c7f06ca780
tolerate unimplemented codegen better
...
llvm-svn: 41424
2007-08-26 05:06:40 +00:00
3d966d6556
Teach emit-llvm for scalars to properly handle compound assignment
...
operators in all their glory :)
llvm-svn: 41373
2007-08-24 21:00:35 +00:00
2da04b3322
completely refactor codegen of scalar expressions out into its own CGExprScalar.cpp file.
...
This patch temporarily breaks compound assignment operators, but greatly simplifies many
things.
llvm-svn: 41355
2007-08-24 05:35:26 +00:00
8ad56707ea
stub out complex -> bool conversion.
...
llvm-svn: 41347
2007-08-24 00:01:20 +00:00
08b15df766
implement passing of complex and aggregates through call args.
...
llvm-svn: 41344
2007-08-23 23:43:33 +00:00
76f4a902d7
Implement parsing and code generation of Objective-C string literals.
...
llvm-svn: 41238
2007-08-21 17:43:55 +00:00
4b0e7873f6
move EmitLoadOfComplex/EmitStoreOfComplex into ComplexExprEmitter.
...
llvm-svn: 41236
2007-08-21 17:28:34 +00:00
3cf417b369
implement comma for complex.
...
llvm-svn: 41235
2007-08-21 17:15:50 +00:00
64be48fede
and/or/xor are invalid for complex, even integer complex apparently.
...
llvm-svn: 41234
2007-08-21 17:12:50 +00:00
bf1bd0dcc7
simplify code slightly
...
llvm-svn: 41233
2007-08-21 17:03:38 +00:00
96d7256d62
reimplement support for complex comparisons, add support for integer complex compares.
...
llvm-svn: 41231
2007-08-21 16:57:55 +00:00
6ce75dff4d
reimplement complex mul
...
llvm-svn: 41226
2007-08-21 16:34:16 +00:00
835635d85a
reimplement addition of complex numbers.
...
llvm-svn: 41215
2007-08-21 04:59:27 +00:00
76ba849ed3
Fix array->pointer decay. This unbreaks test/CodeGen/array.c
...
llvm-svn: 41202
2007-08-20 22:37:10 +00:00
1d8e521022
Add support for code generation of builtins.
...
llvm-svn: 41188
2007-08-20 18:05:56 +00:00
c81614d5d1
Modified ArraySubscriptExpr to have accessors getLHS and getRHS in addition
...
to getBase and getIdx. getBase and getIdx now return a "normalized" view
of the expression (e.g., always "A[4]" instead of possibly "4[A]"). getLHS
and getRHS return the expressions with syntactic fidelity to the original
source code.
Also modified client code of ArraySubscriptExpr, including the AST dumper
and pretty printer, the return-stack value checker, and the LLVM code
generator.
llvm-svn: 41180
2007-08-20 16:18:38 +00:00
6278e6ac30
start splitting out aggregate value computation from EmitExpr into EmitAggExpr.
...
aggregate value and scalar expression computation are very different, this
gets them away from each other. This causes a temporary regression on some
complex number examples.
llvm-svn: 41014
2007-08-11 00:04:45 +00:00
8eab8ff127
fix a codegen bug handling ocuvector element exprs.
...
llvm-svn: 40995
2007-08-10 17:10:08 +00:00
90d9120453
implement initial codegen for aggregate return functions. This implements
...
codegen for:
_Complex double bar(int);
void test(_Complex double*);
void test2(int c) {
_Complex double X;
X = bar(1);
test(&X);
}
llvm-svn: 40993
2007-08-10 17:02:28 +00:00
02697701d9
Fix a bug handling function -> pointer decay and avoid emitting a noop
...
bitcast.
llvm-svn: 40992
2007-08-10 16:33:59 +00:00
87f4836184
now that implicit conversions are explicit, we can eliminate
...
EmitUsualArithmeticConversions.
llvm-svn: 40931
2007-08-08 17:49:18 +00:00
5ebb2fed89
now that implicit conversions are explicit, we can eliminate
...
EmitExprWithUsualUnaryConversions.
llvm-svn: 40929
2007-08-08 17:43:05 +00:00
81a9688e93
Implement codegen for __builtin_choose_expr. For example:
...
struct X { int A; };
void foo() {
struct X s;
int i;
i = __builtin_choose_expr(0, s, i);
}
compiles to:
%tmp = load i32* %i ; <i32> [#uses=1]
store i32 %tmp, i32* %i
wow :)
llvm-svn: 40801
2007-08-04 00:20:15 +00:00
4048005ad8
implement codegen support for __builtin_types_compatible_p
...
llvm-svn: 40788
2007-08-03 17:51:03 +00:00
d268a7a268
Rename AddrLabel and OCUVectorComponent -> AddrLabelExpr and OCUVectorElementExpr respectively. This is for consistency with other expr nodes end with *Expr.
...
llvm-svn: 40785
2007-08-03 17:31:20 +00:00
3a44aa7461
implement codegen for multidest ocuvector expressions, like:
...
vec2.yx = vec2; // reverse
llvm-svn: 40782
2007-08-03 16:37:04 +00:00
41d480e4d8
add codegen support for storing into a single-element ocu lvalue, such as:
...
vec2.x = f;
llvm-svn: 40781
2007-08-03 16:28:33 +00:00
40ff701674
refactor handling of ocuvector lvalue->rvalue codegen into its own method.
...
llvm-svn: 40780
2007-08-03 16:18:34 +00:00
fb837dccac
In the common case where we are shuffling a vector, emit an
...
llvm vector shuffle instead of a bunch of insert/extract operations.
For: vec4 = vec4.yyyy; // splat
Emit:
%tmp1 = shufflevector <4 x float> %tmp, <4 x float> undef, <4 x i32> < i32 1, i32 1, i32 1, i32 1 >
instead of:
%tmp1 = extractelement <4 x float> %tmp, i32 1
%tmp2 = insertelement <4 x float> undef, float %tmp1, i32 0
%tmp3 = extractelement <4 x float> %tmp, i32 1
%tmp4 = insertelement <4 x float> %tmp2, float %tmp3, i32 1
%tmp5 = extractelement <4 x float> %tmp, i32 1
%tmp6 = insertelement <4 x float> %tmp4, float %tmp5, i32 2
%tmp7 = extractelement <4 x float> %tmp, i32 1
%tmp8 = insertelement <4 x float> %tmp6, float %tmp7, i32 3
llvm-svn: 40779
2007-08-03 16:09:33 +00:00
a1036f9155
Add support for scalar-returning element accesses like V.x
...
llvm-svn: 40777
2007-08-03 15:52:31 +00:00
73ab9b3c14
implement lvalue to rvalue conversion for ocuvector components. We can now compile stuff
...
like this:
typedef __attribute__(( ocu_vector_type(4) )) float float4;
float4 test1(float4 V) {
return V.wzyx+V;
}
to:
_test1:
pshufd $27, %xmm0, %xmm1
addps %xmm0, %xmm1
movaps %xmm1, %xmm0
ret
and:
_test1:
mfspr r2, 256
oris r3, r2, 4096
mtspr 256, r3
li r3, lo16(LCPI1_0)
lis r4, ha16(LCPI1_0)
lvx v3, r4, r3
vperm v3, v2, v2, v3
vaddfp v2, v3, v2
mtspr 256, r2
blr
llvm-svn: 40771
2007-08-03 00:16:29 +00:00
9e751cae27
add support for codegen of an OCUVectorComponent as an lvalue.
...
We can now codegen:
vec4.xy;
as nothing!
llvm-svn: 40769
2007-08-02 23:37:31 +00:00
7b9f04eb19
update this to build with LLVM ToT
...
llvm-svn: 40665
2007-08-01 06:24:52 +00:00
625bfc8716
Implement code generation for __func__, __FUNCTION__ and __PRETTY_FUNCTION__
...
llvm-svn: 40162
2007-07-21 05:21:51 +00:00
3f8c6e6e74
implement codegen support for sizeof/alignof
...
llvm-svn: 40009
2007-07-18 18:12:07 +00:00
651f0e9c8f
Add a hack (mirroring llvm-gcc) to pointer difference
...
codegen to compile:
int test(int *A, int *B) {
return A-B;
}
into:
_test:
movl 4(%esp), %eax
subl 8(%esp), %eax
sarl $2, %eax
ret
instead of:
_test:
movl 4(%esp), %eax
subl 8(%esp), %eax
movl %eax, %ecx
sarl $31, %ecx
shrl $30, %ecx
addl %ecx, %eax
sarl $2, %eax
ret
llvm-svn: 39902
2007-07-16 05:43:05 +00:00
0e9d6226ca
Refactor code so that isIntegerConstantExpr has an ASTContext available.
...
llvm-svn: 39884
2007-07-15 23:26:56 +00:00
e97cd7e65c
add FIXME and un-XFAIL test
...
llvm-svn: 39858
2007-07-14 20:05:18 +00:00
4481b428db
A significant refactoring of the type size stuff to also
...
compute type alignment. This info is needed for struct layout.
llvm-svn: 39850
2007-07-14 01:29:45 +00:00
027f21dd72
add a fixme
...
llvm-svn: 39847
2007-07-14 00:01:01 +00:00
d4606aa36e
implement _Complex * == and !=
...
llvm-svn: 39841
2007-07-13 23:33:18 +00:00
983a8bbbb2
Move getSize() out of type, into ASTContext, where it has target info, and
...
where ASTContext can manage caches for struct layout, etc.
llvm-svn: 39835
2007-07-13 22:13:22 +00:00
388cf760e2
implement codegen support for implicit casts.
...
llvm-svn: 39831
2007-07-13 20:25:53 +00:00
6e9d9b350e
"Codegen for Character Literals and Conditional Operator
...
Both in one patch, and the test case that Chris didn't commit last
time is in there too...
I'll split the patch up if somebody wants it split."
Patch by Keith Bauer.
llvm-svn: 39796
2007-07-13 05:18:11 +00:00
fc7634f2ab
"Someone typed "PtrToInt" where they meant "IntToPtr".
...
I've added a tests/CodeGen directory, and a test for this case that
used to fail and now passes."
Patch by Keith Bauer
llvm-svn: 39794
2007-07-13 03:25:53 +00:00
d2b88ab313
Implement codegen for + and - with pointers. Patch contributed by
...
Keith Bauer.
llvm-svn: 39793
2007-07-13 03:05:23 +00:00
dcca487cfa
implement codegen support for pre/post inc/dec.
...
llvm-svn: 39765
2007-07-11 23:43:46 +00:00
c14236b8ae
implement codegen support for the "default argument promotions" (C99 6.5.2.2p6).
...
Not having this prevented promoting float arguments to double when passed
into printf, for example.
llvm-svn: 39727
2007-07-10 22:18:37 +00:00
a779b3df28
implement codegen support for rvalue-only vector subscripts, such as:
...
float4 test(void);
float test2() {
return test()[1];
}
llvm-svn: 39725
2007-07-10 21:58:36 +00:00
08c4b9ffec
Add support for codegen'ing vector subscripts, at least when they are lvalues.
...
llvm-svn: 39724
2007-07-10 21:17:59 +00:00
2ada32ed7d
implement codegen support for FP literals
...
llvm-svn: 39718
2007-07-09 23:03:16 +00:00
47c247e7bf
add codegen support for <<= and >>=.
...
llvm-svn: 39713
2007-06-29 17:26:27 +00:00
b25a94383a
Implement the rest of the compound assignment operators, except shifts.
...
llvm-svn: 39712
2007-06-29 17:03:06 +00:00
cd215f00ee
refactor some code, implement -=
...
llvm-svn: 39711
2007-06-29 16:52:55 +00:00
9369a563b4
Rename ArithAssignBinaryOperator -> CompoundAssignOperator, implement
...
codegen support for +=.
llvm-svn: 39710
2007-06-29 16:31:29 +00:00
e9a6453ded
add some infrastructure for codegen'ing complex numbers. implement addition
...
of complex. We now produce optimal code for:
void test(_Complex float *Y) {
*Y = *Y + *Y;
}
$ clang -emit-llvm cg.c | llvm-as | opt -std-compile-opts | llc -march=x86-64
...
_test:
movss (%rdi), %xmm0
addss %xmm0, %xmm0
movss 4(%rdi), %xmm1
movss %xmm0, (%rdi)
addss %xmm1, %xmm1
movss %xmm1, 4(%rdi)
ret
llvm-svn: 39673
2007-06-22 21:44:33 +00:00
c6395936ae
Split complex types out from being members of BuiltinType to being their own
...
types.
llvm-svn: 39672
2007-06-22 20:56:16 +00:00
f033c147c9
remove location tracking for target info
...
llvm-svn: 39671
2007-06-22 19:05:19 +00:00
09153c0a8c
Build enough support for aggregates to be able to compile this:
...
void test(int A, _Complex float Y) {
_Complex float X;
X = X;
}
llvm-svn: 39669
2007-06-22 18:48:09 +00:00
1fde0b345b
implement codegen of the relational operations
...
One major FIXME though.
llvm-svn: 39666
2007-06-20 18:30:55 +00:00
273c63d450
Implement the equality operators for simple types
...
llvm-svn: 39665
2007-06-20 18:02:30 +00:00
b6984c4854
Hook up global function and variable handling. We can now compile:
...
int X, bar(int,int,int);
short Y;
double foo() {
return bar(X, Y, 3);
}
into:
@X = external global i32 ; <i32*> [#uses=1]
@Y = external global i16 ; <i16*> [#uses=1]
define double @foo() {
entry:
%tmp = load i32* @X ; <i32> [#uses=1]
%tmp1 = load i16* @Y ; <i16> [#uses=1]
%promote = sext i16 %tmp1 to i32 ; <i32> [#uses=1]
%call = tail call i32 @bar( i32 %tmp, i32 %promote, i32 3 ) ; <i32> [#uses=1]
%conv = sitofp i32 %call to double ; <double> [#uses=1]
ret double %conv
}
declare i32 @bar(i32, i32, i32)
llvm-svn: 39663
2007-06-20 04:44:43 +00:00
23b7eb677d
Finally bite the bullet and make the major change: split the clang namespace
...
out of the llvm namespace. This makes the clang namespace be a sibling of
llvm instead of being a child.
The good thing about this is that it makes many things unambiguous. The
bad things is that many things in the llvm namespace (notably data structures
like smallvector) now require an llvm:: qualifier. IMO, libsystem and libsupport
should be split out of llvm into their own namespace in the future, which will fix
this issue.
llvm-svn: 39659
2007-06-15 23:05:46 +00:00
2b228c95aa
implement codegen of calls. We can now compile:
...
double foo( int (*FP)(int, int, int)) {
return FP(1, 2, 3);
}
to:
define double @foo(i32 (i32, i32, i32)* %FP) {
entry:
%call = tail call i32 %FP( i32 1, i32 2, i32 3 )
%conv = sitofp i32 %call to double
ret double %conv
}
llvm-svn: 39658
2007-06-15 21:34:29 +00:00
b48238188e
Fix a nasty bug which caused infinite recursion
...
llvm-svn: 39655
2007-06-15 21:04:38 +00:00
53621a535d
Implement support for formal arguments. We can now compile this:
...
int test(int X, short Y, float Z) {
return (int)(X*Y+Z);
}
to:
define i32 @test(i32 %X, i16 %Y, float %Z) {
entry:
%promote = sext i16 %Y to i32 ; <i32> [#uses=1]
%mul = mul i32 %promote, %X ; <i32> [#uses=1]
%promote3 = sitofp i32 %mul to float ; <float> [#uses=1]
%add = add float %promote3, %Z ; <float> [#uses=1]
%conv = fptosi float %add to i32 ; <i32> [#uses=1]
ret i32 %conv
}
with:
$ clang -emit-llvm t.c | llvm-as | opt -std-compile-opts | llvm-dis
llvm-svn: 39652
2007-06-13 20:44:40 +00:00
f99b3f5ec2
Emit codegen of enum literals.
...
llvm-svn: 39646
2007-06-11 03:52:52 +00:00
d9d2fb1420
Implement array subscripts for non-vla types.
...
llvm-svn: 39622
2007-06-08 23:31:14 +00:00
4347e369b4
implement codegen of string literals.
...
llvm-svn: 39597
2007-06-06 04:54:52 +00:00
83b484b3be
implement the real int/fp conversions
...
llvm-svn: 39596
2007-06-06 04:39:08 +00:00
cf106ab42a
implement support for casts to/from pointers.
...
llvm-svn: 39595
2007-06-06 04:05:39 +00:00
8394d795c3
implement codegen of a bunch more loop constructs and most expressions
...
llvm-svn: 39593
2007-06-05 20:53:16 +00:00
946aa31f02
implement codegen of while stmts and lvalue evaluation of paren exprs :)
...
llvm-svn: 39582
2007-06-05 03:59:43 +00:00
b16f455e8c
Type::isSignedInteger() and isUnsignedInteger() did not properly account for
...
'char', which varies based on the target.
Instead of spreading target knowledge throughout the compiler, bifurcate char
into Char_S and Char_U, and have ASTContext create the right one based on the
target, when it starts up.
llvm-svn: 39577
2007-06-03 07:25:34 +00:00
cf25024828
Implement EmitUsualArithmeticConversions, so we can add shorts to floats and
...
ints to long long etc. For int to longlong, we now get:
%tmp = load i64* %F ; <i64> [#uses=1]
%tmp1 = load i32* %D ; <i32> [#uses=1]
%promote = sext i32 %tmp1 to i64 ; <i64> [#uses=1]
%tmp2 = add i64 %tmp, %promote ; <i64> [#uses=0]
llvm-svn: 39576
2007-06-03 02:02:44 +00:00
c18f9d1aed
Change EmitUsualArithmeticConversions to use EmitExprWithUsualUnaryConversions.
...
This allows us to compile:
register short X;
{
int Y;
return 1+X+Y;
into:
%tmp = load i16* %X ; <i16> [#uses=1]
%promote = sext i16 %tmp to i32 ; <i32> [#uses=1]
%tmp1 = add i32 1, %promote ; <i32> [#uses=1]
%tmp2 = load i32* %Y ; <i32> [#uses=1]
%tmp3 = add i32 %tmp1, %tmp2 ; <i32> [#uses=1]
ret i32 %tmp3
Look at the amazing sign extension, so much happier than an assertion failure. :)
llvm-svn: 39565
2007-06-02 22:51:30 +00:00
6db1fb845a
implement a first hack at codegen'ing the usual unary conversions.
...
This allows us to compile:
int func() {
int A[10];
if (!A) {
to:
define i32 @func() {
entry:
%A = alloca [10 x i32] ; <[10 x i32]*> [#uses=1]
%arraydecay = getelementptr [10 x i32]* %A, i32 0, i32 0 ; <i32*> [#uses=1]
%tobool = icmp ne i32* %arraydecay, null ; <i1> [#uses=1]
%lnot = xor i1 %tobool, true ; <i1> [#uses=1]
br i1 %lnot, label %ifthen, label %ifend
-Chris
llvm-svn: 39564
2007-06-02 22:49:07 +00:00
a45c5af87b
Implement a trivial optimization to reduce the number of compares emitted.
...
For:
register short X;
if (!X) {
We now produce:
%tmp = load i16* %X ; <i16> [#uses=1]
%tobool = icmp ne i16 %tmp, 0 ; <i1> [#uses=1]
%lnot = xor i1 %tobool, true ; <i1> [#uses=1]
br i1 %lnot, label %ifthen, label %ifend
instead of:
%tmp = load i16* %X ; <i16> [#uses=1]
%tobool = icmp ne i16 %tmp, 0 ; <i1> [#uses=1]
%lnot = xor i1 %tobool, true ; <i1> [#uses=1]
%lnot.ext = zext i1 %lnot to i32 ; <i32> [#uses=1]
%tobool1 = icmp ne i32 %lnot.ext, 0 ; <i1> [#uses=1]
br i1 %tobool1, label %ifthen, label %ifend
llvm-svn: 39560
2007-06-02 19:47:04 +00:00
f0106d2578
Refactor EvaluateScalarValueToBool out of if statement emission, so it can
...
be shared.
Implement infrastructure for unary operator emission.
Implement basic logical not support. We now compile:
register short X;
if (!X) {
into:
%tmp = load i16* %X ; <i16> [#uses=1]
%tobool = icmp ne i16 %tmp, 0 ; <i1> [#uses=1]
%lnot = xor i1 %tobool, true ; <i1> [#uses=1]
zext i1 %lnot to i32 ; <i32>:0 [#uses=1]
%tobool1 = icmp ne i32 %0, 0 ; <i1> [#uses=1]
br i1 %tobool1, label %ifthen, label %ifend
llvm-svn: 39559
2007-06-02 19:33:17 +00:00
d7f58867e5
Implement scaffolding for lvalues. Implement block vardecl lvalues.
...
This allows us to translate:
int func() {
register int X;
{
int Y;
return 1+X+Y;
}
}
into:
define i32 @func() {
entry:
%X = alloca i32 ; <i32*> [#uses=1]
%Y = alloca i32 ; <i32*> [#uses=1]
%allocapt = bitcast i32 undef to i32 ; <i32> [#uses=0]
%tmp = load i32* %X ; <i32> [#uses=1]
%tmp1 = add i32 1, %tmp ; <i32> [#uses=1]
%tmp2 = load i32* %Y ; <i32> [#uses=1]
%tmp3 = add i32 %tmp1, %tmp2 ; <i32> [#uses=1]
ret i32 %tmp3
; No predecessors!
ret i32 undef
}
llvm-svn: 39555
2007-06-02 05:24:33 +00:00
db91b16755
stub out codegen of binary plus. We now compile:
...
if (11 + 42) {
to:
%tmp = add i32 11, 42 ; <i32> [#uses=1]
%tobool = icmp ne i32 %tmp, 0 ; <i1> [#uses=1]
br i1 %tobool, label %ifthen, label %ifend
but this doesn't handle any of the interesting/hard stuff yet.
llvm-svn: 39545
2007-06-02 00:16:28 +00:00
e47e440c42
split stmt/expr codegen into their own files.
...
llvm-svn: 39540
2007-06-01 18:02:12 +00:00
Chris Lattner