SystemTap: README Source File

00001 /** @mainpage SystemTap Runtime
-00002 
-00003 @section intro_sec Introduction
-00004 
-00005 This document describes the implementation of the SystemTap Runtime. It is intended for developers
-00006 of the SystemTap Language translator or, possibly TapSet authors. These functions are not directly
-00007 available from the SystemTap Language.
-00008 
-00009 The SystemTap Runtime Library consists of all functions
-00010 and code fragments needed by the compiler/translator to
-00011 include in building a kernel module using kprobes. It
-00012 also include I/O code to transmit its output from the kernel to userspace.
-00013  
-00014 In addition to the library, the runtime includes a SystemTap user-space daemon
-00015 (stpd).  Stpd grabs data sent from the I/O code in the runtime and displays it
-00016 and/or saves it to files. Stpd (or a script invoking it) will handle other issues like
-00017 inserting and removing modules.
-00018 
-00019 Stpd and the I/O code make use of both relayfs and netlink for communication.  For
-00020 kernels without relayfs builtin, it is provided as a standalone module under the runtime directory.
-00021 
-00022 @section design_sec Design
-00023 @subsection impl_sec Implementation
-00024 The library is written in C and is really not a library but a collection of code
-00025 That can be conditionally included in a modules. It may become a library later, but for now
-00026 there are some advantages to being able to change the sizes of static items with simple #defines.
-00027 
-00028 @subsection map_sec Maps (Associative Arrays)
-00029 Maps are implemented as hash lists. It is not expected that users will
-00030 attempt to collect so much data in kernel space that performance problems will require
-00031 more complex solutions such as AVL trees.
-00032 
-00033 Maps are created with _stp_map_new().  Each map can hold only one type of 
-00034 data; int64, string, or statistics.  Each element belonging to a map can have up to 2 keys
-00035 and a value.  Implemented key types are strings and longs.
-00036         
-00037 To simplify the implementation, the functions to set the key and the functions to set the data are separated.
-00038 That means we need only 4 functions to set the key and 3 functions to set the value. 
-00039 
-00040 For example:
-00041 \code
-00042 /* create a map with a max of 100 elements */
-00043 MAP mymap = map_new(100, INT64);
-00044 
-00045 /* mymap[birth year] = 2000 */
-00046 map_key_str (mymap, "birth year");
-00047 map_set_int64 (mymap, 2000);
-00048 \endcode
-00049 
-00050 All elements have a default value of 0 (or NULL).  Elements are only saved to the map when their value is set
-00051 to something nonzero.  This means that querying for the existance of a key is inexpensive because
-00052 no element is created, just a hash table lookup.
-00053 
-00054 @subsection list_sec Lists
-00055 A list is a special map which has internally ascending long integer keys.  Adding a value to
-00056 a list does not require setting a key first. Create a list with _stp_list_new(). Add to it
-00057 with _stp_list_add_str() and _stp_list_add_int64().  Clear it with _stp_list_clear().
-00058 
-00059 @subsection string_sec Strings
-00060 One of the biggest restrictions the library has is that it cannot allocate things like strings off the stack.
-00061 It is also not a good idea to dynamically allocate space for strings with kmalloc().  That leaves us with 
-00062 statically allocated space for strings. This is what is implemented in the String module.  Strings use
-00063 preallocated per-cpu buffers and are safe to use (unlike C strings).
-00064 
-00065 @subsection io_sec I/O
-00066 Generally things are written to a "print buffer" using the internal
-00067 functions _stp_print_xxx().
-00068 \code
-00069 _stp_print ("Output is: ");
-00070 _stp_printf ("pid is %d ", current->pid);
-00071 _stp_printf ("name is %s", current->comm);
-00072 \endcode
-00073 before the probe returns it must call _stp_print_flush().  This
-00074 timestamps the accumulated print buffer and sends it to relayfs.
-00075 When relayfs fills an internal buffer, the user-space daemon is notified
-00076 data is ready and reads a bug per-cpu chunk, which contains a line like:
-00077 \verbatim
-00078 [123456.000002] Output is: pid is 1234 name is bash
-00079 \endverbatim
-00080 
-00081 The user-daemon (stpd) saves this data to a file named something like
-00082 "stpd_cpu2".  When the user hits ^c, a timer expires, or the probe
-00083 module notifies stpd (through a netlink command channel) that it wants
-00084 to terminate, stpd does "system(rmmod)" then collects the last output
-00085 before exiting.
-00086 As an option, if we don't need bulk per-cpu data, we can put
-00087 \code
-00088 #define STP_NETLINK_ONLY
-00089 \endcode
-00090 at the top of the module and all output will go over a netlink channel.
-00091 In the SystemTap language, we will provide some simple functions to control the buffering policy, which
-00092 will control the use of netlink and parameters to relayfs and stpd.
-00093 
-00094 @section status_sec Status
-00095 @li Maps are implemented and tested. Histograms are not yet finished.
-00096 @li Copy_From_User functions are done.
-00097 @li If maps overflow or memory runs out for some reason, globals are set but nothing is done yet.
-00098 I expect to implement a function to tell the system to either ignore it or unload the module and quit.
-00099 @li Stack functions need much improvement.
-00100 
-00101 @section probe_sec Example Probes
-00102 
-00103 Working sample probe code using the runtime is in runtime/probes.
-00104 <a href="dir_000000.html"> Browse probes.</a>
-00105 
-00106 @section todo_sec ToDo 
-00107 \link todo Click Here for Complete List \endlink
-00108 
-00109 @section links Links
-00110 <a href="http://sources.redhat.com/systemtap/">SystemTap Project Page</a>
-00111  */
-