z3fold: limit use of stale list for allocation (5c9bab59) · Commits · 戴 / test

Currently if z3fold couldn't find an unbuddied page it would first try to pull a page off the stale list. The problem with this approach is that we can't 100% guarantee that the page is not processed by the workqueue thread at the same time unless we run cancel_work_sync() on it, which we can't do if we're in an atomic context. So let's just limit stale list usage to non-atomic contexts only. Link: http://lkml.kernel.org/r/47ab51e7-e9c1-d30e-ab17-f734dbc3abce@gmail.com Signed-off-by:

Vitaly Vul <vitaly.vul@sony.com> Reviewed-by:

Andrew Morton <akpm@linux-foundation.org> Cc: <Oleksiy.Avramchenko@sony.com> Signed-off-by:

Andrew Morton <akpm@linux-foundation.org> Signed-off-by:

Linus Torvalds <torvalds@linux-foundation.org>

mm/z3fold.c

+19 −16

Original line number	Diff line number	Diff line
		@@ -620,24 +620,27 @@ lookup:
		bud = FIRST;
		}

		page = NULL;
		if (can_sleep) {
		spin_lock(&pool->stale_lock);
		zhdr = list_first_entry_or_null(&pool->stale,
		struct z3fold_header, buddy);
		/*
		* Before allocating a page, let's see if we can take one from the
		* stale pages list. cancel_work_sync() can sleep so we must make
		* sure it won't be called in case we're in atomic context.
		* Before allocating a page, let's see if we can take one from
		* the stale pages list. cancel_work_sync() can sleep so we
		* limit this case to the contexts where we can sleep
		*/
		if (zhdr && (can_sleep \|\| !work_pending(&zhdr->work))) {
		if (zhdr) {
		list_del(&zhdr->buddy);
		spin_unlock(&pool->stale_lock);
		if (can_sleep)
		cancel_work_sync(&zhdr->work);
		page = virt_to_page(zhdr);
		} else {
		spin_unlock(&pool->stale_lock);
		page = alloc_page(gfp);
		}
		}
		if (!page)
		page = alloc_page(gfp);

		if (!page)
		return -ENOMEM;

Admin message