Lines 43-86
Stokely</a>.</p>
Link Here
|
43 |
<li><a href="#open">Open Discussion</a></li> |
43 |
<li><a href="#open">Open Discussion</a></li> |
44 |
</ul> |
44 |
</ul> |
45 |
|
45 |
|
46 |
<p>NOTE: As usual I missed some names, please add those I missed.</p> |
46 |
<p>NOTE: As usual I missed some names, please add those I missed. During |
|
|
47 |
the discussion names have been abbreviated, for committers their |
48 |
FreeBSD.org username has been used, for non committers the initials |
49 |
are used.</p> |
47 |
|
50 |
|
48 |
<h2>Attending:</h2> |
51 |
<h2>Attending:</h2> |
49 |
|
52 |
|
50 |
<p>In person:</p> |
53 |
<p>Committers In Person:</p> |
|
|
54 |
<ul> |
55 |
<li>Robert Watson (rwatson)</li> |
56 |
<li>Julian Elischer (julian)</li> |
57 |
<li>John Baldwin (jhb)</li> |
58 |
<li>Matt Dillon (dillon)</li> |
59 |
<li>Warner Losh (warner)</li> |
60 |
<li>David O'Brien (obrien)</li> |
61 |
<li>Jeffery Hsu (hsu)</li> |
62 |
<li>Jennifer Yang (jennifer)</li> |
63 |
<li>Bosko Milekic (bmilekic)</li> |
64 |
<li>Alfred Perlstein (alfred)</li> |
65 |
<li>Doug Rabson (dfr)</li> |
66 |
<li>Paul Saab (ps)</li> |
67 |
<li>Brooks Davis (brooks)</li> |
68 |
<li>Murray Stokely (murray)</li> |
69 |
<li>Jonathan Mini (mini)</li> |
70 |
<li>Takanori Watanabe (takawata)</li> |
71 |
<li>Gordon Tetlow (gordon)</li> |
72 |
<li>Gregory Shapiro (gshapiro)</li> |
73 |
<li>Sam Leffler (sam)</li> |
74 |
<li>Bruce Mah (bmah)</li> |
75 |
</ul> |
76 |
|
77 |
<p>Also In Person:</p> |
51 |
<ul> |
78 |
<ul> |
52 |
<li>Robert Watson (RW)</li> |
|
|
53 |
<li>Julian Elischer(JE)</li> |
54 |
<li>John Baldwin(JB)</li> |
55 |
<li>Matt Dillon (MD)</li> |
56 |
<li>Warner Losh (WL)</li> |
57 |
<li>David O'Brian (DO)</li> |
58 |
<li>Jeffery Xu (JX)</li> |
59 |
<li>Jennifer Ying (JY)</li> |
60 |
<li>Bosko Milekic (BM)</li> |
61 |
<li>Alfred Perlstein (AP)</li> |
62 |
<li>Doug Rabson (DR)</li> |
63 |
<li>Paul Saab (PS)</li> |
64 |
<li>Brooks Davis (BD)</li> |
65 |
<li>Murray Stokely (MS)</li> |
66 |
<li>Jonathan Mini (JM)</li> |
67 |
<li>Watanabe ???</li> |
68 |
<li>Gordon Tetlow (GT)</li> |
69 |
<li>Gregory Schapiro (GS)</li> |
70 |
<li>Sam Leffler (SL)</li> |
71 |
<li>Bruce Mah</li> |
72 |
<li>George Neville-Neil (gnn)</li> |
79 |
<li>George Neville-Neil (gnn)</li> |
73 |
<li>Unknown (??)</li> |
80 |
<!--<li>Unknown (??)</li>--> |
74 |
</ul> |
81 |
</ul> |
75 |
|
82 |
|
76 |
<p>On The Phone:</p> |
83 |
<p>On The Phone:</p> |
77 |
<ul> |
84 |
<ul> |
78 |
<li>Alan Cox (AC)</li> |
85 |
<li>Alan Cox (alc)</li> |
79 |
</ul> |
86 |
</ul> |
80 |
|
87 |
|
81 |
<p>Via webcast:</p> |
88 |
<p>Via webcast:</p> |
82 |
|
89 |
|
83 |
<p>??</p> |
90 |
<p>Many people listened in via the stream and chatted on IRC to one |
|
|
91 |
another while the discussion took place.</p> |
84 |
|
92 |
|
85 |
<p>The meeting followed a format where each section was led by an |
93 |
<p>The meeting followed a format where each section was led by an |
86 |
individual and then a discussion ensued. Not all of the discussion |
94 |
individual and then a discussion ensued. Not all of the discussion |
Lines 125-199
perforce and people have to patch it.</p
Link Here
|
125 |
|
133 |
|
126 |
<div class="discussion"> |
134 |
<div class="discussion"> |
127 |
|
135 |
|
128 |
<p><strong class="speaker">RW</strong> : What about userland?</p> |
136 |
<p><strong class="speaker">rwatson</strong> : What about userland?</p> |
129 |
|
137 |
|
130 |
<p><strong class="speaker">JE</strong> : It can run different threads |
138 |
<p><strong class="speaker">julian</strong> : It can run different threads |
131 |
in userland. The primitives are all there it just needs a bit more |
139 |
in userland. The primitives are all there it just needs a bit more |
132 |
help. I would like to put an idea out. Is it a good idea to be able |
140 |
help. I would like to put an idea out. Is it a good idea to be able |
133 |
to have non-threaded programs linking with threaded libraries?</p> |
141 |
to have non-threaded programs linking with threaded libraries?</p> |
134 |
|
142 |
|
135 |
<p><strong class="speaker">RW</strong> : Putting async I/O into such a |
143 |
<p><strong class="speaker">rwatson</strong> : Putting async I/O into such a |
136 |
thing would make sense.</p> |
144 |
thing would make sense.</p> |
137 |
|
145 |
|
138 |
<p><strong class="speaker">JE</strong> : The library would not care |
146 |
<p><strong class="speaker">julian</strong> : The library would not care |
139 |
who was accessing it.</p> |
147 |
who was accessing it.</p> |
140 |
|
148 |
|
141 |
<p><strong class="speaker">RW</strong> : For instance libc could be |
149 |
<p><strong class="speaker">rwatson</strong> : For instance libc could be |
142 |
threaded or not.</p> |
150 |
threaded or not.</p> |
143 |
|
151 |
|
144 |
<p><strong class="speaker">JE</strong> : That would be interesting. I |
152 |
<p><strong class="speaker">julian</strong> : That would be interesting. I |
145 |
don't know if the two interfaces are incompatible.</p> |
153 |
don't know if the two interfaces are incompatible.</p> |
146 |
|
154 |
|
147 |
<p><strong class="speaker">JB</strong> : X does this.</p> |
155 |
<p><strong class="speaker">jhb</strong> : X does this.</p> |
148 |
|
156 |
|
149 |
<p><strong class="speaker">MD</strong> : It is very doable but you |
157 |
<p><strong class="speaker">dillon</strong> : It is very doable but you |
150 |
have to make it non-preemptive. If you're switching non-preemptively |
158 |
have to make it non-preemptive. If you're switching non-preemptively |
151 |
you can use library routines which are non threaded.</p> |
159 |
you can use library routines which are non threaded.</p> |
152 |
|
160 |
|
153 |
<p><strong class="speaker">JE</strong> : If I do what I'm thinking of |
161 |
<p><strong class="speaker">julian</strong> : If I do what I'm thinking of |
154 |
doing then each lib will have its own KSE group.</p> |
162 |
doing then each lib will have its own KSE group.</p> |
155 |
|
163 |
|
156 |
<p><strong class="speaker">MD</strong> : stdio does not have to be |
164 |
<p><strong class="speaker">dillon</strong> : stdio does not have to be |
157 |
thread aware if you don't schedule preemptively. It all matters where |
165 |
thread aware if you don't schedule preemptively. It all matters where |
158 |
it blocks.</p> |
166 |
it blocks.</p> |
159 |
|
167 |
|
160 |
<p><strong class="speaker">JE</strong> : Since you're a non-threaded |
168 |
<p><strong class="speaker">julian</strong> : Since you're a non-threaded |
161 |
program you don't know that.</p> |
169 |
program you don't know that.</p> |
162 |
|
170 |
|
163 |
<p><strong class="speaker">RW</strong> : If you're going to support |
171 |
<p><strong class="speaker">rwatson</strong> : If you're going to support |
164 |
that, libc has to support threads.</p> |
172 |
that, libc has to support threads.</p> |
165 |
|
173 |
|
166 |
<p><strong class="speaker">RW</strong> : It sounds like some |
174 |
<p><strong class="speaker">rwatson</strong> : It sounds like some |
167 |
complexity goes away. Can we use 1 libc with has threading?</p> |
175 |
complexity goes away. Can we use 1 libc with has threading?</p> |
168 |
|
176 |
|
169 |
<p><strong class="speaker">JE</strong> : Do we want to go down this |
177 |
<p><strong class="speaker">julian</strong> : Do we want to go down this |
170 |
path?</p> |
178 |
path?</p> |
171 |
|
179 |
|
172 |
<p><strong class="speaker">RW</strong> : Now or later?</p> |
180 |
<p><strong class="speaker">rwatson</strong> : Now or later?</p> |
173 |
|
181 |
|
174 |
<p><strong class="speaker">JE</strong> : What do I design now to do |
182 |
<p><strong class="speaker">julian</strong> : What do I design now to do |
175 |
this?</p> |
183 |
this?</p> |
176 |
|
184 |
|
177 |
<p><strong class="speaker">JB</strong> : For example libc_r does not |
185 |
<p><strong class="speaker">jhb</strong> : For example libc_r does not |
178 |
work with rfork.</p> |
186 |
work with rfork.</p> |
179 |
|
187 |
|
180 |
<p><strong class="speaker">JE</strong> : The answer is that yes we |
188 |
<p><strong class="speaker">julian</strong> : The answer is that yes we |
181 |
should move forward. Tricky issues, signals...</p> |
189 |
should move forward. Tricky issues, signals...</p> |
182 |
|
190 |
|
183 |
<p><strong class="speaker">WL</strong> : Have people talked about |
191 |
<p><strong class="speaker">warner</strong> : Have people talked about |
184 |
pthread programs and cancellation points?</p> |
192 |
pthread programs and cancellation points?</p> |
185 |
|
193 |
|
186 |
<p><strong class="speaker">JE</strong> : The pthreads library does not |
194 |
<p><strong class="speaker">julian</strong> : The pthreads library does not |
187 |
assume that you're only going to change threads at yield() points. We |
195 |
assume that you're only going to change threads at yield() points. We |
188 |
are going to have cancellation points. There is an unimplemented call |
196 |
are going to have cancellation points. There is an unimplemented call |
189 |
which will be able to send a thread targeted signal even into the |
197 |
which will be able to send a thread targeted signal even into the |
190 |
kernel.</p> |
198 |
kernel.</p> |
191 |
|
199 |
|
192 |
<p><strong class="speaker">JE</strong> : When a thread is scheduled |
200 |
<p><strong class="speaker">julian</strong> : When a thread is scheduled |
193 |
onto a KSE there is a mailbox that the userland thread scheduler |
201 |
onto a KSE there is a mailbox that the userland thread scheduler |
194 |
updates.</p> |
202 |
updates.</p> |
195 |
|
203 |
|
196 |
<p><strong class="speaker">JE</strong> : Is there anyone else who has |
204 |
<p><strong class="speaker">julian</strong> : Is there anyone else who has |
197 |
some time or test it? How many people should test this before I check |
205 |
some time or test it? How many people should test this before I check |
198 |
it in? There is a patch that's continuously updated on my web site to |
206 |
it in? There is a patch that's continuously updated on my web site to |
199 |
be able to patch it to -CURRENT. There is a CVSUP target from cvsup |
207 |
be able to patch it to -CURRENT. There is a CVSUP target from cvsup |
Lines 201-225
be able to patch it to -CURRENT. There
Link Here
|
201 |
freefal there is a pointer there to a web page that explains how to |
209 |
freefal there is a pointer there to a web page that explains how to |
202 |
CVSUP from source.</p> |
210 |
CVSUP from source.</p> |
203 |
|
211 |
|
204 |
<p><strong class="speaker">RW</strong> : What about SMP locking for |
212 |
<p><strong class="speaker">rwatson</strong> : What about SMP locking for |
205 |
this?</p> |
213 |
this?</p> |
206 |
|
214 |
|
207 |
<p><strong class="speaker">JE</strong> : Handled by the proc locking. |
215 |
<p><strong class="speaker">julian</strong> : Handled by the proc locking. |
208 |
Has not been tried on SMP machines yet.</p> |
216 |
Has not been tried on SMP machines yet.</p> |
209 |
|
217 |
|
210 |
<p><strong class="speaker">DO</strong> : What about on Sparc?</p> |
218 |
<p><strong class="speaker">obrien</strong> : What about on Sparc?</p> |
211 |
|
219 |
|
212 |
<p><strong class="speaker">JE</strong> : You may need to stub things |
220 |
<p><strong class="speaker">julian</strong> : You may need to stub things |
213 |
out.</p> |
221 |
out.</p> |
214 |
|
222 |
|
215 |
<p><strong class="speaker">JB</strong> : Is the paper on the web site?</p> |
223 |
<p><strong class="speaker">jhb</strong> : Is the paper on the web site?</p> |
216 |
|
224 |
|
217 |
<p><strong class="speaker">JE</strong> : The updated copy has disappeared.</p> |
225 |
<p><strong class="speaker">julian</strong> : The updated copy has disappeared.</p> |
218 |
|
226 |
|
219 |
<p><strong class="speaker">??</strong> : What's the different between |
227 |
<p><strong class="speaker">unknown</strong> : What's the different between |
220 |
NetBSD and FreeBSD on this?</p> |
228 |
NetBSD and FreeBSD on this?</p> |
221 |
|
229 |
|
222 |
<p><strong class="speaker">JE</strong> : Logically not a tremendous |
230 |
<p><strong class="speaker">julian</strong> : Logically not a tremendous |
223 |
difference but Net follows the paper closely and Free takes the idea |
231 |
difference but Net follows the paper closely and Free takes the idea |
224 |
and makes it into a production system. There were some tough battles |
232 |
and makes it into a production system. There were some tough battles |
225 |
on -arch about this. The tricky point is that the proc structure has |
233 |
on -arch about this. The tricky point is that the proc structure has |
Lines 231-240
end we ended up breaking up the proc str
Link Here
|
231 |
overwhelm the CPU when scheduling threads. This is the major |
239 |
overwhelm the CPU when scheduling threads. This is the major |
232 |
difference.</p> |
240 |
difference.</p> |
233 |
|
241 |
|
234 |
<p><strong class="speaker">JE</strong> : I greatly admire the NetBSD |
242 |
<p><strong class="speaker">julian</strong> : I greatly admire the NetBSD |
235 |
way which is to take an idea and not dilute it.</p> |
243 |
way which is to take an idea and not dilute it.</p> |
236 |
|
244 |
|
237 |
<p><strong class="speaker">JE</strong> : Net is also putting a Solaris |
245 |
<p><strong class="speaker">julian</strong> : Net is also putting a Solaris |
238 |
compatible threads package on top of their scheduler activations in |
246 |
compatible threads package on top of their scheduler activations in |
239 |
the Solaris ABI.</p> |
247 |
the Solaris ABI.</p> |
240 |
</div> |
248 |
</div> |
Lines 247-263
the Solaris ABI.</p>
Link Here
|
247 |
|
255 |
|
248 |
<div class="discussion"> |
256 |
<div class="discussion"> |
249 |
|
257 |
|
250 |
<p><strong class="speaker">JB</strong> : Yesterday we talked about SMP |
258 |
<p><strong class="speaker">jhb</strong> : Yesterday we talked about SMP |
251 |
related things so I'll give a summary and then give a list of things |
259 |
related things so I'll give a summary and then give a list of things |
252 |
for 5.0.</p> |
260 |
for 5.0.</p> |
253 |
|
261 |
|
254 |
<p><strong class="speaker">JB</strong> : The big thing for 5.0 is to |
262 |
<p><strong class="speaker">jhb</strong> : The big thing for 5.0 is to |
255 |
get the network stack out from under Giant.</p> |
263 |
get the network stack out from under Giant.</p> |
256 |
|
264 |
|
257 |
<p><strong class="speaker">JB</strong> : Jefferey Xu and Jennifer Ying |
265 |
<p><strong class="speaker">jhb</strong> : Jefferey Xu and Jennifer Ying |
258 |
were here to talk about this. They have the PCBs checked in now.</p> |
266 |
were here to talk about this. They have the PCBs checked in now.</p> |
259 |
|
267 |
|
260 |
<p><strong class="speaker">JY</strong> : Interface Queues and SynCache |
268 |
<p><strong class="speaker">jennifer</strong> : Interface Queues and SynCache |
261 |
might be done.</p> |
269 |
might be done.</p> |
262 |
</div> |
270 |
</div> |
263 |
|
271 |
|
Lines 275-369
might be done.</p>
Link Here
|
275 |
|
283 |
|
276 |
<div class="discussion"> |
284 |
<div class="discussion"> |
277 |
|
285 |
|
278 |
<p><strong class="speaker">JB</strong> : Aside from network the newbus |
286 |
<p><strong class="speaker">jhb</strong> : Aside from network the newbus |
279 |
locking needs to be done (Warner Losh) and also CAM stuff. No known |
287 |
locking needs to be done (Warner Losh) and also CAM stuff. No known |
280 |
status on CAM. Perhaps CAM is not needed for 5.0</p> |
288 |
status on CAM. Perhaps CAM is not needed for 5.0</p> |
281 |
|
289 |
|
282 |
<p><strong class="speaker">JB</strong> : Disk drive interrupts? Would |
290 |
<p><strong class="speaker">jhb</strong> : Disk drive interrupts? Would |
283 |
help performance. Going to talk to Poul Henning-Kamp</p> |
291 |
help performance. Going to talk to Poul Henning-Kamp</p> |
284 |
|
292 |
|
285 |
<p><strong class="speaker">JB</strong> : Alan Cox is working on the VM |
293 |
<p><strong class="speaker">jhb</strong> : Alan Cox is working on the VM |
286 |
system. Working based on the old Mach stuff. Objective for 5.0 is to |
294 |
system. Working based on the old Mach stuff. Objective for 5.0 is to |
287 |
get zero fill and execute on write to work without Giant. In future |
295 |
get zero fill and execute on write to work without Giant. In future |
288 |
he wants to look at locking down pmap() functions.</p> |
296 |
he wants to look at locking down pmap() functions.</p> |
289 |
|
297 |
|
290 |
<p><strong class="speaker">JB</strong> : Still some stability issues. |
298 |
<p><strong class="speaker">jhb</strong> : Still some stability issues. |
291 |
UMA breaks some assumptions. For instance sockets assume that once |
299 |
UMA breaks some assumptions. For instance sockets assume that once |
292 |
memory is a socket its a socket forever, this is no longer true.</p> |
300 |
memory is a socket its a socket forever, this is no longer true.</p> |
293 |
|
301 |
|
294 |
<p><strong class="speaker">JB</strong> : Talked to Mike Smith about |
302 |
<p><strong class="speaker">jhb</strong> : Talked to Mike Smith about |
295 |
5.0 and have decided to stop adding features so that we can start |
303 |
5.0 and have decided to stop adding features so that we can start |
296 |
clean up 5.0 and make it a real release. This might require hacks.</p> |
304 |
clean up 5.0 and make it a real release. This might require hacks.</p> |
297 |
|
305 |
|
298 |
<p><strong class="speaker">RW</strong> : For example in the UMA case |
306 |
<p><strong class="speaker">rwatson</strong> : For example in the UMA case |
299 |
there could be a flag to just say "don't reclaim this zone" -- this |
307 |
there could be a flag to just say "don't reclaim this zone" -- this |
300 |
would help with issues such as the socket code assuming memory is type |
308 |
would help with issues such as the socket code assuming memory is type |
301 |
stable.</p> |
309 |
stable.</p> |
302 |
|
310 |
|
303 |
<p>Over to AC on the VM system. Nothing to say.</p> |
311 |
<p>Over to alc on the VM system. Nothing to say.</p> |
304 |
|
312 |
|
305 |
<p><strong class="speaker">BM</strong> : As much as I might get hated |
313 |
<p><strong class="speaker">bmilekic</strong> : As much as I might get hated |
306 |
for this. Will preemption stuff go in by 5.0?</p> |
314 |
for this. Will preemption stuff go in by 5.0?</p> |
307 |
|
315 |
|
308 |
<p><strong class="speaker">JB</strong> :No, that's a 6.0 thing. There |
316 |
<p><strong class="speaker">jhb</strong> :No, that's a 6.0 thing. There |
309 |
are things to do first.</p> |
317 |
are things to do first.</p> |
310 |
|
318 |
|
311 |
<p><strong class="speaker">??? Phone</strong> : Could this come in in |
319 |
<p><strong class="speaker">unknown</strong> : Could this come in in |
312 |
the life time of 5.? 5.1?</p> |
320 |
the life time of 5.? 5.1?</p> |
313 |
|
321 |
|
314 |
<p><strong class="speaker">RW</strong> : This is a release issue really.</p> |
322 |
<p><strong class="speaker">rwatson</strong> : This is a release issue really.</p> |
315 |
|
323 |
|
316 |
<p><strong class="speaker">JB</strong> : Yes, the kernel is pre-emptive.</p> |
324 |
<p><strong class="speaker">jhb</strong> : Yes, the kernel is pre-emptive.</p> |
317 |
|
325 |
|
318 |
<p><strong class="speaker">RW</strong> : Perhaps we should talk about |
326 |
<p><strong class="speaker">rwatson</strong> : Perhaps we should talk about |
319 |
is performance goals? What are the comparisons to make? Perhaps head |
327 |
is performance goals? What are the comparisons to make? Perhaps head |
320 |
of 4 with head of 5. We'll see a mix.</p> |
328 |
of 4 with head of 5. We'll see a mix.</p> |
321 |
|
329 |
|
322 |
<p><strong class="speaker">JB</strong> : I need to run benchmarks.</p> |
330 |
<p><strong class="speaker">jhb</strong> : I need to run benchmarks.</p> |
323 |
|
331 |
|
324 |
<p><strong class="speaker">RW</strong> : In terms of SMP features when |
332 |
<p><strong class="speaker">rwatson</strong> : In terms of SMP features when |
325 |
will VM be ready to be measured? I can't put a date on it.</p> |
333 |
will VM be ready to be measured? I can't put a date on it.</p> |
326 |
|
334 |
|
327 |
<p><strong class="speaker">AC</strong> : I think I told John was in |
335 |
<p><strong class="speaker">alc</strong> : I think I told John was in |
328 |
time for release. I'm already doing performance testing so we've |
336 |
time for release. I'm already doing performance testing so we've |
329 |
already started.</p> |
337 |
already started.</p> |
330 |
|
338 |
|
331 |
<p><strong class="speaker">RW</strong> : We'll pick a date to start |
339 |
<p><strong class="speaker">rwatson</strong> : We'll pick a date to start |
332 |
doing measurements. Perhaps 2 or 3 weeks from now.</p> |
340 |
doing measurements. Perhaps 2 or 3 weeks from now.</p> |
333 |
|
341 |
|
334 |
<p><strong class="speaker">AC</strong> : My guess is the the locking |
342 |
<p><strong class="speaker">alc</strong> : My guess is the the locking |
335 |
pmap is going to take some time to shake out. On the other hand the |
343 |
pmap is going to take some time to shake out. On the other hand the |
336 |
next major module we should be working on is machine dependent level. |
344 |
next major module we should be working on is machine dependent level. |
337 |
Last we should try approaching the vmobject level. I'll start |
345 |
Last we should try approaching the vmobject level. I'll start |
338 |
worrying about performance in the near term.</p> |
346 |
worrying about performance in the near term.</p> |
339 |
|
347 |
|
340 |
<p><strong class="speaker">RW</strong> : Will threading improve |
348 |
<p><strong class="speaker">rwatson</strong> : Will threading improve |
341 |
latency or throughput for networking?</p> |
349 |
latency or throughput for networking?</p> |
342 |
|
350 |
|
343 |
<p><strong class="speaker">BM</strong> : I would like if we could |
351 |
<p><strong class="speaker">bmilekic</strong> : I would like if we could |
344 |
actually start before.</p> |
352 |
actually start before.</p> |
345 |
|
353 |
|
346 |
<p><strong class="speaker">RW</strong> : Do you have a timeline for |
354 |
<p><strong class="speaker">rwatson</strong> : Do you have a timeline for |
347 |
the interrupt threading stuff?</p> |
355 |
the interrupt threading stuff?</p> |
348 |
|
356 |
|
349 |
<p><strong class="speaker">BM</strong> : I finished some things last |
357 |
<p><strong class="speaker">bmilekic</strong> : I finished some things last |
350 |
night but there are still issues. In a couple of weeks it should be |
358 |
night but there are still issues. In a couple of weeks it should be |
351 |
ready for first commit.</p> |
359 |
ready for first commit.</p> |
352 |
|
360 |
|
353 |
<p><strong class="speaker">RW</strong> : Informally beginning to |
361 |
<p><strong class="speaker">rwatson</strong> : Informally beginning to |
354 |
measure performance now. What are the right sets of tests? Need to |
362 |
measure performance now. What are the right sets of tests? Need to |
355 |
discuss on -arch.</p> |
363 |
discuss on -arch.</p> |
356 |
|
364 |
|
357 |
<p><strong class="speaker">AC</strong> : It would be nice to have that |
365 |
<p><strong class="speaker">alc</strong> : It would be nice to have that |
358 |
committed to the tools directory.</p> |
366 |
committed to the tools directory.</p> |
359 |
|
367 |
|
360 |
<p><strong class="speaker">JB</strong> : The statistics analysis |
368 |
<p><strong class="speaker">jhb</strong> : The statistics analysis |
361 |
package are we using.</p> |
369 |
package are we using.</p> |
362 |
|
370 |
|
363 |
<p><strong class="speaker">BM</strong> : I have some good success with |
371 |
<p><strong class="speaker">bmilekic</strong> : I have some good success with |
364 |
netpipe for overall measurement.</p> |
372 |
netpipe for overall measurement.</p> |
365 |
|
373 |
|
366 |
<p><strong class="speaker">RW</strong> : Need to be using consistent |
374 |
<p><strong class="speaker">rwatson</strong> : Need to be using consistent |
367 |
compilers because of the compiler change. Also all our debugging |
375 |
compilers because of the compiler change. Also all our debugging |
368 |
stuff will slow down the benchmarking.</p> |
376 |
stuff will slow down the benchmarking.</p> |
369 |
</div> |
377 |
</div> |
Lines 401-411
stuff will slow down the benchmarking.</
Link Here
|
401 |
|
409 |
|
402 |
<div class="discussion"> |
410 |
<div class="discussion"> |
403 |
|
411 |
|
404 |
<p><strong class="speaker">MD</strong> : Debug stuff on 5.0. I think |
412 |
<p><strong class="speaker">dillon</strong> : Debug stuff on 5.0. I think |
405 |
it might be reasonable then to take the space hit and always have the |
413 |
it might be reasonable then to take the space hit and always have the |
406 |
debugging in but turn it on and off with sysctl.</p> |
414 |
debugging in but turn it on and off with sysctl.</p> |
407 |
|
415 |
|
408 |
<p><strong class="speaker">RW</strong> : We should commit an optimized |
416 |
<p><strong class="speaker">rwatson</strong> : We should commit an optimized |
409 |
kernel configuration and benchmarking guidlines to the tree as |
417 |
kernel configuration and benchmarking guidlines to the tree as |
410 |
well.</p> |
418 |
well.</p> |
411 |
</div> |
419 |
</div> |
Lines 414-498
well.</p>
Link Here
|
414 |
|
422 |
|
415 |
<div class="discussion"> |
423 |
<div class="discussion"> |
416 |
|
424 |
|
417 |
<p><strong class="speaker">RW</strong> : I think we should continue |
425 |
<p><strong class="speaker">rwatson</strong> : I think we should continue |
418 |
the performance discussion. We want to accomplish a couple of things. |
426 |
the performance discussion. We want to accomplish a couple of things. |
419 |
One is stability measurement. What are the things we need to be |
427 |
One is stability measurement. What are the things we need to be |
420 |
measuring? What is our definition of useful?</p> |
428 |
measuring? What is our definition of useful?</p> |
421 |
|
429 |
|
422 |
<p><strong class="speaker">Jefferey</strong> : End to end measurement |
430 |
<p><strong class="speaker">hsu</strong> : End to end measurement |
423 |
with gigabit cards. For latency test connections per second. Can use |
431 |
with gigabit cards. For latency test connections per second. Can use |
424 |
ttcp or netbench in ports.</p> |
432 |
ttcp or netbench in ports.</p> |
425 |
|
433 |
|
426 |
<p><strong class="speaker">gnn</strong> : need to make sure we run |
434 |
<p><strong class="speaker">gnn</strong> : need to make sure we run |
427 |
against all of 4.6</p> |
435 |
against all of 4.6</p> |
428 |
|
436 |
|
429 |
<p><strong class="speaker">JE</strong> : Need to really have 3 tests. |
437 |
<p><strong class="speaker">julian</strong> : Need to really have 3 tests. |
430 |
4.6 (forever) 4.x (following updates) and -CURRENT.</p> |
438 |
4.6 (forever) 4.x (following updates) and -CURRENT.</p> |
431 |
|
439 |
|
432 |
<p><strong class="speaker">RW</strong> : There are other dimensions. |
440 |
<p><strong class="speaker">rwatson</strong> : There are other dimensions. |
433 |
Degree of parallelism for instance. We might see degradation in uni |
441 |
Degree of parallelism for instance. We might see degradation in uni |
434 |
but get good stuff in multi case.</p> |
442 |
but get good stuff in multi case.</p> |
435 |
|
443 |
|
436 |
<p><strong class="speaker">JE</strong> : Test for impact of KSE |
444 |
<p><strong class="speaker">julian</strong> : Test for impact of KSE |
437 |
complications as well.</p> |
445 |
complications as well.</p> |
438 |
|
446 |
|
439 |
<p><strong class="speaker">AP</strong> : I think as the results come |
447 |
<p><strong class="speaker">alfred</strong> : I think as the results come |
440 |
through people should not be too worried about it. Perhaps we should |
448 |
through people should not be too worried about it. Perhaps we should |
441 |
benchmark database type stuff as well. Need to do something |
449 |
benchmark database type stuff as well. Need to do something |
442 |
comprehensive.</p> |
450 |
comprehensive.</p> |
443 |
|
451 |
|
444 |
<p><strong class="speaker">DO</strong> : What does the test matrix |
452 |
<p><strong class="speaker">obrien</strong> : What does the test matrix |
445 |
look like? Different architectures and different numbers of |
453 |
look like? Different architectures and different numbers of |
446 |
processors.</p> |
454 |
processors.</p> |
447 |
|
455 |
|
448 |
<p><strong class="speaker">RW</strong> : Can we make a multi-processor |
456 |
<p><strong class="speaker">rwatson</strong> : Can we make a multi-processor |
449 |
run uni-procesor.</p> |
457 |
run uni-procesor.</p> |
450 |
|
458 |
|
451 |
<p><strong class="speaker">AP</strong> : Run queue and scheduler stuff?</p> |
459 |
<p><strong class="speaker">alfred</strong> : Run queue and scheduler stuff?</p> |
452 |
|
460 |
|
453 |
<p><strong class="speaker">JE</strong> : Will talk to Alfred.</p> |
461 |
<p><strong class="speaker">julian</strong> : Will talk to Alfred.</p> |
454 |
|
462 |
|
455 |
<p><strong class="speaker">RW</strong> : Is scalability testing important?</p> |
463 |
<p><strong class="speaker">rwatson</strong> : Is scalability testing important?</p> |
456 |
|
464 |
|
457 |
<p><strong class="speaker">DavidM</strong> : NFS testing.</p> |
465 |
<p><strong class="speaker">obrienM</strong> : NFS testing.</p> |
458 |
|
466 |
|
459 |
<p><strong class="speaker">RW</strong> : What about UI testing?</p> |
467 |
<p><strong class="speaker">rwatson</strong> : What about UI testing?</p> |
460 |
|
468 |
|
461 |
<p><strong class="speaker">JX</strong> : x11perf is the way to do that.</p> |
469 |
<p><strong class="speaker">hsu</strong> : x11perf is the way to do that.</p> |
462 |
|
470 |
|
463 |
<p><strong class="speaker">MD</strong> : Currently we have a directory |
471 |
<p><strong class="speaker">dillon</strong> : Currently we have a directory |
464 |
for regression tests, should we do one for performance tests?</p> |
472 |
for regression tests, should we do one for performance tests?</p> |
465 |
|
473 |
|
466 |
<p><strong class="speaker">gnn</strong> : talk to sleepycat for DB |
474 |
<p><strong class="speaker">gnn</strong> : talk to sleepycat for DB |
467 |
tests, see if they have some</p> |
475 |
tests, see if they have some</p> |
468 |
|
476 |
|
469 |
<p><strong class="speaker">AP</strong> : Really nice to tests DB |
477 |
<p><strong class="speaker">alfred</strong> : Really nice to tests DB |
470 |
applications that are heavily thread dependent.</p> |
478 |
applications that are heavily thread dependent.</p> |
471 |
|
479 |
|
472 |
<p><strong class="speaker">Jefferey</strong> :Apache 2 has threads.</p> |
480 |
<p><strong class="speaker">hsu</strong> :Apache 2 has threads.</p> |
473 |
|
481 |
|
474 |
<p><strong class="speaker">RW</strong> : What about commercial folks? |
482 |
<p><strong class="speaker">rwatson</strong> : What about commercial folks? |
475 |
What do you do.</p> |
483 |
What do you do.</p> |
476 |
|
484 |
|
477 |
<p><strong class="speaker">Paul Saab</strong> : Normally what we end |
485 |
<p><strong class="speaker">ps</strong> : Normally what we end |
478 |
up doing is using the snapshot on some machines and see if the bugs |
486 |
up doing is using the snapshot on some machines and see if the bugs |
479 |
are out. There is no performance testing really.</p> |
487 |
are out. There is no performance testing really.</p> |
480 |
|
488 |
|
481 |
<p><strong class="speaker">RW</strong> : Again, what about performance?</p> |
489 |
<p><strong class="speaker">rwatson</strong> : Again, what about performance?</p> |
482 |
|
490 |
|
483 |
<p><strong class="speaker">Paul Saab</strong> : We've really never had |
491 |
<p><strong class="speaker">ps</strong> : We've really never had |
484 |
one. It's more just bugs. We've just never found the performance to |
492 |
one. It's more just bugs. We've just never found the performance to |
485 |
be a problem.</p> |
493 |
be a problem.</p> |
486 |
|
494 |
|
487 |
<p><strong class="speaker">RW</strong> : We need to create a forum for |
495 |
<p><strong class="speaker">rwatson</strong> : We need to create a forum for |
488 |
talking about performance. We need reproducible test cases.</p> |
496 |
talking about performance. We need reproducible test cases.</p> |
489 |
|
497 |
|
490 |
<p><strong class="speaker">Paul Saab</strong> : There's also other |
498 |
<p><strong class="speaker">ps</strong> : There's also other |
491 |
things. We've been doing lots of looking at this. FreeBSD gets |
499 |
things. We've been doing lots of looking at this. FreeBSD gets |
492 |
kicked down by attacks for instance. We have a lot of tools to get to |
500 |
kicked down by attacks for instance. We have a lot of tools to get to |
493 |
the project though.</p> |
501 |
the project though.</p> |
494 |
|
502 |
|
495 |
<p><strong class="speaker">RW</strong> : I will set up the mailing list.</p> |
503 |
<p><strong class="speaker">rwatson</strong> : I will set up the mailing list.</p> |
496 |
</div> |
504 |
</div> |
497 |
</div> |
505 |
</div> |
498 |
|
506 |
|
Lines 505-519
the project though.</p>
Link Here
|
505 |
|
513 |
|
506 |
<div class="discussion"> |
514 |
<div class="discussion"> |
507 |
|
515 |
|
508 |
<p><strong class="speaker">JB</strong> : Questions about alpha?</p> |
516 |
<p><strong class="speaker">jhb</strong> : Questions about alpha?</p> |
509 |
|
517 |
|
510 |
<p><strong class="speaker">RW</strong> : KSE on alpha?</p> |
518 |
<p><strong class="speaker">rwatson</strong> : KSE on alpha?</p> |
511 |
|
519 |
|
512 |
<p><strong class="speaker">JE</strong> : We have patches so it |
520 |
<p><strong class="speaker">julian</strong> : We have patches so it |
513 |
compiles and runs non-KSE programs. You can have the patched version |
521 |
compiles and runs non-KSE programs. You can have the patched version |
514 |
of the alpha kernel up and running though.</p> |
522 |
of the alpha kernel up and running though.</p> |
515 |
|
523 |
|
516 |
<p><strong class="speaker">RW</strong> : Is the task owned of making |
524 |
<p><strong class="speaker">rwatson</strong> : Is the task owned of making |
517 |
this work on Alpha?</p> |
525 |
this work on Alpha?</p> |
518 |
|
526 |
|
519 |
</div> |
527 |
</div> |
Lines 522-550
this work on Alpha?</p>
Link Here
|
522 |
|
530 |
|
523 |
<div class="discussion"> |
531 |
<div class="discussion"> |
524 |
|
532 |
|
525 |
<p><strong class="speaker">DR</strong> : It works as far as I get to |
533 |
<p><strong class="speaker">dfr</strong> : It works as far as I get to |
526 |
use it. It's not used in production right now.</p> |
534 |
use it. It's not used in production right now.</p> |
527 |
|
535 |
|
528 |
<p><strong class="speaker">PS</strong> : Intel shipped me a quad |
536 |
<p><strong class="speaker">ps</strong> : Intel shipped me a quad |
529 |
processor IA64 board. (McKinley is the name of the board).</p> |
537 |
processor IA64 board. (McKinley is the name of the board).</p> |
530 |
|
538 |
|
531 |
<p><strong class="speaker">RW</strong> : What does it need for 5.0?</p> |
539 |
<p><strong class="speaker">rwatson</strong> : What does it need for 5.0?</p> |
532 |
|
540 |
|
533 |
<p><strong class="speaker">DR</strong> : It works, it works for SMP. |
541 |
<p><strong class="speaker">dfr</strong> : It works, it works for SMP. |
534 |
Self hosts, build worlds. sysinstall compiles but needs more kicking |
542 |
Self hosts, build worlds. sysinstall compiles but needs more kicking |
535 |
to work.</p> |
543 |
to work.</p> |
536 |
|
544 |
|
537 |
<p><strong class="speaker">Paul Saab</strong> : Intel wants us to ship |
545 |
<p><strong class="speaker">ps</strong> : Intel wants us to ship |
538 |
a CD.</p> |
546 |
a CD.</p> |
539 |
|
547 |
|
540 |
<p><strong class="speaker">DR</strong> : There is no thread support |
548 |
<p><strong class="speaker">dfr</strong> : There is no thread support |
541 |
right now (threading library needs to move to get/setcontext rather |
549 |
right now (threading library needs to move to get/setcontext rather |
542 |
than longjmp).</p> |
550 |
than longjmp).</p> |
543 |
|
551 |
|
544 |
<p><strong class="speaker">DR</strong> : Need to move every driver to |
552 |
<p><strong class="speaker">dfr</strong> : Need to move every driver to |
545 |
use BUS DMA for large memory machines to get bounce buffers.</p> |
553 |
use BUS DMA for large memory machines to get bounce buffers.</p> |
546 |
|
554 |
|
547 |
<p><strong class="speaker">JB</strong> : PHK is working on using a new |
555 |
<p><strong class="speaker">jhb</strong> : PHK is working on using a new |
548 |
libwhisk so that sysinstall et al work on all systems.</p> |
556 |
libwhisk so that sysinstall et al work on all systems.</p> |
549 |
|
557 |
|
550 |
</div> |
558 |
</div> |
Lines 553-651
libwhisk so that sysinstall et al work o
Link Here
|
553 |
|
561 |
|
554 |
<div class="discussion"> |
562 |
<div class="discussion"> |
555 |
|
563 |
|
556 |
<p><strong class="speaker">Jake B</strong> : Take control of KSE stuff |
564 |
<p><strong class="speaker">jake</strong> : Take control of KSE stuff |
557 |
on Sparc 64.</p> |
565 |
on Sparc 64.</p> |
558 |
|
566 |
|
559 |
<p><strong class="speaker">RW</strong> : Do we have a Sparc 64 in the |
567 |
<p><strong class="speaker">rwatson</strong> : Do we have a Sparc 64 in the |
560 |
cluster?</p> |
568 |
cluster?</p> |
561 |
|
569 |
|
562 |
<p><strong class="speaker">Jake B</strong> : It's not in the cluster |
570 |
<p><strong class="speaker">jake</strong> : It's not in the cluster |
563 |
yet. It's a serial cluster issue.</p> |
571 |
yet. It's a serial cluster issue.</p> |
564 |
|
572 |
|
565 |
<p><strong class="speaker">RW</strong> : Package building on S64?</p> |
573 |
<p><strong class="speaker">rwatson</strong> : Package building on S64?</p> |
566 |
|
574 |
|
567 |
<p><strong class="speaker">Jake B</strong> : Perhaps a bunch of Ultra |
575 |
<p><strong class="speaker">jake</strong> : Perhaps a bunch of Ultra |
568 |
60s for a package build.</p> |
576 |
60s for a package build.</p> |
569 |
|
577 |
|
570 |
<p><strong class="speaker">David</strong> : 1500 build right now?</p> |
578 |
<p><strong class="speaker">obrien</strong> : 1500 build right now?</p> |
571 |
|
579 |
|
572 |
<p><strong class="speaker">Jake B</strong> : Yes, but a lot of the |
580 |
<p><strong class="speaker">jake</strong> : Yes, but a lot of the |
573 |
same bug in packages are broken.</p> |
581 |
same bug in packages are broken.</p> |
574 |
|
582 |
|
575 |
<p><strong class="speaker">JB</strong> : Timeline for X?</p> |
583 |
<p><strong class="speaker">jhb</strong> : Timeline for X?</p> |
576 |
|
584 |
|
577 |
<p><strong class="speaker">Jake B</strong> : Not really.</p> |
585 |
<p><strong class="speaker">jake</strong> : Not really.</p> |
578 |
|
586 |
|
579 |
<p><strong class="speaker">RW</strong> : In terms of 5.0 how |
587 |
<p><strong class="speaker">rwatson</strong> : In terms of 5.0 how |
580 |
comfortable are you?</p> |
588 |
comfortable are you?</p> |
581 |
|
589 |
|
582 |
<p><strong class="speaker">Jake B</strong> : sysinstall is the only problem.</p> |
590 |
<p><strong class="speaker">jake</strong> : sysinstall is the only problem.</p> |
583 |
</div> |
591 |
</div> |
584 |
|
592 |
|
585 |
<h3>PowerPC</h3> |
593 |
<h3>PowerPC</h3> |
586 |
|
594 |
|
587 |
<div class="discussion"> |
595 |
<div class="discussion"> |
588 |
|
596 |
|
589 |
<p><strong class="speaker">Benno Rice</strong> : I got it up to |
597 |
<p><strong class="speaker">benno</strong> : I got it up to |
590 |
execing a fake init in the simulator and printing "hello world". |
598 |
execing a fake init in the simulator and printing "hello world". |
591 |
Trying to work with real hardware. I now have some semblance of |
599 |
Trying to work with real hardware. I now have some semblance of |
592 |
busdma and am working on other stuff. GEM on iMac is in an embryonic |
600 |
busdma and am working on other stuff. GEM on iMac is in an embryonic |
593 |
state. Should get to NFS mount in a few weeks.</p> |
601 |
state. Should get to NFS mount in a few weeks.</p> |
594 |
|
602 |
|
595 |
<p><strong class="speaker">RW</strong> : How do you feel about your |
603 |
<p><strong class="speaker">rwatson</strong> : How do you feel about your |
596 |
timeline?</p> |
604 |
timeline?</p> |
597 |
|
605 |
|
598 |
<p><strong class="speaker">Benno</strong> : I'm not sure we'll have |
606 |
<p><strong class="speaker">benno</strong> : I'm not sure we'll have |
599 |
something fully workable for 5.0.</p> |
607 |
something fully workable for 5.0.</p> |
600 |
|
608 |
|
601 |
<p><strong class="speaker">RW</strong> : You're not at the point yet |
609 |
<p><strong class="speaker">rwatson</strong> : You're not at the point yet |
602 |
on working on KSE are you?</p> |
610 |
on working on KSE are you?</p> |
603 |
|
611 |
|
604 |
<p><strong class="speaker">Benno</strong> : No, need a useful system |
612 |
<p><strong class="speaker">benno</strong> : No, need a useful system |
605 |
first.</p> |
613 |
first.</p> |
606 |
|
614 |
|
607 |
</div> |
615 |
</div> |
608 |
|
616 |
|
609 |
<h3>AMD64</h3> |
617 |
<h3>Adillon64</h3> |
610 |
|
618 |
|
611 |
<div class="discussion"> |
619 |
<div class="discussion"> |
612 |
|
620 |
|
613 |
<p><strong class="speaker">RW</strong> : I know that we're having |
621 |
<p><strong class="speaker">rwatson</strong> : I know that we're having |
614 |
simulator problems.</p> |
622 |
simulator problems.</p> |
615 |
|
623 |
|
616 |
<p><strong class="speaker">DO</strong> : The issues are about legal |
624 |
<p><strong class="speaker">obrien</strong> : The issues are about legal |
617 |
and NDA. AMD decided on <a href="http://www.freebsdmall.com">FreeBSD |
625 |
and NDA. Adillon decided on <a href="http://www.freebsdmall.com">FreeBSD |
618 |
Mall</a> as the NDA person. I have not had a working simulator since |
626 |
Mall</a> as the NDA person. I have not had a working simulator since |
619 |
September.</p> |
627 |
September.</p> |
620 |
|
628 |
|
621 |
<p><strong class="speaker">Paul</strong> : I can make that happen, as |
629 |
<p><strong class="speaker">ps</strong> : I can make that happen, as |
622 |
well as real hardware.</p> |
630 |
well as real hardware.</p> |
623 |
|
631 |
|
624 |
<p><strong class="speaker">DO</strong> :I've got a cross tool chain in |
632 |
<p><strong class="speaker">obrien</strong> :I've got a cross tool chain in |
625 |
the tree. I have some untested pmap stuff. Hardware has been |
633 |
the tree. I have some untested pmap stuff. Hardware has been |
626 |
available for a month or so. We could boot FreeBSD 4.6 today if only |
634 |
available for a month or so. We could boot FreeBSD 4.6 today if only |
627 |
we had hardware.</p> |
635 |
we had hardware.</p> |
628 |
|
636 |
|
629 |
<p><strong class="speaker">RW</strong> : What do you think about 5.0? |
637 |
<p><strong class="speaker">rwatson</strong> : What do you think about 5.0? |
630 |
Should we discuss that at another date?</p> |
638 |
Should we discuss that at another date?</p> |
631 |
|
639 |
|
632 |
</div> |
640 |
</div> |
633 |
|
641 |
|
634 |
<h3>MIPS</h3> |
642 |
<h3>MIps</h3> |
635 |
|
643 |
|
636 |
<div class="discussion"> |
644 |
<div class="discussion"> |
637 |
|
645 |
|
638 |
<p><strong class="speaker">???</strong> :Juniper offered.</p> |
646 |
<p><strong class="speaker">unknown</strong> :Juniper offered.</p> |
639 |
|
647 |
|
640 |
<p><strong class="speaker">DO</strong> : But we have no hardware.</p> |
648 |
<p><strong class="speaker">obrien</strong> : But we have no hardware.</p> |
641 |
|
649 |
|
642 |
<p><strong class="speaker">???</strong> :Juniper thinks it's OK but |
650 |
<p><strong class="speaker">unknown</strong> :Juniper thinks it's OK but |
643 |
doesn't want to have it rot in the tree.</p> |
651 |
doesn't want to have it rot in the tree.</p> |
644 |
|
652 |
|
645 |
<p><strong class="speaker">BD</strong> : I have a line on a company |
653 |
<p><strong class="speaker">brooks</strong> : I have a line on a company |
646 |
that does compact PCI with R6Ks.</p> |
654 |
that does compact PCI with R6Ks.</p> |
647 |
|
655 |
|
648 |
<p><strong class="speaker">RW</strong> : We're waiting for someone to |
656 |
<p><strong class="speaker">rwatson</strong> : We're waiting for someone to |
649 |
turn up.</p> |
657 |
turn up.</p> |
650 |
|
658 |
|
651 |
</div> |
659 |
</div> |
Lines 660-702
LUNCH
Link Here
|
660 |
<a name="trust"></a> |
668 |
<a name="trust"></a> |
661 |
<h2>Trusted BSD</h2> |
669 |
<h2>Trusted BSD</h2> |
662 |
|
670 |
|
663 |
<p><strong class="speaker">RW</strong> : MAC framework is what is of |
671 |
<p><strong class="speaker">rwatson</strong> : Malc framework is what is of |
664 |
interest today.</p> |
672 |
interest today.</p> |
665 |
|
673 |
|
666 |
<em>See Slides from Robert</em> |
674 |
<em>See Slides from Robert</em> |
667 |
|
675 |
|
668 |
<div class="discussion"> |
676 |
<div class="discussion"> |
669 |
|
677 |
|
670 |
<p><strong class="speaker">JE</strong> : Are the labels the same on |
678 |
<p><strong class="speaker">julian</strong> : Are the labels the same on |
671 |
all structures?</p> |
679 |
all structures?</p> |
672 |
|
680 |
|
673 |
<p><strong class="speaker">RW</strong> : You can modify this but there |
681 |
<p><strong class="speaker">rwatson</strong> : You can modify this but there |
674 |
are issues with memory: is the space needed for a label too large to |
682 |
are issues with memory: is the space needed for a label too large to |
675 |
add to an mbuf header, for example? The label is small, but there |
683 |
add to an mbuf header, for example? The label is small, but there |
676 |
area lot of them?</p> |
684 |
area lot of them?</p> |
677 |
|
685 |
|
678 |
<p><strong class="speaker">BM</strong> : When you're freeing the mbuf |
686 |
<p><strong class="speaker">bmilekic</strong> : When you're freeing the mbuf |
679 |
do you write the label data?</p> |
687 |
do you write the label data?</p> |
680 |
|
688 |
|
681 |
<p><strong class="speaker">RW</strong> : We blank it when we free it.</p> |
689 |
<p><strong class="speaker">rwatson</strong> : We blank it when we free it.</p> |
682 |
|
690 |
|
683 |
<p><strong class="speaker">BM</strong> : I do not think the 36 bytes |
691 |
<p><strong class="speaker">bmilekic</strong> : I do not think the 36 bytes |
684 |
in the mbuf header is a problem.</p> |
692 |
in the mbuf header is a problem.</p> |
685 |
|
693 |
|
686 |
<p><strong class="speaker">JE</strong> : I'm more interested in the |
694 |
<p><strong class="speaker">julian</strong> : I'm more interested in the |
687 |
"why" than the how.</p> |
695 |
"why" than the how.</p> |
688 |
|
696 |
|
689 |
<p><strong class="speaker">RW</strong> : A lot of people are |
697 |
<p><strong class="speaker">rwatson</strong> : A lot of people are |
690 |
interested in this. Some of the things that do interest a lot of |
698 |
interested in this. Some of the things that do interest a lot of |
691 |
people are things like doing on the fly security for a web server.</p> |
699 |
people are things like doing on the fly security for a web server.</p> |
692 |
|
700 |
|
693 |
<p><strong class="speaker">JE</strong> : Is there a black hatted TLA |
701 |
<p><strong class="speaker">julian</strong> : Is there a black hatted TLA |
694 |
interested?</p> |
702 |
interested?</p> |
695 |
|
703 |
|
696 |
<p><strong class="speaker">RW</strong> : Yes and several gov'ts. As |
704 |
<p><strong class="speaker">rwatson</strong> : Yes and several gov'ts. As |
697 |
well as plenty of financial folks.</p> |
705 |
well as plenty of financial folks.</p> |
698 |
|
706 |
|
699 |
<p><strong class="speaker">RW</strong> : There's a lot of userland |
707 |
<p><strong class="speaker">rwatson</strong> : There's a lot of userland |
700 |
stuff that's not done yet.</p> |
708 |
stuff that's not done yet.</p> |
701 |
</div> |
709 |
</div> |
702 |
</div> |
710 |
</div> |
Lines 706-876
stuff that's not done yet.</p>
Link Here
|
706 |
<a name="releng"></a> |
714 |
<a name="releng"></a> |
707 |
<h2>Release Engineering</h2> |
715 |
<h2>Release Engineering</h2> |
708 |
|
716 |
|
709 |
<p><strong class="speaker">MS</strong> : Shows a slide of releases. |
717 |
<p><strong class="speaker">murray</strong> : Shows a slide of releases. |
710 |
4.6 is ready to go but having issues with ISO images. DP1, a lot of |
718 |
4.6 is ready to go but having issues with ISO images. DP1, a lot of |
711 |
goals were met. 1000 packages were building on -CURRENT to get DP1 |
719 |
goals were met. 1000 packages were building on -CURRENT to get DP1 |
712 |
out. Polished 4.2. We need to start making decisions on 5.0. |
720 |
out. Polished 4.2. We need to start making decisions on 5.0. |
713 |
November is still the date we're shooting for. We're going to do a |
721 |
November is still the date we're shooting for. We're going to do a |
714 |
4.7 and a 4.8. DP3?</p> |
722 |
4.7 and a 4.8. DP3?</p> |
715 |
|
723 |
|
716 |
<p>***GET SLIDE FROM MURRAY***</p> |
724 |
<p>***GET samIDE FROM MURRAY***</p> |
717 |
|
725 |
|
718 |
<div class="discussion"> |
726 |
<div class="discussion"> |
719 |
|
727 |
|
720 |
<p><strong class="speaker">MS</strong> : Release engineering area of |
728 |
<p><strong class="speaker">murray</strong> : Release engineering area of |
721 |
the web site www.freebsd.org/releng. For DP2 question about p4 or |
729 |
the web site www.freebsd.org/releng. For DP2 question about p4 or |
722 |
CVS? Will probably use p4 for DP2 as well. USB subsystem? Perl |
730 |
CVS? Will probably use p4 for DP2 as well. USB subsystem? Perl |
723 |
removal? KSE?</p> |
731 |
removal? KSE?</p> |
724 |
|
732 |
|
725 |
<p><strong class="speaker">JE</strong> : KSE should be able to run |
733 |
<p><strong class="speaker">julian</strong> : KSE should be able to run |
726 |
simple tests.</p> |
734 |
simple tests.</p> |
727 |
|
735 |
|
728 |
<p><strong class="speaker">DO</strong> : Is whatever you have |
736 |
<p><strong class="speaker">obrien</strong> : Is whatever you have |
729 |
committed by DP2 be the same as the release.</p> |
737 |
committed by DP2 be the same as the release.</p> |
730 |
|
738 |
|
731 |
<p><strong class="speaker">JE</strong> : It will be a subset.</p> |
739 |
<p><strong class="speaker">julian</strong> : It will be a subset.</p> |
732 |
|
740 |
|
733 |
<p><strong class="speaker">MS</strong> : What will the status be of |
741 |
<p><strong class="speaker">murray</strong> : What will the status be of |
734 |
KSE in userland for 5.0?</p> |
742 |
KSE in userland for 5.0?</p> |
735 |
|
743 |
|
736 |
<p><strong class="speaker">JE</strong> : Can't answer that right |
744 |
<p><strong class="speaker">julian</strong> : Can't answer that right |
737 |
now. We're not removing the old libraries. The userland work will |
745 |
now. We're not removing the old libraries. The userland work will |
738 |
happen between DP2 and release. The next step is MP as well as |
746 |
happen between DP2 and release. The next step is MP as well as |
739 |
UP.</p> |
747 |
UP.</p> |
740 |
|
748 |
|
741 |
<p><strong class="speaker">DO</strong> : Are we heading for a release?</p> |
749 |
<p><strong class="speaker">obrien</strong> : Are we heading for a release?</p> |
742 |
|
750 |
|
743 |
<p><strong class="speaker">MS</strong> : yes.</p> |
751 |
<p><strong class="speaker">murray</strong> : yes.</p> |
744 |
|
752 |
|
745 |
<p><strong class="speaker">DO</strong> : Then we have to stop having |
753 |
<p><strong class="speaker">obrien</strong> : Then we have to stop having |
746 |
major commits.</p> |
754 |
major commits.</p> |
747 |
|
755 |
|
748 |
<p><strong class="speaker">MS</strong> : Yes, the discussion today is |
756 |
<p><strong class="speaker">murray</strong> : Yes, the discussion today is |
749 |
what are the major must have features.</p> |
757 |
what are the major must have features.</p> |
750 |
|
758 |
|
751 |
<p><strong class="speaker">RW</strong> : We need to decide if there |
759 |
<p><strong class="speaker">rwatson</strong> : We need to decide if there |
752 |
are major upcoming problems and reduce risk on things like KSE.</p> |
760 |
are major upcoming problems and reduce risk on things like KSE.</p> |
753 |
|
761 |
|
754 |
<p><strong class="speaker">JE</strong> : That's why I want to get MS 3 |
762 |
<p><strong class="speaker">julian</strong> : That's why I want to get murray 3 |
755 |
in now.</p> |
763 |
in now.</p> |
756 |
|
764 |
|
757 |
<p><strong class="speaker">RW</strong> : Do you think that KSE related |
765 |
<p><strong class="speaker">rwatson</strong> : Do you think that KSE related |
758 |
changes from later milestones are going to be isolated to KSE or |
766 |
changes from later milestones are going to be isolated to KSE or |
759 |
pervasive?</p> |
767 |
pervasive?</p> |
760 |
|
768 |
|
761 |
<p><strong class="speaker">JE</strong> : Hard to say. My guess is |
769 |
<p><strong class="speaker">julian</strong> : Hard to say. My guess is |
762 |
that MS 4 stuff should be less pervasive.</p> |
770 |
that murray 4 stuff should be less pervasive.</p> |
763 |
|
771 |
|
764 |
<p><strong class="speaker">RW</strong> : What happens if KSE just |
772 |
<p><strong class="speaker">rwatson</strong> : What happens if KSE just |
765 |
doesn't work?</p> |
773 |
doesn't work?</p> |
766 |
|
774 |
|
767 |
<p><strong class="speaker">JE</strong> : Well it does work, the |
775 |
<p><strong class="speaker">julian</strong> : Well it does work, the |
768 |
patches work, it's a question of risk. We need to check on new |
776 |
patches work, it's a question of risk. We need to check on new |
769 |
things, like locking two threads in the same process.</p> |
777 |
things, like locking two threads in the same process.</p> |
770 |
|
778 |
|
771 |
<p><strong class="speaker">MD</strong> : KSEs only become fragile when |
779 |
<p><strong class="speaker">dillon</strong> : KSEs only become fragile when |
772 |
pthread uses them. That's the turning point.</p> |
780 |
pthread uses them. That's the turning point.</p> |
773 |
|
781 |
|
774 |
<p><strong class="speaker">DO</strong> : I'd like the rules for the |
782 |
<p><strong class="speaker">obrien</strong> : I'd like the rules for the |
775 |
rest of the summer, I hope we'll talk about that.</p> |
783 |
rest of the summer, I hope we'll talk about that.</p> |
776 |
|
784 |
|
777 |
<p><strong class="speaker">MS</strong> : Earlier is better.</p> |
785 |
<p><strong class="speaker">murray</strong> : Earlier is better.</p> |
778 |
|
786 |
|
779 |
<p><strong class="speaker">JM</strong> : I think the cutoff point for |
787 |
<p><strong class="speaker">mini</strong> : I think the cutoff point for |
780 |
KSE might be MS 3.</p> |
788 |
KSE might be murray 3.</p> |
781 |
|
789 |
|
782 |
<p><strong class="speaker">RW</strong> : It's the kind of thing where |
790 |
<p><strong class="speaker">rwatson</strong> : It's the kind of thing where |
783 |
if we need to back out we can.</p> |
791 |
if we need to back out we can.</p> |
784 |
|
792 |
|
785 |
<p><strong class="speaker">JE</strong> : If you're not going to run |
793 |
<p><strong class="speaker">julian</strong> : If you're not going to run |
786 |
KSEs then you're OK.</p> |
794 |
KSEs then you're OK.</p> |
787 |
|
795 |
|
788 |
<p><strong class="speaker">RW</strong> : I think it's low risk. Let's |
796 |
<p><strong class="speaker">rwatson</strong> : I think it's low risk. Let's |
789 |
avoid the risk is the message.</p> |
797 |
avoid the risk is the message.</p> |
790 |
|
798 |
|
791 |
<p><strong class="speaker">JE</strong> : The next DP2 (where we'd like |
799 |
<p><strong class="speaker">julian</strong> : The next DP2 (where we'd like |
792 |
MS4).</p> |
800 |
murray4).</p> |
793 |
|
801 |
|
794 |
<p><strong class="speaker">AP</strong> : We really need KSE so all |
802 |
<p><strong class="speaker">alfred</strong> : We really need KSE so all |
795 |
this concern about stuff that no one really uses is not a big deal. |
803 |
this concern about stuff that no one really uses is not a big deal. |
796 |
People just need to play catch up. We have performance problems and |
804 |
People just need to play catch up. We have performance problems and |
797 |
we need to solve those.</p> |
805 |
we need to solve those.</p> |
798 |
|
806 |
|
799 |
<p><strong class="speaker">DO</strong> : We quickly need to figure out |
807 |
<p><strong class="speaker">obrien</strong> : We quickly need to figure out |
800 |
our policy on multiple archs.</p> |
808 |
our policy on multiple archs.</p> |
801 |
|
809 |
|
802 |
<p><strong class="speaker">RW</strong> : I briefly want to respond to |
810 |
<p><strong class="speaker">rwatson</strong> : I briefly want to respond to |
803 |
Alfred. We have asserted that KSE will be experimental. It will be |
811 |
Alfred. We have asserted that KSE will be experimental. It will be |
804 |
in and 5.0 will go out but there might be issues.</p> |
812 |
in and 5.0 will go out but there might be issues.</p> |
805 |
|
813 |
|
806 |
<p><strong class="speaker">JB</strong> : Realistically for the network |
814 |
<p><strong class="speaker">jhb</strong> : Realistically for the network |
807 |
stack is that IPv4 sockets will not be giant. But this is only in the |
815 |
stack is that IPv4 sockets will not be giant. But this is only in the |
808 |
network stack world. Several people are working on it.</p> |
816 |
network stack world. Several people are working on it.</p> |
809 |
|
817 |
|
810 |
<p><strong class="speaker">RW</strong> : The GEOM stuff will be |
818 |
<p><strong class="speaker">rwatson</strong> : The GEOM stuff will be |
811 |
enabled by default in 5.0. Sparc depends on it. I do not know what |
819 |
enabled by default in 5.0. Sparc depends on it. I do not know what |
812 |
the impediments are to that though.</p> |
820 |
the impediments are to that though.</p> |
813 |
|
821 |
|
814 |
<p><strong class="speaker">JE</strong> : The kernel stuff is there but |
822 |
<p><strong class="speaker">julian</strong> : The kernel stuff is there but |
815 |
the user space is not. It can't become the default until everything |
823 |
the user space is not. It can't become the default until everything |
816 |
is there.</p> |
824 |
is there.</p> |
817 |
|
825 |
|
818 |
<p><strong class="speaker">WL</strong> : What level of control are you |
826 |
<p><strong class="speaker">warner</strong> : What level of control are you |
819 |
going to exercise over the tree in the coming months?</p> |
827 |
going to exercise over the tree in the coming months?</p> |
820 |
|
828 |
|
821 |
<p><strong class="speaker">MS</strong> : You're going to see more |
829 |
<p><strong class="speaker">murray</strong> : You're going to see more |
822 |
level of control but we expect the requests to be reasonable. It's a |
830 |
level of control but we expect the requests to be reasonable. It's a |
823 |
very open process.</p> |
831 |
very open process.</p> |
824 |
|
832 |
|
825 |
<p><strong class="speaker">JB</strong> : How are we going to address the 5/6 split? |
833 |
<p><strong class="speaker">jhb</strong> : How are we going to address the 5/6 split? |
826 |
|
834 |
|
827 |
<p><strong class="speaker">MS</strong> : Carefully is all I can |
835 |
<p><strong class="speaker">murray</strong> : Carefully is all I can |
828 |
say.</p> |
836 |
say.</p> |
829 |
|
837 |
|
830 |
<p><strong class="speaker">RW</strong> : For 5. 0 we need to have a |
838 |
<p><strong class="speaker">rwatson</strong> : For 5. 0 we need to have a |
831 |
more informed decision. The release engineers will be trying to |
839 |
more informed decision. The release engineers will be trying to |
832 |
reduce the number of large code changes more as time goes by. We |
840 |
reduce the number of large code changes more as time goes by. We |
833 |
don't have to wait for 5.x to be perfectly stable before we branch.</p> |
841 |
don't have to wait for 5.x to be perfectly stable before we branch.</p> |
834 |
|
842 |
|
835 |
<p><strong class="speaker">MS</strong> : Let's move it to more general |
843 |
<p><strong class="speaker">murray</strong> : Let's move it to more general |
836 |
discussion of DP2? Specific technologies.</p> |
844 |
discussion of DP2? Specific technologies.</p> |
837 |
|
845 |
|
838 |
<p><strong class="speaker">BM</strong> : Is there a strategy to lock |
846 |
<p><strong class="speaker">bmilekic</strong> : Is there a strategy to lock |
839 |
other protocols that are not locked down onw?</p> |
847 |
other protocols that are not locked down onw?</p> |
840 |
|
848 |
|
841 |
<p><strong class="speaker">DO</strong> : How much more do we need to |
849 |
<p><strong class="speaker">obrien</strong> : How much more do we need to |
842 |
do before 5.0?</p> |
850 |
do before 5.0?</p> |
843 |
|
851 |
|
844 |
<p><strong class="speaker">JB</strong> : Bug fixing is what we're doing.</p> |
852 |
<p><strong class="speaker">jhb</strong> : Bug fixing is what we're doing.</p> |
845 |
|
853 |
|
846 |
<p><strong class="speaker">RW</strong> : The answer on the network |
854 |
<p><strong class="speaker">rwatson</strong> : The answer on the network |
847 |
stack. We need to choose a strategy on how to handle the other |
855 |
stack. We need to choose a strategy on how to handle the other |
848 |
protocols.</p> |
856 |
protocols.</p> |
849 |
|
857 |
|
850 |
<p><strong class="speaker">DO</strong> : The crux is that socket |
858 |
<p><strong class="speaker">obrien</strong> : The crux is that socket |
851 |
locking must be in 5.0.</p> |
859 |
locking must be in 5.0.</p> |
852 |
|
860 |
|
853 |
<p><strong class="speaker">RW</strong> : There are 2 or 3 problems. |
861 |
<p><strong class="speaker">rwatson</strong> : There are 2 or 3 problems. |
854 |
Routing code is a problem. See earlier discussions.</p> |
862 |
Routing code is a problem. See earlier discussions.</p> |
855 |
|
863 |
|
856 |
<p><strong class="speaker">Doug</strong> : RCng is essentially done. |
864 |
<p><strong class="speaker">dfr</strong> : RCng is essentially done. |
857 |
What it needs is testers.</p> |
865 |
What it needs is testers.</p> |
858 |
|
866 |
|
859 |
<p><strong class="speaker">AP</strong> : What about libh (I think libh |
867 |
<p><strong class="speaker">alfred</strong> : What about libh (I think libh |
860 |
is wrong but this is what I heard)?</p> |
868 |
is wrong but this is what I heard)?</p> |
861 |
|
869 |
|
862 |
<p><strong class="speaker">JB</strong> : It's very far along but not a |
870 |
<p><strong class="speaker">jhb</strong> : It's very far along but not a |
863 |
5.0 thing.</p> |
871 |
5.0 thing.</p> |
864 |
|
872 |
|
865 |
<p><strong class="speaker">WL</strong> : Problems with interrupt |
873 |
<p><strong class="speaker">warner</strong> : Problems with interrupt |
866 |
routing in ACPCI?</p> |
874 |
routing in alcPCI?</p> |
867 |
|
875 |
|
868 |
<p><strong class="speaker">Watanabe</strong> : Cannot handle PCI PCI |
876 |
<p><strong class="speaker">takawata</strong> : Cannot handle PCI PCI |
869 |
interrupt routing. Many 802.11x have this problem.</p> |
877 |
interrupt routing. Many 802.11x have this problem.</p> |
870 |
|
878 |
|
871 |
<p><strong class="speaker">JE</strong> : Is it a problem from Intel?</p> |
879 |
<p><strong class="speaker">julian</strong> : Is it a problem from Intel?</p> |
872 |
|
880 |
|
873 |
<p><strong class="speaker">Watanabe</strong> : This is not an Intel |
881 |
<p><strong class="speaker">takawata</strong> : This is not an Intel |
874 |
problem but a problem on our side. PCI PCI routing code should be |
882 |
problem but a problem on our side. PCI PCI routing code should be |
875 |
added. New code is necessary.</p> |
883 |
added. New code is necessary.</p> |
876 |
|
884 |
|
Lines 879-890
Whiteboard
Link Here
|
879 |
|
887 |
|
880 |
UFS2 rcNG KSE M3 CAM SMPng |
888 |
UFS2 rcNG KSE M3 CAM SMPng |
881 |
|
889 |
|
882 |
GEOM TrustedBSD MAC BusDMA Newbus SMPng |
890 |
GEOM TrustedBSD Malc BusDMA Newbus SMPng |
883 |
|
891 |
|
884 |
C++ Cardbus libwhisk/sysinstall KOBJ? (no!) |
892 |
C++ Cardbus libwhisk/sysinstall KOBJ? (no!) |
885 |
sparc64 |
893 |
sparc64 |
886 |
|
894 |
|
887 |
Perl Removal ACPI Alpha SMP Stability Pkgs for |
895 |
Perl Removal alcPI Alpha SMP Stability Pkgs for |
888 |
sparc64, IA64 |
896 |
sparc64, IA64 |
889 |
|
897 |
|
890 |
devd PCI intr route document hints release docs |
898 |
devd PCI intr route document hints release docs |
Lines 892-917
devd PCI intr route document hints rel
Link Here
|
892 |
platform |
900 |
platform |
893 |
</pre> |
901 |
</pre> |
894 |
|
902 |
|
895 |
<p><strong class="speaker">???</strong> : Firewire?</p> |
903 |
<p><strong class="speaker">unknown</strong> : Firewire?</p> |
896 |
|
904 |
|
897 |
<p><strong class="speaker">RW</strong> : What hardware shipping on |
905 |
<p><strong class="speaker">rwatson</strong> : What hardware shipping on |
898 |
IA64?</p> |
906 |
IA64?</p> |
899 |
|
907 |
|
900 |
<p><strong class="speaker">DR</strong> : Intel stuff</p> |
908 |
<p><strong class="speaker">dfr</strong> : Intel stuff</p> |
901 |
|
909 |
|
902 |
<p><strong class="speaker">RW</strong> : What about on Sparc64?</p> |
910 |
<p><strong class="speaker">rwatson</strong> : What about on Sparc64?</p> |
903 |
|
911 |
|
904 |
<p><strong class="speaker">DO</strong> : Very limited (hme...)</p> |
912 |
<p><strong class="speaker">obrien</strong> : Very limited (hme...)</p> |
905 |
|
913 |
|
906 |
<p><strong class="speaker">RW</strong> : KOBJ extensions discussed at |
914 |
<p><strong class="speaker">rwatson</strong> : KOBJ extensions discussed at |
907 |
BSDCon?</p> |
915 |
BSDCon?</p> |
908 |
|
916 |
|
909 |
<p><strong class="speaker">WL</strong> : Not sure, probably not for |
917 |
<p><strong class="speaker">warner</strong> : Not sure, probably not for |
910 |
5.0. Pervasive, so no.</p> |
918 |
5.0. Pervasive, so no.</p> |
911 |
|
919 |
|
912 |
<p><strong class="speaker">RW</strong> : How broken is C++?</p> |
920 |
<p><strong class="speaker">rwatson</strong> : How broken is C++?</p> |
913 |
|
921 |
|
914 |
<p><strong class="speaker">DO</strong> : Only on sparc64. Don't |
922 |
<p><strong class="speaker">obrien</strong> : Only on sparc64. Don't |
915 |
really know yet, but it's probably a library issue. The compiler is a |
923 |
really know yet, but it's probably a library issue. The compiler is a |
916 |
pre-release snapshot. The diffs are now getting large from May 5 to |
924 |
pre-release snapshot. The diffs are now getting large from May 5 to |
917 |
now. We should attempt to be as far along this gcc branch as possible |
925 |
now. We should attempt to be as far along this gcc branch as possible |
Lines 929-988
come release.</p>
Link Here
|
929 |
|
937 |
|
930 |
<div class="discussion"> |
938 |
<div class="discussion"> |
931 |
|
939 |
|
932 |
<p><strong class="speaker">GT</strong> : Talking about rc.d stuff. |
940 |
<p><strong class="speaker">gordon</strong> : Talking about rc.d stuff. |
933 |
Import from NetBSD. Right now we have patches out there that are |
941 |
Import from NetBSD. Right now we have patches out there that are |
934 |
translated from the current boot order. It's in perforce. After the |
942 |
translated from the current boot order. It's in perforce. After the |
935 |
conference it will go into the mainline. Single toggle for |
943 |
conference it will go into the mainline. Single toggle for |
936 |
booting.</p> |
944 |
booting.</p> |
937 |
|
945 |
|
938 |
<p><strong class="speaker">RW</strong> : How in sync are the bits in |
946 |
<p><strong class="speaker">rwatson</strong> : How in sync are the bits in |
939 |
the new stuff with the old stuff.</p> |
947 |
the new stuff with the old stuff.</p> |
940 |
|
948 |
|
941 |
<p><strong class="speaker">GT</strong> : Last patch is from June 3rd, |
949 |
<p><strong class="speaker">gordon</strong> : Last patch is from June 3rd, |
942 |
but it's tracking closely.</p> |
950 |
but it's tracking closely.</p> |
943 |
|
951 |
|
944 |
<p><strong class="speaker">RW</strong> : What is the schedule for |
952 |
<p><strong class="speaker">rwatson</strong> : What is the schedule for |
945 |
committing to the main tree.</p> |
953 |
committing to the main tree.</p> |
946 |
|
954 |
|
947 |
<p><strong class="speaker">GT</strong> : We have large patches so |
955 |
<p><strong class="speaker">gordon</strong> : We have large patches so |
948 |
we're going to re-import from NetBSD.</p> |
956 |
we're going to re-import from NetBSD.</p> |
949 |
|
957 |
|
950 |
<p><strong class="speaker">RW</strong> : How about you have it done by |
958 |
<p><strong class="speaker">rwatson</strong> : How about you have it done by |
951 |
July 1?</p> |
959 |
July 1?</p> |
952 |
|
960 |
|
953 |
<p><strong class="speaker">GT</strong> : We could probably do that. |
961 |
<p><strong class="speaker">gordon</strong> : We could probably do that. |
954 |
Definitely want to be in DP2.</p> |
962 |
Definitely want to be in DP2.</p> |
955 |
|
963 |
|
956 |
<p><strong class="speaker">GS</strong> : How long will we keep the old |
964 |
<p><strong class="speaker">gshapiro</strong> : How long will we keep the old |
957 |
stuff for?</p> |
965 |
stuff for?</p> |
958 |
|
966 |
|
959 |
<p><strong class="speaker">GT</strong> : We'll keep them both in for a |
967 |
<p><strong class="speaker">gordon</strong> : We'll keep them both in for a |
960 |
while. Not more than 1.5 months though.</p> |
968 |
while. Not more than 1.5 months though.</p> |
961 |
|
969 |
|
962 |
<p><strong class="speaker">JE</strong> : Have you had a look at all at |
970 |
<p><strong class="speaker">julian</strong> : Have you had a look at all at |
963 |
the Mac OS/X startup code?</p> |
971 |
the Mac OS/X startup code?</p> |
964 |
|
972 |
|
965 |
<p><strong class="speaker">GT</strong> : No.</p> |
973 |
<p><strong class="speaker">gordon</strong> : No.</p> |
966 |
|
974 |
|
967 |
<p><strong class="speaker">JE</strong> : Do you deal with dependencies?</p> |
975 |
<p><strong class="speaker">julian</strong> : Do you deal with dependencies?</p> |
968 |
|
976 |
|
969 |
<p><strong class="speaker">GT</strong> : There is meta data in each |
977 |
<p><strong class="speaker">gordon</strong> : There is meta data in each |
970 |
script that says what needs what. There is a program that orders |
978 |
script that says what needs what. There is a program that orders |
971 |
everything correctly.</p> |
979 |
everything correctly.</p> |
972 |
|
980 |
|
973 |
<p><strong class="speaker">???</strong> : How does this effect the rc |
981 |
<p><strong class="speaker">unknown</strong> : How does this effect the rc |
974 |
script for ports install?</p> |
982 |
script for ports install?</p> |
975 |
|
983 |
|
976 |
<p><strong class="speaker">GT</strong> : We could make this available |
984 |
<p><strong class="speaker">gordon</strong> : We could make this available |
977 |
to ports but won't on the first version.</p> |
985 |
to ports but won't on the first version.</p> |
978 |
|
986 |
|
979 |
<p><strong class="speaker">AP</strong> : Can I recommend that you |
987 |
<p><strong class="speaker">alfred</strong> : Can I recommend that you |
980 |
recommend this to ports?</p> |
988 |
recommend this to ports?</p> |
981 |
|
989 |
|
982 |
<p><strong class="speaker">GT</strong> : Yes, the problem is that we |
990 |
<p><strong class="speaker">gordon</strong> : Yes, the problem is that we |
983 |
have so many ports.</p> |
991 |
have so many ports.</p> |
984 |
|
992 |
|
985 |
<p><strong class="speaker">RW</strong> : The reason for this is for |
993 |
<p><strong class="speaker">rwatson</strong> : The reason for this is for |
986 |
rebundlers of FreeBSD in their environments. We don't have to have it |
994 |
rebundlers of FreeBSD in their environments. We don't have to have it |
987 |
for DP2 but it should be an ultimate goal. We might need to have a |
995 |
for DP2 but it should be an ultimate goal. We might need to have a |
988 |
policy statement on this. That at date X all ports must use the new |
996 |
policy statement on this. That at date X all ports must use the new |
Lines 998-1125
system.</p>
Link Here
|
998 |
|
1006 |
|
999 |
<div class="discussion"> |
1007 |
<div class="discussion"> |
1000 |
|
1008 |
|
1001 |
<p><strong class="speaker">SL</strong> : I've been working on hardware |
1009 |
<p><strong class="speaker">sam</strong> : I've been working on hardware |
1002 |
crypto. I'm looking for consensus on getting hardware crypto in the |
1010 |
crypto. I'm looking for consensus on getting hardware crypto in the |
1003 |
kernel. This will not happen in 5.0.</p> |
1011 |
kernel. This will not happen in 5.0.</p> |
1004 |
|
1012 |
|
1005 |
<h3>Syscall vector change for 64bits</h3> |
1013 |
<h3>Syscall vector change for 64bits</h3> |
1006 |
|
1014 |
|
1007 |
|
1015 |
|
1008 |
<p><strong class="speaker">MD</strong> : Two ways to go. Need to |
1016 |
<p><strong class="speaker">dillon</strong> : Two ways to go. Need to |
1009 |
create a new syscall vector. The other is to do a 1 off replacement. |
1017 |
create a new syscall vector. The other is to do a 1 off replacement. |
1010 |
Prefer the former.</p> |
1018 |
Prefer the former.</p> |
1011 |
|
1019 |
|
1012 |
<p><strong class="speaker">RW</strong> : Perhaps we need to create a |
1020 |
<p><strong class="speaker">rwatson</strong> : Perhaps we need to create a |
1013 |
FreeBSD 5 syscall vector. Could be a new ABI.</p> |
1021 |
FreeBSD 5 syscall vector. Could be a new ABI.</p> |
1014 |
|
1022 |
|
1015 |
<p><strong class="speaker">JE</strong> : Aren't there enough other numbers?</p> |
1023 |
<p><strong class="speaker">julian</strong> : Aren't there enough other numbers?</p> |
1016 |
|
1024 |
|
1017 |
<p><strong class="speaker">RW</strong> : That's one way to look at it |
1025 |
<p><strong class="speaker">rwatson</strong> : That's one way to look at it |
1018 |
and other platforms have done that? Is that too heavy weight?</p> |
1026 |
and other platforms have done that? Is that too heavy weight?</p> |
1019 |
|
1027 |
|
1020 |
<p><strong class="speaker">JE</strong> : It sounds that way to me. |
1028 |
<p><strong class="speaker">julian</strong> : It sounds that way to me. |
1021 |
You end up having to replicate the old ones into the new one.</p> |
1029 |
You end up having to replicate the old ones into the new one.</p> |
1022 |
|
1030 |
|
1023 |
<p><strong class="speaker">MD</strong> : The issue is about pollution.</p> |
1031 |
<p><strong class="speaker">dillon</strong> : The issue is about pollution.</p> |
1024 |
|
1032 |
|
1025 |
<p><strong class="speaker">DR</strong> : Seems like too much work for 5.x</p> |
1033 |
<p><strong class="speaker">dfr</strong> : Seems like too much work for 5.x</p> |
1026 |
|
1034 |
|
1027 |
<p><strong class="speaker">JE</strong> : It's more work. There are |
1035 |
<p><strong class="speaker">julian</strong> : It's more work. There are |
1028 |
now two places. Why not talk to OpenBSD?</p> |
1036 |
now two places. Why not talk to OpenBSD?</p> |
1029 |
|
1037 |
|
1030 |
<p><strong class="speaker">RW</strong> : Should there be a BSD API? |
1038 |
<p><strong class="speaker">rwatson</strong> : Should there be a BSD alfredI? |
1031 |
Tough to do across projects.</p> |
1039 |
Tough to do across projects.</p> |
1032 |
|
1040 |
|
1033 |
<p><strong class="speaker">DO</strong> : Who here is going to see that |
1041 |
<p><strong class="speaker">obrien</strong> : Who here is going to see that |
1034 |
through? We have not talked to NetBSD about even SMP.</p> |
1042 |
through? We have not talked to NetBSD about even SMP.</p> |
1035 |
|
1043 |
|
1036 |
<p><strong class="speaker">AP</strong> : Does changing the syscall |
1044 |
<p><strong class="speaker">alfred</strong> : Does changing the syscall |
1037 |
table allow us to do clean up?</p> |
1045 |
table allow us to do clean up?</p> |
1038 |
|
1046 |
|
1039 |
<p><strong class="speaker">RW</strong> : We could do that without |
1047 |
<p><strong class="speaker">rwatson</strong> : We could do that without |
1040 |
doing 64bit syscall table.</p> |
1048 |
doing 64bit syscall table.</p> |
1041 |
|
1049 |
|
1042 |
<h3>5.x ABI stability</h3> |
1050 |
<h3>5.x ABI stability</h3> |
1043 |
|
1051 |
|
1044 |
|
1052 |
|
1045 |
<p><strong class="speaker">RW</strong> : There are new functions in |
1053 |
<p><strong class="speaker">rwatson</strong> : There are new functions in |
1046 |
5.x. At what point do we stop changing?</p> |
1054 |
5.x. At what point do we stop changing?</p> |
1047 |
|
1055 |
|
1048 |
<p><strong class="speaker">DR</strong> : When people start really using it.</p> |
1056 |
<p><strong class="speaker">dfr</strong> : When people start really using it.</p> |
1049 |
|
1057 |
|
1050 |
<p><strong class="speaker">RW</strong> : How do we tell? How did Solaris do it?</p> |
1058 |
<p><strong class="speaker">rwatson</strong> : How do we tell? How did Solaris do it?</p> |
1051 |
|
1059 |
|
1052 |
<p><strong class="speaker">Everyone</strong> : Know one knows.</p> |
1060 |
<p><strong class="speaker">Everyone</strong> : No one knows.</p> |
1053 |
|
1061 |
|
1054 |
<p><strong class="speaker">DR</strong> : It's too hard to add a |
1062 |
<p><strong class="speaker">dfr</strong> : It's too hard to add a |
1055 |
syscall vector. Library issues are a problem.</p> |
1063 |
syscall vector. Library issues are a problem.</p> |
1056 |
|
1064 |
|
1057 |
<p><strong class="speaker">DO</strong> : We can use ELF to handle that.</p> |
1065 |
<p><strong class="speaker">obrien</strong> : We can use ELF to handle that.</p> |
1058 |
|
1066 |
|
1059 |
<p><strong class="speaker">DR</strong> : Let's just add 20 new |
1067 |
<p><strong class="speaker">dfr</strong> : Let's just add 20 new |
1060 |
syscalls instead of adding new work that we don't really really need.</p> |
1068 |
syscalls instead of adding new work that we don't really really need.</p> |
1061 |
|
1069 |
|
1062 |
<p><strong class="speaker">RW</strong> : Punt on lack of time to do |
1070 |
<p><strong class="speaker">rwatson</strong> : Punt on lack of time to do |
1063 |
this.</p> |
1071 |
this.</p> |
1064 |
|
1072 |
|
1065 |
<p><strong class="speaker">MD</strong> : I see DO's point with the |
1073 |
<p><strong class="speaker">dillon</strong> : I see obrien's point with the |
1066 |
libraries but I have done this with time_t at 64 bits.</p> |
1074 |
libraries but I have done this with time_t at 64 bits.</p> |
1067 |
|
1075 |
|
1068 |
<h3>devd</h3> |
1076 |
<h3>devd</h3> |
1069 |
|
1077 |
|
1070 |
|
1078 |
|
1071 |
<p><strong class="speaker">RW</strong> : The devd stuff was to |
1079 |
<p><strong class="speaker">rwatson</strong> : The devd stuff was to |
1072 |
integrate cardbus, newbus, etc.</p> |
1080 |
integrate cardbus, newbus, etc.</p> |
1073 |
|
1081 |
|
1074 |
<p><strong class="speaker">JE</strong> : To monitor requests to mount |
1082 |
<p><strong class="speaker">julian</strong> : To monitor requests to mount |
1075 |
or create new devices.</p> |
1083 |
or create new devices.</p> |
1076 |
|
1084 |
|
1077 |
<p><strong class="speaker">RW</strong> : Is this a 5.0 requirement? |
1085 |
<p><strong class="speaker">rwatson</strong> : Is this a 5.0 requirement? |
1078 |
Is there anyone to do this?</p> |
1086 |
Is there anyone to do this?</p> |
1079 |
|
1087 |
|
1080 |
<p><strong class="speaker">GT (from IRC)</strong> : PHK has patches |
1088 |
<!-- Which Gordon was this ? --> |
|
|
1089 |
<p><strong class="speaker">gordon (from IRC)</strong> : PHK has patches |
1081 |
that make having devd unnecessary.</p> |
1090 |
that make having devd unnecessary.</p> |
1082 |
|
1091 |
|
1083 |
<p><strong class="speaker">BD</strong> : Need something that does what |
1092 |
<p><strong class="speaker">brooks</strong> : Need something that does what |
1084 |
pccardd did. </p> |
1093 |
pccardd did. </p> |
1085 |
|
1094 |
|
1086 |
<p><strong class="speaker">JE</strong> : Need to be able to do this |
1095 |
<p><strong class="speaker">julian</strong> : Need to be able to do this |
1087 |
through a file. </p> |
1096 |
through a file. </p> |
1088 |
|
1097 |
|
1089 |
<p><strong class="speaker">WL</strong> : (from IRC): That's a 6.0 |
1098 |
<p><strong class="speaker">warner</strong> : (from IRC): That's a 6.0 |
1090 |
feature.</p> |
1099 |
feature.</p> |
1091 |
|
1100 |
|
1092 |
<p><strong class="speaker">JE</strong> : It would not be a large step |
1101 |
<p><strong class="speaker">julian</strong> : It would not be a large step |
1093 |
to put something in the middle to handle this.</p> |
1102 |
to put something in the middle to handle this.</p> |
1094 |
|
1103 |
|
1095 |
<p><strong class="speaker">JE</strong> : Sometime in the 5 lifetime we |
1104 |
<p><strong class="speaker">julian</strong> : Sometime in the 5 lifetime we |
1096 |
need this.</p> |
1105 |
need this.</p> |
1097 |
|
1106 |
|
1098 |
<p><strong class="speaker">WL</strong> : There is no way to monitor |
1107 |
<p><strong class="speaker">warner</strong> : There is no way to monitor |
1099 |
events in newbus but it would be easy to add.</p> |
1108 |
events in newbus but it would be easy to add.</p> |
1100 |
|
1109 |
|
1101 |
<p><strong class="speaker">JE</strong> : I'm not sure I understood you |
1110 |
<p><strong class="speaker">julian</strong> : I'm not sure I understood you |
1102 |
correctly.</p> |
1111 |
correctly.</p> |
1103 |
|
1112 |
|
1104 |
<p><strong class="speaker">WL</strong> : What happens now in a PCI is |
1113 |
<p><strong class="speaker">warner</strong> : What happens now in a PCI is |
1105 |
that it makes a call to pci_get_devid() and the driver would say "yes |
1114 |
that it makes a call to pci_get_devid() and the driver would say "yes |
1106 |
I am " or "no I'm not" so you'd have to change each of the busses to |
1115 |
I am " or "no I'm not" so you'd have to change each of the busses to |
1107 |
do this but that's not too tough because we have a small # of |
1116 |
do this but that's not too tough because we have a small # of |
1108 |
busses.</p> |
1117 |
busses.</p> |
1109 |
|
1118 |
|
1110 |
<p><strong class="speaker">JB</strong> : Mike Smith gave us an |
1119 |
<p><strong class="speaker">jhb</strong> : Mike Smith gave us an |
1111 |
informal tour of OS/X. OS/X uses XML to do this. They have the DEVID |
1120 |
informal tour of OS/X. OS/X uses XML to do this. They have the DEVID |
1112 |
in XML.</p> |
1121 |
in XML.</p> |
1113 |
|
1122 |
|
1114 |
<p><strong class="speaker">BD</strong> : I looked at some PCI drivers |
1123 |
<p><strong class="speaker">brooks</strong> : I looked at some PCI drivers |
1115 |
and some work that way but some don't.</p> |
1124 |
and some work that way but some don't.</p> |
1116 |
|
1125 |
|
1117 |
<p><strong class="speaker">JE</strong> : It seems to me we need to not |
1126 |
<p><strong class="speaker">julian</strong> : It seems to me we need to not |
1118 |
have to modify every single driver. If you've got a device that's not |
1127 |
have to modify every single driver. If you've got a device that's not |
1119 |
supported you ask all drivers. At the point when you run out you make |
1128 |
supported you ask all drivers. At the point when you run out you make |
1120 |
an outcall. The outcall returns does a substitution.</p> |
1129 |
an outcall. The outcall returns does a substitution.</p> |
1121 |
|
1130 |
|
1122 |
<p><strong class="speaker">RW</strong> : Time up, time to wrap up.</p> |
1131 |
<p><strong class="speaker">rwatson</strong> : Time up, time to wrap up.</p> |
1123 |
</div> |
1132 |
</div> |
1124 |
|
1133 |
|
1125 |
&footer; |
1134 |
&footer; |