A deadlock can happen when an overloaded system runs ->parallel() in the
context of the current task:

    padata_do_parallel
      ->parallel()
        pcrypt_aead_enc/dec
          padata_do_serial
            spin_lock(&reorder->lock) // BHs still enabled
              <interrupt>
                ...
                  __do_softirq
                    ...
                      padata_do_serial
                        spin_lock(&reorder->lock)

It's a bug for BHs to be on in _do_serial as Steffen points out, so
ensure they're off in the "current task" case like they are in
padata_parallel_worker to avoid this situation.

Reported-by: syzbot+bc05445bc14148d51915@syzkaller.appspotmail.com
Fixes: 4611ce224688 ("padata: allocate work structures for parallel jobs from a pool")
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
---

About the potential deadlock in padata_find_next(), I think this fix
covers us there since that function is only called from
padata_do_serial(), where BHs should now always be off, or
invoke_padata_reorder(), where BHs are off as well.  The _bh spinlock
variants in padata_reorder() seem redundant.

 kernel/padata.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/kernel/padata.c b/kernel/padata.c
index e5819bb8bd1d..97f51e0c1776 100644
--- a/kernel/padata.c
+++ b/kernel/padata.c
@@ -207,14 +207,16 @@ int padata_do_parallel(struct padata_shell *ps,
 	pw = padata_work_alloc();
 	spin_unlock(&padata_works_lock);
 
+	if (!pw) {
+		/* Maximum works limit exceeded, run in the current task. */
+		padata->parallel(padata);
+	}
+
 	rcu_read_unlock_bh();
 
 	if (pw) {
 		padata_work_init(pw, padata_parallel_worker, padata, 0);
 		queue_work(pinst->parallel_wq, &pw->pw_work);
-	} else {
-		/* Maximum works limit exceeded, run in the current task. */
-		padata->parallel(padata);
 	}
 
 	return 0;
-- 
2.38.1

On Wed, Nov 16, 2022 at 08:28:02PM -0500, Daniel Jordan wrote:
> A deadlock can happen when an overloaded system runs ->parallel() in the
> context of the current task:
> 
>     padata_do_parallel
>       ->parallel()
>         pcrypt_aead_enc/dec
>           padata_do_serial
>             spin_lock(&reorder->lock) // BHs still enabled
>               <interrupt>
>                 ...
>                   __do_softirq
>                     ...
>                       padata_do_serial
>                         spin_lock(&reorder->lock)
> 
> It's a bug for BHs to be on in _do_serial as Steffen points out, so
> ensure they're off in the "current task" case like they are in
> padata_parallel_worker to avoid this situation.
> 
> Reported-by: syzbot+bc05445bc14148d51915@syzkaller.appspotmail.com
> Fixes: 4611ce224688 ("padata: allocate work structures for parallel jobs from a pool")
> Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
> Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
> ---
> 
> About the potential deadlock in padata_find_next(), I think this fix
> covers us there since that function is only called from
> padata_do_serial(), where BHs should now always be off, or
> invoke_padata_reorder(), where BHs are off as well.  The _bh spinlock
> variants in padata_reorder() seem redundant.
> 
>  kernel/padata.c | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt