ocfs2-cluster-use-gfp_nofs-for-heartbeat-bio-allocation.patch removed from -mm tree
The quilt patch titled
Subject: ocfs2: cluster: use GFP_NOFS for heartbeat bio allocation
has been removed from the -mm tree. Its filename was
ocfs2-cluster-use-gfp_nofs-for-heartbeat-bio-allocation.patch
This patch was dropped because it was merged into the mm-nonmm-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: Joseph Qi <[email protected]>
Subject: ocfs2: cluster: use GFP_NOFS for heartbeat bio allocation
Date: Fri, 10 Jul 2026 15:17:55 +0800
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.
---
fs/ocfs2/cluster/heartbeat.c | 41 ++++++--------------------------- 1 file changed, 8 insertions(+), 33 deletions(-)
fs/ocfs2/cluster/heartbeat.c renamed from fs/ocfs2/cluster/heartbeat.c~ocfs2-cluster-use-gfp_nofs-for-heartbeat-bio-allocation +8 -33
@@ -522,16 +522,12 @@ static struct bio *o2hb_setup_one_bio(ststruct bio *bio;struct page *page; -/* Testing has shown this allocation to take long enough under -* GFP_KERNEL that the local node can get fenced. It would be -* nicest if we could pre-allocate these bios and avoid this -* all together. */ - bio =bio_alloc (reg_bdev (reg),16 , opf,GFP_ATOMIC ); -if (!bio) { -mlog (ML_ERROR ,"Could not alloc slots BIO!\n" ); - bio =ERR_PTR (-ENOMEM ); - goto bail; - } +/* +* The heartbeat runs in process context and can sleep, so use +* GFP_NOFS. It is backed by the fs_bio_set mempool and thus cannot +* fail, while avoiding recursion back into the filesystem. +*/ + bio =bio_alloc (reg_bdev (reg),16 , opf,GFP_NOFS );/* Must put everything in 512 byte sectors for the bio... */ bio->bi_iter.bi_sector = (reg->hr_start_block + cs) << (bits -9 );
@@ -556,7 +552,6 @@ static struct bio *o2hb_setup_one_bio(st vec_start =0 ; } -bail: *current_slot = cs;return bio; }
@@ -566,7 +561,6 @@ static int o2hb_read_slots(struct o2hb_r unsigned int max_slots) {unsigned int current_slot = begin_slot; -int status;struct o2hb_bio_wait_ctxt wc;struct bio *bio;
@@ -575,30 +569,18 @@ static int o2hb_read_slots(struct o2hb_rwhile (current_slot < max_slots) { bio =o2hb_setup_one_bio (reg, &wc, ¤t_slot, max_slots,REQ_OP_READ ); -if (IS_ERR (bio)) { - status =PTR_ERR (bio); -mlog_errno (status); - goto bail_and_wait; - } -atomic_inc (&wc.wc_num_reqs);submit_bio (bio); } - status =0 ; - -bail_and_wait:o2hb_wait_on_io (&wc); -if (wc.wc_error && !status) - status = wc.wc_error; -return status; +return wc.wc_error; }static int o2hb_issue_node_write (struct o2hb_region *reg,struct o2hb_bio_wait_ctxt *write_wc) { -int status;unsigned int slot;struct bio *bio;
@@ -610,18 +592,11 @@ static int o2hb_issue_node_write(struct bio =o2hb_setup_one_bio (reg, write_wc, &slot, slot+1 ,REQ_OP_WRITE |REQ_SYNC ); -if (IS_ERR (bio)) { - status =PTR_ERR (bio); -mlog_errno (status); - goto bail; - }atomic_inc (&write_wc->wc_num_reqs);submit_bio (bio); - status =0 ; -bail: -return status; +return 0 ; } static u32 o2hb_compute_block_crc_le(struct o2hb_region *reg,
_
Patches currently in -mm which might be from [email protected] are