Message ID | 1579721990-18672-1-git-send-email-yang.shi@linux.alibaba.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | mm: move_pages: report the number of non-attempted pages | expand |
On Thu, 23 Jan 2020 03:39:50 +0800 Yang Shi <yang.shi@linux.alibaba.com> wrote: > Since commit a49bd4d71637 ("mm, numa: rework do_pages_move"), > the semantic of move_pages() was changed to return the number of > non-migrated pages (failed to migration) and the call would be aborted > immediately if migrate_pages() returns positive value. But it didn't > report the number of pages that we even haven't attempted to migrate. > So, fix it by including non-attempted pages in the return value. > > Fixes: a49bd4d71637 ("mm, numa: rework do_pages_move") > Suggested-by: Michal Hocko <mhocko@suse.com> > Cc: Wei Yang <richardw.yang@linux.intel.com> > Cc: <stable@vger.kernel.org> [4.17+] > Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com> > --- > The patch is based off Wei Yang's cleanup patchset: > https://lore.kernel.org/linux-mm/20200122011647.13636-1-richardw.yang@linux.intel.com/T/#t Can you please redo this so it is applicable to current mainline? That will make it more easily backportable and this fix is higher priority than a set of cleanups.
On 1/22/20 2:05 PM, Andrew Morton wrote: > On Thu, 23 Jan 2020 03:39:50 +0800 Yang Shi <yang.shi@linux.alibaba.com> wrote: > >> Since commit a49bd4d71637 ("mm, numa: rework do_pages_move"), >> the semantic of move_pages() was changed to return the number of >> non-migrated pages (failed to migration) and the call would be aborted >> immediately if migrate_pages() returns positive value. But it didn't >> report the number of pages that we even haven't attempted to migrate. >> So, fix it by including non-attempted pages in the return value. >> >> Fixes: a49bd4d71637 ("mm, numa: rework do_pages_move") >> Suggested-by: Michal Hocko <mhocko@suse.com> >> Cc: Wei Yang <richardw.yang@linux.intel.com> >> Cc: <stable@vger.kernel.org> [4.17+] >> Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com> >> --- >> The patch is based off Wei Yang's cleanup patchset: >> https://lore.kernel.org/linux-mm/20200122011647.13636-1-richardw.yang@linux.intel.com/T/#t > Can you please redo this so it is applicable to current mainline? That > will make it more easily backportable and this fix is higher priority > than a set of cleanups. Sure.
diff --git a/mm/migrate.c b/mm/migrate.c index 591f2e5..51b1b76 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1582,7 +1582,7 @@ static int add_page_for_migration(struct mm_struct *mm, unsigned long addr, static int move_pages_and_store_status(struct mm_struct *mm, int node, struct list_head *pagelist, int __user *status, - int start, int nr) + int start, int nr, unsigned long total) { int err; @@ -1590,8 +1590,17 @@ static int move_pages_and_store_status(struct mm_struct *mm, int node, return 0; err = do_move_pages_to_node(mm, pagelist, node); - if (err) + if (err) { + /* + * Possitive err means the number of failed pages to + * migrate. Since we are going to abort and return the + * number of non-migrated pages, so need incude the rest + * of the nr_pages that are not attempted as well. + */ + if (err > 0) + err += total - start - nr - 1; return err; + } err = store_status(status, start, node, nr); return err; } @@ -1640,7 +1649,8 @@ static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes, start = i; } else if (node != current_node) { err = move_pages_and_store_status(mm, current_node, - &pagelist, status, start, i - start); + &pagelist, status, start, i - start, + nr_pages); if (err) goto out; start = i; @@ -1670,7 +1680,7 @@ static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes, goto out_flush; err = move_pages_and_store_status(mm, current_node, &pagelist, - status, start, i - start); + status, start, i - start, nr_pages); if (err) goto out; current_node = NUMA_NO_NODE; @@ -1678,7 +1688,7 @@ static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes, out_flush: /* Make sure we do not overwrite the existing error */ err1 = move_pages_and_store_status(mm, current_node, &pagelist, - status, start, i - start); + status, start, i - start, nr_pages); if (err >= 0) err = err1; out:
Since commit a49bd4d71637 ("mm, numa: rework do_pages_move"), the semantic of move_pages() was changed to return the number of non-migrated pages (failed to migration) and the call would be aborted immediately if migrate_pages() returns positive value. But it didn't report the number of pages that we even haven't attempted to migrate. So, fix it by including non-attempted pages in the return value. Fixes: a49bd4d71637 ("mm, numa: rework do_pages_move") Suggested-by: Michal Hocko <mhocko@suse.com> Cc: Wei Yang <richardw.yang@linux.intel.com> Cc: <stable@vger.kernel.org> [4.17+] Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com> --- The patch is based off Wei Yang's cleanup patchset: https://lore.kernel.org/linux-mm/20200122011647.13636-1-richardw.yang@linux.intel.com/T/#t mm/migrate.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-)