|
Lines 28-39
Link Here
|
| 28 |
<li><a href="#goal">Project Goal</a></li> |
28 |
<li><a href="#goal">Project Goal</a></li> |
| 29 |
<li><a href="#status">Project Status</a></li> |
29 |
<li><a href="#status">Project Status</a></li> |
| 30 |
<li><a href="#history">Project History</a></li> |
30 |
<li><a href="#history">Project History</a></li> |
| 31 |
<li><a href="#plan">Project Plan (Historical)</a></li> |
31 |
<li><a href="#papers">Project Papers</a></li> |
| 32 |
<li><a href="#resources">Resources and Links</a></li> |
|
|
| 33 |
<li><a href="#subsystems">Subsystems and Staffing (Historical)</a></li> |
| 34 |
<li><a href="#tasks">Tasks (Historical)</a></li> |
| 35 |
<li><a href="#issues">Known Issues (Historical)</a></li> |
| 36 |
<li><a href="#news">News (Historical)</a></li> |
| 37 |
</ul> |
32 |
</ul> |
| 38 |
|
33 |
|
| 39 |
<a name="goal"></a> |
34 |
<a name="goal"></a> |
|
Lines 83-90
Link Here
|
| 83 |
As of FreeBSD 6.0, the SMPng Project is considered "complete", in that |
78 |
As of FreeBSD 6.0, the SMPng Project is considered "complete", in that |
| 84 |
the goal of moving to parallelized kernel operation has been met for |
79 |
the goal of moving to parallelized kernel operation has been met for |
| 85 |
most significant parts of the kernel. |
80 |
most significant parts of the kernel. |
| 86 |
This web page is largely historical, and is not up-to-date with respect |
81 |
</p> |
| 87 |
to the completion status of items in the task list, etc.</p> |
82 |
<p> |
|
|
83 |
Work is still ongoing to this day with major advances in performance |
| 84 |
being achieved. |
| 85 |
</p> |
| 88 |
|
86 |
|
| 89 |
<a name="history"></a> |
87 |
<a name="history"></a> |
| 90 |
<h2>Project History</h2> |
88 |
<h2>Project History</h2> |
|
Lines 113-119
Link Here
|
| 113 |
for most relevant code paths, and the FreeBSD 6.0 kernel introduced |
111 |
for most relevant code paths, and the FreeBSD 6.0 kernel introduced |
| 114 |
MPSAFE VFS, as well as wide-spread performance optimization.</p> |
112 |
MPSAFE VFS, as well as wide-spread performance optimization.</p> |
| 115 |
|
113 |
|
| 116 |
<p>Continuing work on SMPng in the 6.x branch has included work sweeping |
114 |
<p>Continued work on SMPng in the 6.x branch included work sweeping |
| 117 |
up the "loose ends" that remain under Giant, such as parts of NFS, and |
115 |
up the "loose ends" that remain under Giant, such as parts of NFS, and |
| 118 |
less commonly used network stack components.</p> |
116 |
less commonly used network stack components.</p> |
| 119 |
|
117 |
|
|
Lines 126-2801
Link Here
|
| 126 |
run-time kernel lock order verifier, and wide-spread use of lock |
124 |
run-time kernel lock order verifier, and wide-spread use of lock |
| 127 |
assertions and run-time invariants testing.</p> |
125 |
assertions and run-time invariants testing.</p> |
| 128 |
|
126 |
|
| 129 |
<a name="plan"></a> |
127 |
<a name="papers"></a> |
| 130 |
<h2>Project Plan (Historical)</h2> |
128 |
<h2>Project Papers</h2> |
| 131 |
|
|
|
| 132 |
<p>This web page contains information relating to the SMPng effort; |
| 133 |
because of the immense scope of the work and rapid pace of development, |
| 134 |
it captures only a subset of what has been done. As the base SMPng |
| 135 |
Project, moving towards fine-grained locking, is complete, this |
| 136 |
information is largely historical, and does not attempt to capture |
| 137 |
more recent work on locking and multi-processor performance.</p> |
| 138 |
|
| 139 |
<p>The task list below is not intended to be complete, but does |
| 140 |
represent a set of relevant and/or important components of the |
| 141 |
overall work. The "Responsible" field identifies a developer who |
| 142 |
has expressed willingness to be responsible for completing the |
| 143 |
identified task; this doesn't preclude others working on it, |
| 144 |
but suggests that coordination with the responsible party might |
| 145 |
be appropriate so as to avoid unnecessary duplication of work, |
| 146 |
and to maximize forward progress. If beginning work on a new |
| 147 |
area of substantial size, or one that appears unclaimed, it may |
| 148 |
be worth dropping an e-mail to the FreeBSD SMP mailing list to |
| 149 |
see if any progress has been made.</p> |
| 150 |
|
| 151 |
<p>The definition of the date field varies depending on the status |
| 152 |
of a task. For completed tasks, it refers to the date completed |
| 153 |
or reported completed. For in-progress tasks, it refers to the |
| 154 |
date of the last update of the entry. For stalled tasks, it |
| 155 |
refers to the date that the task was declared stalled. For |
| 156 |
new tasks, it refers to the date the task was added to the list.</p> |
| 157 |
|
| 158 |
<p>Locking down of individual device drivers is tracked at |
| 159 |
<a href="http://www.FreeBSD.org/projects/busdma/">the busdma and |
| 160 |
SMPng driver conversion web page</a>.</p> |
| 161 |
|
| 162 |
<p>Network stack locking information is available at &a.rwatson;'s |
| 163 |
<a href="http://www.watson.org/~robert/freebsd/netperf/">netperf web |
| 164 |
page</a> and the FreeBSD.org <a href="../projects/netperf/">Netperf |
| 165 |
web page</a>. |
| 166 |
An SMP <a href="../projects/netperf/cluster.html">network performance |
| 167 |
cluster</a> has also been created for the purposes of testing.</p> |
| 168 |
|
| 169 |
<p>Tasks are sorted first by status, then by date.</p> |
| 170 |
|
| 171 |
<a name="resources"></a> |
| 172 |
<h2>Resources and Links</h2> |
| 173 |
|
| 174 |
<ul> |
129 |
<ul> |
| 175 |
<li><p>Most SMP-related discussion takes place on the |
|
|
| 176 |
<a href="mailto:freebsd-smp@FreeBSD.org">freebsd-smp</a> |
| 177 |
mailing list. You can read more about mailing lists in the |
| 178 |
<a href="../doc/en_US.ISO8859-1/books/handbook/eresources.html#ERESOURCES-MAIL"> |
| 179 |
Resources on the Internet</a> appendix of the |
| 180 |
<a href="../doc/en_US.ISO8859-1/books/handbook/index.html">FreeBSD Handbook</a>.</p></li> |
| 181 |
|
| 182 |
<li><p>Steve Passe has been maintaining a |
| 183 |
<a href="http://people.FreeBSD.org/~fsmp/SMP/SMP.html">SMP |
| 184 |
project page</a> that contains additional information, and |
| 185 |
goes back further in time than this web page.</p></li> |
| 186 |
|
| 187 |
<li><p>&a.rwatson; is maintaining a page for SMP-related network |
| 188 |
performance work for the <a href="../projects/netperf/">Netperf |
| 189 |
project</a>. In addition, he has a <a |
| 190 |
href="http://www.watson.org/~robert/freebsd/netperf/">a personal |
| 191 |
web page with a change log and other information</a>.</p></li> |
| 192 |
|
| 193 |
<li><p><a href="http://www.osnews.com/">OSNews</a> has run an |
| 194 |
interview with FreeBSD developers &a.rwatson;, &a.jhb;, and |
| 195 |
&a.scottl;, who talk about many features in 6.x, including recent |
| 196 |
SMPng work, SMPVFS, and more. The <a |
| 197 |
href="http://www.osnews.com/story.php?news_id=10951">article is |
| 198 |
here</a>.</p></li> |
| 199 |
|
| 200 |
<li><p>A series of man pages on kernel synchronization and threading |
| 201 |
primitives can be found here:</p> |
| 202 |
|
| 203 |
<ul> |
| 204 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=mutex&apropos=0&sektion=9">Kernel Mutexes (mutex(9))</a></li> |
| 205 |
|
| 206 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=mtx_pool&apropos=0&sektion=9">Kernel Mutex Pools (mtx_pool(9))</a></li> |
| 207 |
|
| 208 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=critical&apropos=0&sektion=9">Kernel Critical Sections (critical(9))</a></li> |
| 209 |
|
| 210 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=sx&apropos=0&sektion=9">Kernel Shared/Exclusive Locks (sx(9))</a></li> |
| 211 |
|
| 212 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=condvar&apropos=0&sektion=9">Kernel Condition Variables (condvar(9))</a></li> |
| 213 |
|
| 214 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=sema&apropos=0&sektion=9">Kernel Semaphores (sema(9))</a></li> |
| 215 |
|
| 216 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=sleep&apropos=0&sektion=9">Kernel sleep/wakeup API (sleep(9))</a></li> |
| 217 |
|
| 218 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=kthread&apropos=0&sektion=9">Kernel Threads (kthread(9))</a></li> |
| 219 |
|
| 220 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=ithread&apropos=0&sektion=9">Interrupt Threads (ithread(9))</a></li> |
| 221 |
|
| 222 |
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=swi&apropos=0&sektion=9">Software Interrupt Threads (swi(9))</a></li> |
| 223 |
</ul> |
| 224 |
|
| 225 |
<br> |
| 226 |
<p>In addition, the old <a href="http://www.freebsd.org/cgi/man.cgi?query=spl&apropos=0&sektion=9">SPL man page</a> may be of interest, as it reflects the previous SMP synchronization model.</p> |
| 227 |
</li> |
| 228 |
|
| 229 |
<li><p> |
| 230 |
Hiten Pandya's <a |
| 231 |
href="http://storm.uk.FreeBSD.org/~hiten/smp_synch_rules.html"> |
| 232 |
SMP kernel synchronization rules</a>.</p></li> |
| 233 |
|
| 234 |
<li><p>&a.dillon; maintained a |
| 235 |
<a href="http://apollo.backplane.com/FreeBSDSmp/">web page</a> |
| 236 |
that documented the work he was doing on the SMP project.</p></li> |
| 237 |
|
| 238 |
<li> |
130 |
<li> |
| 239 |
<p> |
131 |
<p> |
| 240 |
<a href="http://www.usenix.org/events/bsdcon02/full_papers/baldwin/baldwin_html/index.html"> |
|
|
| 241 |
"Locking in the Multithreaded FreeBSD Kernel" USENIX paper by &a.jhb;</a>. |
| 242 |
</p> |
| 243 |
</li> |
| 244 |
|
| 245 |
<li> |
| 246 |
<p> |
| 247 |
<a href="http://www.usenix.org/publications/library/proceedings/bsdcon03/tech/hsu.html"> |
132 |
<a href="http://www.usenix.org/publications/library/proceedings/bsdcon03/tech/hsu.html"> |
| 248 |
"Reasoning about SMP in FreeBSD" BSDCon'03 paper by &a.hsu;</a>. |
133 |
"Reasoning about SMP in FreeBSD" BSDCon'03 paper by &a.hsu;</a>. |
| 249 |
</p> |
134 |
</p> |
| 250 |
</li> |
135 |
</li> |
| 251 |
|
|
|
| 252 |
<li> |
136 |
<li> |
| 253 |
<p> |
137 |
<p> |
| 254 |
<a href="http://www.usenix.org/publications/library/proceedings/bsdcon03/tech/roberson.html"> |
138 |
<a href="http://www.usenix.org/publications/library/proceedings/bsdcon03/tech/roberson.html"> |
| 255 |
"ULE: A Modern Scheduler for FreeBSD" BSDCon'03 paper by &a.jeff;</a>. |
139 |
"ULE: A Modern Scheduler for FreeBSD" BSDCon'03 paper by &a.jeff;</a>. |
| 256 |
</p> |
140 |
</p> |
| 257 |
</li> |
141 |
</li> |
| 258 |
|
|
|
| 259 |
<li> |
142 |
<li> |
| 260 |
<p> |
143 |
<p> |
| 261 |
<a href="http://www.watson.org/~robert/freebsd/netperf/20051027-eurobsdcon2005-netperf.pdf">"Introduction to Multithreading and Multiprocessing in the FreeBSD SMPng Network Stack" EuroBSDCon 2005 paper by &a.rwatson;</a>. |
144 |
<a href="http://www.watson.org/~robert/freebsd/netperf/20051027-eurobsdcon2005-netperf.pdf">"Introduction to Multithreading and Multiprocessing in the FreeBSD SMPng Network Stack" EuroBSDCon 2005 paper by &a.rwatson;</a>. |
| 262 |
</p> |
145 |
</p> |
| 263 |
</li> |
146 |
</li> |
| 264 |
|
|
|
| 265 |
</ul> |
| 266 |
|
| 267 |
<a name="subsystems"></a> |
| 268 |
<h2>Subsystems and Staffing (Historical)</h2> |
| 269 |
|
| 270 |
<p>This is an incomplete list of high-level kernel subsystems and |
| 271 |
current, active staff working on SMP architecture and stability.</p> |
| 272 |
|
| 273 |
<table class="tblbasic"> |
| 274 |
<tr class="heading"> |
| 275 |
<th> Subsystem </th> |
| 276 |
<th> Status </th> |
| 277 |
<th> Last updated </th> |
| 278 |
<th> Staffing </th> |
| 279 |
</tr> |
| 280 |
|
| 281 |
<tr> |
| 282 |
<td> Newbus </td> |
| 283 |
<td> &status.wip; </td> |
| 284 |
<td> 5 October 2003 </td> |
| 285 |
<td> &a.imp;</td> |
| 286 |
</tr> |
| 287 |
|
| 288 |
<tr> |
| 289 |
<td> VM </td> |
| 290 |
<td> &status.wip; </td> |
| 291 |
<td> 4 October 2003 </td> |
| 292 |
<td> &a.alc;</td> |
| 293 |
</tr> |
| 294 |
|
| 295 |
<tr> |
| 296 |
<td> Buffer cache </td> |
| 297 |
<td> &status.wip; </td> |
| 298 |
<td> 4 October 2003 </td> |
| 299 |
<td> &a.jeff;, &a.phk; </td> |
| 300 |
</tr> |
| 301 |
|
| 302 |
<tr> |
| 303 |
<td> VFS </td> |
| 304 |
<td> &status.done; </td> |
| 305 |
<td> 24 January 2005 </td> |
| 306 |
<td> &a.jeff; </td> |
| 307 |
</tr> |
| 308 |
|
| 309 |
<tr> |
| 310 |
<td> Processes and thread operations </td> |
| 311 |
<td> &status.wip; </td> |
| 312 |
<td> 5 May 2003 </td> |
| 313 |
<td> &a.jhb; </td> |
| 314 |
</tr> |
| 315 |
|
| 316 |
<tr> |
| 317 |
<td> Scheduler </td> |
| 318 |
<td> &status.done; </td> |
| 319 |
<td> 23 April 2004 </td> |
| 320 |
<td> &a.jhb;, &a.jeff; </td> |
| 321 |
</tr> |
| 322 |
|
| 323 |
<tr> |
| 324 |
<td> GEOM </td> |
| 325 |
<td> &status.done; </td> |
| 326 |
<td> 5 February 2003 </td> |
| 327 |
<td> &a.phk; </td> |
| 328 |
</tr> |
| 329 |
|
| 330 |
<tr> |
| 331 |
<td> File descriptors </td> |
| 332 |
<td> &status.done; </td> |
| 333 |
<td> 5 February 2003 </td> |
| 334 |
<td> &a.alfred;, &a.tanimura;, &a.rwatson; </td> |
| 335 |
</tr> |
| 336 |
|
| 337 |
<tr> |
| 338 |
<td> TTY subsystem </td> |
| 339 |
<td> &status.done; </td> |
| 340 |
<td> 20 August 2008 </td> |
| 341 |
<td> &a.ed; </td> |
| 342 |
</tr> |
| 343 |
|
| 344 |
<tr> |
| 345 |
<td> Pipe IPC </td> |
| 346 |
<td> &status.done; </td> |
| 347 |
<td> 4 October 2003 </td> |
| 348 |
<td> &a.alfred; </td> |
| 349 |
</tr> |
| 350 |
|
| 351 |
<tr> |
| 352 |
<td> Socket structures and system calls </td> |
| 353 |
<td> &status.done; </td> |
| 354 |
<td> 25 November 2004 </td> |
| 355 |
<td> &a.sam;, &a.rwatson; </td> |
| 356 |
</tr> |
| 357 |
|
| 358 |
<tr> |
| 359 |
<td> KQueue </td> |
| 360 |
<td> &status.done; </td> |
| 361 |
<td> 24 November 2004 </td> |
| 362 |
<td> &a.jmg;, &a.green; </td> |
| 363 |
</tr> |
| 364 |
|
| 365 |
<tr> |
| 366 |
<td> IPv4 </td> |
| 367 |
<td> &status.done; </td> |
| 368 |
<td> 23 April 2004 </td> |
| 369 |
<td> &a.jennifer;, &a.hsu;, &a.sam;, &a.rwatson;, &a.gnn; </td> |
| 370 |
</tr> |
| 371 |
|
| 372 |
<tr> |
| 373 |
<td> IPv6 </td> |
| 374 |
<td> &status.wip; </td> |
| 375 |
<td> 01 July 2005 </td> |
| 376 |
<td> &a.rwatson;, &a.gnn; </td> |
| 377 |
</tr> |
| 378 |
|
| 379 |
<tr> |
| 380 |
<td> IPX/SPX </td> |
| 381 |
<td> &status.done; </td> |
| 382 |
<td> 09 January 2005 </td> |
| 383 |
<td> &a.rwatson; </td> |
| 384 |
</tr> |
| 385 |
|
| 386 |
<tr> |
| 387 |
<td> netatalk </td> |
| 388 |
<td> &status.done; </td> |
| 389 |
<td> 02 February 2005 </td> |
| 390 |
<td> &a.rwatson; </td> |
| 391 |
</tr> |
| 392 |
|
| 393 |
<tr> |
| 394 |
<td> Network stack infrastructure </td> |
| 395 |
<td> &status.done; </td> |
| 396 |
<td> 28 August 2004 </td> |
| 397 |
<td> &a.hsu;, &a.sam;, &a.rwatson;, &a.mlaier;, &a.luigi;, Maurycy Pawlowski-Wieronski <maurycy@fouk.org>, &a.brooks;, |
| 398 |
&a.rik; </td> |
| 399 |
</tr> |
| 400 |
|
| 401 |
<tr> |
| 402 |
<td> NFS Client </td> |
| 403 |
<td> &status.done; </td> |
| 404 |
<td> 19 May 2006 </td> |
| 405 |
<td> &a.mohans; </td> |
| 406 |
</tr> |
| 407 |
|
| 408 |
<tr> |
| 409 |
<td> NFS Server </td> |
| 410 |
<td> &status.wip; </td> |
| 411 |
<td> 24 February 2007 </td> |
| 412 |
<td> &a.rwatson;, &a.jeff; </td> |
| 413 |
</tr> |
| 414 |
|
| 415 |
</table> |
| 416 |
|
| 417 |
<a name="tasks"></a> |
| 418 |
<h2>Tasks (Historical)</h2> |
| 419 |
|
| 420 |
<p>Following is an incomplete list of general tasks.</p> |
| 421 |
|
| 422 |
<table class="tblbasic"> |
| 423 |
<tr class="heading"> |
| 424 |
<th> Task </th> |
| 425 |
<th> Responsible </th> |
| 426 |
<th> Last updated </th> |
| 427 |
<th> Status </th> |
| 428 |
</tr> |
| 429 |
|
| 430 |
<tr> |
| 431 |
<td> Convert the giant lock from spinning to blocking, add the |
| 432 |
scheduler lock, add per-CPU idle processes. </td> |
| 433 |
<td> &a.dillon; </td> |
| 434 |
<td> 25 June 2000 </td> |
| 435 |
<td> &status.done; </td> |
| 436 |
</tr> |
| 437 |
|
| 438 |
<tr> |
| 439 |
<td> Port the BSD/OS locking primitives (i386). </td> |
| 440 |
<td> &a.jake; </td> |
| 441 |
<td> 3 July 2000 </td> |
| 442 |
<td> &status.done; </td> |
| 443 |
</tr> |
| 444 |
|
| 445 |
<tr> |
| 446 |
<td> Implement heavy-weight interrupt threads (i386). </td> |
| 447 |
<td> &a.grog; </td> |
| 448 |
<td> 3 August 2000 </td> |
| 449 |
<td> &status.done; </td> |
| 450 |
</tr> |
| 451 |
|
| 452 |
<tr> |
| 453 |
<td> Rewrite the low level interrupt code (i386 UP). </td> |
| 454 |
<td> &a.grog; </td> |
| 455 |
<td> 3 August 2000 </td> |
| 456 |
<td> &status.done; </td> |
| 457 |
</tr> |
| 458 |
|
| 459 |
<tr> |
| 460 |
<td> Demonstrated reasonable stability (self-hosted buildworld) |
| 461 |
(i386 UP). </td> |
| 462 |
<td> -smp developers </td> |
| 463 |
<td> 12 August 2000 </td> |
| 464 |
<td> &status.done; </td> |
| 465 |
</tr> |
| 466 |
|
| 467 |
<tr> |
| 468 |
<td> Port the BSD/OS locking primitives (alpha). </td> |
| 469 |
<td> &a.dfr; </td> |
| 470 |
<td> 24 August 2000 </td> |
| 471 |
<td> &status.done; </td> |
| 472 |
</tr> |
| 473 |
|
| 474 |
<tr> |
| 475 |
<td> Stub out (disable) spl()s. </td> |
| 476 |
<td> &a.grog; </td> |
| 477 |
<td> 30 August 2000 </td> |
| 478 |
<td> &status.done; </td> |
| 479 |
</tr> |
| 480 |
|
| 481 |
<tr> |
| 482 |
<td> Port the BSD/OS ktr code. </td> |
| 483 |
<td> &a.grog;, &a.jhb; </td> |
| 484 |
<td> 30 August 2000 </td> |
| 485 |
<td> &status.done; </td> |
| 486 |
</tr> |
| 487 |
|
| 488 |
<tr> |
| 489 |
<td> Rewrite the low level interrupt code (i386 SMP). </td> |
| 490 |
<td> &a.jhb; </td> |
| 491 |
<td> 1 September 2000 </td> |
| 492 |
<td> &status.done; </td> |
| 493 |
</tr> |
| 494 |
|
| 495 |
<tr> |
| 496 |
<td> Demonstrated reasonable stability (self-hosted buildworld) |
| 497 |
(i386 SMP). </td> |
| 498 |
<td> -smp developers </td> |
| 499 |
<td> 6 September 2000 </td> |
| 500 |
<td> &status.done; </td> |
| 501 |
</tr> |
| 502 |
|
| 503 |
<tr> |
| 504 |
<td> Demonstrated reasonable stability (self-hosted buildworld) |
| 505 |
(alpha). </td> |
| 506 |
<td> -smp developers </td> |
| 507 |
<td> 6 September 2000 </td> |
| 508 |
<td> &status.done; </td> |
| 509 |
</tr> |
| 510 |
|
| 511 |
<tr> |
| 512 |
<td> Make malloc and friends thread-safe. </td> |
| 513 |
<td> &a.jasone; </td> |
| 514 |
<td> 10 September 2000 </td> |
| 515 |
<td> &status.done; </td> |
| 516 |
</tr> |
| 517 |
|
| 518 |
<tr> |
| 519 |
<td> Implement msleep(), make tsleep() an msleep() wrapper. </td> |
| 520 |
<td> &a.jake; </td> |
| 521 |
<td> 11 September 2000 </td> |
| 522 |
<td> &status.done; </td> |
| 523 |
</tr> |
| 524 |
|
| 525 |
<tr> |
| 526 |
<td> Make fxp driver thread-safe. </td> |
| 527 |
<td> &a.cp; </td> |
| 528 |
<td> 17 September 2000 </td> |
| 529 |
<td> &status.done; </td> |
| 530 |
</tr> |
| 531 |
|
| 532 |
<tr> |
| 533 |
<td> Make mbuf's thread-safe. </td> |
| 534 |
<td> &a.bmilekic; </td> |
| 535 |
<td> 29 September 2000 </td> |
| 536 |
<td> &status.done; </td> |
| 537 |
</tr> |
| 538 |
|
| 539 |
<tr> |
| 540 |
<td> Lock manager re-work. </td> |
| 541 |
<td> &a.jasone; </td> |
| 542 |
<td> 3 October 2000 </td> |
| 543 |
<td> &status.done; </td> |
| 544 |
</tr> |
| 545 |
|
| 546 |
<tr> |
| 547 |
<td> Implement heavy-weight interrupt threads (alpha). </td> |
| 548 |
<td> &a.jhb;, &a.dfr; </td> |
| 549 |
<td> 5 October 2000 </td> |
| 550 |
<td> &status.done; </td> |
| 551 |
</tr> |
| 552 |
|
| 553 |
<tr> |
| 554 |
<td> Rewrite the low level interrupt code (alpha). </td> |
| 555 |
<td> &a.dfr;, &a.jhb; </td> |
| 556 |
<td> 5 October 2000 </td> |
| 557 |
<td> &status.done; </td> |
| 558 |
</tr> |
| 559 |
|
| 560 |
<tr> |
| 561 |
<td> Process accounting. </td> |
| 562 |
<td> &a.tegge;, &a.jhb; </td> |
| 563 |
<td> 5 October 2000 </td> |
| 564 |
<td> &status.done; </td> |
| 565 |
</tr> |
| 566 |
|
| 567 |
<tr> |
| 568 |
<td> Make ethernet drivers thread-safe. </td> |
| 569 |
<td> &a.wpaul; </td> |
| 570 |
<td> 15 October 2000 </td> |
| 571 |
<td> &status.done; </td> |
| 572 |
</tr> |
| 573 |
|
| 574 |
<tr> |
| 575 |
<td> Make the mutex headers mostly machine-independent. </td> |
| 576 |
<td> &a.jhb; </td> |
| 577 |
<td> 20 October 2000 </td> |
| 578 |
<td> &status.done; </td> |
| 579 |
</tr> |
| 580 |
|
| 581 |
<tr> |
| 582 |
<td> Rename SMP_DEBUG to MUTEX_DEBUG. </td> |
| 583 |
<td> &a.jhb; </td> |
| 584 |
<td> 20 October 2000 </td> |
| 585 |
<td> &status.done; </td> |
| 586 |
</tr> |
| 587 |
|
| 588 |
<tr> |
| 589 |
<td> Give each soft interrupt its own thread. </td> |
| 590 |
<td> &a.cp; </td> |
| 591 |
<td> 25 October 2000 </td> |
| 592 |
<td> &status.done; </td> |
| 593 |
</tr> |
| 594 |
|
| 595 |
<tr> |
| 596 |
<td> Make sf_bufs (sendfile(2)) thread-safe. </td> |
| 597 |
<td> &a.bmilekic; </td> |
| 598 |
<td> 5 November 2000 </td> |
| 599 |
<td> &status.done; </td> |
| 600 |
</tr> |
| 601 |
|
| 602 |
<tr> |
| 603 |
<td> Make the witness code work correctly. </td> |
| 604 |
<td> &a.jhb; </td> |
| 605 |
<td> 18 November 2000 </td> |
| 606 |
<td> &status.done; </td> |
| 607 |
</tr> |
| 608 |
|
| 609 |
<tr> |
| 610 |
<td> Split the ktr-specific code out of db_interface.c. </td> |
| 611 |
<td> &a.jhb; </td> |
| 612 |
<td> 15 December 2000 </td> |
| 613 |
<td> &status.done; </td> |
| 614 |
</tr> |
| 615 |
|
| 616 |
<tr> |
| 617 |
<td> Convert the sio driver to using a spin mutex. </td> |
| 618 |
<td> &a.jhb; </td> |
| 619 |
<td> 18 December 2000 </td> |
| 620 |
<td> &status.done; </td> |
| 621 |
</tr> |
| 622 |
|
| 623 |
<tr> |
| 624 |
<td> Implement condition variables. </td> |
| 625 |
<td> &a.jake;, &a.jasone; </td> |
| 626 |
<td> 15 January 2001 </td> |
| 627 |
<td> &status.done; </td> |
| 628 |
</tr> |
| 629 |
|
| 630 |
<tr> |
| 631 |
<td> Add a flag to mtx_init() (MTX_RECURSE) that denotes |
| 632 |
whether a mutex is allowed to recurse. </td> |
| 633 |
<td> &a.bmilekic; </td> |
| 634 |
<td> 19 January 2001 </td> |
| 635 |
<td> &status.done; </td> |
| 636 |
</tr> |
| 637 |
|
| 638 |
<tr> |
| 639 |
<td> Make the zone allocator thread-safe. </td> |
| 640 |
<td> &a.des; </td> |
| 641 |
<td> 21 January 2001 </td> |
| 642 |
<td> &status.done; </td> |
| 643 |
</tr> |
| 644 |
|
| 645 |
<tr> |
| 646 |
<td> Convert simplelocks to mutexes. </td> |
| 647 |
<td> &a.jasone; </td> |
| 648 |
<td> 24 January 2001 </td> |
| 649 |
<td> &status.done; </td> |
| 650 |
</tr> |
| 651 |
|
| 652 |
<tr> |
| 653 |
<td> Make kernel preemptive with respect to interrupts. </td> |
| 654 |
<td> &a.jake; </td> |
| 655 |
<td> 31 January 2001 </td> |
| 656 |
<td> &status.done; </td> |
| 657 |
</tr> |
| 658 |
|
| 659 |
<tr> |
| 660 |
<td> Cleanup of mutex API. </td> |
| 661 |
<td> &a.bmilekic; </td> |
| 662 |
<td> 8 February 2001 </td> |
| 663 |
<td> &status.done; </td> |
| 664 |
</tr> |
| 665 |
|
| 666 |
<tr> |
| 667 |
<td> Remove COM_LOCK. </td> |
| 668 |
<td> &a.markm; </td> |
| 669 |
<td> 11 February 2001 </td> |
| 670 |
<td> &status.done; </td> |
| 671 |
</tr> |
| 672 |
|
| 673 |
<tr> |
| 674 |
<td> Merge various scheduling classes into one run queue. |
| 675 |
Modify scheduler to support preemptable kernel. </td> |
| 676 |
<td> &a.jake; </td> |
| 677 |
<td> 11 February 2001 </td> |
| 678 |
<td> &status.done; </td> |
| 679 |
</tr> |
| 680 |
|
| 681 |
<tr> |
| 682 |
<td> Make priority propagation work correctly. </td> |
| 683 |
<td> &a.jake; </td> |
| 684 |
<td> 11 February 2001 </td> |
| 685 |
<td> &status.done; </td> |
| 686 |
</tr> |
| 687 |
|
| 688 |
<tr> |
| 689 |
<td> Make most of the interrupt thread code MI and shared |
| 690 |
between hardware and software interrupts. </td> |
| 691 |
<td> &a.jhb; </td> |
| 692 |
<td> 18 February 2001 </td> |
| 693 |
<td> &status.done; </td> |
| 694 |
</tr> |
| 695 |
|
| 696 |
<tr> |
| 697 |
<td> Add protection to struct jail and jail-related functionality. </td> |
| 698 |
<td> &a.rwatson; </td> |
| 699 |
<td> 20 February 2001 </td> |
| 700 |
<td> &status.done; </td> |
| 701 |
</tr> |
| 702 |
|
| 703 |
<tr> |
| 704 |
<td> Implement sx (shared/exclusive) locks. </td> |
| 705 |
<td> &a.jasone; </td> |
| 706 |
<td> 5 March 2001 </td> |
| 707 |
<td> &status.done; </td> |
| 708 |
</tr> |
| 709 |
|
| 710 |
<tr> |
| 711 |
<td> Generalize/improve witness to handle more complex locking |
| 712 |
primitives (mtx, sx). </td> |
| 713 |
<td> &a.jhb; </td> |
| 714 |
<td> 28 March 2001 </td> |
| 715 |
<td> &status.done; </td> |
| 716 |
</tr> |
| 717 |
|
| 718 |
<tr> |
| 719 |
<td> Convert the allproc and proctree locks from lockmgr locks |
| 720 |
to sx locks. </td> |
| 721 |
<td> &a.jhb; </td> |
| 722 |
<td> 28 March 2001 </td> |
| 723 |
<td> &status.done; </td> |
| 724 |
</tr> |
| 725 |
|
| 726 |
<tr> |
| 727 |
<td> Make mbuf system use condition variables instead of |
| 728 |
msleep()/wakeup(). </td> |
| 729 |
<td> &a.bmilekic; </td> |
| 730 |
<td> 2 April 2001 </td> |
| 731 |
<td> &status.done; </td> |
| 732 |
</tr> |
| 733 |
|
| 734 |
<tr> |
| 735 |
<td> Remove <sys/mutex.h> includes from other kernel |
| 736 |
headers such as <vm/vm_zone.h>, <sys/resourcevar.h>, |
| 737 |
<sys/ucred.h>, and <sys/mbuf.h>. </td> |
| 738 |
<td> &a.markm; </td> |
| 739 |
<td> 15 May 2001 </td> |
| 740 |
<td> &status.done; </td> |
| 741 |
</tr> |
| 742 |
|
| 743 |
<tr> |
| 744 |
<td> Cleanup the various mp_machdep.c's, unify various SMP |
| 745 |
API's such as IPI delivery, etc. </td> |
| 746 |
<td> &a.jhb; </td> |
| 747 |
<td> 15 May 2001 </td> |
| 748 |
<td> &status.done; </td> |
| 749 |
</tr> |
| 750 |
|
| 751 |
<tr> |
| 752 |
<td> Make most of the forward_* and forwarded_* functions MI. </td> |
| 753 |
<td> &a.jhb; </td> |
| 754 |
<td> 15 May 2001 </td> |
| 755 |
<td> &status.done; </td> |
| 756 |
</tr> |
| 757 |
|
| 758 |
<tr> |
| 759 |
<td> Complete the MD support for SMP on the Alpha platform.</td> |
| 760 |
<td> &a.gallatin;, &a.dfr;, &a.jhb; </td> |
| 761 |
<td> 15 May 2001 </td> |
| 762 |
<td> &status.done; </td> |
| 763 |
</tr> |
| 764 |
|
| 765 |
<tr> |
| 766 |
<td> Convert select() to use condition variables. </td> |
| 767 |
<td> &a.tanimura; </td> |
| 768 |
<td> 15 May 2001 </td> |
| 769 |
<td> &status.done; </td> |
| 770 |
</tr> |
| 771 |
|
| 772 |
<tr> |
| 773 |
<td> Add a "giant" lock around the VM subsystem. </td> |
| 774 |
<td> &a.alfred; </td> |
| 775 |
<td> 13 June 2001 </td> |
| 776 |
<td> &status.done; </td> |
| 777 |
</tr> |
| 778 |
|
| 779 |
<tr> |
| 780 |
<td> Introduce a modified slab allocator for the mbuf subsystem. </td> |
| 781 |
<td> &a.bmilekic; </td> |
| 782 |
<td> 21 June 2001 </td> |
| 783 |
<td> &status.done; </td> |
| 784 |
</tr> |
| 785 |
|
| 786 |
<tr> |
| 787 |
<td> Add a witness_assert() function to handle lock assertions. </td> |
| 788 |
<td> &a.jhb; </td> |
| 789 |
<td> 27 June 2001 </td> |
| 790 |
<td> &status.done; </td> |
| 791 |
</tr> |
| 792 |
|
| 793 |
<tr> |
| 794 |
<td> Extend sx locks to support try lock operations. </td> |
| 795 |
<td> &a.jhb; </td> |
| 796 |
<td> 27 June 2001 </td> |
| 797 |
<td> &status.done; </td> |
| 798 |
</tr> |
| 799 |
|
| 800 |
<tr> |
| 801 |
<td> Document KTR. </td> |
| 802 |
<td> &a.jhb; </td> |
| 803 |
<td> 28 June 2001 </td> |
| 804 |
<td> &status.done; </td> |
| 805 |
</tr> |
| 806 |
|
| 807 |
<tr> |
| 808 |
<td> Make fork_return, fork_exit, ast, and userret MI. </td> |
| 809 |
<td> &a.jhb; </td> |
| 810 |
<td> 29 June 2001 </td> |
| 811 |
<td> &status.done; </td> |
| 812 |
</tr> |
| 813 |
|
| 814 |
<tr> |
| 815 |
<td> Make sched_lock's savecrit a per-process property saved |
| 816 |
and restored in mi_switch and initialized in fork_exit. </td> |
| 817 |
<td> &a.jhb; </td> |
| 818 |
<td> 30 June 2001 </td> |
| 819 |
<td> &status.done; </td> |
| 820 |
</tr> |
| 821 |
|
| 822 |
<tr> |
| 823 |
<td> Make ast() loop. </td> |
| 824 |
<td> &a.jhb; </td> |
| 825 |
<td> 10 August 2001 </td> |
| 826 |
<td> &status.done; </td> |
| 827 |
</tr> |
| 828 |
|
| 829 |
<tr> |
| 830 |
<td> Add upgrade/downgrade sx lock operations. </td> |
| 831 |
<td> &a.kan;, &a.jasone; </td> |
| 832 |
<td> 13 August 2001 </td> |
| 833 |
<td> &status.done; </td> |
| 834 |
</tr> |
| 835 |
|
| 836 |
<tr> |
| 837 |
<td> Implement semaphores. </td> |
| 838 |
<td> &a.jasone; </td> |
| 839 |
<td> 14 August 2001 </td> |
| 840 |
<td> &status.done; </td> |
| 841 |
</tr> |
| 842 |
|
| 843 |
<tr> |
| 844 |
<td> Add support for upgrade/downgrades in witness. </td> |
| 845 |
<td> &a.jhb; </td> |
| 846 |
<td> 23 August 2001 </td> |
| 847 |
<td> &status.done; </td> |
| 848 |
</tr> |
| 849 |
|
| 850 |
<tr> |
| 851 |
<td> Make most of cpu_wait() and cpu_exit() MI. </td> |
| 852 |
<td> &a.peter; </td> |
| 853 |
<td> 9 September 2001 </td> |
| 854 |
<td> &status.done; </td> |
| 855 |
</tr> |
| 856 |
|
| 857 |
<tr> |
| 858 |
<td> Split NFS into client and server. </td> |
| 859 |
<td> &a.peter; </td> |
| 860 |
<td> 18 Oct 2001 </td> |
| 861 |
<td> &status.done; </td> |
| 862 |
</tr> |
| 863 |
|
| 864 |
<tr> |
| 865 |
<td> Lock taskqueues. </td> |
| 866 |
<td> &a.arr;, &a.jhb; </td> |
| 867 |
<td> 25 October 2001 </td> |
| 868 |
<td> &status.done; </td> |
| 869 |
</tr> |
| 870 |
|
| 871 |
<tr> |
| 872 |
<td> Add a per-thread ucred reference. </td> |
| 873 |
<td> &a.jhb; </td> |
| 874 |
<td> 25 October 2001 </td> |
| 875 |
<td> &status.done; </td> |
| 876 |
</tr> |
| 877 |
|
| 878 |
<tr> |
| 879 |
<td> Make most of the per-CPU stuff MI. </td> |
| 880 |
<td> &a.jhb; </td> |
| 881 |
<td> 11 December 2001 </td> |
| 882 |
<td> &status.done; </td> |
| 883 |
</tr> |
| 884 |
|
| 885 |
<tr> |
| 886 |
<td> Make critical section saved state per-thread instead of per-lock |
| 887 |
so that interlocking spin locks work properly. </td> |
| 888 |
<td> &a.jhb; </td> |
| 889 |
<td> 17 December 2001 </td> |
| 890 |
<td> &status.done; </td> |
| 891 |
</tr> |
| 892 |
|
| 893 |
<tr> |
| 894 |
<td> Replace the APIC-specific imen_mtx with a MI-named icu_lock to |
| 895 |
protect interrupt controllers and associated data within the kernel |
| 896 |
for both i386 and alpha. </td> |
| 897 |
<td> &a.jhb; </td> |
| 898 |
<td> 20 December 2001 </td> |
| 899 |
<td> &status.done; </td> |
| 900 |
</tr> |
| 901 |
|
| 902 |
<tr> |
| 903 |
<td> Use the per-thread critical section nesting level in the mutex |
| 904 |
and interrupt thread code to automatically determine when to not |
| 905 |
preempt. This makes the MTX_NOSWITCH, SWI_SWITCH, and |
| 906 |
SWI_NOSWITCH flags obsolete as the kernel will be able to figure |
| 907 |
out the proper behavior on its own. </td> |
| 908 |
<td> &a.jhb; </td> |
| 909 |
<td> 5 January 2002 </td> |
| 910 |
<td> &status.done; </td> |
| 911 |
</tr> |
| 912 |
|
| 913 |
<tr> |
| 914 |
<td> Lock struct filedesc and struct file. </td> |
| 915 |
<td> &a.tanimura;, &a.alfred; </td> |
| 916 |
<td> 12 January 2002 </td> |
| 917 |
<td> &status.done; </td> |
| 918 |
</tr> |
| 919 |
|
| 920 |
<tr> |
| 921 |
<td> Lock struct pgrp, struct session, and struct sigio. </td> |
| 922 |
<td> &a.tanimura; </td> |
| 923 |
<td> 23 February 2002 </td> |
| 924 |
<td> &status.done; </td> |
| 925 |
</tr> |
| 926 |
|
| 927 |
<tr> |
| 928 |
<td> Lock pipe implementation, but not sigio/fown, VM interactions. </td> |
| 929 |
<td> &a.alfred; </td> |
| 930 |
<td> 27 February 2002 </td> |
| 931 |
<td> &status.done; </td> |
| 932 |
</tr> |
| 933 |
|
| 934 |
<tr> |
| 935 |
<td> Move to explicit reference counting for soft vnode references. |
| 936 |
</td> |
| 937 |
<td> &a.phk; </td> |
| 938 |
<td> 8 March 2002 </td> |
| 939 |
<td> &status.done; </td> |
| 940 |
</tr> |
| 941 |
|
| 942 |
<tr> |
| 943 |
<td> Initialize mutex pools early enough that sx locks can be used |
| 944 |
for VM. </td> |
| 945 |
<td> &a.green; </td> |
| 946 |
<td> 14 March 2002 </td> |
| 947 |
<td> &status.done; </td> |
| 948 |
</tr> |
| 949 |
|
| 950 |
<tr> |
| 951 |
<td> Place a global lock (sellock) around selinfo structures to fix a |
| 952 |
variety of lock order reversals, and make select() MP-safe. </td> |
| 953 |
<td> &a.alfred;, &a.davidc; </td> |
| 954 |
<td> 14 March 2002 </td> |
| 955 |
<td> &status.done; </td> |
| 956 |
</tr> |
| 957 |
|
| 958 |
<tr> |
| 959 |
<td> Push down Giant on read, write, pread, pwrite system calls, |
| 960 |
acquiring Giant in the per-subsystem fileop layer for sockets, VFS, |
| 961 |
etc.</td> |
| 962 |
<td> &a.alfred; </td> |
| 963 |
<td> 15 March 2002 </td> |
| 964 |
<td> &status.done; </td> |
| 965 |
</tr> |
| 966 |
|
| 967 |
<tr> |
| 968 |
<td> Lock down kernel module structures. </td> |
| 969 |
<td> &a.arr; </td> |
| 970 |
<td> 18 March 2002 </td> |
| 971 |
<td> &status.done; </td> |
| 972 |
</tr> |
| 973 |
|
| 974 |
<tr> |
| 975 |
<td> Lock down kernel linker globals.</td> |
| 976 |
<td> &a.arr; </td> |
| 977 |
<td> 18 March 2002 </td> |
| 978 |
<td> &status.done; </td> |
| 979 |
</tr> |
| 980 |
|
| 981 |
<tr> |
| 982 |
<td> Rewrite kernel memory allocator to be a slab allocator that |
| 983 |
uses per-cpu caches. </td> |
| 984 |
<td> &a.jeff; </td> |
| 985 |
<td> 21 March 2002 </td> |
| 986 |
<td> &status.done; </td> |
| 987 |
</tr> |
| 988 |
|
| 989 |
<tr> |
| 990 |
<td> Replace incorrect use of MD critical section API to disable |
| 991 |
interrupts with a specific interrupt disable API. </td> |
| 992 |
<td> &a.imp;, &a.dfr;, &a.benno;, &a.jhb; </td> |
| 993 |
<td> 21 March 2002 </td> |
| 994 |
<td> &status.done; </td> |
| 995 |
</tr> |
| 996 |
|
| 997 |
<tr> |
| 998 |
<td> Lock down access to the shared p_args "process arguments" |
| 999 |
structure through appropriate protection of that structure and |
| 1000 |
references to it. </td> |
| 1001 |
<td> &a.mini; </td> |
| 1002 |
<td> 31 March 2002 </td> |
| 1003 |
<td> &status.done; </td> |
| 1004 |
</tr> |
| 1005 |
|
| 1006 |
<tr> |
| 1007 |
<td> Move from flags/tsleep lock to sx locks to protect sysctl |
| 1008 |
tree from updates during sysctl operations. </td> |
| 1009 |
<td> &a.mini; </td> |
| 1010 |
<td> 1 April 2002 </td> |
| 1011 |
<td> &status.done; </td> |
| 1012 |
</tr> |
| 1013 |
|
| 1014 |
<tr> |
| 1015 |
<td> Create/port userland tool to manage KTR event dumps. </td> |
| 1016 |
<td> &a.jake; </td> |
| 1017 |
<td> 1 April 2002 </td> |
| 1018 |
<td> &status.done; </td> |
| 1019 |
</tr> |
| 1020 |
|
| 1021 |
<tr> |
| 1022 |
<td> Create MTX_SYSINIT and SX_SYSINIT macros that allow for |
| 1023 |
initializing locks that are subsystem independent. </td> |
| 1024 |
<td> &a.arr; </td> |
| 1025 |
<td> 2 April 2002 </td> |
| 1026 |
<td> &status.done; </td> |
| 1027 |
</tr> |
| 1028 |
|
| 1029 |
<tr> |
| 1030 |
<td> Lock down the global securelevel variable. </td> |
| 1031 |
<td> &a.arr; </td> |
| 1032 |
<td> 2 April 2002 </td> |
| 1033 |
<td> &status.done; </td> |
| 1034 |
</tr> |
| 1035 |
|
| 1036 |
<tr> |
| 1037 |
<td> Make grow_stack() MI. Possibly even a macro or inline. </td> |
| 1038 |
<td> &a.alc; </td> |
| 1039 |
<td> 6 April 2002 </td> |
| 1040 |
<td> &status.done; </td> |
| 1041 |
</tr> |
| 1042 |
|
| 1043 |
<tr> |
| 1044 |
<td> Lock use of p_fd, which otherwise can result in corrupted |
| 1045 |
p_fd panics during heavy operation. Start with a global, |
| 1046 |
and move to per-proc locking. </td> |
| 1047 |
<td> &a.alfred;, &a.tanimura; </td> |
| 1048 |
<td> 8 April 2002 </td> |
| 1049 |
<td> &status.done; </td> |
| 1050 |
</tr> |
| 1051 |
|
| 1052 |
<tr> |
| 1053 |
<td> Lock struct pargs. </td> |
| 1054 |
<td> &a.mini; </td> |
| 1055 |
<td> 9 April 2002 </td> |
| 1056 |
<td> &status.done; </td> |
| 1057 |
</tr> |
| 1058 |
|
| 1059 |
<tr> |
| 1060 |
<td> Lock sysctl hierarchy. </td> |
| 1061 |
<td> &a.mini; </td> |
| 1062 |
<td> 9 April 2002 </td> |
| 1063 |
<td> &status.done; </td> |
| 1064 |
</tr> |
| 1065 |
|
| 1066 |
<tr> |
| 1067 |
<td> Make {o,}sigreturn() MPSAFE. </td> |
| 1068 |
<td> &a.alc; </td> |
| 1069 |
<td> 11 April 2002 </td> |
| 1070 |
<td> &status.done; </td> |
| 1071 |
</tr> |
| 1072 |
|
| 1073 |
<tr> |
| 1074 |
<td> Rewrite kernel memory allocator so that Giant is not required |
| 1075 |
for malloc() or free(). </td> |
| 1076 |
<td> &a.jeff; </td> |
| 1077 |
<td> 2 May 2002 </td> |
| 1078 |
<td> &status.done; </td> |
| 1079 |
</tr> |
| 1080 |
|
| 1081 |
<tr> |
| 1082 |
<td> Replace complex shared/exclusive locking scheme in the VM |
| 1083 |
system with a purely exclusive lockmgr locking scheme, simplifying |
| 1084 |
locking and removing potential livelock/deadlock scenarios. </td> |
| 1085 |
<td> &a.green;, &a.alc; </td> |
| 1086 |
<td> 3 May 2002 </td> |
| 1087 |
<td> &status.done; </td> |
| 1088 |
</tr> |
| 1089 |
|
| 1090 |
<tr> |
| 1091 |
<td> Push down Giant into readv/writev system calls in style of |
| 1092 |
read/write/pread/pwrite once malloc no longer requires Giant in |
| 1093 |
the handling of iovec structures for uio. </td> |
| 1094 |
<td> &a.alc; </td> |
| 1095 |
<td> 9 May 2002 </td> |
| 1096 |
<td> &status.done; </td> |
| 1097 |
</tr> |
| 1098 |
|
| 1099 |
<tr> |
| 1100 |
<td> Push down Giant in mprotect(), minherit(), and madvise() so |
| 1101 |
that it is no longer acquired and released directly. </td> |
| 1102 |
<td> &a.alc; </td> |
| 1103 |
<td> 18 May 2002 </td> |
| 1104 |
<td> &status.done; </td> |
| 1105 |
</tr> |
| 1106 |
|
| 1107 |
<tr> |
| 1108 |
<td> Update suser() and p_can*() APIs to accept threads instead of |
| 1109 |
processes. </td> |
| 1110 |
<td> &a.jhb; </td> |
| 1111 |
<td> 18 May 2002 </td> |
| 1112 |
<td> &status.done; </td> |
| 1113 |
</tr> |
| 1114 |
|
| 1115 |
<tr> |
| 1116 |
<td> Broadly transition to td_ucred from p_ucred once KSE dependencies |
| 1117 |
are in place. </td> |
| 1118 |
<td> &a.jhb; </td> |
| 1119 |
<td> 18 May 2002 </td> |
| 1120 |
<td> &status.done; </td> |
| 1121 |
</tr> |
| 1122 |
|
| 1123 |
<tr> |
| 1124 |
<td> Add a witness_sleep() check to uma_zalloc() to catch code |
| 1125 |
calling malloc() or uma_zalloc() while holding non-sleepable |
| 1126 |
locks. </td> |
| 1127 |
<td> &a.jhb; </td> |
| 1128 |
<td> 20 May 2002 </td> |
| 1129 |
<td> &status.done; </td> |
| 1130 |
</tr> |
| 1131 |
|
| 1132 |
<tr> |
| 1133 |
<td> Optimize UP support by changing spin locks to only perform |
| 1134 |
critical section enter and exits. </td> |
| 1135 |
<td> &a.jhb; </td> |
| 1136 |
<td> 21 May 2002 </td> |
| 1137 |
<td> &status.done; </td> |
| 1138 |
</tr> |
| 1139 |
|
| 1140 |
<tr> |
| 1141 |
<td> Make sleep mutexes spin if the current lock holder is |
| 1142 |
executing on another CPU. </td> |
| 1143 |
<td> &a.jhb; </td> |
| 1144 |
<td> 21 May 2002 </td> |
| 1145 |
<td> &status.done; </td> |
| 1146 |
</tr> |
| 1147 |
|
| 1148 |
<tr> |
| 1149 |
<td> Add support for the IA32 pause instruction to spin loops in |
| 1150 |
locks. </td> |
| 1151 |
<td> &a.jhb; </td> |
| 1152 |
<td> 21 May 2002 </td> |
| 1153 |
<td> &status.done; </td> |
| 1154 |
</tr> |
| 1155 |
|
| 1156 |
<tr> |
| 1157 |
<td> Make KTRACE write into tracefiles asynchronously. </td> |
| 1158 |
<td> &a.jhb; </td> |
| 1159 |
<td> 7 June 2002 </td> |
| 1160 |
<td> &status.done; </td> |
| 1161 |
</tr> |
| 1162 |
|
| 1163 |
<tr> |
| 1164 |
<td> Remove Giant from modnext(2), modfnext(2), |
| 1165 |
modstat(2),and modfind(2).</td> |
| 1166 |
<td> &a.arr; </td> |
| 1167 |
<td> 25 June 2002 </td> |
| 1168 |
<td> &status.done; </td> |
| 1169 |
</tr> |
| 1170 |
|
| 1171 |
<tr> |
| 1172 |
<td> Fix synchronization of TLB flushes and invlpg() on x86 SMP. </td> |
| 1173 |
<td> &a.peter; </td> |
| 1174 |
<td> 12 July 2002 </td> |
| 1175 |
<td> &status.done; </td> |
| 1176 |
</tr> |
| 1177 |
|
| 1178 |
<tr> |
| 1179 |
<td> Add KTR(9) tracing for mutex contention. </td> |
| 1180 |
<td> &a.iedowse; </td> |
| 1181 |
<td> 26 Augist 2002 </td> |
| 1182 |
<td> &status.done; </td> |
| 1183 |
</tr> |
| 1184 |
|
| 1185 |
<tr> |
| 1186 |
<td> Make cpu_coredump MI. </td> |
| 1187 |
<td> &a.peter; </td> |
| 1188 |
<td> 7 September 2002 </td> |
| 1189 |
<td> &status.done; </td> |
| 1190 |
</tr> |
| 1191 |
|
| 1192 |
<tr> |
| 1193 |
<td> Add a subsystem lock to the accounting code. </td> |
| 1194 |
<td> &a.arr; </td> |
| 1195 |
<td> 11 September 2002 </td> |
| 1196 |
<td> &status.done; </td> |
| 1197 |
</tr> |
| 1198 |
|
| 1199 |
<tr> |
| 1200 |
<td> Allow KTR(9) to write trace records to alq(9) record facility. |
| 1201 |
</td> |
| 1202 |
<td> &a.jeff; </td> |
| 1203 |
<td> 22 September 2002 </tD> |
| 1204 |
<td> &status.done; </td> |
| 1205 |
</tr> |
| 1206 |
|
| 1207 |
<tr> |
| 1208 |
<td> Create mechanism in cdevsw structure to protect |
| 1209 |
thread-unsafe drivers. </td> |
| 1210 |
<td> &a.phk; </td> |
| 1211 |
<td> 27 September 2002 </td> |
| 1212 |
<td> &status.done; </td> |
| 1213 |
</tr> |
| 1214 |
|
| 1215 |
<tr> |
| 1216 |
<td> Fix SIGXPU and other #if 0'd things in mi_switch(). </td> |
| 1217 |
<td> &a.jhb; </td> |
| 1218 |
<td> 30 September 2002 </td> |
| 1219 |
<td> &status.done; </td> |
| 1220 |
</tr> |
| 1221 |
|
| 1222 |
<tr> |
| 1223 |
<td> Lock down TrustedBSD MAC implementation. </td> |
| 1224 |
<td> &a.rwatson; </td> |
| 1225 |
<td> 11 November 2002 </td> |
| 1226 |
<td> &status.done; </td> |
| 1227 |
</tr> |
| 1228 |
|
| 1229 |
<tr> |
| 1230 |
<td> Lock eventhandlers. </td> |
| 1231 |
<td> &a.msmith;, &a.mini;, &a.jhb; </td> |
| 1232 |
<td> 11 March 2003 </td> |
| 1233 |
<td> &status.done; </td> |
| 1234 |
</tr> |
| 1235 |
|
| 1236 |
<tr> |
| 1237 |
<td> Fix PHOLD() so that it blocks to guarantee PS_INMEM. </td> |
| 1238 |
<td> &a.jhb; </td> |
| 1239 |
<td> 22 April 2003 </td> |
| 1240 |
<td> &status.done; </td> |
| 1241 |
</tr> |
| 1242 |
|
| 1243 |
<tr> |
| 1244 |
<td> Fix various procfs_machdep.c to not use sched_lock. </td> |
| 1245 |
<td> &a.jhb; </td> |
| 1246 |
<td> 22 April 2003 </td> |
| 1247 |
<td> &status.done; </td> |
| 1248 |
</tr> |
| 1249 |
|
| 1250 |
<tr> |
| 1251 |
<td> Lock all references to process credentials and remove Giant |
| 1252 |
from process credential-related system calls. </td> |
| 1253 |
<td> &a.jhb; </td> |
| 1254 |
<td> 1 May 2003 </td> |
| 1255 |
<td> &status.done; </td> |
| 1256 |
</tr> |
| 1257 |
|
| 1258 |
<tr> |
| 1259 |
<td> Merge the procsig and sigacts structures, move the new sigacts |
| 1260 |
structure out of the U-area and add appropriate locking. </td> |
| 1261 |
<td> &a.jhb; </td> |
| 1262 |
<td> 13 May 2003 </td> |
| 1263 |
<td> &status.done; </td> |
| 1264 |
</tr> |
| 1265 |
|
| 1266 |
<tr> |
| 1267 |
<td> Remove Giant from the kill() and killpg() system calls. </td> |
| 1268 |
<td> &a.jhb; </td> |
| 1269 |
<td> 13 May 2003 </td> |
| 1270 |
<td> &status.done; </td> |
| 1271 |
</tr> |
| 1272 |
|
| 1273 |
<tr> |
| 1274 |
<td> Enhance the mutex pool implementation to allow creation and |
| 1275 |
use of multiple, dynamically allocated pools with adjustable pool |
| 1276 |
sizes and mutex options. </td> |
| 1277 |
<td> &a.truckman; </td> |
| 1278 |
<td> 16 July 2003 </td> |
| 1279 |
<td> &status.done; </td> |
| 1280 |
</tr> |
| 1281 |
|
| 1282 |
<tr> |
| 1283 |
<td> Create mutex profiling tool for the kernel so as to measure |
| 1284 |
contention and behavior of kernel mutexes. </td> |
| 1285 |
<td> &a.eivind;, &a.des; </td> |
| 1286 |
<td> 31 March 2002 </td> |
| 1287 |
<td> &status.done; </td> |
| 1288 |
</tr> |
| 1289 |
|
| 1290 |
<tr> |
| 1291 |
<td> Lock down linker_file_t structures in the kernel linker.</td> |
| 1292 |
<td> &a.arr; </td> |
| 1293 |
<td> 19 June 2002 </td> |
| 1294 |
<td> &status.done; </td> |
| 1295 |
</tr> |
| 1296 |
|
| 1297 |
<tr> |
| 1298 |
<td> Lock pipe implementation: VM optimizations. </td> |
| 1299 |
<td> </td> |
| 1300 |
<td> 4 October 2003 </td> |
| 1301 |
<td> &status.done; </td> |
| 1302 |
</tr> |
| 1303 |
|
| 1304 |
<tr> |
| 1305 |
<td> Reimplement i386 interrupt and SMP code so that SMP kernels |
| 1306 |
work on UP boxes and SMP can be enabled in GENERIC. </td> |
| 1307 |
<td> &a.jhb; </td> |
| 1308 |
<td> 3 November 2003 </td> |
| 1309 |
<td> &status.done; </td> |
| 1310 |
</tr> |
| 1311 |
|
| 1312 |
<tr> |
| 1313 |
<td> Implement generic turnstiles to use when blocking on non-sleepable |
| 1314 |
locks. </td> |
| 1315 |
<td> &a.jhb; </td> |
| 1316 |
<td> 11 November 2003 </td> |
| 1317 |
<td> &status.done; </td> |
| 1318 |
</tr> |
| 1319 |
|
| 1320 |
<tr> |
| 1321 |
<td> Split witness_lock() into witness_checkorder() and |
| 1322 |
witness_lock(). witness_checkorder() would be called before |
| 1323 |
acquiring a lock to increase the changes of detecting and warning |
| 1324 |
about a reversal prior to deadlocking. witness_lock() would |
| 1325 |
simply update witness' internal state to note that a lock has |
| 1326 |
been acquired. </td> |
| 1327 |
<td> &a.jhb; </td> |
| 1328 |
<td> 24 January 2004 </td> |
| 1329 |
<td> &status.done; </td> |
| 1330 |
</tr> |
| 1331 |
|
| 1332 |
<tr> |
| 1333 |
<td> Lock per-process resource limits. </td> |
| 1334 |
<td> &a.mtm;, &a.jhb; </td> |
| 1335 |
<td> 4 February 2004 </td> |
| 1336 |
<td> &status.done; </td> |
| 1337 |
</tr> |
| 1338 |
|
| 1339 |
<tr> |
| 1340 |
<td> Implement a sleep queue abstraction to be used by both msleep() |
| 1341 |
and condition variables. This new abstraction should use a hash |
| 1342 |
table of sleep queues with a spin lock on each sleep queue chain |
| 1343 |
similar to turnstile chain locks to make sched_lock finger |
| 1344 |
grained.</td> |
| 1345 |
<td> &a.jhb; </td> |
| 1346 |
<td> 27 February 2004 </td> |
| 1347 |
<td> &status.done; </td> |
| 1348 |
</tr> |
| 1349 |
|
| 1350 |
<tr> |
| 1351 |
<td> Remove Giant from jail(2). </td> |
| 1352 |
<td> &a.arr;, &a.rwatson; </td> |
| 1353 |
<td> 23 April 2004 </td> |
| 1354 |
<td> &status.done; </td> |
| 1355 |
</tr> |
| 1356 |
|
| 1357 |
<tr> |
| 1358 |
<td> Add subsystem locking to NFSv2, NFSv3 server, permitting upcalls |
| 1359 |
and other network-related elements to run Giant-free. </td> |
| 1360 |
<td> &a.rwatson; </td> |
| 1361 |
<td> 24 July 2004 </td> |
| 1362 |
<td> &status.done; </td> |
| 1363 |
</tr> |
| 1364 |
|
| 1365 |
<tr> |
| 1366 |
<td> Add KTR(9) tracing for UMA allocation/free events. </td> |
| 1367 |
<td> &a.rwatson; </td> |
| 1368 |
<td> 05 August 2004 </td> |
| 1369 |
<td> &status.done; </td> |
| 1370 |
</tr> |
| 1371 |
|
| 1372 |
<tr> |
| 1373 |
<td> Add KTR(9) tracing for GEOM I/O events. </td> |
| 1374 |
<td> &a.rwatson; </td> |
| 1375 |
<td> 21 October 2004 </td> |
| 1376 |
<td> &status.done; </td> |
| 1377 |
</tr> |
| 1378 |
|
| 1379 |
<tr> |
| 1380 |
<td> Add KTR(9) tracing for busdma events. </td> |
| 1381 |
<td> &a.rwatson; </td> |
| 1382 |
<td> 23 October 2004 </td> |
| 1383 |
<td> &status.done; </td> |
| 1384 |
</tr> |
| 1385 |
|
| 1386 |
<tr> |
| 1387 |
<td> Add KTR(9) tracing for critical sections. </td> |
| 1388 |
<td> &a.rwatson; </td> |
| 1389 |
<td> 07 November 2004 </td> |
| 1390 |
<td> &status.done; </td> |
| 1391 |
</tr> |
| 1392 |
|
| 1393 |
<tr> |
| 1394 |
<td> Make the kernel fully preemptive. </td> |
| 1395 |
<td> &a.jhb; </td> |
| 1396 |
<td> 24 November 2004 </td> |
| 1397 |
<td> &status.done; </td> |
| 1398 |
</tr> |
| 1399 |
|
| 1400 |
<tr> |
| 1401 |
<td> Lock pipe implementation: sigio/fown-related evil. </td> |
| 1402 |
<td> &a.alfred; </td> |
| 1403 |
<td> 24 November 2004 </td> |
| 1404 |
<td> &status.done; </td> |
| 1405 |
</tr> |
| 1406 |
|
| 1407 |
<tr> |
| 1408 |
<td> Lock down the SysV IPC code. </td> |
| 1409 |
<td> &a.alfred; </td> |
| 1410 |
<td> 24 November 2004 </td> |
| 1411 |
<td> &status.done; </td> |
| 1412 |
</tr> |
| 1413 |
|
| 1414 |
<tr> |
| 1415 |
<td> Lock contention measurement tool to measure heat of various |
| 1416 |
locks, including Giant, and permit more directed performance and |
| 1417 |
locking strategy optimization. </td> |
| 1418 |
<td> &a.rwatson; </td> |
| 1419 |
<td> 24 November 2004 </td> |
| 1420 |
<td> &status.done; </td> |
| 1421 |
</tr> |
| 1422 |
|
| 1423 |
<tr> |
| 1424 |
<td> Add KTR(9) tracing to scheduler run queues. </td> |
| 1425 |
<td> &a.jeff; </td> |
| 1426 |
<td> 26 December 2004 </td> |
| 1427 |
<td> &status.done; </td> |
| 1428 |
</tr> |
| 1429 |
|
| 1430 |
<tr> |
| 1431 |
<td> Review locking strategy and correctness of VFS operations |
| 1432 |
and fix up various failure modes associated with enabling |
| 1433 |
VFS locking assertions. </td> |
| 1434 |
<td> &a.jeff; </td> |
| 1435 |
<td> 01 January 2005 </td> |
| 1436 |
<td> &status.done; </td> |
| 1437 |
</tr> |
| 1438 |
|
| 1439 |
<tr> |
| 1440 |
<td> Document in-vnode locking strategy, clean it up. </td> |
| 1441 |
<td> &a.jeff; </td> |
| 1442 |
<td> 01 January 2005 </td> |
| 1443 |
<td> &status.done; </td> |
| 1444 |
</tr> |
| 1445 |
|
| 1446 |
<tr> |
| 1447 |
<td> Run cross-file system VFS without Giant, acquiring Giant |
| 1448 |
conditionally based on a file system flag. </td> |
| 1449 |
<td> &a.jeff; </td> |
| 1450 |
<td> 01 January 2005 </td> |
| 1451 |
<td> &status.done; </td> |
| 1452 |
</tr> |
| 1453 |
|
| 1454 |
<tr> |
| 1455 |
<td> Run UFS file system MPSAFE. </td> |
| 1456 |
<td> &a.jeff; </td> |
| 1457 |
<td> 01 January 2005 </td> |
| 1458 |
<td> &status.done; </td> |
| 1459 |
</tr> |
| 1460 |
|
| 1461 |
<tr> |
| 1462 |
<td> Add KTR(9) tracing for buffer cache events. </td> |
| 1463 |
<td> &a.jeff; </td> |
| 1464 |
<td> 24 January 2005 </td> |
| 1465 |
<td> &status.done; </td> |
| 1466 |
</tr> |
| 1467 |
|
| 1468 |
<tr> |
| 1469 |
<td> Break out critical section and spin lock APIs, and re-optimize |
| 1470 |
critical sections to not disable interrupts in hardware due to |
| 1471 |
the high cost on some hardware architectures. </td> |
| 1472 |
<td> &a.jhb; </td> |
| 1473 |
<td> 04 April 2005 </td> |
| 1474 |
<td> &status.done; </td> |
| 1475 |
</tr> |
| 1476 |
|
| 1477 |
<tr> |
| 1478 |
<td> Modify uma(9) to use critical sections to protect per-CPU |
| 1479 |
statistics, instead of mutexdes, in order to optimize access. </td> |
| 1480 |
<td> &a.rwatson; </td> |
| 1481 |
<td> 29 April 2005 </td> |
| 1482 |
<td> &status.done; </td> |
| 1483 |
</tr> |
| 1484 |
|
| 1485 |
<tr> |
| 1486 |
<td> Migrate malloc(9) to per-CPU statistics, and use critical |
| 1487 |
sections to optimize access to those statistics. </td> |
| 1488 |
<td> &a.rwatson; </td> |
| 1489 |
<td> 29 May 2005 </td> |
| 1490 |
<td> &status.done; </td> |
| 1491 |
</tr> |
| 1492 |
|
| 1493 |
<tr> |
| 1494 |
<td> Add KTR(9) support for KTR_VFS to trace additional VFS events, |
| 1495 |
rather than mechanically inserted KTR_VOP events. </td> |
| 1496 |
<td> &a.jeff; </td> |
| 1497 |
<td> 11 June 2005 </td> |
| 1498 |
<td> &status.done; </td> |
| 1499 |
</tr> |
| 1500 |
|
| 1501 |
<tr> |
| 1502 |
<td> Push the grabbing of Giant into Linux i386 ABI system calls.</td> |
| 1503 |
<td> &a.jhb; </td> |
| 1504 |
<td> 13 July 2005 </td> |
| 1505 |
<td> &status.done; </td> |
| 1506 |
</tr> |
| 1507 |
|
| 1508 |
<tr> |
| 1509 |
<td> Push the grabbing of Giant into Linux AXP ABI system calls.</td> |
| 1510 |
<td> &a.jhb; </td> |
| 1511 |
<td> 13 July 2005 </td> |
| 1512 |
<td> &status.done; </td> |
| 1513 |
</tr> |
| 1514 |
|
| 1515 |
<tr> |
| 1516 |
<td> Push the grabbing of Giant into SVR4 i386 ABI system calls.</td> |
| 1517 |
<td> &a.jhb; </td> |
| 1518 |
<td> 13 July 2005 </td> |
| 1519 |
<td> &status.done; </td> |
| 1520 |
</tr> |
| 1521 |
|
| 1522 |
<tr> |
| 1523 |
<td> Push the grabbing of Giant into OSF/1 AXP ABI system calls.</td> |
| 1524 |
<td> &a.jhb; </td> |
| 1525 |
<td> 13 July 2005 </td> |
| 1526 |
<td> &status.done; </td> |
| 1527 |
</tr> |
| 1528 |
|
| 1529 |
<tr> |
| 1530 |
<td> Push the grabbing of Giant into IBCS i386 ABI system calls.</td> |
| 1531 |
<td> &a.jhb; </td> |
| 1532 |
<td> 13 July 2005 </td> |
| 1533 |
<td> &status.done; </td> |
| 1534 |
</tr> |
| 1535 |
|
| 1536 |
<tr> |
| 1537 |
<td> Add a new witness check for exiting threads to verify that |
| 1538 |
an exiting thread holds no locks. </td> |
| 1539 |
<td> &a.jhb; </td> |
| 1540 |
<td> 2 September 2005 </td> |
| 1541 |
<td> &status.done; </td> |
| 1542 |
</tr> |
| 1543 |
|
| 1544 |
<tr> |
| 1545 |
<td> Implement atomic_fetchadd() for ints. </td> |
| 1546 |
<td> &a.jhb; </td> |
| 1547 |
<td> 27 September 2005 </td> |
| 1548 |
<td> &status.done; </td> |
| 1549 |
</tr> |
| 1550 |
|
| 1551 |
<tr> |
| 1552 |
<td> Implement a simple reference count API using atomic operations and |
| 1553 |
use this to replace locks that just protect a reference count. </td> |
| 1554 |
<td> &a.jhb; </td> |
| 1555 |
<td> 27 September 2005 </td> |
| 1556 |
<td> &status.done; </td> |
| 1557 |
</tr> |
| 1558 |
|
| 1559 |
<tr> |
| 1560 |
<td> Split the interrupt handler list out of struct ithread into |
| 1561 |
its own structure and only start up kthreads for interrupt |
| 1562 |
vectors that actually have threaded interrupt handlers. </td> |
| 1563 |
<td> &a.jhb; </td> |
| 1564 |
<td> 25 October 2005 </td> |
| 1565 |
<td> &status.done; </td> |
| 1566 |
</tr> |
| 1567 |
|
| 1568 |
<tr> |
| 1569 |
<td> Lock aio(4). </td> |
| 1570 |
<td> &a.davidxu; </td> |
| 1571 |
<td> 22 January 2006 </td> |
| 1572 |
<td> &status.done; </td> |
| 1573 |
</tr> |
| 1574 |
|
| 1575 |
<tr> |
| 1576 |
<td> Implement reader/writer locks. </td> |
| 1577 |
<td> &a.jhb; </td> |
| 1578 |
<td> 27 January 2006 </td> |
| 1579 |
<td> &status.done; </td> |
| 1580 |
</tr> |
| 1581 |
|
| 1582 |
<tr> |
| 1583 |
<td> Remove the MPSAFE syscall flag from the syscall table. </td> |
| 1584 |
<td> &a.jhb; </td> |
| 1585 |
<td> 28 July 2006 </td> |
| 1586 |
<td> &status.done; </td> |
| 1587 |
</tr> |
| 1588 |
|
| 1589 |
<tr> |
| 1590 |
<td> Add sufficient synchronization to printf to avoid printing |
| 1591 |
problems but not so much it doesn't work in debugging contexts. </td> |
| 1592 |
<td> &a.jb; </td> |
| 1593 |
<td> 1 November 2006 </td> |
| 1594 |
<td> &status.done; </td> |
| 1595 |
</tr> |
| 1596 |
|
| 1597 |
<tr> |
| 1598 |
<td> Expand mutex profiling tool to also profile sx locks. </td> |
| 1599 |
<td> &a.kmacy; </td> |
| 1600 |
<td> 11 November 2006 </td> |
| 1601 |
<td> &status.done; </td> |
| 1602 |
</tr> |
| 1603 |
|
| 1604 |
<tr> |
| 1605 |
<td> Lock down the TTY subsystem. </td> |
| 1606 |
<td> &a.ed; </td> |
| 1607 |
<td> 20 August 2008 </td> |
| 1608 |
<td> &status.done; </td> |
| 1609 |
</tr> |
| 1610 |
|
| 1611 |
<!-- WIP --> |
| 1612 |
|
| 1613 |
<tr> |
| 1614 |
<td> Lock struct proc. </td> |
| 1615 |
<td> &a.jhb; </td> |
| 1616 |
<td> 20 February 2001 </td> |
| 1617 |
<td> &status.wip; </td> |
| 1618 |
</tr> |
| 1619 |
|
| 1620 |
<tr> |
| 1621 |
<td> Fix clock locking to be the same on all platforms. </td> |
| 1622 |
<td> &a.jhb; </td> |
| 1623 |
<td> 16 November 2001 </td> |
| 1624 |
<td> &status.wip; </td> |
| 1625 |
</tr> |
| 1626 |
|
| 1627 |
<tr> |
| 1628 |
<td> Make use of process locking and process reference counting |
| 1629 |
to protect debugging interfaces (and procfs). </td> |
| 1630 |
<td> &a.jhb; </td> |
| 1631 |
<td> 27 February 2002 </td> |
| 1632 |
<td> &status.wip; </td> |
| 1633 |
</tr> |
| 1634 |
|
| 1635 |
<tr> |
| 1636 |
<td> Make use of process locking to protect process monitoring |
| 1637 |
sysctls, including those employed by 'ps' and related tools. |
| 1638 |
<td> &a.jhb; </td> |
| 1639 |
<td> 27 February 2002 </td> |
| 1640 |
<td> &status.wip; </td> |
| 1641 |
</tr> |
| 1642 |
|
| 1643 |
<tr> |
| 1644 |
<td> Lock down newbus infrastructure to support driver |
| 1645 |
fine-graining. </td> |
| 1646 |
<td> &a.imp; </td> |
| 1647 |
<td> 28 February 2002 </td> |
| 1648 |
<td> &status.wip; </td> |
| 1649 |
</tr> |
| 1650 |
|
| 1651 |
<tr> |
| 1652 |
<td> SMPng architecture document. </td> |
| 1653 |
<td> &a.jhb;, &a.rwatson; </td> |
| 1654 |
<td> 28 February 2002 </td> |
| 1655 |
<td> &status.wip; </td> |
| 1656 |
</tr> |
| 1657 |
|
| 1658 |
<tr> |
| 1659 |
<td> Move to shared lock for VOP_GETATTR() to reduce blocking during |
| 1660 |
frequent lightweight VFS operations. Modify namei() to provide |
| 1661 |
a LOOKUP_SHARED flag to indicate when the lock required may be |
| 1662 |
shared instead of exclusive. </td> |
| 1663 |
<td> &a.jeff; </td> |
| 1664 |
<td> 11 March 2002 </td> |
| 1665 |
<td> &status.wip; </td> |
| 1666 |
</tr> |
| 1667 |
|
| 1668 |
<tr> |
| 1669 |
<td> Document existing vm_map locking and verify it's correctness. </td> |
| 1670 |
<td> &a.alc; </td> |
| 1671 |
<td> 18 May 2002 </td> |
| 1672 |
<td> &status.wip; </td> |
| 1673 |
</tr> |
| 1674 |
|
| 1675 |
<tr> |
| 1676 |
<td> Document existing vm_object locking and verify it's |
| 1677 |
correctness. </td> |
| 1678 |
<td> &a.alc; </td> |
| 1679 |
<td> 4 May 2002 </td> |
| 1680 |
<td> &status.wip; </td> |
| 1681 |
</tr> |
| 1682 |
|
| 1683 |
<tr> |
| 1684 |
<td> Implement lazy interrupt thread switching (context |
| 1685 |
stealing) on i386. </td> |
| 1686 |
<td> &a.bmilekic;, &a.kan; </td> |
| 1687 |
<td> 10 December 2002 </td> |
| 1688 |
<td> &status.wip; </td> |
| 1689 |
</tr> |
| 1690 |
|
| 1691 |
<tr> |
| 1692 |
<td> Implement lazy interrupt thread switching (context |
| 1693 |
stealing) on sparc64. </td> |
| 1694 |
<td> &a.jake; </td> |
| 1695 |
<td> 10 December 2002 </td> |
| 1696 |
<td> &status.wip; </td> |
| 1697 |
</tr> |
| 1698 |
|
| 1699 |
<tr> |
| 1700 |
<td> Switch from using lockmgr in VM to using a mutex or exclusive |
| 1701 |
sxlock. Push down Giant on all VM except for vm_object/VFS and |
| 1702 |
vm_page/pmap components. </td> |
| 1703 |
<td> &a.alc; </td> |
| 1704 |
<td> 10 December 2002 </td> |
| 1705 |
<td> &status.wip; </td> |
| 1706 |
</tr> |
| 1707 |
|
| 1708 |
<tr> |
| 1709 |
<td> Modify device driver API to permit drivers to more easily |
| 1710 |
split "in interrupt context" and "in interrupt thread" |
| 1711 |
code so as to acknowledge interrupts faster. This will permit |
| 1712 |
lower latency in interrupt handling. </td> |
| 1713 |
<td> &a.piso; </td> |
| 1714 |
<td> 24 February 2007 </td> |
| 1715 |
<td> &status.wip; </td> |
| 1716 |
</tr> |
| 1717 |
|
| 1718 |
<!-- Stalled --> |
| 1719 |
|
| 1720 |
<!-- Not Started --> |
| 1721 |
|
| 1722 |
<tr> |
| 1723 |
<td> Conditionalize atomic ops in the SMP code that are used |
| 1724 |
for debugging statistics. </td> |
| 1725 |
<td> &a.peter; </td> |
| 1726 |
<td> 15 March 2001 </td> |
| 1727 |
<td> &status.new; </td> |
| 1728 |
</tr> |
| 1729 |
|
| 1730 |
<tr> |
| 1731 |
<td> Axe schedcpu() in favor of event driven priority updates as |
| 1732 |
much as possible. </td> |
| 1733 |
<td> </td> |
| 1734 |
<td> 7 September 2001 </td> |
| 1735 |
<td> &status.new; </td> |
| 1736 |
</tr> |
| 1737 |
|
| 1738 |
<tr> |
| 1739 |
<td> Fix *hold (e.g. crhold) to return reference to object. </td> |
| 1740 |
<td> </td> |
| 1741 |
<td> 7 September 2001 </td> |
| 1742 |
<td> &status.new; </td> |
| 1743 |
</tr> |
| 1744 |
|
| 1745 |
<tr> |
| 1746 |
<td> Add witness checking for lockmgr locks. </td> |
| 1747 |
<td> </td> |
| 1748 |
<td> 7 September 2001 </td> |
| 1749 |
<td> &status.new; </td> |
| 1750 |
</tr> |
| 1751 |
|
| 1752 |
<tr> |
| 1753 |
<td> Add ICU spin locks on ia64. </td> |
| 1754 |
<td> </td> |
| 1755 |
<td> 4 January 2002 </td> |
| 1756 |
<td> &status.new; </td> |
| 1757 |
</tr> |
| 1758 |
|
| 1759 |
<tr> |
| 1760 |
<td> Add a witness_sleep() check to copyin/out() and s/fuword(). </td> |
| 1761 |
<td> &a.jhb; </td> |
| 1762 |
<td> 7 June 2002 </td> |
| 1763 |
<td> &status.new; </td> |
| 1764 |
</tr> |
| 1765 |
|
| 1766 |
<tr> |
| 1767 |
<td> Remove Giant from mi_startup() and push Giant down into |
| 1768 |
individual SYSINIT functions. Many SYSINIT functions probably do |
| 1769 |
not need Giant anyway. </td> |
| 1770 |
<td> </td> |
| 1771 |
<td> 21 October 2005 </td> |
| 1772 |
<td> &status.new; </td> |
| 1773 |
</tr> |
| 1774 |
|
| 1775 |
<!-- |
| 1776 |
<tr> |
| 1777 |
<td> </td> |
| 1778 |
<td> </td> |
| 1779 |
<td> D M 2002 </td> |
| 1780 |
<td> &status.new; </td> |
| 1781 |
</tr> |
| 1782 |
|
| 1783 |
--> |
| 1784 |
</table> |
| 1785 |
|
| 1786 |
<p>This table lists the todo subtasks for multithreading the network |
| 1787 |
stack.</p> |
| 1788 |
|
| 1789 |
<table class="tblbasic"> |
| 1790 |
<tr class="heading"> |
| 1791 |
<th> Task </th> |
| 1792 |
<th> Responsible </th> |
| 1793 |
<th> Last updated </th> |
| 1794 |
<th> Status </th> |
| 1795 |
</tr> |
| 1796 |
|
| 1797 |
<tr> |
| 1798 |
<td> Protect network interface queues. </td> |
| 1799 |
<td> &a.jlemon; </td> |
| 1800 |
<td> 24 November 2000 </td> |
| 1801 |
<td> &status.done; </td> |
| 1802 |
</tr> |
| 1803 |
|
| 1804 |
<tr> |
| 1805 |
<td> Lock up IP. </td> |
| 1806 |
<td> &a.jennifer;, &a.hsu; </td> |
| 1807 |
<td> 10 June 2002 </td> |
| 1808 |
<td> &status.done; </td> |
| 1809 |
</tr> |
| 1810 |
|
| 1811 |
<tr> |
| 1812 |
<td> Lock up TCP. </td> |
| 1813 |
<td> &a.jennifer;, &a.hsu;, &a.sam;, &a.rwatson;</td> |
| 1814 |
<td> 24 November 2004 </td> |
| 1815 |
<td> &status.done; </td> |
| 1816 |
</tr> |
| 1817 |
|
| 1818 |
<tr> |
| 1819 |
<td> Lock up UDP. </td> |
| 1820 |
<td> &a.jennifer;, &a.hsu;, &a.rwatson; </td> |
| 1821 |
<td> 24 November 2004 </td> |
| 1822 |
<td> &status.done; </td> |
| 1823 |
</tr> |
| 1824 |
|
| 1825 |
<tr> |
| 1826 |
<td> Lock ifaddr reference counts. </td> |
| 1827 |
<td> &a.hsu; </td> |
| 1828 |
<td> 18 December 2002 </td> |
| 1829 |
<td> &status.done; </td> |
| 1830 |
</tr> |
| 1831 |
|
| 1832 |
<tr> |
| 1833 |
<td> Lock up ifnet list. </td> |
| 1834 |
<td> &a.hsu; </td> |
| 1835 |
<td> 21 December 2002 </td> |
| 1836 |
<td> &status.done; </td> |
| 1837 |
</tr> |
| 1838 |
|
| 1839 |
<tr> |
| 1840 |
<td> Lock radix trees. </td> |
| 1841 |
<td> &a.hsu; </td> |
| 1842 |
<td> 23 December 2002 </td> |
| 1843 |
<td> &status.done; </td> |
| 1844 |
</tr> |
| 1845 |
|
| 1846 |
<tr> |
| 1847 |
<td> Lock up ARP. </td> |
| 1848 |
<td> &a.hsu; </td> |
| 1849 |
<td> 16 January 2003 </td> |
| 1850 |
<td> &status.done; </td> |
| 1851 |
</tr> |
| 1852 |
|
| 1853 |
<tr> |
| 1854 |
<td> Lock up raw IP. </td> |
| 1855 |
<td> &a.sam;, &a.rwatson; </td> |
| 1856 |
<td> 24 July 2004 </td> |
| 1857 |
<td> &status.done; </td> |
| 1858 |
</tr> |
| 1859 |
|
| 1860 |
<tr> |
| 1861 |
<td> Lock divert sockets. </td> |
| 1862 |
<td> &a.sam; </td> |
| 1863 |
<td> 4 October 2003 </td> |
| 1864 |
<td> &status.done; </td> |
| 1865 |
</tr> |
| 1866 |
|
| 1867 |
<tr> |
| 1868 |
<td> Lock ipfw2. </td> |
| 1869 |
<td> &a.sam; </td> |
| 1870 |
<td> 4 October 2003 </td> |
| 1871 |
<td> &status.done; </td> |
| 1872 |
</tr> |
| 1873 |
|
| 1874 |
<tr> |
| 1875 |
<td> Lock DUMMYNET. </td> |
| 1876 |
<td> &a.sam; </td> |
| 1877 |
<td> 4 October 2003 </td> |
| 1878 |
<td> &status.done; </td> |
| 1879 |
</tr> |
| 1880 |
|
| 1881 |
<tr> |
| 1882 |
<td> Lock ethernet bridge. </td> |
| 1883 |
<td> &a.sam; </td> |
| 1884 |
<td> 4 October 2003 </td> |
| 1885 |
<td> &status.done; </td> |
| 1886 |
</tr> |
| 1887 |
|
| 1888 |
<tr> |
| 1889 |
<td> Lock IP fragment queues. </td> |
| 1890 |
<td> &a.rwatson; </td> |
| 1891 |
<td> 4 October 2003 </td> |
| 1892 |
<td> &status.done; </td> |
| 1893 |
</tr> |
| 1894 |
|
| 1895 |
<tr> |
| 1896 |
<td> Lock routing entries. </td> |
| 1897 |
<td> &a.sam; </td> |
| 1898 |
<td> 4 October 2003 </td> |
| 1899 |
<td> &status.done; </td> |
| 1900 |
</tr> |
| 1901 |
|
| 1902 |
<tr> |
| 1903 |
<td> Lock FAST_IPSEC. </td> |
| 1904 |
<td> &a.sam; </td> |
| 1905 |
<td> 4 October 2003 </td> |
| 1906 |
<td> &status.done; </td> |
| 1907 |
</tr> |
| 1908 |
|
| 1909 |
<tr> |
| 1910 |
<td> Permit parallel entry into isr processing. </td> |
| 1911 |
<td> &a.rwatson;, &a.sam; </td> |
| 1912 |
<td> 11 October 2003 </td> |
| 1913 |
<td> &status.done; </td> |
| 1914 |
</tr> |
| 1915 |
|
| 1916 |
<tr> |
| 1917 |
<td> Lock if_disc "discard interface". </td> |
| 1918 |
<td> &a.rwatson; </td> |
| 1919 |
<td> 9 March 2004 </td> |
| 1920 |
<td> &status.done; </td> |
| 1921 |
</tr> |
| 1922 |
|
| 1923 |
<tr> |
| 1924 |
<td> Lock if_faith "IPv6-to-IPv4 TCP relay interface. " </td> |
| 1925 |
<td> &a.sam;, &a.rwatson; </td> |
| 1926 |
<td> 9 March 2004 </td> |
| 1927 |
<td> &status.done; </td> |
| 1928 |
</tr> |
| 1929 |
|
| 1930 |
<tr> |
| 1931 |
<td> Lock if_gif "generic tunnel interface". </td> |
| 1932 |
<td> &a.rwatson; </td> |
| 1933 |
<td> 9 March 2004 </td> |
| 1934 |
<td> &status.done; </td> |
| 1935 |
</tr> |
| 1936 |
|
| 1937 |
<tr> |
| 1938 |
<td> Review ECN tunnel support (ip_ecn.c). </td> |
| 1939 |
<td> &a.rwatson; </td> |
| 1940 |
<td> 9 March 2004 </td> |
| 1941 |
<td> &status.done; </td> |
| 1942 |
</tr> |
| 1943 |
|
| 1944 |
<tr> |
| 1945 |
<td> if_tap global and softc locking. </td> |
| 1946 |
<td> &a.rwatson; </td> |
| 1947 |
<td> 23 April 2004 </td> |
| 1948 |
<td> &status.done; </td> |
| 1949 |
</tr> |
| 1950 |
|
| 1951 |
<tr> |
| 1952 |
<td> if_tun global and softc locking. </td> |
| 1953 |
<td> &a.rwatson; </td> |
| 1954 |
<td> 23 April 2004 </td> |
| 1955 |
<td> &status.done; </td> |
| 1956 |
</tr> |
| 1957 |
|
| 1958 |
<tr> |
| 1959 |
<td> netatalk/aarp.c locking. </td> |
| 1960 |
<td> &a.rwatson; </td> |
| 1961 |
<td> 23 April 2004 </td> |
| 1962 |
<td> &status.done; </td> |
| 1963 |
</tr> |
| 1964 |
|
| 1965 |
<tr> |
| 1966 |
<td> Cache socket MAC label in inpcb label for IPv4 sockets so that |
| 1967 |
the label can be used safely at the inet layer without socket |
| 1968 |
locks. </tD> |
| 1969 |
<td> &a.rwatson; </td> |
| 1970 |
<td> 23 April 2004 </td> |
| 1971 |
<td> &status.done; </td> |
| 1972 |
</tr> |
| 1973 |
|
| 1974 |
<tr> |
| 1975 |
<td> IP encapsulation subroutines (ip_encap.c). </td> |
| 1976 |
<td> &a.rwatson; </td> |
| 1977 |
<td> 23 April 2004 </td> |
| 1978 |
<td> &status.done; </td> |
| 1979 |
</tr> |
| 1980 |
|
| 1981 |
<tr> |
| 1982 |
<td> Lock globals in loopback interface (if_loop.c). </td> |
| 1983 |
<td> &a.rwatson; </td> |
| 1984 |
<td> 23 April 2004 </td> |
| 1985 |
<td> &status.done; </td> |
| 1986 |
</tr> |
| 1987 |
|
| 1988 |
<tr> |
| 1989 |
<td> Use m_tags in if_gif to limit looping configurations, rather |
| 1990 |
than a non-MPSAFE static coutner. </td> |
| 1991 |
<td> &a.ru; </td> |
| 1992 |
<td> 23 April 2004 </td> |
| 1993 |
<td> &status.done; </td> |
| 1994 |
</tr> |
| 1995 |
|
| 1996 |
<tr> |
| 1997 |
<td> netatalk DDP PCB locking. </td> |
| 1998 |
<td> &a.rwatson; </td> |
| 1999 |
<td> 24 July 2004 </td> |
| 2000 |
<td> &status.done; </td> |
| 2001 |
</tr> |
| 2002 |
|
| 2003 |
<tr> |
| 2004 |
<td> Lock up syncache. </td> |
| 2005 |
<td> &a.hsu;, &a.sam; </td> |
| 2006 |
<td> 10 November 2003 </td> |
| 2007 |
<td> &status.done; </td> |
| 2008 |
</tr> |
| 2009 |
|
| 2010 |
<tr> |
| 2011 |
<td> Permit IP forwarding path to run Giant-free. </td> |
| 2012 |
<td> &a.sam; </td> |
| 2013 |
<td> 1 December 2003 </td> |
| 2014 |
<td> &status.done; </td> |
| 2015 |
</tr> |
| 2016 |
|
| 2017 |
<tr> |
| 2018 |
<td> Lock &unix; domain protocols, fifofs. </td> |
| 2019 |
<td> &a.sam;, &a.rwatson; </td> |
| 2020 |
<td> 24 July 2004 </td> |
| 2021 |
<td> &status.done; </td> |
| 2022 |
</tr> |
| 2023 |
|
| 2024 |
<tr> |
| 2025 |
<td> Giant lock over NFS server to protect against so_upcall() w/o Giant </td> |
| 2026 |
<td> &a.rwatson; </td> |
| 2027 |
<td> 24 July 2004 </td> |
| 2028 |
<td> &status.done; </td> |
| 2029 |
</tr> |
| 2030 |
|
| 2031 |
<tr> |
| 2032 |
<td> Lock interface cloning meta-data. </td> |
| 2033 |
<td> &a.brooks; </td> |
| 2034 |
<td> 24 July 2004 </td> |
| 2035 |
<td> &status.done; </td> |
| 2036 |
</tr> |
| 2037 |
|
| 2038 |
<tr> |
| 2039 |
<td> Apply combination of socket and socket buffer locks, label caching |
| 2040 |
to MAC labels on sockets so that they can be used safely without |
| 2041 |
Giant. </td> |
| 2042 |
<td> &a.rwatson; </td> |
| 2043 |
<td> 24 July 2004 </td> |
| 2044 |
<td> &status.done; </td> |
| 2045 |
</tr> |
| 2046 |
|
| 2047 |
<tr> |
| 2048 |
<td> Make routing socket message dispatch use a netisr to avoid |
| 2049 |
re-entering the socket code from the routing code, resolving lock |
| 2050 |
order issues. </td> |
| 2051 |
<td> &a.rwatson; </td> |
| 2052 |
<td> 24 July 2004 </td> |
| 2053 |
<td> &status.done; </td> |
| 2054 |
</tr> |
| 2055 |
|
| 2056 |
<tr> |
| 2057 |
<td> Introduce accept locking to protect accept incomplete and |
| 2058 |
complete queues on listen sockets. </td> |
| 2059 |
<td> &a.rwatson; </td> |
| 2060 |
<td> 24 July 2004 </td> |
| 2061 |
<td> &status.done; </td> |
| 2062 |
</tr> |
| 2063 |
|
| 2064 |
<tr> |
| 2065 |
<td> Break out socket buffer wakeup, socket buffer append, socket |
| 2066 |
state change, socket buffer reserve, flush, etc, calls into |
| 2067 |
_locked() and unlocked versions, and avoid conditional locking. |
| 2068 |
</td> |
| 2069 |
<td> &a.rwatson; </td> |
| 2070 |
<td> 24 July 2004 </td> |
| 2071 |
<td> &status.done; </td> |
| 2072 |
</tr> |
| 2073 |
|
| 2074 |
<tr> |
| 2075 |
<td> Lock down AARP, AppleTalk Address Resolution Protocol. </td> |
| 2076 |
<td> &a.rwatson; </td> |
| 2077 |
<td> 24 July 2004 </td> |
| 2078 |
<td> &status.done; </td> |
| 2079 |
</tr> |
| 2080 |
|
| 2081 |
<tr> |
| 2082 |
<td> Fix pull/push cache data synchronization issues in sosend(), |
| 2083 |
soreceive(), allowing them to run Giant-free. </td> |
| 2084 |
<td> &a.rwatson; </td> |
| 2085 |
<td> 24 July 2004 </td> |
| 2086 |
<td> &status.done; </td> |
| 2087 |
</tr> |
| 2088 |
|
| 2089 |
<tr> |
| 2090 |
<td> Protect socket global counters/limits and generation number |
| 2091 |
with a mutex. </td> |
| 2092 |
<td> &a.rwatson; </td> |
| 2093 |
<td> 24 July 2004 </td> |
| 2094 |
<td> &status.done; </td> |
| 2095 |
</tr> |
| 2096 |
|
| 2097 |
<tr> |
| 2098 |
<td> Lock down unit allocation meta-data in interface related |
| 2099 |
netgraph modules. </td> |
| 2100 |
<td> &a.rwatson; </td> |
| 2101 |
<td> 24 July 2004 </td> |
| 2102 |
<td> &status.done; </td> |
| 2103 |
</tr> |
| 2104 |
|
| 2105 |
<tr> |
| 2106 |
<td> Lock down socket buffer OOB fields across TCP/IP, IPX. </td> |
| 2107 |
<td> &a.rwatson; </td> |
| 2108 |
<td> 24 July 2004 </td> |
| 2109 |
<td> &status.done; </td> |
| 2110 |
</tr> |
| 2111 |
|
| 2112 |
<tr> |
| 2113 |
<td> Add MSG_NBIO so that fifofs can avoid frobbing SO_NBIO in |
| 2114 |
a manner that risks races. </td> |
| 2115 |
<td> &a.truckman; </td> |
| 2116 |
<td> 24 July 2004 </td> |
| 2117 |
<td> &status.done; </td> |
| 2118 |
</tr> |
| 2119 |
|
| 2120 |
<tr> |
| 2121 |
<td> Protect all use of so_count with socket lock. </td> |
| 2122 |
<td> &a.rwatson; </td> |
| 2123 |
<td> 24 July 2004 </td> |
| 2124 |
<td> &status.done; </td> |
| 2125 |
</tr> |
| 2126 |
|
| 2127 |
<tr> |
| 2128 |
<td> Move socket buffer related state from so_state to sb_state so |
| 2129 |
it can be properly locked by the socket buffer mutex. </td> |
| 2130 |
<td> &a.rwatson; </td> |
| 2131 |
<td> 24 July 2004 </td> |
| 2132 |
<td> &status.done; </td> |
| 2133 |
</tr> |
| 2134 |
|
| 2135 |
<tr> |
| 2136 |
<td> Introduce a temporary global lock to lock the if_label field |
| 2137 |
used by the MAC Framework. </tD> |
| 2138 |
<td> &a.rwatson; </td> |
| 2139 |
<td> 24 July 2004 </td> |
| 2140 |
<td> &status.done; </td> |
| 2141 |
</tr> |
| 2142 |
|
| 2143 |
<tr> |
| 2144 |
<td> Push VFS-specific behavior out of fdrop_locked() and acquire |
| 2145 |
Giant in the fo_close per-object methods rather than |
| 2146 |
fdrop_locked(), so that pipes and sockets can run fo_close() |
| 2147 |
Giant-free. </td> |
| 2148 |
<td> &a.rwatson; </td> |
| 2149 |
<td> 24 July 2004 </td> |
| 2150 |
<td> &status.done; </td> |
| 2151 |
</tr> |
| 2152 |
|
| 2153 |
<tr> |
| 2154 |
<td> Push Giant acquisition into fo_stat() file descriptor stat |
| 2155 |
operation, rather than acquiring it in fstat(), so that fstat() |
| 2156 |
on sockets and pipes can run Giant-free. </td> |
| 2157 |
<td> &a.rwatson; </td> |
| 2158 |
<td> 24 July 2004 </td> |
| 2159 |
<td> &status.done; </td> |
| 2160 |
</tr> |
| 2161 |
|
| 2162 |
<tr> |
| 2163 |
<td> Don't hold socket locks over entry to protocol switch methods, |
| 2164 |
allowing protocol methods to acquire socket locks after protocol |
| 2165 |
locks in the lock order. </td> |
| 2166 |
<td> &a.rwatson; </td> |
| 2167 |
<td> 24 July 2004 </td> |
| 2168 |
<td> &status.done; </td> |
| 2169 |
</tr> |
| 2170 |
|
| 2171 |
<tr> |
| 2172 |
<td> Port inpcb mutex locking, assertions from IPv4 to IPv6. </td> |
| 2173 |
<td> &a.rwatson; </td> |
| 2174 |
<td> 8 August 2004 </td> |
| 2175 |
<td> &status.done; </td> |
| 2176 |
</tr> |
| 2177 |
|
| 2178 |
<tr> |
| 2179 |
<td> Add IFF_NEEDSGIANT to allow if_start to run with Giant for |
| 2180 |
specific interfaces. Defer if_start to task queue. </td> |
| 2181 |
<td> &a.rwatson; </td> |
| 2182 |
<td> 8 August 2004 </td> |
| 2183 |
<td> &status.done; </td> |
| 2184 |
</tr> |
| 2185 |
|
| 2186 |
<tr> |
| 2187 |
<td> Push down Giant in stat(), fo_stat() to allow Giant-free stat |
| 2188 |
of pipes, sockets. </td> |
| 2189 |
<td> &a.rwatson; </td> |
| 2190 |
<td> 8 August 2004 </td> |
| 2191 |
<td> &status.done; </td> |
| 2192 |
</tr> |
| 2193 |
|
| 2194 |
<tr> |
| 2195 |
<td> Add TCP lock assertions. </td> |
| 2196 |
<td> &a.rwatson; </td> |
| 2197 |
<td> 24 November 2004 </td> |
| 2198 |
<td> &status.done; </td> |
| 2199 |
</tr> |
| 2200 |
|
| 2201 |
<tr> |
| 2202 |
<td> Lock socket layer. </td> |
| 2203 |
<td> &a.sam;, &a.rwatson; </td> |
| 2204 |
<td> 24 November 2004 </td> |
| 2205 |
<td> &status.done; </td> |
| 2206 |
</tr> |
| 2207 |
|
| 2208 |
<tr> |
| 2209 |
<td> Review TCP timer code. </td> |
| 2210 |
<td> &a.rwatson; </td> |
| 2211 |
<td> 24 November 2004 </td> |
| 2212 |
<td> &status.done; </td> |
| 2213 |
</tr> |
| 2214 |
|
| 2215 |
<tr> |
| 2216 |
<td> Analyze and reduce cost of entropy gathering in network critical |
| 2217 |
paths. </td> |
| 2218 |
<td> &a.rwatson;, &a.markm; </td> |
| 2219 |
<td> 24 November 2004 </td> |
| 2220 |
<td> &status.done; </td> |
| 2221 |
</tr> |
| 2222 |
|
| 2223 |
<tr> |
| 2224 |
<td> Allow code to declare NET_NEEDS_GIANT(), forcing Giant over the |
| 2225 |
network stack if that code is compiled into the kernel. </td> |
| 2226 |
<td> &a.rwatson; </td> |
| 2227 |
<td> 28 August 2004 </td> |
| 2228 |
<td> &status.done; </td> |
| 2229 |
</tr> |
| 2230 |
|
| 2231 |
<tr> |
| 2232 |
<td> Disable Giant over the network stack in the default |
| 2233 |
configuration. </td> |
| 2234 |
<td> &a.rwatson; </td> |
| 2235 |
<td> 28 August 2004 </td> |
| 2236 |
<td> &status.done; </td> |
| 2237 |
</tr> |
| 2238 |
|
| 2239 |
<tr> |
| 2240 |
<td> Additional KTR tracing for UMA, callouts, interrupts, etc. </td> |
| 2241 |
<td> &a.rwatson; </td> |
| 2242 |
<td> 07 November 2004 </td> |
| 2243 |
<td> &status.done; </td> |
| 2244 |
</tr> |
| 2245 |
|
| 2246 |
<tr> |
| 2247 |
<td> Move to using file descriptor reference counts instead of |
| 2248 |
socket reference counts for socket system calls, avoiding extra |
| 2249 |
reference couht operations. </td> |
| 2250 |
<td> &a.rwatson; </td> |
| 2251 |
<td> 24 October 2004 </td> |
| 2252 |
<td> &status.done; </td> |
| 2253 |
</tr> |
| 2254 |
|
| 2255 |
<tr> |
| 2256 |
<td> Lock down netnatm. </td> |
| 2257 |
<td> &a.rwatson; </td> |
| 2258 |
<td> April 2006 </td> |
| 2259 |
<td> &status.done; </td> |
| 2260 |
</tr> |
| 2261 |
|
| 2262 |
<tr> |
| 2263 |
<td> Default to direct dispatch from ithreads </td> |
| 2264 |
<td> &a.rwatson; </td> |
| 2265 |
<td> 28 November 2006 </td> |
| 2266 |
<td> &status.done; </td> |
| 2267 |
</tr> |
| 2268 |
|
| 2269 |
<!-- WIP --> |
| 2270 |
|
| 2271 |
<tr> |
| 2272 |
<td> Lock IPv6. </td> |
| 2273 |
<td> &a.sam;, &a.rwatson;, &a.ume;, &a.mlaier; </td> |
| 2274 |
<td> 8 August 2004 </td> |
| 2275 |
<td> &status.wip; </td> |
| 2276 |
</tr> |
| 2277 |
|
| 2278 |
<tr> |
| 2279 |
<td> if_ppp global, per-softc locking. </td> |
| 2280 |
<td> &a.rwatson;, Maurycy Pawlowski-Wieronski <maurycy@fouk.org> </td> |
| 2281 |
<td> 23 April 2004 </td> |
| 2282 |
<td> &status.wip; </td> |
| 2283 |
</tr> |
| 2284 |
|
| 2285 |
<tr> |
| 2286 |
<td> Lock struct ifnet. </td> |
| 2287 |
<td> &a.mlaier;, &a.luigi;, Maurycy Pawlowski-Wieronski <maurycy@fouk.org> </td> |
| 2288 |
<td> 23 April 2004 </td> |
| 2289 |
<td> &status.wip; </td> |
| 2290 |
</tr> |
| 2291 |
|
| 2292 |
<tr> |
| 2293 |
<td> Lock IPv4, IPv6, atalk interface address lists. </td> |
| 2294 |
<td> &a.mlaier;, &a.rwatson; </td> |
| 2295 |
<td> 8 August 2004 </td> |
| 2296 |
<td> &status.wip; </td> |
| 2297 |
</tr> |
| 2298 |
|
| 2299 |
<tr> |
| 2300 |
<td> Lock consumers of BSD compress (bsd_comp.c) code to protect |
| 2301 |
compression state. </td> |
| 2302 |
<td> &a.rwatson; </td> |
| 2303 |
<td> 23 April 2004 </td> |
| 2304 |
<td> &status.wip; </td> |
| 2305 |
</tr> |
| 2306 |
|
| 2307 |
<tr> |
| 2308 |
<td> Lock global and softc state for six-to-four converter (if_stf.c). </td> |
| 2309 |
<td> &a.rwatson; </td> |
| 2310 |
<td> 23 April 2004 </td> |
| 2311 |
<td> &status.wip; </td> |
| 2312 |
</tr> |
| 2313 |
|
| 2314 |
<tr> |
| 2315 |
<td> Lock down global and softc state for SLIP (if_sl.c). </td> |
| 2316 |
<td> &a.rwatson; </td> |
| 2317 |
<td> 23 April 2004 </td> |
| 2318 |
<td> &status.wip; </td> |
| 2319 |
</tr> |
| 2320 |
|
| 2321 |
<tr> |
| 2322 |
<td> Lock global and softc state for SPPP (if_sppsubr.c). </td> |
| 2323 |
<td> &a.rik;, &a.rwatson; </td> |
| 2324 |
<td> 23 April 2004 </td> |
| 2325 |
<td> &status.wip; </td> |
| 2326 |
</tr> |
| 2327 |
|
| 2328 |
<tr> |
| 2329 |
<td> IGMP locking. </td> |
| 2330 |
<td> &a.rwatson; </td> |
| 2331 |
<td> 23 April 2004 </td> |
| 2332 |
<td> &status.wip; </td> |
| 2333 |
</tr> |
| 2334 |
|
| 2335 |
<tr> |
| 2336 |
<td> IP ID locking. </td> |
| 2337 |
<td> &a.ups; </td> |
| 2338 |
<td> 24 June 2005 </td> |
| 2339 |
<td> &status.wip; </td> |
| 2340 |
</tr> |
| 2341 |
|
| 2342 |
<tr> |
| 2343 |
<td> Locking for polling(4). </td> |
| 2344 |
<td> &a.pjd;, &a.glebius; </td> |
| 2345 |
<td> 19 October 2005 </td> |
| 2346 |
<td> &status.wip; </td> |
| 2347 |
</tr> |
| 2348 |
|
| 2349 |
<tr> |
| 2350 |
<td> BPF locking needs some cleanup, there are some race conditions there relating to interface removal. </td> |
| 2351 |
<td> &a.csjp; </td> |
| 2352 |
<td> 2 June 2006 </td> |
| 2353 |
<td> &status.wip; </td> |
| 2354 |
</tr> |
| 2355 |
|
| 2356 |
<!-- Not Started --> |
| 2357 |
|
| 2358 |
<tr> |
| 2359 |
<td> Reduce contention upon locking a socket buffer by replacing |
| 2360 |
tsleep() and wakeup() with a condvar. </td> |
| 2361 |
<td> &a.tanimura; </td> |
| 2362 |
<td> 21 April 2002 </td> |
| 2363 |
<td> &status.new; </td> |
| 2364 |
</tr> |
| 2365 |
|
| 2366 |
<tr> |
| 2367 |
<td> Lock if_ef "ethernet frame" driver. </td> |
| 2368 |
<td> </td> |
| 2369 |
<td> 9 March 2004 </td> |
| 2370 |
<td> &status.new; </td> |
| 2371 |
</tr> |
| 2372 |
|
| 2373 |
<tr> |
| 2374 |
<td> Further cleanup of socket state machine in order to facilitate finishing socking locking of state transitions. </td> |
| 2375 |
<td> &a.rwatson; </td> |
| 2376 |
<td> 19 October 2005 </td> |
| 2377 |
<td> &status.new; </td> |
| 2378 |
</tr> |
| 2379 |
|
| 2380 |
<tr> |
| 2381 |
<td> Lock KAME IPSEC. </td> |
| 2382 |
<td> </td> |
| 2383 |
<td> 19 October 2005 </td> |
| 2384 |
<td> &status.new; </td> |
| 2385 |
</tr> |
| 2386 |
|
| 2387 |
<tr> |
| 2388 |
<td> Only one of our ATM stacks is MPSAFE, the other two should be deleted or fixed. </td> |
| 2389 |
<td> </td> |
| 2390 |
<td> 19 October 2005 </td> |
| 2391 |
<td> &status.new; </td> |
| 2392 |
</tr> |
| 2393 |
|
| 2394 |
<tr> |
| 2395 |
<td> Lock ND6 (IPv6 Neighbor Discovery). </td> |
| 2396 |
<td> &a.gnn; </td> |
| 2397 |
<td> 19 October 2005 </td> |
| 2398 |
<td> &status.new; </td> |
| 2399 |
</tr> |
| 2400 |
|
| 2401 |
<tr> |
| 2402 |
<td> Lock IPv6 multicast address lists. </td> |
| 2403 |
<td> </td> |
| 2404 |
<td> 19 October 2005 </td> |
| 2405 |
<td> &status.new; </td> |
| 2406 |
</tr> |
| 2407 |
|
| 2408 |
<tr> |
| 2409 |
<td> Lock IPv4 and IPv6 global address lists. </td> |
| 2410 |
<td> </td> |
| 2411 |
<td> 19 October 2005 </td> |
| 2412 |
<td> &status.new; </td> |
| 2413 |
</tr> |
| 2414 |
|
| 2415 |
<tr> |
| 2416 |
<td> Continued cleanup of the stack/device driver ownership and locking for struct ifnet needs to be done. Most fields are now either locked or assigned ownership. Some fields, such as if_flags, need a bit more cleanup due to device drivers modifying stack-owned fields. </td> |
| 2417 |
<td> &a.rwatson; </td> |
| 2418 |
<td> 19 October 2005 </td> |
| 2419 |
<td> &status.new; </td> |
| 2420 |
</tr> |
| 2421 |
|
| 2422 |
<tr> |
| 2423 |
<td> When interfaces are torn down, there are a number of races (not all associated with SMPng) that need to be thought about. </td> |
| 2424 |
<td> </td> |
| 2425 |
<td> 19 October 2005 </td> |
| 2426 |
<td> &status.new; </td> |
| 2427 |
</tr> |
| 2428 |
|
| 2429 |
<tr> |
| 2430 |
<td> Lock if_vlan and inter-layer multicast address manipulation and synchronization in if_vlan. </td> |
| 2431 |
<td> &a.glebius;, &a.yar; </td> |
| 2432 |
<td> 19 October 2005 </td> |
| 2433 |
<td> &status.new; </td> |
| 2434 |
</tr> |
| 2435 |
|
| 2436 |
<tr> |
| 2437 |
<td> Further investigate locking in in_gif and in6_gif. </td> |
| 2438 |
<td> </td> |
| 2439 |
<td> 19 October 2005 </td> |
| 2440 |
<td> &status.new; </td> |
| 2441 |
</tr> |
| 2442 |
|
| 2443 |
<tr> |
| 2444 |
<td> FAST_IPSEC and KAME IPSEC's PF_KEY support likely needs an asynchronous dispatch to prevent socket lock ordering issues similar to what was done for PF_ROUTE. </td> |
| 2445 |
<td> &a.rwatson; </td> |
| 2446 |
<td> 19 October 2005 </td> |
| 2447 |
<td> &status.new; </td> |
| 2448 |
</tr> |
| 2449 |
|
| 2450 |
<tr> |
| 2451 |
<td> Investigate how to eliminate the use of ACCEPT_LOCK(), which currently prevents races in the tear-down of sockets. </td> |
| 2452 |
<td> &a.rwatson; </td> |
| 2453 |
<td> 19 October 2005 </td> |
| 2454 |
<td> &status.new; </td> |
| 2455 |
</tr> |
| 2456 |
|
| 2457 |
<tr> |
| 2458 |
<td> Fix SMP problems with netgraph restructuring. </td> |
| 2459 |
<td> &a.glebius; </td> |
| 2460 |
<td> 19 October 2005 </td> |
| 2461 |
<td> &status.new; </td> |
| 2462 |
</tr> |
| 2463 |
|
| 2464 |
<tr> |
| 2465 |
<td> Verify locking in netgraph nodes and improve where necessary. </td> |
| 2466 |
<td> &a.glebius; </td> |
| 2467 |
<td> 19 October 2005 </td> |
| 2468 |
<td> &status.new; </td> |
| 2469 |
</tr> |
| 2470 |
|
| 2471 |
<tr> |
| 2472 |
<td> More finely-grained locking for pf(4). </td> |
| 2473 |
<td> &a.glebius; </td> |
| 2474 |
<td> 19 October 2005 </td> |
| 2475 |
<td> &status.new; </td> |
| 2476 |
</tr> |
| 2477 |
|
| 2478 |
</table> |
| 2479 |
|
| 2480 |
<a name="issues"></a> |
| 2481 |
<h2>Known Issues (Historical)</h2> |
| 2482 |
|
| 2483 |
<table class="tblbasic"> |
| 2484 |
<tr class="heading"> |
| 2485 |
<th> Issue </th> |
| 2486 |
<th> Last updated </th> |
| 2487 |
<th> Status </th> |
| 2488 |
</tr> |
| 2489 |
|
| 2490 |
<tr> |
| 2491 |
<td> Idle processor time is not charged to the idle processes. </td> |
| 2492 |
<td> 20 September 2000 </td> |
| 2493 |
<td> &status.resolved; </td> |
| 2494 |
</tr> |
| 2495 |
|
| 2496 |
<tr> |
| 2497 |
<td> microuptime creeps backwards. </td> |
| 2498 |
<td> 4 October 2000 </td> |
| 2499 |
<td> &status.resolved; </td> |
| 2500 |
</tr> |
| 2501 |
|
| 2502 |
<tr> |
| 2503 |
<td> microuptime() went backwards </td> |
| 2504 |
<td> 4 October 2000 </td> |
| 2505 |
<td> &status.resolved; </td> |
| 2506 |
</tr> |
| 2507 |
|
| 2508 |
<tr> |
| 2509 |
<td> Process accounting is not accurate (the more CPUs, the |
| 2510 |
closer to correct it is). </td> |
| 2511 |
<td> 5 October 2000 </td> |
| 2512 |
<td> &status.resolved; </td> |
| 2513 |
</tr> |
| 2514 |
|
| 2515 |
<tr> |
| 2516 |
<td> M_DEVBUF is probably the wrong memory pool for interrupt |
| 2517 |
stuff and we should think about creating a new malloc pool for |
| 2518 |
that stuff. </td> |
| 2519 |
<td> 9 February 2001 </td> |
| 2520 |
<td> &status.resolved; </td> |
| 2521 |
</tr> |
| 2522 |
|
| 2523 |
<tr> |
| 2524 |
<td> PC card eject panics due to a race condition in the |
| 2525 |
interrupt thread code. </td> |
| 2526 |
<td> 15 March 2001 </td> |
| 2527 |
<td> &status.resolved; </td> |
| 2528 |
</tr> |
| 2529 |
|
| 2530 |
<tr> |
| 2531 |
<td> SMP x86 boxes are seeing NCPU * 100 clk interrupts and |
| 2532 |
NCPU * 128 rtc interrupts. </td> |
| 2533 |
<td> 15 May 2001 </td> |
| 2534 |
<td> &status.resolved; </td> |
| 2535 |
</tr> |
| 2536 |
|
| 2537 |
<tr> |
| 2538 |
<td> Witness will infinitely recurse when it acquires Giant after |
| 2539 |
sleeping with a sleepable lock. </td> |
| 2540 |
<td> 27 June 2001 </td> |
| 2541 |
<td> &status.resolved; </td> |
| 2542 |
</tr> |
| 2543 |
|
| 2544 |
<tr> |
| 2545 |
<td> Serial gdb does not work if boot_ddb and boot_gdb options |
| 2546 |
are specified. </td> |
| 2547 |
<td> 14 July 2002 </td> |
| 2548 |
<td> &status.resolved; </td> |
| 2549 |
</tr> |
| 2550 |
|
| 2551 |
<tr> |
| 2552 |
<td> Serial gdb does not work at 115200 baud. </td> |
| 2553 |
<td> 14 July 2002 </td> |
| 2554 |
<td> &status.resolved; </td> |
| 2555 |
</tr> |
| 2556 |
|
| 2557 |
<tr> |
| 2558 |
<td> Serial gdb never regains control once 'cont' has been |
| 2559 |
entered. </td> |
| 2560 |
<td> 14 July 2002 </td> |
| 2561 |
<td> &status.resolved; </td> |
| 2562 |
</tr> |
| 2563 |
|
| 2564 |
<!-- unresolved --> |
| 2565 |
|
| 2566 |
<tr> |
| 2567 |
<td> Profiling is broken. </td> |
| 2568 |
<td> 20 February 2001 </td> |
| 2569 |
<td> &status.unresolved; </td> |
| 2570 |
</tr> |
| 2571 |
|
| 2572 |
<!-- |
| 2573 |
<tr> |
| 2574 |
<td> </td> |
| 2575 |
<td> D M 2002 </td> |
| 2576 |
<td> &status.unresolved; </td> |
| 2577 |
</tr> |
| 2578 |
|
| 2579 |
--> |
| 2580 |
</table> |
| 2581 |
|
| 2582 |
<a name="news"></a> |
| 2583 |
<h2>News (Historical)</h2> |
| 2584 |
|
| 2585 |
<p>The remainder of this page is structured as a |
| 2586 |
reverse-chronological log.</p> |
| 2587 |
|
| 2588 |
<!-- |
| 2589 |
<b> D M 2002</b> |
| 2590 |
<ul> |
| 2591 |
<li> <p> </p> </li> |
| 2592 |
</ul> |
| 2593 |
--> |
| 2594 |
<b>28 August 2004</b> |
| 2595 |
<ul> |
| 2596 |
<li> |
147 |
<li> |
| 2597 |
<p>&a.rwatson; threw the switch to change the network stack to run |
148 |
<p> |
| 2598 |
without the Giant lock by default, permitting the network stack to be |
|
|
| 2599 |
run on multiple CPUs at a time, as well as to preempt and be |
| 2600 |
preempted by other code.</p> |
| 2601 |
</li> |
| 2602 |
</ul> |
| 2603 |
<b>30 October 2003</b> |
| 2604 |
<ul> |
| 2605 |
<li> |
| 2606 |
<p>&a.grog; submitted a |
| 2607 |
<a href="http://www.lemis.com/grog/SMPng/Singapore/">FreeBSD SMPng paper</a> |
| 2608 |
to the <a href="http://www.aeosc.org/">Asian Enterprise Open Source |
| 2609 |
Conference</a> in Singapore. The paper presents a historical view of |
| 2610 |
SMPng development through 2001, but omits discussion of more recent |
| 2611 |
progress on the SMPng project, such as substantial performance enhancements |
| 2612 |
resulting from extensive lock pushdown in the storage subsystem, VM |
| 2613 |
subsystem, and major IPC subsystems.</p> |
| 2614 |
</li> |
| 2615 |
</ul> |
| 2616 |
|
| 2617 |
<b>13 January 2002</b> |
| 2618 |
<ul> |
| 2619 |
<li> |
| 2620 |
<p>A <a href = |
| 2621 |
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=122646+0+archive/2002/freebsd-smp/20020113.freebsd-smp"> |
| 2622 |
status report</a> was sent to the -smp mailing list.</p> |
| 2623 |
</li> |
| 2624 |
</ul> |
| 2625 |
|
| 2626 |
<b>15 May 2001</b> |
| 2627 |
<ul> |
| 2628 |
<li> |
| 2629 |
<p>&a.grog; has made his <a href = |
| 2630 |
"http://www.lemis.com/~grog/SMPng/USENIX/">USENIX paper</a> available, |
| 2631 |
which he will present in Boston at the end of June.</p> |
| 2632 |
</li> |
| 2633 |
</ul> |
| 2634 |
|
| 2635 |
<b>22 March 2001</b> |
| 2636 |
<ul> |
| 2637 |
<li> |
| 2638 |
<p>A <a href = |
| 2639 |
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=41302+0+archive/2001/freebsd-smp/20010325.freebsd-smp"> |
| 2640 |
status report</a> was sent to the -smp mailing list.</p> |
| 2641 |
</li> |
| 2642 |
</ul> |
| 2643 |
|
| 2644 |
<b>5 March 2001</b> |
| 2645 |
<ul> |
| 2646 |
<li> |
| 2647 |
<p>A <a href = |
| 2648 |
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=0+0+archive/2001/freebsd-smp/20010311.freebsd-smp"> |
| 2649 |
status report</a> was sent to the -smp mailing list.</p> |
| 2650 |
</li> |
| 2651 |
</ul> |
| 2652 |
|
| 2653 |
<b>24 January 2001</b> |
| 2654 |
<ul> |
| 2655 |
<li> |
| 2656 |
<p>A <a href = |
| 2657 |
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=38125+0+archive/2001/freebsd-smp/20010128.freebsd-smp"> |
| 2658 |
status report</a> was sent to the -smp mailing list.</p> |
| 2659 |
</li> |
| 2660 |
</ul> |
| 2661 |
|
| 2662 |
<b>12 January 2001</b> |
| 2663 |
<ul> |
| 2664 |
<li> |
| 2665 |
<p>A <a href = |
| 2666 |
"http://www.FreeBSD.org/cgi/getmsg.cgi?fetch=45111+48223+/usr/local/www/db/text/2001/freebsd-smp/20010114.freebsd-smp"> |
| 2667 |
status report</a> was sent to the -smp mailing list.</p> |
| 2668 |
</li> |
| 2669 |
</ul> |
| 2670 |
|
| 2671 |
<b>11 October 2000</b> |
| 2672 |
<ul> |
| 2673 |
<li> |
| 2674 |
<p>&a.jhb; and &a.cp; came up with a preliminary list of |
| 2675 |
<a href="http://storm.uk.FreeBSD.org/~hiten/smp_synch_rules.html">rules |
| 2676 |
</a> that should be followed when working on kernel |
| 2677 |
synchronization.</p> |
| 2678 |
</li> |
| 2679 |
</ul> |
| 2680 |
|
| 2681 |
<b>6 September 2000</b> |
| 2682 |
<ul> |
| 2683 |
<li> |
| 2684 |
<p>The SMP code has been committed. All further work is being |
| 2685 |
done in cvs rather than with patches.</p> |
| 2686 |
</li> |
| 2687 |
|
| 2688 |
<li> |
| 2689 |
<p>An updated <a |
| 2690 |
href="http://people.FreeBSD.org/~jasone/smp/smpng_patch_9.7">patch</a> |
| 2691 |
is available for download. This patch is probably what will |
| 2692 |
actually get committed.</p> |
| 2693 |
</li> |
| 2694 |
</ul> |
| 2695 |
|
| 2696 |
<b>5 September 2000</b> |
| 2697 |
<ul> |
| 2698 |
<li> |
| 2699 |
<p>An updated <a |
| 2700 |
href="http://people.FreeBSD.org/~jasone/smp/smpng_patch_9.1">patch</a> |
| 2701 |
is available for download. This patch makes rtc a fast |
| 2702 |
interrupt, uses locked instructions for mutexes in MP kernels, |
| 2703 |
and corrects mtx_*() linkage within modules.</p> |
| 2704 |
</li> |
| 2705 |
</ul> |
| 2706 |
|
| 2707 |
<b>1 September 2000</b> |
| 2708 |
<ul> |
| 2709 |
<li> |
| 2710 |
<p>The code is working for the most part now on i386 (UP and MP). |
| 2711 |
Some additional coding is still necessary for the alpha, |
| 2712 |
which is being done now.</p> |
| 2713 |
</li> |
| 2714 |
</ul> |
| 2715 |
|
| 2716 |
<b>30 August 2000</b> |
| 2717 |
<ul> |
| 2718 |
<li> |
| 2719 |
<p>Updated patches for i386 and alpha are available <a href = |
| 2720 |
"http://people.FreeBSD.org/~jasone/smp/smpng_patch8.6">here</a>.</p> |
| 2721 |
</li> |
| 2722 |
</ul> |
| 2723 |
|
| 2724 |
<b>12 August 2000</b> |
| 2725 |
<ul> |
| 2726 |
<li> |
| 2727 |
<p>Updated patches for i386 are available <a href = |
| 2728 |
"http://www.FreeBSD.org/~jhb/patches/smpng.patch6">here</a>. |
| 2729 |
Process accounting still doesn't work correctly, but a number |
| 2730 |
of other improvements have been made.</p> |
| 2731 |
</li> |
| 2732 |
</ul> |
| 2733 |
|
| 2734 |
<b>3 August 2000</b> |
| 2735 |
<ul> |
| 2736 |
<li> |
| 2737 |
<p>Patches with functional heavy-weight threads for the i386 |
| 2738 |
platform are available <a href = |
| 2739 |
"http://people.FreeBSD.org/~grog/patches4.gz">here</a>. |
| 2740 |
There are a couple of minor issues with this patch set. |
| 2741 |
Specifically, process accounting doesn't work correctly.</p> |
| 2742 |
</li> |
| 2743 |
</ul> |
| 2744 |
|
| 2745 |
<b>6 July 2000</b> |
| 2746 |
<ul> |
| 2747 |
<li> |
| 2748 |
<p>Sheldon Hearn has prepared a mutex(9) man page based on the |
| 2749 |
BSD/OS one, which is available <a href = |
| 2750 |
"http://people.FreeBSD.org/~jake/mutex.9">here</a>.</p> |
| 2751 |
</li> |
| 2752 |
</ul> |
| 2753 |
|
| 2754 |
<b>5 July 2000</b> |
| 2755 |
<ul> |
| 2756 |
<li> |
| 2757 |
<p>&a.jake; put an updated patch <a href = |
| 2758 |
"http://people.FreeBSD.org/~jake/smpng.diff">here</a>.</p> |
| 2759 |
</li> |
| 2760 |
</ul> |
| 2761 |
|
| 2762 |
<b>3 July 2000</b> |
| 2763 |
<ul> |
| 2764 |
<li> |
| 2765 |
<p>&a.jake; has the BSD/OS lock code working now, and has |
| 2766 |
incorporated the pertinent portions of &a.dillon;'s patches |
| 2767 |
(idle processes, some of the schedlock changes, etc.). |
| 2768 |
His patch set is available <a href = |
| 2769 |
"http://people.FreeBSD.org/~jake/smpng2.tar">here</a>.</p> |
| 2770 |
</li> |
| 2771 |
</ul> |
| 2772 |
|
| 2773 |
<b>26 June 2000</b> |
| 2774 |
<ul> |
| 2775 |
<li> |
| 2776 |
<p>&a.cp; has provided the PostScript versions of his |
| 2777 |
presentation slides for the |
| 2778 |
<a href="http://people.FreeBSD.org/~jasone/smp/smp_meeting_slides_1.ps">first day</a> and |
| 2779 |
<a href="http://people.FreeBSD.org/~jasone/smp/smp_meeting_slides_2.ps">second day</a> of the |
| 2780 |
SMP meeting.</p> |
| 2781 |
</li> |
| 2782 |
</ul> |
| 2783 |
|
| 2784 |
<b>25 June 2000</b> |
| 2785 |
<ul> |
| 2786 |
<li> |
| 2787 |
<p>Here's a copy of the |
| 2788 |
<a href="http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=226208+0+archive/2000/freebsd-smp/20000625.freebsd-smp">SMP meeting summary</a> |
| 2789 |
that was posted to the -smp mailing list.</p> |
| 2790 |
</li> |
| 2791 |
</ul> |
| 2792 |
|
| 2793 |
<b>19 June 2000</b> |
| 2794 |
<ul> |
| 2795 |
<li> |
| 2796 |
<p>Here's a copy of the |
| 2797 |
<a href="http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=154884+0+archive/2000/freebsd-current/20000625.freebsd-current">SMP project announcement</a> |
149 |
<a href="http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=154884+0+archive/2000/freebsd-current/20000625.freebsd-current">SMP project announcement</a> |
| 2798 |
that was posted to the -current mailing list.</p> |
150 |
that was posted to the -current mailing list. |
|
|
151 |
</p> |
| 2799 |
</li> |
152 |
</li> |
| 2800 |
</ul> |
153 |
</ul> |