Yes, they put that functionality in 4.9 and now they are using the functionality to make bigmem default for 5.0! They are also tuning bufcache, now I get it Henning!
Start testing guys so the problems are spotted earlier. Did you see the flurry of commits in last 2-4 days. Thanks, you guys rock. On Mon, Apr 4, 2011 at 4:06 AM, Tony Berth <[email protected]> wrote: > Thank you very much for the feedback. Is that related to the following: > > --- > Provided a bus_dmamap_sync() implementation for the amd64 IOMMU. It fixes > bigmem. > --- > > from: > http://www.openbsd.org/plus49.html > > Thanks > > On Mon, Apr 4, 2011 at 8:53 AM, Janne Johansson <[email protected]> wrote: > >> I guess someone listened to you. >> >> http://marc.info/?l=openbsd-cvs&m=130176586700354&w=2 >> >> Next step: >> http://www.openbsd.org/faq/faq5.html >> >> >> 2011/3/30 Tony Berth <[email protected]> >> >>> Thank you for that clarification >>> >>> >>> On Wed, Mar 30, 2011 at 2:52 PM, Janne Johansson <[email protected]>wrote: >>> >>>> >>>> >>>> 2011/3/30 Tony Berth <[email protected]> >>>> >>>>> which translates that the physical 4G limitation is still in place? >>>>> >>>> >>>> Yes, as shown by the 2nd or third line in the dmesg while booting, >>>> somewhere close to 3.5-4G is used only. The MAXDSIZ is how much virtual >>>> memory an app may eat, swap included. >>>> >>>> >>>> -- >>>> To our sweethearts and wives. May they never meet. -- 19th century >>>> toast >>>> >>> >>> >> >> >> -- >> To our sweethearts and wives. May they never meet. -- 19th century toast
