Commit Graph

24816 Commits

Author SHA1 Message Date
Chris Lattner b22eb6304f Add a note
llvm-svn: 28401
2006-05-19 20:55:31 +00:00
Chris Lattner 17f1f1a56c Split the SSE readme items out into their own README.
llvm-svn: 28400
2006-05-19 20:51:43 +00:00
Chris Lattner 427ea6f0a7 Split FP-stack notes out of the main readme. Next up: splitting out SSE.
llvm-svn: 28399
2006-05-19 20:45:52 +00:00
Chris Lattner 240f846495 Move a target-independent note out of the X86 readme.
llvm-svn: 28398
2006-05-19 20:45:08 +00:00
Chris Lattner d6a25a08d1 Particularly ugly code.
llvm-svn: 28397
2006-05-19 19:41:33 +00:00
Chris Lattner 4203df25b8 new testcase
llvm-svn: 28396
2006-05-19 19:34:09 +00:00
Reid Spencer 2452c94df4 Fix a doxygen problem and break lines at 80 columns
llvm-svn: 28395
2006-05-19 19:09:46 +00:00
Reid Spencer 3949d32155 Fix some doxygen usage in these headers.
llvm-svn: 28394
2006-05-19 19:07:54 +00:00
Evan Cheng feca91a516 These can be transformed into lea as well. Not that we use this feature
currently...

llvm-svn: 28393
2006-05-19 18:43:41 +00:00
Evan Cheng 7b8feb27c8 - Use exact-width integer types, e.g. int32_t, to avoid confusion.
- Fix a couple of minor bugs in i16immSExt8 and i16immZExt8.
- Added loadiPTR fragment used for indirect jumps and calls.

llvm-svn: 28392
2006-05-19 18:40:54 +00:00
Chris Lattner c0a65d5ac8 Use class tags instead of struct tags. The coding standards specify this
for public classes for improved win32 compatibility.

llvm-svn: 28391
2006-05-19 17:17:12 +00:00
Evan Cheng 1c8ef9832f Explicitly specify MOV32mi can only be used store 32-bit GV, etc.
llvm-svn: 28390
2006-05-19 07:30:36 +00:00
Evan Cheng 886e8f35aa Now that iPTR is a fully resolved type. We end up losing the type check for
patterns that look like this:

def : Pat<(i32 (X86Wrapper tconstpool  :$dst)), (MOV32ri tconstpool  :$dst)>;

InsertOneTypeCheck should copy the type from the resolved pattern to the
unresolved one as long as there types are different.

llvm-svn: 28389
2006-05-19 07:24:32 +00:00
Rafael Espindola b15597b59a implement movri
add a stub LowerFORMAL_ARGUMENTS

llvm-svn: 28388
2006-05-18 21:45:49 +00:00
Evan Cheng f3cbd7ef31 Added a Flags field to TargetOperandInfo. Currently the only flag is
M_LOOK_UP_PTR_REG_CLASS which allows the register class of the operand to be
resolved via a callback at runtime.

llvm-svn: 28387
2006-05-18 20:44:26 +00:00
Evan Cheng d8e2f6ebc1 lib/Target/Target.td
llvm-svn: 28386
2006-05-18 20:42:07 +00:00
Reid Spencer 8d88fc5ee4 Improve error output. Use the third parameter of the "catch" command to
capture the error output from the exec option. This generally will capture
the stderr messages generated by the tools. This information is then
printed if the test fails. This helps to recognize more quickly what the
error was. Otherwise, this information is lost.

llvm-svn: 28385
2006-05-18 19:42:16 +00:00
Chris Lattner 4cda95b32f add a note
llvm-svn: 28384
2006-05-18 18:26:13 +00:00
Chris Lattner f66e89721d add a note
llvm-svn: 28383
2006-05-18 17:38:16 +00:00
Andrew Lenharth b90055ef24 Fix a bogus gcc warning
llvm-svn: 28382
2006-05-18 17:29:34 +00:00
Evan Cheng 03524c63ff ImmMask should be 3 for a two-bit field; Compact X86II
llvm-svn: 28381
2006-05-18 06:27:15 +00:00
Owen Anderson 08aecf56e7 Fix some think-o's in my last commit. Thanks to Chris for pointing them out.
llvm-svn: 28380
2006-05-18 05:46:08 +00:00
Owen Anderson 9eb1a268c1 Change Module to use TargetData-compatible strings internally.
This is part of the on-going work on PR 761.

llvm-svn: 28379
2006-05-18 02:10:31 +00:00
Evan Cheng 305c49579c getCalleeSaveRegs and getCalleeSaveRegClasses are no long TableGen'd.
llvm-svn: 28378
2006-05-18 00:12:58 +00:00
Evan Cheng 297e1cb10a Remove CalleeSavedRegisters from class Target.
llvm-svn: 28377
2006-05-18 00:09:53 +00:00
Evan Cheng fe72285033 Don't generate getCalleeSaveReg and getCalleeSaveRegClasses anymore.
llvm-svn: 28376
2006-05-18 00:08:46 +00:00
Chris Lattner b848c2457e This test is buggy: printf is a varargs function. This fixes the test with
the PPC JIT

llvm-svn: 28375
2006-05-17 23:43:56 +00:00
Reid Spencer a22a5b382f Make some changes suggested by Chris:
1. Remove the LLVM_DO_NOT_BUILD feature (not needed any more)
2. Ensure that lib/VMCore gets built first. This needs to be done because
   VMCore now uses tblgen to generate the Intrinsics header which are
   needed in other libraries. In parallel builds, this can cause problems.

llvm-svn: 28374
2006-05-17 22:55:35 +00:00
Owen Anderson fc08d5a2a8 Fix a stupid bug when parsing TargetData strings.
llvm-svn: 28373
2006-05-17 21:56:02 +00:00
Chris Lattner b390c47b4b Use DIRS here instead of PARALLEL_DIRS, as VMCore has to be built before the
other dirs (for Intrinsics.gen).  :(

llvm-svn: 28372
2006-05-17 21:54:48 +00:00
Evan Cheng e59042d004 Use generic iPTR instead i32 to represent pointer type.
llvm-svn: 28371
2006-05-17 21:21:41 +00:00
Evan Cheng 7fa58c38c0 Another entry
llvm-svn: 28370
2006-05-17 21:20:51 +00:00
Reid Spencer 1df84eb2d1 Remove the llvm-db and bugpoint restrictions from the win32 platform so
they can be compiled with ming32. The use of fork(2) has been removed.

llvm-svn: 28369
2006-05-17 21:20:50 +00:00
Evan Cheng dcec882286 Remove PointerType from class Target
llvm-svn: 28368
2006-05-17 21:20:27 +00:00
Chris Lattner e4cb4768fa Declare that lowerinvoke doesn't interact with other lowering passes.
Patch written by Domagoj Babic!

llvm-svn: 28367
2006-05-17 21:05:27 +00:00
Evan Cheng 318a68e1ee Typo
llvm-svn: 28366
2006-05-17 20:55:51 +00:00
Chris Lattner 7949c2e8b2 Fix the result of the call to use a correct vbitconvert. There is no need to
use getPackedTypeBreakdown at all here.

llvm-svn: 28365
2006-05-17 20:49:36 +00:00
Chris Lattner 938155ca57 Correct a previous patch which broke CodeGen/PowerPC/vec_call.ll
llvm-svn: 28364
2006-05-17 20:43:21 +00:00
Evan Cheng f5ef47fe74 Remove PointerType from target definition. Use abstract type MVT::iPTR to
represent pointer type.

llvm-svn: 28363
2006-05-17 20:37:59 +00:00
Evan Cheng fabc691ce4 PR736 has already been fixed. Remove XFAIL marker.
llvm-svn: 28362
2006-05-17 20:21:44 +00:00
Evan Cheng 961ef9363d New test case for vector type argument pass by value.
llvm-svn: 28361
2006-05-17 20:20:04 +00:00
Andrew Lenharth 0f524f2050 Fix call_adj.ll
llvm-svn: 28360
2006-05-17 19:24:49 +00:00
Andrew Lenharth 446dbcb5e4 Added sanity check for obviously bogus immediates
llvm-svn: 28359
2006-05-17 19:24:31 +00:00
Andrew Lenharth f8e448f06f Added regression that breaks gcc4 build
llvm-svn: 28358
2006-05-17 19:23:31 +00:00
Evan Cheng 8c6b234ce8 Should pass by reference.
llvm-svn: 28357
2006-05-17 19:07:40 +00:00
Evan Cheng 00bce3f2f4 Another entry
llvm-svn: 28356
2006-05-17 19:05:31 +00:00
Chris Lattner 6353807fdc Add a note about a note
llvm-svn: 28355
2006-05-17 19:02:25 +00:00
Chris Lattner eb755fc1b3 Make PPC call lowering more aggressive, making the isel matching code simple
enough to be autogenerated.

llvm-svn: 28354
2006-05-17 19:00:46 +00:00
Evan Cheng 19aaaca293 Another typo. Pointed out by Nate Begeman.
llvm-svn: 28353
2006-05-17 18:22:14 +00:00
Evan Cheng 751cd7653d Fixed a LowerCallTo and LowerArguments bug. They were introducing illegal
VBIT_VECTOR nodes. There were some confusion about the semantics of
getPackedTypeBreakdown(). e.g. for <4 x f32> it returns 1 and v4f32, not 4,
and f32.

llvm-svn: 28352
2006-05-17 18:16:39 +00:00