tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: pmap_enter(9) rework



On Feb 1, 2013, at 10:29 AM, Toru Nishimura wrote:

> The reason why NB pmap_enter(9) implementation
> so vary among processors are the definition and
> usage of pmap_enter(9) by upper VM layer is incorrect
> or less matched with HW realities.  The most scareful
> situation is the necessity to track the change of existing
> pg attribute inside pmap_enter.  It invovles complicated
> case analysis., but in essense the case analysis is
> common across processors.

Not true.  The PPC OEA pmap is very different because it uses an inverted page 
table.

However, most software managed TLB pmaps can use the common pmap stuff in 
uvm/pmap.

Common PV list handling is something chs is working on.



Home | Main Index | Thread Index | Old Index