[4.14.y][PATCH 1/2] kernfs: deal with kernfs_fill_super() failures

[4.14.y][PATCH 1/2] kernfs: deal with kernfs_fill_super() failures

[Guilherme G. Piccoli]

From: Al Viro <viro@zeniv.linux.org.uk>

commit 82382acec0c97b91830fff7130d0acce4ac4f3f3 upstream.

make sure that info->node is initialized early, so that kernfs_kill_sb()
can list_del() it safely.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Guilherme G. Piccoli <gpiccoli@canonical.com>
---

Hey Al, is there any reason for the absence of this patch in the stable
kernels? We had a report of a crash (NULL-ptr dereference) that seems to be
fixed by this patch - if there isn't a reason, I'd like to propose this one
to be merged on 4.14.y . I've build-tested in x86-64 with defconfig.

Thanks in advance,


Guilherme


 fs/kernfs/mount.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/kernfs/mount.c b/fs/kernfs/mount.c
index 5019058e0f6a..610267585f8f 100644
--- a/fs/kernfs/mount.c
+++ b/fs/kernfs/mount.c
@@ -320,6 +320,7 @@ struct dentry *kernfs_mount_ns(struct file_system_type *fs_type, int flags,
 
 	info->root = root;
 	info->ns = ns;
+	INIT_LIST_HEAD(&info->node);
 
 	sb = sget_userns(fs_type, kernfs_test_super, kernfs_set_super, flags,
 			 &init_user_ns, info);
-- 
2.31.1

[4.14.y][PATCH 2/2] unfuck sysfs_mount()

[Guilherme G. Piccoli]

From: Al Viro <viro@zeniv.linux.org.uk>

commit 7b745a4e4051e1bbce40e0b1c2cf636c70583aa4 upstream.

new_sb is left uninitialized in case of early failures in kernfs_mount_ns(),
and while IS_ERR(root) is true in all such cases, using IS_ERR(root) || !new_sb
is not a solution - IS_ERR(root) is true in some cases when new_sb is true.

Make sure new_sb is initialized (and matches the reality) in all cases and
fix the condition for dropping kobj reference - we want it done precisely
in those situations where the reference has not been transferred into a new
super_block instance.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Guilherme G. Piccoli <gpiccoli@canonical.com>
---

I'd like to protest this patch title heheh
But I think it's better to keep consistency with upstream. It's the same
case as patch 1 of the series, no clear reason for its absence in stable.
Build-tested on x86-64 with defconfig.

Thanks,


Guilherme


 fs/sysfs/mount.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/sysfs/mount.c b/fs/sysfs/mount.c
index 20b8f82e115b..2bbe84d9c0a8 100644
--- a/fs/sysfs/mount.c
+++ b/fs/sysfs/mount.c
@@ -28,7 +28,7 @@ static struct dentry *sysfs_mount(struct file_system_type *fs_type,
 {
 	struct dentry *root;
 	void *ns;
-	bool new_sb;
+	bool new_sb = false;
 
 	if (!(flags & MS_KERNMOUNT)) {
 		if (!kobj_ns_current_may_mount(KOBJ_NS_TYPE_NET))
@@ -38,9 +38,9 @@ static struct dentry *sysfs_mount(struct file_system_type *fs_type,
 	ns = kobj_ns_grab_current(KOBJ_NS_TYPE_NET);
 	root = kernfs_mount_ns(fs_type, flags, sysfs_root,
 				SYSFS_MAGIC, &new_sb, ns);
-	if (IS_ERR(root) || !new_sb)
+	if (!new_sb)
 		kobj_ns_drop(KOBJ_NS_TYPE_NET, ns);
-	else if (new_sb)
+	else if (!IS_ERR(root))
 		root->d_sb->s_iflags |= SB_I_USERNS_VISIBLE;
 
 	return root;
-- 
2.31.1

Re: [4.14.y][PATCH 2/2] unfuck sysfs_mount()

[Greg KH]

On Tue, Jun 22, 2021 at 06:06:22PM -0300, Guilherme G. Piccoli wrote:
> From: Al Viro <viro@zeniv.linux.org.uk>
> 
> commit 7b745a4e4051e1bbce40e0b1c2cf636c70583aa4 upstream.
> 
> new_sb is left uninitialized in case of early failures in kernfs_mount_ns(),
> and while IS_ERR(root) is true in all such cases, using IS_ERR(root) || !new_sb
> is not a solution - IS_ERR(root) is true in some cases when new_sb is true.
> 
> Make sure new_sb is initialized (and matches the reality) in all cases and
> fix the condition for dropping kobj reference - we want it done precisely
> in those situations where the reference has not been transferred into a new
> super_block instance.
> 
> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> Signed-off-by: Guilherme G. Piccoli <gpiccoli@canonical.com>
> ---
> 
> I'd like to protest this patch title heheh
> But I think it's better to keep consistency with upstream. It's the same
> case as patch 1 of the series, no clear reason for its absence in stable.
> Build-tested on x86-64 with defconfig.

Both now queued up, thanks.

greg k-h

Re: [4.14.y][PATCH 2/2] unfuck sysfs_mount()

[Guilherme Piccoli]

Thanks a lot Greg!