Commit Graph

768 Commits

Author SHA1 Message Date
Daniel Dunbar e46506eaea Pull CodeGenFunction::EmitVAArg into target specific ABIInfo classes.
- Missed this file.

llvm-svn: 64238
2009-02-10 21:44:36 +00:00
Daniel Dunbar 2d0746fb97 Pull CodeGenFunction::EmitVAArg into target specific ABIInfo classes.
llvm-svn: 64235
2009-02-10 20:44:09 +00:00
Fariborz Jahanian 4e7ae06b00 Generate ir for ivar offset. This will pass
type-nsobject-attribute.m in the dejagnu test suite
in the nonfragile abi mode.

llvm-svn: 64233
2009-02-10 20:21:06 +00:00
Fariborz Jahanian 21fc74c15a Some refactoring of Ivar offset code gen.
in preparation for nonfragile ivar offset work.

llvm-svn: 64225
2009-02-10 19:02:04 +00:00
Daniel Dunbar 9403cd6d85 Tweak x86-64 ABI to allow reuse for vararg handling.
llvm-svn: 64221
2009-02-10 17:06:09 +00:00
Anders Carlsson 33c1b6528f Remove the last remnants of the Obj-C EH stack code.
llvm-svn: 64205
2009-02-10 06:07:49 +00:00
Anders Carlsson 33747b6c41 Start removing the old Obj-C EH stack now that the cleanup stack is used instead.
llvm-svn: 64203
2009-02-10 05:52:02 +00:00
Sanjiv Gupta e99ad00fd2 Function parameters for PIC16 are like local variables. So use the keyword ".auto." to mangle their names. The working of PIC16AsmPrinter relies on that keyword currently.
llvm-svn: 64198
2009-02-10 04:17:25 +00:00
Daniel Dunbar 0007961241 Support va_arg on _Complex.
gcc compat test suite results (Darwin x86-32 & -64):
--
# of expected passes		1110
# of unexpected failures	74
# of unresolved testcases	168
# of unsupported tests		2

llvm-svn: 64197
2009-02-10 03:03:30 +00:00
Daniel Dunbar 9bfb4de38b ABI: Correctly handle load/store of values which have a different LLVM
memory representation (e.g., bool).
 - This upgrades (downgrades) MultiSource/Applications/ClamAV/clamscan
   to a miscompile and fixes
   SingleSource/UnitTests/2003-05-31-CastToBool.

llvm-svn: 64194
2009-02-10 01:51:39 +00:00
Daniel Dunbar 1d425460d5 Add util Emit{LoadOf,StoreTo}Scalar methods to encapsulate conversion
from LLVM memory type to/from LLVM temporary type.
 - No intended functionality change.

llvm-svn: 64191
2009-02-10 00:57:50 +00:00
Daniel Dunbar d5f1f55e28 Make sure to initialize local variables, even if they were ignored by
ABI.

llvm-svn: 64187
2009-02-10 00:06:49 +00:00
Anders Carlsson 15949b3a32 Use the new cleanup infrastructure for VLAs. The next iteration of patches will remove the old Obj-C EH cleanup code.
llvm-svn: 64161
2009-02-09 20:41:50 +00:00
Anders Carlsson bfee7e921b Use the new cleanup infrastructure for @try/@finally
llvm-svn: 64160
2009-02-09 20:38:58 +00:00
Anders Carlsson 1ac6282c5b Replace a bunch of EmitBranch calls with EmitBranchThroughCleanup. No functionality change (yet).
llvm-svn: 64159
2009-02-09 20:31:03 +00:00
Anders Carlsson ff0bb6ce62 Save and restore the DidCallStackSave variable
llvm-svn: 64157
2009-02-09 20:23:40 +00:00
Anders Carlsson f4478e94b8 Add DidCallStackSave variable to CodeGenFunction.
llvm-svn: 64156
2009-02-09 20:20:56 +00:00
Douglas Gregor 8bf4205c70 Start processing template-ids as types when the template-name refers
to a class template. For example, the template-id 'vector<int>' now
has a nice, sugary type in the type system. What we can do now:

  - Parse template-ids like 'vector<int>' (where 'vector' names a
    class template) and form proper types for them in the type system.
  - Parse icky template-ids like 'A<5>' and 'A<(5 > 0)>' properly,
    using (sadly) a bool in the parser to tell it whether '>' should
    be treated as an operator or not.

This is a baby-step, with major problems and limitations:
  - There are currently two ways that we handle template arguments
  (whether they are types or expressions). These will be merged, and,
  most likely, TemplateArg will disappear.
  - We don't have any notion of the declaration of class template
  specializations or of template instantiations, so all template-ids
  are fancy names for 'int' :)

llvm-svn: 64153
2009-02-09 18:46:07 +00:00
Mike Stump fc49682f20 Add some more documentation. Also reflowed comments to 80 col.
llvm-svn: 64105
2009-02-08 23:14:22 +00:00
Anders Carlsson 9c964ac272 Reuse case destinations.
llvm-svn: 64100
2009-02-08 22:46:50 +00:00
Anders Carlsson f57b9eef62 Always check if we can remove branch fixups, even if the cleanup stack is empty.
llvm-svn: 64099
2009-02-08 22:45:15 +00:00
Anders Carlsson dcb149cbef Add a simplified EmitJumpThroughFinally and use it in CGObjC in preparation of making it use the cleanup stack.
llvm-svn: 64098
2009-02-08 22:25:30 +00:00
Anders Carlsson 76180ea456 Misc fixes to the cleanup stack code.
llvm-svn: 64096
2009-02-08 22:13:37 +00:00
Mike Stump a6dbd7b25a When we're at the stack depth we want, there isn't anything to do.
llvm-svn: 64095
2009-02-08 22:00:53 +00:00
Mike Stump 284d177c7f Wire up break and continue processing to the new stack depth adjuster.
If people could beat on it and let me know if there are any new
semantics required by newer language standards or DRs or any little
details I goofed on, I'd be happy to fix any issues found.

llvm-svn: 64079
2009-02-08 09:22:19 +00:00
Anders Carlsson 66c384ac2e More cleanup stack work, PopCleanupBlock now returns a struct with the switch block and end block.
llvm-svn: 64072
2009-02-08 07:46:24 +00:00
Anders Carlsson ae91d9b140 Split some functions up
llvm-svn: 64069
2009-02-08 03:55:35 +00:00
Anders Carlsson a586ad7f85 CleanupScope needs to push the cleanup block in its destructor
llvm-svn: 64068
2009-02-08 03:22:36 +00:00
Anders Carlsson 3c21dd5a80 Implement support for branch fixups.
llvm-svn: 64064
2009-02-08 01:23:05 +00:00
Anders Carlsson 7d70fd27a4 More cleanup stack work.
llvm-svn: 64059
2009-02-08 00:50:42 +00:00
Anders Carlsson fbfb5e6530 When emitting blocks, keep track of which cleanup scope they have. Minor fixes and cleanup.
llvm-svn: 64053
2009-02-08 00:16:35 +00:00
Anders Carlsson cadb9a6a34 Emit a cleanup block for the cleanup attribute
llvm-svn: 64052
2009-02-07 23:51:38 +00:00
Anders Carlsson be0f76a712 Add support for emitting cleanup blocks. Make EmitCompoundStatement emit cleanup blocks if necessary
llvm-svn: 64051
2009-02-07 23:50:39 +00:00
Anders Carlsson 2cf8c44e43 Add a simple RAII object, to be used for pushing a cleanup entry and make the insertion point be the cleanup block.
llvm-svn: 64048
2009-02-07 23:30:41 +00:00
Mike Stump 0509d9635b Ensure we track all the stack depths for all break and continue points
correctly.  This should lay the ground work to throw the big switch
and start code gening break and continue in the presense of vlas.

llvm-svn: 64046
2009-02-07 23:02:10 +00:00
Anders Carlsson 15cb75a20c Add plumbing for the cleanup stack.
llvm-svn: 64043
2009-02-07 22:53:43 +00:00
Anders Carlsson bf8a1be33c Split the exception object out into its own stack.
llvm-svn: 64032
2009-02-07 21:37:21 +00:00
Anders Carlsson da0e4560a1 Simplify the Objective-C exception handling.
llvm-svn: 64031
2009-02-07 21:26:04 +00:00
Mike Stump 56d2a15829 Format for 80-cols.
llvm-svn: 64030
2009-02-07 20:14:12 +00:00
Mike Stump 6d8a617474 Ensure that we don't miscodegen if vlas creap into the top of the for.
This will allow us to generate break and continue even if vlas are
involved without worry that we'll silently generate bad code.

llvm-svn: 64028
2009-02-07 20:09:00 +00:00
Mike Stump dc0d6be7b9 Arrange to have the correct StackDepth for while statements.
llvm-svn: 64021
2009-02-07 18:08:12 +00:00
Mike Stump 1f8be1b9ac Fit into 80-col.
llvm-svn: 64020
2009-02-07 17:18:33 +00:00
Mike Stump 1d91dd98ad Fixup goto codegen in and around VLAs.
llvm-svn: 64014
2009-02-07 12:52:26 +00:00
Daniel Dunbar ee9e4c274b Set load/store alignment when doing ABI coercions.
- Currently, this is producing poor code, but we prefer correctness
   to performance for now. Eventually we should be able to generally
   avoid having to set the alignment when we control the alignment of
   the alloca.

 - This knocks out 33/1000 failures on my single argument ABI tests,
   down to 22/1000 and 18 of these appear to be gcc bugs. Woot.

llvm-svn: 64001
2009-02-07 02:46:03 +00:00
Daniel Dunbar d404862cf9 When making dummy file entries, the directory name should also be
non-empty.

llvm-svn: 63986
2009-02-07 00:40:41 +00:00
Fariborz Jahanian c76e741622 Fixed an objc2 nonfragile-abi code gen bug.
Now we can say 'hello world' objective-c style
in the nonfragile abi.

llvm-svn: 63981
2009-02-06 23:46:26 +00:00
Fariborz Jahanian 6b7cd6e5f4 objc2's nonfragile abi API for messages sent to 'super'.
llvm-svn: 63959
2009-02-06 20:09:23 +00:00
Daniel Dunbar fe2fb0af5a Use 'compile' instead of 'codegen' when reporting error to user.
llvm-svn: 63952
2009-02-06 19:18:03 +00:00
Fariborz Jahanian 33f66e640a objc2's nonfragile-abi - API selection for when receiver is a class
llvm-svn: 63881
2009-02-05 20:41:40 +00:00
Anders Carlsson 31f8649f83 Follow Eli's advice and store the VLA size with the native size_t type. Fixes PR3491.
llvm-svn: 63879
2009-02-05 19:43:10 +00:00