Eliminate duplicate tracing options and use those defined by the trace passes.
Making the tracing step a module pass that runs before any llc-specific
passes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@3753 91177308-0d34-0410-b5e6-96231b3b80d8
This is required because trace code can insert calls to functions that
are (now) in an LLVM library, not a native library.
This change makes it easier to use the tracing code because the native
link options don't have to change, and it fits in a lot better with the
LLVM model.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@2675 91177308-0d34-0410-b5e6-96231b3b80d8
- Rename runOnMethod to runOnFunction
* Transform getAnalysisUsageInfo into getAnalysisUsage
- Method is now const
- It now takes one AnalysisUsage object to fill in instead of 3 vectors
to fill in
- Pass's now specify which other passes they _preserve_ not which ones
they modify (be conservative!)
- A pass can specify that it preserves all analyses (because it never
modifies the underlying program)
* s/Method/Function/g in other random places as well
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@2333 91177308-0d34-0410-b5e6-96231b3b80d8
written output file. This is important because crashing testcases often write
part of a file out, and the testing harness decides the file is up-to-date next
time the test is run.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@2303 91177308-0d34-0410-b5e6-96231b3b80d8
class. The Method class is obsolete (renamed) and all references to it
are being converted over to Function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@2144 91177308-0d34-0410-b5e6-96231b3b80d8
* Change -trace & -tracem options to use a 3 values enum option
* Change to use new style interface to passes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@1813 91177308-0d34-0410-b5e6-96231b3b80d8
out how to run a collection of passes optimially given their behaviors and
charactaristics.
Convert code to use it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@1507 91177308-0d34-0410-b5e6-96231b3b80d8
this is not a debugging option. But we can export it as assembly
instead of bytecode if -debugtrace is specified.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@889 91177308-0d34-0410-b5e6-96231b3b80d8
Move files from lib/CodeGen/TargetMachine to lib/Target
Move TargetData.h and TargetMachine.h to Target/{Data.h|Machine.h}
Prepare to split TargetMachine.h into several smaller files
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@566 91177308-0d34-0410-b5e6-96231b3b80d8
1. Clean up the TargetMachine structure. No more wierd pointers that have to
be cast around and taken care of by the target.
2. Instruction Scheduling now takes the schedinfo as an argument. The same
should be done with the instinfo, it just isn't now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@565 91177308-0d34-0410-b5e6-96231b3b80d8
1. The TargetMachine structure is free to decide the process a particular target uses to generate code.
2. All of the gooee details of the sparc backend are now localized in the lib/CodeGen/TargetMAchine/Sparc directory. The Sparc.h file that is globally visible is just a stub.
3. The Sparc.h file that esxists now will dissapear entirely someday when we have multiple backends chosen by a factory of some sort.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@559 91177308-0d34-0410-b5e6-96231b3b80d8
