Re: lwp resource limit

[Matt Thomas <matt%3am-software.com@localhost>]

On Jun 13, 2012, at 4:27 AM, Mindaugas Rasiukevicius wrote:

> Matt Thomas <matt%3am-software.com@localhost> wrote:
>> 
>> On Jun 12, 2012, at 8:26 PM, Christos Zoulas wrote:
>> 
>>> On Jun 13,  2:15am, yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi) 
>>> wrote:
>>> -- Subject: Re: lwp resource limit
>>> 
>>> | > Can you explain what you mean by api incompatible? It is used in
>>> | > getrlimit the same way we use it, how can it be different?
>>> | 
>>> | it's semantics (what's counted for) is different, isn't it?
>>> 
>>> I don't know, the pages I looked don't offer details.
>> 
>> The nice thing about our semantics is you can disable threaded programs
>> by setting lwp count to 0 but still allow non-threaded programs to run.
> 
> Yes, but that is not the point.  I think what yamt means is that we should
> check whether our approach (i.e. the way accounting is done) is compatible
> with QNX and if not, then either rename the limit or make it compatible.
> 
> Would somebody be able to check?  Sean?

Looking at QNX docs, it seems that exceeding 
RLIMIT_NTHR causes attempts to create threads 
(e.g. by calling pthread_create()) fails.

Interesting that RLIMIT_NTHR seems to be per-process.