Re: [PATCH 1/2] ocfs2: cluster: use GFP_NOFS for heartbeat bio allocation

[Joseph Qi]

On 7/11/26 8:26 AM, Andrew Morton wrote:
> On Fri, 10 Jul 2026 15:17:55 +0800 Joseph Qi <joseph.qi@xxxxxxxxxxxxxxxxx> wrote:
> 
>> o2hb_setup_one_bio() allocates the heartbeat bio with GFP_ATOMIC. The
>> disk heartbeat runs in the o2hb kernel thread (o2hb_do_disk_heartbeat),
>> which is process context and can sleep, so there is no atomicity
>> requirement here.
>>
>> GFP_ATOMIC lacks __GFP_DIRECT_RECLAIM, so the allocation is not served
>> from the fs_bio_set mempool reserve and can return NULL under memory
>> pressure. A failed heartbeat allocation aborts the heartbeat and can
>> lead to the local node being fenced, which is exactly what the old
>> comment worried about.
>>
>> Use GFP_NOFS instead. It keeps __GFP_DIRECT_RECLAIM so the allocation is
>> backed by the fs_bio_set mempool and cannot fail, while avoiding
>> recursion back into the filesystem during heartbeat I/O. As the
>> allocation can no longer fail, drop the dead ERR_PTR(-ENOMEM) path in
>> o2hb_setup_one_bio() and the now-redundant IS_ERR() handling in its
>> callers.
> 
> fyi, AI review asked a thing:
> 	https://sashiko.dev/#/patchset/20260710071756.3586797-1-joseph.qi@xxxxxxxxxxxxxxxxx

GFP_NOFS won't reintroduce the fencing stall. mempool_alloc strips
__GFP_DIRECT_RECLAIM on the first attempt (mempool_adjust_gfp) and
checks the reserve before any reclaim, so the common path never blocks.
Reclaim happens only if the reserve is fully drained — strictly better
than GFP_ATOMIC failing outright.

And the further improvement is what next patch exactly does.

Thanks,
Joseph