-
Notifications
You must be signed in to change notification settings - Fork 2
/
ChangeLog
5454 lines (3873 loc) · 209 KB
/
ChangeLog
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
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# Copyright (C) 2005-2017 The RTAI project
# This [file] is free software; the RTAI project
# gives unlimited permission to copy and/or distribute it,
# with or without modifications, as long as this notice is preserved.
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY, to the extent permitted by law; without
# even the implied warranty of MERCHANTABILITY or FITNESS FOR A
# PARTICULAR PURPOSE.
2019-03-22 Paolo Mantegazza <[email protected]>
* bas/math/libm.c: make available d2str, always.
2019-03-21 Paolo Mantegazza <[email protected]>
* RTAI public repository <savannah.nongnu.org/projects/rtai> aligned
to its latest in house version.
2018-11-04 Paolo Mantegazza <[email protected]>
* addons/drivers/-> serial/serial.c, comedi/comedi-module.c;
base/ipc/-> bits/bits.c, mbx/mbx.c, mq/mq.c, netrpc/netrpc.c,
sem/sem.c, tbx/tbx.c; base/sched/api.c; addons/rtdm/xn: fixed
error dereferencing pointer to incomplete type ‘struct
task_struct’for Linux >= 4.14.71.
* base/ipc/fifos.c: define CURRENT_TIME for Linux >= 4.14.71.
* base/arch/x86/hal/hal.c: for Linux >= 4.14.71, include
<linux/sched/types.h> to find param defined in "struct sched_param".
* base/include/rtai_shm.h: cast pgd in pud_offset call, to avoid a
compilation error with Linux >= 4.14.71.
* base/sched/sched.c: fix for mmdrop and set_task_state, missing for
Linux >= 4.14.71.
* base/include/asm/rtai_lxrt.h: lazy fpu has gone, any form of fpu
counter used nomore; it is likely that version from 4.4.162 onward
inheredited the same behaviour.
* base/include/asm/rtai_fpu.h: use Linux fpregs_activate, always; in
enable_fpu, code our own clts(), Linux may have it nomore.
2018-05-21 Paolo Mantegazza <[email protected]>
* base/arch/x86/Kconfig: updated to the latest rtai_math state, glibc
used nomore.
* base/math/README.KLIBM: updated to the illustrate the latest set up.
* base/math/libm.c: remove glibc support; fix ccosh and accosh bug in
MUSL; add libm_cmplx.hc for possible missing 32 bits complex support
uclibc-ng.
* base/math/libm_cmplx.hc.gsl: added, possible support for missing 32 bits
complex uclib-ng.
2018-05-03 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: expand infos when echoing of Linuxx syscalls/traps
is enabled in hard real time, extending it to the use of linux syscalls
server.
2018-04-06 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: abort compilation if no TSC is configured.
2018-02-17 Paolo Mantegazza <[email protected]>
* base/math/libm.c, README.KLIBM: added full suport and documentation
for the use of uClibg-ng complex functions
2018-02-15 Paolo Mantegazza <[email protected]>
* base/arch/x86/Kconfig: inform that uclibc-ng is used, in place if
uclibc.
2018-02-14 Paolo Mantegazza <[email protected]>
* base/math/README.KLIBM: updated instructions to support uClibc-ng,
in place of uClib.
2018-02-08 Paolo Mantegazza <[email protected]>
* base/math/libm.c: avoid MUSL caring of complex functions if
the complex option is not configured in RTAI.
2018-02-05 Pierangelo Masarati <[email protected]>
* RTAI 5.1
2018-02-02 Paolo Mantegazza <[email protected]>
* base/math/libm.c: added own missing "void __stack_chk_fail(void)",
who does need it has still to be ascertained.
2018-01-16 Pierangelo Masarati <[email protected]>
* RTAI 5.1-test2
2017-12-27 Tom (tom@shngai) <[email protected]>
* addons/serial/seria.c: do not mess mbx->fbyte in mbxevdrp, use fbyte.
2017-12-27 Paolo Mantegazza <[email protected]>
(pointed out by Duane Jeffery <[email protected]>)
* base/include/rtai_posix.h: call the specific kernel space function
to support timer create.
* base/include/rtai_tasklets.h: declare the specific kernel space
function to support timer create.
* base/tasklets/tasklets.c: recover the missing specific kernel space
function to support timer create
2017-11-22 Pierangelo Masarati <[email protected]>
* RTAI 5.1-test1
2017-11-22 Paolo Mantegazza <[email protected]>
* base/sched/sched.c and base/arch/x86/calibration/calibrate.c:
iif rt_wait_period returns with an error in the calibrating tasks,
set the measured latency to the available estimate.
2017-11-21 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: always use rtai_simuldiv for calibrating
latencies in kernel space.
2017-11-19 Tom (tom@shngai) <[email protected]>
* base/include/asm-x86/rtai_lxrt_32.h and rtai_lxrt_64.h:
* rtai-lab/matlab/rtmain.c, rtai-lab/matrixx/mxp.c,
rtai-lab/scicoslab/rtmain44.c and rtai-lab/scilab/rtmain.c:
fixed set_double for both 32 and 64 bits architectures; the
previous version was Ok only for 32 ones.
2017-11-14 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_lxrt_32.h and rtai_lxrt_64.h:
remove their common code of rt_strncpy_from_user, fix
it to care for count and put it in the common independent
rtai_lxrt.h.
* base/include/asm-x86/rtai_lxrt.h: see above.
2017-11-13 Paolo Mantegazza <[email protected]>
* base/ipc/msg/msg.c: fixed a missing rt_strncpy_from_user in
rt_Name_locate.
* base/include/asm-x86/rtai_lxrt_32.h and rtai_lxrt_64.h:
provide and alternative version of __do_strncpy_from_user,
the existing one does not work on an Intel Ix-6xxx, likely because
of Intel MPX; the change is formally controlled by an ifdef
on the configuration parameter CONFIG_RTAI_USE_STACK_ARGS.
* base/sched/sys.c: comment out, so to be recallable, forcing the end
* base/sched/api.c: conditioned rt_exec_linux_syscall code according
to param CONFIG_RTAI_USE_STACK_ARGS.
2017-11-12 Paolo Mantegazza <[email protected]>
* base/sched/sched.c and
* base/arch/x86/hal/hal.c: undef CONFIG_TRACEPOINTS, to avoid
a missing floating point tracing symbol.
* base/include/asm-x86/rtai_fpu.h: prepared to relax undef
CONFIG_TRACEPOINTS.
2017-11-09 Paolo Mantegazza <[email protected]>
* base/tasklets/taskelets.c: added specific functions to support
posix timers.
* base/include/rtai_posix.h: added specific functions to support
posix timers.
* base/include/asm-x86/rtai_usi.h: fixed to work without exploting
user-kernel arg passing through the stack.
2017-11-07 Paolo Mantegazza <[email protected]>
* base/ipc/netrpc/netrpc.c: fixed args passing from-to user space,
by using standrd copy from-to user space, so to avoid troubles
on any hardware not allowing the simpler memcpy when args are
on the stack.
2017-11-05 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: restructured the coding of intercept_syscall,
fast and normal, to avoid possible compiler issues; return value
for the unblocked case in lxrt_intercept_syscall assigned consistently
with the possibility of directly passing, or not passing, call args
on the stack.
2017-11-05 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/smi-module.c: added a check of the
smi count.
2017-11-04 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/smi-module.c and README.SMI: added the
possibility of using setsmi to provide an SMI chipset, not yet
listed in the module and explained its use in README.SMI
* base/arch/x86/calibration/calibrate.h and calibration_helper.c:
rewritten to support SMI control for chips not yet listed in
smi_module.c.
2017-10-25 Paolo Mantegazza <[email protected]>
* testsuite/switches/switches-module.c: be sure all threads ended
before exiting the module.
2017-10-21 Paolo Mantegazza <[email protected]>
* base/ipc/netrpc/netrpc.c: instrumented to let any user easily
check if memcpy is properly used just for kernel to kernel
copies, within her/his own application.
2017-10-21 Paolo Mantegazza <[email protected]>
* base/ipc/netrpc/netrpc.c: copy _rt_net_rpc call args accordingly
to the calling space (signalled by Tom (tom@shangai) <[email protected]>,
while using qrtailab).
2017-10-21 Tom (tom@shangai) <[email protected]>
* base/arch/x86/calibration/smispv.c: check if an SMI device has
been found.
2017-10-20 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/smi-module.c: use an array of string
pointer reminders in place of a switch.
2017-10-19 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/smi-module.c: fixed to comply with what
stated in README.SMI, extended printout to dmsg what has been
done, some cleanup, added a few chipsets more.
2017-06-14 Paolo Mantegazza <[email protected]>
* addons/drivers/16550A/16550A_pci.h: Do not use the once
DEFINE_PCI_DEVICE_TABLE, eventually disappeared from the most
recent kernels used by RTAI.
2017-05-29 Paolo Mantegazza <[email protected]>
* addons/comedi/rtai_comedi.h, kcomedi-module.c: fixes to comply
with the way RTAI syscall args are exchanged between kernel and
user space,
* addons/drivers/serial/rtai_serial.h, serial.c: fixes to comply
with the way RTAI syscall args are exchanged between kernel and
user space,
* base/include/rtai_fifos.h and base/ipc/fifos/fifos.c: fixes to
comply with the way RTAI syscall args are exchanged between kernel
and user space,
2017-05-24 Aßfalg Josef <[email protected]>
Paolo Mantegazza <[email protected]>
* base/arch/x86/Kconfig: added configure param RTAI_USE_STACK_ARGS,
to be used to let the user choose how RTAI syscall args are
exchanged between kernel and user space.
* base/arch/x86/hal/hal.c: fixes to comply with the way RTAI syscall
args are exchanged between kernel and user space; print a message
showing the way parameters will be exchanged.
* base/include/rtai_lxrt.h: calls for task infos and exectime made to
comply with the way args are exchanged between kernel and user space,
* base/sched/sched.c, lxrt.c, api.c: fixes to comply with the way RTAI
syscall args are exchanged between kernel and user space,
* base/ipc/shm/shm.c: fixes to comply with the way RTAI syscall args
are exchanged between kernel and user space,
2017-05-16 Paolo Mantegazza <[email protected]>
* addons/rtdm/xn.h: keep rtdm unchanged, as far as possible, by
providing a couple of no more available patch legacy functions.
2017-05-15 Paolo Mantegazza <[email protected]>
* base/include/rtai_netrpc.h: to avoid troubles with linux > 4.1.18
set a safe number of MAX SOCKS and STUBS; a newer version of netrpc
will allow more with linux > 4.1.18 also.
* base/ipc/netrpc.c: refurbished version, supporting a large number
of SOCKS and STUBS also with linux > 4.1.18; it works as the former
release for any linux version, but needs a change to the module
for making it effective the supoort of a larger number of SOCKS
and STUBS with linux > 4.1.18.
2017-05-15 Paolo Mantegazza <[email protected]>
* addons/rtdm/xn.h: define SPIN_LOCK_UNLOCKED, if it is not defined.
2017-05-15 Pierangelo Masarati <[email protected]>
* README.CONF_RMRKS: added missing file
2017-05-09 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: fixed wrong module parm declaration for
kernel_latency and user_latency. Don't look for calibrated
latencies if any of kernel_latency and user_latency is not zero.
2016-05-05 Pierangelo Masarati <[email protected]>
* RTAI 5.0
2017-05-04 Paolo Mantegazza <[email protected]>
* README.CONF_RMRKS: a few remarks-hints about latencies and how to
avoid them.
2017-04-21 Paolo Mantegazza <[email protected]>
* addons/rtdm/rtai_taskq.c: anticipate a null task pointer check in
the while loop termination test of rt_taskq_ready_all.
2017-04-20 Paolo Mantegazza <[email protected]>
* addons/rtdm/all files: extensively revised and updated.
2017-04-18 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: as for the previous version, use a spinlock
to make rtai_rdtsc 32 bits atomic.
2017-04-16 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_hal_32.h and rtai_hal_64.h: take care
of the refurbished synchronization of the TCSes.
* base/sched/sys.c: get rid of too old kernel versions support.
* base/ipc/fifos/fifos.c: get rid of too old kernel versions support.
* base/shm/kvmem.c: get rid of too old kernel versions support.
* base/include/asm-x86: make got rid of the nomore needed 32-64 bits
specific fpu headers.
2017-04-15 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: the TSC alignment hack is now closer
to what found in arch/ia64/kernel/smpboot.c; got rid of useless
older Linux versions; exported sync_tsc for possible independent
uses.
* base/arch/x86/hal/GNUmakefile.am and GNUmakefile.in: get make rid
of no more used, arch dependent, hal_32.c and hal_64.c.
2017-04-14 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: extend multi processor TSCs alignment
to x86_64.
* base/include/asm-x86/rtai_hal_64.h: added support for correcting
TSCs misalignments.
* base/arch/x86/Kconfig: edited to be up to date with the latest
TSC alignment stuff.
2017-04-01 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: added the missing setting of the epoch
origin (rt_gettimeorig) when hard timers are started, needed
when REALTIME mode is used and some further cleaning more.
(BTW, it is not a 1st of April fool).
2017-03-28 Paolo Mantegazza <[email protected]>
* base/include/rtai_sanity.h: allow using gcc version up to 5.
2017-03-25 Paolo Mantegazza <[email protected]>
* base/include/rtai_nam2num.h: added the construct extern "C" { }
to make it C++ friendlier.
* base/include/rtai_sched.h: added the declaration of the function
set_rtext, to avoid repeating it when used in a few files.
* base/ipc/netrpc.c: use the new declaration of set_rtext, omitting
a nomore used argument.
* base/sched/sys.c: use the new declaration of set_rtext, omitting
a nomore used argument.
* base/sched/sched.c: wiped out unused code, shortening the call
list, because of a nomore used arg.
2017-03-21 Paolo Mantegazza <[email protected]>
* base/include/rtai_shm.h: fixed a few warnings and cleaned old
stuff.
* addons/comedi/rtai_comedi.h: fixed warnings and a couple of
distributed functions error, appearing only in C++.
2017-03-19 Paolo Mantegazza <[email protected]>
* base/wd/README.WATCHDOG: edited to inform that oneshot is the
only hard timer mode in use now.
* base/include/rtai_wd.h: wiped away a comment suggesting possible
non oneshot hard time mode.
* base/wd/wd.c: force wd_OneShot to be always 1, oneshot is the
only hard timer mode in use now.
* base/sched/api.c: no more references to any hard timer mode but
oneshot, be it in comments or code.
* base/sched/sched.c: cleaned dead code, wiped out anything
referencing non oneshot hard timer mode, the only one in use
now.
2017-03-15 Paolo Mantegazza <[email protected]>
* base/include/rtai_shm.h: wiped unused code away.
* base/include/rtai_schedcore.h: wiped unused code away, cleaned
the code related to the nomore used periodic mode.
2017-03-12 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_fpu.h: update for newer kernels,
code reworked, compacted and stremlined, no more per 32-64
bit archs specific includes.
2017-03-11 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: make it compile with FPU configured as
disabled.
2017-03-10 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: update for newer kernels.
* base/sched/sys.c: update for newer kernels.
* base/include/asm-x86/rtai_lxrt.h: update for newer kernels.
2017-03-09 Paolo Mantegazza <[email protected]>
* base/include/rtai_lxrt.h, rtai_msg.h, rtai_netrpc.h, rtai_sem.h: fixed
a bunch of unharmful warnings, quite annoying in large compilations,
especially under g++, e.g (q)rtailab.
2016-12-07 Paolo Mantegazza <[email protected]>
* base/arch/x86/ihal/hal.c: fixed inconsistent declaration of
cpu_isolated_map for the UP case.
2016-08-21 Sebastian Kuzminsky <[email protected]>
* base/math/README.KLIB: fixed a few typos.
* base/sched/sched.c: set Linux current task state using its suuport,
in place of doing it by hand.
* base/sched/api.c: nullify lxrt_list after vfreeing what it points at;
fix the setting of rtai task extension in the old registry way; avoid
a redundant spin unlock when hash_ins_adr is satisfied in hash_ins_name;
add supposedly missing spin unlocks in functions: hash_ins_adr,
hash_ins_name, hash_find_name, hash_find_adr (just a formal error,
a branch of the if-then-else construct should always be satisfied
anyhow).
2016-04-29 Pierangelo Masarati <[email protected]>
* RTAI 5.0-test2
2016-04-20 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: abort compilation if CONFIG_IPIPE_LEGACY,
trivial way to avoid undue confusion and troubles in the use of
RTAI.
2016-04-20 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/smi-module.c: updated SMI chipset
list.
2016-04-19 Pierangelo Masarati <[email protected]>
* base/arch/x86/calibration/GNUmakefile.am: conditional compilation
of smispv
* configure.in: added CONFIG_RTAI_LATENCY_SELF_CALIBRATION_METRICS;
conditional compilation of smispv
2016-04-18 Marco Morandini <[email protected]>
* base/arch/x86/calibration/smispv.c: added interactive SMI
supervisor-monitor in user space; requires: libncurses, libpci,
libsensors.
2016-04-11 Paolo Mantegazza <[email protected]>
* base/arch/x86/Kconfig: added parameter to chooce latency calibration
metrics.
* base/arch/x86/calibration/calibrate.c: calibrate user space latency
according to the chosen metrics.
* base/sched/sched.c: calibrate kernel space latency according to the
chosen metrics.
* base/arch/x86/hal/hal.c: set allowed cpus to be portable across
differing Linux versions, fix printk warnings and linux version
dependency of cpu_isolated_mask.
2016-04-05 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/calibrate.c: added a fading memory
option for the user space latency calibration and an option
for playing with an alternative Hinf filtering.
* base/sched/sched.c: added a fading memory option for the kernel
space latency calibration.
2016-04-04 Pierangelo Masarati <[email protected]>
* base/arch/x86/Kconfig, base/arch/x86/defconfig, configure.in:
redefined self-calibration macros and defaults.
* base/sched/sched.c: added Kalman filter based kernel latency
calibration, as for user space but with integer arithmetics.
2016-04-02 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/calibrate.c: added a calibration through a
Kalman filter.
2016-03-31 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/calibrate.c: save the scheduling latencies
in nanoseconds, to make it easier their editing.
* base/sched/sched.c: echo diagnose of the scheduling latencies in
nanoseconds, to make it easier their reading; be sure of the kernel
thread server end at rmmod rtai_sched.
2016-03-30 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibration/calibrate.c: rewrite user space support
for calibrating scheduling latencies on the fly.
* base/include/rtai_lxrt.h: recast the kernel_calibrator function
as a simple interface to provide kernel and user space latencies,
in clock count units.
* base/include/rtai_sched.h: add per task calibrated latency variable.
* base/include/rtai_schedcore.h: wake up timed tasks, immediately
accounting for the available per task schedling latency.
* base/sched/sched.c: rework the all kernel/user space self calibration
stuff.
* base/sched/sys.c: recast the kernel_calibrator function as a simple
interface to provide kernel and user space latencies, in clock count
units.
2016-03-29 Paolo Mantegazza <[email protected]>
Pierangelo Masarati <[email protected]>
* base/arch/x86/Kconfig and configure.in: removed old self
calibration parameters, substituted with the new ones,
i.e.: self calibration period and time.
* base/sched/GNUmakefile.am: removed old self calibration
stuff
* base/include/rtai_defs.h.in and base/include/GNUmakefile.am:
generate macro for installation path
2016-03-28 Paolo Mantegazza <[email protected]>
* testsuite/user/latency.c: reworked heading lines and added chache
warmup.
* testsuite/kthreads/display.c and latency.c: reworked heading lines
and added chache warmup.
* testsuite/kern/display.c and latency.c: reworked heading lines and
added chache warmup.
2016-03-23 Paolo Mantegazza <[email protected]>
+* base/sched/sched.c: hard timers started-stopped in the scheduler,
oneshot only, related user calls emptied.
2016-03-12 Paolo Mantegazza <[email protected]>
* addons/rtdm/rtdm.h: a, likely provisional, fix to support
Linux-4.x.x.
2016-03-12 Emmanuel Pacaud <[email protected]>
Paolo Mantegazza <[email protected]>
* base/sched/sched.c and sys.c: fix the wrong placement of a task
always on all available CPUs when more than 1 is allowed, without
carying of the allowed mask.
2016-03-10 Paolo Mantegazza <[email protected]>
* base/ipc/fifos/fifos.c: recover the macro defining f_dentry from
< LINUX 4.0.0 kernels, to avoid recoding.
* base/arch/x86/hal/hal.c: treat cpu_isolated map according to the
kernel version, changed with LINUX 4.x.x.
* base/ipc/netrpc/netrpc.c: move rtdm.h inclusion where HARD_NET is
to be used.
2015-12-10 Paolo Mantegazza <[email protected]>
* base/arch/x86/defconfig: set the use of new error return code by
default; do not configure COMEDI usage innuser space by default.
2015-12-04 Pierangelo Masarati <[email protected]>
* RTAI 5.0-test1
2015-12-03 Paolo Mantegazza <[email protected]>
* base/math/libm.c: added an internal implementation of cpow for MUSL,
it fixes a segfault when the MUSL one is used for 32 bits; added
an internal MUSL implementation of: gamma_r, gamma, gammaf_r,
gammaf, whose sources are available in MUSL while the related entries
do not appear in libc.a.
2015-11-30 Paolo Mantegazza <[email protected]>
* base/math/libm.c: adjusted a missing macro in the test enabling the
use of complex numbers support; added comments about the complex
mul/div supported codes.
2015-11-29 Paolo Mantegazza <[email protected]>
Marco Morandini <[email protected]>
* base/math/libm.c: added support for linking to MUSL libc.a, which
embeds libm.a.
* base/math/export_musl.h: adds export math entries supported through
MUSL libc.a.
* base/math/README.KLIBM: updated with explanations for using MUSL
libc.a.
* base/arch/x86/Kconfig: updated to allow configuring math support
for MUSL.
2015-11-25 Paolo Mantegazza <[email protected]>
* base/sched/sys.c: fix the setting of the return value of the
IS_HARD enquiry so that it cannot return wrong values.
2015-11-14 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: brought rtai_hirq_dispatcher here, where the
possible deferred sched in intr belongs; see the related item
at RTAI config time.
* base/arch/x86/hal/hal.c: simplify rtai_hirq_dispatch for non
scheduling times, sched.c will take it. Ensure that no TSC syncing
is undertaken on x86_64.
* base/ipc/fifos/fifos.c: to avoide an annoying warning, fixed a not
needed initialization of struct rt_fifo_get_info_structr at
RTF_GET_FIFO_INFO.
2015-11-11 John Wilkes <[email protected]>
2015-11-11 Paolo Mantegazza <[email protected]>
* addons/rtdm/drvlib.c: fixed map/unmapping to work with Linux 3,
updated with a few services more.
* addons/rtdm/rtdm_drvlib.h: updated with a few needed declarations.
2015-10-28 Pierangelo Masarati <[email protected]>
* RTAI 5.0
2015-09-24 Pierangelo Masarati <[email protected]>
* RTAI 4.1.1 (4.1 closure release)
2015-05-09 Paolo Mantegazza <[email protected]>
* base/addons/rtdm/rtdm_driver.h, xn.h: support changing, deprecated
and not, smp_mb__before_atomic versions.
2015-04-30 Pierangelo Masarati <[email protected]>
* base/math/export_newlib.h, base/math/export_uclibc.h: added
missing files
* base/math/GNUmakefile.am: added missing dependences
2015-04-18 Paul Corner <[email protected]>
* base/scripts/rtai-config.in: added datarootdir; applied and checked
by Pierangelo (Masarati).
2015-04-13 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: if CONFIG_SMP is not defined set up an
internal declaration for it.
2015-04-07 Pierangelo Masarati <[email protected]>
* base/include/GNUmakefile.am: make selfcalibration work also
for VPATH builds
2015-04-03 Paolo Mantegazza <[email protected]>
* base/include/rtai.h: insert dummy rt_mount_rtai and rt_umount_rtai
to make comedi making happy.
2015-03-25 Shahbaz Youssefi <[email protected]>
* rtai-py/rtai_lxrt.py: fixed nam2num and num2nam typos, argstype
used in place argtypes.
2015-03-22 Paolo Mantegazza <[email protected]>
* base/sched/sys.c: revised lxrt_invoke, argument list and catching
Linux signal at entering moved to syscall interception in sched.c.
* base/sched/sched.c: revised linux signal wakeup and RTAI-Linux
syscalls intercpting.
2015-03-19 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: omit argument in macros
"...fast_flush_pipeline", not needed any more and return though linux
after intercepting syscalls, theu can never be in hard real time
without the scheduler.
* base/include/rtai_schedcore.h: make TASK_HARDREALTIME and
TASK_RTAISRVSLEEP as TASK_INTERRUPTIBLE.
* base/include/asm-x86/rtai_lxrt.h: experiment with the possibility of
* base/sched/api.c: split the function rt_task_masked_unblock in a
basic unprotected function, to be used in specifically protected
forms, both in api.c and wherever it is needed.
2015-03-17 Paolo Mantegazza <[email protected]>
* base/ipc/msg/msg.c: improve the check of the sending task in
rt_Receive.
2015-03-06 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: type rtai_hirq_dispatcher as void, no
return value used by the function calling it.
* base/sched/sched.c: removed the unused code related to the old way
of using a list of tasks to transition to hardmode
(lxrt_intercept_schedule_tail).
2015-03-06 Pierangelo Masarati <[email protected]>
* RTAI 4.1
2015-03-04 Pierangelo Masarati <[email protected]>
* configure.in, configure: revised to support self calibration and
revised configuration paraeters.
* base/sched/selfcalibrate.sh: added to support self calibration.
* base/arch/x86/Kconfig: restructured calibration parameters structure
names and help, both to clean no more used names and to support self
calibration.
* base/arch/x86/calibration/calibrate.c: modified to be usable both
for self and user guided calibrations.
2015-02-27 Pierangelo Masarati <[email protected]>
* configure.in, configure, testsuite/kthreads/*/GNUmakefile.*,
testsuite/kthreads/*/Makefile.kbuild: add prefix kthreads_ to
modules testsuite/kthreads, showing the use of hardened Linux
kernel threads in the same way as user space threads.
* base/arch/x86/calibration/GNUmakefile.*: module rtai_calibrate.ko
no longer built; build calibrate instead.
2015-02-22 Paolo Mantegazza <[email protected]>
* base/sched/sys.c: avoid waiting on a Linux sem when a kthread
creation is requested from hard real; the Linux task structure
address is returned through a call argument.
2015-02-21 Paolo Mantegazza <[email protected]>
* testsuite/kthreads/latency/display.c and latency-module.c,
testsuite/kthreads/preempt/display.c and preempt-module.c,
testsuite/kthreads/switches/switches-module.c: set of examples
showing the usage of RTAI enabled soft-hardened Linux kthreads.
* base/sched/sys.c: added a Linux kthread server, to allow creating
hadr-soft RTAI Linux kernel threads from RTAI own kernel tasks and
Linux real time hardened kernel thread, without loosing hard real
time execution; it implies a revision of what prviously committed.
* base/sched/sched.c: run and termininate, at init-exiting the Linux
kthread server.
2015-02-19 Paolo Mantegazza <[email protected]>
* base/arch/x86/calibrate/calibrate.c: do not self calibrate if
SCHED_LATENCY is set already; update also .rtai-config.
* base/sched/sched.c: do not convert negative kern-user busy waits.
* base/arch/x86/Kconfig: display kern busy wait again.
2015-02-18 Paolo Mantegazza <[email protected]>
* base/sched/sched.c and sys.c: busy timed sync waits only if there
is a sched_latency.
* base/sched/sys.c: set sched_latency also when sched_latency is zero.
* base/include/rtai_schedcore.h: enable timed sync busy waits at wake
up only if there is a sched_latency.
* base/arch/x86/calibrate/calibrate.c: recoded for use within the
make self tuning.
2015-02-16 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_fpu_64.h: cleaned legacy oldnames.
2015-02-15 Paolo Mantegazza <[email protected]>
* base/sched/sys.c: rt_thread_create returns an error when called from
an RTAI own kernel task; use a wrapper function and a LINUX kernel
sem to grant the new kthread is running before retunring,
substitutes the previous polled wait for an RTAI enabled new khread.
2015-02-14 Paolo Mantegazza <[email protected]>
* addons/drivers/serial/serial.c and serialP.h: cleaned legacy
oldnames.
* addons/rtdm/drvlib.c, module.c, rtdm_driver.h and xn.h: cleaned
legacy oldnames, aligned with renewed variable and structure names.
* base/arch/x86/Kconfig: prepared for a few next config changes.
* base/arch/x86/hal/hal/c: cleaned legacy oldnames and macros, aligned
with renewed variable and structure names.
* base/include/rtai_hal_names.h: reduced to what is still needed with
the latest patches and a more direct use of its raw names.
* base/include/asm-x86/rtai_hal.h: redefined struct calibration_data,
cleaned legacy oldnames, messy redundant macros and unused parts of
code; defined new macros, aligned with renewed variable and structure
names.
* base/include/asm-x86/rtai_lxrt.h: cleaned legacy oldnames, messy
redundant macros and unused parts of code; defined new macros,
aligned with renewed variable and structure names.
* base/include/asm-x86/rtai_vectors.h: cleaned of vectors and ipis
macros definition, introduced a reminder of what in use now.
* base/sched/api.c: cleaned legacy oldnames and cared of the no more
used rt_times tunables, now rt_smp_times[0].
* base/sched/sched.c: cleaned legacy oldnames, messy redundant macros;
aligned with renewed variable, macros and structure names; cared of
the no more used rt_times tunables, now rt_smp_times[0].
* base/sched/sys.c: cleaned legacy oldnames, aligned with renewed
variable and structure names.
* base/tasklets/tasklets.c: cleaned legacy oldnames, aligned with
renewed variable and structure names.
* base/wd/wd.c: cleaned legacy oldnames, aligned with renewed variable
and structure names.
* testsuite/kern/latency/latency-module.c,
testsuite/kern/preempt/preempt-module.c,
testsuite/kern/switches/switches-module.c: cleaned legacy oldnames,
aligned with renewed variable and structure names.
2015-02-11 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: got rid of rt_times variable, beware
just the variable, not of the structure with the same name.
2015-02-08 Paolo Mantegazza <[email protected]>
* base/sched/sys.c: rt_thread_create poll waits till the created
thread is rtai enabled.
2015-02-07 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: added the possibility of testing different ways
to set active_mm for kernel threads.
2015-02-06 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: avoid loosing the active_mm stolen for kernel
threads working in hard real time, by incrementing its mm_count
when it is acquired in fast_schedule, then mmdrop active_mm any
time they are returned to soft mode in give_back_to_linux.
2015-02-05 Paolo Mantegazza <[email protected]>
* base/include/rtai_sched.h: added the new declaration for
rt_kthread_init, hiding the old one. Still to be cleaned a bit.
* base/include/rtai_schedcore.h: hided the old declaration of
rt_kthread_init. Awaiting some cleaning.
* base/sched/sched.c: made fast_schedule support hard transitions
for kernel threads, without touching the context switch in use.
* base/sched/sys.c: reinstated support for hard-soft kernel threads,
emulating the scheme used in kernel space. Beware of the way of
calling RTAI services in the case of soft threads. See
soft calling functions. The legacy RESERVOIR option awaits a new
helper module. TBD: the releated testsuite.
2015-02-04 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: got rid of all old timer request/free.
* base/sched/sched.c: got rid of all old start/stop_rt_timer and
likes. Added fastcall hook.
2015-02-02 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_hal.h: macro rtai_tskext modified so to
include the related LINUX task, and to use either ptd in task_struct
and thread_info.
* base/ipc/msg/msg.c;
base/ipc/shm/shm.c;
addons/rtdm/xn.h;
base/sched/api.c, sched.c, sys.c: use the new macro rtai_tskext,
modified so to include the related LINUX task.
2015-02-01 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: force oneshot mode ALWAYS, through a bad, but
temporary, hack.
2015-01-31 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_lxrt.h: gather the common parts of 32-64
bits archs, unify the definition of union rtai_lxrt_t.
* base/include/asm-x86/rtai_lxrt_32.h: keap only the 32 bits specific
definitions and services, required copying __do_strncpy_from_user
from LINUX, as for 64 bits.
* base/include/asm-x86/rtai_lxrt_64.h: keap only the 64 bits specific
definitions and services.
* base/include/asm-x86/rtai_fpu.h: gather the common parts of 32-64
* base/include/asm-x86/rtai_fpu_32.h and rtai_fpu_64.h: keap only the
32-64 bits specific bits archs.
2015-01-31 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c: significantly refurbished, x86-32-64
unified, patch's hard timers services; from now on only the oneshot
mode must be used.
2015-01-28 Pierangelo Masarati <[email protected]>
* RTAI 4.1-test1
2015-01-27 Paolo Mantegazza <[email protected]>
* base/include/asm/rtai_hal.h: unify 32-64 bits independent code here
and revise real time spin locks and global lock.
* base/include/asm/rtai_hal_32.h and rtai_hal_64.h: left with only
32-64 bits arch specific code.
* base/sched/sched.c: use non rtai_cli protected global lock get and
release from rtai_hal.h.
* base/arch/x86/hal/hal_32.c: redefine rtai_cpu_lock according to the
global lock structure.
* addons/rtdm/xn.h: make xnlock_t comply with the new global_lock
structure declaration.
2015-01-24 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal_32.c: exploit oneshot hardware timing setting
from the modified patch, working but still to be finished along with
some cleaning more.
* base/include/asm-x86/rtai_hal_32.h: optionally set the oneshot delay
by optionally using ipipe_timer_set if X86_FEATURE_TSC_DEADLINE_TIMER.
* base/sched/sched.c: do not provide rt_linux_hrt_set_mode any more,
use the one found of rtai_hal.c.
2015-01-22 Paolo Mantegazza <[email protected]>
* base/sched/sched.c: serialize symultaneous uses of steal_from_linux
on the same cpu, directly pass fast_schedule as a migration_hook for
the patch; removed rt_daemonize stuff.
* base/sched/sys.c: removed rt_daemonize stuff.
2015-01-18 Paolo Mantegazza <[email protected]>
* addons/rtdm/rtdm_driver.h and xn.h: a few changes to work also
without legacy constraints.
* base/arch/x86/hal/hal_32.c and hal_64.c: a series of changes to work
also without legacy constraints and some cleaning.
* base/include/rtai_hal_names.h: cleaned up, also in relation to a few
changes to work also without legacy constraints.
* base/include/rtai_hal_schecore.h: a few changes to work also without
legacy constraints.
* base/include/asm-x86/rtai_hal_32.h: a few changes to work also without
legacy constraints.
* base/sched/sys.c: grant zeroing of ptd and use the patch support
* base/sched/sched.c: some cleaning and changes to work also without
legacy constraints.
2015-01-18 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal_32.c and hal_64.c: warn for the incosistency:
IsolCpusMask && (IsolCpusMask != cpu_isolated_map).
2015-01-14 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal.c, hal_32.c and hal_64.c: if IsolCpusMask is
null use that of Linux, so that there is nomore the need to set it
in RTAI.
2015-01-14 Paolo Mantegazza <[email protected]>
* base/arch/x86/hal/hal_32.c and hal_64.c : avoid reprogramming the
apic timer mode when tsc-deadline is available.
2015-01-14 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_hal_32.h and rtai_hal_64.h: reworked
rt_set_timer_delay.
* base/include/asm-x86/rtai_hal.h: reorder a set of declarations.
* base/sched/sched.c: avoid multiple oneshot delay units conversion,
unified in rt_set_timer_delay.
2015-01-14 Paolo Mantegazza <[email protected]>
* base/include/asm-x86/rtai_lxrt_32.h and rtai_lxrt_64.h: update
caring of the Linux version when manupulating the fpu_counter in
context switches.
* base/sched/sched.c: care of Linux version in setting need_resched.
2015-01-14 Pierangelo Masarati <[email protected]>
* configure.in: fix linux source dir extraction when further args
follow -C, thanks to Lukas Anzinger <[email protected]>
* configure: rebuild
2015-01-14 Pierangelo Masarati <[email protected]>
* base/sched/{GNUmakefile.am,sched.c}: use MODULE_ALIAS()
instead of manual symlinks for legacy schedulers,
thanks to Lukas Anzinger <[email protected]>
* base/sched/GNUmakefile.in: regenerate
2015-01-14 Pierangelo Masarati <[email protected]>
* configure.in: fix linux source dir extraction when further args
follow -C, thanks to Lukas Anzinger <[email protected]>
* configure: rebuild
2015-01-14 Pierangelo Masarati <[email protected]>
* addons/rdtm/GNUmakefile.*: updated for Linux >= 3.10.0.
2015-01-14 Paolo Mantegazza <[email protected]>
* addons/rtdm/vfile.h, vfile.c and proc.c: updated for Linux >= 3.10.0,
lifted from native nucleus and rtdm.
* addons/rtdm/rtdm_driver.h: added declaration to support the new proc
file system.
2015-01-14 Paolo Mantegazza <[email protected]>