
graph where each basic block is connected to an outgoing and an ingoing bundle.
These bundles are useful for identifying regions of the CFG for live range
splitting.

llvm-svn: 122301
2010-12-21 01:50:21 +00:00
Jakob Stoklund Olesen
86786c46c2 Use IntEqClasses to compute connected components of live intervals.
llvm-svn: 122296
2010-12-21 00:48:17 +00:00
Jakob Stoklund Olesen
26b4e8ec0a Add ADT/IntEqClasses.h as a light-weight implementation of EquivalenceClasses.h.
This implementation already exists as ConnectedVNInfoEqClasses in
LiveInterval.cpp, and it seems to be generally useful to have a light-weight way
of forming equivalence classes of small integers.

IntEqClasses doesn't allow enumeration of the elements in a class.

llvm-svn: 122293
2010-12-21 00:04:46 +00:00
Owen Anderson
308ca4e488 Speculatively revert the use of DenseMap in LazyValueInfo, which may be causing Linux self-host failures.
llvm-svn: 122291
2010-12-20 23:53:19 +00:00
Owen Anderson
b481f13b97 Attempt to appease the DragonEgg buildbots.
llvm-svn: 122288
2010-12-20 23:23:18 +00:00
Owen Anderson
c1eee5ac49 Revert r122114 (CallbackVH observing use-list changes) because it caused severe slowdowns on the Linux self-host configuration.
llvm-svn: 122279
2010-12-20 22:28:03 +00:00
Nate Begeman
c7dfecb10e Implement feedback from Bruno on making pblendvb an x86-specific ISD node in addition to being an intrinsic, and convert
lowering to use it.  Hopefully the pattern fragment is doing the right thing with XMM0, looks correct in testing.

llvm-svn: 122277
2010-12-20 22:04:24 +00:00
Owen Anderson
006186c519 Convert one of LVI's primary maps to a DenseMap, now that we know are more assured of iterator stability.
llvm-svn: 122273
2010-12-20 21:30:54 +00:00
Wesley Peck
e8ec7a4d1f Teach the MBlaze disassembler to disassemble special purpose registers.
llvm-svn: 122269
2010-12-20 21:18:04 +00:00
Roman Divacky
42b3eee794 Set the value of absolute symbols.
llvm-svn: 122268
2010-12-20 21:14:39 +00:00