Re: scheduling while atomic in z3fold

[Mike Galbraith]

On Wed, 2020-12-09 at 07:13 +0100, Mike Galbraith wrote:
> On Wed, 2020-12-09 at 00:26 +0100, Vitaly Wool wrote:
> > Hi Mike,
> >
> > On 2020-12-07 16:41, Mike Galbraith wrote:
> > > On Mon, 2020-12-07 at 16:21 +0100, Vitaly Wool wrote:
> > >> On Mon, Dec 7, 2020 at 1:34 PM Mike Galbraith <efault@xxxxxx> wrote:
> > >>>
> > >>
> > >>> Unfortunately, that made zero difference.
> > >>
> > >> Okay, I suggest that you submit the patch that changes read_lock() to
> > >> write_lock() in __release_z3fold_page() and I'll ack it then.
> > >> I would like to rewrite the code so that write_lock is not necessary
> > >> there but I don't want to hold you back and it isn't likely that I'll
> > >> complete this today.
> > >
> > > Nah, I'm in no rush... especially not to sign off on "Because the
> > > little voices in my head said this bit should look like that bit over
> > > yonder, and testing _seems_ to indicate they're right about that" :)
> > >
> > > 	-Mike
> > >
> >
> > okay, thanks. Would this make things better:
>
> Yup, z3fold became RT tolerant with this (un-munged and) applied.

Below is the other change that any RT users of z3fold will need.

mm, z3fold: Remove preempt disabled sections for RT

Replace get_cpu_ptr() with migrate_disable()+this_cpu_ptr() so RT can take
spinlocks that become sleeping locks.

Signed-off-by Mike Galbraith <efault@xxxxxx>
---
 mm/z3fold.c |   17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

--- a/mm/z3fold.c
+++ b/mm/z3fold.c
@@ -617,14 +617,16 @@ static inline void add_to_unbuddied(stru
 {
 	if (zhdr->first_chunks == 0 || zhdr->last_chunks == 0 ||
 			zhdr->middle_chunks == 0) {
-		struct list_head *unbuddied = get_cpu_ptr(pool->unbuddied);
-
+		struct list_head *unbuddied;
 		int freechunks = num_free_chunks(zhdr);
+
+		migrate_disable();
+		unbuddied = this_cpu_ptr(pool->unbuddied);
 		spin_lock(&pool->lock);
 		list_add(&zhdr->buddy, &unbuddied[freechunks]);
 		spin_unlock(&pool->lock);
 		zhdr->cpu = smp_processor_id();
-		put_cpu_ptr(pool->unbuddied);
+		migrate_enable();
 	}
 }

@@ -861,8 +863,9 @@ static inline struct z3fold_header *__z3
 	int chunks = size_to_chunks(size), i;

 lookup:
+	migrate_disable();
 	/* First, try to find an unbuddied z3fold page. */
-	unbuddied = get_cpu_ptr(pool->unbuddied);
+	unbuddied = this_cpu_ptr(pool->unbuddied);
 	for_each_unbuddied_list(i, chunks) {
 		struct list_head *l = &unbuddied[i];

@@ -880,7 +883,7 @@ static inline struct z3fold_header *__z3
 		    !z3fold_page_trylock(zhdr)) {
 			spin_unlock(&pool->lock);
 			zhdr = NULL;
-			put_cpu_ptr(pool->unbuddied);
+			migrate_enable();
 			if (can_sleep)
 				cond_resched();
 			goto lookup;
@@ -894,7 +897,7 @@ static inline struct z3fold_header *__z3
 		    test_bit(PAGE_CLAIMED, &page->private)) {
 			z3fold_page_unlock(zhdr);
 			zhdr = NULL;
-			put_cpu_ptr(pool->unbuddied);
+			migrate_enable();
 			if (can_sleep)
 				cond_resched();
 			goto lookup;
@@ -909,7 +912,7 @@ static inline struct z3fold_header *__z3
 		kref_get(&zhdr->refcount);
 		break;
 	}
-	put_cpu_ptr(pool->unbuddied);
+	migrate_enable();

 	if (!zhdr) {
 		int cpu;