Re: [PATCH] coroutine-sigaltstack: Keep SIGUSR2 handler up

[Laszlo Ersek]

On 01/25/21 11:57, Max Reitz wrote:
> On 23.01.21 01:41, Laszlo Ersek wrote:
>> On 01/22/21 22:26, Laszlo Ersek wrote:
>>
>>> I'm drifting towards an overhaul of coroutine-sigaltstack, based on my
>>> personal understanding of POSIX, but given that I can absolutely not
>>> *test* coroutine-sigaltstack on the platforms where it actually matters,
>>> an "overhaul" by me would be reckless.
>>>
>>> I didn't expect these skeletons when I first read Max's "Thread safety
>>> of coroutine-sigaltstack" email :/
>>>
>>> Max, after having worked on top of your patch for a few hours, I
>>> officially endorse your mutex approach. I can't encourage you or myself
>>> to touch this code, in good conscience. It's not that it's "bad"; it's
>>> inexplicable and (to me) untestable.
> 
> On one hand, that’s too bad; on the other perhaps it’s just for the
> better to get all of this out of our minds again (for now)... O:)
> 
>> I'm attaching a patch (based on 0e3246263068). I'm not convinced that I
>> should take responsibility for this, given the lack of testability on my
>> end. So I'm not posting it stand-alone even as an RFC. I've built it and
>> have booted one of my existent domains with it, but that's all.
> 
> FWIW, it looks good to me.  We should keep it in mind if in the future
> for some reason sigaltstack becomes more important, but for now I’m not
> too sad to abort any improvement efforts.

OK -- so do you plan to post your mutex approach stand-alone? (Sorry if
that's already been done and I've missed it.)

I actually feel somewhat safe regarding my patch, as long as it is
*only* executed on Linux hosts, and on other hosts that claim
conformance to POSIX.

But, of course, this whole thing has come up because of OSX... and I
don't have the first clue about OSX. (For example whether sigsuspend()
on OSX is prone to a spurious wakeup in response to SIGCONT.)

Thanks
Laszlo