Chris Lattner [Fri, 3 Mar 2006 00:19:58 +0000 (00:19 +0000)]
remove the read/write port/io intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26479
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 3 Mar 2006 00:16:23 +0000 (00:16 +0000)]
Remove these testcases. These intrinsics are going away
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26478
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 3 Mar 2006 00:07:20 +0000 (00:07 +0000)]
Update the llvm.memset/memcpy/memmove intrinsics to reflect the recent split
into types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26477
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 3 Mar 2006 00:00:25 +0000 (00:00 +0000)]
Split memcpy/memset/memmove intrinsics into i32/i64 versions, resolving
PR709, and paving the way for future progress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26476
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 23:59:12 +0000 (23:59 +0000)]
Simplify the autoupgrade interface
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26475
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 23:58:40 +0000 (23:58 +0000)]
Split memcpy/memset/memmove intrinsics into i32/i64 versions, resolving
PR709, and paving the way for future progress.
Significantly refactor autoupgrading code, to handle the more complex case
(where we upgrade one argument in a function), and fix some bugs in it.
Testcase here: llvm/test/Regression/Bytecode/memcpy.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26474
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 23:56:23 +0000 (23:56 +0000)]
new testcases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26473
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 22:34:38 +0000 (22:34 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26472
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 2 Mar 2006 21:48:34 +0000 (21:48 +0000)]
Add a regression test for bug 478.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26471
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 2 Mar 2006 21:38:29 +0000 (21:38 +0000)]
- Fixed some priority calculation bugs that were causing bug 478. Among them:
a predecessor appearing more than once in the operand list was counted as
multiple predecessor; priority1 should be updated during scheduling;
CycleBound was updated after the node is inserted into priority queue; one
of the tie breaking condition was flipped.
- Take into consideration of two address opcodes. If a predecessor is a def&use
operand, it should have a higher priority.
- Scheduler should also favor floaters, i.e. nodes that do not have real
predecessors such as MOV32ri.
- The scheduling fixes / tweaks fixed bug 478:
.text
.align 4
.globl _f
_f:
movl 4(%esp), %eax
movl 8(%esp), %ecx
movl %eax, %edx
imull %ecx, %edx
imull %eax, %eax
imull %ecx, %ecx
addl %eax, %ecx
leal (%ecx,%edx,2), %eax
ret
It is also a slight performance win (1% - 3%) for most tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26470
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 06:50:58 +0000 (06:50 +0000)]
Generalize the REM folding code to handle another case Nick Lewycky
pointed out: realize the AND can provide factors and look through Casts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26469
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 06:50:04 +0000 (06:50 +0000)]
add a couple more cases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26468
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 2 Mar 2006 00:21:41 +0000 (00:21 +0000)]
Fix a compilation error with GCC 4.1. Thanks to Vladimir Merzliakov for
pointing this out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26467
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 23:52:37 +0000 (23:52 +0000)]
Support for enumerations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26466
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 22:45:12 +0000 (22:45 +0000)]
testcase that crashed the new CFE
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26465
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 22:18:09 +0000 (22:18 +0000)]
Don't print llvm constant in assmebly file. Assembler won't like comments that
span multiple lines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26463
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 22:17:00 +0000 (22:17 +0000)]
Back out my last check-in. Wrong place to fix it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26462
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 22:00:59 +0000 (22:00 +0000)]
AsmWriter should not print LLVM constant in comment. Assembler won't like
multi-line comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26461
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 21:47:21 +0000 (21:47 +0000)]
Fix CodeGen/Generic/2006-03-01-dagcombineinfloop.ll, an infinite loop
in the dag combiner on 176.gcc on x86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26459
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 21:46:53 +0000 (21:46 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26458
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 20:49:44 +0000 (20:49 +0000)]
Remove extra comma from enum list.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26457
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 20:48:20 +0000 (20:48 +0000)]
Remove comma from enum list.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26456
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 20:39:36 +0000 (20:39 +0000)]
Switch back to using actual dwarf tags. Simplifies code without loss to other
debug forms.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26455
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 19:55:35 +0000 (19:55 +0000)]
Fix a typo evan noticed
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26454
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 18:20:30 +0000 (18:20 +0000)]
Use context and not compile unit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26453
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 18:13:05 +0000 (18:13 +0000)]
I guess I can handle large type sizes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26452
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Wed, 1 Mar 2006 17:53:02 +0000 (17:53 +0000)]
Basic array support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26451
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 07:14:48 +0000 (07:14 +0000)]
Implement CodeGen/PowerPC/or-addressing-mode.ll, which is also PR668.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26450
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 07:13:56 +0000 (07:13 +0000)]
load (x|y) -> load (x+y) iff x and y have no common bits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26449
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 06:36:20 +0000 (06:36 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26448
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 05:50:56 +0000 (05:50 +0000)]
Compile this:
void foo(float a, int *b) { *b = a; }
to this:
_foo:
fctiwz f0, f1
stfiwx f0, 0, r4
blr
instead of this:
_foo:
fctiwz f0, f1
stfd f0, -8(r1)
lwz r2, -4(r1)
stw r2, 0(r4)
blr
This implements CodeGen/PowerPC/stfiwx.ll, and also incidentally does the
right thing for GCC bugzilla 26505.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26447
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 05:49:05 +0000 (05:49 +0000)]
new testcase. These functions shouldn't touch the stack if stfiwx use
is enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26446
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 04:57:39 +0000 (04:57 +0000)]
Use a target-specific dag-combine to implement CodeGen/PowerPC/fp-int-fp.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26445
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 04:56:33 +0000 (04:56 +0000)]
new testcase. There should be no accesses to the stack for these functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26444
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 04:53:38 +0000 (04:53 +0000)]
Add support for target-specific dag combines
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26443
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 04:52:55 +0000 (04:52 +0000)]
Add interfaces for targets to provide target-specific dag combiner optimizations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26442
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 04:03:14 +0000 (04:03 +0000)]
Add a new AddToWorkList method, start using it
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26441
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 03:44:24 +0000 (03:44 +0000)]
Pull shifts by a constant through multiplies (a form of reassociation),
implementing Regression/CodeGen/X86/mul-shift-reassoc.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26440
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 1 Mar 2006 03:43:38 +0000 (03:43 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26439
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 01:11:20 +0000 (01:11 +0000)]
Vector op lowering.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26438
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 01:10:52 +0000 (01:10 +0000)]
New vector type v2f32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26437
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 01:09:54 +0000 (01:09 +0000)]
Vector ops lowering.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26436
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 01:06:22 +0000 (01:06 +0000)]
New type v2f32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26435
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 00:58:54 +0000 (00:58 +0000)]
Missing a cast previously.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26434
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 00:55:26 +0000 (00:55 +0000)]
- Added v2f32, not used by any target currently. Only for testing purpose.
- Minor bug fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26433
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 1 Mar 2006 00:51:13 +0000 (00:51 +0000)]
- Added VConstant as an abstract version of ConstantVec.
- All abstrct vector nodes must have # of elements and element type as their
first two operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26432
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 28 Feb 2006 23:57:45 +0000 (23:57 +0000)]
Add a test case for left shift by 1. We should not be using lea for this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26431
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 28 Feb 2006 23:38:49 +0000 (23:38 +0000)]
Another entry.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26430
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 28 Feb 2006 21:13:57 +0000 (21:13 +0000)]
Don't match x << 1 to LEAL. It's better to emit x + x.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26429
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Tue, 28 Feb 2006 20:15:07 +0000 (20:15 +0000)]
Add const, volatile, restrict support.
Add array of debug descriptor support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26428
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 19:47:20 +0000 (19:47 +0000)]
Fix a regression in a patch from a couple of days ago. This fixes
Transforms/InstCombine/2006-02-28-Crash.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26427
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 19:46:56 +0000 (19:46 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26426
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 19:12:58 +0000 (19:12 +0000)]
8 spaces -> tab. Reported by Wink Saville
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26425
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 16:39:56 +0000 (16:39 +0000)]
evan's recent x86 isel improvements have fixed this, though not in the way
originally envisioned :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26422
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 07:08:22 +0000 (07:08 +0000)]
Add a subtarget feature for the stfiwx instruction. I know the G5 has it,
but I don't know what other PPC impls do. If someone could update the proc
table, I would appreciate it :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26421
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 06:54:19 +0000 (06:54 +0000)]
new testcases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26420
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 06:49:37 +0000 (06:49 +0000)]
Compile:
unsigned foo4(unsigned short *P) { return *P & 255; }
unsigned foo5(short *P) { return *P & 255; }
to:
_foo4:
lbz r3,1(r3)
blr
_foo5:
lbz r3,1(r3)
blr
not:
_foo4:
lhz r2, 0(r3)
rlwinm r3, r2, 0, 24, 31
blr
_foo5:
lhz r2, 0(r3)
rlwinm r3, r2, 0, 24, 31
blr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26419
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 06:36:04 +0000 (06:36 +0000)]
remove implemented item
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26418
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 06:35:35 +0000 (06:35 +0000)]
Fold "and (LOAD P), 255" -> zextload. This allows us to compile:
unsigned foo3(unsigned *P) { return *P & 255; }
as:
_foo3:
lbz r3, 3(r3)
blr
instead of:
_foo3:
lwz r2, 0(r3)
rlwinm r3, r2, 0, 24, 31
blr
and:
unsigned short foo2(float a) { return a; }
as:
_foo2:
fctiwz f0, f1
stfd f0, -8(r1)
lhz r3, -2(r1)
blr
instead of:
_foo2:
fctiwz f0, f1
stfd f0, -8(r1)
lwz r2, -4(r1)
rlwinm r3, r2, 0, 16, 31
blr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26417
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 06:23:04 +0000 (06:23 +0000)]
fold (sra (sra x, c1), c2) -> (sra x, c1+c2)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26416
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 05:49:21 +0000 (05:49 +0000)]
Implement rem.ll:test[7-9] and PR712
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26415
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 05:48:56 +0000 (05:48 +0000)]
new testcases for PR712
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26414
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 05:40:55 +0000 (05:40 +0000)]
Simplify some code now that the RHS of a rem can't be 0
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26413
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 05:30:48 +0000 (05:30 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26412
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 28 Feb 2006 05:30:45 +0000 (05:30 +0000)]
Rearrange some code, fold "rem X, 0", implementing rem.ll:test6
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26411
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 23:45:39 +0000 (23:45 +0000)]
Add support for output memory constraints.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26410
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Mon, 27 Feb 2006 22:37:23 +0000 (22:37 +0000)]
Qualify dwarf namespace inside llvm namespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26409
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Mon, 27 Feb 2006 22:08:36 +0000 (22:08 +0000)]
readme updates
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26405
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Mon, 27 Feb 2006 20:37:42 +0000 (20:37 +0000)]
Partial enabling of functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26404
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 20:09:23 +0000 (20:09 +0000)]
Don't print constant initializers, they may span lines now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26403
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Mon, 27 Feb 2006 17:27:12 +0000 (17:27 +0000)]
Supporting multiple compile units.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26402
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Mon, 27 Feb 2006 12:43:29 +0000 (12:43 +0000)]
Re-orging file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26401
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Mon, 27 Feb 2006 10:33:53 +0000 (10:33 +0000)]
Pretty print large struct constants.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26400
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Mon, 27 Feb 2006 10:29:04 +0000 (10:29 +0000)]
Removed dependency on how operands are printed (want multi-line.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26399
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 05:39:00 +0000 (05:39 +0000)]
Use -emit-llvm -S to get .ll file output from llvm-gcc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26397
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 02:38:23 +0000 (02:38 +0000)]
Merge two almost-identical pieces of code.
Make this code more powerful by using ComputeMaskedBits instead of looking
for an AND operand. This lets us fold this:
int %test23(int %a) {
%tmp.1 = and int %a, 1
%tmp.2 = seteq int %tmp.1, 0
%tmp.3 = cast bool %tmp.2 to int ;; xor tmp1, 1
ret int %tmp.3
}
into: xor (and a, 1), 1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26396
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 02:36:19 +0000 (02:36 +0000)]
new testcases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26395
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 01:44:11 +0000 (01:44 +0000)]
Fold (A^B) == A -> B == 0
and (A-B) == A -> B == 0
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26394
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 01:43:02 +0000 (01:43 +0000)]
New testcases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26393
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 01:00:42 +0000 (01:00 +0000)]
Implement bit propagation through sub nodes, this (re)implements
PowerPC/div-2.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26392
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 01:00:12 +0000 (01:00 +0000)]
Reenable this
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26391
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 00:39:31 +0000 (00:39 +0000)]
remove some completed notes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26390
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 00:36:27 +0000 (00:36 +0000)]
Check RHS simplification before LHS simplification to avoid infinitely looping
on PowerPC/small-arguments.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26389
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 00:22:28 +0000 (00:22 +0000)]
Just like we use the RHS of an AND to simplify the LHS, use the LHS to
simplify the RHS. This allows for the elimination of many thousands of
ands from multisource, and compiles CodeGen/PowerPC/and-elim.ll:test2
into this:
_test2:
srwi r2, r3, 1
xori r3, r2, 40961
blr
instead of this:
_test2:
rlwinm r2, r3, 31, 17, 31
xori r2, r2, 40961
rlwinm r3, r2, 0, 16, 31
blr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26388
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 27 Feb 2006 00:20:23 +0000 (00:20 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26387
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 26 Feb 2006 23:36:02 +0000 (23:36 +0000)]
Add a bunch of missed cases. Perhaps the most significant of which is that
assertzext produces zero bits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26386
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 26 Feb 2006 19:57:54 +0000 (19:57 +0000)]
Fold (X|C1)^C2 -> X^(C1|C2) when possible. This implements
InstCombine/or.ll:test23.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26385
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 26 Feb 2006 19:55:30 +0000 (19:55 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26384
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Sun, 26 Feb 2006 10:16:05 +0000 (10:16 +0000)]
Reverting. Didn't realize some developers were embedding constants in their
target assembler code gen.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26383
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sun, 26 Feb 2006 09:12:34 +0000 (09:12 +0000)]
ConstantPoolIndex is now the displacement portion of the address (rather
than base).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26382
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sun, 26 Feb 2006 08:36:57 +0000 (08:36 +0000)]
Print ConstantPoolSDNode offset field.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26381
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sun, 26 Feb 2006 08:28:12 +0000 (08:28 +0000)]
Fixed ConstantPoolIndex operand asm print bug. This fixed 2005-07-17-INT-To-FP
and 2005-05-12-Int64ToFP.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26380
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Laskey [Sat, 25 Feb 2006 12:27:03 +0000 (12:27 +0000)]
Format large struct constants for readability.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26379
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 10:16:10 +0000 (10:16 +0000)]
New test case: use lea for imul by some constants.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26378
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 10:15:22 +0000 (10:15 +0000)]
lea.ll is XFAIL until we implement convertToThreeAddress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26377
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 10:09:08 +0000 (10:09 +0000)]
* Cleaned up addressing mode matching code.
* Cleaned up and tweaked LEA cost analysis code. Removed some hacks.
* Handle ADD $X, c to MOV32ri $X+c. These patterns cannot be autogen'd and
they need to be matched before LEA.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26376
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 10:04:07 +0000 (10:04 +0000)]
Updates.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26375
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 10:02:21 +0000 (10:02 +0000)]
* Allow mul, shl nodes to be codegen'd as LEA (if appropriate).
* Add patterns to handle GlobalAddress, ConstantPool, etc.
MOV32ri to materialize these nodes in registers.
ADD32ri to handle %reg + GA, etc.
MOV32mi to handle store GA, etc. to memory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26374
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 09:56:50 +0000 (09:56 +0000)]
ConstantPoolIndex is now the displacement field of addressing mode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26373
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sat, 25 Feb 2006 09:55:19 +0000 (09:55 +0000)]
Added a common about the need for X86ISD::Wrapper.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26372
91177308-0d34-0410-b5e6-
96231b3b80d8