smb: client: refactor ACL setting control flow in id_mode_to_cifs_acl()
Subsystems (3)
PATCH 1/3 smb: client: refactor ACL setting control flow in id_mode_to_cifs_acl()
From: Ralph Boehme <[email protected]>
Refactor the control flow in id_mode_to_cifs_acl() to reduce nesting and prevent error code overwriting.
Instead of wrapping the call to ops->set_acl() in a conditional block, introduce early exits (goto id_mode_to_cifs_acl_exit) when build_sec_desc() fails or ops->set_acl is NULL. This ensures that any actual error returned by build_sec_desc() is not overwritten with -EOPNOTSUPP.
--- fs/smb/client/cifsacl.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-)
fs/smb/client/cifsacl.c +10 -6
@@ -1834,14 +1834,18 @@ id_mode_to_cifs_acl(struct inode *inode, const char *path, __u64 *pnmode,cifs_dbg (NOISY ,"build_sec_desc rc: %d\n" , rc); -if (ops->set_acl ==NULL ) - rc = -EOPNOTSUPP ; +if (rc !=0 ) + goto id_mode_to_cifs_acl_exit; -if (!rc) { -/* Set the security descriptor */ - rc = ops->set_acl (pnntsd, nsecdesclen, inode, path, aclflag); -cifs_dbg (NOISY ,"set_cifs_acl rc: %d\n" , rc); +if (ops->set_acl ==NULL ) { + rc = -EOPNOTSUPP ; + goto id_mode_to_cifs_acl_exit; } + +/* Set the security descriptor */ + rc = ops->set_acl (pnntsd, nsecdesclen, inode, path, aclflag); +cifs_dbg (NOISY ,"set_cifs_acl rc: %d\n" , rc); + id_mode_to_cifs_acl_exit:cifs_put_tlink (tlink);
Signature · 1 line
2.53.0
PATCH 2/3 smb/client: fix security flag calculation when setting security descriptors
From: Ralph Boehme <[email protected]>
In id_mode_to_cifs_acl(), aclflag was initialized to CIFS_ACL_DACL by default. This forced the client to request setting the DACL even when only an ownership (chown) or group (chgrp) change was being performed.
Let build_sec_desc() do the proper flag calculation by initializing aclflag to 0. build_sec_desc() sets the appropriate bits (CIFS_ACL_OWNER, CIFS_ACL_GROUP, or CIFS_ACL_DACL) depending on what actually changed.
If build_sec_desc() results in aclflag being 0 (meaning no changes were mapped), exit early to avoid sending an empty security descriptor update to the server.
--- fs/smb/client/cifsacl.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
fs/smb/client/cifsacl.c +6 -1
@@ -1738,7 +1738,7 @@ id_mode_to_cifs_acl(struct inode *inode, const char *path, __u64 *pnmode, kuid_t uid, kgid_t gid) { int rc = 0; -int aclflag =CIFS_ACL_DACL ;/* default flag to set */ +int aclflag =0 ;__u32 secdesclen =0 ;__u32 nsecdesclen =0 ;__u32 dacloffset =0 ;
@@ -1837,6 +1837,11 @@ id_mode_to_cifs_acl(struct inode *inode, const char *path, __u64 *pnmode,if (rc !=0 ) goto id_mode_to_cifs_acl_exit; +if (aclflag ==0 ) { +cifs_dbg (FYI ,"set_cifs_acl aclflag=0, no change mapped\n" ); + goto id_mode_to_cifs_acl_exit; + } +if (ops->set_acl ==NULL ) { rc = -EOPNOTSUPP ; goto id_mode_to_cifs_acl_exit;
Signature · 1 line
2.53.0
PATCH 3/3 smb/client: fix chown/chgrp with SMB3 POSIX Extensions
In reply to Steve FrenchFrom: Ralph Boehme <[email protected]>
Ownership (chown) and group (chgrp) modifications were being ignored when mounting with SMB3 POSIX Extensions unless CIFS_MOUNT_CIFS_ACL or CIFS_MOUNT_MODE_FROM_SID were also explicitly set.
Fix this by checking for posix_extensions in cifs_setattr_nounix() when updating UID and GID, ensuring that id_mode_to_cifs_acl() is called to map and set the ownership/group information on the server.
--- fs/smb/client/inode.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
fs/smb/client/inode.c +2 -1
@@ -3376,7 +3376,8 @@ cifs_setattr_nounix(struct dentry *direntry, struct iattr *attrs)if (attrs->ia_valid &ATTR_GID ) gid = attrs->ia_gid; -if (sbflags & (CIFS_MOUNT_CIFS_ACL |CIFS_MOUNT_MODE_FROM_SID )) { +if ((sbflags & (CIFS_MOUNT_CIFS_ACL |CIFS_MOUNT_MODE_FROM_SID )) || +cifs_sb_master_tcon (cifs_sb)->posix_extensions) {if (uid_valid (uid) ||gid_valid (gid)) { mode =NO_CHANGE_64 ; rc = id_mode_to_cifs_acl(inode, full_path, &mode,
Signature · 1 line
2.53.0