RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 04:02:41 +01:00
Code Issues Projects Releases Wiki Activity
43,170 Commits 4 Branches 5 Tags 704 MiB
a3b718a3c9
Commit Graph

746 Commits

Author SHA1 Message Date
Eli Friedman
46782c75fe Remove unnecessary target lines. 
llvm-svn: 52260
 2008-06-13 22:10:32 +00:00
Eli Friedman
11d4c94933 Don't skip over instructions other than loads that might read memory 
when trying to sink stores.

llvm-svn: 52259
 2008-06-13 22:02:12 +00:00
Dan Gohman
7f1fcc4339 Protect ChangeCompareStride from situations in which it is possible 
for it to generate use-before-def IR, such as in this testcase.

llvm-svn: 52258
 2008-06-13 21:43:41 +00:00
Eli Friedman
d38a639deb Make sure SimplifyStoreAtEndOfBlock doesn't mess with loops; the 
structure checks are incorrect if the blocks aren't distinct.
Fixes PR2435.

llvm-svn: 52257
 2008-06-13 21:17:49 +00:00
Evan Cheng
66ce588b87 Fix some tests. 
llvm-svn: 52245
 2008-06-12 21:23:38 +00:00
Evan Cheng
b0d847cf05 Revert 52223. 
llvm-svn: 52243
 2008-06-12 20:55:39 +00:00
Matthijs Kooijman
3a11ccc589 Add line continuation character so the avoid dup loop header test actually runs. 
llvm-svn: 52228
 2008-06-12 08:49:04 +00:00
Evan Cheng
9b0c0a0f00 Avoid duplicating loop header which leads to unnatural loops (and just seem like general badness to me, likely to cause code explosion). 
Patch by Florian Brandner.

llvm-svn: 52223
 2008-06-11 19:07:54 +00:00
Matthijs Kooijman
0f9df32e12 Teach instruction combining about the extractvalue. It can succesfully fold 
useless insert-extract chains, similar to how it folds them for vectors.

Add a testcase for this.

llvm-svn: 52217
 2008-06-11 14:05:05 +00:00
Matthijs Kooijman
3488e4542b Ignore stderr for some more tests that expect warnings there. 
This fixes 2 testcases.

llvm-svn: 52184
 2008-06-10 16:13:38 +00:00
Matthijs Kooijman
e8fb62fb3c Fix some escaping and quoting in RUN lines, mainly involving { and <. In two 
cases quoting of <{ didn't work out, so I changed the grep to check for }>
instead.

This fixes 7 testcases that were not properly running before.

llvm-svn: 52182
 2008-06-10 16:04:47 +00:00
Matthijs Kooijman
15ab3c5f19 Let some more tests ignore expected output on stderr. 
Also, use > %t instead of -o %t for output in one test since that also works
when %t already exists.

This fixes 6 testcases.

llvm-svn: 52178
 2008-06-10 15:04:14 +00:00
Dan Gohman
9eace09bfa Fix two more not-grep tests that were missing llvm-dis. 
llvm-svn: 52159
 2008-06-09 22:36:45 +00:00
Duncan Sands
a15ae3d239 Test that prune-eh doesn't make deductions based 
on bodies of functions with weak linkage.

llvm-svn: 52141
 2008-06-09 11:28:41 +00:00
Chris Lattner
806f0a8411 lower calls to abs to inline code, PR2337 
llvm-svn: 52138
 2008-06-09 08:26:51 +00:00
Chris Lattner
7864575654 Fix PR2411, where ip constant prop would propagate the 
result of a weak function.

llvm-svn: 52137
 2008-06-09 07:58:07 +00:00
Chris Lattner
4a896996cb Limit the icmp+phi merging optimization to the cases where it is profitable: 
don't make i1 phis when it won't be possible to eliminate them.

llvm-svn: 52097
 2008-06-08 20:52:11 +00:00
Evan Cheng
c7ed1b9258 Speculatively execute a block when the the block is the then part of a triangle shape and it contains a single, side effect free, cheap instruction. The branch is eliminated by adding a select instruction. i.e. 
Turn                                                                                                                                                                                                       
BB:                                                                                                                                                                                                        
    %t1 = icmp                                                                                                                                                                                             
    br i1 %t1, label %BB1, label %BB2                                                                                                                                                                      
BB1:                                                                                                                                                                                                       
    %t3 = add %t2, c                                                                                                                                                                                       
    br label BB2                                                                                                                                                                                           
BB2:                                                                                                                                                                                                       
=>                                                                                                                                                                                                         
BB:                                                                                                                                                                                                        
    %t1 = icmp                                                                                                                                                                                             
    %t4 = add %t2, c                                                                                                                                                                                       
    %t3 = select i1 %t1, %t2, %t3

llvm-svn: 52073
 2008-06-07 08:52:29 +00:00
Evan Cheng
bc28ef2028 Fix run line. 
llvm-svn: 52072
 2008-06-07 08:40:16 +00:00
Zhou Sheng
d7b035ee2b Add a test case for opt -instcombine bug fix in revision 52003. 
llvm-svn: 52004
 2008-06-05 14:25:11 +00:00
Matthijs Kooijman
6e1c286f53 Learn ScalarReplAggregrates how stores and loads of first class aggregrates 
work and how to replace them into individual values. Also, when trying to
replace an aggregrate that is used by load or store with a single (large)
integer, don't crash (but don't replace the aggregrate either).

Also adds a testcase for both structs and arrays.

llvm-svn: 51997
 2008-06-05 12:51:53 +00:00
Matthijs Kooijman
775c91b2f5 Let StructRetPromotion check if all if its users are really calls or invokesn, 
not other instructions. This fixes a crash with the added testcase.

llvm-svn: 51992
 2008-06-05 08:57:20 +00:00
Matthijs Kooijman
df97b7b4a2 Let StructRetPromotion check if it's users are really calling it and not 
passing its pointer. Fixes test with added testcase.

llvm-svn: 51991
 2008-06-05 08:48:32 +00:00
Owen Anderson
3f738eb65b Testcase for LoopIndexSplit and DomFrontier. 
llvm-svn: 51916
 2008-06-03 18:32:27 +00:00
Devang Patel
b1798d2be0 Update dom tree. Fix PR 2372. 
llvm-svn: 51887
 2008-06-02 22:52:56 +00:00
Owen Anderson
7700de3137 Fix two issues that Eli Friedman pointed out, where would misoptimized code like: 
char a[200];
init(a, a+200);

OR

int a[200];
char* b = (char*)a;
char* c = (char*)a;
foo(b, c);

llvm-svn: 51850
 2008-06-01 22:26:26 +00:00
Owen Anderson
d194f76cb4 Test for PR2401 
llvm-svn: 51849
 2008-06-01 21:55:55 +00:00
Duncan Sands
d14212a3e1 When simplifying a call to a bitcast function, tighten up 
the conditions for performing the transform when only the
function declaration is available: no longer allow turning
i32 into i64 for example.  Only allow changing between
pointer types, and between pointer types and integers of
the same size.  For return values ptr -> intptr was already
allowed; I added ptr -> ptr and intptr -> ptr while there.
As shown by a recent objc testcase, changing the way
parameters/return values are passed can be fatal when calling
code written in assembler that directly manipulates call
arguments and return values unless the transform has no
impact on the way they are passed at the codegen level.
While it is possible to imagine an ABI that treats integers
of pointer size differently to pointers, I don't think LLVM
supports any so the transform should now be safe while still
being useful.

llvm-svn: 51834
 2008-06-01 07:38:42 +00:00
Nick Lewycky
1bcd80adf7 Peer through sext/zext when looking for not(cmp). 
llvm-svn: 51819
 2008-05-31 19:01:33 +00:00
Nick Lewycky
b30afdb62b Add more i1 optimizations. add, sub, mul, s/udiv on i1 are now simplified away. 
llvm-svn: 51817
 2008-05-31 17:59:52 +00:00
Nick Lewycky
cdcdcddc85 Adding i1 is always Xor. 
llvm-svn: 51816
 2008-05-31 17:10:28 +00:00
Owen Anderson
bd3940abc7 Replace the old ADCE implementation with a new one that more simply solves 
the one case that ADCE catches that normal DCE doesn't: non-induction variable
loop computations.

This implementation handles this problem without using postdominators.

llvm-svn: 51668
 2008-05-29 08:45:13 +00:00
Chris Lattner
7a7da4f9c3 Implement PR2370: memmove(x,x,size) -> noop. 
llvm-svn: 51636
 2008-05-28 05:30:41 +00:00
Nick Lewycky
744dad8004 "ret (constexpr)" can't be folded into a Constant. Add a method to 
Analysis/ConstantFolding to fold ConstantExpr's, then make instcombine use it
to try to use targetdata to fold constant expressions on void instructions.

Also extend the icmp(inttoptr, inttoptr) folding to handle the case where
int size != ptr size.

llvm-svn: 51559
 2008-05-25 20:56:15 +00:00
Chris Lattner
3def8b4e53 Fix a serious brain-o. Obviously no-one reviewed my patch :( 
This fixes PR2359

llvm-svn: 51536
 2008-05-24 04:06:28 +00:00
Chris Lattner
bde5fd685d Fix PR2358 by resolving calls with undef arguments to overdefined. 
llvm-svn: 51535
 2008-05-24 03:59:33 +00:00
Dan Gohman
2412469191 Remove lingering references to .llx and .tr in the tests. 
llvm-svn: 51500
 2008-05-23 21:15:35 +00:00
Matthijs Kooijman
cf417144f6 Restucture a part of the SimplifyCFG pass and include a testcase. 
The SimplifyCFG pass looks at basic blocks that contain only phi nodes,
followed by an unconditional branch. In a lot of cases, such a block (BB) can
be merged into their successor (Succ).

This merging is performed by TryToSimplifyUncondBranchFromEmptyBlock. It does
this by taking all phi nodes in the succesor block Succ and expanding them to
include the predecessors of BB. Furthermore, any phi nodes in BB are moved to
Succ and expanded to include the predecessors of Succ as well.

Before attempting this merge, CanPropagatePredecessorsForPHIs checks to see if
all phi nodes can be properly merged. All functional changes are made to
this function, only comments were updated in
TryToSimplifyUncondBranchFromEmptyBlock.

In the original code, CanPropagatePredecessorsForPHIs looks quite convoluted
and more like stack of checks added to handle different kinds of situations
than a comprehensive check. In particular the first check in the function did
some value checking for the case that BB and Succ have a common predecessor,
while the last check in the function simply rejected all cases where BB and
Succ have a common predecessor. The first check was still useful in the case
that BB did not contain any phi nodes at all, though, so it was not completely
useless.

Now, CanPropagatePredecessorsForPHIs is restructured to to look a lot more
similar to the code that actually performs the merge. Both functions now look
at the same phi nodes in about the same order.  Any conflicts (phi nodes with
different values for the same source) that could arise from merging or moving
phi nodes are detected. If no conflicts are found, the merge can happen.

Apart from only restructuring the checks, two main changes in functionality
happened.

Firstly, the old code rejected blocks with common predecessors in most cases.
The new code performs some extra checks so common predecessors can be handled
in a lot of cases. Wherever common predecessors still pose problems, the
blocks are left untouched.

Secondly, the old code rejected the merge when values (phi nodes) from BB were
used in any other place than Succ. However, it does not seem that there is any
situation that would require this check. Even more, this can be proven.

Consider that BB is a block containing of a single phi node "%a" and a branch
to Succ. Now, since the definition of %a will dominate all of its uses, BB
will dominate all blocks that use %a. Furthermore, since the branch from BB to
Succ is unconditional, Succ will also dominate all uses of %a.

Now, assume that one predecessor of Succ is not dominated by BB (and thus not
dominated by Succ). Since at least one use of %a (but in reality all of them)
is reachable from Succ, you could end up at a use of %a without passing
through it's definition in BB (by coming from X through Succ). This is a
contradiction, meaning that our original assumption is wrong. Thus, all
predecessors of Succ must also be dominated by BB (and thus also by Succ).

This means that moving the phi node %a from BB to Succ does not pose any
problems when the two blocks are merged, and any use checks are not needed.

llvm-svn: 51478
 2008-05-23 09:09:41 +00:00
Nick Lewycky
6a16ace643 Constant integer vectors may also be negated. 
llvm-svn: 51476
 2008-05-23 04:54:45 +00:00
Nick Lewycky
bd2da8098d Revert X + X --> X * 2 optz'n which pessimizes heavily on x86. 
llvm-svn: 51474
 2008-05-23 04:34:58 +00:00
Nick Lewycky
427209006f Implement X + X for vectors. 
llvm-svn: 51472
 2008-05-23 04:14:51 +00:00
Nick Lewycky
e62259c369 Fix a recently added optimization to not crash on vectors. 
llvm-svn: 51471
 2008-05-23 03:26:47 +00:00
Dan Gohman
67e1a58e22 Generalize the new code in instcombine's ComputeNumSignBits for handling 
and/or to handle more cases (such as this add-sitofp.ll testcase), and
port it to selectiondag's ComputeNumSignBits.

llvm-svn: 51469
 2008-05-23 02:28:01 +00:00
Gabor Greif
b03785f0cd Eliminate questionable syntax for stdin redirection. This probably also speeds things up a bit. 
llvm-svn: 51357
 2008-05-20 22:07:21 +00:00
Chris Lattner
821dc30131 Fix PR2346 by marking vaarg as volatile so that licm doesn't try to 
hoist them.

llvm-svn: 51356
 2008-05-20 22:05:28 +00:00
Dan Gohman
7d78d53d2a Oops, commit the version of this test that actually works. 
llvm-svn: 51351
 2008-05-20 21:19:36 +00:00
Dan Gohman
b48d4a75f6 Port SelectionDAG's ComputeNumSignBits-using code to instcombine, 
now that instcombine also has ComputeNumSignBits.

llvm-svn: 51350
 2008-05-20 21:01:12 +00:00
Gabor Greif
807c2df887 sabre brings to my attention that the 'tr' suffix is also obsolete 
llvm-svn: 51349
 2008-05-20 21:00:03 +00:00
Gabor Greif
d8a4dbb5da Rename the last test with .llx extension to .ll, resolve duplicate test by renaming to isnan2. Now that no test has llx ending there is no need to search for them from dg.exp too. 
llvm-svn: 51328
 2008-05-20 19:52:04 +00:00
Chris Lattner
b387fd90fc Teach instcombine 4 new xforms: 
(add (sext x), cst) --> (sext (add x, cst'))
  (add (sext x), (sext y)) --> (sext (add int x, y))
  (add double (sitofp x), fpcst) --> (sitofp (add int x, intcst))
  (add double (sitofp x), (sitofp y)) --> (sitofp (add int x, y))

This generally reduces conversions.  For example MiBench/telecomm-gsm
gets these simplifications:

HACK2: 	%tmp67.i142.i.i = sext i16 %tmp6.i141.i.i to i32		; <i32> [#uses=1]
	%tmp23.i139.i.i = sext i16 %tmp2.i138.i.i to i32		; <i32> [#uses=1]
	%tmp8.i143.i.i = add i32 %tmp67.i142.i.i, %tmp23.i139.i.i		; <i32> [#uses=3]
HACK2: 	%tmp67.i121.i.i = sext i16 %tmp6.i120.i.i to i32		; <i32> [#uses=1]
	%tmp23.i118.i.i = sext i16 %tmp2.i117.i.i to i32		; <i32> [#uses=1]
	%tmp8.i122.i.i = add i32 %tmp67.i121.i.i, %tmp23.i118.i.i		; <i32> [#uses=3]
HACK2: 	%tmp67.i.i190.i = sext i16 %tmp6.i.i189.i to i32		; <i32> [#uses=1]
	%tmp23.i.i187.i = sext i16 %tmp2.i.i186.i to i32		; <i32> [#uses=1]
	%tmp8.i.i191.i = add i32 %tmp67.i.i190.i, %tmp23.i.i187.i		; <i32> [#uses=3]
HACK2: 	%tmp67.i173.i.i.i = sext i16 %tmp6.i172.i.i.i to i32		; <i32> [#uses=1]
	%tmp23.i170.i.i.i = sext i16 %tmp2.i169.i.i.i to i32		; <i32> [#uses=1]
	%tmp8.i174.i.i.i = add i32 %tmp67.i173.i.i.i, %tmp23.i170.i.i.i		; <i32> [#uses=3]
HACK2: 	%tmp67.i152.i.i.i = sext i16 %tmp6.i151.i.i.i to i32		; <i32> [#uses=1]
	%tmp23.i149.i.i.i = sext i16 %tmp2.i148.i.i.i to i32		; <i32> [#uses=1]
	%tmp8.i153.i.i.i = add i32 %tmp67.i152.i.i.i, %tmp23.i149.i.i.i		; <i32> [#uses=3]
HACK2: 	%tmp67.i.i.i.i = sext i16 %tmp6.i.i.i.i to i32		; <i32> [#uses=1]
	%tmp23.i.i5.i.i = sext i16 %tmp2.i.i.i.i to i32		; <i32> [#uses=1]
	%tmp8.i.i7.i.i = add i32 %tmp67.i.i.i.i, %tmp23.i.i5.i.i		; <i32> [#uses=3]


This also fixes a bug in ComputeNumSignBits handling select and
makes it more aggressive with and/or.

llvm-svn: 51302
 2008-05-20 05:46:13 +00:00
Devang Patel
9f385d71c2 Do not erase induction variable increment if it is used outside the loop. 
llvm-svn: 51280
 2008-05-19 22:23:55 +00:00
Chris Lattner
63c384df1e convert fptosi(sitofp x) -> x if the fp value has enough bits in its mantissa 
to accurately represent the integer.  This triggers 9 times in 471.omnetpp,
though 8 of those seem to be inlined from the same place.

llvm-svn: 51271
 2008-05-19 20:25:04 +00:00
Chris Lattner
1435b94f62 Fold FP comparisons where one operand is converted from an integer 
type and the other operand is a constant into integer comparisons.
This happens surprisingly frequently (e.g. 10 times in 471.omnetpp),
which are things like this:

	%tmp8283 = sitofp i32 %tmp82 to double	
	%tmp1013 = fcmp ult double %tmp8283, 0.0

Clearly comparing tmp82 against i32 0 is cheaper here.

this also triggers 8 times in gobmk, including this one:

	%tmp375376 = sitofp i32 %tmp375 to double
	%tmp377 = fcmp ogt double %tmp375376, 8.150000e+01

which is comparing an integer against 81.5 :).

llvm-svn: 51268
 2008-05-19 20:18:56 +00:00
Chris Lattner
510a6b249c be more aggressive about transforming add -> or when the operands have no 
intersecting bits.  This triggers all over the place, for example in lencode,
with adds of stuff like:

	%tmp580 = mul i32 %tmp579, 2	
	%tmp582 = and i32 %b8, 1
and

	%tmp28 = shl i32 %abs.i, 1		
	%sign.0 = select i1 %tmp23, i32 1, i32 0
and
	%tmp344 = shl i32 %tmp343, 2	
	%tmp346 = and i32 %tmp96, 3

etc.

llvm-svn: 51263
 2008-05-19 20:01:56 +00:00
Duncan Sands
7b84c36791 Fix PR2341 - when the length is 4 use an i32 not 
an i16!  Cleaned up trailing whitespace while there.

llvm-svn: 51240
 2008-05-19 09:27:24 +00:00
Chris Lattner
8c0f0a0e6c Fix PR2339 
llvm-svn: 51226
 2008-05-18 04:11:26 +00:00
Chris Lattner
8871489ae7 remove empty file? 
llvm-svn: 51225
 2008-05-18 04:10:18 +00:00
Nick Lewycky
46e3a168c0 Revert constant-folding change that will miscompile in some cases. 
llvm-svn: 51223
 2008-05-17 19:00:05 +00:00
Nick Lewycky
1df40102a9 Constant fold inttoptr and ptrtoint. 
llvm-svn: 51216
 2008-05-17 09:03:26 +00:00
Evan Cheng
76aaaf62e8 Fix test. 
llvm-svn: 51191
 2008-05-16 17:08:51 +00:00
Owen Anderson
55b78bc887 Move this test from ADCE to loop deletion, where it is more appropriate. 
llvm-svn: 51181
 2008-05-16 04:34:19 +00:00
Owen Anderson
79a25ff8ec Use loop deletion instead of ADCE in these tests. 
llvm-svn: 51180
 2008-05-16 04:33:37 +00:00
Owen Anderson
d282184d18 Use loop deletion instead of ADCE for removing loops. 
llvm-svn: 51178
 2008-05-16 04:27:38 +00:00
Chris Lattner
00e8e1e258 implement PR2328. 
llvm-svn: 51176
 2008-05-16 02:59:42 +00:00
Bill Wendling
c1d9f9604b Situations can arise when you have a function called that returns a 'void', but 
is bitcast to return a floating point value. The result of the instruction may
not be used by the program afterwards, and LLVM will happily remove all
instructions except the call. But, on some platforms, if a value is returned as
a floating point, it may need to be removed from the stack (like x87). Thus, we
can't get rid of the bitcast even if there isn't a use of the value.

llvm-svn: 51134
 2008-05-14 22:45:20 +00:00
Devang Patel
047ba6df54 Simplify internalize pass. Add test case. 
Patch by Matthijs Kooijman!

llvm-svn: 51114
 2008-05-14 20:01:01 +00:00
Dale Johannesen
676a1d026b Fix for PR 2323, infinite loop in tail dup. 
llvm-svn: 51063
 2008-05-13 20:06:43 +00:00
Owen Anderson
f67c06279b Add a testcase for non-local CSE of read-only calls. 
llvm-svn: 51025
 2008-05-13 08:17:44 +00:00
Duncan Sands
15622620d3 Testcase for PR2303. 
llvm-svn: 50951
 2008-05-10 16:43:10 +00:00
Chris Lattner
02ca137915 Implement PR2298. This transforms: 
~x < ~y --> y < x
   -x == -y --> x == y

llvm-svn: 50882
 2008-05-09 05:19:28 +00:00
Chris Lattner
4c1ef3628b More than just loads can read from memory: readonly calls like strlen 
also need to be checked for memory modifying instructions before we
can sink them.  THis fixes the second half of PR2297.

llvm-svn: 50860
 2008-05-08 17:37:37 +00:00
Chris Lattner
cba8b4c7e8 Make instcombine's DSE respect loads as well as stores. It is not safe to 
delete the first store in:

store x -> p
load p
store y -> p

This is for PR2297.

llvm-svn: 50859
 2008-05-08 17:20:30 +00:00
Dan Gohman
6ea87fa437 Fix a bug in the ComputeMaskedBits logic for multiply. 
llvm-svn: 50793
 2008-05-07 00:35:55 +00:00
Owen Anderson
2dccdcf2f2 Testcase for r50770. 
llvm-svn: 50771
 2008-05-06 21:01:34 +00:00
Dan Gohman
faf9df7227 Correct the value of LowBits in srem and urem handling in 
ComputeMaskedBits.

llvm-svn: 50692
 2008-05-06 00:51:48 +00:00
Chris Lattner
50d16c2939 Fix a crash when threading a block that includes a MRV call result. 
DemoteRegToStack doesn't work with MRVs yet, because it relies on the
ability to load/store things.

This fixes PR2285.

llvm-svn: 50667
 2008-05-05 20:21:22 +00:00
Dan Gohman
27156711ef Fix a mistake in the computation of leading zeros for udiv. 
llvm-svn: 50591
 2008-05-02 21:30:02 +00:00
Chris Lattner
96467cc665 strength reduce exp2 into ldexp, rdar://5852514 
llvm-svn: 50586
 2008-05-02 18:43:35 +00:00
Dan Gohman
04e2b94842 Update old-style syntax in some "not grep" tests. 
llvm-svn: 50560
 2008-05-01 23:50:07 +00:00
Dale Johannesen
4ab8b00dfa New test for bug fixed in 50545. 
llvm-svn: 50548
 2008-05-01 22:50:14 +00:00
Dan Gohman
793c9fed45 Fix an overaggressive SimplifyDemandedBits optimization on urem. This 
fixes the 254.gap regression on x86 and the 403.gcc regression on x86-64.

llvm-svn: 50537
 2008-05-01 19:13:24 +00:00
Chris Lattner
9d73228708 fix typo 
llvm-svn: 50519
 2008-05-01 06:16:48 +00:00
Chris Lattner
9a678d6f55 instcombine does memset optzns. 
llvm-svn: 50518
 2008-05-01 06:16:38 +00:00
Chris Lattner
926efd9174 simplifylibcalls doesn't optimize llvm.memmove, instcombine does. 
llvm-svn: 50517
 2008-05-01 06:14:24 +00:00
Chris Lattner
d4bf588b85 move some tests from libcall optimizer suite. 
llvm-svn: 50516
 2008-05-01 06:13:48 +00:00
Owen Anderson
f8c80ca156 Move this test to LoopDeletion, where it now passes. 
llvm-svn: 50474
 2008-04-30 07:17:22 +00:00
Chris Lattner
15195e00ee move lowering of llvm.memset -> store from simplify libcalls 
to instcombine.

llvm-svn: 50472
 2008-04-30 06:39:11 +00:00
Chris Lattner
ce01263bff no reason for simplifylibcalls to simplify intrinsics, instcombine does 
a fine job.

llvm-svn: 50470
 2008-04-30 06:12:15 +00:00
Chris Lattner
a62a4d407a remove redundant check. 
llvm-svn: 50469
 2008-04-30 06:06:37 +00:00
Owen Anderson
2caa79ae70 Fix a bug in memcpyopt where the memcpy-memcpy transform was never being applied because 
we were checking for it in the wrong order.  This caused a miscompilation because the
return slot optimization assumes that the call it is dealing with is NOT a memcpy.

llvm-svn: 50444
 2008-04-29 21:26:06 +00:00
Chris Lattner
5bd55b0885 don't eliminate load from volatile value on paths where the load is dead. 
This fixes the second half of PR2262

llvm-svn: 50430
 2008-04-29 17:28:22 +00:00
Chris Lattner
4b5d48a3f0 make this test reduced and *valid* 
llvm-svn: 50429
 2008-04-29 17:25:32 +00:00
Chris Lattner
7099f3c400 fix a subtle volatile handling bug. 
llvm-svn: 50428
 2008-04-29 17:13:43 +00:00
Chris Lattner
51fe8415da don't delete the last store to an alloca if the store is volatile. 
llvm-svn: 50390
 2008-04-29 04:58:38 +00:00
Dan Gohman
9e4db7f0bd Fix DSE to not eliminate volatile loads with no uses. 
llvm-svn: 50370
 2008-04-28 19:51:27 +00:00
Dan Gohman
1b7238e6e4 Teach InstCombine's ComputeMaskedBits what SelectionDAG's 
ComputeMaskedBits knows about cttz, ctlz, and ctpop. Teach
SelectionDAG's ComputeMaskedBits what InstCombine's knows
about SRem. And teach them both some things about high bits
in Mul, UDiv, URem, and Sub. This allows instcombine and
dagcombine to eliminate sign-extension operations in
several new cases.

llvm-svn: 50358
 2008-04-28 17:02:21 +00:00
Chris Lattner
ede7e89144 Fix PR2256, yet another miscompilation in simplifycfg of i 
multiple return values.

Bill, please pull this into Tak.

llvm-svn: 50332
 2008-04-28 00:19:07 +00:00
Chris Lattner
2798e42a9f When SRoA'ing a global variable, make sure the new globals get the 
appropriate alignment.  This fixes a miscompilation of 252.eon on
x86-64 (rdar://5891920).

Bill, please pull this into Tak.

llvm-svn: 50308
 2008-04-26 07:40:11 +00:00
Nick Lewycky
1f831c0f57 Remove 'unwinds to' support from mainline. This patch undoes r47802 r47989 
r48047 r48084 r48085 r48086 r48088 r48096 r48099 r48109 and r48123.

llvm-svn: 50265
 2008-04-25 16:53:59 +00:00
Chris Lattner
1a6268f776 Don't infininitely thread branches when a threaded edge 
goes back to the block, e.g.:

  Threading edge through bool from 'bb37.us.thread3829' to 'bb37.us' with cost: 1, across block:

bb37.us:		; preds = %bb37.us.thread3829, %bb37.us, %bb33
	%D1361.1.us = phi i32 [ %tmp36, %bb33 ], [ %D1361.1.us, %bb37.us ], [ 0, %bb37.us.thread3829 ]		; <i32> [#uses=2]
	%tmp39.us = icmp eq i32 %D1361.1.us, 0		; <i1> [#uses=1]
	br i1 %tmp39.us, label %bb37.us, label %bb42.us

llvm-svn: 50251
 2008-04-25 04:12:29 +00:00
Chris Lattner
be35a0c224 Split some code out of the main SimplifyCFG loop into its own function. 
Fix said code to handle merging return instructions together correctly
when handling multiple return values.

llvm-svn: 50199
 2008-04-24 00:01:19 +00:00
Chris Lattner
721ea7ca10 Rewrite multiple return value handling in SCCP. Before, the -sccp pass 
would turn every getresult instruction into undef.  This helps with
rdar://5778210

llvm-svn: 50140
 2008-04-23 05:38:20 +00:00
Chris Lattner
0dd624d232 remove this testcase. It isn't testing loop rotate, it is testing all 
of -std-compile-opts and is now failing because other passes are generating
IR that looks different to input of loop rotate.  Devang, please 
introduce a testcase that only runs loop rotate.

llvm-svn: 50136
 2008-04-23 05:36:04 +00:00
Chris Lattner
be858fc296 make this test more interesting. 
llvm-svn: 50128
 2008-04-23 03:49:32 +00:00
Chris Lattner
d059ac2e32 distill down the essense of this test. 
llvm-svn: 50125
 2008-04-23 03:03:42 +00:00
Dale Johannesen
547a55caf1 new test 
llvm-svn: 50123
 2008-04-23 01:22:22 +00:00
Evan Cheng
680839e258 Don't do: "(X & 4) >> 1 == 2 --> (X & 4) == 4" if there are more than one uses of the shift result. 
llvm-svn: 50118
 2008-04-23 00:38:06 +00:00
Chris Lattner
e304ae5621 Start doing the significantly useful part of jump threading: handle cases 
where a comparison has a phi input and that phi is a constant.  For example,
stuff like:

  Threading edge through bool from 'bb2149' to 'bb2231' with cost: 1, across block:
bb2237:		; preds = %bb2231, %bb2149
	%tmp2328.rle = phi i32 [ %tmp2232, %bb2231 ], [ %tmp2232439, %bb2149 ]		; <i32> [#uses=2]
	%done.0 = phi i32 [ %done.2, %bb2231 ], [ 0, %bb2149 ]		; <i32> [#uses=1]
	%tmp2239 = icmp eq i32 %done.0, 0		; <i1> [#uses=1]
	br i1 %tmp2239, label %bb2231, label %bb2327

or

bb38.i298:		; preds = %bb33.i295, %bb1693
	%tmp39.i296.rle = phi %struct.ibox* [ null, %bb1693 ], [ %tmp39.i296.rle1109, %bb33.i295 ]		; <%struct.ibox*> [#uses=2]
	%minspan.1.i291.reg2mem.1 = phi i32 [ 32000, %bb1693 ], [ %minspan.0.i288, %bb33.i295 ]		; <i32> [#uses=1]
	%tmp40.i297 = icmp eq %struct.ibox* %tmp39.i296.rle, null		; <i1> [#uses=1]
	br i1 %tmp40.i297, label %implfeeds.exit311, label %bb43.i301

This triggers thousands of times in spec.

llvm-svn: 50110
 2008-04-22 21:40:39 +00:00
Chris Lattner
c59cf9c8da Dig through multiple levels of AND to thread jumps if needed. 
llvm-svn: 50106
 2008-04-22 20:46:09 +00:00
Chris Lattner
dcbc6443ae Teach jump threading to thread through blocks like: 
br (and X, phi(Y, Z, false)), label L1, label L2

This triggers once on 252.eon and 6 times on 176.gcc.  Blocks 
in question often look like this:

bb262:		; preds = %bb261, %bb248
	%iftmp.251.0 = phi i1 [ true, %bb261 ], [ false, %bb248 ]		; <i1> [#uses=4]
	%tmp270 = icmp eq %struct.rtx_def* %tmp.0.i, null		; <i1> [#uses=1]
	%bothcond = or i1 %iftmp.251.0, %tmp270		; <i1> [#uses=1]
	br i1 %bothcond, label %bb288, label %bb273

In this case, it is clear that it doesn't matter if tmp.0.i is null when coming from bb261.  When coming from bb248, it is all that matters.


Another random example:

check_asm_operands.exit:		; preds = %check_asm_operands.exit.thr_comm, %bb30.i, %bb12.i, %bb6.i413
	%tmp.0.i420 = phi i1 [ true, %bb6.i413 ], [ true, %bb12.i ], [ true, %bb30.i ], [ false, %check_asm_operands.exit.thr_comm ; <i1> [#uses=1]
	call void @llvm.stackrestore( i8* %savedstack ) nounwind 
	%tmp4389 = icmp eq i32 %added_sets_1.0, 0		; <i1> [#uses=1]
	%tmp4394 = icmp eq i32 %added_sets_2.0, 0		; <i1> [#uses=1]
	%bothcond80 = and i1 %tmp4389, %tmp4394		; <i1> [#uses=1]
	%bothcond81 = and i1 %bothcond80, %tmp.0.i420		; <i1> [#uses=1]
	br i1 %bothcond81, label %bb4398, label %bb4397

Here is the case from 252.eon:

bb290.i.i:		; preds = %bb23.i57.i.i, %bb8.i39.i.i, %bb100.i.i, %bb100.i.i, %bb85.i.i110
	%myEOF.1.i.i = phi i1 [ true, %bb100.i.i ], [ true, %bb100.i.i ], [ true, %bb85.i.i110 ], [ true, %bb8.i39.i.i ], [ false, %bb23.i57.i.i ]		; <i1> [#uses=2]
	%i.4.i.i = phi i32 [ %i.1.i.i, %bb85.i.i110 ], [ %i.0.i.i, %bb100.i.i ], [ %i.0.i.i, %bb100.i.i ], [ %i.3.i.i, %bb8.i39.i.i ], [ %i.3.i.i, %bb23.i57.i.i ]		; <i32> [#uses=3]
	%tmp292.i.i = load i8* %tmp16.i.i100, align 1		; <i8> [#uses=1]
	%tmp293.not.i.i = icmp ne i8 %tmp292.i.i, 0		; <i1> [#uses=1]
	%bothcond.i.i = and i1 %tmp293.not.i.i, %myEOF.1.i.i		; <i1> [#uses=1]
	br i1 %bothcond.i.i, label %bb202.i.i, label %bb301.i.i
  Factoring out 3 common predecessors.

On the path from any blocks other than bb23.i57.i.i, the load and compare 
are dead.

llvm-svn: 50096
 2008-04-22 07:05:46 +00:00
Chris Lattner
4638234905 add a basic testcase. 
llvm-svn: 50093
 2008-04-22 06:35:14 +00:00
Chris Lattner
14be19cf1e optimize "p != gep p, ..." better. This allows us to compile 
getelementptr-seteq.ll into:

define i1 @test(i64 %X, %S* %P) {
	%C = icmp eq i64 %X, -1		; <i1> [#uses=1]
	ret i1 %C
}

instead of:

define i1 @test(i64 %X, %S* %P) {
	%A.idx.mask = and i64 %X, 4611686018427387903		; <i64> [#uses=1]
	%C = icmp eq i64 %A.idx.mask, 4611686018427387903		; <i1> [#uses=1]
	ret i1 %C
}

And fixes the second half of PR2235.  This speeds up the insertion sort
case by 45%, from 1.12s to 0.77s.  In practice, this will significantly
speed up for loops structured like:

for (double *P = Base + N; P != Base; --P)
  ...

Which happens frequently for C++ iterators.

llvm-svn: 50079
 2008-04-22 02:53:33 +00:00
Owen Anderson
bc6046416f Refactor memcpyopt based on Chris' suggestions. Consolidate several functions 
and simplify code that was fallout from the separation of memcpyopt and gvn.

llvm-svn: 50034
 2008-04-21 07:45:10 +00:00
Chris Lattner
a9d8d647ca rename *.llx -> *.ll, last batch. 
llvm-svn: 49971
 2008-04-19 22:32:52 +00:00
Owen Anderson
64fc7a4268 XFAIL this test for the moment. The real solution is to prevent ADCE 
from transforming loops and adding a separate loop pass for removing
loops with know trip counts.  Until that happens, ADCE is miscompiling this code.

llvm-svn: 49769
 2008-04-16 04:25:42 +00:00
Owen Anderson
15e930588a Add testcase for PR2213. 
llvm-svn: 49517
 2008-04-11 05:13:32 +00:00
Dan Gohman
318d9a6605 Teach InstCombine's ComputeMaskedBits to handle pointer expressions 
in addition to integer expressions. Rewrite GetOrEnforceKnownAlignment
as a ComputeMaskedBits problem, moving all of its special alignment
knowledge to ComputeMaskedBits as low-zero-bits knowledge.

Also, teach ComputeMaskedBits a few basic things about Mul and PHI
instructions.

This improves ComputeMaskedBits-based simplifications in a few cases,
but more noticeably it significantly improves instcombine's alignment
detection for loads, stores, and memory intrinsics.

llvm-svn: 49492
 2008-04-10 18:43:06 +00:00
Chris Lattner
be01a5f699 Generalize getUnaryFloatFunction to handle any FP unary function, automatically 
figuring out the suffix to use.  implement pow(2,x) -> exp2(x).

llvm-svn: 49437
 2008-04-09 17:48:11 +00:00
Chris Lattner
5d0cbe7d22 remove capital letter from test name. 
llvm-svn: 49436
 2008-04-09 17:46:36 +00:00
Owen Anderson
ca7e0e21f3 Factor a bunch of functionality related to memcpy and memset transforms out of 
GVN and into its own pass.

llvm-svn: 49419
 2008-04-09 08:23:16 +00:00
Chris Lattner
976ea8990e many cleanups to the pow optimizer. Allow it to handle powf, 
add support for  pow(x, 2.0) -> x*x.

llvm-svn: 49411
 2008-04-09 00:07:45 +00:00
Gabor Greif
80acb912a9 merge r48768 from branches/ggreif/parallelized-test 
llvm-svn: 49382
 2008-04-08 15:22:41 +00:00
Chris Lattner
12cecbbb25 add a testcase for forming memset from noncontiguous stores. 
llvm-svn: 48938
 2008-03-29 04:51:35 +00:00
Evan Cheng
563b265f37 Handle a special case xor undef, undef -> 0. Technically this should be transformed to undef. But this is such a common idiom (misuse) we are going to handle it. 
llvm-svn: 48791
 2008-03-25 20:07:13 +00:00
Tanya Lattner
b6a27ed83f Byebye llvm-upgrade! 
llvm-svn: 48762
 2008-03-25 04:26:08 +00:00
Devang Patel
425514c509 Add incoming value from header only if phi node has any use inside the loop. 
llvm-svn: 48738
 2008-03-24 20:16:14 +00:00
Chris Lattner
97e4d98c2d apparently tclsh doesn't lex like bash. Weird. 
llvm-svn: 48732
 2008-03-24 17:41:57 +00:00
Chris Lattner
3a6d3372f5 pass the option so this test tests the right thing. 
llvm-svn: 48731
 2008-03-24 17:36:38 +00:00
Evan Cheng
1d63708523 Transform (zext (or (icmp), (icmp))) to (or (zext (cimp), (zext icmp))) if at least one of the (zext icmp) can be transformed to eliminate an icmp. 
llvm-svn: 48715
 2008-03-24 00:21:34 +00:00
Owen Anderson
2f91173e40 Use normal naming convention for test. 
llvm-svn: 48693
 2008-03-22 21:08:33 +00:00
Chris Lattner
16f62d36e8 implement an initial hack at a straight-line store -> memset optimization. 
This fires dozens of times across spec and multisource, but I don't know
if it actually speeds stuff up.  Hopefully the testers will show something
nice :)

llvm-svn: 48680
 2008-03-22 05:37:16 +00:00
Chris Lattner
96cdf21ed4 Teach masked value is zero about add and sub, and use MVIZ to 
simplify things like (X & 4) >> 1 == 2  --> (X & 4) == 4.

since it is obvious that the shift doesn't remove any bits.

llvm-svn: 48631
 2008-03-21 05:19:58 +00:00
Tanya Lattner
52e5896b3f Upgrade tests. 
llvm-svn: 48538
 2008-03-19 07:28:33 +00:00
Tanya Lattner
f0dc625b4f Upgrade tests. 
llvm-svn: 48536
 2008-03-19 05:39:35 +00:00
Tanya Lattner
be72114460 Upgrade tests to not use llvm-upgrade. 
llvm-svn: 48530
 2008-03-19 04:36:04 +00:00
Tanya Lattner
88c66b1027 Upgrade tests to not use llvm-upgrade. 
llvm-svn: 48529
 2008-03-19 04:14:49 +00:00
Tanya Lattner
0721aaa8f7 Remove llvm-upgrade and update tests. 
llvm-svn: 48527
 2008-03-19 03:47:13 +00:00
Tanya Lattner
9bd47b05dd Upgrade tests to not use llvm-upgrade. 
llvm-svn: 48484
 2008-03-18 04:14:37 +00:00
Tanya Lattner
60b7b29981 Upgrade tests to not use llvm-upgrade. 
llvm-svn: 48483
 2008-03-18 03:45:45 +00:00
Bill Wendling
7d054f8b3f The inst combining of inttoptr into GEP with one index was using the bit size of 
the type instead of the byte size. This was causing troublesome mis-compilations.

True to form, this took 2 days to find and is a one-line fix. :-P

llvm-svn: 48354
 2008-03-14 05:12:19 +00:00
Owen Anderson
6c2454d9d1 Fix a bug in GVN that Duncan noticed, where we potentially need to insert a 
pointer bitcast when performing return slot optimization.

llvm-svn: 48343
 2008-03-13 22:07:10 +00:00
Owen Anderson
5887233a3f Improve the return slot optimization to be both more aggressive (not limited to sret parameters), and 
safer (when the passed pointer might be invalid).  Thanks to Duncan and Chris for the idea behind this, 
and extra thanks to Duncan for helping me work out the trap-safety.

llvm-svn: 48280
 2008-03-12 07:37:44 +00:00
Devang Patel
9e4d8236fc Fix attribute handling. 
llvm-svn: 48262
 2008-03-12 00:07:03 +00:00
Devang Patel
54b2c77a4e Handle multiple ret values. 
llvm-svn: 48254
 2008-03-11 22:24:29 +00:00
Dan Gohman
34ea45f10d Check to see if a two-entry PHI block can be simplified 
before trying to merge the block into its predecessors.
This allows two-entry-phi-return.ll to be simplified
into a single basic block.

llvm-svn: 48252
 2008-03-11 21:53:06 +00:00
Dan Gohman
05907c9f06 Make this test more challenging to help it avoid being 
optimized away before it tests what it is intended to test.

llvm-svn: 48251
 2008-03-11 21:47:57 +00:00
Devang Patel
6c189c16b5 Initial multiple return values support. 
llvm-svn: 48210
 2008-03-11 05:46:42 +00:00
Dan Gohman
2674ad5bed Upgrade this test. 
llvm-svn: 48207
 2008-03-11 02:19:59 +00:00
Devang Patel
659c71126e Simplify 
llvm-svn: 48163
 2008-03-10 18:38:30 +00:00
Tanya Lattner
8ac346c316 Remove llvm-upgrade and update tests. 
llvm-svn: 48137
 2008-03-10 07:21:50 +00:00
Nick Lewycky
5dd879d5b5 Turn unwind_to into "unwinds to". 
llvm-svn: 48123
 2008-03-10 02:20:00 +00:00
Tanya Lattner
9aa573954c Remove llvm-upgrade and update tests. 
llvm-svn: 48103
 2008-03-09 08:16:40 +00:00
Nick Lewycky
4732d3c9b7 Firstly, having a BranchInst isn't exclusive with having an unwind_to. 
Secondly, we have to check whether the branch is actually pointing to the block
with the unwind in it. We could have gotten here because of the unwind_to alone.

llvm-svn: 48099
 2008-03-09 07:50:37 +00:00
Nick Lewycky
e93a1819dc A BB that unwind_to an "unwind" inst is that same as one that doesn't unwind_to 
at all.

llvm-svn: 48096
 2008-03-09 07:36:38 +00:00
Nick Lewycky
d98272094e Update the inliner and simplifycfg to handle unwind_to. 
llvm-svn: 48086
 2008-03-09 05:10:13 +00:00
Nick Lewycky
980738baf2 Prune the unwind_to labels on BBs that don't need them. Another step in the 
removal of invoke, PR1269.

llvm-svn: 48084
 2008-03-09 04:55:16 +00:00
Devang Patel
9543238662 Update inliner to handle functions that return multiple values. 
llvm-svn: 48020
 2008-03-07 20:06:16 +00:00
Devang Patel
a6ec13ef6b Place for sret promotion tests. 
llvm-svn: 48016
 2008-03-07 20:00:15 +00:00
Nick Lewycky
c6edcb8904 Commit the testcase too. 
llvm-svn: 47988
 2008-03-06 06:50:03 +00:00
Nick Lewycky
f249c5d5ad Don't try to simplify urem and srem using arithmetic rules that don't work 
under modulo (overflow). Fixes PR1933.

llvm-svn: 47987
 2008-03-06 06:48:30 +00:00
Devang Patel
84e9abbb60 Use cast instead of dyn_cast. 
Update test to use multiple return value directly, instead of relying on -sretpromotion.

llvm-svn: 47907
 2008-03-04 21:45:28 +00:00
Devang Patel
a14b91d6a3 Handle multiple return values. 
llvm-svn: 47904
 2008-03-04 21:15:15 +00:00
Tanya Lattner
c072619922 Remove llvm-upgrade and update test cases. 
llvm-svn: 47793
 2008-03-01 09:15:35 +00:00
Chris Lattner
75f5d14574 fix a bug Anders ran into where scalarrepl would crash when promoting 
a union containing a vector and an array whose elements were smaller than
the vector elements.  this means we need to compile the load of the 
array elements into an extract element plus a truncate.

llvm-svn: 47752
 2008-02-29 07:12:06 +00:00
Chris Lattner
5aeccb7353 Folding or(fcmp,fcmp) only works if the operands of the fcmps are the same fp type. 
llvm-svn: 47750
 2008-02-29 06:09:11 +00:00
Owen Anderson
a39b18223f Add PR number to testcase. 
llvm-svn: 47640
 2008-02-26 23:16:11 +00:00
Owen Anderson
eadd074b22 Fix an issue where GVN had the sizes of the two memcpy's reverse, resulting 
in an invalid transformation.

llvm-svn: 47639
 2008-02-26 23:06:17 +00:00
Chris Lattner
9d19e558b7 fix this test so that the fn name doesn't match the regex 
llvm-svn: 47608
 2008-02-26 18:13:51 +00:00
Gabor Greif
a8990a9cac Really feed llvm-as with the testcase, do not let it read from stdin. This fixes the hangs seen on solaris10. 
llvm-svn: 47604
 2008-02-26 13:37:13 +00:00
Owen Anderson
6eafd532ab Fix an issue where GVN was performing the return slot optimization when it was 
not safe.  This is fixed by more aggressively checking that the return slot is
not used elsewhere in the function.

llvm-svn: 47544
 2008-02-25 04:08:09 +00:00
Owen Anderson
432abc0479 Fix an issue where GVN would try to use an instruction before its definition when performing return slot optimization. 
llvm-svn: 47541
 2008-02-25 00:40:41 +00:00
Zhou Sheng
5f4e29db89 Testcase for Revision 47478. 
llvm-svn: 47531
 2008-02-23 10:59:51 +00:00
Nick Lewycky
1f3c58df08 Correctly fold divide-by-constant, even when faced with overflow. 
llvm-svn: 47287
 2008-02-18 22:48:05 +00:00
Chris Lattner
a6dc8aaa3f make this just a bit more strict. 
llvm-svn: 47274
 2008-02-18 17:33:10 +00:00
Owen Anderson
7b092ea631 Add support to GVN for performing sret return slot optimization. This means that, if an sret function tail calls 
another sret function, it should pass its own sret parameter to the tail callee, allowing it to fill in the correct
return value.  llvm-gcc does not emit this by default.  Instead, it allocates space in the caller for the sret of
the tail call and then uses memcpy to copy the result into the caller's sret parameter.  This optimization detects
and optimizes that case.

llvm-svn: 47265
 2008-02-18 09:24:53 +00:00
Chris Lattner
9851db050b optimize away stackrestore calls that have no intervening alloca or call. 
llvm-svn: 47258
 2008-02-18 06:12:38 +00:00
Chris Lattner
dd489bf3eb upgrade this test. 
llvm-svn: 47257
 2008-02-18 06:11:00 +00:00
Chris Lattner
2fa904b3af Fold (-x + -y) -> -(x+y) which promotes better association, fixing 
the second half of PR2047

llvm-svn: 47244
 2008-02-17 21:03:36 +00:00
Chris Lattner
4a34461d64 Split up subtracts into add+negate if they have a reassociable use or operand 
that is also a subtract.  This implements PR2047 and Transforms/Reassociate/subtest2.ll 

llvm-svn: 47241
 2008-02-17 20:51:26 +00:00
Chris Lattner
7cd16f34b3 upgrade and simplify this test. 
llvm-svn: 47240
 2008-02-17 20:48:43 +00:00
Duncan Sands
cd0325cce1 Remove any 'nest' parameter attributes if the function 
is not passed as an argument to a trampoline intrinsic.

llvm-svn: 47220
 2008-02-16 20:56:04 +00:00
Devang Patel
8af92b942c If loop header is also loop exiting block then OrigPN is incoming value for B loop header. 
Fixes PR 2030.

llvm-svn: 47141
 2008-02-14 23:18:47 +00:00
Chris Lattner
b8bf200b77 Fix PR2029 
llvm-svn: 47129
 2008-02-14 19:18:13 +00:00
Nick Lewycky
12098ea401 Testcase for PR2032. 
llvm-svn: 47113
 2008-02-14 07:15:11 +00:00
Devang Patel
38de2d11a5 A loop latch phi node may have uses inside loop, not just in loop header. 
llvm-svn: 47093
 2008-02-13 22:23:07 +00:00
Devang Patel
1e71afe2df While moving exit condition, do not drop loop latch on the floor. 
llvm-svn: 47089
 2008-02-13 22:06:36 +00:00
Devang Patel
d48bbbf07b Keep track of exit value operand number when operands are swapped. 
llvm-svn: 47082
 2008-02-13 19:48:48 +00:00
Eli Friedman
cf6273d015 Add a note pointing to PR1996. 
llvm-svn: 47055
 2008-02-13 07:56:04 +00:00
Eli Friedman
75220639b6 Add test for PR1996. (This is my first time adding a test for a 
transform, so please review.)

llvm-svn: 47050
 2008-02-13 06:55:57 +00:00
Owen Anderson
274aa2846e Re-apply the patch to improve the optimizations of memcpy's, with several 
bugs fixed.  This now passes PPC bootstrap.

llvm-svn: 47026
 2008-02-12 21:15:18 +00:00
Devang Patel
9aba178666 Fix PR 1995. 
llvm-svn: 46898
 2008-02-08 22:49:13 +00:00
Bill Wendling
8a28ab4b1f Temporarily reverting: 
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080128/057882.html

This is causing a miscompilation on PPC G5 and just now seeing it on iMac x86-64.

llvm-svn: 46822
 2008-02-06 20:03:07 +00:00
Chris Lattner
96deed5d4d Fix a bug compiling PR1978 (perhaps not the only one though) which 
was incorrectly simplifying "x == (gep x, 1, i)" into false, even 
though i could be negative.  As it turns out, all the code to 
handle this already existed, we just need to disable the incorrect
optimization case and let the general case handle it.

llvm-svn: 46739
 2008-02-05 04:45:32 +00:00
Owen Anderson
e2150dfe24 Make this test more aggressive, to cover recent improvements. 
llvm-svn: 46695
 2008-02-04 04:55:24 +00:00
Owen Anderson
aaba6f96da Allow GVN to hack on memcpy's, making them open to further optimization. 
llvm-svn: 46693
 2008-02-04 02:59:58 +00:00
Nick Lewycky
81cc718a25 Tag this test with the PR reference. 
llvm-svn: 46688
 2008-02-03 16:35:19 +00:00
Nick Lewycky
febd3642ce There are some cases where icmp(add) can be folded into a new icmp. Handle them. 
llvm-svn: 46687
 2008-02-03 16:33:09 +00:00
Duncan Sands
454a8eaee9 Don't drop function/call return attributes like 'nounwind'. 
llvm-svn: 46645
 2008-02-01 20:37:16 +00:00
Owen Anderson
ad9a347656 Make DSE much more aggressive by performing DCE earlier. Update a testcase to reflect this increased aggressiveness. 
llvm-svn: 46542
 2008-01-30 01:24:47 +00:00
Chris Lattner
83227e350d Fix a bug where scalarrepl would discard offset if type would match. 
In practice this can only happen on code with already undefined behavior, 
but this is still a good thing to handle correctly.

llvm-svn: 46539
 2008-01-30 00:39:15 +00:00
Chris Lattner
e08ec140e0 Don't let globalopt hack on volatile loads or stores. 
llvm-svn: 46523
 2008-01-29 19:01:37 +00:00
Chris Lattner
83f411c586 eliminate additions of 0.0 when they are obviously dead. This has to be careful to 
avoid turning -0.0 + 0.0 -> -0.0 which is incorrect.

llvm-svn: 46499
 2008-01-29 06:52:45 +00:00
Owen Anderson
2688087c9c Add a testcase for eliminating memcpy's at the end of functions. Forgot to commit this with my last commit. 
llvm-svn: 46497
 2008-01-29 06:40:32 +00:00
Devang Patel
86ff705c22 Filter loops that subtract induction variables. 
These loops are not yet handled.

Fix PR 1912.

llvm-svn: 46484
 2008-01-29 02:20:41 +00:00
Chris Lattner
20854cf4e7 this test is now compiled into the right thing. 
llvm-svn: 46454
 2008-01-28 17:38:46 +00:00
Nick Lewycky
6b070b1b93 Handle some more combinations of extend and icmp. Fixes PR1940. 
llvm-svn: 46431
 2008-01-28 03:48:02 +00:00
Chris Lattner
359756ea4b Fix PR1932 by disabling an xform invalid for fdiv. 
llvm-svn: 46429
 2008-01-28 00:58:18 +00:00
Chris Lattner
7250586ec9 Fix PR1938 by forcing the code that uses an undefined value to branch one 
way or the other.  Rewriting the code itself prevents subsequent analysis
passes from making contradictory conclusions about the code that could 
cause an infeasible path to be made feasible.

llvm-svn: 46427
 2008-01-28 00:32:30 +00:00
Nick Lewycky
cd28ef8950 Be more careful modifying the use_list while also iterating through it. 
llvm-svn: 46417
 2008-01-27 18:35:00 +00:00
Duncan Sands
e77256b325 Revert r46393: readonly/readnone functions are no 
longer allowed to write through byval arguments.

llvm-svn: 46416
 2008-01-27 18:12:58 +00:00
Bill Wendling
0e2b8c2c45 The CorrelatedExpressions pass is now no more. 
llvm-svn: 46409
 2008-01-27 06:13:32 +00:00
Chris Lattner
aa553aa0c1 Fold fptrunc(add (fpextend x), (fpextend y)) -> add(x,y), as GCC does. 
llvm-svn: 46406
 2008-01-27 05:29:54 +00:00
Duncan Sands
9fae964ef7 Invert this test, because it is wrong if we allow 
readonly functions to use byval parameters as local
storage (how much do we want this?).

llvm-svn: 46399
 2008-01-26 12:33:01 +00:00
Owen Anderson
a4ff15c69f DeadStoreElimination can treat byval parameters as if there were alloca's for the purpose of removing end-of-function stores. 
llvm-svn: 46351
 2008-01-25 10:10:33 +00:00
Nick Lewycky
78780f175b Multiply can be evaluated in a different type, so long as the target type has 
a smaller bitwidth.

llvm-svn: 46244
 2008-01-22 05:08:48 +00:00
Evan Cheng
ca434b34ff Test case for varargs parameter attribute issue I just fixed. 
llvm-svn: 46127
 2008-01-17 07:26:31 +00:00
Chris Lattner
c32be7e6b2 Fix arg promotion to propagate the correct attrs on the calls to 
promoted functions.  This is important for varargs calls in 
particular.  Thanks to duncan for providing a great testcase.

llvm-svn: 46108
 2008-01-17 01:17:03 +00:00
Devang Patel
a8d63b0b30 Do not strip llvm.used values. 
llvm-svn: 46045
 2008-01-16 03:33:05 +00:00
Chris Lattner
f0dd2bb4e6 add a test to ensure that argpromote of one argument doesn't 
break the byval attr on some other argument.

llvm-svn: 46025
 2008-01-15 22:38:12 +00:00
Duncan Sands
81e35b4d47 I noticed that the trampoline straightening transformation could 
drop attributes on varargs call arguments.  Also, it could generate
invalid IR if the transformed call already had the 'nest' attribute
somewhere (this can never happen for code coming from llvm-gcc,
but it's a theoretical possibility).  Fix both problems.

llvm-svn: 45973
 2008-01-14 19:52:09 +00:00
Chris Lattner
efb498eddb Fix the miscompilation of MiBench/consumer-lame that was exposed by Evan's 
byval work.  This miscompilation is due to the program indexing an array out
of range and us doing a transformation that broke this.

llvm-svn: 45949
 2008-01-14 02:09:12 +00:00
Chris Lattner
d22a5f6314 Turn a memcpy from a double* into a load/store of double instead of 
a load/store of i64.  The later prevents promotion/scalarrepl of the
source and dest in many cases.

This fixes the 300% performance regression of the byval stuff on 
stepanov_v1p2.

llvm-svn: 45945
 2008-01-14 00:28:35 +00:00
Chris Lattner
32eae5daa5 Fix PR1907, a nasty miscompilation because instcombine didn't 
realize that ne & sgt  was a signed comparison (it was only 
looking at whether the left compare was signed).

llvm-svn: 45937
 2008-01-13 20:59:02 +00:00
Duncan Sands
7414cc131b When turning a call to a bitcast function into a direct call, 
if this becomes a varargs call then deal correctly with any
parameter attributes on the newly vararg call arguments.

llvm-svn: 45931
 2008-01-13 08:02:44 +00:00
Chris Lattner
d90840eddc we don't have to make an explicit copy of a byval argument when 
inlining a function if we know that the function does not write
to *any* memory.  This implements test/Transforms/Inline/byval2.ll

llvm-svn: 45912
 2008-01-12 18:54:29 +00:00
Duncan Sands
6f49217a5e When DAE drops the varargs part of a function, ensure any 
attributes on the vararg call arguments are also dropped.

llvm-svn: 45892
 2008-01-11 23:13:45 +00:00
Chris Lattner
a1246ba8ed Teach argpromote to ruthlessly hack small byval structs when it can 
get away with it, which exposes opportunities to eliminate the memory
objects entirely.  For example, we now compile byval.ll to:

define internal void @f1(i32 %b.0, i64 %b.1) {
entry:
	%tmp2 = add i32 %b.0, 1		; <i32> [#uses=0]
	ret void
}

define i32 @main() nounwind  {
entry:
	call void @f1( i32 1, i64 2 )
	ret i32 0
}

This seems like it would trigger a lot for code that passes around small
structs (e.g. SDOperand's or _Complex)...

llvm-svn: 45886
 2008-01-11 22:31:41 +00:00
Chris Lattner
bf51fecdc4 When inlining a functino with a byval argument, make an explicit 
copy of it in case the callee modifies the struct.

llvm-svn: 45853
 2008-01-11 06:09:30 +00:00
Chris Lattner
67f581b344 Implement PR1795, an instcombine hack for forming GEPs with integer pointer arithmetic. 
llvm-svn: 45745
 2008-01-08 07:23:51 +00:00
Duncan Sands
fd975e4b3d The transform that tries to turn calls to bitcast functions into 
direct calls bails out unless caller and callee have essentially
equivalent parameter attributes.  This is illogical - the callee's
attributes should be of no relevance here.  Rework the logic, which
incidentally fixes a crash when removed arguments have attributes.

llvm-svn: 45658
 2008-01-06 18:27:01 +00:00
Duncan Sands
b8489f09a2 When transforming a call to a bitcast function into 
a direct call with cast parameters and cast return
value (if any), instcombine was prepared to cast any
non-void return value into any other, whether castable
or not.  Add a new predicate for testing whether casting
is valid, and check it both for the return value and
(as a cleanup) for the parameters.

llvm-svn: 45657
 2008-01-06 10:12:28 +00:00
Chris Lattner
7e1c3aa702 remove a couple more unsafe xforms in the face of overflow. 
llvm-svn: 45613
 2008-01-05 01:22:42 +00:00
Chris Lattner
d4c66656a1 Fix PR1896 
llvm-svn: 45568
 2008-01-04 05:04:53 +00:00
Chris Lattner
26b89fd30a don't hoist FP additions into unconditional adds + selects. This 
could theoretically introduce a trap, but is also a performance issue.
This speeds up ptrdist/ks by 8%.

llvm-svn: 45533
 2008-01-03 07:25:26 +00:00
Bill Wendling
e6898cbe7a Update this testcase. The output needs to be disabled to pass. 
llvm-svn: 45478
 2008-01-01 01:34:36 +00:00
Chris Lattner
2369d2f4ab dead calls to llvm.stacksave can be deleted, even though they 
have potential side-effects.

llvm-svn: 45392
 2007-12-29 00:59:12 +00:00
Chris Lattner
e26989231e upgrade this test 
llvm-svn: 45391
 2007-12-29 00:57:06 +00:00
Devang Patel
d601ada369 Test -simplifycfg only. 
llvm-svn: 45389
 2007-12-28 22:59:48 +00:00
Owen Anderson
ddb5d73f0a Add a testcase for my recent InstCombine fix, written by Nicholas. 
llvm-svn: 45386
 2007-12-28 21:08:43 +00:00
Chris Lattner
d64df490ca implement InstCombine/shift-trunc-shift.ll. This allows 
us to compile:
#include <math.h>
int t1(double d) { return signbit(d); }

into:

_t1:
	movd	%xmm0, %rax
	shrq	$63, %rax
	ret

instead of:

_t1:
	movd	%xmm0, %rax
	shrq	$32, %rax
	shrl	$31, %eax
	ret

on x86-64.

llvm-svn: 45311
 2007-12-22 09:07:47 +00:00
Devang Patel
e035f776e9 If succ has succ itself as one of the predecessors then do 
not merge current bb and succ even if bb's terminator is
unconditional branch to succ.

llvm-svn: 45305
 2007-12-22 01:32:53 +00:00
Duncan Sands
85ca85c070 Make DAE not wipe out attributes on calls, and not drop 
return attributes on the floor.  In the case of a call
to a varargs function where the varargs arguments are
being removed, any call attributes on those arguments
need to be dropped.  I didn't do this because I plan to
make it illegal to have such attributes (see next patch).
With this change, compiling the gcc filter2 eh test at -O0
and then running opt -std-compile-opts on it results in
a correctly working program (compiling at -O1 or higher
results in the test failing due to a problem with how we
output eh info into the IR).

llvm-svn: 45285
 2007-12-21 19:16:16 +00:00
Christopher Lamb
7ca648a7b1 Implement review feedback, including additional transforms 
(icmp slt (sub A B) 1) -> (icmp sle A B)
icmp sgt (sub A B) -1) -> (icmp sge A B)

and add testcase.

llvm-svn: 45256
 2007-12-20 07:21:11 +00:00
Duncan Sands
56f3add5b7 When inlining through an 'nounwind' call, mark inlined 
calls 'nounwind'.  It is important for correct C++
exception handling that nounwind markings do not get
lost, so this transformation is actually needed for
correctness.

llvm-svn: 45218
 2007-12-19 21:13:37 +00:00
Christopher Lamb
d56318b885 Remove an orthogonal transformation of the selection condition from my most recent submission. 
llvm-svn: 45169
 2007-12-18 20:30:28 +00:00
Christopher Lamb
437b4d229e Fix typos. 
llvm-svn: 45159
 2007-12-18 09:45:40 +00:00
Christopher Lamb
aeb76743dc Fold certain additions through selects (and their compares) so as to eliminate subtractions. This code is often produced by the SMAX expansion in SCEV. 
This implements test/Transforms/InstCombine/2007-12-18-AddSelCmpSub.ll

llvm-svn: 45158
 2007-12-18 09:34:41 +00:00
Duncan Sands
3a0d757bd5 Make invokes of inline asm legal. Teach codegen 
how to lower them (with no attempt made to be
efficient, since they should only occur for
unoptimized code).

llvm-svn: 45108
 2007-12-17 18:08:19 +00:00
Duncan Sands
bf62f62058 Make instcombine promote inline asm calls to 'nounwind' 
calls.  Remove special casing of inline asm from the
inliner.  There is a potential problem: the verifier
rejects invokes of inline asm (not sure why).  If an
asm call is not marked "nounwind" in some .ll, and
instcombine is not run, but the inliner is run, then
an illegal module will be created.  This is bad but
I'm not sure what the best approach is.  I'm tempted
to remove the check in the verifier...

llvm-svn: 45073
 2007-12-16 15:51:49 +00:00
Wojciech Matyjewicz
8bb1d9e67c 1. "Upgrage" comments. 
2. Using zero-extended value of Scale and unsigned division is safe provided
   that Scale doesn't have the sign bit set.
   Previously these 2 instructions:
        %p = bitcast [100 x {i8,i8,i8}]* %x to i8*
        %q = getelementptr i8* %p, i32 -4
   were combined into:
        %q = getelementptr [100 x { i8, i8, i8 }]* %x, i32 0,
               i32 1431655764, i32 0
   what was incorrect.

llvm-svn: 44936
 2007-12-12 15:21:32 +00:00