70ed924e18
Teach PHIElimination to handle <undef> operands.
...
When a PHI use is <undef>, don't emit a copy in the predecessor block,
but insert an IMPLICIT_DEF instruction instead. This ensures that
virtual register uses are always jointly dominated by defs, even if some
of them are IMPLICIT_DEF.
llvm-svn: 159121
2012-06-25 03:36:12 +00:00
6b556f824d
Handle <undef> operands in TwoAddressInstructionPass.
...
When the source register to a 2-addr instruction is undefined, there is
no need to attempt any transformations - simply replace the source
register with the destination register.
This also comes up when lowering IMPLICIT_DEF instructions - make sure
the <undef> flag is moved to the new partial register def operand:
%vreg8<def> = INSERT_SUBREG %vreg9<undef>, %vreg0<kill>, sub_16bit
rewrite undef:
%vreg8<def> = INSERT_SUBREG %vreg8<undef>, %vreg0<kill>, sub_16bit
convert to:
%vreg8:sub_16bit<def,read-undef> = COPY %vreg0<kill>
llvm-svn: 159120
2012-06-25 03:27:12 +00:00
2e22e6a361
%RCX is not a function live-out in eh.return functions.
...
The function live-out registers must be live at all function returns,
and %RCX is only used by eh.return. When a function also has a normal
return, only %RAX holds a return value.
This fixes PR13188.
llvm-svn: 159116
2012-06-24 15:53:01 +00:00
704de074b8
llvm/lib: [CMake] Add explicit dependency to intrinsics_gen.
...
llvm-svn: 159112
2012-06-24 13:32:01 +00:00
3099ce9489
Allow controlling vectorization of boolean values separately from other integer types.
...
These are used as the result of comparisons, and often handled differently from larger integer types.
llvm-svn: 159111
2012-06-24 13:28:01 +00:00
0a045bbe4e
Remove dyn_cast + dereference pattern by replacing it with a cast and changing
...
the safety check to look for the same type we're going to actually cast to.
Fixes PR13180!
llvm-svn: 159110
2012-06-24 10:15:42 +00:00
fd5e6e7db1
Remove intrinsic specific instructions for (V)CVTPS2DQ and replace with patterns.
...
llvm-svn: 159109
2012-06-24 07:07:16 +00:00
b925230fb1
Remove intrinsic specific instructions for (V)CVTPS2DQ and replace with patterns.
...
llvm-svn: 159108
2012-06-24 06:55:37 +00:00
f48ec7a708
Fix build failures from r159106.
...
llvm-svn: 159107
2012-06-24 06:08:31 +00:00
bab2b89944
Remove intrinsic specific instructions for CVTPD2PS and replace with just patterns.
...
llvm-svn: 159106
2012-06-24 05:44:31 +00:00
3cee08ce7d
Remove intrinsic specific instructions for CVTPD2DQ. Replace with patterns.
...
llvm-svn: 159105
2012-06-24 05:33:24 +00:00
b74ae9c5b2
Tab to spaces. No functionality change.
...
llvm-svn: 159104
2012-06-24 04:07:14 +00:00
38f4c93837
llvm/lib/Support/Errno.cpp: [Win32] Fix usage of strerror_s().
...
FYI, two arguments of strerror_s() is provided in C++ on msvc.
llvm-svn: 159103
2012-06-24 03:51:04 +00:00
cca44e219f
VMCore/CMakeLists.txt: [CMake][MSVC] Add "/Og-" to Function.cpp on msvc10. Otherwise, it took over 20 minutes to compile.
...
FIXME: Suppressing optimizations to core libraries would not be good thing.
llvm-svn: 159097
2012-06-24 03:48:29 +00:00
bfb07fb562
Remove a dangling reference to a deleted instruction. Fixes PR13185!
...
llvm-svn: 159096
2012-06-24 01:44:08 +00:00
3c680dec8a
Remove code i'd been testing with but didn't mean to commit. Oops
...
llvm-svn: 159094
2012-06-24 00:08:36 +00:00
fe212e762f
DAG legalisation can now handle illegal fma vector types by scalarisation
...
llvm-svn: 159092
2012-06-24 00:05:44 +00:00
a899cc15f1
Remove intrinsic specific instructions for (V)CVTDQ2PS. Use a Pat instead instead.
...
llvm-svn: 159090
2012-06-23 22:33:14 +00:00
4b06b1a0ee
Allow BBVectorize to fuse compare instructions.
...
llvm-svn: 159088
2012-06-23 21:52:50 +00:00
7e9415220a
Make CVTDQ2PS instruction use SSE2 predicate instead of SSE1. No functional change because there are no patterns in the instructions. Also fix a typo in a comment.
...
llvm-svn: 159087
2012-06-23 20:52:45 +00:00
24e3418215
Move CVTPD2DQ to use SSE2 predicate instead of SSE3. Move DQ2PD and PD2DQ to the SSE2 section of the file.
...
llvm-svn: 159086
2012-06-23 20:15:42 +00:00
53ffe55a66
Add a microoptimization note.
...
llvm-svn: 159082
2012-06-23 15:19:31 +00:00
ac9fb36c31
Clean-up after r159077.
...
Remove temporary GlobalVariable constructors now that Clang has been
updated (r159078).
llvm-svn: 159079
2012-06-23 12:14:23 +00:00
cbe34b4cc9
Extend the IL for selecting TLS models (PR9788)
...
This allows the user/front-end to specify a model that is better
than what LLVM would choose by default. For example, a variable
might be declared as
@x = thread_local(initialexec) global i32 42
if it will not be used in a shared library that is dlopen'ed.
If the specified model isn't supported by the target, or if LLVM can
make a better choice, a different model may be used.
llvm-svn: 159077
2012-06-23 11:37:03 +00:00
8e00efeace
Optimized usage of new SwitchInst case values (IntegersSubset type) in Local.cpp, Execution.cpp and BitcodeWriter.cpp.
...
I got about 1% of compile-time improvement on my machines (Ubuntu 11.10 i386 and Ubuntu 12.04 x64).
llvm-svn: 159076
2012-06-23 10:58:58 +00:00
8c03ea79c4
Use correct memory types for (V)CVTDQ2PD instructions.
...
llvm-svn: 159075
2012-06-23 08:30:27 +00:00
2361cd9897
Silence an unused variable warning on release builds.
...
llvm-svn: 159074
2012-06-23 08:09:30 +00:00
1cac50bc5e
Compress flags in X86 op folding to reduce space in static tables.
...
llvm-svn: 159073
2012-06-23 08:01:18 +00:00
d9c7d0dda4
Make helper method static since it doesn't use anything in the class.
...
llvm-svn: 159071
2012-06-23 04:58:41 +00:00
431f1e7192
Remove intrinsic specific instructions for 128-bit (V)CVTDQ2PD. Replace with intrinsic patterns. Mem forms omitted because the load size is only 64-bits.
...
llvm-svn: 159070
2012-06-23 04:23:36 +00:00
502e4c6ac4
Teach LiveVariables to handle <undef> operands.
...
It's simple: Don't treat <undef> operands as uses, and don't assume a
virtual register has a defining instruction unless a real use has been
seen.
llvm-svn: 159061
2012-06-23 02:23:00 +00:00
a3088f09b3
Handle aliases to tls variables in all architectures, not just x86.
...
llvm-svn: 159058
2012-06-23 00:30:03 +00:00
68c2f9a9a7
(sub X, imm) gets canonicalized to (add X, -imm)
...
There are patterns to handle immediates when they fit in the immediate field.
e.g. %sub = add i32 %x, -123
=> sub r0, r0, #123
Add patterns to catch immediates that do not fit but should be materialized
with a single movw instruction rather than movw + movt pair.
e.g. %sub = add i32 %x, -65535
=> movw r1, #65535
sub r0, r0, r1
rdar://11726136
llvm-svn: 159057
2012-06-23 00:29:06 +00:00
de8c6fb24f
BoundsChecking: attach debug info to traps to make my life a bit more sane
...
llvm-svn: 159055
2012-06-23 00:12:34 +00:00
087affe2f3
ARM: Add a better diagnostic for some out of range immediates.
...
As an example of how the custom DiagnosticType can be used to provide
better operand-mismatch diagnostics, add a custom diagnostic for
the imm0_15 operand class used for several system instructions.
Update the tests to expect the improved diagnostic.
rdar://8987109
llvm-svn: 159051
2012-06-22 23:56:48 +00:00
460e94d842
Add support for the PPC isel instruction.
...
The isel (integer select) instruction is supported on the 440 and A2
embedded cores and on the POWER7.
llvm-svn: 159045
2012-06-22 23:10:08 +00:00
a127fc780a
Remove ProcessImplicitDefs.h which was unused.
...
The ProcessImplicitDefs class can be local to its implementation file.
llvm-svn: 159041
2012-06-22 22:27:36 +00:00
b033dede17
Also verify the def index for early clobbers.
...
llvm-svn: 159039
2012-06-22 22:23:58 +00:00
f5cdea3d79
Whitespace.
...
llvm-svn: 159035
2012-06-22 22:07:19 +00:00
4fa84ba8b9
Delete a boring statistic.
...
llvm-svn: 159030
2012-06-22 20:40:15 +00:00
c61edda0ab
Store live intervals in an IndexedMap.
...
It is both smaller and faster than DenseMap.
llvm-svn: 159029
2012-06-22 20:37:52 +00:00
8db5547252
Revert r158679 - use case is unclear (and it increases the memory footprint).
...
Original commit message:
Allow up to 64 functional units per processor itinerary.
This patch changes the type used to hold the FU bitset from unsigned to uint64_t.
This will be needed for some upcoming PowerPC itineraries.
llvm-svn: 159027
2012-06-22 20:27:13 +00:00
f5bd6c6510
EmitZerofill should take a 64-bit size or else it's chopping off large zero-filled global. rdar://11729134
...
llvm-svn: 159023
2012-06-22 20:14:46 +00:00
48828bb402
Fix a crash in --debug code.
...
Don't try to print out the live range of a physreg.
llvm-svn: 159021
2012-06-22 19:51:41 +00:00
48a1647c93
Don't depend on live ranges being present.
...
DBG_VALUE instructions could be referring to non-existing virtual
registers.
llvm-svn: 159020
2012-06-22 18:51:35 +00:00
8a833649e5
Simplify handleMove() a bit.
...
There is no need to check for physreg live ranges. They don't exist any
more.
llvm-svn: 159019
2012-06-22 18:38:57 +00:00
37e797fedc
Stop computing physreg live ranges.
...
Everyone is using on-demand regunit ranges now.
llvm-svn: 159018
2012-06-22 18:20:50 +00:00
bbad269a3e
Remove some redundant LIS->hasInterval() checks.
...
These functions only operate on virtual registers now, and they all have
live ranges.
llvm-svn: 159015
2012-06-22 17:49:44 +00:00
7809578cfe
Use MRI::isConstantPhysReg() to check remat feasibility.
...
Don't depend on LiveIntervals::hasInterval() to determine if a physreg
is reserved and constant.
llvm-svn: 159013
2012-06-22 17:31:01 +00:00
3244963ecc
Use regunit liveness to guide LiveDebugVariables.
...
This should produce the same results as using physreg liveness directly.
llvm-svn: 159009
2012-06-22 17:15:32 +00:00
Jakob Stoklund Olesen