Commit Graph

20468 Commits

Author SHA1 Message Date
Chris Lattner
85dd3be735 Move CreateStackTemporary out to SelectionDAG
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42995 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 17:48:57 +00:00
Chris Lattner
37ce9df0da add a new CreateStackTemporary helper method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42994 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 17:47:20 +00:00
Chris Lattner
cfc24007b1 implement promotion of BR_CC operands, fixing bisort on ppc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42992 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 17:16:12 +00:00
Chris Lattner
1296e9d306 updates from duncan
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42991 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 16:46:29 +00:00
Devang Patel
b9e981304b Achieve same result but use fewer lines of code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42985 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 15:31:35 +00:00
Neil Booth
686700e325 Fast-track obviously over-large and over-small exponents during decimal->
integer conversion.  In some such cases this makes us one or two orders
of magnitude faster than NetBSD's libc.  Glibc seems to have a similar
fast path.

Also, tighten up some upper bounds to save a bit of memory.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42984 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 15:00:55 +00:00
Duncan Sands
65f8d3bf6b Fix some typos. Call getTypeToTransformTo rather than
getTypeToExpandTo.  The difference is that
getTypeToExpandTo gives the final result of expansion
(eg: i128 -> i32 on a 32 bit machine) while
getTypeToTransformTo does just one step (i128 -> i64).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42982 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 13:30:18 +00:00
Chris Lattner
01d029b82c One mundane change: Change ReplaceAllUsesOfValueWith to *optionally*
take a deleted nodes vector, instead of requiring it.

One more significant change:  Implement the start of a legalizer that
just works on types.  This legalizer is designed to run before the 
operation legalizer and ensure just that the input dag is transformed
into an output dag whose operand and result types are all legal, even
if the operations on those types are not.

This design/impl has the following advantages:

1. When finished, this will *significantly* reduce the amount of code in
   LegalizeDAG.cpp.  It will remove all the code related to promotion and
   expansion as well as splitting and scalarizing vectors.
2. The new code is very simple, idiomatic, and modular: unlike 
   LegalizeDAG.cpp, it has no 3000 line long functions. :)
3. The implementation is completely iterative instead of recursive, good
   for hacking on large dags without blowing out your stack.
4. The implementation updates nodes in place when possible instead of 
   deallocating and reallocating the entire graph that points to some 
   mutated node.
5. The code nicely separates out handling of operations with invalid 
   results from operations with invalid operands, making some cases
   simpler and easier to understand.
6. The new -debug-only=legalize-types option is very very handy :), 
   allowing you to easily understand what legalize types is doing.

This is not yet done.  Until the ifdef added to SelectionDAGISel.cpp is
enabled, this does nothing.  However, this code is sufficient to legalize
all of the code in 186.crafty, olden and freebench on an x86 machine.  The
biggest issues are:

1. Vectors aren't implemented at all yet
2. SoftFP is a mess, I need to talk to Evan about it.
3. No lowering to libcalls is implemented yet.
4. Various operations are missing etc.
5. There are FIXME's for stuff I hax0r'd out, like softfp.

Hey, at least it is a step in the right direction :).  If you'd like to help,
just enable the #ifdef in SelectionDAGISel.cpp and compile code with it.  If
this explodes it will tell you what needs to be implemented.  Help is 
certainly appreciated.

Once this goes in, we can do three things:

1. Add a new pass of dag combine between the "type legalizer" and "operation
   legalizer" passes.  This will let us catch some long-standing isel issues
   that we miss because operation legalization often obfuscates the dag with
   target-specific nodes.
2. We can rip out all of the type legalization code from LegalizeDAG.cpp,
   making it much smaller and simpler.  When that happens we can then 
   reimplement the core functionality left in it in a much more efficient and
   non-recursive way.
3. Once the whole legalizer is non-recursive, we can implement whole-function
   selectiondags maybe...



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42981 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 06:10:22 +00:00
Chris Lattner
3cb9351e8a One xform performed by LegalizeDAG is transformation of "store of fp" to "store of int".
Make two changes:
1) only xform "store of f32" if i32 is a legal type for the target.
2) only xform "store of f64" if either i64 or i32 are legal for the target.
3) if i64 isn't legal, manually lower to 2 stores of i32 instead of letting a
   later pass of legalize do it.  This is ugly, but helps future changes I'm 
   about to commit.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42980 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 05:46:06 +00:00
Chris Lattner
3cab1c72b5 avoid an APFloat copy.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42979 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 05:34:10 +00:00
Chris Lattner
59afba0fe9 Add a (disabled by default) way to view the ID of a node.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42978 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 05:32:43 +00:00
Dale Johannesen
6e64473603 Handle PPC long double in CBackend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42972 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-15 01:05:37 +00:00
Chris Lattner
7a3c85583d remove misleading comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42970 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 20:35:12 +00:00
Chris Lattner
fb5b11046b If a target doesn't have HasMULHU or HasUMUL_LOHI, ExpandOp would return
without lo/hi set.  Fall through to making a libcall instead.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42969 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 18:35:05 +00:00
Neil Booth
e5e0194583 Consolidate logic for creating NaNs. Silence compiler warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42966 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 10:39:51 +00:00
Neil Booth
caf19d79e5 Whether arithmetic is supported is a property of the semantics. Make it
so, and clean up the checks by putting them in an inline function.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42965 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 10:29:28 +00:00
Neil Booth
1870f29c6c Separate out parsing of decimal number. Use this to only allocate
memory for the significand once up-front.  Also ignore insignificant
trailing zeroes; this saves unnecessary multiplications later.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42964 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 10:16:12 +00:00
Evan Cheng
3644601563 Unbreak x86-64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42962 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 10:09:39 +00:00
Evan Cheng
34729256e8 When coalescing an EXTRACT_SUBREG and the dst register is a physical register,
the source register will be coalesced to the super register of the LHS. Properly
merge in the live ranges of the resulting coalesced interval that were part of
the original source interval to the live interval of the super-register.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42961 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 10:08:34 +00:00
Evan Cheng
347d39f1fd Revert 42908 for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42960 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 05:57:21 +00:00
Dale Johannesen
296c176141 Fix type mismatch error in PPC Altivec (only causes
a problem when asserts are on).  From vecLib.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42959 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 01:58:32 +00:00
Dale Johannesen
5927d8e94d Disable some compile-time optimizations on PPC
long double.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42958 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-14 01:56:47 +00:00
Duncan Sands
7fef59f7e5 Clarify that fastcc has a problem with nested function
trampolines, rather than with nested functions themselves.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42955 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 07:38:37 +00:00
Chris Lattner
e33544ce55 Enhance the truncstore optimization code to handle shifted
values and propagate demanded bits through them in simple cases.

This allows this code:
void foo(char *P) {
   strcpy(P, "abc");
}
to compile to:

_foo:
        ldrb r3, [r1]
        ldrb r2, [r1, #+1]
        ldrb r12, [r1, #+2]!
        ldrb r1, [r1, #+1]
        strb r1, [r0, #+3]
        strb r2, [r0, #+1]
        strb r12, [r0, #+2]
        strb r3, [r0]
        bx lr

instead of:

_foo:
        ldrb r3, [r1, #+3]
        ldrb r2, [r1, #+2]
        orr r3, r2, r3, lsl #8
        ldrb r2, [r1, #+1]
        ldrb r1, [r1]
        orr r2, r1, r2, lsl #8
        orr r3, r2, r3, lsl #16
        strb r3, [r0]
        mov r2, r3, lsr #24
        strb r2, [r0, #+3]
        mov r2, r3, lsr #16
        strb r2, [r0, #+2]
        mov r3, r3, lsr #8
        strb r3, [r0, #+1]
        bx lr

testcase here: test/CodeGen/ARM/truncstore-dag-combine.ll

This also helps occasionally for X86 and other cases not involving 
unaligned load/stores.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42954 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 06:58:48 +00:00
Chris Lattner
2b4c279a8e Add a simple optimization to simplify the input to
truncate and truncstore instructions, based on the 
knowledge that they don't demand the top bits.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42952 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 06:35:54 +00:00
Neil Booth
d1a23d573d If the power of 5 is exact, and the reciprocal exact, the error is zero not one half-ulps. This prevents an infinite loop in rare cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42950 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 03:34:08 +00:00
Evan Cheng
7f56625447 Local spiller optimization:
Turn this:
movswl  %ax, %eax
movl    %eax, -36(%ebp)
xorl    %edi, -36(%ebp)
into
movswl  %ax, %eax
xorl    %edi, %eax
movl    %eax, -36(%ebp)
by unfolding the load / store xorl into an xorl and a store when we know the
value in the spill slot is available in a register. This doesn't change the
number of instructions but reduce the number of times memory is accessed.

Also unfold some load folding instructions and reuse the value when similar
situation presents itself.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42947 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 02:50:24 +00:00
Evan Cheng
106e8020bd Change unfoldMemoryOperand(). User is now responsible for passing in the
register used by the unfolded instructions. User can also specify whether to
unfold the load, the store, or both.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42946 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 02:35:06 +00:00
Evan Cheng
fa9457276a Optionally create a MachineInstr without default implicit operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42945 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 02:23:01 +00:00
Arnold Schwaighofer
92226dd19f Correcting the corrections. Bad bad baaad emacs!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42935 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 21:53:12 +00:00
Arnold Schwaighofer
48abc5cf6b Corrected many typing errors. And removed 'nest' parameter handling
for fastcc from X86CallingConv.td.  This means that nested functions
are not supported for calling convention 'fastcc'.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42934 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 21:30:57 +00:00
Devang Patel
7b9e1d251f Dest type is always i8 *. This allows some simplification.
Do not filter memmove.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42930 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 20:10:21 +00:00
Duncan Sands
2e4d675d93 Due to the new tail call optimization, trampolines can no
longer be created for fastcc functions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42925 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 19:37:31 +00:00
Dale Johannesen
f646774edd ppc long double. Implement fabs and fneg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42924 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 19:02:17 +00:00
Evan Cheng
1c5d83c14e Update.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42922 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 18:22:55 +00:00
Chris Lattner
a79dd432b3 Fix a bug in my patch last night that broke InstCombine/2007-10-12-Crash.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42920 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 18:05:47 +00:00
Dale Johannesen
6e63e09236 Implement i64->ppcf128 conversions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42919 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 17:52:03 +00:00
Evan Cheng
48ff282dd4 Did mean to leave this in. INSERT_SUBREG isn't being coalesced yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42916 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 17:16:50 +00:00
Neil Booth
6ac7016eb1 Remove duplicate comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42913 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 16:05:57 +00:00
Neil Booth
96c7471b39 Implement correctly-rounded decimal->binary conversion, i.e. conversion
from user input strings.

Such conversions are more intricate and subtle than they may appear;
it is unlikely I have got it completely right first time.  I would
appreciate being informed of any bugs and incorrect roundings you
might discover.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42912 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 16:02:31 +00:00
Neil Booth
5b8e0c5fd3 Remove a field that was never used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42911 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 15:35:10 +00:00
Neil Booth
7a951ca548 If we're trying to be arbitrary precision, unsigned char clearly won't cut it. Needed for dec->bin conversions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42910 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 15:33:27 +00:00
Neil Booth
1e8390d8d6 Don't attempt to mask no bits
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42909 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 15:31:31 +00:00
Dan Gohman
8ddde0a151 Change the names used for internal labels to use the current
function symbol name instead of a codegen-assigned function
number.

Thanks Evan! :-)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42908 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 14:53:36 +00:00
Dan Gohman
a9b511187b Fix some corner cases with vectors in copyToRegs and copyFromRegs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42907 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 14:33:11 +00:00
Dan Gohman
9e04c82c0e Add support to SplitVectorOp for powi, where the second operand
is a scalar integer.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42906 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 14:13:46 +00:00
Dan Gohman
f0d0089340 Mark vector ctpop, cttz, and ctlz as Expand on x86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42905 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 14:09:42 +00:00
Dan Gohman
a3f269f7f6 Mark vector pow, ctpop, cttz, and ctlz as Expand on PowerPC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42904 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 14:08:57 +00:00
Evan Cheng
52c7ff7067 Restrict EXTRACT_SUBREG coalescing to avoid negative performance impact.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42903 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 09:15:53 +00:00
Evan Cheng
32dfbeada7 EXTRACT_SUBREG coalescing support. The coalescer now treats EXTRACT_SUBREG like
(almost) a register copy. However, it always coalesced to the register of the
RHS (the super-register). All uses of the result of a EXTRACT_SUBREG are sub-
register uses which adds subtle complications to load folding, spiller rewrite,
etc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42899 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-12 08:50:34 +00:00