btrfs: stripe_alloc: do not re-queue a trapped group for reclaim every commit

btrfs: stripe_alloc: do not re-queue a trapped group for reclaim every commit

The reclaim trigger runs at every commit-time rescan, so a group that
stayed above the threshold because its relocation could not get space
(or was still in progress) was handed to the worker again and again:
258 "relocating block group" messages in 24 minutes on one filesystem
while the trapped total moved once.  Remember when a group was last
queued and leave it alone for five minutes.

Assisted-by: Claude:claude-fable-5