Kernel bug in iomap_read_inline_data() fixed by the following patch is hit on older stables 4.19/5.4/5.10.
The patch failed to be initially backported into stable branches older than 5.15 due to the upstream commit 7db354444ad8 ("gfs2: Cosmetic gfs2_dinode_{in,out} cleanup").
Now it can be cleanly applied to the 4.19/5.4/5.10 stable branches.
commit 70376c7ff31221f1d21db5611d8209e677781d3a upstream.
Check if the inode size of stuffed (inline) inodes is within the allowed range when reading inodes from disk (gfs2_dinode_in()). This prevents us from on-disk corruption.
The two checks in stuffed_readpage() and gfs2_unstuffer_page() that just truncate inline data to the maximum allowed size don't actually make sense, and they can be removed now as well.
Reported-by: syzbot+7bb81dfa9cda07d9cd9d@syzkaller.appspotmail.com Signed-off-by: Andreas Gruenbacher agruenba@redhat.com [pchelkin@ispras.ru: adjust the inode variable inside gfs2_dinode_in with the format used before upstream commit 7db354444ad8 ("gfs2: Cosmetic gfs2_dinode_{in,out} cleanup")] Signed-off-by: Fedor Pchelkin pchelkin@ispras.ru --- fs/gfs2/aops.c | 2 -- fs/gfs2/bmap.c | 3 --- fs/gfs2/glops.c | 3 +++ 3 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 530659554870..a0430da033b3 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -451,8 +451,6 @@ static int stuffed_readpage(struct gfs2_inode *ip, struct page *page) return error;
kaddr = kmap_atomic(page); - if (dsize > gfs2_max_stuffed_size(ip)) - dsize = gfs2_max_stuffed_size(ip); memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize); memset(kaddr + dsize, 0, PAGE_SIZE - dsize); kunmap_atomic(kaddr); diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index b4fde3a8eeb4..eaee95d2ad14 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -69,9 +69,6 @@ static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh, void *kaddr = kmap(page); u64 dsize = i_size_read(inode);
- if (dsize > gfs2_max_stuffed_size(ip)) - dsize = gfs2_max_stuffed_size(ip); - memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize); memset(kaddr + dsize, 0, PAGE_SIZE - dsize); kunmap(page); diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c index bf539eab92c6..db28c240dae3 100644 --- a/fs/gfs2/glops.c +++ b/fs/gfs2/glops.c @@ -454,6 +454,9 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) ip->i_depth = (u8)depth; ip->i_entries = be32_to_cpu(str->di_entries);
+ if (gfs2_is_stuffed(ip) && ip->i_inode.i_size > gfs2_max_stuffed_size(ip)) + goto corrupt; + if (S_ISREG(ip->i_inode.i_mode)) gfs2_set_aops(&ip->i_inode);
From: Andreas Gruenbacher agruenba@redhat.com
commit 70376c7ff31221f1d21db5611d8209e677781d3a upstream.
Check if the inode size of stuffed (inline) inodes is within the allowed range when reading inodes from disk (gfs2_dinode_in()). This prevents us from on-disk corruption.
The two checks in stuffed_readpage() and gfs2_unstuffer_page() that just truncate inline data to the maximum allowed size don't actually make sense, and they can be removed now as well.
Reported-by: syzbot+7bb81dfa9cda07d9cd9d@syzkaller.appspotmail.com Signed-off-by: Andreas Gruenbacher agruenba@redhat.com [pchelkin@ispras.ru: adjust the inode variable inside gfs2_dinode_in with the format used before upstream commit 7db354444ad8 ("gfs2: Cosmetic gfs2_dinode_{in,out} cleanup")] Signed-off-by: Fedor Pchelkin pchelkin@ispras.ru --- v2: add missed From: tag
fs/gfs2/aops.c | 2 -- fs/gfs2/bmap.c | 3 --- fs/gfs2/glops.c | 3 +++ 3 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 530659554870..a0430da033b3 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -451,8 +451,6 @@ static int stuffed_readpage(struct gfs2_inode *ip, struct page *page) return error;
kaddr = kmap_atomic(page); - if (dsize > gfs2_max_stuffed_size(ip)) - dsize = gfs2_max_stuffed_size(ip); memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize); memset(kaddr + dsize, 0, PAGE_SIZE - dsize); kunmap_atomic(kaddr); diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index b4fde3a8eeb4..eaee95d2ad14 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -69,9 +69,6 @@ static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh, void *kaddr = kmap(page); u64 dsize = i_size_read(inode);
- if (dsize > gfs2_max_stuffed_size(ip)) - dsize = gfs2_max_stuffed_size(ip); - memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize); memset(kaddr + dsize, 0, PAGE_SIZE - dsize); kunmap(page); diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c index bf539eab92c6..db28c240dae3 100644 --- a/fs/gfs2/glops.c +++ b/fs/gfs2/glops.c @@ -454,6 +454,9 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) ip->i_depth = (u8)depth; ip->i_entries = be32_to_cpu(str->di_entries);
+ if (gfs2_is_stuffed(ip) && ip->i_inode.i_size > gfs2_max_stuffed_size(ip)) + goto corrupt; + if (S_ISREG(ip->i_inode.i_mode)) gfs2_set_aops(&ip->i_inode);
On Fri, Mar 24, 2023 at 11:26:15PM +0300, Fedor Pchelkin wrote:
From: Andreas Gruenbacher agruenba@redhat.com
commit 70376c7ff31221f1d21db5611d8209e677781d3a upstream.
Check if the inode size of stuffed (inline) inodes is within the allowed range when reading inodes from disk (gfs2_dinode_in()). This prevents us from on-disk corruption.
The two checks in stuffed_readpage() and gfs2_unstuffer_page() that just truncate inline data to the maximum allowed size don't actually make sense, and they can be removed now as well.
Reported-by: syzbot+7bb81dfa9cda07d9cd9d@syzkaller.appspotmail.com Signed-off-by: Andreas Gruenbacher agruenba@redhat.com [pchelkin@ispras.ru: adjust the inode variable inside gfs2_dinode_in with the format used before upstream commit 7db354444ad8 ("gfs2: Cosmetic gfs2_dinode_{in,out} cleanup")] Signed-off-by: Fedor Pchelkin pchelkin@ispras.ru
v2: add missed From: tag
Now queued up, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org
-
Fedor Pchelkin -
Greg Kroah-Hartman