[PATCH] 8006508 : Wrong frame constructor is called in os_linux_x86.cpp

[Coleen Phillimore]

It makes sense to call the default constructor in this case.   But there 
are other cases just like this.   Do they need to be changed also?

./os_cpu/bsd_x86/vm/os_bsd_x86.cpp:    return frame(NULL, NULL, NULL);
./os_cpu/linux_sparc/vm/os_linux_sparc.cpp:    return frame(NULL, 
frame::unpatchable, NULL);
./os_cpu/linux_x86/vm/os_linux_x86.cpp:    return frame(NULL, NULL, NULL);
./os_cpu/solaris_sparc/vm/os_solaris_sparc.cpp:    return frame(NULL, 
NULL, NULL);
./os_cpu/windows_x86/vm/os_windows_x86.cpp:  if (func == NULL) return 
frame(NULL, NULL, NULL);
./os_cpu/windows_x86/vm/os_windows_x86.cpp:    return frame(NULL, NULL, 
NULL);

Thanks,
Coleen

On 1/17/2013 1:32 PM, Jeremy Manson wrote:
> As discussed with David last month.  The frame constructor that is 
> currently getting called in this case fails if NULL is passed as a PC. 
>  The fix is to call a constructor that does not expect a PC.
>
> Apologies for the lack of a reproducible test case, but it's basically 
> impossible to reproduce without a very peculiar and platform-dependent 
> compilation strategy.
>
> # HG changeset patch
> # User jeremymanson
> # Date 1358447322 28800
> # Node ID b155e0a0e6f030aea5fbdf8cd341efcb2f097a76
> # Parent  1a3e54283c54aaa8b3437813e8507fbdc966e5b6
> 8006508 : Wrong frame constructor is called in os_linux_x86.cpp
>
> diff --git a/src/os_cpu/linux_x86/vm/os_linux_x86.cpp 
> b/src/os_cpu/linux_x86/vm/os_linux_x86.cpp
> --- a/src/os_cpu/linux_x86/vm/os_linux_x86.cpp
> +++ b/src/os_cpu/linux_x86/vm/os_linux_x86.cpp
> @@ -189,7 +189,7 @@
>                  CAST_FROM_FN_PTR(address, os::current_frame));
>    if (os::is_first_C_frame(&myframe)) {
>      // stack is not walkable
> -    return frame(NULL, NULL, NULL);
> +    return frame();
>    } else {
>      return os::get_sender_for_C_frame(&myframe);
>    }
>