[PATCH] D60039: Fix the bug of garbage collection of siod.

Mon Apr 1 19:47:34 PDT 2019

LuoYuanke added a comment.

In D60039#1450876 <https://reviews.llvm.org/D60039#1450876>, @craig.topper wrote:

> I believe this is saving rbp but mangling it while doing so in a reversible way.
>
>   0x00007ffff6f93fa3 <+3>:     mov    %rbp,%rax
>   0x00007ffff6f93fa6 <+6>:     xor    %fs:0x30,%rax
>   0x00007ffff6f93faf <+15>:    rol    $0x11,%rax
>   0x00007ffff6f93fb3 <+19>:    mov    %rax,0x8(%rdi)

Yes. It is demangled on __longjmp. But anyway in the jmp buffer the rbp value is mangled, so it is not object pointer any more.

  (gdb) disassemble __longjmp
  Dump of assembler code for function __longjmp:
     0x00007ffff6f94090 <+0>:     mov    0x30(%rdi),%r8
     0x00007ffff6f94094 <+4>:     mov    0x8(%rdi),%r9
     0x00007ffff6f94098 <+8>:     mov    0x38(%rdi),%rdx
     0x00007ffff6f9409c <+12>:    ror    $0x11,%r8
     0x00007ffff6f940a0 <+16>:    xor    %fs:0x30,%r8
     0x00007ffff6f940a9 <+25>:    ror    $0x11,%r9
     0x00007ffff6f940ad <+29>:    xor    %fs:0x30,%r9
     0x00007ffff6f940b6 <+38>:    ror    $0x11,%rdx
     0x00007ffff6f940ba <+42>:    xor    %fs:0x30,%rdx
     0x00007ffff6f940c3 <+51>:    nop
     0x00007ffff6f940c4 <+52>:    mov    (%rdi),%rbx
     0x00007ffff6f940c7 <+55>:    mov    0x10(%rdi),%r12
     0x00007ffff6f940cb <+59>:    mov    0x18(%rdi),%r13
     0x00007ffff6f940cf <+63>:    mov    0x20(%rdi),%r14
     0x00007ffff6f940d3 <+67>:    mov    0x28(%rdi),%r15
     0x00007ffff6f940d7 <+71>:    mov    %esi,%eax
     0x00007ffff6f940d9 <+73>:    mov    %r8,%rsp
     0x00007ffff6f940dc <+76>:    mov    %r9,%rbp
     0x00007ffff6f940df <+79>:    nop
     0x00007ffff6f940e0 <+80>:    jmpq   *%rdx
  End of assembler dump.

Repository:
  rT test-suite

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D60039/new/

https://reviews.llvm.org/D60039