RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 12:12:47 +01:00
Code Issues Projects Releases Wiki Activity
52,114 Commits 4 Branches 5 Tags 704 MiB
07d521ed99
Commit Graph

340 Commits

Author SHA1 Message Date
Evan Cheng
07d521ed99 Fix a pasto. Also simplify for Bill's benefit. 
llvm-svn: 82505
 2009-09-22 01:48:19 +00:00
Evan Cheng
a6d602a5c1 Clean up spill weight computation. Also some changes to give loop induction 
variable increment / decrement slighter high priority. 

This has major impact on some micro-benchmarks. On MultiSource/Applications
and spec tests, it's a minor win. It also reduce 256.bzip instruction count
by 8%, 55 on 164.gzip on i386 / Darwin.

llvm-svn: 82485
 2009-09-21 21:12:25 +00:00
Evan Cheng
3edb8b18a5 Fix PR4986. "r1024 = insert_subreg r1024, undef, 2" cannot be turned in an implicit_def. Instead, it's an identity copy so it should be eliminated. Also make sure to update livevariable kill information. 
llvm-svn: 82436
 2009-09-21 04:32:32 +00:00
Bob Wilson
9bcea11785 Convert more tests to FileCheck. 
llvm-svn: 81915
 2009-09-15 20:58:02 +00:00
Sandeep Patel
7727a68464 Fix superreg use in ARMAsmPrinter. Approved by Anton Korobeynikov. 
llvm-svn: 81878
 2009-09-15 17:53:11 +00:00
Anton Korobeynikov
fa4c7562d5 Define proper subreg sets for arm - this should fix bunch of subtle problems 
with subreg - superreg mapping and also fix PR4965.

llvm-svn: 81657
 2009-09-13 00:59:43 +00:00
Dan Gohman
8e55f0f55b Remove an unnecessary -f. 
llvm-svn: 81546
 2009-09-11 18:41:06 +00:00
Dan Gohman
f2c290dfa6 Convert more tests to avoid llvm-as. 
llvm-svn: 81545
 2009-09-11 18:36:27 +00:00
Bob Wilson
7b39f31422 Don't swap the operands of a subtraction when trying to create a 
post-decrement load/store.

llvm-svn: 81464
 2009-09-10 22:09:31 +00:00
Bob Wilson
877a857b4b Fix pr4939: Change FPCCToARMCC to translate SETOLE to ARMCC::LS. 
See the bug report for details.

llvm-svn: 81397
 2009-09-09 23:14:54 +00:00
Dan Gohman
142428ce64 Eliminate more uses of llvm-as and llvm-dis. 
llvm-svn: 81293
 2009-09-09 00:09:15 +00:00
Anton Korobeynikov
2b6ef7724e Unbreak getOnesVector() / getZeroVector() to use valid ARM extended imm's. 
llvm-svn: 81262
 2009-09-08 22:51:43 +00:00
Anton Korobeynikov
0b3a620d60 Add NEON 'laned' operations. This fixes another bunch of gcc testsuite fails and 
makes the code faster.

llvm-svn: 81220
 2009-09-08 15:22:32 +00:00
Daniel Dunbar
2a64e85835 Remove stale greps. 
llvm-svn: 80986
 2009-09-04 05:07:52 +00:00
Bob Wilson
25410ac604 Convert tests to FileCheck. 
llvm-svn: 80983
 2009-09-04 04:07:19 +00:00
Bob Wilson
9e02907942 Convert a test to FileCheck. 
llvm-svn: 80975
 2009-09-04 00:32:31 +00:00
Evan Cheng
41e87f2f13 Reference to hidden symbols do not have to go through non-lazy pointer in non-pic mode. rdar://7187172. 
llvm-svn: 80904
 2009-09-03 07:04:02 +00:00
Anton Korobeynikov
7125d63acf More missed vdup patterns 
llvm-svn: 80838
 2009-09-02 21:21:28 +00:00
Bob Wilson
6972a16bbc Add support for generating code for vst{234}lane intrinsics. 
llvm-svn: 80707
 2009-09-01 18:51:56 +00:00
Bob Wilson
75b2b04e1e Fix incorrect declarations of intrinsics in this test. 
llvm-svn: 80705
 2009-09-01 18:50:43 +00:00
Bob Wilson
d638cc8869 Add test for vld{234}_lane instructions. 
llvm-svn: 80658
 2009-09-01 04:27:10 +00:00
Bob Wilson
03f5a5bfff Fix pr4843: When an instruction has multiple destination registers that are 
tied to different source registers, the TwoAddressInstructionPass needs to
be smarter.  Change it to check before replacing a source register whether
that source register is tied to a different destination register, and if so,
defer handling it until a subsequent iteration.

llvm-svn: 80654
 2009-09-01 04:18:40 +00:00
Jim Grosbach
4e0e9a4870 SJLJ is arm/darwin only for now. force the triple for the test 
llvm-svn: 80651
 2009-09-01 02:34:49 +00:00
Jim Grosbach
9a220088ac Clean up LSDA name generation and use for SJLJ exception handling. This 
makes an eggregious hack somewhat more palatable. Bringing the LSDA forward
and making it a GV available for reference would be even better, but is
beyond the scope of what I'm looking to solve at this point.

Objective C++ code could generate function names that broke the previous
scheme. This fixes that.

llvm-svn: 80649
 2009-09-01 01:57:56 +00:00
David Goodwin
0fc3764297 Don't mark a register live at an undef use. 
llvm-svn: 80621
 2009-08-31 20:47:02 +00:00
Anton Korobeynikov
17529667db Add missed pattern 
llvm-svn: 80502
 2009-08-30 19:06:39 +00:00
Anton Korobeynikov
a261afbf14 EXTRACT_VECTOR_ELEMENT can have result type different from element type. 
Remove the assertion and generalize the code for ARM NEON stuff.

llvm-svn: 80498
 2009-08-30 17:14:54 +00:00
Anton Korobeynikov
b2e6f5eed4 Do not assert on too wide splats we don't support. 
llvm-svn: 80409
 2009-08-29 00:08:18 +00:00
Anton Korobeynikov
9fd6082c10 Add missed extract_element pattern 
llvm-svn: 80408
 2009-08-28 23:41:26 +00:00
Evan Cheng
d7a07ab112 Let Darwin linker auto-synthesize stubs and lazy-pointers. This deletes a bunch of nasty code in ARM asm printer. 
llvm-svn: 80404
 2009-08-28 23:18:09 +00:00
Evan Cheng
2d5d3700e9 v4, v5 does not support sxtb / sxth. 
llvm-svn: 80322
 2009-08-28 00:31:43 +00:00
Anton Korobeynikov
cb0fdc4505 scalar_to_vector is fully legal now (implemented as subreg accesses) 
llvm-svn: 80249
 2009-08-27 16:04:47 +00:00
Anton Korobeynikov
e17a92c545 Ok, sometimes it's profitable to turn scalar_to_vector stuff into subreg access. 
Add a testcase.

llvm-svn: 80246
 2009-08-27 14:51:42 +00:00
Evan Cheng
984f8efcaa Fix PR4789. Teach eliminateFrameIndex how to handle VLDRQ and VSTRQ which cannot fold any immediate offset. 
llvm-svn: 80191
 2009-08-27 01:23:50 +00:00
Bob Wilson
c7d92cfb15 Convert some more Neon tests to FileCheck. 
llvm-svn: 80120
 2009-08-26 18:11:50 +00:00
Anton Korobeynikov
1c904039ce Expand scalar_to_vector - we don't have any isel logic for it now 
llvm-svn: 80107
 2009-08-26 16:26:09 +00:00
David Goodwin
047f69da86 Fixup register kills after scheduling. 
llvm-svn: 80002
 2009-08-25 17:03:05 +00:00
Dan Gohman
bf08e82d8e Remove obsolete -f flags. 
llvm-svn: 79992
 2009-08-25 15:38:29 +00:00
Dale Johannesen
add8a314dd Split test into 3. 
llvm-svn: 79926
 2009-08-24 17:51:19 +00:00
Eli Friedman
79615641f1 Make x86 test actually test x86 code generation. Fix the 
construct on ARM, which was breaking by coincidence, and add a similar 
testcase for ARM.

llvm-svn: 79719
 2009-08-22 03:13:10 +00:00
Bob Wilson
79c0af15d0 Use CHECK-NEXT to make sure we're only getting one copy of each shuffle 
instruction.

llvm-svn: 79702
 2009-08-22 00:13:23 +00:00
Bob Wilson
6d4400e852 Match VTRN, VZIP, and VUZP shuffles. Restore the tests for these operations, 
now using shuffles instead of intrinsics.

llvm-svn: 79673
 2009-08-21 20:54:19 +00:00
Bob Wilson
0da4ec0046 Add some tests for vext.16 and vext.32. 
llvm-svn: 79638
 2009-08-21 16:35:24 +00:00
Bob Wilson
c046b62f1a Remove Neon intrinsics for VZIP, VUZP, and VTRN. We will represent these as 
vector shuffles.  Temporarily remove the tests for these operations until the
new implementation is working.

llvm-svn: 79579
 2009-08-21 00:01:42 +00:00
Bob Wilson
fae9057bf0 Add support for Neon VEXT (vector extract) shuffles. 
This is derived from a patch by Anton Korzh.  I modified it to recognize
the VEXT shuffles during legalization and lower them to a target-specific
DAG node.

llvm-svn: 79428
 2009-08-19 17:03:43 +00:00
Bill Wendling
962adec4ee Reapply r79127. It was fixed by d0k. 
llvm-svn: 79136
 2009-08-15 21:21:19 +00:00
Bill Wendling
bfebbb6477 Revert r79127. It was causing compilation errors. 
llvm-svn: 79135
 2009-08-15 21:14:01 +00:00
Evan Cheng
5d841097a9 Change allowsUnalignedMemoryAccesses to take type argument since some targets 
support unaligned mem access only for certain types. (Should it be size
instead?)

ARM v7 supports unaligned access for i16 and i32, some v6 variants support it
as well.

llvm-svn: 79127
 2009-08-15 19:23:44 +00:00
Jakob Stoklund Olesen
7f4ef2d59a Refine EarlyClobber assert in register scavenger. 
It is legal for an inline asm operand to use an earlyclobber register if the
use operand is tied to the earlyclobber operand. The issue is discussed here:

  http://gcc.gnu.org/ml/gcc/1999-04n/msg00431.html

We should perhaps let only the machine code verifier worry about these finer
details. EarlyClobber operands are not really interesting to the scavenger.

This fixes PR4528 for the third time.

llvm-svn: 79122
 2009-08-15 18:16:58 +00:00
Jakob Stoklund Olesen
8f6660c417 Don't setCalleeSavedInfoValid() until spills are interted. 
In a naked function, the flag is never set and getPristineRegs() returns an
empty list. That means naked functions are able to clobber callee saved
registers, but that is the whole point of naked functions.

This fixes PR4716.

llvm-svn: 79096
 2009-08-15 13:10:46 +00:00