• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1Coverage-Area: Clocks and Timers (TMR)
2Area-Code: tmr
3
4/*
5COVERAGE DOCUMENT FOR TIMERS
6============================
7
8This document lists the items that will be tested for POSIX timers.
9
10	Item to Test		Status
11==========================================================
12||	Definitions		DONE			||
13||	Behavior		DONE with TMR and CS    ||
14||				tags planned.		||
15||				see table below		||
16||	Functionality		STOPPED FOR NOW		||
17||	Stress			STOPPED FOR NOW		||
18==========================================================
19*/
20
21// Function		Complete?	Priority
22// =================================================
23clock			YES		MED
24clock_getcpuclockid	YES		LOW
25clock_getres		YES		MED
26clock_gettime		YES*		MED
27clock_settime		YES*		MED
28clock_nanosleep		YES*		HIGH
29nanosleep		YES		HIGH
30timer_create		YES*		HIGH
31timer_delete		YES		MED
32timer_getoverrun	YES		HIGH
33timer_gettime		YES		MED
34timer_settime		YES		MED
35// ===================================================
36
37// *Incomplete because process time clocks functionality not in current version
38// under test.
39
40/*
41NOTES
42=====
43See the BUILD file in the main directory for the most up-to-date
44information on how to compile tests from within the framework.
45
46Some additional notes:
47If you want to get CLOCK_MONOTONIC, though, one option that worked
48with HRT 2.5.54bk6 and below and RH7.3 was to include
49/usr/src/linux/Documentation/high-res-timers/lib/posix_time.h
50and turn on -D_POSIX_TIMERS.
51
52Testing BIN List
53================
54Items which I'd like to test more in the future (possibly in functional or
55stress tests).
56
57- Absolute timers and clock changes:
58  - test setting the time for a timer where the interval is large
59  - also, test where the delta between old and new clock time is large
60    => forces division of large numbers, could have inaccuracies (would
61       probably? be a stress test)
62
63- test clock_nanosleep() or nanosleep() with a variety of different signals
64  being sent to interrupt it (esp. if they have no user-handler)
65  -> probably also stress
66
67- Also, there may be some ideas for stress tests in the Base Definitions
68document.
69
70POSIX Timers Implementations
71============================
72glibc Timers in user-space - http://ftp.gnu.org/gnu/glibc/
73High Res Timers            - http://high-res-timers.sourceforge.net
74Alternate High Res Timers  - http://high-res-timers.sourceforge.net (look
75                             on mailing list for Jim Houston's alternate
76                             implementation -- patch that goes on top
77                             of George Anzinger's HRT)
78*/
79
80Maintainer: Julie Fleischer
81