summaryrefslogtreecommitdiff
path: root/source3
diff options
context:
space:
mode:
authorHolger Hetterich <hhetter@novell.com>2010-01-23 22:03:22 +0100
committerJim McDonough <jmcd@samba.org>2010-03-16 09:52:09 -0400
commit81c6b878b1cb665d7dd4b365af82a8c15b099d38 (patch)
treec45b57288498edd7281ad3adbb44493d4f9c99e9 /source3
parenta45db5948050b7a94181e0579fb9fc9f651aed74 (diff)
downloadsamba-81c6b878b1cb665d7dd4b365af82a8c15b099d38.tar.gz
samba-81c6b878b1cb665d7dd4b365af82a8c15b099d38.tar.bz2
samba-81c6b878b1cb665d7dd4b365af82a8c15b099d38.zip
Added an exact description of the V2 protocol.
I don't think it should have it's place the man page, because this is developer information.
Diffstat (limited to 'source3')
-rw-r--r--source3/modules/vfs_smb_traffic_analyzer.c62
1 files changed, 62 insertions, 0 deletions
diff --git a/source3/modules/vfs_smb_traffic_analyzer.c b/source3/modules/vfs_smb_traffic_analyzer.c
index f2735ec9fe..226611ca64 100644
--- a/source3/modules/vfs_smb_traffic_analyzer.c
+++ b/source3/modules/vfs_smb_traffic_analyzer.c
@@ -27,6 +27,68 @@ enum sock_type {INTERNET_SOCKET = 0, UNIX_DOMAIN_SOCKET};
#define LOCAL_PATHNAME "/var/tmp/stadsocket"
+
+/*
+ * Protocol version 2.0 description
+ *
+ * The following table shows the exact assembly of the 2.0 protocol.
+ *
+ * -->Header<--
+ * The protocol header is always send first, and contains various
+ * information about the data block to come.
+ * The header is always of fixed length, and will be send unencrypted.
+ *
+ * Byte Number/Bytes Description
+ * 00-02 Contains always the string "V2."
+ * 03 This byte contains a possible subrelease number of the
+ * protocol. This enables the receiver to make a version
+ * check to ensure the compatibility and allows us to
+ * release 2.x versions of the protocol with bugfixes or
+ * enhancements.
+ * 04 Usually, this byte contains the character '0'. If the
+ * VFS module is configured for anonymization, this is
+ * set to 'A'. This information can be useful for the
+ * receiver.
+ * 05 Usually, this byte contains the character '0'. If the
+ * VFS module is configured for encryption of the data,
+ * this byte is set to 'E'.
+ * 06-09 These bytes contain the character '0' by default, and
+ * are reserved for possible future extensions. They have
+ * no function in 2.0.
+ * 10-27 17 bytes containing a string representation of the
+ * number of bytes to come in the following data block.
+ * It is right aligned and filled from the left with '0'.
+ *
+ * -->Data Block<--
+ * The data block is send immediately after the header was send. It's length
+ * is exactly what was given in bytes 11-28 from in the header.
+ *
+ * The data block may be send encrypted.
+ *
+ * To make the data block easy for the receiver to read, it is divided into
+ * several sub-blocks, each with it's own header of four byte length. In each
+ * of the sub-headers, a string representation of the length of this block is
+ * to be found.
+ *
+ * Thus the formal structure is very simple:
+ *
+ * [HEADER]data[HEADER]data[HEADER]data[END]
+ *
+ * whereas [END] is exactly at the position given in bytes 11-28 of the
+ * header.
+ *
+ * Some data the VFS module is capturing is of use for any VFS operation.
+ * Therefore, there is a "common set" of data, that will be send with any
+ * data block. The following provides a list of this data.
+ * - the VFS function identifier (see VFS function ifentifier table below).
+ * - a timestamp to the millisecond.
+ * - the username (as text) who runs the VFS operation.
+ * - the SID of the user who run the VFS operation.
+ * - the domain under which the VFS operation has happened.
+ *
+ */
+
+
/* VFS Functions identifier table. In protocol version 2, every vfs */
/* function is given a unique id. */
enum vfs_id {