This is a note to let you know that I've just added the patch titled
smb3: fix unusable share after force unmount failure
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:
smb3-fix-unusable-share-after-force-unmount-failure.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 <stable@xxxxxxxxxxxxxxx> know about it.
>From 491eafce1a51c457701351a4bf40733799745314 Mon Sep 17 00:00:00 2001
From: Steve French <stfrench@xxxxxxxxxxxxx>
Date: Thu, 23 Mar 2023 16:20:02 -0500
Subject: smb3: fix unusable share after force unmount failure
From: Steve French <stfrench@xxxxxxxxxxxxx>
commit 491eafce1a51c457701351a4bf40733799745314 upstream.
If user does forced unmount ("umount -f") while files are still open
on the share (as was seen in a Kubernetes example running on SMB3.1.1
mount) then we were marking the share as "TID_EXITING" in umount_begin()
which caused all subsequent operations (except write) to fail ... but
unfortunately when umount_begin() is called we do not know yet that
there are open files or active references on the share that would prevent
unmount from succeeding. Kubernetes had example when they were doing
umount -f when files were open which caused the share to become
unusable until the files were closed (and the umount retried).
Fix this so that TID_EXITING is not set until we are about to send
the tree disconnect (not at the beginning of forced umounts in
umount_begin) so that if "umount -f" fails (due to open files or
references) the mount is still usable.
Cc: stable@xxxxxxxxxxxxxxx
Reviewed-by: Shyam Prasad N <sprasad@xxxxxxxxxxxxx>
Reviewed-by: Paulo Alcantara (SUSE) <pc@xxxxxxxxxxxxx>
Signed-off-by: Steve French <stfrench@xxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
fs/cifs/cifsfs.c | 9 ++++++---
fs/cifs/cifssmb.c | 6 ++----
fs/cifs/connect.c | 1 +
fs/cifs/smb2pdu.c | 8 ++------
4 files changed, 11 insertions(+), 13 deletions(-)
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -730,13 +730,16 @@ static void cifs_umount_begin(struct sup
spin_lock(&tcon->tc_lock);
if ((tcon->tc_count > 1) || (tcon->status == TID_EXITING)) {
/* we have other mounts to same share or we have
- already tried to force umount this and woken up
+ already tried to umount this and woken up
all waiting network requests, nothing to do */
spin_unlock(&tcon->tc_lock);
spin_unlock(&cifs_tcp_ses_lock);
return;
- } else if (tcon->tc_count == 1)
- tcon->status = TID_EXITING;
+ }
+ /*
+ * can not set tcon->status to TID_EXITING yet since we don't know if umount -f will
+ * fail later (e.g. due to open files). TID_EXITING will be set just before tdis req sent
+ */
spin_unlock(&tcon->tc_lock);
spin_unlock(&cifs_tcp_ses_lock);
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -85,13 +85,11 @@ cifs_reconnect_tcon(struct cifs_tcon *tc
/*
* only tree disconnect, open, and write, (and ulogoff which does not
- * have tcon) are allowed as we start force umount
+ * have tcon) are allowed as we start umount
*/
spin_lock(&tcon->tc_lock);
if (tcon->status == TID_EXITING) {
- if (smb_command != SMB_COM_WRITE_ANDX &&
- smb_command != SMB_COM_OPEN_ANDX &&
- smb_command != SMB_COM_TREE_DISCONNECT) {
+ if (smb_command != SMB_COM_TREE_DISCONNECT) {
spin_unlock(&tcon->tc_lock);
cifs_dbg(FYI, "can not send cmd %d while umounting\n",
smb_command);
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -2365,6 +2365,7 @@ cifs_put_tcon(struct cifs_tcon *tcon)
WARN_ON(tcon->tc_count < 0);
list_del_init(&tcon->tcon_list);
+ tcon->status = TID_EXITING;
spin_unlock(&tcon->tc_lock);
spin_unlock(&cifs_tcp_ses_lock);
--- a/fs/cifs/smb2pdu.c
+++ b/fs/cifs/smb2pdu.c
@@ -225,13 +225,9 @@ smb2_reconnect(__le16 smb2_command, stru
spin_lock(&tcon->tc_lock);
if (tcon->status == TID_EXITING) {
/*
- * only tree disconnect, open, and write,
- * (and ulogoff which does not have tcon)
- * are allowed as we start force umount.
+ * only tree disconnect allowed when disconnecting ...
*/
- if ((smb2_command != SMB2_WRITE) &&
- (smb2_command != SMB2_CREATE) &&
- (smb2_command != SMB2_TREE_DISCONNECT)) {
+ if (smb2_command != SMB2_TREE_DISCONNECT) {
spin_unlock(&tcon->tc_lock);
cifs_dbg(FYI, "can not send cmd %d while umounting\n",
smb2_command);
Patches currently in stable-queue which might be from stfrench@xxxxxxxxxxxxx are
queue-6.1/smb3-lower-default-deferred-close-timeout-to-address-perf-regression.patch
queue-6.1/smb3-fix-unusable-share-after-force-unmount-failure.patch
queue-6.1/ksmbd-add-low-bound-validation-to-fsctl_query_alloca.patch
queue-6.1/ksmbd-fix-possible-refcount-leak-in-smb2_open.patch
queue-6.1/ksmbd-add-low-bound-validation-to-fsctl_set_zero_dat.patch
