summaryrefslogtreecommitdiffstats
path: root/llvm/utils
Commit message (Collapse)AuthorAgeFilesLines
* Fix FileCheck crash when fuzzy scanning starting at the end of the file.Daniel Dunbar2009-11-291-1/+1
| | | | llvm-svn: 90065
* Sketch TableGen disassembler emitter, based on patch by Sean Callanan.Daniel Dunbar2009-11-254-0/+66
| | | | llvm-svn: 89833
* Delete some dead and non-obvious code.Dan Gohman2009-11-242-55/+0
| | | | llvm-svn: 89729
* Remove ISD::DEBUG_LOC and ISD::DBG_LABEL, which are no longer used.Dan Gohman2009-11-231-1/+0
| | | | | | | | Note that "hasDotLocAndDotFile"-style debug info was already broken; people wanting this functionality should implement it in the AsmPrinter/DwarfWriter code. llvm-svn: 89711
* FileCheck, PR5239: Try to find the intended match on failures, but looking for aDaniel Dunbar2009-11-221-4/+60
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | good nearby fuzzy match. Frequently the input is nearly correct, and just showing the user the a nearby sensible match is enough to diagnose the problem. - The "fuzzyness" is pretty simple and arbitrary, but worked on my three test cases. If you encounter problems, or places you think FileCheck should have guessed but didn't, please add test cases to PR5239. For example, previously FileCheck would report this: -- t.cpp:21:55: error: expected string not found in input // CHECK: define void @_Z2f25f2_s1([[i64_i64_ty]] %a0) ^ <stdin>:19:30: note: scanning from here define void @_Z2f15f1_s1(%1) nounwind { ^ <stdin>:19:30: note: with variable "i64_i64_ty" equal to "%0" -- and now it also reports this: -- <stdin>:27:1: note: possible intended match here define void @_Z2f25f2_s1(%0) nounwind { ^ -- which makes it clear that the CHECK just has an extra ' %a0' in it, without having to check the input. llvm-svn: 89631
* FileCheck: When a string using variable references fails to match, printDaniel Dunbar2009-11-221-4/+50
| | | | | | | additional information about the current definitions of the variables used in the string. llvm-svn: 89628
* Allow '_' in FileCheck variable names, it is nice to have at least oneDaniel Dunbar2009-11-221-2/+3
| | | | | | | separate character. - Chris, OK? llvm-svn: 89626
* Fix whitespace.Bob Wilson2009-11-223-288/+267
| | | | llvm-svn: 89582
* Fix pr5470. Tablegen handles template arguments by temporarily setting theirBob Wilson2009-11-221-8/+20
| | | | | | | | | | | values, resolving references to them, and then removing the definitions. If a template argument is set to an undefined value, we need to resolve references to that argument to an explicit undefined value. The current code leaves the reference to the template argument as it is, which causes an assertion failure later when the definition of the template argument is removed. llvm-svn: 89581
* Fix some spelling in comments.Bob Wilson2009-11-212-4/+4
| | | | llvm-svn: 89566
* Avoid a redundant assertion.Bob Wilson2009-11-211-2/+1
| | | | llvm-svn: 89565
* TableGen/OptParser: When ordering options, make "sentinel" options appear beforeDaniel Dunbar2009-11-191-3/+10
| | | | | | everything else. llvm-svn: 89368
* De-bork CMake buildDouglas Gregor2009-11-181-0/+1
| | | | llvm-svn: 89272
* TableGen: Add initial backend for clang Driver's option parsing.Daniel Dunbar2009-11-183-10/+248
| | | | llvm-svn: 89245
* lit: Fix exclude dirs functionality.Daniel Dunbar2009-11-181-2/+3
| | | | llvm-svn: 89210
* Make X86-64 in the Large model always emit 64-bit calls.Jeffrey Yasskin2009-11-161-2/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The large code model is documented at http://www.x86-64.org/documentation/abi.pdf and says that calls should assume their target doesn't live within the 32-bit pc-relative offset that fits in the call instruction. To do this, we turn off the global-address->target-global-address conversion in X86TargetLowering::LowerCall(). The first attempt at this broke the lazy JIT because it can separate the movabs(imm->reg) from the actual call instruction. The lazy JIT receives the address of the movabs as a relocation and needs to record the return address from the call; and then when that call happens, it needs to patch the movabs with the newly-compiled target. We could thread the call instruction into the relocation and record the movabs<->call mapping explicitly, but that seems to require at least as much new complication in the code generator as this change. To fix this, we make lazy functions _always_ go through a call stub. You'd think we'd only have to force lazy calls through a stub on difficult platforms, but that turns out to break indirect calls through a function pointer. The right fix for that is to distinguish between calls and address-of operations on uncompiled functions, but that's complex enough to leave for someone else to do. Another attempt at this defined a new CALL64i pseudo-instruction, which expanded to a 2-instruction sequence in the assembly output and was special-cased in the X86CodeEmitter's emitInstruction() function. That broke indirect calls in the same way as above. This patch also removes a hack forcing Darwin to the small code model. Without far-call-stubs, the small code model requires things of the JITMemoryManager that the DefaultJITMemoryManager can't provide. Thanks to echristo for lots of testing! llvm-svn: 88984
* lit: Factor a new OneCommandPerFileTest out of SyntaxCheckTest.Daniel Dunbar2009-11-152-12/+42
| | | | | | | - Used for running a single fixed command on a directory of files, with the option of deriving a temporary input file from the test source. llvm-svn: 88844
* Remove duplicate implementation of excludes functionality, and support excludingDaniel Dunbar2009-11-151-14/+6
| | | | | | | | dirnames. Also, add support for the 'unsupported' config property. llvm-svn: 88838
* lit: Add --repeat=N option, for running each test N times.Daniel Dunbar2009-11-153-4/+27
| | | | | | - Currently just useful for timing, although it could be extended as one (bad) way to deal with flaky tests. llvm-svn: 88827
* Implement DISABLE_INLINE for MSVC. This required changing the position in allBenjamin Kramer2009-11-141-5/+4
| | | | | | forward declaration and patching tblgen to emit it right. Patch by Amine Khaldi! llvm-svn: 88798
* Added getSubRegIndex(A,B) that returns subreg index of A to B. Use it to ↵Evan Cheng2009-11-141-0/+18
| | | | | | replace broken code in VirtRegRewriter. llvm-svn: 88753
* Show command-line args and features passed into backend in debug output. ↵Sandeep Patel2009-11-111-0/+4
| | | | | | Approved by Evan Cheng. llvm-svn: 86797
* lit: Add ExampleTests, for testing lit and demonstrating test suite features.Daniel Dunbar2009-11-1026-0/+555
| | | | llvm-svn: 86654
* lit: Fix bug in --show-suites which accidentally override the list of tests.Daniel Dunbar2009-11-101-2/+2
| | | | llvm-svn: 86653
* Add a 'zkill' script, which is more-or-less a fancy (although not necessarilyDaniel Dunbar2009-11-081-0/+276
| | | | | | | | | very robust) version of killall. Because I like making shiny new wheels out of spare parts. For use by buildbots when people insist on making cc1 infinite loop. :) llvm-svn: 86484
* It is invalid to infer the value type from the result #0 of the nodeAnton Korobeynikov2009-11-081-1/+1
| | | | | | since the instruction might use the other result of different type. llvm-svn: 86462
* NNT: Remove DejaGNU test from NewNightlyTest reports, this aspect of testing isDaniel Dunbar2009-11-081-73/+6
| | | | | | handled by buildbots now. llvm-svn: 86454
* lit: Hardcode whence seek value, os.SEEK_END isn't always available.Daniel Dunbar2009-11-081-1/+1
| | | | llvm-svn: 86449
* lit: Warn when a test suite contains no tests.Daniel Dunbar2009-11-081-7/+12
| | | | llvm-svn: 86448
* lit: Drop require_and_and support.Daniel Dunbar2009-11-082-21/+6
| | | | llvm-svn: 86447
* lit: Workaround a Win32/subprocess bug when appending.Daniel Dunbar2009-11-081-0/+3
| | | | llvm-svn: 86437
* lit: Preserve the PATHEXT variable when running subcommands, this is ↵Daniel Dunbar2009-11-081-0/+1
| | | | | | important on Win32 llvm-svn: 86436
* clang++ points out that this is pointless.Chris Lattner2009-11-061-1/+1
| | | | llvm-svn: 86239
* NewNighlytTest: Fix timestamp format to actually make sense (it was missing ↵Daniel Dunbar2009-11-061-1/+2
| | | | | | the hour). llvm-svn: 86229
* NewNightlyTest: Add -noclean option, which doesn't run 'make clean' before ↵Daniel Dunbar2009-11-061-6/+8
| | | | | | | | building LLVM (for testing). Also, switch to always running 'make clean' in the test-suite directories. llvm-svn: 86228
* NewNightlyTest: Unbreak passing the build directory via a positional argument.Daniel Dunbar2009-11-061-7/+7
| | | | llvm-svn: 86227
* NewNightlyTest: Add -llvmgccdir as alternative to environment variable.Daniel Dunbar2009-11-061-8/+15
| | | | llvm-svn: 86226
* Don't print a redundant tab for inline asm, and do use the new printKill.Dan Gohman2009-11-061-2/+1
| | | | llvm-svn: 86206
* Teach lit's SyntaxCheckTest two new tricks:Douglas Gregor2009-11-051-1/+16
| | | | | | | - skip .svn directories - add a set of excluded filenames so we can easily skip tests llvm-svn: 86185
* lit: Add --param NAME=VALUE option, for test suite specific use (to communicateDaniel Dunbar2009-11-052-2/+18
| | | | | | arbitrary command line arguments to the test suite). llvm-svn: 86137
* Look for llvm-gcc under /Developer/usr/bin first.Evan Cheng2009-11-041-1/+1
| | | | llvm-svn: 86023
* Use llvm-gcc on newer Darwins.Bill Wendling2009-11-031-1/+1
| | | | llvm-svn: 85963
* mark some constant global const.Chris Lattner2009-11-031-7/+8
| | | | llvm-svn: 85910
* lit: Update Clang's test style to use XFAIL: and XTARGET: lines that matchDaniel Dunbar2009-11-031-24/+26
| | | | | | LLVM's tests. llvm-svn: 85882
* Support updating 'llvm_add_target' lists as well.Ted Kremenek2009-11-031-1/+1
| | | | llvm-svn: 85860
* Missing bit of universal built + hostedJim Grosbach2009-11-021-1/+1
| | | | llvm-svn: 85785
* Do not infer the target type for COPY_TO_REGCLASS from dest regclass, this ↵Anton Korobeynikov2009-11-021-1/+0
| | | | | | | won't work if it can contain several types. Require explicit result type for the node for now. This fixes PR5364. PS: It seems that blackfin usage of copy_to_regclass is completely bogus! llvm-svn: 85766
* The universal SDKROOT should only be assigned when hosted. Otherwise theJim Grosbach2009-10-311-0/+1
| | | | | | SDKROOT can refer to the target when we're building for the host. llvm-svn: 85672
* Allow cross target buildJim Grosbach2009-10-301-18/+67
| | | | llvm-svn: 85611
* Initial target-independent CodeGen support for BlockAddresses.Dan Gohman2009-10-301-0/+1
| | | | llvm-svn: 85556
OpenPOWER on IntegriCloud