]> git.ipfire.org Git - thirdparty/kernel/linux.git/commitdiff
xfs: make the RT rsum_cache mandatory
authorChristoph Hellwig <hch@lst.de>
Fri, 30 Aug 2024 22:36:52 +0000 (15:36 -0700)
committerDarrick J. Wong <djwong@kernel.org>
Sun, 1 Sep 2024 15:58:19 +0000 (08:58 -0700)
Currently the RT mount code simply ignores an allocation failure for the
rsum_cache.  The code mostly works fine with it, but not having it leads
to nasty corner cases in the growfs code that we don't really handle
well.  Switch to failing the mount if we can't allocate the memory, the
file system would not exactly be useful in such a constrained environment
to start with.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
fs/xfs/xfs_rtalloc.c

index d28395abdd02914d97ac6870cc620d9c83b3f525..26eab1b408c8ad9cfe147e0df93680029114f2d9 100644 (file)
@@ -767,21 +767,20 @@ out_trans_cancel:
        return error;
 }
 
-static void
+static int
 xfs_alloc_rsum_cache(
-       xfs_mount_t     *mp,            /* file system mount structure */
-       xfs_extlen_t    rbmblocks)      /* number of rt bitmap blocks */
+       struct xfs_mount        *mp,
+       xfs_extlen_t            rbmblocks)
 {
        /*
         * The rsum cache is initialized to the maximum value, which is
         * trivially an upper bound on the maximum level with any free extents.
-        * We can continue without the cache if it couldn't be allocated.
         */
        mp->m_rsum_cache = kvmalloc(rbmblocks, GFP_KERNEL);
-       if (mp->m_rsum_cache)
-               memset(mp->m_rsum_cache, -1, rbmblocks);
-       else
-               xfs_warn(mp, "could not allocate realtime summary cache");
+       if (!mp->m_rsum_cache)
+               return -ENOMEM;
+       memset(mp->m_rsum_cache, -1, rbmblocks);
+       return 0;
 }
 
 /*
@@ -939,8 +938,11 @@ xfs_growfs_rt(
                goto out_unlock;
 
        rsum_cache = mp->m_rsum_cache;
-       if (nrbmblocks != sbp->sb_rbmblocks)
-               xfs_alloc_rsum_cache(mp, nrbmblocks);
+       if (nrbmblocks != sbp->sb_rbmblocks) {
+               error = xfs_alloc_rsum_cache(mp, nrbmblocks);
+               if (error)
+                       goto out_unlock;
+       }
 
        /*
         * Allocate a new (fake) mount/sb.
@@ -1268,7 +1270,9 @@ xfs_rtmount_inodes(
        if (error)
                goto out_rele_summary;
 
-       xfs_alloc_rsum_cache(mp, sbp->sb_rbmblocks);
+       error = xfs_alloc_rsum_cache(mp, sbp->sb_rbmblocks);
+       if (error)
+               goto out_rele_summary;
        return 0;
 
 out_rele_summary: