[llvm-bugs] [Bug 43975] New: Incorrect call site values for parameters produced by MOVZXi on AArch64
via llvm-bugs
llvm-bugs at lists.llvm.org
Tue Nov 12 04:11:08 PST 2019
https://bugs.llvm.org/show_bug.cgi?id=43975
Bug ID: 43975
Summary: Incorrect call site values for parameters produced by
MOVZXi on AArch64
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: DebugInfo
Assignee: unassignedbugs at nondot.org
Reporter: david.stenberg at ericsson.com
CC: jdevlieghere at apple.com, keith.walker at arm.com,
llvm-bugs at lists.llvm.org,
paul_robinson at playstation.sony.com
Reproduced on 636412bf31fc4285d63faeea67e4d71e44e3799c.
Compiling the following code:
#include <stdint.h>
extern void call1(int64_t);
extern void call2(int64_t);
extern void call3(int64_t);
extern void call4(int64_t);
void caller() {
call1(UINT64_C(1) << 0);
call2(UINT64_C(1) << 16);
call3(UINT64_C(1) << 32);
call4(UINT64_C(1) << 48);
}
using the following command:
clang movz.c --target=aarch64 -O1 -g -Xclang -femit-debug-entry-values -c
gives the following call site entries for the calls to call[1-4]:
$ llvm-dwarfdump movz.o | grep -E '(abstract_origin|call_site_value)'
DW_AT_abstract_origin (0x0000002a "call1")
DW_AT_GNU_call_site_value (DW_OP_lit1)
DW_AT_abstract_origin (0x0000003e "call2")
DW_AT_GNU_call_site_value (DW_OP_constu 0x10000)
DW_AT_abstract_origin (0x0000004b "call3")
DW_AT_GNU_call_site_value (DW_OP_lit1)
DW_AT_abstract_origin (0x00000058 "call4")
DW_AT_GNU_call_site_value (DW_OP_constu 0x10000)
As seen, the values passed to call3 and call4 are incorrect; 1 instead of
0x100000000 and 0x10000 instead of 0x1000000000000.
The issue seems to be that in AArch64InstrInfo::describeLoadedValue() a 32-bit
shift is used when calculating the immediate:
5762 int Immediate = MI.getOperand(1).getImm();
5763 int Shift = MI.getOperand(2).getImm();
5764 return ParamLoadedValue(MachineOperand::CreateImm(Immediate <<
Shift),
5765 nullptr);
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20191112/a231e1e9/attachment.html>
More information about the llvm-bugs
mailing list