Re: [Fwd: Re: [LTP]: Test Case Design for move_pages()]]

["Michael Kerrisk" <mtk.manpages@xxxxxxxxxxxxxx>]

[CC += cl@xxxxxxxxxxxxxxxxxxxx]
On Thu, Jul 3, 2008 at 2:35 PM, Subrata Modak<subrata@xxxxxxxxxxxxxxxxxx> wrote:> On Thu, 2008-07-03 at 09:04 +0530, Vijay Kumar wrote:>> Subrata Modak wrote:>> > Would you like to incorporate Andi´s review comments inthe next patch.>> > On Mon, 2008-06-30 at 14:39 +0200, Andi Kleen wrote:>> > > > +void>> > > > +verify_pages_on_nodes(int *status, unsigned int num, int *nodes)>> > > > +{>> > > > +       unsigned int i;>> > > > +>> > > > +       for (i = 0; i < num; i++) {>> > > > +               if (status[i] != nodes[i]) {>> > > > +                       tst_resm(TFAIL, "page %d on node %d, ">> > > > +                                "expected on node %d", i,>> > > > +                                status[i], nodes[i]);>> > > > +                       return;>> > > > +               }>> > >>> > > The correct way to verify this would be to use get_mempolicy(...,>> > > MPOL_F_NODE|MPOL_F_ADDR, &node) on each page and check if the node>> > > is correct.>>>> OK, I will make this change and provide an updated patch.>>>> It would also be great if someone can explain the interaction between>> mlock and move_pages. The man page>> http://linux.die.net/man/2/move_pages says that the status[] will be>> set to -EPERM if the page is mlocked. But that does not seem to>> happen, and an mlocked page gets happily moved around. So what is the>> actual expected behaviour?>> I would rather loop in Michael directly here. Michael, can you please> comment on Vijayś question on move_pages man page !
Here, you encounter some dusty corners.
That page is not yet part of my set, though it really should be (*),since it is a syscall, and several related pages are in man-pagesalready.Probably, the page that Vijay is referring to comes from:http://oss.sgi.com/projects/libnuma/
I myself can't help you with the details of move_pages(2).
The man page and (I imagine also) the syscall itself were written byChristoph Lameter, whose email address recently changed (I've CCed hisnew address).  Perhaps he can provide more insight.
Cheers,
Michael
(*) Getting this page into man-pages is on my TODO list:http://www.kernel.org/doc/man-pages/missing_pages.html
-- Michael KerriskLinux man-pages maintainer; http://www.kernel.org/doc/man-pages/man-pages online: http://www.kernel.org/doc/man-pages/online_pages.htmlFound a bug? http://www.kernel.org/doc/man-pages/reporting_bugs.html˙ôčş{.nÇ+?ˇ?Ž?­?+%?Ë˙ąéÝś;Ľ?w˙ş{.nÇ+?ˇĽ?{ąţ{Śkř§ś?ĄÜ¨}Š?˛Ć zÚ&j:+v?¨ţřŻůŽwĽţ?ŕ2?Ţ?¨č­Ú&˘)ߥŤaśÚ˙˙űŕzżäzšŢ?ú+?ů???ݢj˙?wčţf