221 Commits

Author SHA1 Message Date
Chad Rosier
46d8fc9d6b [ms-inline asm] Set the SemaCallback in the TargetAsmParser.
llvm-svn: 166310
2012-10-19 20:36:37 +00:00
Chad Rosier
db97454f8e [ms-inline asm] Set the MCTargetAsmParser as paring MS-style inline asm.
llvm-svn: 166293
2012-10-19 17:58:45 +00:00
Chad Rosier
b91b3f88ad [ms-inline asm] Add a size argument to the LookupInlineAsmIdentifier() callback,
which will be used by the asm matcher in the near future.

llvm-svn: 166221
2012-10-18 20:27:06 +00:00
Chad Rosier
791bbe298d [ms-inline asm] Have the LookupInlineAsmIdentifier() callback function return a
*NamedDecl.  In turn, build the expressions after we're finished parsing the
asm.  This avoids a crasher if the lookup fails.

llvm-svn: 166213
2012-10-18 19:39:37 +00:00
Chad Rosier
ce2bcbf8c0 [ms-inline asm] Move most of the AsmParsing logic in clang back into the MC
layer.  Use the new ParseMSInlineAsm() API and add an implementation of the
MCAsmParserSemaCallback interface.

llvm-svn: 166184
2012-10-18 15:49:40 +00:00
Chad Rosier
f8037a1fb6 [ms-inline asm] Move some logic around to simplify the interface between the
front-end and the AsmParser.  No functional change intended.

llvm-svn: 166063
2012-10-16 21:55:39 +00:00
Chad Rosier
4a0054f383 [ms-inline asm] Rework the front-end to use the API introduced in r165946.
-The front-end now builds a single assembly string and feeds it to the
 AsmParser.  The front-end iterates on a per statement basis by calling the
 ParseStatement() function.  Please note, the calling of ParseStatement() and
 and any notion of MCAsmParsedOperands will be sunk into the MC layer in the
 near future.  I plan to expose more basic APIs such as getClobbers, etc.

-The enumeration of the AsmString expressions have been reworked to use SMLocs
 rather than assembly Pieces, which were being parsed in the front-end.

-The test case, t8(), was modified due to r129223.  I'll have to find a way to
 work around things such as these.

Sorry for the large commit, but breaking this in multiple smaller commits proved
too irritating.

llvm-svn: 165957
2012-10-15 19:56:10 +00:00
Chad Rosier
ce17b0f5b8 Update for r165858.
llvm-svn: 165859
2012-10-13 00:26:22 +00:00
Chad Rosier
aac08d2c3d [ms-inline asm] Remove a bunch of parsing code from the front-end.
llvm-svn: 165851
2012-10-12 23:32:10 +00:00
Chad Rosier
7d1ec922ff Update for r165833.
llvm-svn: 165834
2012-10-12 22:53:52 +00:00
Chad Rosier
941faac455 [ms-inline asm] Remove a bunch of parsing code from the front-end, which will be
reimplemented in the AsmParser where it belongs.

llvm-svn: 165825
2012-10-12 21:37:00 +00:00
Benjamin Kramer
25a3470371 Remove unused variable.
llvm-svn: 165793
2012-10-12 14:01:58 +00:00
Chad Rosier
98b4164503 Remove dead code introduced in r165751.
llvm-svn: 165772
2012-10-12 00:02:48 +00:00
Chad Rosier
dc12c45058 [ms-inline asm] Remove a bunch of parsing code from the front-end, which will be
reimplemented in the AsmParser where it belongs.

llvm-svn: 165752
2012-10-11 21:44:41 +00:00
Chad Rosier
e66f42634f [ms-inline asm] Lookup the IdentifierInfo using the Idents table and remove the
now unused static helper function.

The test case needs to be remove temporarily until I can better filter memory
operands that aren't actual variable reference.

llvm-svn: 165751
2012-10-11 21:28:29 +00:00
Chad Rosier
0c1b62779e Revert 165058, per Jim request. This requires further discussion.
llvm-svn: 165070
2012-10-02 23:38:55 +00:00
Chad Rosier
09d7710e04 [ms-inline asm] Rewrite the symbol references as wildcard MCParsedAsmOperands.
A Sema lookup is used to determine the size of the variable, which is in turn
used during wildcard matching.

llvm-svn: 165058
2012-10-02 21:49:14 +00:00
Chad Rosier
04d5664295 [ms-inline asm] Enhance the isSimpleMSAsm() function to handle operands with pointer size
directives (e.g., dword ptr [eax]).

llvm-svn: 165031
2012-10-02 18:51:05 +00:00
Chad Rosier
0ede0986b9 No need to call the InitializeAll* functions.
llvm-svn: 165025
2012-10-02 16:41:12 +00:00
Chad Rosier
bb90c414c4 [ms-inline asm] Use the convertToMapAndConstraints() function in the front-end.
Rework the logic to account for the fact that we no longer create a MCInst.

llvm-svn: 164980
2012-10-01 23:45:59 +00:00
Bob Wilson
24719e3236 Examine the last, not the first, instruction from the MC matcher.
If an MS-style inline asm is matched to multiple instructions, e.g., with a
a WAIT-prefix, then we need to examine the operands of the last instruction
instruction, not the prefix instruction.

llvm-svn: 164608
2012-09-25 16:30:16 +00:00
Chad Rosier
93b3eb7793 Update for r164567.
llvm-svn: 164568
2012-09-24 22:58:50 +00:00
Bob Wilson
317be45091 Replace an assertion with an error for empty __asm statements.
llvm-svn: 164551
2012-09-24 19:57:59 +00:00
Bob Wilson
0fc64f0b11 Fix a comment typo and clean up formatting.
llvm-svn: 164550
2012-09-24 19:57:55 +00:00
Chad Rosier
90cad09ba5 [ms-inline asm] Use mnemonicIsValid() in a helpful assert.
llvm-svn: 164421
2012-09-21 22:22:39 +00:00
Chad Rosier
f83315fd4a Add a few FIXMEs.
llvm-svn: 163758
2012-09-13 00:20:37 +00:00
Chad Rosier
b261a50b40 [ms-inline asm] Handle the enumeration of input and output expressions in a
more robust way to address a few FIXMEs.

The initial implementation, r163342, built the IR asm string and then tried to
patch things on the fly without enough context.  Specifically, it didn't skip
mnemonics nor did it track with assembly instruction an expression was related
to.  The new implementation patches the operands and then builds the final
IR string.

llvm-svn: 163756
2012-09-13 00:06:55 +00:00
Chad Rosier
57cd91f683 [ms-inline asm] If we have a single asm operand that maps to multiple
MCOperands then iterate over all of then when computing clobbers, inputs and
outputs.

On x86 the 1-to-many mapping is a memory operand that includes a BaseReg(reg),
MemScale(imm), MemIndexReg(reg), an Expr(MCExpr or imm) and a MemSegReg(reg).
Invalid register (Op.getReg() == 0) are not considered when computing clobber.

llvm-svn: 163728
2012-09-12 18:14:25 +00:00
Chad Rosier
a4f52e16a5 [ms-inline asm] isDef/NumDefs in this context only refer to register definitions, not memory definitions.
llvm-svn: 163666
2012-09-11 23:53:48 +00:00
Chad Rosier
911d50fe4e [ms-inline asm] As of r163657 this check is unnecessary.
llvm-svn: 163665
2012-09-11 23:48:00 +00:00
Chad Rosier
ee94a2ebd5 [ms-inline asm] Don't consider tokens or immediates when computing clobbers, inputs and outputs.
llvm-svn: 163657
2012-09-11 23:13:15 +00:00
Chad Rosier
bd7fdadb90 [ms-inline asm] Add $$ before numeric constants in the IR.
llvm-svn: 163581
2012-09-11 00:51:28 +00:00
Chad Rosier
46b0a0adbe [ms-inline asm] Output empty asm statements for the directives we don't
handle.  Otherwise, the AsmParser will explode if we try to generate an
object files.

llvm-svn: 163345
2012-09-06 19:56:25 +00:00
Chad Rosier
c05d2c908d [ms-inline asm] The IR representation of inline assembly enumerates the input
and output expressions much like that in GNU-style inline assembly. Output
expressions are first.  Do this for MS-style inline asms.

llvm-svn: 163342
2012-09-06 19:35:00 +00:00
Chad Rosier
87b3d03cc2 Update for r163187.
llvm-svn: 163188
2012-09-05 01:16:06 +00:00
Chad Rosier
dea7964ddb [ms-inline asm] The MCInstrDesc only tracks register definitions. For now,
assume that if the 1st operands is an expression and the instruction mayStore,
then it is a memory definition.

llvm-svn: 163144
2012-09-04 16:39:38 +00:00
Chad Rosier
5c8822d3ff Fix 80-column violation.
llvm-svn: 163142
2012-09-04 15:58:44 +00:00
Chad Rosier
93380f63c9 [ms-inline asm] Use the new GetMCInstOperandNum() interface in the front-end.
llvm-svn: 163126
2012-09-03 20:40:52 +00:00
Chad Rosier
fd648a0320 Removed unused argument.
llvm-svn: 163105
2012-09-03 03:16:15 +00:00
Chad Rosier
30c0cb2f30 [ms-inline asm] Update for the MatchInstruction API change in r163101.
llvm-svn: 163102
2012-09-03 02:30:13 +00:00
Chad Rosier
6f0a9a2e2f Revert r162424 now that the scoping issue has been fixed in r162444.
llvm-svn: 162797
2012-08-28 22:08:58 +00:00
Chad Rosier
7ae6360758 Tidy up.
llvm-svn: 162786
2012-08-28 20:35:06 +00:00
Chad Rosier
3dd7bf2c86 [ms-inline asm] Add constraints to MSAsmStmt. We don't currently compute
the constraints, so in the interim we speculatively assume a 'r' constraint.
This is expected to work for most cases on x86.

llvm-svn: 162784
2012-08-28 20:28:20 +00:00
Chad Rosier
de70e0ef45 [ms-inline asm] As part of a larger refactoring, rename AsmStmt to GCCAsmStmt.
No functional change intended.

llvm-svn: 162632
2012-08-25 00:11:56 +00:00
Chad Rosier
175ea24e38 [ms-inline asm] Change the -fenable-experimental-ms-inline-asm option from a
CodeGen option to a LangOpt option.  In turn, hoist the guard into the parser 
so that we avoid the new (and fairly unstable) Sema/AST/CodeGen logic.  This
should restore the behavior of clang to that prior to r158325.
<rdar://problem/12163681>

llvm-svn: 162602
2012-08-24 21:42:51 +00:00
Benjamin Kramer
c5d00f61b6 Remove unnecessary calls to c_str.
llvm-svn: 162590
2012-08-24 20:43:21 +00:00
Chad Rosier
1e17cf9c22 [ms-inline asm] Generate the Input/Output expressions using
Sema::ActOnIdExpression().

llvm-svn: 162564
2012-08-24 16:38:58 +00:00
Chad Rosier
ab24b6e24e Fix 80-column violation.
llvm-svn: 162560
2012-08-24 15:51:10 +00:00
Chad Rosier
fe31e626ca [ms-inline asm] Add the basic APIs for Exprs to the MSAsmStmt AST. Next we need
generate the Input/Output expressions using Sema::ActOnIdExpression().

llvm-svn: 162509
2012-08-24 00:07:09 +00:00
Benjamin Kramer
cc4c49dd63 Now that ASTMultiPtr is nothing more than a array reference, make it a MutableArrayRef.
This required changing all get() calls to data() and using the simpler constructors.

llvm-svn: 162501
2012-08-23 23:38:35 +00:00