[9fans] exiting processes in Plan9

[9fans] exiting processes in Plan9

[Roman V. Shaposhnick]

As you, probably, have seen from my previous post I had a situation
where a process waiting for a Rread message was several times notified
by writing "interrupt" to the notepg. For each note there was a separate
clunk message generated ( but never replied ). Since the only place in
Plan9 kernel where clunk messages are generated is port/devmnt.c:mntclunk().

I can safely assume that this was nothing else but a manifestation of
closing a file ( on which read() was hanging ) from pexit, right ?

If so, it means that I had a situation, where pexit was called several
times for the same process. The question here, is:

    Why did it happen ? Why in Plan9 process that is doing pexit can still
    receive messages, which result in even more calls to pexit ?

I could be wrong in my evaluation of the origin of all those clunks -- in
that case, please be so kind to explain what really happened.

Thanks,
Roman.

Re: [9fans] exiting processes in Plan9

[Roman V. Shaposhnick]

Please disregard those questions -- it was just a temporary glitch.

Thanks,
Roman.

> As you, probably, have seen from my previous post I had a situation
> where a process waiting for a Rread message was several times notified
> by writing "interrupt" to the notepg. For each note there was a separate
> clunk message generated ( but never replied ). Since the only place in
> Plan9 kernel where clunk messages are generated is port/devmnt.c:mntclunk().
>
> I can safely assume that this was nothing else but a manifestation of
> closing a file ( on which read() was hanging ) from pexit, right ?
>
> If so, it means that I had a situation, where pexit was called several
> times for the same process. The question here, is:
>
>     Why did it happen ? Why in Plan9 process that is doing pexit can still
>     receive messages, which result in even more calls to pexit ?
>
> I could be wrong in my evaluation of the origin of all those clunks -- in
> that case, please be so kind to explain what really happened.
>
> Thanks,
> Roman.