Displaying 4 results from an estimated 4 matches for "mipsjitinfo".
2012 Dec 28
1
[LLVMdev] [PATCH] LLVM build failure on OpenBSD/mips64
I'm just passing this patch on to the list as I know Brian is busy and I
had proded him to file the bug report.
http://llvm.org/bugs/show_bug.cgi?id=14721
"binutils on OpenBSD is quite old, which causes the build on mips64el to fail on
MipsJITInfo.cpp - it complains of illegal operands. The attachment patches it
so that it will build. All the patch does is replace the register aliases with
the actual register (i.e. changing $sp to $29). The registers are identical to
both mips and mips64, so it should not disrupt 32-bit Mips builds."
-...
2013 Nov 24
2
[LLVMdev] Strange i386 cross build error.
...g version 3.5 (trunk)
Target: i386-ellcc-linux
Thread model: posix
ecc: note: diagnostic msg: PLEASE submit a bug report to
http://llvm.org/bugs/ and include the crash backtrace, preprocessed
source, and associated run script.
llvm[3]: Compiling MipsJITInfo.cpp for Release+Asserts build
ecc: note: diagnostic msg:
********************
PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:
Preprocessed source(s) and associated run script(s) are
located at:
ecc: note: diagnostic msg: /tmp/A...
2013 Apr 01
0
[LLVMdev] proposed change to class BasicTTI and dual mode mips16/32 working
...t MipsFrameLowering> FrameLowering16;
+ OwningPtr<const MipsTargetLowering> TLInfo16;
+ OwningPtr<const MipsInstrInfo> InstrInfoSE;
+ OwningPtr<const MipsFrameLowering> FrameLoweringSE;
+ OwningPtr<const MipsTargetLowering> TLInfoSE;
MipsSelectionDAGInfo TSInfo;
MipsJITInfo JITInfo;
@@ -47,6 +55,8 @@ public:
virtual ~MipsTargetMachine() {}
+ virtual void addAnalysisPasses(PassManagerBase &PM);
+
virtual const MipsInstrInfo *getInstrInfo() const
{ return InstrInfo.get(); }
virtual const TargetFrameLowering *getFrameLowering() const
@@ -73,6 +83,1...
2013 Apr 01
3
[LLVMdev] proposed change to class BasicTTI and dual mode mips16/32 working
On Thu, Mar 28, 2013 at 12:22 PM, Nadav Rotem <nrotem at apple.com> wrote:
> IMHO the right way to handle target function attributes is to
> re-initialize the target machine and TTI for every function (if the
> attributes changed). Do you have another solution in mind ?
I don't really understand this.
TargetMachine and TTI may be quite expensive to initialize. Doing so for