For checking if a symbol >= value, we need to check if symbol == value || symbol
> value. When checking symbol > value and we know that symbol != value, the path
is infeasible only if value == maximum integer.
For checking if a symbol <= value, we need to check if symbol == value || symbol
< value. When checking symbol < value and we know that symbol != value, the path
is infeasible only if value == minimum integer.
Updated test case exercising this logic: we only prune paths if the values are
unsigned.
llvm-svn: 56354
'symbol operator-reverse int'. This patch is a combination of code from
Zhongxing Xu and myself (Zhongxing noticed this bug for the cases of
relational operators).
llvm-svn: 56351
Unfortunately this means removing one regression test
of GlobalsModRef because I couldn't work out how to
perform it without MarkModRef.
llvm-svn: 56342
can get the readnone/readonly attributes, and gives them it.
The plan is to remove markmodref (which did the same thing
by querying GlobalsModRef) and delete the analogous
functionality from GlobalsModRef.
llvm-svn: 56341
catches a fair number of common cases. Note that this currently
causes Fast-ISel to leave behind lots of dead instructions.
Those will be dealt with in subsequent commits.
llvm-svn: 56320
copy of the BURRList scheduler, but with several parts ripped
out, such as backtracking, online topological sort maintenance
(needed by backtracking), the priority queue, and Sethi-Ullman
number computation and maintenance (needed by the priority
queue). As a result of all this, it generates somewhat lower
quality code, but that's its tradeoff for running about 30%
faster than list-burr in -fast mode in many cases.
This is somewhat experimental. Moving forward, major pieces of
this can be refactored with pieces in common with
ScheduleDAGRRList.cpp.
llvm-svn: 56307
"Method accepting NSError** argument should have non-void return value to indicate that an error occurred."
Test case written, but the header needs to be delta-debugged reduced. Will commit shortly.
llvm-svn: 56297
- Enables use of ABIArgInfo::Expand when needed. This greatly
improves our x86 ABI compatibility.
- As the infrastructure for target specific ABI handling isn't built
yet, this change means ABI compatibility on other platforms is once
again broken in a different way than before.
- Upcoming: Figure out how to refactor ABI handling into
targets. More documentation.
llvm-svn: 56293
over having it in a register. And wait until after checking type
legality before requesting that the callee address be placed in a
register. Also, fix support for calls with void return type.
This speeds up fast-isel isel time by about 15% and reduces
instruction counts by about 3% overall on certain testcases. It also
changes many indirect calls to direct calls.
llvm-svn: 56292
with an earlyclobber operand elsewhere. Propagate
this bit and the earlyclobber bit through SDISel.
Change linear-scan RA not to allocate regs in a way
that conflicts with an earlyclobber. See also comments.
llvm-svn: 56290
Ensure that ABIArgInfo::Default implementation for aggregrate types
is sensible (StructRet on return and ByVal on arguments).
No functionality change.
llvm-svn: 56289
Ted Kremenek