1 .. SPDX-License-Identifier: (GPL-2.0+ OR CC-BY
2 .. See the bottom of this file for additional
3
4 Handling regressions
5 ++++++++++++++++++++
6
7 *We don't cause regressions* -- this document
8 Linux kernel development" means in practice fo
9 Documentation/admin-guide/reporting-regression
10 user's point of view; if you never read that t
11 before continuing here.
12
13 The important bits (aka "The TL;DR")
14 ====================================
15
16 #. Ensure subscribers of the `regression maili
17 (regressions@lists.linux.dev) quickly becom
18 report:
19
20 * When receiving a mailed report that did
21 loop by immediately sending at least a b
22 CCed.
23
24 * Forward or bounce any reports submitted
25
26 #. Make the Linux kernel regression tracking b
27 is optional, but recommended):
28
29 * For mailed reports, check if the reporte
30 introduced: v5.13..v5.14-rc1``. If not,
31 list in CC) containing a paragraph like
32 when the issue started to happen::
33
34 #regzbot ^introduced: 1f2e3d4c5b6a
35
36 * When forwarding reports from a bug track
37 above), include a paragraph like the fol
38
39 #regzbot introduced: v5.13..v5.14-rc1
40 #regzbot from: Some N. Ice Human <some.h
41 #regzbot monitor: http://some.bugtracke
42
43 #. When submitting fixes for regressions, add
44 description pointing to all places where th
45 mandated by Documentation/process/submittin
46 :ref:`Documentation/process/5.Posting.rst <
47 only fixing part of the issue that caused t
48 "Link:" tags instead. regzbot currently mak
49 two.
50
51 #. Try to fix regressions quickly once the cul
52 for most regressions should be merged withi
53 resolved within two or three days.
54
55
56 All the details on Linux kernel regressions re
57 ==============================================
58
59
60 The important basics in more detail
61 -----------------------------------
62
63
64 What to do when receiving regression reports
65 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
66
67 Ensure the Linux kernel's regression tracker a
68 `regression mailing list <https://lore.kernel.
69 (regressions@lists.linux.dev) become aware of
70
71 * When you receive a report by mail that did
72 it into the loop by sending at least a brie
73 try to ensure it gets CCed again in case yo
74 the list.
75
76 * If a report submitted in a bug tracker hits
77 to the list. Consider checking the list arc
78 already forwarded the report as instructed
79 Documentation/admin-guide/reporting-issues.
80
81 When doing either, consider making the Linux k
82 "regzbot" immediately start tracking the issue
83
84 * For mailed reports, check if the reporter i
85 ``#regzbot introduced: 1f2e3d4c5b6a``. If n
86 regressions list in CC) with a paragraph li
87
88 #regzbot ^introduced: v5.13..v5.14-rc1
89
90 This tells regzbot the version range in whi
91 you can specify a range using commit-ids as
92 in case the reporter bisected the culprit.
93
94 Note the caret (^) before the "introduced":
95 parent mail (the one you reply to) as the i
96 you want to see tracked; that's important,
97 for patches with "Closes:" tags pointing to
98 lore.kernel.org.
99
100 * When forwarding a regression reported to a
101 with these regzbot commands::
102
103 #regzbot introduced: 1f2e3d4c5b6a
104 #regzbot from: Some N. Ice Human <some.h
105 #regzbot monitor: http://some.bugtracke
106
107 Regzbot will then automatically associate p
108 contain "Closes:" tags pointing to your mai
109
110 What's important when fixing regressions
111 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
112
113 You don't need to do anything special when sub
114 remember to do what Documentation/process/subm
115 :ref:`Documentation/process/5.Posting.rst <dev
116 Documentation/process/stable-kernel-rules.rst
117
118 * Point to all places where the issue was rep
119
120 Closes: https://lore.kernel.org/r/30th.
121 Closes: https://bugzilla.kernel.org/sho
122
123 If you are only fixing part of the issue, y
124 described in the first document mentioned a
125 both of these equivalently and considers th
126
127 * Add a "Fixes:" tag to specify the commit ca
128
129 * If the culprit was merged in an earlier dev
130 the fix for backporting using the ``Cc: sta
131
132 All this is expected from you and important wh
133 these tags are of great value for everyone (yo
134 into the issue weeks, months, or years later.
135 tools and scripts used by other kernel develop
136 these tools is regzbot, which heavily relies o
137 reports for regression with changes resolving
138
139 Expectations and best practices for fixing reg
140 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
141
142 As a Linux kernel developer, you are expected
143 situations where a regression caused by a rece
144 only these options:
145
146 * Run a kernel with a regression that impacts
147
148 * Switch to an older or newer kernel series.
149
150 * Continue running an outdated and thus poten
151 than three weeks after the regression's cul
152 should be less than two. And it ought to be
153 severe or affects many users -- either in g
154 environments.
155
156 How to realize that in practice depends on var
157 rules of thumb as a guide.
158
159 In general:
160
161 * Prioritize work on regressions over all oth
162 latter concerns a severe issue (e.g. acute
163 bricked hardware, ...).
164
165 * Expedite fixing mainline regressions that r
166 mainline, stable, or longterm release (eith
167
168 * Do not consider regressions from the curren
169 till the end of the cycle, as the issue mig
170 CI systems from testing mainline now or gen
171
172 * Work with the required care to avoid additi
173 resolving an issue then might take longer t
174
175 On timing once the culprit of a regression is
176
177 * Aim to mainline a fix within two or three d
178 bothering many users -- either in general o
179 particular hardware environment, distributi
180
181 * Aim to mainline a fix by Sunday after the n
182 into a recent mainline, stable, or longterm
183 backport); if the culprit became known earl
184 resolve, try to mainline the fix within the
185
186 * For other regressions, aim to mainline fixe
187 within the next three weeks. One or two Sun
188 regression is something people can live wit
189 mild performance regression.
190
191 * It's strongly discouraged to delay mainlini
192 merge window, except when the fix is extrao
193 culprit was mainlined more than a year ago.
194
195 On procedure:
196
197 * Always consider reverting the culprit, as i
198 dangerous way to fix a regression. Don't wo
199 variant later: that should be straight-forw
200 through review once already.
201
202 * Try to resolve any regressions introduced i
203 twelve months before the current developmen
204 regressions to be handled like those from t
205 bears unusual risks.
206
207 * Consider CCing Linus on discussions or patc
208 tangly. Do the same in precarious or urgent
209 subsystem maintainer might be unavailable.
210 know such a regression made it into a mainl
211
212 * For urgent regressions, consider asking Lin
213 from the mailing list: he is totally fine w
214 fixes. Ideally though such requests should
215 subsystem maintainers or come directly from
216
217 * In case you are unsure if a fix is worth th
218 a new mainline release, send Linus a mail w
219 CC; in it, summarize the situation while as
220 the fix straight from the list. He then him
221 needed even postpone the release. Such requ
222 in accordance with the subsystem maintainer
223
224 Regarding stable and longterm kernels:
225
226 * You are free to leave regressions to the st
227 time occurred with mainline or were fixed t
228
229 * If a regression made it into a proper mainl
230 twelve months, ensure to tag the fix with "
231 "Fixes:" tag alone does not guarantee a bac
232 in case you know the culprit was backported
233
234 * When receiving reports about regressions in
235 series, please evaluate at least briefly if
236 mainline as well -- and if that seems likel
237 doubt, ask the reporter to check mainline.
238
239 * Whenever you want to swiftly resolve a regr
240 into a proper mainline, stable, or longterm
241 mainline; when appropriate thus involve Lin
242 above). That's because the stable team norm
243 any changes that cause the same problems in
244
245 * In case of urgent regression fixes you migh
246 backporting by dropping the stable team a n
247 this is especially advisable during merge w
248 the fix otherwise might land at the end of
249
250 On patch flow:
251
252 * Developers, when trying to reach the time p
253 to account for the time it takes to get fix
254 Linus, ideally with them being in linux-nex
255 fix is urgent, make it obvious to ensure ot
256
257 * Reviewers, you are kindly asked to assist d
258 periods mentioned above by reviewing regres
259
260 * Subsystem maintainers, you likewise are enc
261 of regression fixes. Thus evaluate if skipp
262 the particular fix. Also consider sending g
263 usual when needed. And try to avoid holding
264 weekends -- especially when the fix is mark
265
266
267 More aspects regarding regressions developers
268 ----------------------------------------------
269
270
271 How to deal with changes where a risk of regre
272 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
273
274 Evaluate how big the risk of regressions is, f
275 search in Linux distributions and Git forges.
276 developers or projects likely to be affected t
277 proposed change; if problems surface, maybe so
278 can be found.
279
280 If the risk of regressions in the end seems to
281 with the change, but let all involved parties
282 sure your patch description makes this aspect
283 merged, tell the Linux kernel's regression tra
284 list about the risk, so everyone has the chang
285 trickle in. Depending on the risk, you also mi
286 maintainer to mention the issue in his mainlin
287
288 What else is there to known about regressions?
289 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
290
291 Check out Documentation/admin-guide/reporting-
292 of other aspects you want might want to be awa
293
294 * the purpose of the "no regressions" rule
295
296 * what issues actually qualify as regression
297
298 * who's in charge for finding the root cause
299
300 * how to handle tricky situations, e.g. when
301 security fix or when fixing a regression mi
302
303 Whom to ask for advice when it comes to regres
304 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
305
306 Send a mail to the regressions mailing list (r
307 CCing the Linux kernel's regression tracker (r
308 issue might better be dealt with in private, f
309
310
311 More about regression tracking and regzbot
312 ------------------------------------------
313
314
315 Why the Linux kernel has a regression tracker,
316 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
317
318 Rules like "no regressions" need someone to en
319 they are broken either accidentally or on purp
320 true for the Linux kernel as well. That's why
321 keep an eye on things as the Linux kernel's re
322 occasionally helped by other people. Neither o
323 that's why regression tracking is done on a be
324
325 Earlier attempts to manually track regressions
326 frustrating work, which is why they were aband
327 this from happening again, Thorsten developed
328 with the long term goal to automate regression
329 everyone involved.
330
331 How does regression tracking work with regzbot
332 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
333
334 The bot watches for replies to reports of trac
335 it's looking out for posted or committed patch
336 with "Closes:" tags; replies to such patch pos
337 Combined this data provides good insights into
338 process.
339
340 Regzbot tries to do its job with as little ove
341 reporters and developers. In fact, only report
342 duty: they need to tell regzbot about the regr
343 introduced`` command outlined above; if they d
344 take care of that using ``#regzbot ^introduced
345
346 For developers there normally is no extra work
347 sure to do something that was expected long be
348 links to the patch description pointing to all
349
350 Do I have to use regzbot?
351 ~~~~~~~~~~~~~~~~~~~~~~~~~
352
353 It's in the interest of everyone if you do, as
354 Torvalds partly rely on regzbot's tracking in
355 deciding to release a new version or extend th
356 need to be aware of all unfixed regression; to
357 into the weekly reports sent by regzbot.
358
359 Do I have to tell regzbot about every regressi
360 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
361
362 Ideally yes: we are all humans and easily forg
363 important unexpectedly comes up -- for example
364 kernel or something in real life that's keepin
365 while. Hence, it's best to tell regzbot about
366 immediately write a fix and commit it to a tre
367 kernel series.
368
369 How to see which regressions regzbot tracks cu
370 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
371
372 Check `regzbot's web-interface <https://linux-
373 for the latest info; alternatively, `search fo
374 <https://lore.kernel.org/lkml/?q=%22Linux+regr
375 which regzbot normally sends out once a week o
376 few hours before Linus usually publishes new (
377
378 What places is regzbot monitoring?
379 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
380
381 Regzbot is watching the most important Linux m
382 repositories of linux-next, mainline, and stab
383
384 What kind of issues are supposed to be tracked
385 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
386
387 The bot is meant to track regressions, hence p
388 regular issues. But it's okay for the Linux ke
389 use regzbot to track severe issues, like repor
390 or internal errors (Panic, Oops, BUG(), warnin
391
392 Can I add regressions found by CI systems to r
393 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
394
395 Feel free to do so, if the particular regressi
396 use cases and thus might be noticed by users;
397 regzbot for theoretical regressions unlikely t
398 usage.
399
400 How to interact with regzbot?
401 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
402
403 By using a 'regzbot command' in a direct or in
404 regression report. These commands need to be i
405 need to be separated from the rest of the mail
406
407 One such command is ``#regzbot introduced: <ve
408 regzbot consider your mail as a regressions re
409 already described above; ``#regzbot ^introduce
410 such command, which makes regzbot consider the
411 regression which it starts to track.
412
413 Once one of those two commands has been utiliz
414 used in direct or indirect replies to the repo
415 of the `introduced` commands or in replies to
416 or itself is a reply to that mail:
417
418 * Set or update the title::
419
420 #regzbot title: foo
421
422 * Monitor a discussion or bugzilla.kernel.org
423 the issue or a fix are discussed -- for exa
424 the regression::
425
426 #regzbot monitor: https://lore.kernel.o
427
428 Monitoring only works for lore.kernel.org a
429 will consider all messages in that thread o
430 process.
431
432 * Point to a place with further details of in
433 or a ticket in a bug tracker that are sligh
434 topic::
435
436 #regzbot link: https://bugzilla.kernel.
437
438 * Mark a regression as fixed by a commit that
439 landed::
440
441 #regzbot fix: 1f2e3d4c5d
442
443 * Mark a regression as a duplicate of another
444
445 #regzbot dup-of: https://lore.kernel.or
446
447 * Mark a regression as invalid::
448
449 #regzbot invalid: wasn't a regression,
450
451 Is there more to tell about regzbot and its co
452 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
453
454 More detailed and up-to-date information about
455 kernel's regression tracking bot can be found
456 `project page <https://gitlab.com/knurd42/regz
457 contains a `getting started guide <https://git
458 and `reference documentation <https://gitlab.c
459 which both cover more details than the above s
460
461 Quotes from Linus about regression
462 ----------------------------------
463
464 Find below a few real life examples of how Lin
465 be handled:
466
467 * From `2017-10-26 (1/2)
468 <https://lore.kernel.org/lkml/CA+55aFwiiQYJ+
469
470 If you break existing user space setups
471
472 It's not ok to say "but we'll fix the u
473
474 Really. NOT OK.
475
476 [...]
477
478 The first rule is:
479
480 - we don't cause regressions
481
482 and the corollary is that when regressi
483 them and fix them, instead of blaming u
484
485 The fact that you have apparently been
486 three weeks means that I will revert, a
487 requests until the people involved unde
488 is done.
489
490 * From `2017-10-26 (2/2)
491 <https://lore.kernel.org/lkml/CA+55aFxW7NMAM
492
493 People should basically always feel lik
494 and simply not have to worry about it.
495
496 I refuse to introduce "you can only upd
497 update that other program" kind of limi
498 work for you, the rule is that it conti
499
500 There have been exceptions, but they ar
501 generally have some major and fundament
502 that were basically entirely unavoidabl
503 avoid them. Maybe we can't practically
504 after it is decades old and nobody uses
505 more. Maybe there's a serious security
506 and people actually depended on that fu
507 there was some fundamental other breaka
508 flag day for very core and fundamental
509
510 And notice that this is very much about
511
512 Behavioral changes happen, and maybe we
513 feature any more. There's a number of f
514 are printed out as zeroes, simply becau
515 the kernel any more, or because showing
516 an information leak). But the numbers g
517 the code that used to parse the fields
518 see everything they used to see, and so
519 but things still _work_, even if they m
520 (or no longer relevant) information.
521
522 But if something actually breaks, then
523 reverted. And it gets fixed in the *ker
524 your user space then". It was a kernel
525 problem, it needs to be the kernel that
526 have a "upgrade in place" model. We don
527 user space".
528
529 And I seriously will refuse to take cod
530 understand and honor this very simple r
531
532 This rule is also not going to change.
533
534 And yes, I realize that the kernel is "
535 proud of it.
536
537 I have seen, and can point to, lots of
538 break that use case in order to make pr
539 undocumented behavior, it sucks to be y
540 do what you want to do, and you have to
541 way", and I simply don't think that's a
542 alpha releases that have experimental u
543 up for. The kernel hasn't been in that
544 decades.
545
546 We do API breakage _inside_ the kernel
547 internal problems by saying "you now ne
548 about internal kernel API's, and the pe
549 obviously have to fix up all the in-ker
550 can say "I now broke the API you used,
551 up". Whoever broke something gets to fi
552
553 And we simply do not break user space.
554
555 * From `2020-05-21
556 <https://lore.kernel.org/all/CAHk-=wiVi7mSrs
557
558 The rules about regressions have never
559 documented behavior, or where the code
560
561 The rules about regressions are always
562
563 Users are literally the _only_ thing th
564
565 No amount of "you shouldn't have used t
566 undefined, it's your own fault your app
567 simply because of a kernel bug" is at a
568
569 Now, reality is never entirely black-an
570 like "serious security issue" etc that
571 that may break user space. But even the
572 really have other options that would al
573
574 And obviously, if users take years to e
575 broke, or if we have sane ways to work
576 doesn't make for too much trouble for u
577 handful of users, and they can use a ke
578 around it" kind of things) we've also b
579
580 But no, "that was documented to be brok
581 code was in staging or because the man-
582 irrelevant. If staging code is so usefu
583 that means that it's basically regular
584 "please clean this up".
585
586 The other side of the coin is that peop
587 stability" are entirely wrong. API's do
588 any changes to an API you like - as lon
589
590 Again, the regression rule is not about
591 API's, and not about the phase of the m
592
593 It's entirely about "we caused problems
594
595 * From `2017-11-05
596 <https://lore.kernel.org/all/CA+55aFzUvbGjD8
597
598 And our regression rule has never been
599 That would mean that we could never mak
600
601 For example, we do things like add new
602 time, which we then sometimes even add
603 directory.
604
605 So clearly behavior changes all the tim
606 regression per se.
607
608 The rule for a regression for the kerne
609 workflow breaks. Not some test. Not a "
610 X, now I can't".
611
612 * From `2018-08-03
613 <https://lore.kernel.org/all/CA+55aFwWZX=CXm
614
615 YOU ARE MISSING THE #1 KERNEL RULE.
616
617 We do not regress, and we do not regres
618
619 And the reason you state for your opini
620 are wrong.
621
622 Your "good reasons" are pure and utter
623
624 The whole point of "we do not regress"
625 the kernel and never have to worry abou
626
627 > Kernel had a bug which has been fixed
628
629 That is *ENTIRELY* immaterial.
630
631 Guys, whether something was buggy or no
632
633 Why?
634
635 Bugs happen. That's a fact of life. Arg
636 something because we were fixing a bug"
637 tens of bugs every single day, thinking
638 we can break something is simply NOT TR
639
640 So bugs simply aren't even relevant to
641 they get found, they get fixed, and it
642 break users".
643
644 Because the only thing that matters IS
645
646 How hard is that to understand?
647
648 Anybody who uses "but it was buggy" as
649 the point. As far as the USER was conce
650 worked for him/her.
651
652 Maybe it worked *because* the user had
653 maybe it worked because the user didn't
654 matter. It worked for the user.
655
656 Breaking a user workflow for a "bug" is
657 for breakage you can imagine.
658
659 It's basically saying "I took something
660 but now it's better". Do you not see ho
661 is?
662
663 And without users, your program is not
664 piece of code that you might as well th
665
666 Seriously. This is *why* the #1 rule fo
667 don't break users". Because "I fixed a
668 ARGUMENT if that bug fix broke a user s
669 MUCH BIGGER bug by "fixing" something t
670 even care about.
671
672 And dammit, we upgrade the kernel ALL T
673 other programs at all. It is absolutely
674 and dependencies are horribly bad.
675
676 And it is also required simply because
677 upgrade random other tools that I don't
678 the kernel, and I want any of my users
679 time.
680
681 So no. Your rule is COMPLETELY wrong. I
682 without upgrading some other random bin
683
684 * From `2021-06-05
685 <https://lore.kernel.org/all/CAHk-=wiUVqHN76
686
687 THERE ARE NO VALID ARGUMENTS FOR REGRES
688
689 Honestly, security people need to under
690 a success case of security. It's a fail
691
692 Yes, "not working" may be secure. But s
693
694 * From `2011-05-06 (1/3)
695 <https://lore.kernel.org/all/BANLkTim9YvResB
696
697 Binary compatibility is more important.
698
699 And if binaries don't use the interface
700 parse it wrongly - see the fairly recen
701 /proc/self/mountinfo), then it's a regr
702
703 And regressions get reverted, unless th
704 similar that makes us go "Oh Gods, we r
705
706 I don't understand why this simple logi
707 developers to understand. Reality matte
708 NOT AT ALL.
709
710 If you made an interface that can be us
711 interface description, then we're stuck
712 simply doesn't matter.
713
714 You could help fix the tools, and try t
715 issues that way. There aren't that many
716
717 From `2011-05-06 (2/3)
718 <https://lore.kernel.org/all/BANLkTi=KVXjKR8
719
720 it's clearly NOT an internal tracepoint
721 used by powertop.
722
723 From `2011-05-06 (3/3)
724 <https://lore.kernel.org/all/BANLkTinazaXRdG
725
726 We have programs that use that ABI and
727
728 * From `2012-07-06 <https://lore.kernel.org/al
729
730 > Now this got me wondering if Debian _
731 > standard distro userspace.
732
733 Oh, if the kernel breaks some standard
734 of people run Debian unstable
735
736 * From `2019-09-15
737 <https://lore.kernel.org/lkml/CAHk-=wiP4K8DR
738
739 One _particularly_ last-minute revert i
740 the version change itself) done just be
741 it's very annoying, it's perhaps also i
742
743 What's instructive about it is that I r
744 actually buggy. In fact, it was doing e
745 and did it very well. In fact it did it
746 improved IO patterns it caused then end
747 regression due to a real bug in a compl
748
749 The actual details of that regression a
750 revert out as instructive, though. It's
751 example of what counts as a regression,
752 regressions" kernel rule means. The rev
753 API's, and it didn't introduce any new
754 another problem, and as such caused a k
755 user. So it got reverted.
756
757 The point here being that we revert bas
758 not based on some "it changes the ABI"
759 The problem was really pre-existing, an
760 trigger before. The better IO patterns
761 happened to expose an old bug, and peop
762 previously benign behavior of that old
763
764 And never fear, we'll re-introduce the
765 patterns once we've decided just how to
766 bad interaction with an interface that
767 to rely on incidental behavior for befo
768 to hash through how to do that (there a
769 patches by three different developers b
770 be more coming...). In the meantime, I
771 the problem to users for this release,
772 re-introduced (perhaps even backported
773 consensus about the issue it exposed.
774
775 Take-away from the whole thing: it's no
776 kernel-userspace ABI, or fix a bug, or
777 "should never have worked in the first
778 something breaks existing users' workfl
779
780 Anyway, that was my little aside on the
781 it's that "first rule of kernel program
782 worth just bringing it up every once in
783
784 ..
785 end-of-content
786 ..
787 This text is available under GPL-2.0+ or CC
788 of the file. If you want to distribute this
789 please use "The Linux kernel developers" fo
790 this as source:
791 https://git.kernel.org/pub/scm/linux/kernel
792 ..
793 Note: Only the content of this RST file as
794 is available under CC-BY-4.0, as versions o
795 (for example by the kernel's build system)
796 files which use a more restrictive license.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.