Re: page color initialization

2011-09-30 Thread Thor Lancelot Simon
On Fri, Sep 30, 2011 at 03:36:14PM +0900, Masao Uebayashi wrote: > > - is there any architecture which allows MP with different cache-line sizes? At a guess, this can probably happen on an SGI Challenge (a real Challenge, that is, a DM, L, or XL, not a "Challenge S"). It is possible it can also

Re: page color initialization

2011-09-29 Thread Masao Uebayashi
On Fri, Sep 30, 2011 at 1:13 PM, matthew green wrote: > >> The current code seems to allow to change uvmexp.ncolors at >> configure() -> cpu_attach() -> uvm_page_recolor().  I think changing >> uvmexp.ncolors after uvm_init() is too late, and allocating such >> fundamental data like page freelist

re: page color initialization

2011-09-29 Thread matthew green
> The current code seems to allow to change uvmexp.ncolors at > configure() -> cpu_attach() -> uvm_page_recolor(). I think changing > uvmexp.ncolors after uvm_init() is too late, and allocating such > fundamental data like page freelist buckets that late is best to > avoid. > > How about: > > -

page color initialization

2011-09-29 Thread Masao Uebayashi
The current code seems to allow to change uvmexp.ncolors at configure() -> cpu_attach() -> uvm_page_recolor(). I think changing uvmexp.ncolors after uvm_init() is too late, and allocating such fundamental data like page freelist buckets that late is best to avoid. How about: - determine cache si