Commit Graph

10991 Commits

Author SHA1 Message Date
Jakob Stoklund Olesen
cfa7134a9c Basic rematerialization during splitting.
Whenever splitting wants to insert a copy, it checks if the value can be
rematerialized cheaply instead.

Missing features:
- Delete instructions when all uses have been rematerialized.
- Truncate live ranges to the remaining uses after rematerialization.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118702 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-10 19:31:50 +00:00
Andrew Trick
f4baeaf848 RABasic is nearly functionally complete. There are a few remaining
benchmarks hitting an assertion.
Adds LiveIntervalUnion::collectInterferingVRegs.
Fixes "late spilling" by checking for any unspillable live vregs among
all physReg aliases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118701 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-10 19:18:47 +00:00
Jakob Stoklund Olesen
b80e973c95 Simplify the LiveRangeEdit::canRematerializeAt() interface a bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118661 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-10 01:05:12 +00:00
Rafael Espindola
34be396a12 Fixed version of 118639 with an extra assert to catch similar problems
earlier. Implicit bool -> int conversions are evil!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118651 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 23:42:07 +00:00
Andrew Trick
071d1c063f Adds RABasic verification and tracing.
(retry now that the windows build is green)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118630 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 21:04:34 +00:00
Matt Beaumont-Gay
e33daaaaf9 Add a trivial virtual dtor to AbstractRegisterDescription to appease
-Wnon-virtual-dtor.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118616 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 19:56:25 +00:00
Andrew Trick
32d9aeb1df Reverting r118604. Windows build broke.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118613 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 19:47:51 +00:00
Andrew Trick
39a5ce46a5 Adds RABasic verification and tracing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118604 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 19:01:17 +00:00
Dan Gohman
6a559cd695 Fix DAGCombiner to avoid folding a sext-in-reg or similar through a shl
in order to fold it into a load.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118471 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 01:54:35 +00:00
Dale Johannesen
71365d3774 Fix an inline asm pasto from 117667; was preventing
{i64, i64} from matching i128.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118465 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 01:15:07 +00:00
Andrew Trick
e141a4960f Adds support for spilling previously allocated live intervals to
handle cases in which a register is unavailable for spill code.
Adds LiveIntervalUnion::extract. While processing interferences on a
live virtual register, reuses the same Query object for each
physcial reg.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118423 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-08 18:02:08 +00:00
Che-Liang Chiou
6da24ca51d Add registry hook for assembly text output
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118394 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-08 02:21:17 +00:00
Benjamin Kramer
f788854d20 Prune includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118342 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-06 11:45:59 +00:00
Duncan Sands
69300a2f47 When passing a parameter using the 'byval' mechanism, inline code needs to be used
to perform the copy, which may be of lots of memory [*].  It would be good if the
fall-back code generated something reasonable, i.e. did the copy in a loop, rather
than vast numbers of loads and stores.  Add a note about this.  Currently target
specific code seems to always kick in so this is more of a theoretical issue rather
than a practical one now that X86 has been fixed.
[*] It's amazing how often people pass mega-byte long arrays by copy...


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118275 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-05 15:20:29 +00:00
Rafael Espindola
7387345016 Add 118023 back, but with proper spelling for .uleb128/.sleb128.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118254 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-04 18:17:08 +00:00
Rafael Espindola
537ebeaffc Revert previous patch. Some targets don't support uleb and say
they do :-(

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118250 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-04 17:04:24 +00:00
Rafael Espindola
40f6514ff0 MCize.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118249 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-04 16:32:18 +00:00
Duncan Sands
1e96bab329 In the calling convention logic, ValVT is always a legal type,
and as such can be represented by an MVT - the more complicated
EVT is not needed.  Use MVT for ValVT everywhere.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118245 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-04 10:49:57 +00:00
Jakob Stoklund Olesen
8c42f48d02 Disable fancy splitting during spilling unless -extra-spiller-splits is given.
This way, InlineSpiller does the same amount of splitting as the standard
spiller. Splitting should really be guided by the register allocator, and
doesn't belong in the spiller at all.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118216 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-04 00:32:32 +00:00
Eric Christopher
d81f17acb4 Just return undef for invalid masks or elts, and since we're doing that,
just do it earlier too.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118195 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 20:44:42 +00:00
Jakob Stoklund Olesen
964bc25e5a Let RegAllocBasic require MachineDominators - they are already available and
splitting needs them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118194 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 20:39:26 +00:00
Jakob Stoklund Olesen
376dcbd6c2 Tag debug output as regalloc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118193 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 20:39:23 +00:00
Duncan Sands
cdfad36b40 Simplify uses of MVT and EVT. An MVT can be compared directly
with a SimpleValueType, while an EVT supports equality and
inequality comparisons with SimpleValueType.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118169 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 12:17:33 +00:00
Duncan Sands
1440e8b918 Inside the calling convention logic LocVT is always a simple
value type, so there is no point in passing it around using
an EVT.  Use the simpler MVT everywhere.  Rather than trying
to propagate this information maximally in all the code that
using the calling convention stuff, I chose to do a mainly
low impact change instead.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118167 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 11:35:31 +00:00
Eric Christopher
caebdd44f3 If we have an undef mask our Elt will be -1 for our access, handle
this by using an undef as a pointer.

Fixes rdar://8625016


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118164 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 09:36:40 +00:00
Dan Gohman
394d6298bc Fix DAGCombiner to avoid going into an infinite loop when it
encounters (and:i64 (shl:i64 (load:i64), 1), 0xffffffff).
This fixes rdar://8606584.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118143 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 01:47:46 +00:00
Evan Cheng
8239daf7c8 Two sets of changes. Sorry they are intermingled.
1. Fix pre-ra scheduler so it doesn't try to push instructions above calls to
   "optimize for latency". Call instructions don't have the right latency and
   this is more likely to use introduce spills.
2. Fix if-converter cost function. For ARM, it should use instruction latencies,
   not # of micro-ops since multi-latency instructions is completely executed
   even when the predicate is false. Also, some instruction will be "slower"
   when they are predicated due to the register def becoming implicit input.
   rdar://8598427


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118135 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-03 00:45:17 +00:00
Andrew Trick
4638852696 Fixes <rdar://problem/8612856>: During postRAsched, the antidependence
breaker needs to check all definitions of the antidepenent register to
avoid multiple defs of the same new register.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118032 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-02 18:16:45 +00:00
Devang Patel
c8821046f8 Simplify.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118027 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-02 17:37:00 +00:00
Devang Patel
8bc9ef77b7 If value map does not have register for an argument then try to find frame index before giving up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118022 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-02 17:19:03 +00:00
Devang Patel
a90b305af9 Use frameindex, if available, as a last resort to emit debug info for a parameter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118020 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-02 17:01:30 +00:00
Jakob Stoklund Olesen
371e82bf51 Don't try to split weird critical edges that really aren't:
BB#1: derived from LLVM BB %bb.nph28
    Live Ins: %AL
    Predecessors according to CFG: BB#0
	TEST8rr %reg16384<kill>, %reg16384, %EFLAGS<imp-def>; GR8:%reg16384
	JNE_4 <BB#2>, %EFLAGS<imp-use,kill>
	JMP_4 <BB#2>
    Successors according to CFG: BB#2 BB#2

These double CFG edges only ever occur in bugpoint-generated code, so there is
no need to attempt something clever.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117992 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-02 00:58:37 +00:00
Jakob Stoklund Olesen
8bcf7603f7 MachineLICM should not claim to be preserving the CFG when it can split critical
edges on demand.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117982 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 23:59:55 +00:00
Jakob Stoklund Olesen
962c71089d Be more precise about verifying missing kill flags.
It is legal for an instruction to have two operands using the same register,
only one a kill. This is interpreted as a kill.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117981 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 23:59:53 +00:00
Jakob Stoklund Olesen
3d4114c464 When inserting copies during splitting, always use the parent register as the
source, and let rewrite() clean it up.

This way, kill flags on the inserted copies are fixed as well during rewrite().

We can't just assume that all the copies we insert are going to be kills since
critical edges into loop headers sometimes require both source and dest to be
live out of a block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117980 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 23:59:48 +00:00
Jakob Stoklund Olesen
1c163d2a06 Add kill flag verification.
At least X86FloatingPoint requires correct kill flags after register allocation,
and targets using register scavenging benefit. Conservative kill flags are not
enough.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117960 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 21:51:31 +00:00
Jakob Stoklund Olesen
79cb53ce95 Update kill flags while rewriting instructions after splitting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117959 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 21:51:29 +00:00
Bill Wendling
40a5eb18b0 When we look at instructions to convert to setting the 's' flag, we need to look
at more than those which define CPSR. You can have this situation:

(1)  subs  ...
(2)  sub   r6, r5, r4
(3)  movge ...
(4)  cmp   r6, 0
(5)  movge ...

We cannot convert (2) to "subs" because (3) is using the CPSR set by
(1). There's an analogous situation here:

(1)  sub   r1, r2, r3
(2)  sub   r4, r5, r6
(3)  cmp   r4, ...
(5)  movge ...
(6)  cmp   r1, ...
(7)  movge ...

We cannot convert (1) to "subs" because of the intervening use of CPSR.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117950 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 20:41:43 +00:00
Jakob Stoklund Olesen
a37d5cf342 Don't assign new registers created during a split to the same stack slot, but
give them individual stack slots once the are actually spilled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117945 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 19:49:57 +00:00
Jakob Stoklund Olesen
e8f0823a68 Add basic LiveStacks verification.
When an instruction refers to a spill slot with a LiveStacks entry, check that
the spill slot is live at the instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117944 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 19:49:52 +00:00
Bill Wendling
9e7cb3d243 The testcase is now XFAILed. Sorry about the breakage.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117904 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-01 05:50:55 +00:00
Eric Christopher
b698a56dd4 Revert r117876 for now, it's causing more testsuite failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117879 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-31 22:42:55 +00:00
Bill Wendling
b3747706fc Disable the peephole optimizer until 186.crafty on armv6 is fixed. This is what
looks like is happening:

Without the peephole optimizer:
  (1)   sub     r6, r6, #32
        orr     r12, r12, lr, lsl r9
        orr     r2, r2, r3, lsl r10
  (x)   cmp     r6, #0
        ldr     r9, LCPI2_10
        ldr     r10, LCPI2_11
  (2)   sub     r8, r8, #32
  (a)   movge   r12, lr, lsr r6
  (y)   cmp     r8, #0
LPC2_10:
        ldr     lr, [pc, r10]
  (b)   movge   r2, r3, lsr r8

With the peephole optimizer:
        ldr     r9, LCPI2_10
        ldr     r10, LCPI2_11
  (1*)  subs    r6, r6, #32
  (2*)  subs    r8, r8, #32
  (a*)  movge   r12, lr, lsr r6
  (b*)  movge   r2, r3, lsr r8

(1) is used by (x) for the conditional move at (a). (2) is used by (y) for the
conditional move at (b). After the peephole optimizer, these the flags resulting
from (1*) are ignored and only the flags from (2*) are considered for both
conditional moves.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117876 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-31 22:07:12 +00:00
Nicolas Geoffray
7a5a3f75eb Attach a GCModuleInfo to a MachineFunction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117867 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-31 20:38:38 +00:00
Jakob Stoklund Olesen
e5f4e9fb2b Include MachineBasicBlock numbers in viewCFG() output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117765 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-30 01:26:19 +00:00
Jakob Stoklund Olesen
5b4bb218d1 Make sure copies are inserted after any exception handling labels at the top of
a basic block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117764 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-30 01:26:16 +00:00
Jakob Stoklund Olesen
92095e7b3f Add SkipPHIsAndLabels from PHIElimination to MachineBasicBlock. It is needed
elsewhere.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117763 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-30 01:26:14 +00:00
Jakob Stoklund Olesen
ab56647e6b Disable more of physical register live intervals verification.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117762 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-30 01:26:11 +00:00
Jakob Stoklund Olesen
7d57753df3 Print out register class of spilled register.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117761 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-30 01:26:09 +00:00
Evan Cheng
189c1ec4c1 Teach machine cse to eliminate instructions with multiple physreg uses and defs. rdar://8610857.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117745 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 23:36:03 +00:00
Bob Wilson
46455096be Remove DAG combiner patch to fold vector splats. Instcombiner does it now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117720 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 22:03:02 +00:00
Jakob Stoklund Olesen
db4eec3d2c Fix sign error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117677 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 18:21:18 +00:00
Evan Cheng
089751535d Avoiding overly aggressive latency scheduling. If the two nodes share an
operand and one of them has a single use that is a live out copy, favor the
one that is live out. Otherwise it will be difficult to eliminate the copy
if the instruction is a loop induction variable update. e.g.

BB:
sub r1, r3, #1
str r0, [r2, r3]
mov r3, r1
cmp
bne BB

=>

BB:
str r0, [r2, r3]
sub r3, r3, #1
cmp
bne BB

This fixed the recent 256.bzip2 regression.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117675 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 18:09:28 +00:00
Jakob Stoklund Olesen
3ccfce0e35 Don't transfer unused values to the new intervals formed by splitting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117673 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 17:47:49 +00:00
Benjamin Kramer
8a8e26fb18 Silence Release build warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117671 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 17:40:05 +00:00
Jakob Stoklund Olesen
6d309059a7 Teach ConnectedVNInfoEqClasses::Classify to deal with unused values.
We don't want unused values forming their own equivalence classes, so we lump
them all together in one class, and then merge them with the class of the last
used value.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117670 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 17:37:29 +00:00
Jakob Stoklund Olesen
c94fcb1507 Never propagate the idom value out of a block that defines its own value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117669 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 17:37:25 +00:00
John Thompson
44ab89eb37 Inline asm multiple alternative constraints development phase 2 - improved basic logic, added initial platform support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117667 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 17:29:13 +00:00
Bill Wendling
49ad731adf This may be an ARM target, so check for _Unwind_SjLj_Resume.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117643 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 07:46:01 +00:00
Jakob Stoklund Olesen
eeaae20be0 Fix broken equivalence class calculation. We could probably also use
EquvivalenceClasses.h except it looks like overkill when elements are continuous
integers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117631 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 00:40:59 +00:00
Jakob Stoklund Olesen
cb367778c0 Print out the connected components in the verifier after complaining about their
multiplicity.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117630 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 00:40:57 +00:00
Jakob Stoklund Olesen
1f46a0ac66 Run a verification pass before any splitting to better distribute blame.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117629 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-29 00:40:55 +00:00
Devang Patel
4d7f9a00a5 Ignore empty blocks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117615 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 22:11:59 +00:00
Jakob Stoklund Olesen
8c4564287d One day, physical register live ranges will be sensible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117602 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 20:44:22 +00:00
Jakob Stoklund Olesen
e1dde7b05a Replace SplitKit SSA update with an iterative algorithm very similar to the one
in SSAUpdaterImpl.h

Verifying live intervals revealed that the old method was completely wrong, and
we need an iterative approach to calculating PHI placemant. Fortunately, we have
MachineDominators available, so we don't have to compute that over and over
like SSAUpdaterImpl.h must.

Live-out values are cached between calls to mapValue() and computed in a greedy
way, so most calls will be working with very small block sets.

Thanks to Bob for explaining how this should work.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117599 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 20:34:52 +00:00
Jakob Stoklund Olesen
d68f458244 Make MachineDominators available for SplitEditor. We are going to need it for
proper SSA updating.

This doesn't cause MachineDominators to be recomputed since we are already
requiring MachineLoopInfo which uses dominators as well.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117598 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 20:34:50 +00:00
Jakob Stoklund Olesen
26b92be922 Add a temporary command line option to verify machine code after each spill or
split.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117597 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 20:34:47 +00:00
Devang Patel
bc2bb9bbd0 Do not work too hard to find type's file info. There is a special field to record file info.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117588 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 19:50:08 +00:00
Devang Patel
3cabc9d2c9 Technically DIFile scope should also be handled here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117563 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 17:30:52 +00:00
Bob Wilson
0f1db1a6c6 Teach the DAG combiner to fold a splat of a splat. Radar 8597790.
Also do some minor refactoring to reduce indentation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117558 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 17:06:14 +00:00
Evan Cheng
7e2fe9150f Re-commit 117518 and 117519 now that ARM MC test failures are out of the way.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117531 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 06:47:08 +00:00
Evan Cheng
9e08ee5d16 Revert 117518 and 117519 for now. They changed scheduling and cause MC tests to fail. Ugh.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117520 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 02:00:25 +00:00
Evan Cheng
7c88cdcc3b Fix a major bug in operand latency computation. The use index must be adjusted
by the number of defs first for it to match the instruction itinerary.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117518 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-28 01:46:29 +00:00
Evan Cheng
de5fa932b9 Putting r117193 back except for the compile time cost. Rather than assuming fallthroughs uses all registers, just gather the union of all successor liveins.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117506 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 23:17:17 +00:00
Michael J. Spencer
3931b54a5f COFF: Add IMAGE_SCN_MEM_READ to text sections.
There are currently 100 references to COFF::IMAGE_SCN in 6 files
and 11 different functions. Section to attribute mapping really
needs to happen in one place to avoid problems like this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117473 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 18:52:29 +00:00
Michael J. Spencer
579d7a3dcc Fix whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117472 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 18:52:20 +00:00
Jim Grosbach
0ed257c036 Formatting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117453 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 16:30:18 +00:00
Jakob Stoklund Olesen
c95c1465fd Handle critical loop predecessors by making both inside and outside registers
live out.

This doesn't prevent us from inserting a loop preheader later on, if that is
better.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117424 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 00:39:07 +00:00
Jakob Stoklund Olesen
0960a650b7 Compute critical loop predecessors in the same way as critical loop exits.
Critical edges going into a loop are not as bad as critical exits. We can handle
them by splitting the critical edge, or by having both inside and outside
registers live out of the predecessor.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117423 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 00:39:05 +00:00
Jakob Stoklund Olesen
8c593f9173 Physical registers trivially have multiple connected components all the time.
Only virtuals should be requires to be connected.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117422 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-27 00:39:01 +00:00
Dale Johannesen
1de4aa904e Use a MemIntrinsicSDNode for ISD::PREFETCH, which touches
memory, so a MachineMemOperand is useful (not propagated
into the MachineInstr yet).  No functional change except
for dump output.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117413 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 23:11:10 +00:00
Andrew Trick
3d26d5d524 Remove the vector of live vregs. I thought we would need to track
them, but hopefully we won't. And this is not the right data structure
to do it anyway.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117412 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 22:58:24 +00:00
Jakob Stoklund Olesen
3a0e0715a5 After splitting, compute connected components of all new registers, not just for
the remainder register.

Example:

bb0:
  x = 1
bb1:
  use(x)
  ...
  x = 2
  jump bb1

When x is isolated in bb1, the inner part breaks into two components, x1 and x2:

bb0:
  x0 = 1
bb1:
  x1 = x0
  use(x1)
  ...
  x2 = 2
  x0 = x2
  jump bb1

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117408 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 22:36:09 +00:00
Jakob Stoklund Olesen
501dc42245 Verify that live intervals are connected. If there are multiple connected
components, each should get its own virtual register.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117407 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 22:36:07 +00:00
Jakob Stoklund Olesen
f1354ae95a Call RenumberValues for all new registers created during splitting. This is
necessary to get correct hasPHIKill flags.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117406 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 22:36:05 +00:00
Jakob Stoklund Olesen
79c0262fa8 Preserve PHIDef bits in cloned values during splitting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117405 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 22:36:02 +00:00
Devang Patel
7e13efad38 Assign source ordering to nodes created for StoreInst.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117404 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 22:14:52 +00:00
Jakob Stoklund Olesen
f4a1e1a69f Teach MachineBasicBlock::print() to annotate instructions and blocks with
SlotIndexes when available.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117392 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 20:21:46 +00:00
Jakob Stoklund Olesen
dbcc2e119d Remmeber to print full live interval on verification error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117391 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 20:21:43 +00:00
Andrew Trick
e16eecc323 Jakob's review of the basic register allocator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117384 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 18:34:01 +00:00
Devang Patel
cbbe287f8a s/beginScope/beginInstruction/g
s/endScope/endInstruction/g



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117376 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 17:49:02 +00:00
Jakob Stoklund Olesen
e459d55f28 Don't verify physical registers going into landing pads.
Magic is happening that we don't understand.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117370 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 16:49:23 +00:00
Evan Cheng
c8141dfc7f Use instruction itinerary to determine what instructions are 'cheap'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117348 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 02:08:50 +00:00
Nick Lewycky
a568d66512 For statistics that are only used in functions declared in !NDEBUG, wrap the
declarations in !NDEBUG to avoid -Wunused-variable warnings. Patch by
Matt Beaumont-Gay!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117345 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 00:51:57 +00:00
Jakob Stoklund Olesen
0a12b801b5 InlineSpiller can also update LiveStacks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117338 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 00:11:35 +00:00
Jakob Stoklund Olesen
2d17293dd0 Make the spiller responsible for updating the LiveStacks analysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117337 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 00:11:33 +00:00
Bob Wilson
7c730e7790 When the "true" and "false" blocks of a diamond if-conversion are the same,
do not double-count the duplicate instructions by counting once from the
beginning and again from the end.  Keep track of where the duplicates from
the beginning ended and don't go past that point when counting duplicates
at the end.  Radar 8589805.

This change causes one of the MC/ARM/simple-fp-encoding tests to produce
different (better!) code without the vmovne instruction being tested.
I changed the test to produce vmovne and vmoveq instructions but moving
between register files in the opposite direction.  That's not quite the same
but predicated versions of those instructions weren't being tested before,
so at least the test coverage is not any worse, just different.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117333 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 00:02:24 +00:00
Bob Wilson
2ad40d3494 Change if-conversion to keep track of the extra cost due to microcoded
instructions separately from the count of non-predicated instructions.  The
instruction count is used in places to determine how many instructions to
copy, predicate, etc. and things get confused if that count includes the
extra cost for microcoded ops.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117332 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-26 00:02:21 +00:00
Evan Cheng
38fd9f2a2d Neuter r117193 as it causes significant post-ra scheduler compile time regression.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117329 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-25 23:56:21 +00:00
Devang Patel
df8370bd70 Simplify.
Do not count use of sdisel for single call instruction.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117316 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-25 21:31:46 +00:00
Devang Patel
948f7d0216 Add counters to count basic blocks and machine basic blocks with out of order line number info.
Add counters to count how many basic blocks are entirely selected by fastisel.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117310 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-25 20:55:43 +00:00
Devang Patel
9a31f0f398 Add simple counter to count no. of basic blocks without any line number information. At -O0, these basic block coule cause less than optimial debugging experience.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117307 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-25 20:45:32 +00:00
Jakob Stoklund Olesen
d3b4895414 In which I learn how to forward declare template classes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117272 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-25 17:27:30 +00:00
Chandler Carruth
19e57025d4 Move the remaining attribute macros to systematic names based on the attribute
name and prefixed with 'LLVM_'.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117203 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-23 08:40:19 +00:00
Chandler Carruth
2b3ef8b065 Fix a likely bug in an assertion by adding parentheses around '||'. This bug
was found by a GCC warning. ;]


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117199 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-23 07:46:14 +00:00
Evan Cheng
ec6906ba47 Properly model the latency of register defs which are 1) function returns or
2) live-outs.

Previously the post-RA schedulers completely ignore these dependencies since
returns, branches, etc. are all scheduling barriers. This patch model the
latencies between instructions being scheduled and the barriers. It also
handle calls by marking their register uses.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117193 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-23 02:10:46 +00:00
Jakob Stoklund Olesen
7871687604 Verify LiveIntervals against the CFG, ensuring that live-in values are live-out
of all predecessors.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117191 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-23 00:49:09 +00:00
Andrew Trick
f433106054 Nonvirtual dtor that was accessible enough to be bad.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117180 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 23:33:19 +00:00
Andrew Trick
14e8d71cc9 This is a prototype of an experimental register allocation
framework. It's purpose is not to improve register allocation per se,
but to make it easier to develop powerful live range splitting. I call
it the basic allocator because it is as simple as a global allocator
can be but provides the building blocks for sophisticated register
allocation with live range splitting. 

A minimal implementation is provided that trivially spills whenever it
runs out of registers. I'm checking in now to get high-level design
and style feedback. I've only done minimal testing. The next step is
implementing a "greedy" allocation algorithm that does some register
reassignment and makes better splitting decisions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117174 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 23:09:15 +00:00
Jakob Stoklund Olesen
3bf7cf9f0e Add more verification of LiveIntervals.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117170 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 22:48:58 +00:00
Jakob Stoklund Olesen
2bfb324684 Be more strict about detecting multi-use blocks for isolation.
When a block has exactly two uses and the register is both live-in and live-out,
don't isolate the block. We would be inserting two copies, so we haven't really
made any progress.

If the live-in and live-out values separate into disconnected components after
splitting, we would be making progress. We can't detect that for now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117169 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 22:48:56 +00:00
Evan Cheng
c8f46c45a0 Unbreak build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117155 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 21:49:09 +00:00
Evan Cheng
b179b46cc5 Transfer implicit ops when forming load multiple and return instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117151 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 21:29:58 +00:00
Jakob Stoklund Olesen
2c1442e1b2 Be more strict when detecting critical edges before loop splitting.
An exit block with a critical edge must only have predecessors in the loop, or
just before the loop. This guarantees that the inserted copies in the loop
predecessors dominate the exit block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117144 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 20:28:23 +00:00
Jakob Stoklund Olesen
532de3dc6e Add print methods
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117143 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-22 20:28:21 +00:00
Michael J. Spencer
391b43bf33 X86: Base _fltused on the FunctionType of the called value instead of the potentially null "CalledFunction". Thanks Duncan!
This is needed for indirect calls.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117061 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-21 20:49:23 +00:00
Jakob Stoklund Olesen
9d999772df Don't include the destination interval in the union when computing
Parent - union(Y, ...). Doh.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117042 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-21 18:47:08 +00:00
Jakob Stoklund Olesen
0a7bbcb0cc Permit landing pad successor blocks when verifying basic blocks that end in an
unconditional branch.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117041 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-21 18:47:06 +00:00
Duncan Sands
9bc20ca728 The variable liTRC is not used for anything useful, zap it
(gcc-4.6 warning).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117022 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-21 16:04:43 +00:00
Duncan Sands
6abfa46987 Remove some variables that are never really used
(gcc-4.6 warns about these).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117021 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-21 16:03:28 +00:00
Michael J. Spencer
dd0a00a6e3 CodeGen-Windows: Only emit _fltused if a VarArg function is called with floating point args.
This should be the minimum set of functions that could possibly need it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116978 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-21 00:08:21 +00:00
Jakob Stoklund Olesen
f1583ae84a Remember to keep track of rematted values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116962 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-20 22:50:42 +00:00
Evan Cheng
134982daa9 More accurate estimate / tracking of register pressure.
- Initial register pressure in the loop should be all the live defs into the
  loop. Not just those from loop preheader which is often empty.
- When an instruction is hoisted, update register pressure from loop preheader
  to the original BB.
- Treat only use of a virtual register as kill since the code is still SSA.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116956 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-20 22:03:58 +00:00
Jakob Stoklund Olesen
080c316ff8 Move some of the InlineSpiller rematerialization code into LiveRangeEdit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116951 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-20 22:00:51 +00:00
Dale Johannesen
e93d99cf07 Remove Synthesizable from the Type system; as MMX vector
types are no longer Legal on X86, we don't need it.
No functional change.  8499854.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116947 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-20 21:32:10 +00:00
Jakob Stoklund Olesen
9b25940474 When SimpleRegisterCoalescing is trimming kill flags on a physical register
operand, also check if subregisters are killed.

Add <imp-def> operands for subregisters that remain alive after a super register
is killed.

I don't have a testcase for this that reproduces on trunk. <rdar://problem/8441758>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116940 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-20 18:45:55 +00:00
Dan Gohman
f96e4bd2a3 Make CodeGen TBAA-aware.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116890 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-20 00:31:05 +00:00
Jim Grosbach
e4ad387a5a Add a pre-dispatch SjLj EH hook on the unwind edge for targets to do any
setup they require. Use this for ARM/Darwin to rematerialize the base
pointer from the frame pointer when required. rdar://8564268

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116879 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 23:27:08 +00:00
Jakob Stoklund Olesen
2594746045 Shrink MachineOperand from 40 to 32 bytes on 64-bit hosts.
Pull an unsigned out of the Contents union such that it has the same size as two
pointers and no padding.

Arrange members such that the Contents union and all pointers can be 8-byte
aligned without padding.

This speeds up code generation by 0.8% on a 64-bit host. 32-bit hosts should be
unaffected.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116857 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 20:56:32 +00:00
Evan Cheng
2312842de0 Re-enable register pressure aware machine licm with fixes. Hoist() may have
erased the instruction during LICM so UpdateRegPressureAfter() should not
reference it afterwards.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116845 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 18:58:51 +00:00
Owen Anderson
081c34b725 Get rid of static constructors for pass registration. Instead, every pass exposes an initializeMyPassFunction(), which
must be called in the pass's constructor.  This function uses static dependency declarations to recursively initialize
the pass's dependencies.

Clients that only create passes through the createFooPass() APIs will require no changes.  Clients that want to use the
CommandLine options for passes will need to manually call the appropriate initialization functions in PassInitialization.h
before parsing commandline arguments.

I have tested this with all standard configurations of clang and llvm-gcc on Darwin.  It is possible that there are problems
with the static dependencies that will only be visible with non-standard options.  If you encounter any crash in pass
registration/creation, please send the testcase to me directly.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116820 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 17:21:58 +00:00
Daniel Dunbar
9869413802 Revert r116781 "- Add a hook for target to determine whether an instruction def
is", which breaks some nightly tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116816 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 17:14:24 +00:00
NAKAMURA Takumi
3b3b0ebd2e lib/CodeGen/TargetLoweringObjectFileImpl.cpp: Tweak to emit ".{section}${name}" instead of ".{section}$linkonce_{name}" for linkonce sections.
It seems GNU ld/PECOFF relies on section names, linking with g++'s libstdc++.a would fail.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116791 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 03:24:42 +00:00
Andrew Trick
dc986d2462 Fix for machine licm assert: RCCost <= RegPressure[RCId]
in MultiSource/Benchmarks/VersaBench/beamformer/beamformer.
SmallSet.insert returns true if the element is inserted.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116790 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 02:50:50 +00:00
Evan Cheng
11e8b74a7a - Add a hook for target to determine whether an instruction def is
"long latency" enough to hoist even if it may increase spilling. Reloading
  a value from spill slot is often cheaper than performing an expensive
  computation in the loop. For X86, that means machine LICM will hoist
  SQRT, DIV, etc. ARM will be somewhat aggressive with VFP and NEON
  instructions.
- Enable register pressure aware machine LICM by default.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116781 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-19 00:55:07 +00:00
Bill Wendling
b41ee96d76 Don't recompute MachineRegisterInfo in the Optimize* method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116750 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-18 21:22:31 +00:00
Dan Gohman
4021fccf9f Add TypeBasedAliasAnalysis to the standard pass lists. Note that it
is currently inert by default.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116732 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-18 18:50:27 +00:00
Dan Gohman
c1be92f3bb Make BasicAliasAnalysis a normal AliasAnalysis implementation which
does normal initialization and normal chaining. Change the default
AliasAnalysis implementation to NoAlias.

Update StandardCompileOpts.h and friends to explicitly request
BasicAliasAnalysis.

Update tests to explicitly request -basicaa.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116720 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-18 18:04:47 +00:00
Jim Grosbach
c0c7c2b606 Trivial grammar tweak.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116710 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-18 16:29:26 +00:00
Michael J. Spencer
84ac4d5a2a X86-Windows: Emit an undefined global __fltused symbol when targeting Windows
if any floating point arguments are passed to an external function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116665 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-16 08:25:41 +00:00
Michael J. Spencer
e70c526d59 Whitespace!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116664 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-16 08:25:21 +00:00
Evan Cheng
03a9fdf2e7 More machine LICM work. It now tracks register pressure for path from preheader to current BB and use the information determine whether hoisting is worthwhile.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116654 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-16 02:20:26 +00:00
Jakob Stoklund Olesen
50cef5829d Remove unused accessor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116580 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-15 16:06:40 +00:00
Jakob Stoklund Olesen
9d5d48bc41 Eliminate curli from SplitEditor. Use the LiveRangeEdit reference instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116547 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-15 00:34:01 +00:00
Jakob Stoklund Olesen
2a0180fbff Move stack slot assignments into LiveRangeEdit.
All registers created during splitting or spilling are assigned to the same
stack slot as the parent register.

When splitting or rematting, we may not spill at all. In that case the stack
slot is still assigned, but it will be dead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116546 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-15 00:16:55 +00:00
Jakob Stoklund Olesen
a17768f582 Create a new LiveRangeEdit class to keep track of the new registers created when
splitting or spillling, and to help with rematerialization.

Use LiveRangeEdit in InlineSpiller and SplitKit. This will eventually make it
possible to share remat code between InlineSpiller and SplitKit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116543 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-14 23:49:52 +00:00
Jakob Stoklund Olesen
ab00e9f498 Only split around a loop if the live range has uses outside the loop periphery.
Before we would also split around a loop if any peripheral block had multiple
uses. This could cause repeated splitting when splitting a different live range
would insert uses into the periphery.

Now -spiller=inline passes the nightly test suite again.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116494 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-14 18:26:45 +00:00
Evan Cheng
0e673919f0 Register pressure and instruction latency aware machine LICM. Work in progress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116465 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-14 01:16:09 +00:00
Owen Anderson
325e264355 Analysis groups need to initialize their default implementations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116441 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-13 21:49:58 +00:00
Owen Anderson
2ab36d3502 Begin adding static dependence information to passes, which will allow us to
perform initialization without static constructors AND without explicit initialization
by the client.  For the moment, passes are required to initialize both their
(potential) dependencies and any passes they preserve.  I hope to be able to relax
the latter requirement in the future.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116334 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-12 19:48:12 +00:00