[go: up one dir, main page]

File: CHANGES

package info (click to toggle)
schedtool 1.3.0-3
  • links: PTS, VCS
  • area: main
  • in suites: buster
  • size: 204 kB
  • sloc: ansic: 523; makefile: 85
file content (263 lines) | stat: -rw-r--r-- 7,205 bytes parent folder | download | duplicates (4)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
Tue Aug  6 14:42:36 CEST 2002
v0.3

-added SCHED_FIFO, SCHED_RR, static priorities
-updated documentation
-included some scheduler overview (sched-design.txt)
-misc typo-fixes
-added Makefile for fast compiling / installing
-fixed logical errors in CHECK_RANGE_PRIO, CHECK_RANGE_POLICY
	(must've been quite late ;-)


Sat Aug 10 16:26:19 CEST 2002
v0.5

-man-page is ready and also installed via make install
-README restructured
-Wondering about the fast version-progress?
The whole package is now roughly half of what I've been dreaming of.
Still missing are e.g. some numbers.


Fri Aug 30 16:52:27 CEST 2002
v0.6 (only for me)

-minor source (-h / -v switches), README and Makefile (target: zipman,
unzipman) fixes


Wed Sep  4 10:22:31 CEST 2002
v0.7

-huge README updates with some numbers, improvements
-return-value updates: return values represent the number of failed
sched_setscheduler()-calls with 0 being everything okay.


Sun Sep 22 00:18:09 CEST 2002
v0.8

-some code work, #define SCHED_*, replace 0 with SCHED_NORMAL and such minor
stuff
-clarified the meaning of "higher" in README
-asked for util-linux inclusion


Mon Nov 11 17:26:45 CET 2002
v0.8b

-minor cleanups


Tue Aug 26 19:53:41 CEST 2003
v0.9

-implement better error handling to return exactly (well I tried at my
best knowledge) what went wrong during the sched_setscheduler()-call
-set_process() no longer automatically calls print_process()
-minor README/man-page updates regarding kernel versions 2.5 and 2.6


Sun Aug 31 16:14:35 CEST 2003
v0.95

-add support for CPU-affinity
-simplify error-handling in code
-overhaul the man-page


Tue Dec 24 21:31:59 CET 2003 (Tue Jan  6 13:37:32 CET 2004)
v0.96

-it is bad practice to have the "prerequisites" containing technical
stuff before the "about / info"-section explaing what the hell this
even is. README changed accordingly.
-moved installing into file INSTALL
-Con made a suggestion for using schedtool as a wrapper to start processes
with adjusted scheduling-params. Here you go, Con (option -e for (e)xec)!
-added parenthesis for safety in print_process in the big 
if( ... || ... || ... ) --now--> if( (...) || (...) || (...) )
-fixed an argument-BUG ("prio >= prio_min" instead of "p >= ...") in 
#define CHECK_RANGE_PRIO which did not change behaviour, but was wrong.
-made my debug-printing accessible via -DDEBUG_ME
-merry christmas


Thu Jan 15 14:41:06 CET 2004
v0.97

-changed -e(xecute) semantics from fork() to in-place exec() much like RML's
tools do it. This means that the schedtool-process will be overwritten by 
the specified command.
-added padding for better readable output; may change to dynamic spacing in 
future versions


Wed Jan 21 15:44:59 CET 2004
v0.98

-there, people, you have it: your changed cmdline parsing, no need for quoting
multi-argument-commands, just plain 
#> schedtool -3 -e cpu-hog -x1 -l2
THIS MEANS ALSO THAT YOU NO LONGER CAN MIX ARGS AND OPTIONS LIKE:
#> schedtool 1 2 3 4 -3 5 6
-restructured the help-output to be more "comprehendable" ... as an admin I get
serious pain reading it now
-and, hooray, schedtool is now silent "as that is standard unix behaviour"
unless you specify -v, which will print for each setting PID and so on, which
is indeed what I prefer to have.

Now that you have it all like you wanted, 
	send me money or even better, hot chicks.


Wed Jan 28 13:51:05 CET 2004
v0.99

-build schedtool with affinity is now the default
-changed Makefile a bit; new target no_affinity
-included support for SCHED_ISO (patch by Con Kolivas)
-update docs to reflect changes


Sun Apr  4 10:38:17 CEST 2004
v0.99.1

-support nice operations; schedtool is now capable of setting ALL scheduling
attributes!
-minor change in engine (move execute into the loop to eliminate duplicate 
code; the code is slowly getting a mess with all these checks here and there)
-documentation /man-page update; man-page now lists SCHED_*
-Makefile: new target install-doc, installs some docs into $PREFIX/share/doc/schedtool


Tue Apr 20 11:45:19 CEST 2004
v1.0

-released 1.0; more documentation updates and clarifications


Wed Jun 30 20:27:29 CEST 2004
v1.1.0

-added "NO WARRANTY" to help output
-add raw mode for the kernel hackers; allows direct input of the numerical 
policy value and changed docs accordingly.


Mon Jul  5 14:12:35 CEST 2004
v1.1.1

-added a list mode for affinity, additional to the bitmask mode
-identified 2 FIXMEs


Sun Aug  1 14:00:58 CEST 2004

-fixed one fixme: check affinity input for isdigit()
-changed the CHECK_RANGE_PRIO macro a bit for code cleanup
-moved affinity setting to the end of the loop, because may fail with ENOSYS


Sun Sep 19 19:50:09 CEST 2004
v1.1.2

-added missing padding for NICE in output
-changed custom "policy not implemented" error to "value out of range / policy
not implemented" to indicate wrong values when setting affinity or such


Tue Sep 28 20:59:42 CEST 2004
v1.2

-finally try to not break affinity-compiled binaries on non-affinity kernels
(2nd FIXME fixed)
-use getpid() instead of PID 0 (== current process) when in execute mode


Sun Oct  3 18:14:46 CEST 2004
v1.2.1

-REALLY DO NOT break affinity stuff; sched_getaffinity() suddenly returns 4 on
success instead of 0 as declared in the man-page on at least kernels 2.6.8+
-pad nice-levels correctly; they can be negative, as such we need %3d in 
printf()


Tue Oct 12 13:04:53 CEST 2004
v1.2.2

-adjust copyright
-correct typo in man-page
-correctly figure out affinity; I was missing an #include <linux/unistd.h>


Sun Oct 17 00:41:43 CEST 2004
v1.2.3

-get the affinity-ifdef-stuff right. Oh my.
-really up version to 1.2.3 (I'm getting old)
-remove superfluous break in schedtool.c::engine(), execute part
-correct another typo in man-page


Sat Nov 27 00:27:23 CET 2004
v1.2.4

-add a "probe" mode (-r), similar to sched-utils; lists prio_min/max
-add helper-functions get_prio_min_max() and print_prio_min_max()


Sat Feb  5 19:59:44 CET 2005
v1.2.5

-fix NULL pointer printing when policy is out-of-range of TAB[] ; instead,
print numeric value + <UNKNOWN>
-behave more nicely with SCHED_ISO and newer -ck kernels (inspired by cr7)


Tue Jan 17 18:50:04 CET 2006
v1.2.6

-add support for SCHED_IDLEPRIO in 2.6.16-ck kernels
-update documentation accordingly; remove old stuff about SCHED_BATCH which is
in mainline 2.6 now


Mon Sep 30 22:29:48 CEST 2006
v1.2.7

-update man-page: get section(8) right again, mention affinity list mode


Tue Nov 14 21:22:18 CET 2006
v1.2.8

-try to initialize aff_mask in a more portable way
-minor man-page update
-change SCHED_IDLE to SCHED_IDLEPRIO everywhere
-change MANPATH to $PREFIX/share/man


Sun Nov 26 15:27:12 CET 2006
v1.2.9

-spelling fixes for README; thanks to Sven Hartrumpf
-Makefile fix for installing man-page (wrong path); thanks to Adam Sampson


Fri Jul 20 20:24:24 CEST 2007
v1.2.10

-fix incorrect use of $DISTDIR in Makefile for RPM-bulding
 thanks to Adel Gadllah
-fix version to 1.2.10 (forgot for 1.2.9)


Mon Sep 15 15:35:29 CEST 2008
v1.3.0

-bring schedtool to new glibc affinity api (cpu_set_t) & make it work under new glibcs again
-bug-fix behaviour when 1 wrong PID was given - all other (valid) PIDs would run into error