mirrors/llvm-project
0
0
Fork 0
mirror of https://github.com/llvm/llvm-project.git synced 2025-05-18 01:56:07 +00:00
Code Issues Projects Releases Wiki Activity
22,868 Commits 1,258 Branches 288 Tags
Commit Graph

173 Commits

Author SHA1 Message Date
Chris Lattner
b32664583b Compile llvm.stacksave/restore into STACKSAVE/STACKRESTORE nodes, and allow 
targets to custom expand them as they desire.

llvm-svn: 25273
 2006-01-13 02:50:02 +00:00
Chris Lattner
6c9c250dcd Add "support" for stacksave/stackrestore to the dag isel 
llvm-svn: 25268
 2006-01-13 02:24:42 +00:00
Robert Bocchino
2c966e7617 Added selection DAG support for the extractelement operation. 
llvm-svn: 25179
 2006-01-10 19:04:57 +00:00
Jim Laskey
219d559824 Applied some recommend changes from sabre. The dominate one beginning "let the 
pass manager do it's thing."  Fixes crash when compiling -g files and suppresses
dwarf statements if no debug info is present.

llvm-svn: 25100
 2006-01-04 22:28:25 +00:00
Chris Lattner
44c07ed61a enable the gep isel opt 
llvm-svn: 24910
 2005-12-21 19:36:36 +00:00
Chris Lattner
803a575616 Lower ConstantAggregateZero into zeros 
llvm-svn: 24890
 2005-12-21 02:43:26 +00:00
Jim Laskey
7c462768ed Added source file/line correspondence for dwarf (PowerPC only at this point.) 
llvm-svn: 24748
 2005-12-16 22:45:29 +00:00
Chris Lattner
5d4e61dd87 Don't lump the filename and working dir together 
llvm-svn: 24697
 2005-12-13 17:40:33 +00:00
Chris Lattner
9e8b633ec1 Accept and ignore prefetches for now 
llvm-svn: 24678
 2005-12-12 22:51:16 +00:00
Chris Lattner
f1a54c0d14 Minor tweak to get isel opt 
llvm-svn: 24663
 2005-12-11 09:05:13 +00:00
Chris Lattner
be73d6eece improve code insertion in two ways: 
1. Only forward subst offsets into loads and stores, not into arbitrary
   things, where it will likely become a load.
2. If the source is a cast from pointer, forward subst the cast as well,
   allowing us to fold the cast away (improving cases when the cast is
   from an alloca or global).

This hasn't been fully tested, but does appear to further reduce register
pressure and improve code.  Lets let the testers grind on it a bit. :)

llvm-svn: 24640
 2005-12-08 08:00:12 +00:00
Nate Begeman
ae89d862f5 Fix a crash where ConstantVec nodes were being generated with the wrong 
type when the target did not support them.  Also teach Legalize how to
expand ConstantVecs.

This allows us to generate

_test:
        lwz r2, 12(r3)
        lwz r4, 8(r3)
        lwz r5, 4(r3)
        lwz r6, 0(r3)
        addi r2, r2, 4
        addi r4, r4, 3
        addi r5, r5, 2
        addi r6, r6, 1
        stw r2, 12(r3)
        stw r4, 8(r3)
        stw r5, 4(r3)
        stw r6, 0(r3)
        blr

For:

void %test(%v4i *%P) {
        %T = load %v4i* %P
        %S = add %v4i %T, <int 1, int 2, int 3, int 4>
        store %v4i %S, %v4i * %P
        ret void
}

On PowerPC.

llvm-svn: 24633
 2005-12-07 19:48:11 +00:00
Nate Begeman
41b1cdc771 Teach the SelectionDAG ISel how to turn ConstantPacked values into 
constant nodes with vector types.  Also teach the asm printer how to print
ConstantPacked constant pool entries.  This allows us to generate altivec
code such as the following, which adds a vector constantto a packed float.

LCPI1_0:  <4 x float> < float 0.0e+0, float 0.0e+0, float 0.0e+0, float 1.0e+0 >
        .space  4
        .space  4
        .space  4
        .long   1065353216      ; float 1
        .text
        .align  4
        .globl  _foo
_foo:
        lis r2, ha16(LCPI1_0)
        la r2, lo16(LCPI1_0)(r2)
        li r4, 0
        lvx v0, r4, r2
        lvx v1, r4, r3
        vaddfp v0, v1, v0
        stvx v0, r4, r3
        blr

For the llvm code:

void %foo(<4 x float> * %a) {
entry:
  %tmp1 = load <4 x float> * %a;
  %tmp2 = add <4 x float> %tmp1, < float 0.0, float 0.0, float 0.0, float 1.0 >
  store <4 x float> %tmp2, <4 x float> *%a
  ret void
}

llvm-svn: 24616
 2005-12-06 06:18:55 +00:00
Chris Lattner
3539778883 Fix the #1 code quality problem that I have seen on X86 (and it also affects 
PPC and other targets).  In a particular, consider code like this:

struct Vector3 { double x, y, z; };
struct Matrix3 { Vector3 a, b, c; };
double dot(Vector3 &a, Vector3 &b) {
   return a.x * b.x  +  a.y * b.y  +  a.z * b.z;
}
Vector3 mul(Vector3 &a, Matrix3 &b) {
   Vector3 r;
   r.x = dot( a, b.a );
   r.y = dot( a, b.b );
   r.z = dot( a, b.c );
   return r;
}
void transform(Matrix3 &m, Vector3 *x, int n) {
   for (int i = 0; i < n; i++)
      x[i] = mul( x[i], m );
}

we compile transform to a loop with all of the GEP instructions for indexing
into 'm' pulled out of the loop (9 of them).  Because isel occurs a bb at a time
we are unable to fold the constant index into the loads in the loop, leading to
PPC code that looks like this:

LBB3_1: ; no_exit.preheader
        li r2, 0
        addi r6, r3, 64        ;; 9 values live across the loop body!
        addi r7, r3, 56
        addi r8, r3, 48
        addi r9, r3, 40
        addi r10, r3, 32
        addi r11, r3, 24
        addi r12, r3, 16
        addi r30, r3, 8
LBB3_2: ; no_exit
        lfd f0, 0(r30)
        lfd f1, 8(r4)
        fmul f0, f1, f0
        lfd f2, 0(r3)        ;; no constant indices folded into the loads!
        lfd f3, 0(r4)
        lfd f4, 0(r10)
        lfd f5, 0(r6)
        lfd f6, 0(r7)
        lfd f7, 0(r8)
        lfd f8, 0(r9)
        lfd f9, 0(r11)
        lfd f10, 0(r12)
        lfd f11, 16(r4)
        fmadd f0, f3, f2, f0
        fmul f2, f1, f4
        fmadd f0, f11, f10, f0
        fmadd f2, f3, f9, f2
        fmul f1, f1, f6
        stfd f0, 0(r4)
        fmadd f0, f11, f8, f2
        fmadd f1, f3, f7, f1
        stfd f0, 8(r4)
        fmadd f0, f11, f5, f1
        addi r29, r4, 24
        stfd f0, 16(r4)
        addi r2, r2, 1
        cmpw cr0, r2, r5
        or r4, r29, r29
        bne cr0, LBB3_2 ; no_exit

uh, yuck.  With this patch, we now sink the constant offsets into the loop, producing
this code:

LBB3_1: ; no_exit.preheader
        li r2, 0
LBB3_2: ; no_exit
        lfd f0, 8(r3)
        lfd f1, 8(r4)
        fmul f0, f1, f0
        lfd f2, 0(r3)
        lfd f3, 0(r4)
        lfd f4, 32(r3)       ;; much nicer.
        lfd f5, 64(r3)
        lfd f6, 56(r3)
        lfd f7, 48(r3)
        lfd f8, 40(r3)
        lfd f9, 24(r3)
        lfd f10, 16(r3)
        lfd f11, 16(r4)
        fmadd f0, f3, f2, f0
        fmul f2, f1, f4
        fmadd f0, f11, f10, f0
        fmadd f2, f3, f9, f2
        fmul f1, f1, f6
        stfd f0, 0(r4)
        fmadd f0, f11, f8, f2
        fmadd f1, f3, f7, f1
        stfd f0, 8(r4)
        fmadd f0, f11, f5, f1
        addi r6, r4, 24
        stfd f0, 16(r4)
        addi r2, r2, 1
        cmpw cr0, r2, r5
        or r4, r6, r6
        bne cr0, LBB3_2 ; no_exit

This is much nicer as it reduces register pressure in the loop a lot.  On X86,
this takes the function from having 9 spilled registers to 2.  This should help
some spec programs on X86 (gzip?)

This is currently only enabled with -enable-gep-isel-opt to allow perf testing
tonight.

llvm-svn: 24606
 2005-12-05 07:10:48 +00:00
Chris Lattner
8782b782cd dbg.stoppoint returns a value, don't forget to init it 
llvm-svn: 24583
 2005-12-03 18:50:48 +00:00
Nate Begeman
1064d6ec43 First chunk of actually generating vector code for packed types. These 
changes allow us to generate the following code:

_foo:
        li r2, 0
        lvx v0, r2, r3
        vaddfp v0, v0, v0
        stvx v0, r2, r3
        blr

for this llvm:

void %foo(<4 x float>* %a) {
entry:
        %tmp1 = load <4 x float>* %a
        %tmp2 = add <4 x float> %tmp1, %tmp1
        store <4 x float> %tmp2, <4 x float>* %a
        ret void
}

llvm-svn: 24534
 2005-11-30 08:22:07 +00:00
Reid Spencer
3fd1b4c9bf Fix a problem with llvm-ranlib that (on some platforms) caused the archive 
file to become corrupted due to interactions between mmap'd memory segments
and file descriptors closing. The problem is completely avoiding by using
a third temporary file.

Patch provided by Evan Jones

llvm-svn: 24527
 2005-11-30 05:21:10 +00:00
Chris Lattner
435b402e1f Add support for a new STRING and LOCATION node for line number support, patch 
contributed by Daniel Berlin, with a few cleanups here and there by me.

llvm-svn: 24515
 2005-11-29 06:21:05 +00:00
Nate Begeman
d37c13154a Check in code to scalarize arbitrarily wide packed types for some simple 
vector operations (load, add, sub, mul).

This allows us to codegen:
void %foo(<4 x float> * %a) {
entry:
  %tmp1 = load <4 x float> * %a;
  %tmp2 = add <4 x float> %tmp1, %tmp1
  store <4 x float> %tmp2, <4 x float> *%a
  ret void
}

on ppc as:
_foo:
        lfs f0, 12(r3)
        lfs f1, 8(r3)
        lfs f2, 4(r3)
        lfs f3, 0(r3)
        fadds f0, f0, f0
        fadds f1, f1, f1
        fadds f2, f2, f2
        fadds f3, f3, f3
        stfs f0, 12(r3)
        stfs f1, 8(r3)
        stfs f2, 4(r3)
        stfs f3, 0(r3)
        blr

llvm-svn: 24484
 2005-11-22 18:16:00 +00:00
Nate Begeman
07890bbec4 Rather than attempting to legalize 1 x float, make sure the SD ISel never 
generates it.  Make MVT::Vector expand-only, and remove the code in
Legalize that attempts to legalize it.

The plan for supporting N x Type is to continually epxand it in ExpandOp
until it gets down to 2 x Type, where it will be scalarized into a pair of
scalars.

llvm-svn: 24482
 2005-11-22 01:29:36 +00:00
Chris Lattner
19baba67b5 Unbreak codegen of bools. This should fix the llc/jit/llc-beta failures 
from last night.

llvm-svn: 24427
 2005-11-19 18:40:42 +00:00
Nate Begeman
b2e089c31b Teach LLVM how to scalarize packed types. Currently, this only works on 
packed types with an element count of 1, although more generic support is
coming.  This allows LLVM to turn the following code:

void %foo(<1 x float> * %a) {
entry:
  %tmp1 = load <1 x float> * %a;
  %tmp2 = add <1 x float> %tmp1, %tmp1
  store <1 x float> %tmp2, <1 x float> *%a
  ret void
}

Into:

_foo:
        lfs f0, 0(r3)
        fadds f0, f0, f0
        stfs f0, 0(r3)
        blr

llvm-svn: 24416
 2005-11-19 00:36:38 +00:00
Nate Begeman
127321b14c Split out the shift code from visitBinary. 
llvm-svn: 24412
 2005-11-18 07:42:56 +00:00
Chris Lattner
f2b62f317c when debugging lower dbg intrinsics to calls 
llvm-svn: 24377
 2005-11-16 07:22:30 +00:00
Andrew Lenharth
de1b5d6baa added a chain output 
llvm-svn: 24306
 2005-11-11 22:48:54 +00:00
Andrew Lenharth
01aa56397d continued readcyclecounter support 
llvm-svn: 24300
 2005-11-11 16:47:30 +00:00
Chris Lattner
cd6f0f47f2 Refactor intrinsic lowering stuff out of visitCall 
llvm-svn: 24261
 2005-11-09 19:44:01 +00:00
Chris Lattner
41fd6d5d27 Fix CodeGen/X86/shift-folding.ll:test3 on X86 
llvm-svn: 24256
 2005-11-09 16:50:40 +00:00
Chris Lattner
b7cad90e55 Avoid creating a token factor node in trivially redundant cases. This 
eliminates almost one node per block in common cases.

llvm-svn: 24254
 2005-11-09 05:03:03 +00:00
Chris Lattner
43535a19b1 Handle GEP's a bit more intelligently. Fold constant indices early and 
turn power-of-two multiplies into shifts early to improve compile time.

llvm-svn: 24253
 2005-11-09 04:45:33 +00:00
Nate Begeman
3ee3e69556 Add the necessary support to the ISel to allow targets to codegen the new 
alignment information appropriately.  Includes code for PowerPC to support
fixed-size allocas with alignment larger than the stack.  Support for
arbitrarily aligned dynamic allocas coming soon.

llvm-svn: 24224
 2005-11-06 09:00:38 +00:00
Chris Lattner
6871b23d02 Significantly simplify this code and make it more aggressive. Instead of having 
a special case hack for X86, make the hack more general: if an incoming argument
register is not used in any block other than the entry block, don't copy it to
a vreg.  This helps us compile code like this:

%struct.foo = type { int, int, [0 x ubyte] }
int %test(%struct.foo* %X) {
        %tmp1 = getelementptr %struct.foo* %X, int 0, uint 2, int 100
        %tmp = load ubyte* %tmp1                ; <ubyte> [#uses=1]
        %tmp2 = cast ubyte %tmp to int          ; <int> [#uses=1]
        ret int %tmp2
}

to:

_test:
        lbz r3, 108(r3)
        blr

instead of:

_test:
        lbz r2, 108(r3)
        or r3, r2, r2
        blr

The (dead) copy emitted to copy r3 into a vreg for extra-block uses was
increasing the live range of r3 past the load, preventing the coallescing.

This implements CodeGen/PowerPC/reg-coallesce-simple.ll

llvm-svn: 24115
 2005-10-30 19:42:35 +00:00
Nate Begeman
78afac2ddd Add the ability to lower return instructions to TargetLowering. This 
allows us to lower legal return types to something else, to meet ABI
requirements (such as that i64 be returned in two i32 regs on Darwin/ppc).

llvm-svn: 23802
 2005-10-18 23:23:37 +00:00
Chris Lattner
0a71a9ac86 Fix Generic/2005-10-18-ZeroSizeStackObject.ll by not requesting a zero 
sized stack object if either the array size or the type size is zero.

llvm-svn: 23801
 2005-10-18 22:14:06 +00:00
Chris Lattner
8396a308a7 remove hack 
llvm-svn: 23797
 2005-10-18 22:11:42 +00:00
Chris Lattner
bcfebebf22 Enable Nate's excellent DAG combiner work by default. This allows the 
removal of a bunch of ad-hoc and crufty code from SelectionDAG.cpp.

llvm-svn: 23682
 2005-10-10 16:47:10 +00:00
Chris Lattner
6bd8fd09b6 make sure that -view-isel-dags is the input to the isel, not the input to 
the second phase of dag combining

llvm-svn: 23631
 2005-10-05 06:09:10 +00:00
Jeff Cohen
f8a5e5ae6e Fix VC++ warnings. 
llvm-svn: 23579
 2005-10-01 03:57:14 +00:00
Chris Lattner
6f3b577ee6 Add FP versions of the binary operators, keeping the int and fp worlds seperate. 
Though I have done extensive testing, it is possible that this will break
things in configs I can't test.  Please let me know if this causes a problem
and I'll fix it ASAP.

llvm-svn: 23504
 2005-09-28 22:28:18 +00:00
Chris Lattner
0fd8f9fbc9 If the target prefers it, use _setjmp/_longjmp should be used instead of setjmp/longjmp for llvm.setjmp/llvm.longjmp. 
llvm-svn: 23481
 2005-09-27 22:15:53 +00:00
Chris Lattner
d4382f0afa If a function has liveins, and if the target requested that they be plopped 
into particular vregs, emit copies into the entry MBB.

llvm-svn: 23331
 2005-09-13 19:30:54 +00:00
Nate Begeman
007c650699 Add an option to the DAG Combiner to enable it for beta runs, and turn on 
that option for PowerPC's beta.

llvm-svn: 23253
 2005-09-07 00:15:36 +00:00
Chris Lattner
b0b4ec5655 Don't create zero sized stack objects even for array allocas with a zero 
number of elements.

llvm-svn: 23219
 2005-09-02 18:41:28 +00:00
Chris Lattner
b6cde17d29 Fix the release build, noticed by Eric van Riet Paap 
llvm-svn: 23215
 2005-09-02 07:09:28 +00:00
Chris Lattner
a66403dbf7 For values that are live across basic blocks and need promotion, use ANY_EXTEND 
instead of ZERO_EXTEND to eliminate extraneous extensions.  This eliminates
dead zero extensions on formal arguments and other cases on PPC, implementing
the newly tightened up test/Regression/CodeGen/PowerPC/small-arguments.ll test.

llvm-svn: 23205
 2005-09-02 00:19:37 +00:00
Chris Lattner
975f5c9f46 It is NDEBUG not _NDEBUG 
llvm-svn: 23186
 2005-09-01 18:44:10 +00:00
Chris Lattner
075250bda1 Disable this code, which broke many tests last night 
llvm-svn: 23114
 2005-08-27 16:16:51 +00:00
Chris Lattner
e7a2998064 Don't copy regs that are only used in the entry block into a vreg. This 
changes the code generated for:

short %test(short %A) {
  %B = xor short %A, -32768
  ret short %B
}

to:

_test:
        xori r2, r3, 32768
        xoris r2, r2, 65535
        extsh r3, r2
        blr

instead of:

_test:
        rlwinm r2, r3, 0, 16, 31
        xori r2, r3, 32768
        xoris r2, r2, 65535
        extsh r3, r2
        blr

llvm-svn: 23109
 2005-08-26 22:49:59 +00:00
Chris Lattner
13d7c252e5 Call the InsertAtEndOfBasicBlock hook if the usesCustomDAGSchedInserter 
flag is set on an instruction.

llvm-svn: 23098
 2005-08-26 20:54:47 +00:00
Chris Lattner
99282c7b92 Make -view-isel-dags show the dag before instruction selecting, in case 
the target isel crashes due to unimplemented features like calls :)

llvm-svn: 22997
 2005-08-24 00:34:29 +00:00