public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
* [merged mm-hotfixes-stable] ocfs2-fix-the-issue-with-discontiguous-allocation-in-the-global_bitmap.patch removed from -mm tree
@ 2025-05-08  6:40 Andrew Morton
  0 siblings, 0 replies; only message in thread
From: Andrew Morton @ 2025-05-08  6:40 UTC (permalink / raw)
  To: mm-commits, stable, piaojun, mark, junxiao.bi, joseph.qi, jlbec,
	gechangwei, gautham.ananthakrishna, heming.zhao, akpm


The quilt patch titled
     Subject: ocfs2: fix the issue with discontiguous allocation in the global_bitmap
has been removed from the -mm tree.  Its filename was
     ocfs2-fix-the-issue-with-discontiguous-allocation-in-the-global_bitmap.patch

This patch was dropped because it was merged into the mm-hotfixes-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

------------------------------------------------------
From: Heming Zhao <heming.zhao@suse.com>
Subject: ocfs2: fix the issue with discontiguous allocation in the global_bitmap
Date: Mon, 14 Apr 2025 14:01:23 +0800

commit 4eb7b93e0310 ("ocfs2: improve write IO performance when
fragmentation is high") introduced another regression.

The following ocfs2-test case can trigger this issue:
> discontig_runner.sh => activate_discontig_bg.sh => resv_unwritten:
> ${RESV_UNWRITTEN_BIN} -f ${WORK_PLACE}/large_testfile -s 0 -l \
> $((${FILE_MAJOR_SIZE_M}*1024*1024))

In my env, test disk size (by "fdisk -l <dev>"):
> 53687091200 bytes, 104857600 sectors.

Above command is:
> /usr/local/ocfs2-test/bin/resv_unwritten -f \
> /mnt/ocfs2/ocfs2-activate-discontig-bg-dir/large_testfile -s 0 -l \
> 53187969024

Error log:
> [*] Reserve 50724M space for a LARGE file, reserve 200M space for future test.
> ioctl error 28: "No space left on device"
> resv allocation failed Unknown error -1
> reserve unwritten region from 0 to 53187969024.

Call flow:
__ocfs2_change_file_space //by ioctl OCFS2_IOC_RESVSP64
 ocfs2_allocate_unwritten_extents //start:0 len:53187969024
  while()
   + ocfs2_get_clusters //cpos:0, alloc_size:1623168 (cluster number)
   + ocfs2_extend_allocation
     + ocfs2_lock_allocators
     |  + choose OCFS2_AC_USE_MAIN & ocfs2_cluster_group_search
     |
     + ocfs2_add_inode_data
        ocfs2_add_clusters_in_btree
         __ocfs2_claim_clusters
          ocfs2_claim_suballoc_bits
          + During the allocation of the final part of the large file
	    (after ~47GB), no chain had the required contiguous
            bits_wanted. Consequently, the allocation failed.

How to fix:
When OCFS2 is encountering fragmented allocation, the file system should
stop attempting bits_wanted contiguous allocation and instead provide the
largest available contiguous free bits from the cluster groups.

Link: https://lkml.kernel.org/r/20250414060125.19938-2-heming.zhao@suse.com
Fixes: 4eb7b93e0310 ("ocfs2: improve write IO performance when fragmentation is high")
Signed-off-by: Heming Zhao <heming.zhao@suse.com>
Reported-by: Gautham Ananthakrishna <gautham.ananthakrishna@oracle.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Jun Piao <piaojun@huawei.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 fs/ocfs2/suballoc.c |   38 ++++++++++++++++++++++++++++++++------
 fs/ocfs2/suballoc.h |    1 +
 2 files changed, 33 insertions(+), 6 deletions(-)

--- a/fs/ocfs2/suballoc.c~ocfs2-fix-the-issue-with-discontiguous-allocation-in-the-global_bitmap
+++ a/fs/ocfs2/suballoc.c
@@ -698,10 +698,12 @@ static int ocfs2_block_group_alloc(struc
 
 	bg_bh = ocfs2_block_group_alloc_contig(osb, handle, alloc_inode,
 					       ac, cl);
-	if (PTR_ERR(bg_bh) == -ENOSPC)
+	if (PTR_ERR(bg_bh) == -ENOSPC) {
+		ac->ac_which = OCFS2_AC_USE_MAIN_DISCONTIG;
 		bg_bh = ocfs2_block_group_alloc_discontig(handle,
 							  alloc_inode,
 							  ac, cl);
+	}
 	if (IS_ERR(bg_bh)) {
 		status = PTR_ERR(bg_bh);
 		bg_bh = NULL;
@@ -1794,6 +1796,7 @@ static int ocfs2_search_chain(struct ocf
 {
 	int status;
 	u16 chain;
+	u32 contig_bits;
 	u64 next_group;
 	struct inode *alloc_inode = ac->ac_inode;
 	struct buffer_head *group_bh = NULL;
@@ -1819,10 +1822,21 @@ static int ocfs2_search_chain(struct ocf
 	status = -ENOSPC;
 	/* for now, the chain search is a bit simplistic. We just use
 	 * the 1st group with any empty bits. */
-	while ((status = ac->ac_group_search(alloc_inode, group_bh,
-					     bits_wanted, min_bits,
-					     ac->ac_max_block,
-					     res)) == -ENOSPC) {
+	while (1) {
+		if (ac->ac_which == OCFS2_AC_USE_MAIN_DISCONTIG) {
+			contig_bits = le16_to_cpu(bg->bg_contig_free_bits);
+			if (!contig_bits)
+				contig_bits = ocfs2_find_max_contig_free_bits(bg->bg_bitmap,
+						le16_to_cpu(bg->bg_bits), 0);
+			if (bits_wanted > contig_bits && contig_bits >= min_bits)
+				bits_wanted = contig_bits;
+		}
+
+		status = ac->ac_group_search(alloc_inode, group_bh,
+				bits_wanted, min_bits,
+				ac->ac_max_block, res);
+		if (status != -ENOSPC)
+			break;
 		if (!bg->bg_next_group)
 			break;
 
@@ -1982,6 +1996,7 @@ static int ocfs2_claim_suballoc_bits(str
 	victim = ocfs2_find_victim_chain(cl);
 	ac->ac_chain = victim;
 
+search:
 	status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits,
 				    res, &bits_left);
 	if (!status) {
@@ -2022,6 +2037,16 @@ static int ocfs2_claim_suballoc_bits(str
 		}
 	}
 
+	/* Chains can't supply the bits_wanted contiguous space.
+	 * We should switch to using every single bit when allocating
+	 * from the global bitmap. */
+	if (i == le16_to_cpu(cl->cl_next_free_rec) &&
+	    status == -ENOSPC && ac->ac_which == OCFS2_AC_USE_MAIN) {
+		ac->ac_which = OCFS2_AC_USE_MAIN_DISCONTIG;
+		ac->ac_chain = victim;
+		goto search;
+	}
+
 set_hint:
 	if (status != -ENOSPC) {
 		/* If the next search of this group is not likely to
@@ -2365,7 +2390,8 @@ int __ocfs2_claim_clusters(handle_t *han
 	BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
 
 	BUG_ON(ac->ac_which != OCFS2_AC_USE_LOCAL
-	       && ac->ac_which != OCFS2_AC_USE_MAIN);
+	       && ac->ac_which != OCFS2_AC_USE_MAIN
+	       && ac->ac_which != OCFS2_AC_USE_MAIN_DISCONTIG);
 
 	if (ac->ac_which == OCFS2_AC_USE_LOCAL) {
 		WARN_ON(min_clusters > 1);
--- a/fs/ocfs2/suballoc.h~ocfs2-fix-the-issue-with-discontiguous-allocation-in-the-global_bitmap
+++ a/fs/ocfs2/suballoc.h
@@ -29,6 +29,7 @@ struct ocfs2_alloc_context {
 #define OCFS2_AC_USE_MAIN  2
 #define OCFS2_AC_USE_INODE 3
 #define OCFS2_AC_USE_META  4
+#define OCFS2_AC_USE_MAIN_DISCONTIG  5
 	u32    ac_which;
 
 	/* these are used by the chain search */
_

Patches currently in -mm which might be from heming.zhao@suse.com are



^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2025-05-08  6:40 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-05-08  6:40 [merged mm-hotfixes-stable] ocfs2-fix-the-issue-with-discontiguous-allocation-in-the-global_bitmap.patch removed from -mm tree Andrew Morton

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox