Re: [PATCH v2 1/2] mm/vmalloc: Use batched page requests in bulk-allocator

[Uladzislau Rezki]

On Wed, Jul 07, 2021 at 10:42:29AM +0200, Michal Hocko wrote:
> On Tue 06-07-21 13:26:53, Andrew Morton wrote:
> > On Mon,  5 Jul 2021 19:05:36 +0200 "Uladzislau Rezki (Sony)" <urezki@xxxxxxxxx> wrote:
> > 
> > > In case of simultaneous vmalloc allocations, for example it is 1GB and
> > > 12 CPUs my system is able to hit "BUG: soft lockup" for !CONFIG_PREEMPT
> > > kernel.
> > > 
> > > <snip>
> > > ...
> > >
> > > are obtained, i.e. do batched page requests adding cond_resched() meanwhile
> > > to reschedule. Batched value is hard-coded and is 100 pages per call.
> > > 
> > > Signed-off-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
> > 
> > Can we please have a Fixes: for this?
> 
> Is this a fix for any actual real life problem? I mean allocating 1GB of
> vmalloc space back and forth sounds like a stretch to me.
>  
It is not a real scenario. I simulated it by the stress-suite tests. So the
Fixes tag is not needed, IMHO.

> > Is this fix important enough for 4.14-rcx?  I think so...
> 
> I do not think so. This is an improvement so that vmalloc behaves more
> sanely for those abusers...
>
A bulk-allocator has recently been introduced, so 4.x does not have it,
i.e. this change is not applicable and 4.x kernel does not suffer from
it.

--
Vlad Rezki