Patch "xfs: add lock protection when remove perag from radix tree" has been added to the 6.1-stable tree

This is a note to let you know that I've just added the patch titled

    xfs: add lock protection when remove perag from radix tree

to the 6.1-stable tree which can be found at:
    http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     xfs-add-lock-protection-when-remove-perag-from-radix-tree.patch
and it can be found in the queue-6.1 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <[email protected]> know about it.


From [email protected] Thu Mar 13 21:26:45 2025
From: Leah Rumancik <[email protected]>
Date: Thu, 13 Mar 2025 13:25:44 -0700
Subject: xfs: add lock protection when remove perag from radix tree
To: [email protected]
Cc: [email protected], Long Li <[email protected]>, Christoph Hellwig <[email protected]>, "Darrick J. Wong" <[email protected]>, Chandan Babu R <[email protected]>, Catherine Hoang <[email protected]>, Greg Kroah-Hartman <[email protected]>, Leah Rumancik <[email protected]>
Message-ID: <[email protected]>

From: Long Li <[email protected]>

[ Upstream commit 07afd3173d0c6d24a47441839a835955ec6cf0d4 ]

[ 6.1: resolved conflict in xfs_ag.c ]

Take mp->m_perag_lock for deletions from the perag radix tree in
xfs_initialize_perag to prevent racing with tagging operations.
Lookups are fine - they are RCU protected so already deal with the
tree changing shape underneath the lookup - but tagging operations
require the tree to be stable while the tags are propagated back up
to the root.

Right now there's nothing stopping radix tree tagging from operating
while a growfs operation is progress and adding/removing new entries
into the radix tree.

Hence we can have traversals that require a stable tree occurring at
the same time we are removing unused entries from the radix tree which
causes the shape of the tree to change.

Likely this hasn't caused a problem in the past because we are only
doing append addition and removal so the active AG part of the tree
is not changing shape, but that doesn't mean it is safe. Just making
the radix tree modifications serialise against each other is obviously
correct.

Signed-off-by: Long Li <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Reviewed-by: "Darrick J. Wong" <[email protected]>
Signed-off-by: Chandan Babu R <[email protected]>
Signed-off-by: Catherine Hoang <[email protected]>
Acked-by: Darrick J. Wong <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Signed-off-by: Leah Rumancik <[email protected]>
Acked-by: "Darrick J. Wong" <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
---
 fs/xfs/libxfs/xfs_ag.c |    4 ++++
 1 file changed, 4 insertions(+)

--- a/fs/xfs/libxfs/xfs_ag.c
+++ b/fs/xfs/libxfs/xfs_ag.c
@@ -345,13 +345,17 @@ xfs_initialize_perag(
 	return 0;
 
 out_remove_pag:
+	spin_lock(&mp->m_perag_lock);
 	radix_tree_delete(&mp->m_perag_tree, index);
+	spin_unlock(&mp->m_perag_lock);
 out_free_pag:
 	kmem_free(pag);
 out_unwind_new_pags:
 	/* unwind any prior newly initialized pags */
 	for (index = first_initialised; index < agcount; index++) {
+		spin_lock(&mp->m_perag_lock);
 		pag = radix_tree_delete(&mp->m_perag_tree, index);
+		spin_unlock(&mp->m_perag_lock);
 		if (!pag)
 			break;
 		xfs_buf_hash_destroy(pag);


Patches currently in stable-queue which might be from [email protected] are

queue-6.1/xfs-fix-confusing-xfs_extent_item-variable-names.patch
queue-6.1/xfs-fix-32-bit-truncation-in-xfs_compute_rextslog.patch
queue-6.1/xfs-transfer-recovered-intent-item-ownership-in-iop_recover.patch
queue-6.1/xfs-initialise-di_crc-in-xfs_log_dinode.patch
queue-6.1/xfs-consider-minlen-sized-extents-in-xfs_rtallocate_extent_block.patch
queue-6.1/xfs-don-t-leak-recovered-attri-intent-items.patch
queue-6.1/xfs-remove-unused-fields-from-struct-xbtree_ifakeroot.patch
queue-6.1/xfs-fix-bounds-check-in-xfs_defer_agfl_block.patch
queue-6.1/xfs-ensure-logflagsp-is-initialized-in-xfs_bmap_del_extent_real.patch
queue-6.1/xfs-convert-rt-bitmap-extent-lengths-to-xfs_rtbxlen_t.patch
queue-6.1/xfs-pass-refcount-intent-directly-through-the-log-intent-code.patch
queue-6.1/xfs-fix-perag-leak-when-growfs-fails.patch
queue-6.1/xfs-pass-the-xfs_defer_pending-object-to-iop_recover.patch
queue-6.1/xfs-update-dir3-leaf-block-metadata-after-swap.patch
queue-6.1/xfs-use-deferred-frees-for-btree-block-freeing.patch
queue-6.1/xfs-make-rextslog-computation-consistent-with-mkfs.patch
queue-6.1/xfs-pass-xfs_extent_free_item-directly-through-the-log-intent-code.patch
queue-6.1/xfs-move-the-xfs_rtbitmap.c-declarations-to-xfs_rtbitmap.h.patch
queue-6.1/xfs-recompute-growfsrtfree-transaction-reservation-while-growing-rt-volume.patch
queue-6.1/xfs-reserve-less-log-space-when-recovering-log-intent-items.patch
queue-6.1/xfs-pass-the-xfs_bmbt_irec-directly-through-the-log-intent-code.patch
queue-6.1/xfs-force-all-buffers-to-be-written-during-btree-bulk-load.patch
queue-6.1/xfs-reset-xfs_attr_incomplete-filter-on-node-removal.patch
queue-6.1/xfs-add-lock-protection-when-remove-perag-from-radix-tree.patch
queue-6.1/xfs-use-xfs_defer_pending-objects-to-recover-intent-items.patch
queue-6.1/xfs-pass-per-ag-references-to-xfs_free_extent.patch
queue-6.1/xfs-validate-block-number-being-freed-before-adding-to-xefi.patch
queue-6.1/xfs-don-t-allow-overly-small-or-large-realtime-volumes.patch
queue-6.1/xfs-remove-conditional-building-of-rt-geometry-validator-functions.patch