Messages in this thread | | | Date | Tue, 3 Jan 2006 11:30:53 -0800 (PST) | From | Christoph Lameter <> | Subject | Re: [PATCH 6/9] clockpro-clockpro.patch |
| |
On Sat, 31 Dec 2005, Marcelo Tosatti wrote:
> > > > + * res | h/c | tst | ref || Hcold | Hhot | Htst || Flt > > > > + * ----+-----+-----+-----++-------+------+------++----- > > > > + * 1 | 1 | 0 | 1 ||=1101 | 1100 |=1101 || > > > > + * 1 | 1 | 0 | 0 ||=1100 | 1000 |=1100 || > > > > + * ----+-----+-----+-----++-------+------+------++----- > > > > + * 1 | 0 | 1 | 1 || 1100 | 1001 | 1001 || > > > > + * 1 | 0 | 1 | 0 ||X0010 | 1000 | 1000 || > > > > + * 1 | 0 | 0 | 1 || 1010 |=1001 |=1001 || > > > > + * 1 | 0 | 0 | 0 ||X0000 |=1000 |=1000 || > > > > + * ----+-----+-----+-----++-------+------+------++----- > > > > + * ----+-----+-----+-----++-------+------+------++----- > > > > + * 0 | 0 | 1 | 1 || | | || 1100 > > > > + * 0 | 0 | 1 | 0 ||=0010 |X0000 |X0000 || > > > > + * 0 | 0 | 0 | 1 || | | || 1010 > > state table, it describes how (in the original paper) the three hands > > modify the page state. Given the state in the first four columns, the > > next three columns give a new state for each hand; hand cold, hot and > > test. The last column describes the action of a pagefault. > > > > Ex. given a resident cold page in its test period that is referenced > > (1011): > > - Hand cold will make it 1100, that is, a resident hot page; > > - Hand hot will make it 1001, that is, a resident cold page with a > > reference; and > > - Hand test will also make it 1001. > > > > (The prefixes '=' and 'X' are used to indicate: not changed, and remove > > from list - that can be either move from resident->non-resident or > > remove altogether). > > I see - can you add this info to the patch?
Hmm.. This looks as if it would be better to manage the page state as a bitmap rather than individual bits. Could we put this state in an integer variable and do an array lookup to get to the next state? - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |