I've got this complaint from coverity: ----- Forwarded message from scan-admin@xxxxxxxxxxxx ----- ________________________________________________________________________________________________________ *** CID 1600337: (UNINIT) /fs/attr.c: 311 in setattr_copy_mgtime() 305 else if (ia_valid & ATTR_ATIME) 306 inode_set_atime_to_ts(inode, now); 307 308 if (ia_valid & ATTR_MTIME_SET) 309 inode_set_mtime_to_ts(inode, attr->ia_mtime); 310 else if (ia_valid & ATTR_MTIME) >>> CID 1600337: (UNINIT) >>> Using uninitialized value "now" when calling "inode_set_mtime_to_ts". 311 inode_set_mtime_to_ts(inode, now); 312 } 313 314 /** 315 * setattr_copy - copy simple metadata updates into the generic inode 316 * @idmap: idmap of the mount the inode was found from /fs/attr.c: 306 in setattr_copy_mgtime() 300 WARN_ON_ONCE(ia_valid & ATTR_MTIME); 301 } 302 303 if (ia_valid & ATTR_ATIME_SET) 304 inode_set_atime_to_ts(inode, attr->ia_atime); 305 else if (ia_valid & ATTR_ATIME) >>> CID 1600337: (UNINIT) >>> Using uninitialized value "now" when calling "inode_set_atime_to_ts". 306 inode_set_atime_to_ts(inode, now); 307 308 if (ia_valid & ATTR_MTIME_SET) 309 inode_set_mtime_to_ts(inode, attr->ia_mtime); 310 else if (ia_valid & ATTR_MTIME) 311 inode_set_mtime_to_ts(inode, now); Now we WARN if ATTR_MTIME is set without ATTR_CTIME but still it might be good to place some sane value in 'now' so that we don't set the timestamp to garbage and more importantly setting ATTR_ATIME without ATTR_CTIME looks as a possible thing? Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR
