Date: Wed, 08 May 2019 06:54:04 +1000
From: matthew green <m...@eterna.com.au>
Message-ID: <7543.1557262...@splode.eterna.com.au>
| > Log Message:
| > Use the max limit (aka maxfiles or the moral equivalent of OPEN_MAX) which
| > makes poll(2) align with the Posix documentation (which allows EINVAL if
| > nfds > OPEN_MAX). From: Anthony Mallet
|
| this seems to still miss the case kre was talking about:
It does, but there also needs to be a limit somewhere (unless we add
more complicated code - which we could do for unusualk cases),
In practice soming near this limit is very rare, and as long as we
have some documented (and ideally posix approved) behaviour, it should
be OK.
My only issue currently is that rlimit.max can be reduced as well,
so I think we should have a value that we always permit, regardless
of that limit (ie: max(rlimit.max, 1000) or something).
kre
