RAID5: check_reshape() shouldn't call mddev_suspend

[ Upstream commit 27a353c0 ]

check_reshape() is called from raid5d thread. raid5d thread shouldn't
call mddev_suspend(), because mddev_suspend() waits for all IO finish
but IO is handled in raid5d thread, we could easily deadlock here.

This issue is introduced by
738a2738 ("md/raid5: fix allocation of 'scribble' array.")

Cc: stable@vger.kernel.org (v4.1+)
Reported-and-tested-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Reviewed-by: NeilBrown <neilb@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>

drivers/md/raid5.c

@@ -2080,6 +2080,14 @@ static int resize_chunks(struct r5conf *conf, int new_disks, int new_sectors)
 	unsigned long cpu;
 	int err = 0;
 
+	/*
+	 * Never shrink. And mddev_suspend() could deadlock if this is called
+	 * from raid5d. In that case, scribble_disks and scribble_sectors
+	 * should equal to new_disks and new_sectors
+	 */
+	if (conf->scribble_disks >= new_disks &&
+	    conf->scribble_sectors >= new_sectors)
+		return 0;
 	mddev_suspend(conf->mddev);
 	get_online_cpus();
 	for_each_present_cpu(cpu) {
@@ -2101,6 +2109,10 @@ static int resize_chunks(struct r5conf *conf, int new_disks, int new_sectors)
 	}
 	put_online_cpus();
 	mddev_resume(conf->mddev);
+	if (!err) {
+		conf->scribble_disks = new_disks;
+		conf->scribble_sectors = new_sectors;
+	}
 	return err;
 }
 
@@ -6366,6 +6378,12 @@ static int raid5_alloc_percpu(struct r5conf *conf)
 	}
 	put_online_cpus();
 
+	if (!err) {
+		conf->scribble_disks = max(conf->raid_disks,
+			conf->previous_raid_disks);
+		conf->scribble_sectors = max(conf->chunk_sectors,
+			conf->prev_chunk_sectors);
+	}
 	return err;
 }
 

drivers/md/raid5.h

@@ -501,6 +501,8 @@ struct r5conf {
 					      * conversions
 					      */
 	} __percpu *percpu;
+	int scribble_disks;
+	int scribble_sectors;
 #ifdef CONFIG_HOTPLUG_CPU
 	struct notifier_block	cpu_notify;
 #endif