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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
|
<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE html PUBLIC "-//FreeBSD//DTD XHTML 1.0 Transitional-Based Extension//EN"
"http://www.FreeBSD.org/XML/share/xml/xhtml10-freebsd.dtd" [
<!ENTITY email 'hackers'>
<!ENTITY break '<div xmlns="http://www.w3.org/1999/xhtml"><hr/><center>BREAK</center><hr/></div>'>
<!ENTITY title "Usenix 2002 FreeBSD Developer Summit III">
]>
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>&title;</title>
<cvs:keyword xmlns:cvs="http://www.FreeBSD.org/XML/CVS">$FreeBSD$</cvs:keyword>
</head>
<body class="navinclude.about">
<p>The third FreeBSD Developer Summit was held on June 11-12, 2002, at
the Monterey Marriott in Monterey, CA, immediately prior to the USENIX
Annual Technical Conference at the same location. The FreeBSD
Developer Summit was sponsored by DARPA, the <a
href="https://www.freebsdfoundation.org">FreeBSD Foundation</a>, <a
href="http://www.freebsdmall.com/">FreeBSD Mall</a>, Network
Associates Laboratories, and AT&T. Notes were taken by George
Neville-Neil, <a href="http://people.FreeBSD.org/~bmah">Bruce Mah</a>,
and <a href="https://www.FreeBSD.org/~rwatson">Robert Watson</a>.
Markup by <a href="https://www.FreeBSD.org/~murray">Murray
Stokely</a>.</p>
<p>These notes cover day 2, which began at 9:30am, and ended at 5:00pm.</p>
<h2>Agenda</h2>
<ul>
<li>Opening Comments</li>
<li><a href="#kse">KSE</a></li>
<li>Break</li>
<li><a href="#smp">SMPng</a></li>
<li>PGP key signing</li>
<li><a href="#hw">New Hardware Architectures</a></li>
<li>Lunch</li>
<li><a href="#trust">TrustedBSD</a></li>
<li><a href="#releng">Release Engineering</a></li>
<li>Break</li>
<li><a href="#rc">RCng</a></li>
<li><a href="#open">Open Discussion</a></li>
</ul>
<p>NOTE: As usual I missed some names, please add those I missed.</p>
<h2>Attending:</h2>
<p>Committers in person:</p>
<ul>
<li>Robert Watson (rwatson)</li>
<li>Julian Elischer (julian)</li>
<li>John Baldwin (jhb)</li>
<li>Matt Dillon (dillon)</li>
<li>David O'Brien (obrien)</li>
<li>Jeffrey Hsu (hsu)</li>
<li>Jennifer Yang (jennifer)</li>
<li>Bosko Milekic (bmilekic)</li>
<li>Alfred Perlstein (alfred)</li>
<li>Doug Rabson (dfr)</li>
<li>Paul Saab (ps)</li>
<li>Brooks Davis (brooks)</li>
<li>Murray Stokely (murray)</li>
<li>Jonathan Mini (mini)</li>
<li>Takanori Watanabe (takawata)</li>
<li>Gordon Tetlow (gordon)</li>
<li>Gregory Shapiro (gshapiro)</li>
<li>Sam Leffler (sam)</li>
<li>Bruce Mah (bmah)</li>
<li>David Malone (dwmalone)</li>
<li>Ian Dowse (iedowse)</li>
</ul>
<p>Also in person:</p>
<ul>
<li>George Neville-Neil (gnn)</li>
</ul>
<p>On The Phone:</p>
<ul>
<li>Alan Cox (alc)</li>
<li>Warner Losh (warner)</li>
</ul>
<p>Via webcast:</p>
<p>Many people listened into the webcast and discussed the topics on
IRC.</p>
<p>The meeting followed a format where each section was led by an
individual and then a discussion ensued. Not all of the discussion
was caught but we have tried to make those sections
understandable.</p>
<div class="topic">
<hr/>
<h2>09:30 Opening Remarks</h2>
</div>
<div class="topic">
<hr/>
<a name="kse"></a>
<h2>KSE - Julian Elischer</h2>
<p>KSE has not changed much since the last summit (Feb). The major
change is that upcalls works more like signals instead of like fork().
That is to say that you give the system a function pointer to call
instead of having the "return twice" semantics so that it supports
all architectures.</p>
<p>The names in the system are deliberately different from other
threading packages. This was to reduce confusion during
development.</p>
<p>The process structure has been broken into 4 parts. This is in
-CURRENT at the moment. It's still "really" one structure but is
being accessed as 4 different ones.</p>
<p>Looking for more people to run the code.</p>
<p>Right now we're rewriting to clean up how the functions work.</p>
<p>Other architectures can be stubbed out as well.</p>
<p>Right now there is no support for Sparc or IA64 but he would like
to commit now. Not committing now means that it has to come out of
perforce and people have to patch it.</p>
<h3>Discussion</h3>
<div class="discussion">
<p><strong class="speaker">rwatson</strong> : What about userland?</p>
<p><strong class="speaker">julian</strong> : It can run different threads
in userland. The primitives are all there it just needs a bit more
help. I would like to put an idea out. Is it a good idea to be able
to have non-threaded programs linking with threaded libraries?</p>
<p><strong class="speaker">rwatson</strong> : Putting async I/O into such a
thing would make sense.</p>
<p><strong class="speaker">julian</strong> : The library would not care
who was accessing it.</p>
<p><strong class="speaker">rwatson</strong> : For instance libc could be
threaded or not.</p>
<p><strong class="speaker">julian</strong> : That would be interesting. I
don't know if the two interfaces are incompatible.</p>
<p><strong class="speaker">jhb</strong> : X does this.</p>
<p><strong class="speaker">dillon</strong> : It is very doable but you
have to make it non-preemptive. If you're switching non-preemptively
you can use library routines which are non threaded.</p>
<p><strong class="speaker">julian</strong> : If I do what I'm thinking of
doing then each lib will have its own KSE group.</p>
<p><strong class="speaker">dillon</strong> : stdio does not have to be
thread aware if you don't schedule preemptively. It all matters where
it blocks.</p>
<p><strong class="speaker">julian</strong> : Since you're a non-threaded
program you don't know that.</p>
<p><strong class="speaker">rwatson</strong> : If you're going to support
that, libc has to support threads.</p>
<p><strong class="speaker">rwatson</strong> : It sounds like some
complexity goes away. Can we use 1 libc with has threading?</p>
<p><strong class="speaker">julian</strong> : Do we want to go down this
path?</p>
<p><strong class="speaker">rwatson</strong> : Now or later?</p>
<p><strong class="speaker">julian</strong> : What do I design now to do
this?</p>
<p><strong class="speaker">jhb</strong> : For example libc_r does not
work with rfork.</p>
<p><strong class="speaker">julian</strong> : The answer is that yes we
should move forward. Tricky issues, signals...</p>
<p><strong class="speaker">warner</strong> : Have people talked about
pthread programs and cancellation points?</p>
<p><strong class="speaker">julian</strong> : The pthreads library does not
assume that you're only going to change threads at yield() points. We
are going to have cancellation points. There is an unimplemented call
which will be able to send a thread targeted signal even into the
kernel.</p>
<p><strong class="speaker">julian</strong> : When a thread is scheduled
onto a KSE there is a mailbox that the userland thread scheduler
updates.</p>
<p><strong class="speaker">julian</strong> : Is there anyone else who has
some time or test it? How many people should test this before I check
it in? There is a patch that's continuously updated on my web site to
be able to patch it to -CURRENT. There is a CVSUP target from cvsup
10 which will bring down the sources. If you go to my web page on
freefal there is a pointer there to a web page that explains how to
CVSUP from source.</p>
<p><strong class="speaker">rwatson</strong> : What about SMP locking for
this?</p>
<p><strong class="speaker">julian</strong> : Handled by the proc locking.
Has not been tried on SMP machines yet.</p>
<p><strong class="speaker">obrien</strong> : What about on Sparc?</p>
<p><strong class="speaker">julian</strong> : You may need to stub things
out.</p>
<p><strong class="speaker">jhb</strong> : Is the paper on the web site?</p>
<p><strong class="speaker">julian</strong> : The updated copy has disappeared.</p>
<p><strong class="speaker">unknown</strong> : What's the different between
NetBSD and FreeBSD on this?</p>
<p><strong class="speaker">julian</strong> : Logically not a tremendous
difference but Net follows the paper closely and Free takes the idea
and makes it into a production system. There were some tough battles
on -arch about this. The tricky point is that the proc structure has
to be broken into 4 instead of 2. If you want to be able to do POSIX
you need to be able to treated as different processes but in other
systems you need to group the threads. You wind up with two classes
of threads. You need another structure to do the aggregation. In the
end we ended up breaking up the proc structure into 4 pieces to not
overwhelm the CPU when scheduling threads. This is the major
difference.</p>
<p><strong class="speaker">julian</strong> : I greatly admire the NetBSD
way which is to take an idea and not dilute it.</p>
<p><strong class="speaker">julian</strong> : Net is also putting a Solaris
compatible threads package on top of their scheduler activations in
the Solaris ABI.</p>
</div>
</div>
<div class="topic">
<hr/>
<a name="smp"></a>
<h2>SMPng - John Baldwin</h2>
<div class="discussion">
<p><strong class="speaker">jhb</strong> : Yesterday we talked about SMP
related things so I'll give a summary and then give a list of things
for 5.0.</p>
<p><strong class="speaker">jhb</strong> : The big thing for 5.0 is to
get the network stack out from under Giant.</p>
<p><strong class="speaker">jhb</strong> : Jeffrey Hsu and Jennifer Yang
were here to talk about this. They have the PCBs checked in now.</p>
<p><strong class="speaker">jennifer</strong> : Interface Queues and SynCache
might be done.</p>
</div>
<p>The remaining chunks of the network code are:</p>
<ul>
<li>Routing Table Locking</li>
<li>Socket Buffers</li>
<li>ifa structures</li>
<li>struct ifnet</li>
<li>locking ipfw and dummynet</li>
<li>other (non IPv4) protocols</li>
<li>Netgraph</li>
</ul>
<div class="discussion">
<p><strong class="speaker">jhb</strong> : Aside from network the newbus
locking needs to be done (Warner Losh) and also CAM stuff. No known
status on CAM. Perhaps CAM is not needed for 5.0</p>
<p><strong class="speaker">jhb</strong> : Disk drive interrupts? Would
help performance. Going to talk to Poul Henning-Kamp</p>
<p><strong class="speaker">jhb</strong> : Alan Cox is working on the VM
system. Working based on the old Mach stuff. Objective for 5.0 is to
get zero fill and execute on write to work without Giant. In future
he wants to look at locking down pmap() functions.</p>
<p><strong class="speaker">jhb</strong> : Still some stability issues.
UMA breaks some assumptions. For instance sockets assume that once
memory is a socket its a socket forever, this is no longer true.</p>
<p><strong class="speaker">jhb</strong> : Talked to Mike Smith about
5.0 and have decided to stop adding features so that we can start
clean up 5.0 and make it a real release. This might require hacks.</p>
<p><strong class="speaker">rwatson</strong> : For example in the UMA case
there could be a flag to just say "don't reclaim this zone" -- this
would help with issues such as the socket code assuming memory is type
stable.</p>
<p>Over to Alan Cox on the VM system.</p>
<p><strong class="speaker">alc</strong> : Nothing to say.</p>
<p><strong class="speaker">bmilekic</strong> : As much as I might get hated
for this. Will preemption stuff go in by 5.0?</p>
<p><strong class="speaker">jhb</strong> :No, that's a 6.0 thing. There
are things to do first.</p>
<p><strong class="speaker">unknown</strong> : Could this come in in
the life time of 5.? 5.1?</p>
<p><strong class="speaker">rwatson</strong> : This is a release issue really.</p>
<p><strong class="speaker">jhb</strong> : Yes, the kernel is pre-emptive.</p>
<p><strong class="speaker">rwatson</strong> : Perhaps we should talk about
is performance goals? What are the comparisons to make? Perhaps head
of 4 with head of 5. We'll see a mix.</p>
<p><strong class="speaker">jhb</strong> : I need to run benchmarks.</p>
<p><strong class="speaker">rwatson</strong> : In terms of SMP features when
will VM be ready to be measured? I can't put a date on it.</p>
<p><strong class="speaker">alc</strong> : I think I told John was in
time for release. I'm already doing performance testing so we've
already started.</p>
<p><strong class="speaker">rwatson</strong> : We'll pick a date to start
doing measurements. Perhaps 2 or 3 weeks from now.</p>
<p><strong class="speaker">alc</strong> : My guess is the the locking
pmap is going to take some time to shake out. On the other hand the
next major module we should be working on is machine dependent level.
Last we should try approaching the vmobject level. I'll start
worrying about performance in the near term.</p>
<p><strong class="speaker">rwatson</strong> : Will threading improve
latency or throughput for networking?</p>
<p><strong class="speaker">bmilekic</strong> : I would like if we could
actually start before.</p>
<p><strong class="speaker">rwatson</strong> : Do you have a timeline for
the interrupt threading stuff?</p>
<p><strong class="speaker">bmilekic</strong> : I finished some things last
night but there are still issues. In a couple of weeks it should be
ready for first commit.</p>
<p><strong class="speaker">rwatson</strong> : Informally beginning to
measure performance now. What are the right sets of tests? Need to
discuss on -arch.</p>
<p><strong class="speaker">alc</strong> : It would be nice to have that
committed to the tools directory.</p>
<p><strong class="speaker">jhb</strong> : The statistics analysis
package are we using.</p>
<p><strong class="speaker">bmilekic</strong> : I have some good success with
netpipe for overall measurement.</p>
<p><strong class="speaker">rwatson</strong> : Need to be using consistent
compilers because of the compiler change. Also all our debugging
stuff will slow down the benchmarking.</p>
</div>
<p>Benchmark Ideas</p>
<ul>
<li>chroot of 4.6 area for testing</li>
<li>buildworld</li>
<li>build X11R4 w/gcc 2.95</li>
<li>netpipe on loopback</li>
<li>end to end testing on on gigE cards
(throughput,connections/sec)</li>
<li>thread stuff like mySQL</li>
<li>Kirk's FS benchmarks</li>
<li>NFS Tests (nfsstone)</li>
<li>Sleepycat DB tests?</li>
</ul>
<p>Tests to be run on:</p>
<ul>
<li>4.6-RELEASE</li>
<li>5.0-DP1</li>
<li>5.0-DP2</li>
<li>5-CURRENT</li>
<li>4-STABLE</li>
</ul>
<p>Targets:</p>
<ul>
<li>alpha, i386</li>
<li>UP, 2/4-way SMP, SMP with one processor</li>
<li>low/high memory</li>
<li>slow/fast CPU</li>
</ul>
<div class="discussion">
<p><strong class="speaker">dillon</strong> : Debug stuff on 5.0. I think
it might be reasonable then to take the space hit and always have the
debugging in but turn it on and off with sysctl.</p>
<p><strong class="speaker">rwatson</strong> : We should commit an optimized
kernel configuration and benchmarking guidelines to the tree as
well.</p>
</div>
&break;
<div class="discussion">
<p><strong class="speaker">rwatson</strong> : I think we should continue
the performance discussion. We want to accomplish a couple of things.
One is stability measurement. What are the things we need to be
measuring? What is our definition of useful?</p>
<p><strong class="speaker">hsu</strong> : End to end measurement
with gigabit cards. For latency test connections per second. Can use
ttcp or netbench in ports.</p>
<p><strong class="speaker">gnn</strong> : need to make sure we run
against all of 4.6</p>
<p><strong class="speaker">julian</strong> : Need to really have 3 tests.
4.6 (forever) 4.x (following updates) and -CURRENT.</p>
<p><strong class="speaker">rwatson</strong> : There are other dimensions.
Degree of parallelism for instance. We might see degradation in uni
but get good stuff in multi case.</p>
<p><strong class="speaker">julian</strong> : Test for impact of KSE
complications as well.</p>
<p><strong class="speaker">alfred</strong> : I think as the results come
through people should not be too worried about it. Perhaps we should
benchmark database type stuff as well. Need to do something
comprehensive.</p>
<p><strong class="speaker">obrien</strong> : What does the test matrix
look like? Different architectures and different numbers of
processors.</p>
<p><strong class="speaker">rwatson</strong> : Can we make a multi-processor
run uni-processor.</p>
<p><strong class="speaker">alfred</strong> : Run queue and scheduler stuff?</p>
<p><strong class="speaker">julian</strong> : Will talk to Alfred.</p>
<p><strong class="speaker">rwatson</strong> : Is scalability testing
important?</p>
<p><strong class="speaker">obrienM</strong> : NFS testing.</p>
<p><strong class="speaker">rwatson</strong> : What about UI testing?</p>
<p><strong class="speaker">hsu</strong> : x11perf is the way to do that.</p>
<p><strong class="speaker">dillon</strong> : Currently we have a directory
for regression tests, should we do one for performance tests?</p>
<p><strong class="speaker">gnn</strong> : talk to sleepycat for DB
tests, see if they have some</p>
<p><strong class="speaker">alfred</strong> : Really nice to tests DB
applications that are heavily thread dependent.</p>
<p><strong class="speaker">hsu</strong> :Apache 2 has threads.</p>
<p><strong class="speaker">rwatson</strong> : What about commercial folks?
What do you do.</p>
<p><strong class="speaker">ps</strong> : Normally what we end
up doing is using the snapshot on some machines and see if the bugs
are out. There is no performance testing really.</p>
<p><strong class="speaker">rwatson</strong> : Again, what about performance?</p>
<p><strong class="speaker">ps</strong> : We've really never had
one. It's more just bugs. We've just never found the performance to
be a problem.</p>
<p><strong class="speaker">rwatson</strong> : We need to create a forum for
talking about performance. We need reproducible test cases.</p>
<p><strong class="speaker">ps</strong> : There's also other
things. We've been doing lots of looking at this. FreeBSD gets
kicked down by attacks for instance. We have a lot of tools to get to
the project though.</p>
<p><strong class="speaker">rwatson</strong> : I will set up the mailing
list.</p>
</div>
</div>
<div class="topic">
<hr/>
<a name="hw"></a>
<h2>New Hardware Architectures</h2>
<h3>Alpha</h3>
<div class="discussion">
<p><strong class="speaker">jhb</strong> : Questions about alpha?</p>
<p><strong class="speaker">rwatson</strong> : KSE on alpha?</p>
<p><strong class="speaker">julian</strong> : We have patches so it
compiles and runs non-KSE programs. You can have the patched version
of the alpha kernel up and running though.</p>
<p><strong class="speaker">rwatson</strong> : Is the task owned of making
this work on Alpha?</p>
</div>
<h3>IA64</h3>
<div class="discussion">
<p><strong class="speaker">dfr</strong> : It works as far as I get to
use it. It's not used in production right now.</p>
<p><strong class="speaker">ps</strong> : Intel shipped me a quad
processor IA64 board. (McKinley is the name of the board).</p>
<p><strong class="speaker">rwatson</strong> : What does it need for 5.0?</p>
<p><strong class="speaker">dfr</strong> : It works, it works for SMP.
Self hosts, build worlds. sysinstall compiles but needs more kicking
to work.</p>
<p><strong class="speaker">ps</strong> : Intel wants us to ship
a CD.</p>
<p><strong class="speaker">dfr</strong> : There is no thread support
right now (threading library needs to move to get/setcontext rather
than longjmp).</p>
<p><strong class="speaker">dfr</strong> : Need to move every driver to
use BUS DMA for large memory machines to get bounce buffers.</p>
<p><strong class="speaker">jhb</strong> : PHK is working on using a new
libwhisk so that sysinstall et al work on all systems.</p>
</div>
<h3>Sparc64</h3>
<div class="discussion">
<p><strong class="speaker">jake</strong> : Take control of KSE stuff
on Sparc 64.</p>
<p><strong class="speaker">rwatson</strong> : Do we have a Sparc 64 in the
cluster?</p>
<p><strong class="speaker">jake</strong> : It's not in the cluster
yet. It's a serial cluster issue.</p>
<p><strong class="speaker">rwatson</strong> : Package building on S64?</p>
<p><strong class="speaker">jake</strong> : Perhaps a bunch of Ultra
60s for a package build.</p>
<p><strong class="speaker">obrien</strong> : 1500 build right now?</p>
<p><strong class="speaker">jake</strong> : Yes, but a lot of the
same bug in packages are broken.</p>
<p><strong class="speaker">jhb</strong> : Timeline for X?</p>
<p><strong class="speaker">jake</strong> : Not really.</p>
<p><strong class="speaker">rwatson</strong> : In terms of 5.0 how
comfortable are you?</p>
<p><strong class="speaker">jake</strong> : sysinstall is the only problem.</p>
</div>
<h3>PowerPC</h3>
<div class="discussion">
<p><strong class="speaker">benno</strong> : I got it up to
execing a fake init in the simulator and printing "hello world".
Trying to work with real hardware. I now have some semblance of
busdma and am working on other stuff. GEM on iMac is in an embryonic
state. Should get to NFS mount in a few weeks.</p>
<p><strong class="speaker">rwatson</strong> : How do you feel about your
timeline?</p>
<p><strong class="speaker">benno</strong> : I'm not sure we'll have
something fully workable for 5.0.</p>
<p><strong class="speaker">rwatson</strong> : You're not at the point yet
on working on KSE are you?</p>
<p><strong class="speaker">benno</strong> : No, need a useful system
first.</p>
</div>
<h3>AMD64</h3>
<div class="discussion">
<p><strong class="speaker">rwatson</strong> : I know that we're having
simulator problems.</p>
<p><strong class="speaker">obrien</strong> : The issues are about legal
and NDA. AMD decided on <a href="http://www.freebsdmall.com">FreeBSD
Mall</a> as the NDA person. I have not had a working simulator since
September.</p>
<p><strong class="speaker">ps</strong> : I can make that happen, as
well as real hardware.</p>
<p><strong class="speaker">obrien</strong> :I've got a cross tool chain in
the tree. I have some untested pmap stuff. Hardware has been
available for a month or so. We could boot FreeBSD 4.6 today if only
we had hardware.</p>
<p><strong class="speaker">rwatson</strong> : What do you think about 5.0?
Should we discuss that at another date?</p>
</div>
<h3>MIPs</h3>
<div class="discussion">
<p><strong class="speaker">unknown</strong> :Juniper offered.</p>
<p><strong class="speaker">obrien</strong> : But we have no hardware.</p>
<p><strong class="speaker">unknown</strong> :Juniper thinks it's OK but
doesn't want to have it rot in the tree.</p>
<p><strong class="speaker">brooks</strong> : I have a line on a company
that does compact PCI with R6Ks.</p>
<p><strong class="speaker">rwatson</strong> : We're waiting for someone to
turn up.</p>
</div>
</div>
<hr/>
LUNCH
<hr/>
<div class="topic">
<hr/>
<a name="trust"></a>
<h2>Trusted BSD</h2>
<p><strong class="speaker">rwatson</strong> : Malc framework is what is of
interest today.</p>
<em>See Slides from Robert</em>
<div class="discussion">
<p><strong class="speaker">julian</strong> : Are the labels the same on
all structures?</p>
<p><strong class="speaker">rwatson</strong> : You can modify this but there
are issues with memory: is the space needed for a label too large to
add to an mbuf header, for example? The label is small, but there
area lot of them?</p>
<p><strong class="speaker">bmilekic</strong> : When you're freeing the mbuf
do you write the label data?</p>
<p><strong class="speaker">rwatson</strong> : We blank it when we free it.</p>
<p><strong class="speaker">bmilekic</strong> : I do not think the 36 bytes
in the mbuf header is a problem.</p>
<p><strong class="speaker">julian</strong> : I'm more interested in the
"why" than the how.</p>
<p><strong class="speaker">rwatson</strong> : A lot of people are
interested in this. Some of the things that do interest a lot of
people are things like doing on the fly security for a web server.</p>
<p><strong class="speaker">julian</strong> : Is there a black hatted TLA
interested?</p>
<p><strong class="speaker">rwatson</strong> : Yes and several gov'ts. As
well as plenty of financial folks.</p>
<p><strong class="speaker">rwatson</strong> : There's a lot of userland
stuff that's not done yet.</p>
</div>
</div>
<div class="topic">
<hr/>
<a name="releng"></a>
<h2>Release Engineering</h2>
<p><strong class="speaker">murray</strong> : Shows a slide of releases.
4.6 is ready to go but having issues with ISO images. DP1, a lot of
goals were met. 1000 packages were building on -CURRENT to get DP1
out. Polished 4.2. We need to start making decisions on 5.0.
November is still the date we're shooting for. We're going to do a
4.7 and a 4.8. DP3?</p>
<p>***GET SLIDE FROM MURRAY***</p>
<div class="discussion">
<p><strong class="speaker">murray</strong> : Release engineering area of
the web site www.FreeBSD.org/releng. For DP2 question about p4 or
CVS? Will probably use p4 for DP2 as well. USB subsystem? Perl
removal? KSE?</p>
<p><strong class="speaker">julian</strong> : KSE should be able to run
simple tests.</p>
<p><strong class="speaker">obrien</strong> : Is whatever you have
committed by DP2 be the same as the release.</p>
<p><strong class="speaker">julian</strong> : It will be a subset.</p>
<p><strong class="speaker">murray</strong> : What will the status be of
KSE in userland for 5.0?</p>
<p><strong class="speaker">julian</strong> : Can't answer that right
now. We're not removing the old libraries. The userland work will
happen between DP2 and release. The next step is MP as well as
UP.</p>
<p><strong class="speaker">obrien</strong> : Are we heading for a release?</p>
<p><strong class="speaker">murray</strong> : yes.</p>
<p><strong class="speaker">obrien</strong> : Then we have to stop having
major commits.</p>
<p><strong class="speaker">murray</strong> : Yes, the discussion today is
what are the major must have features.</p>
<p><strong class="speaker">rwatson</strong> : We need to decide if there
are major upcoming problems and reduce risk on things like KSE.</p>
<p><strong class="speaker">julian</strong> : That's why I want to get milestone 3
in now.</p>
<p><strong class="speaker">rwatson</strong> : Do you think that KSE related
changes from later milestones are going to be isolated to KSE or
pervasive?</p>
<p><strong class="speaker">julian</strong> : Hard to say. My guess is
that milestone 4 stuff should be less pervasive.</p>
<p><strong class="speaker">rwatson</strong> : What happens if KSE just
doesn't work?</p>
<p><strong class="speaker">julian</strong> : Well it does work, the
patches work, it's a question of risk. We need to check on new
things, like locking two threads in the same process.</p>
<p><strong class="speaker">dillon</strong> : KSEs only become fragile when
pthread uses them. That's the turning point.</p>
<p><strong class="speaker">obrien</strong> : I'd like the rules for the
rest of the summer, I hope we'll talk about that.</p>
<p><strong class="speaker">murray</strong> : Earlier is better.</p>
<p><strong class="speaker">mini</strong> : I think the cutoff point for
KSE might be milestone 3.</p>
<p><strong class="speaker">rwatson</strong> : It's the kind of thing where
if we need to back out we can.</p>
<p><strong class="speaker">julian</strong> : If you're not going to run
KSEs then you're OK.</p>
<p><strong class="speaker">rwatson</strong> : I think it's low risk. Let's
avoid the risk is the message.</p>
<p><strong class="speaker">julian</strong> : The next DP2 (where we'd like
milestone 4).</p>
<p><strong class="speaker">alfred</strong> : We really need KSE so all
this concern about stuff that no one really uses is not a big deal.
People just need to play catch up. We have performance problems and
we need to solve those.</p>
<p><strong class="speaker">obrien</strong> : We quickly need to figure out
our policy on multiple archs.</p>
<p><strong class="speaker">rwatson</strong> : I briefly want to respond to
Alfred. We have asserted that KSE will be experimental. It will be
in and 5.0 will go out but there might be issues.</p>
<p><strong class="speaker">jhb</strong> : Realistically for the network
stack is that IPv4 sockets will not be giant. But this is only in the
network stack world. Several people are working on it.</p>
<p><strong class="speaker">rwatson</strong> : The GEOM stuff will be
enabled by default in 5.0. Sparc depends on it. I do not know what
the impediments are to that though.</p>
<p><strong class="speaker">julian</strong> : The kernel stuff is there but
the user space is not. It can't become the default until everything
is there.</p>
<p><strong class="speaker">warner</strong> : What level of control are you
going to exercise over the tree in the coming months?</p>
<p><strong class="speaker">murray</strong> : You're going to see more
level of control but we expect the requests to be reasonable. It's a
very open process.</p>
<p><strong class="speaker">jhb</strong> : How are we going to address the
5/6 split?</p>
<p><strong class="speaker">murray</strong> : Carefully is all I can
say.</p>
<p><strong class="speaker">rwatson</strong> : For 5. 0 we need to have a
more informed decision. The release engineers will be trying to
reduce the number of large code changes more as time goes by. We
don't have to wait for 5.x to be perfectly stable before we branch.</p>
<p><strong class="speaker">murray</strong> : Let's move it to more general
discussion of DP2? Specific technologies.</p>
<p><strong class="speaker">bmilekic</strong> : Is there a strategy to lock
other protocols that are not locked down onw?</p>
<p><strong class="speaker">obrien</strong> : How much more do we need to
do before 5.0?</p>
<p><strong class="speaker">jhb</strong> : Bug fixing is what we're doing.</p>
<p><strong class="speaker">rwatson</strong> : The answer on the network
stack. We need to choose a strategy on how to handle the other
protocols.</p>
<p><strong class="speaker">obrien</strong> : The crux is that socket
locking must be in 5.0.</p>
<p><strong class="speaker">rwatson</strong> : There are 2 or 3 problems.
Routing code is a problem. See earlier discussions.</p>
<p><strong class="speaker">dfr</strong> : RCng is essentially done.
What it needs is testers.</p>
<p><strong class="speaker">alfred</strong> : What about libh (I think libh
is wrong but this is what I heard)?</p>
<p><strong class="speaker">jhb</strong> : It's very far along but not a
5.0 thing.</p>
<p><strong class="speaker">warner</strong> : Problems with interrupt
routing in ACPI?</p>
<p><strong class="speaker">takawata</strong> : Cannot handle PCI<->PCI
interrupt routing. Many 802.11x have this problem.</p>
<p><strong class="speaker">julian</strong> : Is it a problem from Intel?</p>
<p><strong class="speaker">takawata</strong> : This is not an Intel
problem but a problem on our side. PCI<->PCI routing code should be
added. New code is necessary.</p>
<pre>
Whiteboard
UFS2 rcNG KSE M3 CAM SMPng
GEOM TrustedBSD Malc BusDMA Newbus SMPng
C++ Cardbus libwhisk/sysinstall KOBJ? (no!)
sparc64
Perl Removal ACPI Alpha SMP Stability Pkgs for
sparc64, IA64
devd PCI intr route document hints release docs
for new
platform
</pre>
<p><strong class="speaker">unknown</strong> : Firewire?</p>
<p><strong class="speaker">rwatson</strong> : What hardware shipping on
IA64?</p>
<p><strong class="speaker">dfr</strong> : Intel stuff</p>
<p><strong class="speaker">rwatson</strong> : What about on Sparc64?</p>
<p><strong class="speaker">obrien</strong> : Very limited (hme...)</p>
<p><strong class="speaker">rwatson</strong> : KOBJ extensions discussed at
BSDCon?</p>
<p><strong class="speaker">warner</strong> : Not sure, probably not for
5.0. Pervasive, so no.</p>
<p><strong class="speaker">rwatson</strong> : How broken is C++?</p>
<p><strong class="speaker">obrien</strong> : Only on sparc64. Don't
really know yet, but it's probably a library issue. The compiler is a
pre-release snapshot. The diffs are now getting large from May 5 to
now. We should attempt to be as far along this gcc branch as possible
come release.</p>
</div>
</div>
&break;
<div class="topic">
<hr/>
<a name="rc"></a>
<h2>rc.d</h2>
</div>
<div class="discussion">
<p><strong class="speaker">gordon</strong> : Talking about rc.d stuff.
Import from NetBSD. Right now we have patches out there that are
translated from the current boot order. It's in perforce. After the
conference it will go into the mainline. Single toggle for
booting.</p>
<p><strong class="speaker">rwatson</strong> : How in sync are the bits in
the new stuff with the old stuff.</p>
<p><strong class="speaker">gordon</strong> : Last patch is from June 3rd,
but it's tracking closely.</p>
<p><strong class="speaker">rwatson</strong> : What is the schedule for
committing to the main tree.</p>
<p><strong class="speaker">gordon</strong> : We have large patches so
we're going to re-import from NetBSD.</p>
<p><strong class="speaker">rwatson</strong> : How about you have it done by
July 1?</p>
<p><strong class="speaker">gordon</strong> : We could probably do that.
Definitely want to be in DP2.</p>
<p><strong class="speaker">gshapiro</strong> : How long will we keep the old
stuff for?</p>
<p><strong class="speaker">gordon</strong> : We'll keep them both in for a
while. Not more than 1.5 months though.</p>
<p><strong class="speaker">julian</strong> : Have you had a look at all at
the Mac OS/X startup code?</p>
<p><strong class="speaker">gordon</strong> : No.</p>
<p><strong class="speaker">julian</strong> : Do you deal with dependencies?</p>
<p><strong class="speaker">gordon</strong> : There is meta data in each
script that says what needs what. There is a program that orders
everything correctly.</p>
<p><strong class="speaker">unknown</strong> : How does this effect the rc
script for ports install?</p>
<p><strong class="speaker">gordon</strong> : We could make this available
to ports but won't on the first version.</p>
<p><strong class="speaker">alfred</strong> : Can I recommend that you
recommend this to ports?</p>
<p><strong class="speaker">gordon</strong> : Yes, the problem is that we
have so many ports.</p>
<p><strong class="speaker">rwatson</strong> : The reason for this is for
rebundlers of FreeBSD in their environments. We don't have to have it
for DP2 but it should be an ultimate goal. We might need to have a
policy statement on this. That at date X all ports must use the new
system.</p>
</div>
<div class="topic">
<hr/>
<a name="open"></a>
<h2>Other Issues</h2>
</div>
<div class="discussion">
<p><strong class="speaker">sam</strong> : I've been working on hardware
crypto. I'm looking for consensus on getting hardware crypto in the
kernel. This will not happen in 5.0.</p>
<h3>Syscall vector change for 64bits</h3>
<p><strong class="speaker">dillon</strong> : Two ways to go. Need to
create a new syscall vector. The other is to do a 1 off replacement.
Prefer the former.</p>
<p><strong class="speaker">rwatson</strong> : Perhaps we need to create a
FreeBSD 5 syscall vector. Could be a new ABI.</p>
<p><strong class="speaker">julian</strong> : Aren't there enough other
numbers?</p>
<p><strong class="speaker">rwatson</strong> : That's one way to look at it
and other platforms have done that? Is that too heavy weight?</p>
<p><strong class="speaker">julian</strong> : It sounds that way to me.
You end up having to replicate the old ones into the new one.</p>
<p><strong class="speaker">dillon</strong> : The issue is about pollution.</p>
<p><strong class="speaker">dfr</strong> : Seems like too much work for 5.x</p>
<p><strong class="speaker">julian</strong> : It's more work. There are
now two places. Why not talk to OpenBSD?</p>
<p><strong class="speaker">rwatson</strong> : Should there be a BSD alfredI?
Tough to do across projects.</p>
<p><strong class="speaker">obrien</strong> : Who here is going to see that
through? We have not talked to NetBSD about even SMP.</p>
<p><strong class="speaker">alfred</strong> : Does changing the syscall
table allow us to do clean up?</p>
<p><strong class="speaker">rwatson</strong> : We could do that without
doing 64bit syscall table.</p>
<h3>5.x ABI stability</h3>
<p><strong class="speaker">rwatson</strong> : There are new functions in
5.x. At what point do we stop changing?</p>
<p><strong class="speaker">dfr</strong> : When people start really using it.</p>
<p><strong class="speaker">rwatson</strong> : How do we tell? How did
Solaris do it?</p>
<p><strong class="speaker">Everyone</strong> : No one knows.</p>
<p><strong class="speaker">dfr</strong> : It's too hard to add a
syscall vector. Library issues are a problem.</p>
<p><strong class="speaker">obrien</strong> : We can use ELF to handle that.</p>
<p><strong class="speaker">dfr</strong> : Let's just add 20 new
syscalls instead of adding new work that we don't really really need.</p>
<p><strong class="speaker">rwatson</strong> : Punt on lack of time to do
this.</p>
<p><strong class="speaker">dillon</strong> : I see obrien's point with the
libraries but I have done this with time_t at 64 bits.</p>
<h3>devd</h3>
<p><strong class="speaker">rwatson</strong> : The devd stuff was to
integrate cardbus, newbus, etc.</p>
<p><strong class="speaker">julian</strong> : To monitor requests to mount
or create new devices.</p>
<p><strong class="speaker">rwatson</strong> : Is this a 5.0 requirement?
Is there anyone to do this?</p>
<!-- Which Gordon was this ? -->
<p><strong class="speaker">gordon (from IRC)</strong> : PHK has patches
that make having devd unnecessary.</p>
<p><strong class="speaker">brooks</strong> : Need something that does what
pccardd did.</p>
<p><strong class="speaker">julian</strong> : Need to be able to do this
through a file.</p>
<p><strong class="speaker">warner</strong> : (from IRC): That's a 6.0
feature.</p>
<p><strong class="speaker">julian</strong> : It would not be a large step
to put something in the middle to handle this.</p>
<p><strong class="speaker">julian</strong> : Sometime in the 5 lifetime we
need this.</p>
<p><strong class="speaker">warner</strong> : There is no way to monitor
events in newbus but it would be easy to add.</p>
<p><strong class="speaker">julian</strong> : I'm not sure I understood you
correctly.</p>
<p><strong class="speaker">warner</strong> : What happens now in a PCI is
that it makes a call to pci_get_devid() and the driver would say "yes
I am " or "no I'm not" so you'd have to change each of the busses to
do this but that's not too tough because we have a small # of
busses.</p>
<p><strong class="speaker">jhb</strong> : Mike Smith gave us an
informal tour of OS/X. OS/X uses XML to do this. They have the DEVID
in XML.</p>
<p><strong class="speaker">brooks</strong> : I looked at some PCI drivers
and some work that way but some don't.</p>
<p><strong class="speaker">julian</strong> : It seems to me we need to not
have to modify every single driver. If you've got a device that's not
supported you ask all drivers. At the point when you run out you make
an outcall. The outcall returns does a substitution.</p>
<p><strong class="speaker">rwatson</strong> : Time up, time to wrap up.</p>
</div>
</body>
</html>
|
