thank you mulyadi, i was looking for further insight, and from the stack trace, found this function - cpuset_migrate_mm()'s comment here: /* * cpuset_migrate_mm * * Migrate memory region from one set of nodes to another. * * Temporarilly set tasks mems_allowed to target nodes of migration, * so that the migration code can allocate pages on these nodes. * * Call holding cgroup_mutex, so current's cpuset won't change * during this call, as manage_mutex holds off any cpuset_attach() * calls. Therefore we don't need to take task_lock around the * call to guarantee_online_mems(), as we know no one is changing * our task's cpuset. * * Hold callback_mutex around the two modifications of our tasks * mems_allowed to synchronize with cpuset_mems_allowed(). * * While the mm_struct we are migrating is typically from some * other task, the task_struct mems_allowed that we are hacking * is for our current task, which must allocate new pages for that * migrating memory region. * * We call cpuset_update_task_memory_state() before hacking * our tasks mems_allowed, so that we are assured of being in * sync with our tasks cpuset, and in particular, callbacks to * cpuset_update_task_memory_state() from nested page allocations * won't see any mismatch of our cpuset and task mems_generation * values, so won't overwrite our hacked tasks mems_allowed * nodemask. */ So migration is from node to node. thanks. On Wed, Jun 18, 2008 at 10:57 AM, Mulyadi Santosa <mulyadi.santosa@xxxxxxxxx> wrote: > Hi... > > On Tue, Jun 17, 2008 at 9:51 PM, Peter Teoh <htmldeveloper@xxxxxxxxx> wrote: >> I am interested in this: >> >> why do u do page migration? > > AFAIK, (i also study about it briefly during reading about > ZONE_MOVEABLE), it happens mostly in NUMA with hot pluggable memory > capability. Think when you want to unplug certain memory chip...you > have to "migrate" the page, right? Migrate here, IMHO means: > a. copying the page's content somewhere > b. fixing the PTE so that it points to the new page that hold the > copied data while maintaining consistent virtual address > > regards, > > Mulyadi. > -- Regards, Peter Teoh -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
