bcachefs: Don't block on allocations when only writing to specific device

Since the copygc thread is now global and not per device, we're not
freeing up space on any one device in bounded time - and indeed we never
really were, since rebalance wasn't moving data around between devices
with that objective.

Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

diff --git a/fs/bcachefs/io.c b/fs/bcachefs/io.c
index 0f4c5c3c408c74e20f0d966fcf48b3ce34a0ddb4..52501714985501696167f33a2a9d7806636d1740 100644 (file)
--- a/fs/bcachefs/io.c
+++ b/fs/bcachefs/io.c
@@ -1090,6 +1090,11 @@ again:
                        goto err;
                }
 
+               /*
+                * The copygc thread is now global, which means it's no longer
+                * freeing up space on specific disks, which means that
+                * allocations for specific disks may hang arbitrarily long:
+                */
                wp = bch2_alloc_sectors_start(c,
                        op->target,
                        op->opts.erasure_code,
@@ -1099,7 +1104,8 @@ again:
                        op->nr_replicas_required,
                        op->alloc_reserve,
                        op->flags,
-                       (op->flags & BCH_WRITE_ALLOC_NOWAIT) ? NULL : cl);
+                       (op->flags & (BCH_WRITE_ALLOC_NOWAIT|
+                                     BCH_WRITE_ONLY_SPECIFIED_DEVS)) ? NULL : cl);
                EBUG_ON(!wp);
 
                if (unlikely(IS_ERR(wp))) {