RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-26 22:42:46 +02:00
Code Issues Projects Releases Wiki Activity
60,450 Commits 4 Branches 5 Tags 704 MiB
6976c543cd
Commit Graph

60450 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Jakob Stoklund Olesen
6976c543cd Enable a bunch more -regalloc=fast tests 
llvm-svn: 103531
 2010-05-12 00:11:24 +00:00
Jakob Stoklund Olesen
b3ca444697 Avoid scoping issues, fix buildbots 
llvm-svn: 103530
 2010-05-12 00:11:19 +00:00
Dan Gohman
03e407ed83 Add initial kill flag support to FastISel. 
llvm-svn: 103529
 2010-05-11 23:54:07 +00:00
Daniel Dunbar
5d81e1a0bf Make Clang happy. 
llvm-svn: 103528
 2010-05-11 23:53:13 +00:00
Daniel Dunbar
a033860d55 MC/Mach-O/x86_64: Fix PCrel adjustment for x86_64, which was using the fixup 
offset instead of the fixup address as intended.

llvm-svn: 103527
 2010-05-11 23:53:11 +00:00
Daniel Dunbar
7d7b6cf3b7 MC/Mach-O: As Kevin pointed out, 'Address' is really an offset -- rename to clarify. 
llvm-svn: 103526
 2010-05-11 23:53:07 +00:00
Daniel Dunbar
e43676c164 MC/Mach-O: Fix a crash on invalid. 
llvm-svn: 103525
 2010-05-11 23:53:05 +00:00
Jeffrey Yasskin
6768d483a0 Fix PR6951 by fixing Module leaks in bugpoint. 
llvm-svn: 103523
 2010-05-11 23:25:16 +00:00
Jakob Stoklund Olesen
e6eee8913f Store the Dirty bit in the LiveReg structure instead of a bit vector. 
llvm-svn: 103522
 2010-05-11 23:24:47 +00:00
Jakob Stoklund Olesen
063844f706 Keep track of the last place a live virtreg was used. 
This allows us to add accurate kill markers, something the scavenger likes.
Add some more tests from ARM that needed this.

llvm-svn: 103521
 2010-05-11 23:24:45 +00:00
Dan Gohman
2ace137eb9 Don't set kill flags on uses of CopyFromReg nodes. InstrEmitter doesn't 
create separate virtual registers for CopyFromReg values, so uses of
them don't necessarily kill the value.

llvm-svn: 103519
 2010-05-11 21:59:14 +00:00
Evan Cheng
e46afad99b Avoid breaking vstd when reg_sequence is not used. 
llvm-svn: 103513
 2010-05-11 21:07:36 +00:00
Jakob Stoklund Olesen
99d5d74fb0 One more -regalloc=fast test 
llvm-svn: 103509
 2010-05-11 20:51:07 +00:00
Jakob Stoklund Olesen
bea7fa3416 Silence warning 
llvm-svn: 103508
 2010-05-11 20:51:04 +00:00
Bill Wendling
4769f124cd Simplify this logic of creating a default Features object. 
llvm-svn: 103507
 2010-05-11 20:46:04 +00:00
Jakob Stoklund Olesen
e27902ac68 Simplify the tracking of used physregs to a bulk bitor followed by a transitive 
closure after allocating all blocks.

Add a few more test cases for -regalloc=fast.

llvm-svn: 103500
 2010-05-11 20:30:28 +00:00
Dan Gohman
1ebf23a692 Revert r103493, materializing functions in the regular PassManager. 
It works in simple cases, but it isn't a general solution.

llvm-svn: 103499
 2010-05-11 20:30:00 +00:00
Duncan Sands
7d5e4152c3 I got tired of VISIBILITY_HIDDEN colliding with the gcc enum. Rename it 
to LLVM_LIBRARY_VISIBILITY and introduce LLVM_GLOBAL_VISIBILITY, which is
the opposite, for future use by dragonegg.

llvm-svn: 103495
 2010-05-11 20:16:09 +00:00
Dan Gohman
4c44e690d7 Teach the regular pass manager how to materialize functions as needed. 
llvm-svn: 103493
 2010-05-11 19:58:43 +00:00
Dan Gohman
5f389dacab Remove the "WantsWholeFile" concept, as it's no longer needed. CBE 
and the others use the regular addPassesToEmitFile hook now, and
llc no longer needs a bunch of redundant code to handle the
whole-file case.

llvm-svn: 103492
 2010-05-11 19:57:55 +00:00
Dan Gohman
ca2df906ae Trim #includes and forward declarations. 
llvm-svn: 103489
 2010-05-11 19:11:43 +00:00
Jakob Stoklund Olesen
442e38c4de Mostly rewrite RegAllocFast. 
Sorry for the big change. The path leading up to this patch had some TableGen
changes that I didn't want to commit before I knew they were useful. They
weren't, and this version does not need them.

The fast register allocator now does no liveness calculations. Instead it relies
on kill flags provided by isel. (Currently those kill flags are also ignored due
to isel bugs). The allocation algorithm is supposed to work with any subset of
valid kill flags. More kill flags simply means fewer spills inserted.

Registers are allocated from a working set that contains no aliases. That means
most allocations can be done directly without expensive alias checks. When the
working set runs out of registers we do the full alias check to find new free
registers.

llvm-svn: 103488
 2010-05-11 18:54:45 +00:00
Dan Gohman
259488cf67 Fix a comment. 
llvm-svn: 103483
 2010-05-11 18:03:41 +00:00
Dan Gohman
fb6f4da0e0 Implement a bunch more TargetSelectionDAGInfo infrastructure. 
Move EmitTargetCodeForMemcpy, EmitTargetCodeForMemset, and
EmitTargetCodeForMemmove out of TargetLowering and into
SelectionDAGInfo to exercise this.

llvm-svn: 103481
 2010-05-11 17:31:57 +00:00
Daniel Dunbar
7670f69e42 MC/Mach-O x86_64: Switch to using fragment atom symbol. 
- This eliminates getAtomForAddress() (which was a linear search) and
   simplifies getAtom().
 - This also fixes some correctness problems where local labels at the same
   address as non-local labels could be assigned to the wrong atom.

llvm-svn: 103480
 2010-05-11 17:22:50 +00:00
Daniel Dunbar
2b45036b3b Test commit. 
llvm-svn: 103479
 2010-05-11 17:22:45 +00:00
Tanya Lattner
0b1b95f2c1 Test commit. 
llvm-svn: 103478
 2010-05-11 16:47:42 +00:00
Tanya Lattner
2cc9bc56c9 Test commit. 
llvm-svn: 103477
 2010-05-11 16:46:45 +00:00
Dan Gohman
eaacb8cb1f Remove the TargetLowering::getSubtarget() virtual function, which 
was unused. TargetMachine::getSubtarget() is used instead.

llvm-svn: 103474
 2010-05-11 16:21:03 +00:00
Kalle Raiskila
e302300b51 Make SPU backend not assert on jump tables. 
llvm-svn: 103466
 2010-05-11 11:00:02 +00:00
Evan Cheng
11130a0a22 Select @llvm.trap to the special B with 1111 condition (i.e. trap) instruction. 
llvm-svn: 103459
 2010-05-11 07:26:32 +00:00
Douglas Gregor
2f2491405a Fixes for Microsoft Visual Studio 2010, from Steven Watanabe! 
llvm-svn: 103457
 2010-05-11 06:17:44 +00:00
Bill Wendling
03d3cef073 Don't create a StringRef with a NULL value. 
llvm-svn: 103455
 2010-05-11 01:33:39 +00:00
Sean Callanan
e79322802a Extended the edis "IsBranch" property to call 
instructions as well.  Added support for checking
this to the llvm-mc tester as well.

llvm-svn: 103454
 2010-05-11 01:27:08 +00:00
Evan Cheng
5311f25b0b Model some vst3 and vst4 with reg_sequence. 
llvm-svn: 103453
 2010-05-11 01:19:40 +00:00
Bill Wendling
ed91169004 The getDefaultSubtargetFeatures method of SubtargetFeature did actually return a 
string of features for that target. However LTO was using that string to pass
into the "create target machine" stuff. That stuff needed the feature string to
be in a particular form. In particular, it needed the CPU specified first and
then the attributes. If there isn't a CPU specified, it required it to be blank
-- e.g., ",+altivec". Yuck.

Modify the getDefaultSubtargetFeatures method to be a non-static member
function. For all attributes for a specific subtarget, it will add them in like
normal. It will also take a CPU string so that it can satisfy this horrible
syntax.

llvm-svn: 103451
 2010-05-11 00:30:02 +00:00
Evan Cheng
97fc5247c8 It's not safe eliminate copies where src and dst have different sub-register indices. 
llvm-svn: 103450
 2010-05-11 00:20:03 +00:00
Evan Cheng
b4ac902aee Ensure REG_SEQUENCE source operands are unique. 
llvm-svn: 103449
 2010-05-11 00:04:31 +00:00
Daniel Dunbar
d2c14ce38d MC/Mach-O: Fix another mismatch with .weak_definition, we shouldn't use a 
scattered relocation entry with a .weak_definition.

llvm-svn: 103443
 2010-05-10 23:15:20 +00:00
Daniel Dunbar
c54539c121 MC/Mach-O: Factor out doesSymbolRequireExternRelocation. 
llvm-svn: 103442
 2010-05-10 23:15:13 +00:00
Evan Cheng
de3482f4af Indentation. 
llvm-svn: 103441
 2010-05-10 23:08:19 +00:00
Devang Patel
23f7323e7f Enable multiple Compile Units in one module. 
This means now 'llvm-ld a.bc b.bc' will preserve debug info appropriately.

llvm-svn: 103439
 2010-05-10 22:49:55 +00:00
Daniel Dunbar
00ffbb63da MC/Mach-O: Explicitly track atoms, as represented by their defining symbol, for each fragment (not yet used). 
llvm-svn: 103438
 2010-05-10 22:45:09 +00:00
Evan Cheng
91747dc5dc Model some vld3 instructions with REG_SEQUENCE. 
llvm-svn: 103437
 2010-05-10 21:26:24 +00:00
Evan Cheng
119fec2a67 It's not safe to propagate implicit_def that defines part of a register. 
llvm-svn: 103436
 2010-05-10 21:25:30 +00:00
Evan Cheng
040646a78e Clear RegSequences vector after eliminating REG_SEQUENCE instructions. 
llvm-svn: 103435
 2010-05-10 21:24:55 +00:00
Chris Lattner
a0f98775e9 this really is needed. :( 
llvm-svn: 103434
 2010-05-10 21:23:48 +00:00
Chris Lattner
bf56c070f3 just remove this, it isn't needed. 
llvm-svn: 103432
 2010-05-10 21:01:47 +00:00
Chris Lattner
622f6de228 simplify more. 
llvm-svn: 103431
 2010-05-10 20:59:18 +00:00
Chris Lattner
2ab3a7d574 Simplify by using startswith instead of substr. 
llvm-svn: 103430
 2010-05-10 20:58:42 +00:00