Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Jeffle Xu <jefflexu@linux.alibaba.com>
To:     xiang@kernel.org, chao@kernel.org, linux-erofs@lists.ozlabs.org
Cc:     linux-kernel@vger.kernel.org
Subject: [PATCH] erofs: leave compressed inodes unsupported in fscache mode for now
Date:   Wed, 25 May 2022 15:59:50 +0800
Message-Id: <20220525075950.21535-1-jefflexu@linux.alibaba.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

erofs over fscache doesn't support the compressed layout yet. It will
cause NULL crash if there are compressed inodes contained when working
in fscache mode.

So far in the erofs based container image distribution scenarios
(RAFS v6), the compressed images are downloaded to local and then
decompressed as a erofs image of uncompressed layout. Then the erofs
image is mounted in fscache mode and serves as the container image. Thus
the current implementation won't break the container image distribution
scenarios.

The fscache support for the compressed layout is still under
development. Anyway, to avoid the potential crash, let's leave the
compressed inodes unsupported in fscache mode until we support it later.

Fixes: 1442b02b66ad ("erofs: implement fscache-based data read for non-inline layout")
Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
---
 fs/erofs/inode.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index bcc8335b46b3..95a403720e8c 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -288,7 +288,10 @@ static int erofs_fill_inode(struct inode *inode, int isdir)
 	}
 
 	if (erofs_inode_is_data_compressed(vi->datalayout)) {
-		err = z_erofs_fill_inode(inode);
+		if (!erofs_is_fscache_mode(inode->i_sb))
+			err = z_erofs_fill_inode(inode);
+		else
+			err = -EOPNOTSUPP;
 		goto out_unlock;
 	}
 	inode->i_mapping->a_ops = &erofs_raw_access_aops;
-- 
2.27.0