[RFC][PATCH 13/16] writeback: restructure bdi forker loop a little
in [Kernel]
|
Prev: [RFC][PATCH 10/16] writeback: simplify bdi code a little
Next: [RFC][PATCH 16/16] writeback: prevent unnecessary bdi threads wakeups
From: Artem Bityutskiy on 16 Jul 2010 09:00 From: Artem Bityutskiy <Artem.Bityutskiy(a)nokia.com> This patch re-structures the loop which walks bdi a little. This is really just a micro-step towards the coming change where the forker thread will kill the bdi threads. Just a tiny preparation which should simplify reviewing and make it easier to see what I changed and catch mistakes by reviewing. To put it differently, the final patch which moves the inactive bdi threads exiting logic will be smaller if I do this little preparation. Smaller patches are easier to review. This patch also adds the 'bdi_cap_flush_forker(bdi)' condition to the loop. The reason for this is that the forker thread can start _before_ the BID_pending flag is set (see 'bdi_register()'), so the WARN() statement will fire for the default bdi. I observed this warning the system boots up. This patch also amends comments a little. Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy(a)nokia.com> --- mm/backing-dev.c | 23 +++++++++++------------ 1 files changed, 11 insertions(+), 12 deletions(-) diff --git a/mm/backing-dev.c b/mm/backing-dev.c index 0123d6f..18d7c22 100644 --- a/mm/backing-dev.c +++ b/mm/backing-dev.c @@ -340,24 +340,23 @@ static int bdi_forker_thread(void *ptr) spin_lock_bh(&bdi_lock); set_current_state(TASK_INTERRUPTIBLE); - /* - * Check if any existing bdi's have dirty data without - * a thread registered. If so, set that up. - */ list_for_each_entry_safe(bdi, tmp, &bdi_list, bdi_list) { - if (!bdi_cap_writeback_dirty(bdi)) - continue; - if (bdi->wb.task) - continue; - if (!bdi_has_dirty_io(bdi)) + if (!bdi_cap_writeback_dirty(bdi) || + bdi_cap_flush_forker(bdi)) continue; WARN(!test_bit(BDI_registered, &bdi->state), "bdi %p/%s is not registered!\n", bdi, bdi->name); - list_del_rcu(&bdi->bdi_list); - fork = true; - break; + /* + * If the bdi has work to do, but the thread does not + * exist - create it. + */ + if (!bdi->wb.task && bdi_has_dirty_io(bdi)) { + list_del_rcu(&bdi->bdi_list); + fork = true; + break; + } } spin_unlock_bh(&bdi_lock); -- 1.7.1.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |