6ea2aecb |
1 | Linux Trace Toolkit |
2 | |
3 | Mathieu Desnoyers 17-05-2004 |
4 | |
5 | |
6 | This document explains how the lttvwindow API could process the event requests |
7 | of the viewers, merging event requests and hook lists to benefit from the fact |
8 | that process_traceset can call multiple hooks for the same event. |
9 | |
10 | First, we will explain the detailed process of event delivery in the current |
11 | framework. We will then study its strengths and weaknesses. |
12 | |
13 | In a second time, a framework where the events requests are dealt by the main |
14 | window with fine granularity will be described. We will then discussed the |
15 | advantages and inconvenients over the first framework. |
16 | |
17 | |
18 | 1. (Actual) Boundaryless event reading |
19 | |
20 | Actually, viewers request events in a time interval from the main window. They |
21 | also specify a (not so) maximum number of events to be delivered. In fact, the |
22 | number of events to read only gives a stop point, from where only events with |
23 | the same timestamp will be delivered. |
24 | |
25 | Viewers register hooks themselves in the traceset context. When merging read |
26 | requests in the main window, all hooks registered by viewers will be called for |
27 | the union of all the read requests, because the main window has no control on |
28 | hook registration. |
29 | |
30 | The main window calls process_traceset on its own for all the intervals |
31 | requested by all the viewers. It must not duplicate a read of the same time |
32 | interval : it could be very hard to filter by viewers. So, in order to achieve |
33 | this, time requests are sorted by start time, and process_traceset is called for |
34 | each time request. We keep the last event time between each read : if the start |
35 | time of the next read is lower than the time reached, we continue the reading |
36 | from the actual position. |
37 | |
38 | We deal with specific number of events requests (infinite end time) by |
39 | garantying that, starting from the time start of the request, at least that |
40 | number of events will be read. As we can't do it efficiently without interacting |
41 | very closely with process_traceset, we always read the specified number of |
42 | events requested starting from the current position when we answer to a request |
43 | based on the number of events. |
44 | |
45 | The viewers have to filter events delivered by traceset reading, because they |
46 | can be asked by another viewer for a totally (or partially) different time |
47 | interval. |
48 | |
49 | |
50 | Weaknesses |
51 | |
52 | - process_middle does not guarantee the number of events read |
53 | |
54 | First of all, a viewer that requests events to process_traceset has no garantee |
55 | that it will get exactly what it asked for. For example, a direct call to |
56 | traceset_middle for a specific number of events will delived _at least_ that |
57 | quantity of events, plus the ones that have the same timestamp that the last one |
58 | has. |
59 | |
60 | - Border effects |
61 | |
62 | Viewer's writers will have to deal with a lot of border effects caused by the |
318585ee |
63 | particularities of the reading. They will be required to select the information |
64 | they need from their input by filtering. |
6ea2aecb |
65 | |
318585ee |
66 | - Lack of encapsulation and difficulty of testing |
6ea2aecb |
67 | |
68 | The viewer's writer will have to take into account all the border effects caused |
69 | by the interaction with other modules. This means that event if a viewer works |
70 | well alone or with another viewer, it's possible that new bugs arises when a new |
318585ee |
71 | viewer comes around. So, even if a perfect testbench works well for a viewer, it |
72 | does not confirm that no new bug will arise when another viewer is loaded at the |
73 | same moment asking for different time intervals. |
6ea2aecb |
74 | |
75 | |
76 | - Duplication of the work |
77 | |
78 | Time based filters and counters of events will have to be implemented at the |
79 | viewer's side, which is a duplication of the functionnalities that would |
80 | normally be expected from the tracecontext API. |
81 | |
82 | - Lack of control over the data input |
83 | |
84 | As we expect module's writers to prefer to be as close as possible from the raw |
85 | datas, making them interact with a lower level library that gives them a data |
86 | input that they only control by further filtering of the input is not |
87 | appropriated. We should expect some reluctancy from them about using this API |
88 | because of this lack of control on the input. |
89 | |
90 | - Speed cost |
91 | |
92 | All hooks of all viewers will be called for all the time intervals. So, if we |
93 | have a detailed events list and a control flow view, asking both for different |
94 | time intervals, the detailed events list will have to filter all the events |
95 | delivered originally to the control flow view. This can be a case occuring quite |
96 | often. |
97 | |
98 | |
99 | |
100 | Strengths |
101 | |
102 | - Simple concatenation of time intervals at the main window level. |
103 | |
104 | Having the opportunity of delivering more events than necessary to the viewers |
105 | means that we can concatenate time intervals and number of events requested |
106 | fairly easily, while being hard to determine if some specific cases will be |
318585ee |
107 | wrong, in depth testing being impossible. |
6ea2aecb |
108 | |
109 | - No duplication of the tracecontext API |
110 | |
111 | Viewers deal directly with the tracecontext API for registering hooks, removing |
112 | a layer of encapsulation. |
113 | |
114 | |
115 | |
116 | |
117 | |
118 | 2. (Proposed) Strict boundaries events reading |
119 | |
120 | The idea behind this method is to provide exactly the events requested by the |
121 | viewers to them, no more, no less. |
122 | |
6ea2aecb |
123 | It uses the new API for process traceset suggested in the document |
124 | process_traceset_strict_boundaries.txt. |
125 | |
126 | It also means that the lttvwindow API will have to deal with viewer's hooks. |
127 | Those will not be allowed to add them directly in the context. They will give |
128 | them to the lttvwindow API, along with the time interval or the position and |
129 | number of events. The lttvwindow API will have to take care of adding and |
130 | removing hooks for the different time intervals requested. That means that hooks |
131 | insertion and removal will be done between each traceset processing based on |
132 | the time intervals and event positions related to each hook. We must therefore |
133 | provide a simple interface for hooks passing between the viewers and the main |
318585ee |
134 | window, make them easier to manage from the main window. A modification to the |
135 | LttvHooks type solves this problem. |
6ea2aecb |
136 | |
137 | |
138 | Architecture |
139 | |
140 | Added to the lttvwindow API : |
141 | |
142 | |
589a505d |
143 | void lttvwindow_events_request |
fc9fa653 |
144 | ( MainWindow *main_win, |
3c502bdc |
145 | EventsRequest *events_request); |
6ea2aecb |
146 | |
8646cedb |
147 | void lttvwindow_events_request |
148 | ( MainWindow *main_win, |
149 | EventsRequest events_request); |
6ea2aecb |
150 | |
8646cedb |
151 | void lttvwindow_events_request_remove_all |
152 | ( MainWindow *main_win, |
153 | gpointer viewer); |
6ea2aecb |
154 | |
6ea2aecb |
155 | |
8646cedb |
156 | Internal functions : |
6ea2aecb |
157 | |
158 | - lttvwindow_process_pending_requests |
159 | |
318585ee |
160 | |
8646cedb |
161 | Events Requests Removal |
318585ee |
162 | |
8646cedb |
163 | A new API function will be necessary to let viewers remove all event requests |
164 | they have made previously. By allowing this, no more out of bound requests will |
165 | be serviced : a viewer that sees its time interval changed before the first |
166 | servicing is completed can clear its previous events requests and make a new |
167 | one for the new interval needed, considering the finished chunks as completed |
168 | area. |
318585ee |
169 | |
8646cedb |
170 | It is also very useful for dealing with the viewer destruction case : the viewer |
171 | just has to remove its events requests from the main window before it gets |
172 | destroyed. |
318585ee |
173 | |
318585ee |
174 | |
8646cedb |
175 | Permitted GTK Events Between Chunks |
318585ee |
176 | |
8646cedb |
177 | All GTK Events will be enabled between chunks. This is due to the fact that the |
178 | background processing and a high priority request are seen as the same case. |
179 | While a background processing is in progress, the whole graphical interface must |
180 | be enabled. |
318585ee |
181 | |
8646cedb |
182 | We needed to deal with the coherence of background processing and diverse GTK |
183 | events anyway. This algorithm provides a generalized way to deal with any type |
184 | of request and any GTK events. |
318585ee |
185 | |
318585ee |
186 | |
8646cedb |
187 | Background Computation Request |
318585ee |
188 | |
8646cedb |
189 | The types of background computation that can be requested by a viewer : state |
190 | computation (main window scope) or viewer specific background computation. |
318585ee |
191 | |
8646cedb |
192 | A background computation request is asked via lttvwindow_events_request, with a |
193 | priority field set with a low priority. |
6ea2aecb |
194 | |
8646cedb |
195 | If a lttvwindow_events_request_remove_all is done on the viewer pointer, it will |
196 | not affect the state computation as no viewer pointer will have been passed in |
197 | the initial request. This is the expected result. For the background processings |
198 | that call viewer's hooks, they will be removed. |
6ea2aecb |
199 | |
200 | |
206ea1f4 |
201 | A New "Redraw" Button |
202 | |
203 | It will be used to redraw the viewers entirely. It is useful to restart the |
204 | servicing after a "stop" action. |
205 | |
206 | A New "Continue" Button |
207 | |
208 | It will tell the viewers to send requests for damaged areas. It is useful to |
209 | complete the servicing after a "stop" action. |
210 | |
211 | |
6ea2aecb |
212 | |
8646cedb |
213 | Implementation |
6ea2aecb |
214 | |
6ea2aecb |
215 | |
8646cedb |
216 | - Type LttvHooks |
6ea2aecb |
217 | |
8646cedb |
218 | see hook_prio.txt |
69381fc7 |
219 | |
8646cedb |
220 | The viewers will just have to pass hooks to the main window through this type, |
221 | using the hook.h interface to manipulate it. Then, the main window will add |
222 | them and remove them from the context to deliver exactly the events requested by |
223 | each viewer through process traceset. |
69381fc7 |
224 | |
225 | |
8646cedb |
226 | - lttvwindow_events_request |
69381fc7 |
227 | |
8646cedb |
228 | It adds the an EventsRequest struct to the array of time requests |
229 | pending and registers a pending request for the next g_idle if none is |
230 | registered. The viewer can access this structure during the read as its |
231 | hook_data. Only the stop_flag can be changed by the viewer through the |
232 | event hooks. |
69381fc7 |
233 | |
234 | typedef LttvEventsRequestPrio guint; |
235 | |
236 | typedef struct _EventsRequest { |
237 | gpointer viewer_data; |
238 | gboolean servicing; /* service in progress: TRUE */ |
239 | LttvEventsRequestPrio prio; /* Ev. Req. priority */ |
240 | LttTime start_time; /* Unset : { 0, 0 } */ |
241 | LttvTracesetContextPosition *start_position; /* Unset : num_traces = 0 */ |
242 | gboolean stop_flag; /* Continue:TRUE Stop:FALSE */ |
243 | LttTime end_time; /* Unset : { 0, 0 } */ |
244 | guint num_events; /* Unset : G_MAXUINT */ |
245 | LttvTracesetContextPosition *end_position; /* Unset : num_traces = 0 */ |
246 | LttvHooks *before_traceset; /* Unset : NULL */ |
247 | LttvHooks *before_trace; /* Unset : NULL */ |
248 | LttvHooks *before_tracefile;/* Unset : NULL */ |
249 | LttvHooks *event; /* Unset : NULL */ |
250 | LttvHooksById *event_by_id; /* Unset : NULL */ |
251 | LttvHooks *after_tracefile; /* Unset : NULL */ |
252 | LttvHooks *after_trace; /* Unset : NULL */ |
253 | LttvHooks *after_traceset; /* Unset : NULL */ |
254 | LttvHooks *before_chunk; /* Unset : NULL */ |
255 | LttvHooks *after_chunk /* Unset : NULL */ |
256 | } EventsRequest; |
257 | |
258 | |
8646cedb |
259 | |
260 | - lttvwindow_events_request_remove_all |
261 | |
262 | It removes all the events requests from the pool that has their "viewer" field |
263 | maching the viewer pointer given in argument. |
264 | |
265 | It calls the traceset/trace/tracefile end hooks for each request removed. |
266 | |
267 | |
268 | - lttvwindow_process_pending_requests |
269 | |
270 | This internal function gets called by g_idle, taking care of the pending |
271 | requests. It is responsible for concatenation of time intervals and position |
272 | requests. It does it with the following algorithm organizing process traceset |
273 | calls. Here is the detailed description of the way it works : |
274 | |
275 | |
276 | |
277 | - Revised Events Requests Servicing Algorithm (v2) |
278 | |
69381fc7 |
279 | The reads are splitted in chunks. After a chunk is over, we want to check if |
280 | there is a GTK Event pending and execute it. It can add or remove events |
281 | requests from the event requests list. If it happens, we want to start over |
282 | the algorithm from the beginning. |
283 | |
284 | Two levels of priority exists. High priority and low priority. High prio |
285 | requests are serviced first, even if lower priority requests has lower start |
286 | time or position. |
287 | |
288 | |
289 | Data structures necessary : |
290 | |
291 | List of requests added to context : list_in |
292 | List of requests not added to context : list_out |
293 | |
294 | Initial state : |
295 | |
296 | list_in : empty |
297 | list_out : many events requests |
298 | |
299 | |
300 | A. While list_in !empty and list_out !empty and !GTK Event pending |
301 | 1. If list_in is empty (need a seek) |
302 | 1.1 Add requests to list_in |
303 | 1.1.1 Find all time requests with the highest priority and lowest start |
304 | time in list_out (ltime) |
305 | 1.1.2 Find all position requests with the highest priority and lowest |
306 | position in list_out (lpos) |
307 | 1.1.3 If lpos.prio > ltime.prio |
308 | || (lpos.prio == ltime.prio && lpos.start time < ltime) |
309 | - Add lpos to list_in, remove them from list_out |
310 | 1.1.4 Else, (lpos.prio < ltime.prio |
311 | ||(lpos.prio == ltime.prio && lpos.start time >= ltime)) |
312 | - Add ltime to list_in, remove them from list_out |
313 | 1.2 Seek |
314 | 1.2.1 If first request in list_in is a time request |
e6359327 |
315 | - If first req in list_in start time != current time |
316 | - Seek to that time |
69381fc7 |
317 | 1.2.2 Else, the first request in list_in is a position request |
e6359327 |
318 | - If first req in list_in pos != current pos |
319 | - If the position is the same than the saved state, restore state |
320 | - Else, seek to that position |
69381fc7 |
321 | 1.3 Add hooks and call begin for all list_in members |
322 | 1.3.1 If !servicing |
323 | - begin hooks called |
324 | - servicing = TRUE |
325 | 1.3.2 call before_chunk |
326 | 1.3.3 events hooks added |
327 | 2. Else, list_in is not empty, we continue a read |
328 | 2.1 For each req of list_out |
329 | - if req.start time == current context time |
330 | - Add to list_in, remove from list_out |
331 | - If !servicing |
332 | - Call begin |
333 | - servicing = TRUE |
334 | - Call before_chunk |
335 | - events hooks added |
336 | - if req.start position == current position |
337 | - Add to list_in, remove from list_out |
338 | - If !servicing |
339 | - Call begin |
340 | - servicing = TRUE |
341 | - Call before_chunk |
342 | - events hooks added |
343 | |
344 | 3. Find end criterions |
345 | 3.1 End time |
346 | 3.1.1 Find lowest end time in list_in |
347 | 3.1.2 Find lowest start time in list_out (>= than current time*) |
348 | * To eliminate lower prio requests |
349 | 3.1.3 Use lowest of both as end time |
350 | 3.2 Number of events |
351 | 3.2.1 Find lowest number of events in list_in |
352 | 3.2.2 Use min(CHUNK_NUM_EVENTS, min num events in list_in) as num_events |
353 | 3.3 End position |
354 | 3.3.1 Find lowest end position in list_in |
355 | 3.3.2 Find lowest start position in list_out (>= than current |
356 | position) |
357 | 3.3.3 Use lowest of both as end position |
358 | |
359 | 4. Call process traceset middle |
360 | 4.1 Call process traceset middle (Use end criterion found in 3) |
361 | * note : end criterion can also be viewer's hook returning TRUE |
362 | 5. After process traceset middle |
363 | - if current context time > traceset.end time |
364 | - For each req in list_in |
365 | - Call end for req |
366 | - Remove events hooks for req |
367 | - remove req from list_in |
368 | 5.1 For each req in list_in |
369 | - req.num -= count |
370 | - if req.num == 0 |
371 | - Call end for req |
372 | - Remove events hooks for req |
373 | - remove req from list_in |
374 | - if current context time > req.end time |
375 | - Call end for req |
376 | - Remove events hooks for req |
377 | - remove req from list_in |
378 | - if req.end pos == current pos |
379 | - Call end for req |
380 | - Remove events hooks for req |
381 | - remove req from list_in |
382 | - if req.stop_flag == TRUE |
383 | - Call end for req |
384 | - Remove events hooks for req |
385 | - remove req from list_in |
386 | - if exists one events requests in list_out that has |
387 | higher priority and time != current time |
388 | - Use current position as start position for req |
389 | - Remove start time from req |
390 | - Call after_chunk for req |
391 | - Remove event hooks for req |
392 | - Put req back in list_out, remove from list_in |
393 | - Save current state into saved_state. |
394 | |
395 | B. When interrupted |
396 | 1. for each request in list_in |
397 | 1.1. Use current postition as start position |
398 | 1.2. Remove start time |
399 | 1.3. Call after_chunk |
400 | 1.4. Remove event hooks |
401 | 1.5. Put it back in list_out |
402 | 2. Save current state into saved_state. |
403 | 2.1 Free old saved state. |
404 | 2.2 save current state. |
405 | |
406 | |
407 | |
408 | |
409 | |
410 | Notes : |
411 | End criterions for process traceset middle : |
412 | If the criterion is reached, event is out of boundaries and we return. |
413 | Current time >= End time |
414 | Event count > Number of events |
415 | Current position >= End position |
416 | Last hook list called returned TRUE |
417 | |
418 | The >= for position is necessary to make ensure consistency between start time |
419 | requests and positions requests that happens to be at the exact same start time |
420 | and position. |
421 | |
422 | We only keep one saved state in memory. If, for example, a low priority |
423 | servicing is interrupted, a high priority is serviced, then the low priority |
424 | will use the saved state to start back where it was instead of seeking to the |
425 | time. In the very specific case where a low priority servicing is interrupted, |
426 | and then a high priority servicing on top of it is also interrupted, well, the |
427 | low priority will loose its state and will have to seek back. It should not |
428 | occur often. The solution to it would be to save one state per priority. |
429 | |
430 | |
8646cedb |
431 | |
432 | |
433 | |
434 | |
435 | Weaknesses |
436 | |
437 | - None (nearly?) :) |
438 | |
439 | |
440 | Strengths |
441 | |
442 | - Removes the need for filtering of information supplied to the viewers. |
443 | |
444 | - Viewers have a better control on their data input. |
445 | |
446 | - Solves all the weaknesses idenfied in the actual boundaryless traceset |
447 | reading. |
448 | |
449 | - Background processing available. |
450 | |