ltt/branches/poly/doc/developer/lttng-lttv-roadmap.html

   1 <html>
   2 <body>
   3 <center><big><big>LTTV & LTTng roadmap<small><small></center>
   4 <br>
   5 <br>
   6 Here are the roadmaps for the LTTV and LTTng development. I use a priority indice
   7 for the TODO items :<br>
   8 (1) : very high priority<br>
   9 (10): lowest priority<br>
  10 <br>
  11 <br>
  12 Dependencies are written between brackets [ ].<br>
  13 The # symbol marks who is currently working on the item.<br>
  14 The % symbol marks who is interested in the realisation of the item.<br>
  15 <br>
  16 <big>LTTV Roadmap<small><br>
  17 <br>
  18 * TODO<br>
  19 (1) Marker integration. Change the byte pair "facility, event" id for a short
  20 combining the informatinon.<br>
  21         # <A HREF="mailto:Mathieu Desnoyers <compudj@krystal.dyndns.org>">Mathieu Desnoyers</A><br>
  22 (3) Add cluster support.<br>
  23         # <A HREF="mailto:Eric Clement <eric.clement@polymtl.ca>">Eric Clement</A><br>
  24 (3) Make LTTV aware of type formats (visual separators) defined in the XML
  25 file.<br>
  26         # <A HREF="mailto:Gabriel Matni <gabrien.matni@polymtl.ca>">Gabriel Matni</A><br>
  27 (3) Use a per architecture enumeration for traps.<br>
  28 (4) Statistics per time window.<br>
  29 (4) Add Xen per physical CPU view.<br>
  30 (4) Add Xen per vcpu view.<br>
  31 (4) Disable plugins when threshold reached (i.e. too much process in control
  32 flow view). Draw, and, when the threshold is reached, stop drawing. The global
  33 statistics view can inhibit showing the per process stats.<br>
  34 (4) Add a visual artifact : PID 0 could be named swapper instead of UNNAMED for
  35 cpus > 0.<br>
  36 (4) Add event specific fields support to filter.<br>
  37 (4) Add a periodic event interval view. (useful to verify event periodicity)<br>
  38 (4) create a graphical per cpu activity view.<br>
  39 (4) Filter by target process.<br>
  40 (4) Compensate for time spent in probes in LTTV analysis.<br>
  41 (4) Add CPU, network, disk, memory usage histogram. [Per interval statistics]<br>
  42 (5) Add Python scripting hooks.<br>
  43 (5) Add GUI interface to take an hybrid trace.<br>
  44 (5) Flight recorder : start lttd automatically upon GUI trace control stop.<br>
  45 (5) Automatically detect traces with too much processes and disable faulty operations.<br>
  46 (5) Event sequence detector (inspired from regular expressions).<br>
  47 (7) Create a hardware counter viewer (low cost rate counters : L1 cache miss,
  48 page faults, interrupts...). This will be a generalisation of the event rate
  49 view into a view of the evolution of a user definable event field.<br>
  50 <br>
  51 * TO FIX<br>
  52 (10) Add cancel button to LTTV filter GUI window.<br>
  53 (10) Sometimes, in the control flow view, a process with 0 creation time is
  54 created in addition to the real process itself. Seems to be caused by end of
  55 process life.<br>
  56 (10) Statistics do not take in account the time spent in the mode present at
  57 the beginning of the trace. Example : real time spent in system call on behalf
  58 of process 0.<br>
  59 <br>
  60 <br>
  61 <big>LTT Next Generation Roadmap<small><br>
  62 <br>
  63 * TODO<br>
  64 (1) efficient dynamic event filtering while recording trace.<br>
  65         % Google<br>
  66         % Sensis Corp. <A HREF="mailto:Bish, Tim <Tim.Bish@Sensis.com>">Tim Bish</A><br>
  67         # <A HREF="mailto:Mathieu Desnoyers <compudj@krystal.dyndns.org>">Mathieu Desnoyers</A><br>
  68         - Partly implemented: per marker selection through /proc interface.<br>
  69         - Must still implement per active trace filtering.<br>
  70 (1) Add Xen support. (Trace buffer desallocation needs to be fixed)<br>
  71         # <A HREF="mailto:Mathieu Desnoyers <compudj@krystal.dyndns.org>">Mathieu Desnoyers</A><br>
  72 (1) Integrate SystemTAP logging with LTTng.<br>
  73 (4) instrument kernel bottom half irqsave, seqlocks, semaphores, mutexes, brlock.<br>
  74 (4) integrate NPTL instrumentation (see
  75 <A HREF="http://nptltracetool.sourceforge.net/">PTT</A>).<br>
  76 (4) Probe calibration kernel module.<br>
  77 (4) Make page faults detect nested fault without nesting 4 times in the page
  78 fault handler.<br>
  79 (5) Support CPUs with scalable frequency.<br>
  80 (5) Add boot time tracing support.<br>
  81 (5) Integrate LTTng and lttd with LKCD.<br>
  82 (7) Integrate periodical dump of perfctr hardware counters.<br>
  83 (8) Integrate periodical dump of SystemTAP computed information.<br>
  84 (9) Add support for setjmp/longjmp and jump tables instrumentation to
  85 ltt-instrument-functions.<br>
  86 (9) Make ltt-usertrace-fast support internal heartbeat timer.<br>
  87 (3) port LTTng to :<br>
  88 alpha<BR>
  89 user-mode Linux<BR>
  90 Xen<BR>
  91 # IBM<BR>
  92 S/390<BR>
  93 RTLinux<BR>
  94 % Wind River for 2.6.14<BR>
  95 sparc64<BR>
  96 # Wind River<BR>
  97 sh4<br>
  98 <br>
  99 <br>
 100 <br>
 101 * TODO (low priority)<br>
 102 enhance RPM packages for lttng kernel<br>
 103 Integrate header generation (genevent) in kernel build system.<br>
 104 Export channels via network sockets instead of writing them to disk.<br>
 105 Export buffers with time constraint for "live" visualisation. Use
 106 ltt_force_switch periodically from a timer to insure slow channels do not
 107 interfere with viewing.<br>
 108 Have an optional round-robin mode to write information into multiple channels
 109 from the same source.<br>
 110 <br>
 111 * Need to be discussed<br>
 112 Use 2.6.14 RelayFS control files.<br>
 113 Drop ltt-module-register and ltt-module-unregister, use exported variables.<br>
 114 drop ltt_filter_control, use functions pointers instead.<br>
 115 Merge facilities headers into one big header.<br>
 116 Change the name of XML files from XML to something else.<br>
 117 Remove ltt-base.c.<br>
 118 <br>
 119 Mathieu Desnoyers<br>
 120
 121
 122 </body>
 123 </html>