[llvm-bugs] [Bug 34028] New: aarch64 .cfi_* using regno 33 gets: lib/MC/MCRegisterInfo.cpp:87: int llvm::MCRegisterInfo::getLLVMRegNum(unsigned int, bool) const: Assertion `I != M+Size && I->FromReg == RegNum && "Invalid RegNum"' failed.

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=34028

            Bug ID: 34028
           Summary: aarch64 .cfi_* using regno 33 gets:
                    lib/MC/MCRegisterInfo.cpp:87: int
                    llvm::MCRegisterInfo::getLLVMRegNum(unsigned int,
                    bool) const: Assertion `I != M+Size && I->FromReg ==
                    RegNum && "Invalid RegNum"' failed.
           Product: libraries
           Version: trunk
          Hardware: Other
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: AArch64
          Assignee: unassignedbugs at nondot.org
          Reporter: roland at hack.frob.com
                CC: llvm-bugs at lists.llvm.org, phosek at chromium.org

Created attachment 18891
  --> https://bugs.llvm.org/attachment.cgi?id=18891&action=edit
test case, preprocessed assembly source

clang  -cc1as -triple aarch64-fuchsia -filetype obj -main-file-name
exceptions.S -target-cpu cortex-a53 -target-feature +crc -target-feature
-fp-armv8 -target-feature -crypto -target-feature -neon -target-feature
+reserve-x18 -debug-info-kind=limited -dwarf-version=4 -mrelocation-model
static exceptions.s

yields:

lib/MC/MCRegisterInfo.cpp:87: int llvm::MCRegisterInfo::getLLVMRegNum(unsigned
int, bool) const: Assertion `I != M+Size && I->FromReg == RegNum && "Invalid
RegNum"' failed.

It's triggered by using DWARF register number 33 in .cfi_* directives.  I
couldn't get a simple test case using the same directives with 33 to exhibit
the crash.
33 is a valid DWARF register number for aarch64, cf
http://infocenter.arm.com/help/topic/com.arm.doc.ihi0057b/IHI0057B_aadwarf64.pdf
section 3.1; but it's not a normal register the assembler understands by name,
and the compiler would never generate CFI referring to it.
In hand-written assembly for kernel code, it's entirely valid.

-- 
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/20170802/bee92cca/attachment.html>