Re: nios2: is the ptrace ABI correct?

[Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>]

On 02/24/2015 04:25 PM, Arnd Bergmann wrote:
> On Tuesday 24 February 2015 12:28:41 Ezequiel Garcia wrote:
>>
>> Gah, no, you are right. I got confused.
>>
>> So it would be OK to avoid remove pt_regs from the uapi headers?
>> How does this affect the signal handling nios2 implementation?
>>
> 
> We have a number of architectures that don't provide this structure:
> 
> $ git grep -L pt_regs arch/*/include/uapi/asm/ptrace.h
> arch/frv/include/uapi/asm/ptrace.h
> arch/metag/include/uapi/asm/ptrace.h
> arch/openrisc/include/uapi/asm/ptrace.h
> arch/s390/include/uapi/asm/ptrace.h
> 
> so I'd assume it's ok in general not to have it. However, on
> nios2, struct pt_regs is embedded inside of struct sigcontext.
> If I read the code in arch/nios2/kernel/signal.c correctly,
> this is actually a bug and you should use a different structure
> there too, because pt_regs does not match the layout of the
> stack either. This means that the (rare) user programs that
> would know about the architecture to modify signal stacks
> are currently broken.
> 

/me is more confused now

In arch/nios2/include/asm/ucontext.h

struct ucontext {
        unsigned long     uc_flags;
        struct ucontext  *uc_link;
        stack_t           uc_stack;
        struct mcontext   uc_mcontext;
        sigset_t          uc_sigmask;
};

And in include/uapi/asm-generic/ucontext.h:

struct ucontext {
        unsigned long     uc_flags;
        struct ucontext  *uc_link;
        stack_t           uc_stack;
        struct sigcontext uc_mcontext;
        sigset_t          uc_sigmask;
};

Which one is the one that userspace sees? And why does the kernel has
two different structures?

Given this oddities, I'm wondering how troublesome would be to just
re-do this and break the ptrace and signal ABI. For instance, just
pushing pt_regs in PTRACE_GETREGSET would make things much clearer.

I guess Linus would burn me for even suggesting to breaking users... but
do we have any users at all? This arch has just been mainlined. Altera's
out-of-tree is already ABI-incompatible with mainline so that's not an
issue.

The only one using this ABI is gdb, which is easily fixed.
-- 
Ezequiel Garcia, VanguardiaSur
www.vanguardiasur.com.ar