[RFC] [PATCH] Cgroup based OOM killer controller

Evgeniy Polyakov zbr at ioremap.net
Thu Jan 22 15:25:32 PST 2009


On Fri, Jan 23, 2009 at 01:53:04AM +0300, Evgeniy Polyakov (zbr at ioremap.net) wrote:
> > I'm quite certain you've spent more time writing emails to me than merging 
> > the patch and testing its possibilities, given your lack of understanding 
> > of its very basic concepts.
> 
> How cute :)
> Any other technical arguments of the same strength?

Its my turn now for the professional statements, let's start with this
technical side: you wanted to send a signal for the process to be killed,
but this will need
1. to allocate a signal, which will deadlock
2. no need to do this for sigill, but it will not work if process is in
unkillable state, while oom-killer clears iirc

Now to the oom-handler: if it will want to free some memory, it will
have to call a syscall with some pointer, which in turn may be in the
swapped area, so handler will be locked.

-- 
	Evgeniy Polyakov


More information about the Containers mailing list