http://llvm.org/bugs/show_bug.cgi?id=17047
Bug ID: 17047
Summary: Assertion `LC != RTLIB::UNKNOWN_LIBCALL &&
"Unsupported FP_EXTEND!"' failed
Product: new-bugs
Version: trunk
Hardware: Other
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
Classification: Unclassified
Created attachment 11123
--> http://llvm.org/bugs/attachment.cgi?id=11123&action=edit
Original source code
I am using clang/llvm 3.3, and am running into an error when running llc on
bytecode generated from OpenCL C. I attach the source code below; this code
loads half (half precision) values from memory and converts them to floats.
I also attach the bytecode below.
I am running on a Raspberry Pi running Raspbian, which is an
armv6-unknown-linux-gnueabihf.
I run llc with the command
{{{
llc vload_half.cl.bc
}}}
which reports the error
{{{
llc: /home/pi/src/llvm-3.3/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp:374:
llvm::SDValue llvm::DAGTypeLegalizer::SoftenFloatRes_FP_EXTEND(llvm::SDNode*):
Assertion `LC != RTLIB::UNKNOWN_LIBCALL && "Unsupported FP_EXTEND!"' failed.
Stack dump:
0. Program arguments: llc vload_half.cl.bc
1. Running pass 'Function Pass Manager' on module 'vload_half.cl.bc'.
2. Running pass 'ARM Instruction Selection' on function
'@_Z10vload_halfjPKU3AS1Dh'
Aborted
}}}
The failing function is un-mangled via c++filt to "vload_half(unsigned int,
half AS1 const*)".
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
LLVMbugs mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
