These patches are backport patches to support directory(v2) lease and additional bug fixes for linux-5.15.y.
Namjae Jeon (8): ksmbd: add support for key exchange ksmbd: set epoch in create context v2 lease ksmbd: set v2 lease capability ksmbd: downgrade RWH lease caching state to RH for directory ksmbd: send v2 lease break notification for directory ksmbd: lazy v2 lease break on smb2_write() ksmbd: avoid duplicate opinfo_put() call on error of smb21_lease_break_ack() ksmbd: fix wrong allocation size update in smb2_open()
fs/Kconfig | 4 +- fs/ksmbd/oplock.c | 115 ++++++++++++++++++++++++++++++++++++++----- fs/ksmbd/oplock.h | 8 ++- fs/ksmbd/smb2ops.c | 9 ++-- fs/ksmbd/smb2pdu.c | 61 +++++++++++++---------- fs/ksmbd/smb2pdu.h | 1 + fs/ksmbd/vfs.c | 3 ++ fs/ksmbd/vfs_cache.c | 13 ++++- fs/ksmbd/vfs_cache.h | 3 ++ 9 files changed, 171 insertions(+), 46 deletions(-)
[ Upstream commit f9929ef6a2a55f03aac61248c6a3a987b8546f2a ]
When mounting cifs client, can see the following warning message.
CIFS: decode_ntlmssp_challenge: authentication has been weakened as server does not support key exchange
To remove this warning message, Add support for key exchange feature to ksmbd. This patch decrypts 16-byte ciphertext value sent by the client using RC4 with session key. The decrypted value is the recovered secondary key that will use instead of the session key for signing and sealing.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/Kconfig b/fs/Kconfig index a6313a969bc5..971339ecc1a2 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -369,8 +369,8 @@ source "fs/ksmbd/Kconfig"
config SMBFS_COMMON tristate - default y if CIFS=y - default m if CIFS=m + default y if CIFS=y || SMB_SERVER=y + default m if CIFS=m || SMB_SERVER=m
source "fs/coda/Kconfig" source "fs/afs/Kconfig"
On Tue, Dec 26, 2023 at 07:53:26PM +0900, Namjae Jeon wrote:
[ Upstream commit f9929ef6a2a55f03aac61248c6a3a987b8546f2a ]
When mounting cifs client, can see the following warning message.
CIFS: decode_ntlmssp_challenge: authentication has been weakened as server does not support key exchange
To remove this warning message, Add support for key exchange feature to ksmbd. This patch decrypts 16-byte ciphertext value sent by the client using RC4 with session key. The decrypted value is the recovered secondary key that will use instead of the session key for signing and sealing.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com
fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/Kconfig b/fs/Kconfig index a6313a969bc5..971339ecc1a2 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -369,8 +369,8 @@ source "fs/ksmbd/Kconfig"
config SMBFS_COMMON tristate
- default y if CIFS=y
- default m if CIFS=m
- default y if CIFS=y || SMB_SERVER=y
- default m if CIFS=m || SMB_SERVER=m
source "fs/coda/Kconfig" source "fs/afs/Kconfig"
This looks really weird: the hunk above is in the original upstream patch, but what happened to the rest of the upstream code?
This change doesn't do what the message describing it says it does.
2023-12-26 23:59 GMT+09:00, Sasha Levin sashal@kernel.org:
On Tue, Dec 26, 2023 at 07:53:26PM +0900, Namjae Jeon wrote:
[ Upstream commit f9929ef6a2a55f03aac61248c6a3a987b8546f2a ]
When mounting cifs client, can see the following warning message.
CIFS: decode_ntlmssp_challenge: authentication has been weakened as server does not support key exchange
To remove this warning message, Add support for key exchange feature to ksmbd. This patch decrypts 16-byte ciphertext value sent by the client using RC4 with session key. The decrypted value is the recovered secondary key that will use instead of the session key for signing and sealing.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com
fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/Kconfig b/fs/Kconfig index a6313a969bc5..971339ecc1a2 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -369,8 +369,8 @@ source "fs/ksmbd/Kconfig"
config SMBFS_COMMON tristate
- default y if CIFS=y
- default m if CIFS=m
- default y if CIFS=y || SMB_SERVER=y
- default m if CIFS=m || SMB_SERVER=m
source "fs/coda/Kconfig" source "fs/afs/Kconfig"
This looks really weird: the hunk above is in the original upstream patch, but what happened to the rest of the upstream code?
This change doesn't do what the message describing it says it does.
There was a problem(omitted some changes) in the previous backport patch, I didn't know what to do, so I just sent a patch like this. Should I add it again after reverting the patch or just updating the patch description?
-- Thanks, Sasha
On Wed, Dec 27, 2023 at 05:05:10AM +0900, Namjae Jeon wrote:
2023-12-26 23:59 GMT+09:00, Sasha Levin sashal@kernel.org:
On Tue, Dec 26, 2023 at 07:53:26PM +0900, Namjae Jeon wrote:
[ Upstream commit f9929ef6a2a55f03aac61248c6a3a987b8546f2a ]
When mounting cifs client, can see the following warning message.
CIFS: decode_ntlmssp_challenge: authentication has been weakened as server does not support key exchange
To remove this warning message, Add support for key exchange feature to ksmbd. This patch decrypts 16-byte ciphertext value sent by the client using RC4 with session key. The decrypted value is the recovered secondary key that will use instead of the session key for signing and sealing.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com
fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/Kconfig b/fs/Kconfig index a6313a969bc5..971339ecc1a2 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -369,8 +369,8 @@ source "fs/ksmbd/Kconfig"
config SMBFS_COMMON tristate
- default y if CIFS=y
- default m if CIFS=m
- default y if CIFS=y || SMB_SERVER=y
- default m if CIFS=m || SMB_SERVER=m
source "fs/coda/Kconfig" source "fs/afs/Kconfig"
This looks really weird: the hunk above is in the original upstream patch, but what happened to the rest of the upstream code?
This change doesn't do what the message describing it says it does.
There was a problem(omitted some changes) in the previous backport patch, I didn't know what to do, so I just sent a patch like this. Should I add it again after reverting the patch or just updating the patch description?
Given that this was due to an issue with the backport, I'd say just write a new commit message explaining what happened, and point the fixes tag to c5049d2d73b2 ("ksmbd: add support for key exchange")
2023-12-27 8:01 GMT+09:00, Sasha Levin sashal@kernel.org:
On Wed, Dec 27, 2023 at 05:05:10AM +0900, Namjae Jeon wrote:
2023-12-26 23:59 GMT+09:00, Sasha Levin sashal@kernel.org:
On Tue, Dec 26, 2023 at 07:53:26PM +0900, Namjae Jeon wrote:
[ Upstream commit f9929ef6a2a55f03aac61248c6a3a987b8546f2a ]
When mounting cifs client, can see the following warning message.
CIFS: decode_ntlmssp_challenge: authentication has been weakened as server does not support key exchange
To remove this warning message, Add support for key exchange feature to ksmbd. This patch decrypts 16-byte ciphertext value sent by the client using RC4 with session key. The decrypted value is the recovered secondary key that will use instead of the session key for signing and sealing.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com
fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/Kconfig b/fs/Kconfig index a6313a969bc5..971339ecc1a2 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -369,8 +369,8 @@ source "fs/ksmbd/Kconfig"
config SMBFS_COMMON tristate
- default y if CIFS=y
- default m if CIFS=m
- default y if CIFS=y || SMB_SERVER=y
- default m if CIFS=m || SMB_SERVER=m
source "fs/coda/Kconfig" source "fs/afs/Kconfig"
This looks really weird: the hunk above is in the original upstream patch, but what happened to the rest of the upstream code?
This change doesn't do what the message describing it says it does.
There was a problem(omitted some changes) in the previous backport patch, I didn't know what to do, so I just sent a patch like this. Should I add it again after reverting the patch or just updating the patch description?
Given that this was due to an issue with the backport, I'd say just write a new commit message explaining what happened, and point the fixes tag to c5049d2d73b2 ("ksmbd: add support for key exchange")
Okay, I will do that, Thanks for your answer!
-- Thanks, Sasha
[ Upstream commit d045850b628aaf931fc776c90feaf824dca5a1cf ]
To support v2 lease(directory lease), ksmbd set epoch in create context v2 lease response.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/oplock.c | 5 ++++- fs/ksmbd/oplock.h | 1 + 2 files changed, 5 insertions(+), 1 deletion(-)
diff --git a/fs/ksmbd/oplock.c b/fs/ksmbd/oplock.c index 1cf2d2a3746a..ed639b7b6509 100644 --- a/fs/ksmbd/oplock.c +++ b/fs/ksmbd/oplock.c @@ -104,7 +104,7 @@ static int alloc_lease(struct oplock_info *opinfo, struct lease_ctx_info *lctx) lease->duration = lctx->duration; memcpy(lease->parent_lease_key, lctx->parent_lease_key, SMB2_LEASE_KEY_SIZE); lease->version = lctx->version; - lease->epoch = 0; + lease->epoch = le16_to_cpu(lctx->epoch); INIT_LIST_HEAD(&opinfo->lease_entry); opinfo->o_lease = lease;
@@ -1032,6 +1032,7 @@ static void copy_lease(struct oplock_info *op1, struct oplock_info *op2) SMB2_LEASE_KEY_SIZE); lease2->duration = lease1->duration; lease2->flags = lease1->flags; + lease2->epoch = lease1->epoch++; }
static int add_lease_global_list(struct oplock_info *opinfo) @@ -1364,6 +1365,7 @@ void create_lease_buf(u8 *rbuf, struct lease *lease) memcpy(buf->lcontext.LeaseKey, lease->lease_key, SMB2_LEASE_KEY_SIZE); buf->lcontext.LeaseFlags = lease->flags; + buf->lcontext.Epoch = cpu_to_le16(++lease->epoch); buf->lcontext.LeaseState = lease->state; memcpy(buf->lcontext.ParentLeaseKey, lease->parent_lease_key, SMB2_LEASE_KEY_SIZE); @@ -1423,6 +1425,7 @@ struct lease_ctx_info *parse_lease_state(void *open_req) memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE); lreq->req_state = lc->lcontext.LeaseState; lreq->flags = lc->lcontext.LeaseFlags; + lreq->epoch = lc->lcontext.Epoch; lreq->duration = lc->lcontext.LeaseDuration; memcpy(lreq->parent_lease_key, lc->lcontext.ParentLeaseKey, SMB2_LEASE_KEY_SIZE); diff --git a/fs/ksmbd/oplock.h b/fs/ksmbd/oplock.h index 4b0fe6da7694..ad31439c61fe 100644 --- a/fs/ksmbd/oplock.h +++ b/fs/ksmbd/oplock.h @@ -34,6 +34,7 @@ struct lease_ctx_info { __le32 flags; __le64 duration; __u8 parent_lease_key[SMB2_LEASE_KEY_SIZE]; + __le16 epoch; int version; };
[ Upstream commit 18dd1c367c31d0a060f737d48345747662369b64 ]
Set SMB2_GLOBAL_CAP_DIRECTORY_LEASING to ->capabilities to inform server support directory lease to client.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/oplock.c | 4 ---- fs/ksmbd/smb2ops.c | 9 ++++++--- 2 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/fs/ksmbd/oplock.c b/fs/ksmbd/oplock.c index ed639b7b6509..24716dbe7108 100644 --- a/fs/ksmbd/oplock.c +++ b/fs/ksmbd/oplock.c @@ -1105,10 +1105,6 @@ int smb_grant_oplock(struct ksmbd_work *work, int req_op_level, u64 pid, bool prev_op_has_lease; __le32 prev_op_state = 0;
- /* not support directory lease */ - if (S_ISDIR(file_inode(fp->filp)->i_mode)) - return 0; - opinfo = alloc_opinfo(work, pid, tid); if (!opinfo) return -ENOMEM; diff --git a/fs/ksmbd/smb2ops.c b/fs/ksmbd/smb2ops.c index 9138e1c29b22..c69943d96565 100644 --- a/fs/ksmbd/smb2ops.c +++ b/fs/ksmbd/smb2ops.c @@ -222,7 +222,8 @@ void init_smb3_0_server(struct ksmbd_conn *conn) conn->signing_algorithm = SIGNING_ALG_AES_CMAC;
if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES) - conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING; + conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING | + SMB2_GLOBAL_CAP_DIRECTORY_LEASING;
if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION && conn->cli_cap & SMB2_GLOBAL_CAP_ENCRYPTION) @@ -246,7 +247,8 @@ void init_smb3_02_server(struct ksmbd_conn *conn) conn->signing_algorithm = SIGNING_ALG_AES_CMAC;
if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES) - conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING; + conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING | + SMB2_GLOBAL_CAP_DIRECTORY_LEASING;
if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION || (!(server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION_OFF) && @@ -271,7 +273,8 @@ int init_smb3_11_server(struct ksmbd_conn *conn) conn->signing_algorithm = SIGNING_ALG_AES_CMAC;
if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES) - conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING; + conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING | + SMB2_GLOBAL_CAP_DIRECTORY_LEASING;
if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION || (!(server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION_OFF) &&
[ Upstream commit eb547407f3572d2110cb1194ecd8865b3371a7a4 ]
RWH(Read + Write + Handle) caching state is not supported for directory. ksmbd downgrade it to RH for directory if client send RWH caching lease state.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/oplock.c | 9 +++++++-- fs/ksmbd/oplock.h | 2 +- fs/ksmbd/smb2pdu.c | 8 ++++---- 3 files changed, 12 insertions(+), 7 deletions(-)
diff --git a/fs/ksmbd/oplock.c b/fs/ksmbd/oplock.c index 24716dbe7108..600312e2c6c2 100644 --- a/fs/ksmbd/oplock.c +++ b/fs/ksmbd/oplock.c @@ -1398,10 +1398,11 @@ void create_lease_buf(u8 *rbuf, struct lease *lease) /** * parse_lease_state() - parse lease context containted in file open request * @open_req: buffer containing smb2 file open(create) request + * @is_dir: whether leasing file is directory * * Return: oplock state, -ENOENT if create lease context not found */ -struct lease_ctx_info *parse_lease_state(void *open_req) +struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir) { struct create_context *cc; struct smb2_create_req *req = (struct smb2_create_req *)open_req; @@ -1419,7 +1420,11 @@ struct lease_ctx_info *parse_lease_state(void *open_req) struct create_lease_v2 *lc = (struct create_lease_v2 *)cc;
memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE); - lreq->req_state = lc->lcontext.LeaseState; + if (is_dir) + lreq->req_state = lc->lcontext.LeaseState & + ~SMB2_LEASE_WRITE_CACHING_LE; + else + lreq->req_state = lc->lcontext.LeaseState; lreq->flags = lc->lcontext.LeaseFlags; lreq->epoch = lc->lcontext.Epoch; lreq->duration = lc->lcontext.LeaseDuration; diff --git a/fs/ksmbd/oplock.h b/fs/ksmbd/oplock.h index ad31439c61fe..672127318c75 100644 --- a/fs/ksmbd/oplock.h +++ b/fs/ksmbd/oplock.h @@ -109,7 +109,7 @@ void opinfo_put(struct oplock_info *opinfo);
/* Lease related functions */ void create_lease_buf(u8 *rbuf, struct lease *lease); -struct lease_ctx_info *parse_lease_state(void *open_req); +struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir); __u8 smb2_map_lease_to_oplock(__le32 lease_state); int lease_read_to_write(struct oplock_info *opinfo);
diff --git a/fs/ksmbd/smb2pdu.c b/fs/ksmbd/smb2pdu.c index 7ce5746f9167..6a3cd6ea3af1 100644 --- a/fs/ksmbd/smb2pdu.c +++ b/fs/ksmbd/smb2pdu.c @@ -2729,10 +2729,6 @@ int smb2_open(struct ksmbd_work *work) } }
- req_op_level = req->RequestedOplockLevel; - if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) - lc = parse_lease_state(req); - if (le32_to_cpu(req->ImpersonationLevel) > le32_to_cpu(IL_DELEGATE_LE)) { pr_err("Invalid impersonationlevel : 0x%x\n", le32_to_cpu(req->ImpersonationLevel)); @@ -3212,6 +3208,10 @@ int smb2_open(struct ksmbd_work *work) need_truncate = 1; }
+ req_op_level = req->RequestedOplockLevel; + if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) + lc = parse_lease_state(req, S_ISDIR(file_inode(filp)->i_mode)); + share_ret = ksmbd_smb_check_shared_mode(fp->filp, fp); if (!test_share_config_flag(work->tcon->share_conf, KSMBD_SHARE_FLAG_OPLOCKS) || (req_op_level == SMB2_OPLOCK_LEVEL_LEASE &&
[ Upstream commit d47d9886aeef79feba7adac701a510d65f3682b5 ]
If client send different parent key, different client guid, or there is no parent lease key flags in create context v2 lease, ksmbd send lease break to client.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/oplock.c | 56 ++++++++++++++++++++++++++++++++++++++++---- fs/ksmbd/oplock.h | 4 ++++ fs/ksmbd/smb2pdu.c | 7 ++++++ fs/ksmbd/smb2pdu.h | 1 + fs/ksmbd/vfs_cache.c | 13 +++++++++- fs/ksmbd/vfs_cache.h | 2 ++ 6 files changed, 77 insertions(+), 6 deletions(-)
diff --git a/fs/ksmbd/oplock.c b/fs/ksmbd/oplock.c index 600312e2c6c2..df6f00f58f19 100644 --- a/fs/ksmbd/oplock.c +++ b/fs/ksmbd/oplock.c @@ -102,6 +102,7 @@ static int alloc_lease(struct oplock_info *opinfo, struct lease_ctx_info *lctx) lease->new_state = 0; lease->flags = lctx->flags; lease->duration = lctx->duration; + lease->is_dir = lctx->is_dir; memcpy(lease->parent_lease_key, lctx->parent_lease_key, SMB2_LEASE_KEY_SIZE); lease->version = lctx->version; lease->epoch = le16_to_cpu(lctx->epoch); @@ -543,12 +544,13 @@ static struct oplock_info *same_client_has_lease(struct ksmbd_inode *ci, /* upgrading lease */ if ((atomic_read(&ci->op_count) + atomic_read(&ci->sop_count)) == 1) { - if (lease->state == - (lctx->req_state & lease->state)) { + if (lease->state != SMB2_LEASE_NONE_LE && + lease->state == (lctx->req_state & lease->state)) { lease->state |= lctx->req_state; if (lctx->req_state & SMB2_LEASE_WRITE_CACHING_LE) lease_read_to_write(opinfo); + } } else if ((atomic_read(&ci->op_count) + atomic_read(&ci->sop_count)) > 1) { @@ -900,7 +902,8 @@ static int oplock_break(struct oplock_info *brk_opinfo, int req_op_level) lease->new_state = SMB2_LEASE_READ_CACHING_LE; } else { - if (lease->state & SMB2_LEASE_HANDLE_CACHING_LE) + if (lease->state & SMB2_LEASE_HANDLE_CACHING_LE && + !lease->is_dir) lease->new_state = SMB2_LEASE_READ_CACHING_LE; else @@ -1082,6 +1085,48 @@ static void set_oplock_level(struct oplock_info *opinfo, int level, } }
+void smb_send_parent_lease_break_noti(struct ksmbd_file *fp, + struct lease_ctx_info *lctx) +{ + struct oplock_info *opinfo; + struct ksmbd_inode *p_ci = NULL; + + if (lctx->version != 2) + return; + + p_ci = ksmbd_inode_lookup_lock(fp->filp->f_path.dentry->d_parent); + if (!p_ci) + return; + + read_lock(&p_ci->m_lock); + list_for_each_entry(opinfo, &p_ci->m_op_list, op_entry) { + if (!opinfo->is_lease) + continue; + + if (opinfo->o_lease->state != SMB2_OPLOCK_LEVEL_NONE && + (!(lctx->flags & SMB2_LEASE_FLAG_PARENT_LEASE_KEY_SET_LE) || + !compare_guid_key(opinfo, fp->conn->ClientGUID, + lctx->parent_lease_key))) { + if (!atomic_inc_not_zero(&opinfo->refcount)) + continue; + + atomic_inc(&opinfo->conn->r_count); + if (ksmbd_conn_releasing(opinfo->conn)) { + atomic_dec(&opinfo->conn->r_count); + continue; + } + + read_unlock(&p_ci->m_lock); + oplock_break(opinfo, SMB2_OPLOCK_LEVEL_NONE); + opinfo_conn_put(opinfo); + read_lock(&p_ci->m_lock); + } + } + read_unlock(&p_ci->m_lock); + + ksmbd_inode_put(p_ci); +} + /** * smb_grant_oplock() - handle oplock/lease request on file open * @work: smb work @@ -1420,10 +1465,11 @@ struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir) struct create_lease_v2 *lc = (struct create_lease_v2 *)cc;
memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE); - if (is_dir) + if (is_dir) { lreq->req_state = lc->lcontext.LeaseState & ~SMB2_LEASE_WRITE_CACHING_LE; - else + lreq->is_dir = true; + } else lreq->req_state = lc->lcontext.LeaseState; lreq->flags = lc->lcontext.LeaseFlags; lreq->epoch = lc->lcontext.Epoch; diff --git a/fs/ksmbd/oplock.h b/fs/ksmbd/oplock.h index 672127318c75..b64d1536882a 100644 --- a/fs/ksmbd/oplock.h +++ b/fs/ksmbd/oplock.h @@ -36,6 +36,7 @@ struct lease_ctx_info { __u8 parent_lease_key[SMB2_LEASE_KEY_SIZE]; __le16 epoch; int version; + bool is_dir; };
struct lease_table { @@ -54,6 +55,7 @@ struct lease { __u8 parent_lease_key[SMB2_LEASE_KEY_SIZE]; int version; unsigned short epoch; + bool is_dir; struct lease_table *l_lb; };
@@ -125,4 +127,6 @@ struct oplock_info *lookup_lease_in_table(struct ksmbd_conn *conn, int find_same_lease_key(struct ksmbd_session *sess, struct ksmbd_inode *ci, struct lease_ctx_info *lctx); void destroy_lease_table(struct ksmbd_conn *conn); +void smb_send_parent_lease_break_noti(struct ksmbd_file *fp, + struct lease_ctx_info *lctx); #endif /* __KSMBD_OPLOCK_H */ diff --git a/fs/ksmbd/smb2pdu.c b/fs/ksmbd/smb2pdu.c index 6a3cd6ea3af1..c1dde4204366 100644 --- a/fs/ksmbd/smb2pdu.c +++ b/fs/ksmbd/smb2pdu.c @@ -3222,6 +3222,13 @@ int smb2_open(struct ksmbd_work *work) } } else { if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) { + /* + * Compare parent lease using parent key. If there is no + * a lease that has same parent key, Send lease break + * notification. + */ + smb_send_parent_lease_break_noti(fp, lc); + req_op_level = smb2_map_lease_to_oplock(lc->req_state); ksmbd_debug(SMB, "lease req for(%s) req oplock state 0x%x, lease state 0x%x\n", diff --git a/fs/ksmbd/smb2pdu.h b/fs/ksmbd/smb2pdu.h index e1d0849ee68f..912bd94257ec 100644 --- a/fs/ksmbd/smb2pdu.h +++ b/fs/ksmbd/smb2pdu.h @@ -737,6 +737,7 @@ struct create_posix_rsp { #define SMB2_LEASE_WRITE_CACHING_LE cpu_to_le32(0x04)
#define SMB2_LEASE_FLAG_BREAK_IN_PROGRESS_LE cpu_to_le32(0x02) +#define SMB2_LEASE_FLAG_PARENT_LEASE_KEY_SET_LE cpu_to_le32(0x04)
#define SMB2_LEASE_KEY_SIZE 16
diff --git a/fs/ksmbd/vfs_cache.c b/fs/ksmbd/vfs_cache.c index 774a387fccce..2528ce8aeebb 100644 --- a/fs/ksmbd/vfs_cache.c +++ b/fs/ksmbd/vfs_cache.c @@ -86,6 +86,17 @@ static struct ksmbd_inode *ksmbd_inode_lookup(struct ksmbd_file *fp) return __ksmbd_inode_lookup(fp->filp->f_path.dentry); }
+struct ksmbd_inode *ksmbd_inode_lookup_lock(struct dentry *d) +{ + struct ksmbd_inode *ci; + + read_lock(&inode_hash_lock); + ci = __ksmbd_inode_lookup(d); + read_unlock(&inode_hash_lock); + + return ci; +} + int ksmbd_query_inode_status(struct dentry *dentry) { struct ksmbd_inode *ci; @@ -198,7 +209,7 @@ static void ksmbd_inode_free(struct ksmbd_inode *ci) kfree(ci); }
-static void ksmbd_inode_put(struct ksmbd_inode *ci) +void ksmbd_inode_put(struct ksmbd_inode *ci) { if (atomic_dec_and_test(&ci->m_count)) ksmbd_inode_free(ci); diff --git a/fs/ksmbd/vfs_cache.h b/fs/ksmbd/vfs_cache.h index 8325cf4527c4..4d4938d6029b 100644 --- a/fs/ksmbd/vfs_cache.h +++ b/fs/ksmbd/vfs_cache.h @@ -138,6 +138,8 @@ struct ksmbd_file *ksmbd_lookup_foreign_fd(struct ksmbd_work *work, u64 id); struct ksmbd_file *ksmbd_lookup_fd_slow(struct ksmbd_work *work, u64 id, u64 pid); void ksmbd_fd_put(struct ksmbd_work *work, struct ksmbd_file *fp); +struct ksmbd_inode *ksmbd_inode_lookup_lock(struct dentry *d); +void ksmbd_inode_put(struct ksmbd_inode *ci); struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id); struct ksmbd_file *ksmbd_lookup_fd_cguid(char *cguid); struct ksmbd_file *ksmbd_lookup_fd_inode(struct dentry *dentry);
[ Upstream commit c2a721eead71202a0d8ddd9b56ec8dce652c71d1 ]
Don't immediately send directory lease break notification on smb2_write(). Instead, It postpones it until smb2_close().
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/oplock.c | 45 ++++++++++++++++++++++++++++++++++++++++++-- fs/ksmbd/oplock.h | 1 + fs/ksmbd/vfs.c | 3 +++ fs/ksmbd/vfs_cache.h | 1 + 4 files changed, 48 insertions(+), 2 deletions(-)
diff --git a/fs/ksmbd/oplock.c b/fs/ksmbd/oplock.c index df6f00f58f19..2da256259722 100644 --- a/fs/ksmbd/oplock.c +++ b/fs/ksmbd/oplock.c @@ -396,8 +396,8 @@ void close_id_del_oplock(struct ksmbd_file *fp) { struct oplock_info *opinfo;
- if (S_ISDIR(file_inode(fp->filp)->i_mode)) - return; + if (fp->reserve_lease_break) + smb_lazy_parent_lease_break_close(fp);
opinfo = opinfo_get(fp); if (!opinfo) @@ -1127,6 +1127,47 @@ void smb_send_parent_lease_break_noti(struct ksmbd_file *fp, ksmbd_inode_put(p_ci); }
+void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp) +{ + struct oplock_info *opinfo; + struct ksmbd_inode *p_ci = NULL; + + rcu_read_lock(); + opinfo = rcu_dereference(fp->f_opinfo); + rcu_read_unlock(); + + if (!opinfo->is_lease || opinfo->o_lease->version != 2) + return; + + p_ci = ksmbd_inode_lookup_lock(fp->filp->f_path.dentry->d_parent); + if (!p_ci) + return; + + read_lock(&p_ci->m_lock); + list_for_each_entry(opinfo, &p_ci->m_op_list, op_entry) { + if (!opinfo->is_lease) + continue; + + if (opinfo->o_lease->state != SMB2_OPLOCK_LEVEL_NONE) { + if (!atomic_inc_not_zero(&opinfo->refcount)) + continue; + + atomic_inc(&opinfo->conn->r_count); + if (ksmbd_conn_releasing(opinfo->conn)) { + atomic_dec(&opinfo->conn->r_count); + continue; + } + read_unlock(&p_ci->m_lock); + oplock_break(opinfo, SMB2_OPLOCK_LEVEL_NONE); + opinfo_conn_put(opinfo); + read_lock(&p_ci->m_lock); + } + } + read_unlock(&p_ci->m_lock); + + ksmbd_inode_put(p_ci); +} + /** * smb_grant_oplock() - handle oplock/lease request on file open * @work: smb work diff --git a/fs/ksmbd/oplock.h b/fs/ksmbd/oplock.h index b64d1536882a..5b93ea9196c0 100644 --- a/fs/ksmbd/oplock.h +++ b/fs/ksmbd/oplock.h @@ -129,4 +129,5 @@ int find_same_lease_key(struct ksmbd_session *sess, struct ksmbd_inode *ci, void destroy_lease_table(struct ksmbd_conn *conn); void smb_send_parent_lease_break_noti(struct ksmbd_file *fp, struct lease_ctx_info *lctx); +void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp); #endif /* __KSMBD_OPLOCK_H */ diff --git a/fs/ksmbd/vfs.c b/fs/ksmbd/vfs.c index a89529b21c86..173a488bfeee 100644 --- a/fs/ksmbd/vfs.c +++ b/fs/ksmbd/vfs.c @@ -517,6 +517,9 @@ int ksmbd_vfs_write(struct ksmbd_work *work, struct ksmbd_file *fp, } }
+ /* Reserve lease break for parent dir at closing time */ + fp->reserve_lease_break = true; + /* Do we need to break any of a levelII oplock? */ smb_break_all_levII_oplock(work, fp, 1);
diff --git a/fs/ksmbd/vfs_cache.h b/fs/ksmbd/vfs_cache.h index 4d4938d6029b..a528f0cc775a 100644 --- a/fs/ksmbd/vfs_cache.h +++ b/fs/ksmbd/vfs_cache.h @@ -105,6 +105,7 @@ struct ksmbd_file { struct ksmbd_readdir_data readdir_data; int dot_dotdot[2]; unsigned int f_state; + bool reserve_lease_break; };
static inline void set_ctx_actor(struct dir_context *ctx,
[ Upstream commit 658609d9a618d8881bf549b5893c0ba8fcff4526 ]
opinfo_put() could be called twice on error of smb21_lease_break_ack(). It will cause UAF issue if opinfo is referenced on other places.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/smb2pdu.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/ksmbd/smb2pdu.c b/fs/ksmbd/smb2pdu.c index c1dde4204366..e1c640ed7acc 100644 --- a/fs/ksmbd/smb2pdu.c +++ b/fs/ksmbd/smb2pdu.c @@ -8221,6 +8221,11 @@ static void smb21_lease_break_ack(struct ksmbd_work *work) le32_to_cpu(req->LeaseState)); }
+ if (ret < 0) { + rsp->hdr.Status = err; + goto err_out; + } + lease_state = lease->state; opinfo->op_state = OPLOCK_STATE_NONE; wake_up_interruptible_all(&opinfo->oplock_q); @@ -8228,11 +8233,6 @@ static void smb21_lease_break_ack(struct ksmbd_work *work) wake_up_interruptible_all(&opinfo->oplock_brk); opinfo_put(opinfo);
- if (ret < 0) { - rsp->hdr.Status = err; - goto err_out; - } - rsp->StructureSize = cpu_to_le16(36); rsp->Reserved = 0; rsp->Flags = 0;
[ Upstream commit a9f106c765c12d2f58aa33431bd8ce8e9d8a404a ]
When client send SMB2_CREATE_ALLOCATION_SIZE create context, ksmbd update old size to ->AllocationSize in smb2 create response. ksmbd_vfs_getattr() should be called after it to get updated stat result.
Signed-off-by: Namjae Jeon linkinjeon@kernel.org Signed-off-by: Steve French stfrench@microsoft.com --- fs/ksmbd/smb2pdu.c | 36 ++++++++++++++++++------------------ 1 file changed, 18 insertions(+), 18 deletions(-)
diff --git a/fs/ksmbd/smb2pdu.c b/fs/ksmbd/smb2pdu.c index e1c640ed7acc..8875c04e8382 100644 --- a/fs/ksmbd/smb2pdu.c +++ b/fs/ksmbd/smb2pdu.c @@ -2516,7 +2516,7 @@ static void smb2_new_xattrs(struct ksmbd_tree_connect *tcon, const struct path * da.flags = XATTR_DOSINFO_ATTRIB | XATTR_DOSINFO_CREATE_TIME | XATTR_DOSINFO_ITIME;
- rc = ksmbd_vfs_set_dos_attrib_xattr(mnt_user_ns(path->mnt), path, &da, false); + rc = ksmbd_vfs_set_dos_attrib_xattr(mnt_user_ns(path->mnt), path, &da, true); if (rc) ksmbd_debug(SMB, "failed to store file attribute into xattr\n"); } @@ -3182,23 +3182,6 @@ int smb2_open(struct ksmbd_work *work) goto err_out; }
- rc = ksmbd_vfs_getattr(&path, &stat); - if (rc) - goto err_out; - - if (stat.result_mask & STATX_BTIME) - fp->create_time = ksmbd_UnixTimeToNT(stat.btime); - else - fp->create_time = ksmbd_UnixTimeToNT(stat.ctime); - if (req->FileAttributes || fp->f_ci->m_fattr == 0) - fp->f_ci->m_fattr = - cpu_to_le32(smb2_get_dos_mode(&stat, le32_to_cpu(req->FileAttributes))); - - if (!created) - smb2_update_xattrs(tcon, &path, fp); - else - smb2_new_xattrs(tcon, &path, fp); - if (file_present || created) ksmbd_vfs_kern_path_unlock(&parent_path, &path);
@@ -3299,6 +3282,23 @@ int smb2_open(struct ksmbd_work *work) } }
+ rc = ksmbd_vfs_getattr(&path, &stat); + if (rc) + goto err_out1; + + if (stat.result_mask & STATX_BTIME) + fp->create_time = ksmbd_UnixTimeToNT(stat.btime); + else + fp->create_time = ksmbd_UnixTimeToNT(stat.ctime); + if (req->FileAttributes || fp->f_ci->m_fattr == 0) + fp->f_ci->m_fattr = + cpu_to_le32(smb2_get_dos_mode(&stat, le32_to_cpu(req->FileAttributes))); + + if (!created) + smb2_update_xattrs(tcon, &path, fp); + else + smb2_new_xattrs(tcon, &path, fp); + memcpy(fp->client_guid, conn->ClientGUID, SMB2_CLIENT_GUID_SIZE);
rsp->StructureSize = cpu_to_le16(89);
linux-stable-mirror@lists.linaro.org
-
Namjae Jeon -
Sasha Levin