Re: uvm: shrinking amap kmem consumption

2016-04-22 Thread Stefan Kempf
Stefan Kempf wrote: > Next try. A problem with the last version was that amaps for shared > mappings still would grow pretty large: establishing a shared mapping > causes uvm to immediately allocate an amap for the whole range. I messed up the last diff. tb@ found and fixed bug in the libkvm bits

Re: uvm: shrinking amap kmem consumption

2016-04-22 Thread Stefan Kempf
Next try. A problem with the last version was that amaps for shared mappings still would grow pretty large: establishing a shared mapping causes uvm to immediately allocate an amap for the whole range. That's a problem when changing vmm to use shared memory to allow vmd(8) to access the memory of

RELAYD_ANCHOR as a relayd.conf option

2016-04-22 Thread Philipp Buehler
Moin, while trying to push rdomain setups a bit further, I noticed that relayd is using a fixed anchor. For the pre-rdomain days this was sufficient, but nowadays that might look a bit different. Some dance with 'match pftag', carefully crafted (read:unique) rdr-subanchor-names can make the

Re: synaptics: two-finger scrolling and coasting

2016-04-22 Thread Alexandr Shadchin
On Fri, Apr 22, 2016 at 05:24:29AM +0200, Sebastien Marie wrote: > On Thu, Apr 21, 2016 at 09:36:33PM +0200, Ulf Brosziewski wrote: > > Ping? > > > > It isn't a severe bug and it doesn't concern a spectacular feature, > > but shouldn't we fix it? > > > > I run it too since 1 week I think, and I