f09c372191
Added support for FP Registers
...
llvm-svn: 52079
2008-06-07 21:32:41 +00:00
4595a72c79
Update VC project files, patch by Cédric Venet!
...
llvm-svn: 52077
2008-06-07 19:33:54 +00:00
89200c9177
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
003b4b0cd2
Fix run line.
...
llvm-svn: 52072
2008-06-07 08:40:16 +00:00
1a0835017a
Revert r52046. It broke cbe on x86 / Mac OS X.
...
llvm-svn: 52071
2008-06-07 07:50:29 +00:00
f6743d70ab
CodeGen support for insertvalue and extractvalue, and for loads and
...
stores of aggregate values.
llvm-svn: 52069
2008-06-07 02:02:36 +00:00
0bd08cf64c
Connect successors before creating the DAG node for the branch. This has
...
no visible functionality change, but enables a future patch where node creation
will update the CFG if it decides to create an unconditional rather than a conditional branch.
llvm-svn: 52067
2008-06-07 00:00:23 +00:00
cbcc495c1b
get rid of ExtractValueInst::init's Value argument, it is already passed to the UnaryInstruction ctor
...
llvm-svn: 52064
2008-06-06 21:06:32 +00:00
0b8f2c53a2
Typo.
...
llvm-svn: 52062
2008-06-06 21:00:10 +00:00
21ba184b27
make ExtractValueInst derived from UnaryInstruction
...
llvm-svn: 52061
2008-06-06 20:28:12 +00:00
c324be32c4
Enable stack coloring by default.
...
llvm-svn: 52057
2008-06-06 19:52:44 +00:00
9bf9110d93
PPC preferred loop alignment is 16.
...
llvm-svn: 52056
2008-06-06 19:50:46 +00:00
8549e4ca07
LoopSimplify preserves AA.
...
llvm-svn: 52053
2008-06-06 17:50:58 +00:00
656a917ad9
Print debug message only if there are dead passes.
...
llvm-svn: 52052
2008-06-06 17:50:36 +00:00
f7647e432c
Include loop deletion and the ADCE rewrite in the release notes.
...
llvm-svn: 52048
2008-06-06 16:23:15 +00:00
089018fb05
Testcase for PR2418
...
llvm-svn: 52047
2008-06-06 16:08:56 +00:00
f69bc3df9b
Handle assembler identifiers specially in CBE. This fixes PR2418.
...
llvm-svn: 52046
2008-06-06 16:08:26 +00:00
f1123e58fc
Tighten up the abstraction slightly.
...
llvm-svn: 52045
2008-06-06 12:49:32 +00:00
13237ac3b9
Wrap MVT::ValueType in a struct to get type safety
...
and better control the abstraction. Rename the type
to MVT. To update out-of-tree patches, the main
thing to do is to rename MVT::ValueType to MVT, and
rewrite expressions like MVT::getSizeInBits(VT) in
the form VT.getSizeInBits(). Use VT.getSimpleVT()
to extract a MVT::SimpleValueType for use in switch
statements (you will get an assert failure if VT is
an extended value type - these shouldn't exist after
type legalization).
This results in a small speedup of codegen and no
new testsuite failures (x86-64 linux).
llvm-svn: 52044
2008-06-06 12:08:01 +00:00
1152ca9101
As Chris suggested, handle the situation if ShAmt larger than BitWidth,
...
otherwise, opt might crash.
llvm-svn: 52041
2008-06-06 08:32:05 +00:00
6d7a144453
Refine stack slot interval weight computation.
...
llvm-svn: 52040
2008-06-06 07:54:39 +00:00
5b3e396c24
Turn stdout into binary mode during bitcode emission.
...
This is necessary on windows targets, since stdout is in text mode there.
Patch by Julien Lerouge!
llvm-svn: 52038
2008-06-06 07:24:01 +00:00
5e8644071d
Provide hooks for libgcc symbols' address resolution inside lli on mingw32.
...
Patch by Julien Lerouge!
llvm-svn: 52037
2008-06-06 07:20:07 +00:00
1a6e0d613f
Added custom isel for MUL, SDIVREM, UDIVREM, SMUL_LOHI and UMUL_LOHI nodes
...
MUL is not anymore directly matched because its a pseudoinstruction.
LogicI class fixed to zero-extend immediates.
llvm-svn: 52036
2008-06-06 06:37:31 +00:00
4eed3afda0
Added custom SELECT_CC lowering
...
Added special isel for ADDE,SUBE and new patterns to match SUBC,ADDC
llvm-svn: 52031
2008-06-06 00:58:26 +00:00
1d304a99d7
Revert 52002.
...
llvm-svn: 52030
2008-06-05 23:57:06 +00:00
3816aa7ef7
Expose a public interface to this pass.
...
llvm-svn: 52029
2008-06-05 23:45:18 +00:00
0282d0247e
Let libcall semantics decide whether it knows about functions
...
with definitions or not.
llvm-svn: 52028
2008-06-05 23:38:34 +00:00
c7007a7211
Warn of potential violations of strict aliasing rules.
...
llvm-svn: 52027
2008-06-05 23:00:08 +00:00
9e76c047d1
Don't break strict aliasing.
...
llvm-svn: 52026
2008-06-05 22:59:21 +00:00
33aac00892
Describe loop index split.
...
llvm-svn: 52021
2008-06-05 21:44:00 +00:00
ae408aa83f
Wording fixes. Thanks Bill!
...
llvm-svn: 52017
2008-06-05 18:45:33 +00:00
05d1cd99b3
Remove debugging code.
...
llvm-svn: 52016
2008-06-05 18:43:34 +00:00
80ae8a7194
fix typos, tweak stuff, make it to validate
...
llvm-svn: 52015
2008-06-05 18:39:01 +00:00
65dc2b2e06
Use the newly created helper on LiveIntervals.
...
llvm-svn: 52013
2008-06-05 17:22:53 +00:00
35e2dfe1cf
Add a helper for constructing new live ranges that ended from an instruction to the end of its MBB.
...
llvm-svn: 52012
2008-06-05 17:15:43 +00:00
b0ec18bd81
Move i128 on x86-64 from the codegen section to the x86-specific section.
...
llvm-svn: 52010
2008-06-05 16:15:39 +00:00
4277e7e49a
Add a note about improved alignment detection.
...
llvm-svn: 52009
2008-06-05 15:49:29 +00:00
23254a8f83
Add a note mentioning --view-sunit-dags improvements.
...
llvm-svn: 52008
2008-06-05 15:46:36 +00:00
46105567f7
Expand the bullet item about i128 support and APInt usage in codegen.
...
llvm-svn: 52007
2008-06-05 15:43:11 +00:00
c775e462a8
Add a test case for opt -instcombine bug fix in revision 52003.
...
llvm-svn: 52004
2008-06-05 14:25:11 +00:00
fbe1dc240c
If BitWidth equals to ShtAmt, the RHSKnownZero[BitWidth-ShiftAmt-1] will
...
crash the opt. Just fix this.
Test case in llvm/test/Transforms/InstCombine/2008-06-05-ashr-crash.ll
llvm-svn: 52003
2008-06-05 14:23:44 +00:00
5eef32ba94
Change the Verifier to support returning first class aggregrates.
...
Add a testcase for functions returning first class aggregrates.
llvm-svn: 52002
2008-06-05 14:00:36 +00:00
3f8de8b130
Add a test case for APInt bug fix in r51999.
...
llvm-svn: 52000
2008-06-05 13:42:21 +00:00
1247c07742
As comments said, for negative value, the arithmetic
...
over-shift-right should return -1. So here it should be signed-extended,
when bitwidth larger than 64.
test case: llvm/test/ExecutionEngine/2008-06-05-APInt-OverAShr.ll
llvm-svn: 51999
2008-06-05 13:27:38 +00:00
0702adc65e
Escape "$#".
...
llvm-svn: 51998
2008-06-05 12:54:44 +00:00
812989b147
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
03a4ad4db9
Move this to the proper section.
...
I misread Optimizer Improvements as Other Improvements for some reason.
llvm-svn: 51996
2008-06-05 12:51:50 +00:00
03d9625286
Adding release notes for bindings improvements.
...
llvm-svn: 51995
2008-06-05 12:48:32 +00:00
e0c5adc158
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
463f86639d
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
230d6fbfeb
Use use_iterator::getOperandNo instead of CallSite::hasArgument to check if a
...
function is passed as an argument instead of called. Also do this check a bit
earlier.
llvm-svn: 51990
2008-06-05 08:34:25 +00:00
b0dffd691c
* Make CallSite::hasArgument const and let it take a const parameter.
...
llvm-svn: 51989
2008-06-05 08:04:58 +00:00
6cb64030e5
more updates and random notes, including changes up through Week-of-Mon-20080324.
...
llvm-svn: 51988
2008-06-05 08:02:49 +00:00
5afc2740b7
Update comments and documentation to reflect that GCSE and ValueNumbering are
...
deprecated by the GVN and GVNPRE passes.
llvm-svn: 51983
2008-06-05 07:55:49 +00:00
18fbf1bc37
Cast because I think I know what I'm doing. There are warnings about converting
...
from 64-bit to 32-bit.
llvm-svn: 51982
2008-06-05 07:35:27 +00:00
63a2cb59b0
Fix 80 column violations.
...
llvm-svn: 51981
2008-06-05 07:26:15 +00:00
b80d63268a
Remove unused parameter. Some clients might use -Werror when compiling and
...
cause drama.
llvm-svn: 51980
2008-06-05 07:24:08 +00:00
618a99f920
add llvmc2 notes, stack realignment.
...
llvm-svn: 51979
2008-06-05 06:57:39 +00:00
a1a4c9a227
various status updates.
...
llvm-svn: 51978
2008-06-05 06:35:40 +00:00
b711222b27
Start adding 2.3 content.
...
llvm-svn: 51977
2008-06-05 06:25:56 +00:00
976b1eee81
Fix a memcpy lowering bug. Even though the memcpy alignment is smaller than the desired alignment, the frame destination alignment may still be larger than the desired alignment. Don't change its alignment to something smaller.
...
llvm-svn: 51970
2008-06-04 23:37:54 +00:00
61c7f2a633
Remove unneeded #include.
...
llvm-svn: 51955
2008-06-04 18:28:10 +00:00
09ebde227e
Fix inst_iterator example.
...
llvm-svn: 51954
2008-06-04 18:20:42 +00:00
9e27316337
Oops. Should not be enabled by default.
...
llvm-svn: 51953
2008-06-04 18:09:20 +00:00
c596ec04e1
Rewrite a bunch of the CBE's inline asm code, giving it the
...
ability to handle indirect input operands. This fixes PR2407.
llvm-svn: 51952
2008-06-04 18:03:28 +00:00
3a697779fe
Correctly construct live intervals for the copies we inserted into the predecessors of a block containing a PHI.
...
llvm-svn: 51950
2008-06-04 17:55:58 +00:00
805cde4fbf
Revert this.
...
llvm-svn: 51949
2008-06-04 17:21:44 +00:00
2353f35989
Replace two manual loops with calls to CallSite::hasArguments (no functional changes).
...
llvm-svn: 51947
2008-06-04 16:57:50 +00:00
d901e5842f
Add CallSite::hasArgument to allow for seeing if a call passes a certain value as an argument quickly.
...
llvm-svn: 51946
2008-06-04 16:31:12 +00:00
76d8dec5db
Add a Name parameter to two of the init methods of GetElementPointer to make the name setting more consistent.
...
llvm-svn: 51945
2008-06-04 16:14:12 +00:00
a6bb22e9bc
Fix example to be valid LLVM assembly.
...
llvm-svn: 51942
2008-06-04 15:46:35 +00:00
cfd41dbd5e
Implement the two constructors in InsertValueInst and ExtractValueInst.
...
Add a Name argment to two init methods in these classes as well to make things
a bit more consistent.
llvm-svn: 51937
2008-06-04 14:40:55 +00:00
12a0222a01
Add a stack slot coloring pass. Not yet enabled.
...
llvm-svn: 51934
2008-06-04 09:18:41 +00:00
a15e7eb29b
LowerSubregs should not clobber any analysis.
...
llvm-svn: 51933
2008-06-04 09:17:16 +00:00
eecdf659e8
Move #include to right place.
...
llvm-svn: 51932
2008-06-04 09:16:33 +00:00
c5b3a3bea5
Register if-converter pass for -debug-pass.
...
llvm-svn: 51931
2008-06-04 09:15:51 +00:00
93af6ce129
More pass manager debugging outputs.
...
llvm-svn: 51930
2008-06-04 09:13:31 +00:00
fc3c489b52
Change packed struct layout so that field sizes
...
are the same as in unpacked structs, only field
positions differ. This only matters for structs
containing x86 long double or an apint; it may
cause backwards compatibility problems if someone
has bitcode containing a packed struct with a
field of one of those types.
The issue is that only 10 bytes are needed to
hold an x86 long double: the store size is 10
bytes, but the ABI size is 12 or 16 bytes (linux/
darwin) which comes from rounding the store size
up by the alignment. Because it seemed silly not
to pack an x86 long double into 10 bytes in a
packed struct, this is what was done. I now
think this was a mistake. Reserving the ABI size
for an x86 long double field even in a packed
struct makes things more uniform: the ABI size is
now always used when reserving space for a type.
This means that developers are less likely to
make mistakes. It also makes life easier for the
CBE which otherwise could not represent all LLVM
packed structs (PR2402).
Front-end people might need to adjust the way
they create LLVM structs - see following change
to llvm-gcc.
llvm-svn: 51928
2008-06-04 08:21:45 +00:00
6449690da4
Add #includes required by GCC 4.3, thanks for Zhongxing Xu
...
for reporting this.
llvm-svn: 51926
2008-06-04 04:46:14 +00:00
326a03732e
Some Mips minor fixes
...
Added support for mips little endian arch => mipsel
llvm-svn: 51923
2008-06-04 01:45:25 +00:00
c777d9e6fe
We need to subtract one from this index because live ranges are open at the end.
...
llvm-svn: 51922
2008-06-04 00:38:56 +00:00
a7d8649f78
Fix spellnig error
...
llvm-svn: 51917
2008-06-03 19:13:20 +00:00
6c871c0866
Testcase for LoopIndexSplit and DomFrontier.
...
llvm-svn: 51916
2008-06-03 18:32:27 +00:00
2df82e7cec
LoopIndexSplit can sometimes result in cases where a block in its own domfrontier.
...
Don't crash when we encounter one of these.
llvm-svn: 51915
2008-06-03 18:29:48 +00:00
4c75de7f39
Expand documentation of StringConstantPrefix.
...
llvm-svn: 51911
2008-06-03 18:15:03 +00:00
83e468ae23
Prevent a crash in debug dumps.
...
llvm-svn: 51910
2008-06-03 18:14:29 +00:00
355b74acc2
Add StringConstantPrefix to control what the
...
assembler names of string constants look like.
llvm-svn: 51909
2008-06-03 18:09:06 +00:00
c0e9ff6e52
Find a better place to output hex constants corresponding to integers.
...
llvm-svn: 51904
2008-06-03 15:39:51 +00:00
2dd8fdc78a
Fixed bug in bad behavior in calculateFrameObjectOffsets,
...
the solution commited is different from the previous patch to
avoid int and unsigned comparison
llvm-svn: 51899
2008-06-03 08:46:59 +00:00
097826643b
Do not run loop-aligner at -fast (e.g. -O0).
...
llvm-svn: 51898
2008-06-03 06:56:08 +00:00
499c119d7c
Revert this patch
...
llvm-svn: 51897
2008-06-03 06:18:19 +00:00
25b3b49dbc
nounwindify.
...
llvm-svn: 51893
2008-06-03 01:21:11 +00:00
ab85d6b6af
"Unable to schedule <A> required by <B>" is more helpful then
...
"Unable to handle Pass that requires lower level Analysis pass"
llvm-svn: 51892
2008-06-03 01:20:02 +00:00
bb4720c43f
Add debugging aid.
...
llvm-svn: 51891
2008-06-03 01:02:16 +00:00
2ad7e7341c
Fix whitespace in whitespace-significant pseudocode in a comment.
...
llvm-svn: 51890
2008-06-03 00:57:21 +00:00
3db11c2bd9
Constant folding for insertvalue and extractvalue.
...
llvm-svn: 51889
2008-06-03 00:15:20 +00:00
7314d0ee3c
Update dom tree. Fix PR 2372.
...
llvm-svn: 51887
2008-06-02 22:52:56 +00:00
057240f4f0
Fold adds and subtracts of zero immediately, instead of waiting
...
for dagcombine to do this.
llvm-svn: 51886
2008-06-02 22:27:05 +00:00
5793cbcf44
Minor cosmetic patch so that the hex equivalent of a decimal
...
constant shows up in the assembly language output. Helps with
debugging without a HP calculator having to be handy.
llvm-svn: 51885
2008-06-02 22:19:12 +00:00
Bruno Cardoso Lopes