Re: PQ_LAUNDRY: unexpected behaviour

On 5 Jan 2017, at 0:17, Matthew Macy wrote:

>  ---- On Mon, 02 Jan 2017 06:01:50 -0800 Jonathan Anderson 
> <jonathan_at_freebsd.org> wrote ----
>> Hi all,
>>
>> I'm seeing some unexpected PQ_LAUNDRY behaviour on something fairly 
>> close
>> to -CURRENT (drm-next-4.7 with an IFC on 26 Dec). Aside from the use 
>> of
>> not-quite-CURRENT, it's also very possible that I don't understand 
>> how the
>> laundry queue is supposed to work. Nonetheless, I thought I'd check 
>> whether
>> there is a tunable I should change, an issue with the laundry queue 
>> itself,
>> etc.
>>
>> After running X overnight (i915 can now run overnight on 
>> drm-next-4.7!), I
>> end up with a little over half of my system memory in the laundry 
>> queue and
>> a bunch of swap utilization. Even after closing X and shutting down 
>> lots of
>> services, I see the following in top:
>
>
> Please try the drm-next branch now. Up until very recently, the 
> shrinkers responsible for culling ttm/gem allocations were never run. 
> I've now implemented the shrinker, but it's driven from vm_lowmem, so 
> you'll probably still see what looks like a leak until you hit low 
> memory conditions. The shrinker should probably be run from 
> uma_timeout, but there isn't an eventhandler for that and I haven't 
> looked any further.
>
> -M

Hi,

I am now testing the `drm-next` branch, but I'm finding it crashes much 
more frequently (a la 
https://github.com/FreeBSDDesktop/freebsd-base-graphics/issues/96) than 
`drm-next-4.7`. While the 4.7 branch would sometimes only last a few 
minutes, it would sometimes run for a day or more. On `drm-next`, 
however, I think I'm yet to have 20 minutes of uptime. So, I haven't run 
into the memory shrinker yet because I haven't had enough uptime to use 
lots of memory. :) I will continue testing... any specific things I 
ought to be doing?


Jon
--
jonathan_at_FreeBSD.org