Re: [PATCH v3] kernel/signal.c: fix BUG_ON with SIG128 (MIPS)

[Oleg Nesterov <oleg@redhat.com>]

On 06/21, David Daney wrote:
>
> On 06/21/2013 01:22 PM, Oleg Nesterov wrote:
>> On 06/21, David Daney wrote:
>>>
>>> On 06/21/2013 06:39 AM, James Hogan wrote:
>>>> Therefore add sig_to_exitcode() and exitcode_to_sig() functions which
>>>> map signal numbers > 126 to exit code 126 and puts the remainder (i.e.
>>>> sig - 126) in higher bits. This allows WIFSIGNALED() to return true for
>>>> both SIG127 and SIG128, and allows WTERMSIG to be later updated to read
>>>> the correct signal number for SIG127 and SIG128.
>>>
>>> I really hate this approach.
>>>
>>> Can we just change the ABI to reduce the number of signals so that all
>>> the standard C library wait related macros don't have to be changed?
>>>
>>> Think about it, any user space program using signal numbers 127 and 128
>>> doesn't work correctly as things exist today, so removing those two will
>>> be no great loss.
>>
>> Oh, I agree.
>>
>> Besides, this changes ABI anyway. And if we change it we can do this in
>> a more clean way, afaics. MIPS should simply use 2 bytes in exit_code for
>> signal number.
>
> Wouldn't that break *all* existing programs that use signals?  Perhaps I
> misunderstand what you are suggesting.

Of course this will break the userspace more than the original patch,
that is why I said "And yes, this means that WIFSIGNALED/etc should
be updated".

> I am proposing that we just reduce the number of usable signals such
> that existing libc status checking macros/functions don't change in any
> way.

And I fully agree! Absolutely, sorry for confusion.


What I tried to say, _if_ we change the ABI instead, lets make this
change sane.

To me this hack is not sane. And btw, the patch doesn't look complete.
Say, wait_task_zombie() should do exitcode_to_sig() for ->si_status.

Oleg.