Commit Graph

2228 Commits

Author SHA1 Message Date
Chris Lattner
727552bb89 If not internalizing, don't mark llvm.global[cd]tors const, as a fix for a
hypothetical future boog.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25430 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-19 00:46:54 +00:00
Chris Lattner
ee9e14cb8a Don't internalize llvm.global[cd]tor unless there are uses of it. This
unbreaks front-ends that don't use __main (like the new CFE).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25429 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-19 00:40:39 +00:00
Chris Lattner
c4e8c9f318 Make sure that cloning a module clones its target triple and dependent
library list as well.  This should help bugpoint.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25424 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-18 21:32:45 +00:00
Robert Bocchino
956fd7254f Constant folding support for the insertelement operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25407 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-17 20:07:07 +00:00
Robert Bocchino
8fcf01ead0 Lowerpacked and SCCP support for the insertelement operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25406 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-17 20:06:55 +00:00
Chris Lattner
f91e97ae81 Clean up the FFS optimization code, and make it correctly create the appropriate
unsigned llvm.cttz.* intrinsic, fixing the 2005-05-11-Popcount-ffs-fls regression
last night.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25398 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-17 18:27:17 +00:00
Reid Spencer
0b118206bf For PR411:
This patch is an incremental step towards supporting a flat symbol table.
It de-overloads the intrinsic functions by providing type-specific intrinsics
and arranging for automatically upgrading from the old overloaded name to
the new non-overloaded name. Specifically:
  llvm.isunordered -> llvm.isunordered.f32, llvm.isunordered.f64
  llvm.sqrt -> llvm.sqrt.f32, llvm.sqrt.f64
  llvm.ctpop -> llvm.ctpop.i8, llvm.ctpop.i16, llvm.ctpop.i32, llvm.ctpop.i64
  llvm.ctlz -> llvm.ctlz.i8, llvm.ctlz.i16, llvm.ctlz.i32, llvm.ctlz.i64
  llvm.cttz -> llvm.cttz.i8, llvm.cttz.i16, llvm.cttz.i32, llvm.cttz.i64
New code should not use the overloaded intrinsic names. Warnings will be
emitted if they are used.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25366 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-16 21:12:35 +00:00
Chris Lattner
72223eea95 fix a crash due to missing parens
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25363 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-16 19:47:21 +00:00
Chris Lattner
57a47d6a4b This pass has never worked correctly. Remove.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25349 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-16 01:06:00 +00:00
Chris Lattner
432a205769 Let the inliner update the callgraph to reflect the changes it makes, instead
of doing it ourselves.  This fixes Transforms/Inline/2006-01-14-CallGraphUpdate.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25321 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-14 20:09:18 +00:00
Chris Lattner
468fb1df7d Teach the inliner to update the CallGraph itself, and have it add edges to
llvm.stacksave/restore when it inserts calls to them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25320 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-14 20:07:50 +00:00
Chris Lattner
516a74c01a FunctionPass's cannot do IPO things.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25315 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-14 19:30:35 +00:00
Nate Begeman
6fb3bd6a65 Add bswap intrinsics as documented in the Language Reference
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25309 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-14 01:25:24 +00:00
Robert Bocchino
1d7456d0bf Added instcombine support for extractelement.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25299 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 22:48:06 +00:00
Chris Lattner
5522037136 it is ok to dce stacksave.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25295 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 21:31:54 +00:00
Chris Lattner
a728ddc815 Do a simple instcombine xforms to delete llvm.stackrestore cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25294 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 21:28:09 +00:00
Chris Lattner
8b0ea313d9 Simplify this a tiny bit by using the new IntrinsicInst functionality.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25292 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 20:11:04 +00:00
Chris Lattner
ccca3ca85f Permit inlining functions that contain dynamic allocations now that
InlineFunction handles this case safely.  This implements
Transforms/Inline/dynamic_alloca_test.ll.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25288 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 19:35:43 +00:00
Chris Lattner
bf229f488a If inlining a call to a function that contains dynamic allocas, wrap the
resultant code with llvm.stacksave/llvm.stackrestore intrinsics.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25286 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 19:34:14 +00:00
Chris Lattner
1fdf4a859f Use ClonedCodeInfo to avoid another walk over the inlined code, this this
time in common C cases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25285 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 19:18:11 +00:00
Chris Lattner
727d1dd587 Use the ClonedCodeInfo object to avoid scans of the inlined code when
it doesn't contain any calls.  This is a fairly common case for C++ code,
so it will probably speed up the inliner marginally in these cases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25284 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 19:15:15 +00:00
Chris Lattner
cd4d339ec1 Refactor a bunch of invoke handling stuff out into a new function
"HandleInlinedInvoke".  No functionality change.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25283 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 19:05:59 +00:00
Chris Lattner
a4c29d2037 Allow the code cloning interfaces to capture some important info about the
code being cloned if the client wants.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25281 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 18:39:17 +00:00
Chris Lattner
21f20558d6 Fix a bug I noticed by inspection: if the first instruction in the inlined
function was not an alloca, we wouldn't check the entry block for any allocas,
leading to increased stack space in some cases.  In practice, allocas are almost
always at the top of the block, so this was never noticed.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25280 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 18:16:48 +00:00
Chris Lattner
8acb249725 Fix 80 column violations
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25279 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-13 18:06:56 +00:00
Chris Lattner
dd9e956605 Preserve and update ETForest. Patch by Daniel Berlin
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25203 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-11 05:11:13 +00:00
Chris Lattner
88cac3d2f7 Switch these to using ETForest instead of DominatorSet to compute itself.
Patch written by Daniel Berlin!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25202 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-11 05:10:20 +00:00
Chris Lattner
19ef3d5be2 Switch this to using ETForest instead of DominatorSet to compute itself.
Patch written by Daniel Berlin!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25201 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-11 05:09:40 +00:00
Robert Bocchino
b52ee7f5ff Added support for the extractelement operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25181 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-10 19:05:34 +00:00
Robert Bocchino
56107e21a5 Added lower packed support for the extractelement operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25180 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-10 19:05:05 +00:00
Chris Lattner
baec98d00b Teach loopsimplify to update et-forest. Patch contributed by Daniel Berlin!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25153 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-09 08:03:08 +00:00
Chris Lattner
e8d56c536d fix some 176.gcc miscompilation from my previous patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25137 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-07 01:32:28 +00:00
Chris Lattner
4f637d4db1 silence some bogus gcc warnings on fenris
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25130 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-06 17:59:59 +00:00
Chris Lattner
ad0124c188 Enhance the shift-shift folding code to allow a no-op cast to occur in between
the shifts.

This allows us to fold this (which is the 'integer add a constant' sequence
from cozmic's scheme compmiler):

int %x(uint %anf-temporary776) {
        %anf-temporary777 = shr uint %anf-temporary776, ubyte 1
        %anf-temporary800 = cast uint %anf-temporary777 to int
        %anf-temporary804 = shl int %anf-temporary800, ubyte 1
        %anf-temporary805 = add int %anf-temporary804, -2
        %anf-temporary806 = or int %anf-temporary805, 1
        ret int %anf-temporary806
}

into this:

int %x(uint %anf-temporary776) {
        %anf-temporary776 = cast uint %anf-temporary776 to int
        %anf-temporary776.mask1 = add int %anf-temporary776, -2
        %anf-temporary805 = or int %anf-temporary776.mask1, 1
        ret int %anf-temporary805
}

note that instcombine already knew how to eliminate the AND that the two
shifts fold into.  This is tested by InstCombine/shift.ll:test26

-Chris


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25128 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-06 07:52:12 +00:00
Chris Lattner
830ed03e8b Simplify the code a bit more
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25126 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-06 07:22:22 +00:00
Chris Lattner
4d5542ce6e Extract a bunch of code out of visitShiftInst into FoldShiftByConstant. No
functionality changes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25125 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-06 07:12:35 +00:00
Chris Lattner
4eb40df1a8 Pull inline methods out of the pass class definition to make it easier to
read the code.

Do not internalize debugger anchors.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25067 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-03 19:13:17 +00:00
Duraid Madina
b5186853f8 getting there...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25021 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-26 13:48:44 +00:00
Chris Lattner
6860f6a01c Fix Transforms/ScalarRepl/2005-12-14-UnionPromoteCrash.ll, a crash on undefined
behavior in 126.gcc on big-endian systems.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24708 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-14 17:23:59 +00:00
Reid Spencer
2f1890792c Improve ResolveFunctions to:
a) use better local variable names (OldMT -> OldFT) where "M" is used to
   mean "Function" (perhaps it was previously "Method"?)
b) print out the module identifier in a warning message so that it is
   possible to track down in which module the error occurred.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24698 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-13 19:56:51 +00:00
Chris Lattner
a188894d67 Implement a little hack for parity with GCC on crafty. This speeds up
186.crafty by about 16% (from 15.109s to 13.045s) on my system.

This turns allocas with unions/casts into scalars.  For example crafty has
something like this:

    union doub {
      unsigned short i[4];
      long long d;
    };
int f(long long a) {
  return ((union doub){.d=a}).i[1];
}

Instead of generating loads and stores to an alloca, we now promote the
whole thing to a scalar long value.

This implements: Transforms/ScalarRepl/AggregatePromote.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24667 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-12 07:19:13 +00:00
Chris Lattner
e08dc62b1a getRawValue zero extens for unsigned values, use getsextvalue so that we
know that small negative values fit into the immediate field of addressing
modes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24608 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-05 18:23:57 +00:00
Chris Lattner
6b44ba2803 Wrap a long line, never internalize llvm.used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24602 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-05 05:07:38 +00:00
Chris Lattner
8e75ee212f Fix SimplifyCFG/2005-12-03-IncorrectPHIFold.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24581 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-03 18:25:58 +00:00
Chris Lattner
2bd4cb597a Fix a bug where we didn't realize that vaarg reads memory. This fixes
Transforms/DeadStoreElimination/2005-11-30-vaarg.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24545 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-30 19:38:22 +00:00
Andrew Lenharth
8dc2d50988 a few more comments on the interfaces and functions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24500 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-28 18:10:59 +00:00
Andrew Lenharth
bb227c1b79 Added documented rsprofiler interface. Also remove new profiler passes, the
old ones have been updated to implement the interface.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24499 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-28 18:00:38 +00:00
Jeff Cohen
3523f6e7a4 Fix VC++ warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24496 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-28 06:45:57 +00:00
Andrew Lenharth
701f5ac73c Random sampling (aka Arnold and Ryder) profiling. This is still preliminary, but it works on spec on x86 and alpha. The idea is to allow profiling passes to remember what profiling they inserted, then a random sampling framework is inserted which consists of duplicated basic blocks (without profiling), such that at each backedge in the program and entry into every function, the framework chooses whether to use the instrumented code or the instrumentation free code. The goal of such a framework is to make it reasonably cheap to do random sampling of very expensive profiling products (such as load-value profiling).
The code is organized into 3 parts (2 passes)
1) a linked set of profiling passes, which implement an analysis group (linked, like alias analysis are).  These insert profiling into the program, and remember what they inserted, so that at a later time they can be queried about any instruction.

2) a pass that handles inserting the random sampling framework.  This also has options to control how random samples are choosen.  Currently implemented are Global counters, register allocated global counters, and read cycle counter (see? there was a reason for it).

The profiling passes are almost identical to the existing ones (block, function, and null profiling is supported right now), and they are valid passes without the sampling framework (hence the existing passes can be unified with the new ones, not done yet).

Some things are a bit ugly still, but that should be fixed up soon enough.

Other todo? making the counter values not "magic 2^16 -1" values, but dynamically choosable.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24493 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-28 00:58:09 +00:00
Andrew Lenharth
b0826529f8 since reg2mem requires it, might as well mention that it preserves it
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24491 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-25 16:04:54 +00:00