summaryrefslogtreecommitdiffstats
path: root/llvm/utils
Commit message (Collapse)AuthorAgeFilesLines
* Input files should go before all other options.Mikhail Glushenkov2009-10-081-14/+24
| | | | | | Important, for example, when calling 'gcc a.o b.o c.o -lD -lE -lF'. llvm-svn: 83524
* Add instruction flags: hasExtraSrcRegAllocReq and hasExtraDefRegAllocReq. WhenEvan Cheng2009-10-013-0/+6
| | | | | | | | | | set, these flags indicate the instructions source / def operands have special register allocation requirement that are not captured in their register classes. Post-allocation passes (e.g. post-alloc scheduler) should not change their allocations. e.g. ARM::LDRD require the two definitions to be allocated even / odd register pair. llvm-svn: 83196
* B&I's buildit forces a PATH that omits /Developer. Temporarily addStuart Hastings2009-09-281-1/+3
| | | | | | /Developer/usr/bin to the PATH when looking for llvm-gcc. llvm-svn: 83028
* For B&I-style builds, tweak build_llvm script to prefer LLVM-G++ ifStuart Hastings2009-09-281-4/+10
| | | | | | available. Override by setting CC and CXX in the environment. llvm-svn: 83024
* Introduce the TargetInstrInfo::KILL machine instruction and get rid of theJakob Stoklund Olesen2009-09-284-9/+11
| | | | | | | | | | unused DECLARE instruction. KILL is not yet used anywhere, it will replace TargetInstrInfo::IMPLICIT_DEF in the places where IMPLICIT_DEF is just used to alter liveness of physical registers. llvm-svn: 83006
* Don't traverse into .svn directories.Daniel Dunbar2009-09-281-1/+1
| | | | llvm-svn: 82978
* Add a way to query the number of input files.Mikhail Glushenkov2009-09-281-5/+21
| | | | llvm-svn: 82957
* A bit prettier formatting.Mikhail Glushenkov2009-09-281-3/+3
| | | | llvm-svn: 82955
* implement and document support for filecheck variables. ThisChris Lattner2009-09-271-43/+172
| | | | | | | | | | allows matching and remembering a string and then matching and verifying that the string occurs later in the file. Change X86/xor.ll to use this in some cases where the test was checking for an arbitrary register allocation decision. llvm-svn: 82891
* remove support for "NoSub" from regex. It seems like a minor optimizationChris Lattner2009-09-261-1/+1
| | | | | | and makes the API more annoying. Add a Regex::getNumMatches() method. llvm-svn: 82877
* Provide proper masks for neon perfect shuffle table.Anton Korobeynikov2009-09-251-6/+6
| | | | | | I definitely need to read documentation better :( llvm-svn: 82813
* Improve MachineMemOperand handling.Dan Gohman2009-09-251-18/+11
| | | | | | | | | | | | | | | | | | | | | - Allocate MachineMemOperands and MachineMemOperand lists in MachineFunctions. This eliminates MachineInstr's std::list member and allows the data to be created by isel and live for the remainder of codegen, avoiding a lot of copying and unnecessary translation. This also shrinks MemSDNode. - Delete MemOperandSDNode. Introduce MachineSDNode which has dedicated fields for MachineMemOperands. - Change MemSDNode to have a MachineMemOperand member instead of its own fields with the same information. This introduces some redundancy, but it's more consistent with what MachineInstr will eventually want. - Ignore alignment when searching for redundant loads for CSE, but remember the greatest alignment. Target-specific code which previously used MemOperandSDNodes with generic SDNodes now use MemIntrinsicSDNodes, with opcodes in a designated range so that the SelectionDAG framework knows that MachineMemOperand information is available. llvm-svn: 82794
* Rename getTargetNode to getMachineNode, for consistency with theDan Gohman2009-09-251-3/+3
| | | | | | | | naming scheme used in SelectionDAG, where there are multiple kinds of "target" nodes, but "machine" nodes are nodes which represent a MachineInstr. llvm-svn: 82790
* reject attempts to use ()'s in patterns, these are reserved for filecheck.Chris Lattner2009-09-251-3/+15
| | | | llvm-svn: 82780
* reimplement the regex matching strategy by building a singleChris Lattner2009-09-251-88/+51
| | | | | | | | | | | | | | | | | regex and matching it instead of trying to match chunks at a time. Matching chunks at a time broke with check lines like CHECK: foo {{.*}}bar because the .* would eat the entire rest of the line and bar would never match. Now we just escape the fixed strings for the user, so that something like: CHECK: a() {{.*}}??? is matched as: CHECK: {{a\(\) .*\?\?\?}} transparently "under the covers". llvm-svn: 82779
* special case Patterns that are a single fixed string. This is a microscopicChris Lattner2009-09-251-0/+16
| | | | | | perf win and is needed for future changes. llvm-svn: 82777
* filecheck should not match a \n with a .Chris Lattner2009-09-251-3/+1
| | | | llvm-svn: 82758
* turn a std::pair into a real class.Chris Lattner2009-09-251-23/+34
| | | | llvm-svn: 82754
* add and document regex support for FileCheck. You can now do stuff like:Chris Lattner2009-09-241-11/+116
| | | | | | | | ; CHECK: movl {{%e[a-z][xi]}}, %eax or whatever. llvm-svn: 82717
* Use CanonicalizeInputFile to canonicalize the entire buffer containing theChris Lattner2009-09-241-57/+40
| | | | | | | CHECK strings, instead of canonicalizing the patterns directly. This allows Pattern to just contain a StringRef instead of std::string. llvm-svn: 82713
* change 'not' matching to use Pattern, move pattern parsing logic intoChris Lattner2009-09-241-32/+51
| | | | | | the Pattern class. llvm-svn: 82712
* refactor out the match string into its own Pattern class.Chris Lattner2009-09-241-33/+50
| | | | llvm-svn: 82711
* Make the end-of-itinerary mark explicit. Some cleanup.David Goodwin2009-09-241-5/+3
| | | | llvm-svn: 82709
* Add count/not tools as executables.Daniel Dunbar2009-09-247-1/+121
| | | | | | - Apparently, I'm willing to do incredibly stupid things in the name of portability. llvm-svn: 82685
* lit: When executing shell scripts internally, don't allow piped stderr on anyDaniel Dunbar2009-09-221-3/+21
| | | | | | | commands except the last one, instead redirect the stderr to a temporary file. This sidesteps a potential deadlocking issue. llvm-svn: 82538
* Add a magic LLVM_DISABLE_CRT_DEBUG environment variable which we check in ↵Daniel Dunbar2009-09-221-0/+1
| | | | | | | | RegisterHandler and use to disable the Win32 crash dialogs. These are a major blocker to any kind of automated testing. Also, tweak the 'lit' test runner to set this variable unconditionally. llvm-svn: 82537
* Actually use the arguments with the resolved executable path.Daniel Dunbar2009-09-221-1/+1
| | | | llvm-svn: 82527
* lit: Don't use close_fds=True on Windows.Daniel Dunbar2009-09-221-1/+5
| | | | llvm-svn: 82521
* lit: When executing commands internally, perform PATH resolution ourselves.Daniel Dunbar2009-09-221-2/+23
| | | | llvm-svn: 82520
* Use raw_ostream::indent instead of passing strings.Mikhail Glushenkov2009-09-211-145/+163
| | | | llvm-svn: 82456
* fix a FileCheck bug where:Chris Lattner2009-09-211-3/+3
| | | | | | | | | | ; CHECK: foo ; CHECK-NOT: foo ; CHECK: bar would always fail. llvm-svn: 82424
* rewrite CountNumNewlinesBetween to be in terms of StringRef.Chris Lattner2009-09-201-13/+15
| | | | llvm-svn: 82410
* implement and document support for CHECK-NOTChris Lattner2009-09-201-8/+50
| | | | llvm-svn: 82408
* rewrite FileCheck in terms of StringRef instead of manual pointer pairs.Chris Lattner2009-09-201-68/+37
| | | | llvm-svn: 82407
* Follow googletest logic for suppressing warnings in unittests/UnitTestMain.Daniel Dunbar2009-09-202-2/+4
| | | | llvm-svn: 82373
* convert some stuff to StringRef to avoid temporary std::strings.Chris Lattner2009-09-182-47/+47
| | | | llvm-svn: 82244
* add a comment.Chris Lattner2009-09-181-0/+5
| | | | llvm-svn: 82236
* lit: Add a custom test format for use in clang.Daniel Dunbar2009-09-162-1/+54
| | | | llvm-svn: 81987
* lit: When finding nested test suites, check first in the execpath in case thereDaniel Dunbar2009-09-151-13/+22
| | | | | | is a site configuration. llvm-svn: 81902
* Add a valgrind suppressions file for x86_64/linux/4.3.3.Daniel Dunbar2009-09-141-0/+23
| | | | llvm-svn: 81766
* lit: Give test formats control over test discovery.Daniel Dunbar2009-09-145-26/+100
| | | | llvm-svn: 81751
* slightly increase prettiness.Chris Lattner2009-09-141-2/+2
| | | | llvm-svn: 81742
* emit the register table as a massive string to avoid relocations.Chris Lattner2009-09-141-5/+16
| | | | llvm-svn: 81741
* move StringToOffsetTable out to its own header.Chris Lattner2009-09-142-56/+77
| | | | llvm-svn: 81740
* factor string table generation out to its own class. This changesChris Lattner2009-09-141-50/+73
| | | | | | the encoding of the AsmStrs table saving a byte or two. llvm-svn: 81739
* eliminate the TargetRegisterDesc::AsmName field, the asmprinters now have ↵Chris Lattner2009-09-131-6/+1
| | | | | | this table. llvm-svn: 81728
* Move unittest driver to utils/unittest/UnitTestMain.Daniel Dunbar2009-09-133-1/+35
| | | | | | | - This eliminates a race between building the unittests and linking the UnitTestMain library. llvm-svn: 81719
* the tblgen produced 'getRegisterName' method does not accessChris Lattner2009-09-131-1/+1
| | | | | | the object, make it static instead of const. llvm-svn: 81711
* make tblgen produce a function that returns the name for a physreg.Chris Lattner2009-09-132-8/+49
| | | | | | Nothing is using this info yet. llvm-svn: 81707
* tests: Add a %abs_tmp substitution which is guaranteed to be a full path.Daniel Dunbar2009-09-131-0/+2
| | | | llvm-svn: 81662
OpenPOWER on IntegriCloud