From: NeilBrown <neilb@suse.de>
Date: Wed, 27 Oct 2021 15:12:30 +1100
Subject: [PATCH] SUNRPC/auth: async tasks mustn't block waiting for memory
Patch-mainline: Not yet - undergoing review
References: bsc#1191876 bsc#1192866

When memory is short, new worker threads cannot be created and we depend
on the minimum one rpciod thread to be able to handle everything.  So it
must not block waiting for memory.

mempools are particularly a problem as memory can only be released back
to the mempool by an async rpc task running.  If all available workqueue
threads are waiting on the mempool, no thread is available to return
anything.

lookup_cred() can block on a mempool or kmalloc - and this can cause
deadlocks.  So add a new RPCAUTH_LOOKUP flag for async lookups and don't
block on memory.  If the -ENOMEM gets back to call_refreshresult(), wait
a short while and try again.  HZ>>4 is chosen as it is used elsewhere
for -ENOMEM retries.

Signed-off-by: NeilBrown <neilb@suse.de>
Acked-by: NeilBrown <neilb@suse.com>

---
 include/linux/sunrpc/auth.h    |    1 +
 net/sunrpc/auth.c              |    6 +++++-
 net/sunrpc/auth_generic.c      |    5 ++++-
 net/sunrpc/auth_gss/auth_gss.c |    6 +++++-
 net/sunrpc/auth_unix.c         |    5 ++++-
 net/sunrpc/clnt.c              |    3 +++
 6 files changed, 22 insertions(+), 4 deletions(-)

--- a/include/linux/sunrpc/auth.h
+++ b/include/linux/sunrpc/auth.h
@@ -116,6 +116,7 @@ struct rpc_auth_create_args {
 /* Flags for rpcauth_lookupcred() */
 #define RPCAUTH_LOOKUP_NEW		0x01	/* Accept an uninitialised cred */
 #define RPCAUTH_LOOKUP_RCU		0x02	/* lock-less lookup */
+#define RPCAUTH_LOOKUP_ASYNC		0x04	/* Don't block waiting for memory */
 
 /*
  * Client authentication ops
--- a/net/sunrpc/auth.c
+++ b/net/sunrpc/auth.c
@@ -673,6 +673,8 @@ rpcauth_bind_root_cred(struct rpc_task *
 
 	dprintk("RPC: %5u looking up %s cred\n",
 		task->tk_pid, task->tk_client->cl_auth->au_ops->au_name);
+	if (RPC_IS_ASYNC(task))
+		lookupflags |= RPCAUTH_LOOKUP_ASYNC;
 	return auth->au_ops->lookup_cred(auth, &acred, lookupflags);
 }
 
@@ -683,6 +685,8 @@ rpcauth_bind_new_cred(struct rpc_task *t
 
 	dprintk("RPC: %5u looking up %s cred\n",
 		task->tk_pid, auth->au_ops->au_name);
+	if (RPC_IS_ASYNC(task))
+		lookupflags |= RPCAUTH_LOOKUP_ASYNC;
 	return rpcauth_lookupcred(auth, lookupflags);
 }
 
@@ -694,7 +698,7 @@ rpcauth_bindcred(struct rpc_task *task,
 	int lookupflags = 0;
 
 	if (flags & RPC_TASK_ASYNC)
-		lookupflags |= RPCAUTH_LOOKUP_NEW;
+		lookupflags |= RPCAUTH_LOOKUP_NEW | RPCAUTH_LOOKUP_ASYNC;
 	if (cred != NULL)
 		new = cred->cr_ops->crbind(task, cred, lookupflags);
 	else if (flags & RPC_TASK_ROOTCREDS)
--- a/net/sunrpc/auth_generic.c
+++ b/net/sunrpc/auth_generic.c
@@ -103,7 +103,10 @@ generic_hash_cred(struct auth_cred *acre
 static struct rpc_cred *
 generic_lookup_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
 {
-	return rpcauth_lookup_credcache(&generic_auth, acred, flags, GFP_KERNEL);
+	gfp_t gfp = GFP_NOFS;
+	if (flags & RPCAUTH_LOOKUP_ASYNC)
+		gfp = GFP_NOWAIT | __GFP_NOWARN;
+	return rpcauth_lookup_credcache(&generic_auth, acred, flags, gfp);
 }
 
 static struct rpc_cred *
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -1315,7 +1315,11 @@ gss_hash_cred(struct auth_cred *acred, u
 static struct rpc_cred *
 gss_lookup_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
 {
-	return rpcauth_lookup_credcache(auth, acred, flags, GFP_NOFS);
+	gfp_t gfp = GFP_NOFS;
+
+	if (flags & RPCAUTH_LOOKUP_ASYNC)
+		gfp = GFP_NOWAIT | __GFP_NOWARN;
+	return rpcauth_lookup_credcache(auth, acred, flags, gfp);
 }
 
 static struct rpc_cred *
--- a/net/sunrpc/auth_unix.c
+++ b/net/sunrpc/auth_unix.c
@@ -69,7 +69,10 @@ unx_hash_cred(struct auth_cred *acred, u
 static struct rpc_cred *
 unx_lookup_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
 {
-	return rpcauth_lookup_credcache(auth, acred, flags, GFP_NOFS);
+	gfp_t gfp = GFP_NOFS;
+	if (flags & RPCAUTH_LOOKUP_ASYNC)
+		gfp = GFP_NOWAIT | __GFP_NOWARN;
+	return rpcauth_lookup_credcache(auth, acred, flags, gfp);
 }
 
 static struct rpc_cred *
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -1708,6 +1708,9 @@ call_refreshresult(struct rpc_task *task
 		dprintk("RPC: %5u %s: retry refresh creds\n",
 				task->tk_pid, __func__);
 		return;
+	case -ENOMEM:
+		rpc_delay(task, HZ >> 4);
+		return;
 	}
 	dprintk("RPC: %5u %s: refresh creds failed with error %d\n",
 				task->tk_pid, __func__, status);