There are no clang or llvm* tests for EABI and no EABI buildbots.
*There is a single backend test that specifies EABI but it tests MIPS16.
Depends on D20678
Paths
| Differential D20679
[mips] Kill 'support' for untested EABI. ClosedPublic Authored by dsanders on May 26 2016, 7:10 AM.
Details Summary There are no clang or llvm* tests for EABI and no EABI buildbots. *There is a single backend test that specifies EABI but it tests MIPS16. Depends on D20678
Diff Detail
Event Timelinedsanders updated this object. Comment Actions LGTM Tools.cpp contains the following code. We need to remove eabi here too. if (CPUName.empty()) { // Deduce CPU name from ABI name. CPUName = llvm::StringSwitch<const char *>(ABIName) .Cases("o32", "eabi", DefMips32CPU) .Cases("n32", "n64", DefMips64CPU) .Default(""); } This revision is now accepted and ready to land.May 26 2016, 7:20 AM dsanders added a child revision: D20680: [mips] Slightly simplify MipsTargetInfo::setDataLayout(). NFC..May 26 2016, 7:22 AM dsanders edited edge metadata. Comment ActionsRemoved the 'eabi' from Tools.cpp this revealed three driver tests for passing Comment Actions Just to double-check: Do you still agree we should kill this now that the number of tests covering EABI-specific behaviour is non-zero? I think we should on the basis that we still have no backend tests and no buildbots. The only tests we have check that the driver passes -mabi=eabi to the assembler. Comment Actions
I agree. Closed by commit rL270998: [mips] Kill 'support' for untested EABI. (authored by dsanders). · Explain WhyMay 27 2016, 7:36 AM This revision was automatically updated to reflect the committed changes.
Revision Contents
Diff 58782 cfe/trunk/lib/Basic/Targets.cpp
cfe/trunk/lib/Driver/Tools.cpp
cfe/trunk/test/Driver/freebsd-mips-as.c
cfe/trunk/test/Driver/mips-abi.c
cfe/trunk/test/Driver/mips-as.c
|