1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 The tip tree handbook 3 The tip tree handbook
4 ===================== 4 =====================
5 5
6 What is the tip tree? 6 What is the tip tree?
7 --------------------- 7 ---------------------
8 8
9 The tip tree is a collection of several subsys 9 The tip tree is a collection of several subsystems and areas of
10 development. The tip tree is both a direct dev 10 development. The tip tree is both a direct development tree and a
11 aggregation tree for several sub-maintainer tr 11 aggregation tree for several sub-maintainer trees. The tip tree gitweb URL
12 is: https://git.kernel.org/pub/scm/linux/kerne 12 is: https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git
13 13
14 The tip tree contains the following subsystems 14 The tip tree contains the following subsystems:
15 15
16 - **x86 architecture** 16 - **x86 architecture**
17 17
18 The x86 architecture development takes pl 18 The x86 architecture development takes place in the tip tree except
19 for the x86 KVM and XEN specific parts wh 19 for the x86 KVM and XEN specific parts which are maintained in the
20 corresponding subsystems and routed direc 20 corresponding subsystems and routed directly to mainline from
21 there. It's still good practice to Cc the 21 there. It's still good practice to Cc the x86 maintainers on
22 x86-specific KVM and XEN patches. 22 x86-specific KVM and XEN patches.
23 23
24 Some x86 subsystems have their own mainta 24 Some x86 subsystems have their own maintainers in addition to the
25 overall x86 maintainers. Please Cc the o 25 overall x86 maintainers. Please Cc the overall x86 maintainers on
26 patches touching files in arch/x86 even w 26 patches touching files in arch/x86 even when they are not called out
27 by the MAINTAINER file. 27 by the MAINTAINER file.
28 28
29 Note, that ``x86@kernel.org`` is not a ma 29 Note, that ``x86@kernel.org`` is not a mailing list. It is merely a
30 mail alias which distributes mails to the 30 mail alias which distributes mails to the x86 top-level maintainer
31 team. Please always Cc the Linux Kernel m 31 team. Please always Cc the Linux Kernel mailing list (LKML)
32 ``linux-kernel@vger.kernel.org``, otherwi 32 ``linux-kernel@vger.kernel.org``, otherwise your mail ends up only in
33 the private inboxes of the maintainers. 33 the private inboxes of the maintainers.
34 34
35 - **Scheduler** 35 - **Scheduler**
36 36
37 Scheduler development takes place in the 37 Scheduler development takes place in the -tip tree, in the
38 sched/core branch - with occasional sub-t 38 sched/core branch - with occasional sub-topic trees for
39 work-in-progress patch-sets. 39 work-in-progress patch-sets.
40 40
41 - **Locking and atomics** 41 - **Locking and atomics**
42 42
43 Locking development (including atomics an 43 Locking development (including atomics and other synchronization
44 primitives that are connected to locking) 44 primitives that are connected to locking) takes place in the -tip
45 tree, in the locking/core branch - with o 45 tree, in the locking/core branch - with occasional sub-topic trees
46 for work-in-progress patch-sets. 46 for work-in-progress patch-sets.
47 47
48 - **Generic interrupt subsystem and interru 48 - **Generic interrupt subsystem and interrupt chip drivers**:
49 49
50 - interrupt core development happens in t 50 - interrupt core development happens in the irq/core branch
51 51
52 - interrupt chip driver development also 52 - interrupt chip driver development also happens in the irq/core
53 branch, but the patches are usually app 53 branch, but the patches are usually applied in a separate maintainer
54 tree and then aggregated into irq/core 54 tree and then aggregated into irq/core
55 55
56 - **Time, timers, timekeeping, NOHZ and rel 56 - **Time, timers, timekeeping, NOHZ and related chip drivers**:
57 57
58 - timekeeping, clocksource core, NTP and 58 - timekeeping, clocksource core, NTP and alarmtimer development
59 happens in the timers/core branch, but 59 happens in the timers/core branch, but patches are usually applied in
60 a separate maintainer tree and then agg 60 a separate maintainer tree and then aggregated into timers/core
61 61
62 - clocksource/event driver development ha 62 - clocksource/event driver development happens in the timers/core
63 branch, but patches are mostly applied 63 branch, but patches are mostly applied in a separate maintainer tree
64 and then aggregated into timers/core 64 and then aggregated into timers/core
65 65
66 - **Performance counters core, architecture 66 - **Performance counters core, architecture support and tooling**:
67 67
68 - perf core and architecture support deve 68 - perf core and architecture support development happens in the
69 perf/core branch 69 perf/core branch
70 70
71 - perf tooling development happens in the 71 - perf tooling development happens in the perf tools maintainer
72 tree and is aggregated into the tip tre 72 tree and is aggregated into the tip tree.
73 73
74 - **CPU hotplug core** 74 - **CPU hotplug core**
75 75
76 - **RAS core** 76 - **RAS core**
77 77
78 Mostly x86-specific RAS patches are colle 78 Mostly x86-specific RAS patches are collected in the tip ras/core
79 branch. 79 branch.
80 80
81 - **EFI core** 81 - **EFI core**
82 82
83 EFI development in the efi git tree. The 83 EFI development in the efi git tree. The collected patches are
84 aggregated in the tip efi/core branch. 84 aggregated in the tip efi/core branch.
85 85
86 - **RCU** 86 - **RCU**
87 87
88 RCU development happens in the linux-rcu 88 RCU development happens in the linux-rcu tree. The resulting changes
89 are aggregated into the tip core/rcu bran 89 are aggregated into the tip core/rcu branch.
90 90
91 - **Various core code components**: 91 - **Various core code components**:
92 92
93 - debugobjects 93 - debugobjects
94 94
95 - objtool 95 - objtool
96 96
97 - random bits and pieces 97 - random bits and pieces
98 98
99 99
100 Patch submission notes 100 Patch submission notes
101 ---------------------- 101 ----------------------
102 102
103 Selecting the tree/branch 103 Selecting the tree/branch
104 ^^^^^^^^^^^^^^^^^^^^^^^^^ 104 ^^^^^^^^^^^^^^^^^^^^^^^^^
105 105
106 In general, development against the head of th 106 In general, development against the head of the tip tree master branch is
107 fine, but for the subsystems which are maintai 107 fine, but for the subsystems which are maintained separately, have their
108 own git tree and are only aggregated into the 108 own git tree and are only aggregated into the tip tree, development should
109 take place against the relevant subsystem tree 109 take place against the relevant subsystem tree or branch.
110 110
111 Bug fixes which target mainline should always 111 Bug fixes which target mainline should always be applicable against the
112 mainline kernel tree. Potential conflicts agai 112 mainline kernel tree. Potential conflicts against changes which are already
113 queued in the tip tree are handled by the main 113 queued in the tip tree are handled by the maintainers.
114 114
115 Patch subject 115 Patch subject
116 ^^^^^^^^^^^^^ 116 ^^^^^^^^^^^^^
117 117
118 The tip tree preferred format for patch subjec 118 The tip tree preferred format for patch subject prefixes is
119 'subsys/component:', e.g. 'x86/apic:', 'x86/mm 119 'subsys/component:', e.g. 'x86/apic:', 'x86/mm/fault:', 'sched/fair:',
120 'genirq/core:'. Please do not use file names o 120 'genirq/core:'. Please do not use file names or complete file paths as
121 prefix. 'git log path/to/file' should give you 121 prefix. 'git log path/to/file' should give you a reasonable hint in most
122 cases. 122 cases.
123 123
124 The condensed patch description in the subject 124 The condensed patch description in the subject line should start with a
125 uppercase letter and should be written in impe 125 uppercase letter and should be written in imperative tone.
126 126
127 127
128 Changelog 128 Changelog
129 ^^^^^^^^^ 129 ^^^^^^^^^
130 130
131 The general rules about changelogs in the :ref 131 The general rules about changelogs in the :ref:`Submitting patches guide
132 <describe_changes>`, apply. 132 <describe_changes>`, apply.
133 133
134 The tip tree maintainers set value on followin 134 The tip tree maintainers set value on following these rules, especially on
135 the request to write changelogs in imperative 135 the request to write changelogs in imperative mood and not impersonating
136 code or the execution of it. This is not just 136 code or the execution of it. This is not just a whim of the
137 maintainers. Changelogs written in abstract wo 137 maintainers. Changelogs written in abstract words are more precise and
138 tend to be less confusing than those written i 138 tend to be less confusing than those written in the form of novels.
139 139
140 It's also useful to structure the changelog in 140 It's also useful to structure the changelog into several paragraphs and not
141 lump everything together into a single one. A 141 lump everything together into a single one. A good structure is to explain
142 the context, the problem and the solution in s 142 the context, the problem and the solution in separate paragraphs and this
143 order. 143 order.
144 144
145 Examples for illustration: 145 Examples for illustration:
146 146
147 Example 1:: 147 Example 1::
148 148
149 x86/intel_rdt/mbm: Fix MBM overflow handle 149 x86/intel_rdt/mbm: Fix MBM overflow handler during hot cpu
150 150
151 When a CPU is dying, we cancel the worker 151 When a CPU is dying, we cancel the worker and schedule a new worker on a
152 different CPU on the same domain. But if t 152 different CPU on the same domain. But if the timer is already about to
153 expire (say 0.99s) then we essentially dou 153 expire (say 0.99s) then we essentially double the interval.
154 154
155 We modify the hot cpu handling to cancel t 155 We modify the hot cpu handling to cancel the delayed work on the dying
156 cpu and run the worker immediately on a di 156 cpu and run the worker immediately on a different cpu in same domain. We
157 do not flush the worker because the MBM ov !! 157 donot flush the worker because the MBM overflow worker reschedules the
158 worker on same CPU and scans the domain->c 158 worker on same CPU and scans the domain->cpu_mask to get the domain
159 pointer. 159 pointer.
160 160
161 Improved version:: 161 Improved version::
162 162
163 x86/intel_rdt/mbm: Fix MBM overflow handle 163 x86/intel_rdt/mbm: Fix MBM overflow handler during CPU hotplug
164 164
165 When a CPU is dying, the overflow worker i 165 When a CPU is dying, the overflow worker is canceled and rescheduled on a
166 different CPU in the same domain. But if t 166 different CPU in the same domain. But if the timer is already about to
167 expire this essentially doubles the interv 167 expire this essentially doubles the interval which might result in a non
168 detected overflow. 168 detected overflow.
169 169
170 Cancel the overflow worker and reschedule 170 Cancel the overflow worker and reschedule it immediately on a different CPU
171 in the same domain. The work could be flus 171 in the same domain. The work could be flushed as well, but that would
172 reschedule it on the same CPU. 172 reschedule it on the same CPU.
173 173
174 Example 2:: 174 Example 2::
175 175
176 time: POSIX CPU timers: Ensure that variab 176 time: POSIX CPU timers: Ensure that variable is initialized
177 177
178 If cpu_timer_sample_group returns -EINVAL, 178 If cpu_timer_sample_group returns -EINVAL, it will not have written into
179 *sample. Checking for cpu_timer_sample_gro 179 *sample. Checking for cpu_timer_sample_group's return value precludes the
180 potential use of an uninitialized value of 180 potential use of an uninitialized value of now in the following block.
181 Given an invalid clock_idx, the previous c 181 Given an invalid clock_idx, the previous code could otherwise overwrite
182 *oldval in an undefined manner. This is no 182 *oldval in an undefined manner. This is now prevented. We also exploit
183 short-circuiting of && to sample the timer 183 short-circuiting of && to sample the timer only if the result will
184 actually be used to update *oldval. 184 actually be used to update *oldval.
185 185
186 Improved version:: 186 Improved version::
187 187
188 posix-cpu-timers: Make set_process_cpu_tim 188 posix-cpu-timers: Make set_process_cpu_timer() more robust
189 189
190 Because the return value of cpu_timer_samp 190 Because the return value of cpu_timer_sample_group() is not checked,
191 compilers and static checkers can legitima 191 compilers and static checkers can legitimately warn about a potential use
192 of the uninitialized variable 'now'. This 192 of the uninitialized variable 'now'. This is not a runtime issue as all
193 call sites hand in valid clock ids. 193 call sites hand in valid clock ids.
194 194
195 Also cpu_timer_sample_group() is invoked u 195 Also cpu_timer_sample_group() is invoked unconditionally even when the
196 result is not used because *oldval is NULL 196 result is not used because *oldval is NULL.
197 197
198 Make the invocation conditional and check 198 Make the invocation conditional and check the return value.
199 199
200 Example 3:: 200 Example 3::
201 201
202 The entity can also be used for other purp 202 The entity can also be used for other purposes.
203 203
204 Let's rename it to be more generic. 204 Let's rename it to be more generic.
205 205
206 Improved version:: 206 Improved version::
207 207
208 The entity can also be used for other purp 208 The entity can also be used for other purposes.
209 209
210 Rename it to be more generic. 210 Rename it to be more generic.
211 211
212 212
213 For complex scenarios, especially race conditi 213 For complex scenarios, especially race conditions and memory ordering
214 issues, it is valuable to depict the scenario 214 issues, it is valuable to depict the scenario with a table which shows
215 the parallelism and the temporal order of even 215 the parallelism and the temporal order of events. Here is an example::
216 216
217 CPU0 CPU1 217 CPU0 CPU1
218 free_irq(X) interrupt 218 free_irq(X) interrupt X
219 spin_lock( 219 spin_lock(desc->lock)
220 wake irq t 220 wake irq thread()
221 spin_unloc 221 spin_unlock(desc->lock)
222 spin_lock(desc->lock) 222 spin_lock(desc->lock)
223 remove action() 223 remove action()
224 shutdown_irq() 224 shutdown_irq()
225 release_resources() thread_han 225 release_resources() thread_handler()
226 spin_unlock(desc->lock) access r 226 spin_unlock(desc->lock) access released resources.
227 ^^^^^^^^ 227 ^^^^^^^^^^^^^^^^^^^^^^^^^
228 synchronize_irq() 228 synchronize_irq()
229 229
230 Lockdep provides similar useful output to depi 230 Lockdep provides similar useful output to depict a possible deadlock
231 scenario:: 231 scenario::
232 232
233 CPU0 CP 233 CPU0 CPU1
234 rtmutex_lock(&rcu->rt_mutex) 234 rtmutex_lock(&rcu->rt_mutex)
235 spin_lock(&rcu->rt_mutex.wait_lock) 235 spin_lock(&rcu->rt_mutex.wait_lock)
236 lo 236 local_irq_disable()
237 sp 237 spin_lock(&timer->it_lock)
238 sp 238 spin_lock(&rcu->mutex.wait_lock)
239 --> Interrupt 239 --> Interrupt
240 spin_lock(&timer->it_lock) 240 spin_lock(&timer->it_lock)
241 241
242 242
243 Function references in changelogs 243 Function references in changelogs
244 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 244 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
245 245
246 When a function is mentioned in the changelog, 246 When a function is mentioned in the changelog, either the text body or the
247 subject line, please use the format 'function_ 247 subject line, please use the format 'function_name()'. Omitting the
248 brackets after the function name can be ambigu 248 brackets after the function name can be ambiguous::
249 249
250 Subject: subsys/component: Make reservation_ 250 Subject: subsys/component: Make reservation_count static
251 251
252 reservation_count is only used in reservatio 252 reservation_count is only used in reservation_stats. Make it static.
253 253
254 The variant with brackets is more precise:: 254 The variant with brackets is more precise::
255 255
256 Subject: subsys/component: Make reservation_ 256 Subject: subsys/component: Make reservation_count() static
257 257
258 reservation_count() is only called from rese 258 reservation_count() is only called from reservation_stats(). Make it
259 static. 259 static.
260 260
261 261
262 Backtraces in changelogs 262 Backtraces in changelogs
263 ^^^^^^^^^^^^^^^^^^^^^^^^ 263 ^^^^^^^^^^^^^^^^^^^^^^^^
264 264
265 See :ref:`backtraces`. 265 See :ref:`backtraces`.
266 266
267 Ordering of commit tags 267 Ordering of commit tags
268 ^^^^^^^^^^^^^^^^^^^^^^^ 268 ^^^^^^^^^^^^^^^^^^^^^^^
269 269
270 To have a uniform view of the commit tags, the 270 To have a uniform view of the commit tags, the tip maintainers use the
271 following tag ordering scheme: 271 following tag ordering scheme:
272 272
273 - Fixes: 12char-SHA1 ("sub/sys: Original subj 273 - Fixes: 12char-SHA1 ("sub/sys: Original subject line")
274 274
275 A Fixes tag should be added even for change 275 A Fixes tag should be added even for changes which do not need to be
276 backported to stable kernels, i.e. when add 276 backported to stable kernels, i.e. when addressing a recently introduced
277 issue which only affects tip or the current 277 issue which only affects tip or the current head of mainline. These tags
278 are helpful to identify the original commit 278 are helpful to identify the original commit and are much more valuable
279 than prominently mentioning the commit whic 279 than prominently mentioning the commit which introduced a problem in the
280 text of the changelog itself because they c 280 text of the changelog itself because they can be automatically
281 extracted. 281 extracted.
282 282
283 The following example illustrates the diffe 283 The following example illustrates the difference::
284 284
285 Commit 285 Commit
286 286
287 abcdef012345678 ("x86/xxx: Replace foo 287 abcdef012345678 ("x86/xxx: Replace foo with bar")
288 288
289 left an unused instance of variable foo a 289 left an unused instance of variable foo around. Remove it.
290 290
291 Signed-off-by: J.Dev <j.dev@mail> 291 Signed-off-by: J.Dev <j.dev@mail>
292 292
293 Please say instead:: 293 Please say instead::
294 294
295 The recent replacement of foo with bar le 295 The recent replacement of foo with bar left an unused instance of
296 variable foo around. Remove it. 296 variable foo around. Remove it.
297 297
298 Fixes: abcdef012345678 ("x86/xxx: Replace 298 Fixes: abcdef012345678 ("x86/xxx: Replace foo with bar")
299 Signed-off-by: J.Dev <j.dev@mail> 299 Signed-off-by: J.Dev <j.dev@mail>
300 300
301 The latter puts the information about the p 301 The latter puts the information about the patch into the focus and
302 amends it with the reference to the commit 302 amends it with the reference to the commit which introduced the issue
303 rather than putting the focus on the origin 303 rather than putting the focus on the original commit in the first place.
304 304
305 - Reported-by: ``Reporter <reporter@mail>`` 305 - Reported-by: ``Reporter <reporter@mail>``
306 306
307 - Closes: ``URL or Message-ID of the bug repo <<
308 <<
309 - Originally-by: ``Original author <original-a 307 - Originally-by: ``Original author <original-author@mail>``
310 308
311 - Suggested-by: ``Suggester <suggester@mail>`` 309 - Suggested-by: ``Suggester <suggester@mail>``
312 310
313 - Co-developed-by: ``Co-author <co-author@mail 311 - Co-developed-by: ``Co-author <co-author@mail>``
314 312
315 Signed-off-by: ``Co-author <co-author@mail>` !! 313 Signed-off: ``Co-author <co-author@mail>``
316 314
317 Note, that Co-developed-by and Signed-off-b 315 Note, that Co-developed-by and Signed-off-by of the co-author(s) must
318 come in pairs. 316 come in pairs.
319 317
320 - Signed-off-by: ``Author <author@mail>`` 318 - Signed-off-by: ``Author <author@mail>``
321 319
322 The first Signed-off-by (SOB) after the las 320 The first Signed-off-by (SOB) after the last Co-developed-by/SOB pair is the
323 author SOB, i.e. the person flagged as auth 321 author SOB, i.e. the person flagged as author by git.
324 322
325 - Signed-off-by: ``Patch handler <handler@mail 323 - Signed-off-by: ``Patch handler <handler@mail>``
326 324
327 SOBs after the author SOB are from people h 325 SOBs after the author SOB are from people handling and transporting
328 the patch, but were not involved in develop 326 the patch, but were not involved in development. SOB chains should
329 reflect the **real** route a patch took as 327 reflect the **real** route a patch took as it was propagated to us,
330 with the first SOB entry signalling primary 328 with the first SOB entry signalling primary authorship of a single
331 author. Acks should be given as Acked-by li 329 author. Acks should be given as Acked-by lines and review approvals
332 as Reviewed-by lines. 330 as Reviewed-by lines.
333 331
334 If the handler made modifications to the pa 332 If the handler made modifications to the patch or the changelog, then
335 this should be mentioned **after** the chan 333 this should be mentioned **after** the changelog text and **above**
336 all commit tags in the following format:: 334 all commit tags in the following format::
337 335
338 ... changelog text ends. 336 ... changelog text ends.
339 337
340 [ handler: Replaced foo by bar and update 338 [ handler: Replaced foo by bar and updated changelog ]
341 339
342 First-tag: ..... 340 First-tag: .....
343 341
344 Note the two empty new lines which separate 342 Note the two empty new lines which separate the changelog text and the
345 commit tags from that notice. 343 commit tags from that notice.
346 344
347 If a patch is sent to the mailing list by a 345 If a patch is sent to the mailing list by a handler then the author has
348 to be noted in the first line of the change 346 to be noted in the first line of the changelog with::
349 347
350 From: Author <author@mail> 348 From: Author <author@mail>
351 349
352 Changelog text starts here.... 350 Changelog text starts here....
353 351
354 so the authorship is preserved. The 'From:' 352 so the authorship is preserved. The 'From:' line has to be followed
355 by a empty newline. If that 'From:' line is 353 by a empty newline. If that 'From:' line is missing, then the patch
356 would be attributed to the person who sent 354 would be attributed to the person who sent (transported, handled) it.
357 The 'From:' line is automatically removed w 355 The 'From:' line is automatically removed when the patch is applied
358 and does not show up in the final git chang 356 and does not show up in the final git changelog. It merely affects
359 the authorship information of the resulting 357 the authorship information of the resulting Git commit.
360 358
361 - Tested-by: ``Tester <tester@mail>`` 359 - Tested-by: ``Tester <tester@mail>``
362 360
363 - Reviewed-by: ``Reviewer <reviewer@mail>`` 361 - Reviewed-by: ``Reviewer <reviewer@mail>``
364 362
365 - Acked-by: ``Acker <acker@mail>`` 363 - Acked-by: ``Acker <acker@mail>``
366 364
367 - Cc: ``cc-ed-person <person@mail>`` 365 - Cc: ``cc-ed-person <person@mail>``
368 366
369 If the patch should be backported to stable 367 If the patch should be backported to stable, then please add a '``Cc:
370 stable@vger.kernel.org``' tag, but do not C 368 stable@vger.kernel.org``' tag, but do not Cc stable when sending your
371 mail. 369 mail.
372 370
373 - Link: ``https://link/to/information`` 371 - Link: ``https://link/to/information``
374 372
375 For referring to an email posted to the ker !! 373 For referring to an email on LKML or other kernel mailing lists,
376 use the lore.kernel.org redirector URL:: !! 374 please use the lore.kernel.org redirector URL::
377 375
378 Link: https://lore.kernel.org/email-messa !! 376 https://lore.kernel.org/r/email-message@id
379 377
380 This URL should be used when referring to r !! 378 The kernel.org redirector is considered a stable URL, unlike other email
381 topics, related patch sets, or other notabl !! 379 archives.
382 A convenient way to associate ``Link:`` tra <<
383 message is to use markdown-like bracketed n <<
384 <<
385 A similar approach was attempted before a <<
386 effort [1], but the initial implementatio <<
387 regressions [2], so it was backed out and <<
388 <<
389 Link: https://lore.kernel.org/some-msgid@ <<
390 Link: https://bugzilla.example.org/bug/12 <<
391 <<
392 You can also use ``Link:`` trailers to indi <<
393 patch when applying it to your git tree. In <<
394 dedicated ``patch.msgid.link`` domain inste <<
395 This practice makes it possible for automat <<
396 which link to use to retrieve the original <<
397 example:: <<
398 380
399 Link: https://patch.msgid.link/patch-sour !! 381 Maintainers will add a Link tag referencing the email of the patch
>> 382 submission when they apply a patch to the tip tree. This tag is useful
>> 383 for later reference and is also used for commit notifications.
400 384
401 Please do not use combined tags, e.g. ``Report 385 Please do not use combined tags, e.g. ``Reported-and-tested-by``, as
402 they just complicate automated extraction of t 386 they just complicate automated extraction of tags.
403 387
404 388
405 Links to documentation 389 Links to documentation
406 ^^^^^^^^^^^^^^^^^^^^^^ 390 ^^^^^^^^^^^^^^^^^^^^^^
407 391
408 Providing links to documentation in the change 392 Providing links to documentation in the changelog is a great help to later
409 debugging and analysis. Unfortunately, URLs o 393 debugging and analysis. Unfortunately, URLs often break very quickly
410 because companies restructure their websites f 394 because companies restructure their websites frequently. Non-'volatile'
411 exceptions include the Intel SDM and the AMD A 395 exceptions include the Intel SDM and the AMD APM.
412 396
413 Therefore, for 'volatile' documents, please cr 397 Therefore, for 'volatile' documents, please create an entry in the kernel
414 bugzilla https://bugzilla.kernel.org and attac 398 bugzilla https://bugzilla.kernel.org and attach a copy of these documents
415 to the bugzilla entry. Finally, provide the UR 399 to the bugzilla entry. Finally, provide the URL of the bugzilla entry in
416 the changelog. 400 the changelog.
417 401
418 Patch resend or reminders 402 Patch resend or reminders
419 ^^^^^^^^^^^^^^^^^^^^^^^^^ 403 ^^^^^^^^^^^^^^^^^^^^^^^^^
420 404
421 See :ref:`resend_reminders`. 405 See :ref:`resend_reminders`.
422 406
423 Merge window 407 Merge window
424 ^^^^^^^^^^^^ 408 ^^^^^^^^^^^^
425 409
426 Please do not expect patches to be reviewed or !! 410 Please do not expect large patch series to be handled during the merge
427 maintainers around or during the merge window. !! 411 window or even during the week before. Such patches should be submitted in
428 to all but urgent fixes during this time. The !! 412 mergeable state *at* *least* a week before the merge window opens.
429 window closes and a new -rc1 kernel has been r !! 413 Exceptions are made for bug fixes and *sometimes* for small standalone
430 !! 414 drivers for new hardware or minimally invasive patches for hardware
431 Large series should be submitted in mergeable !! 415 enablement.
432 before the merge window opens. Exceptions are <<
433 *sometimes* for small standalone drivers for n <<
434 invasive patches for hardware enablement. <<
435 416
436 During the merge window, the maintainers inste 417 During the merge window, the maintainers instead focus on following the
437 upstream changes, fixing merge window fallout, 418 upstream changes, fixing merge window fallout, collecting bug fixes, and
438 allowing themselves a breath. Please respect t 419 allowing themselves a breath. Please respect that.
439 420
>> 421 The release candidate -rc1 is the starting point for new patches to be
>> 422 applied which are targeted for the next merge window.
>> 423
440 So called _urgent_ branches will be merged int 424 So called _urgent_ branches will be merged into mainline during the
441 stabilization phase of each release. 425 stabilization phase of each release.
442 426
443 427
444 Git 428 Git
445 ^^^ 429 ^^^
446 430
447 The tip maintainers accept git pull requests f 431 The tip maintainers accept git pull requests from maintainers who provide
448 subsystem changes for aggregation in the tip t 432 subsystem changes for aggregation in the tip tree.
449 433
450 Pull requests for new patch submissions are us 434 Pull requests for new patch submissions are usually not accepted and do not
451 replace proper patch submission to the mailing 435 replace proper patch submission to the mailing list. The main reason for
452 this is that the review workflow is email base 436 this is that the review workflow is email based.
453 437
454 If you submit a larger patch series it is help 438 If you submit a larger patch series it is helpful to provide a git branch
455 in a private repository which allows intereste 439 in a private repository which allows interested people to easily pull the
456 series for testing. The usual way to offer thi 440 series for testing. The usual way to offer this is a git URL in the cover
457 letter of the patch series. 441 letter of the patch series.
458 442
459 Testing 443 Testing
460 ^^^^^^^ 444 ^^^^^^^
461 445
462 Code should be tested before submitting to the 446 Code should be tested before submitting to the tip maintainers. Anything
463 other than minor changes should be built, boot 447 other than minor changes should be built, booted and tested with
464 comprehensive (and heavyweight) kernel debuggi 448 comprehensive (and heavyweight) kernel debugging options enabled.
465 449
466 These debugging options can be found in kernel 450 These debugging options can be found in kernel/configs/x86_debug.config
467 and can be added to an existing kernel config 451 and can be added to an existing kernel config by running:
468 452
469 make x86_debug.config 453 make x86_debug.config
470 454
471 Some of these options are x86-specific and can 455 Some of these options are x86-specific and can be left out when testing
472 on other architectures. 456 on other architectures.
473 457
474 .. _maintainer-tip-coding-style: 458 .. _maintainer-tip-coding-style:
475 459
476 Coding style notes 460 Coding style notes
477 ------------------ 461 ------------------
478 462
479 Comment style 463 Comment style
480 ^^^^^^^^^^^^^ 464 ^^^^^^^^^^^^^
481 465
482 Sentences in comments start with an uppercase 466 Sentences in comments start with an uppercase letter.
483 467
484 Single line comments:: 468 Single line comments::
485 469
486 /* This is a single line comment */ 470 /* This is a single line comment */
487 471
488 Multi-line comments:: 472 Multi-line comments::
489 473
490 /* 474 /*
491 * This is a properly formatted 475 * This is a properly formatted
492 * multi-line comment. 476 * multi-line comment.
493 * 477 *
494 * Larger multi-line comments should b 478 * Larger multi-line comments should be split into paragraphs.
495 */ 479 */
496 480
497 No tail comments (see below): !! 481 No tail comments:
498 482
499 Please refrain from using tail comments. Tai 483 Please refrain from using tail comments. Tail comments disturb the
500 reading flow in almost all contexts, but esp 484 reading flow in almost all contexts, but especially in code::
501 485
502 if (somecondition_is_true) /* Don't pu 486 if (somecondition_is_true) /* Don't put a comment here */
503 dostuff(); /* Neither here */ 487 dostuff(); /* Neither here */
504 488
505 seed = MAGIC_CONSTANT; /* Nor here */ 489 seed = MAGIC_CONSTANT; /* Nor here */
506 490
507 Use freestanding comments instead:: 491 Use freestanding comments instead::
508 492
509 /* This condition is not obvious witho 493 /* This condition is not obvious without a comment */
510 if (somecondition_is_true) { 494 if (somecondition_is_true) {
511 /* This really needs to be doc 495 /* This really needs to be documented */
512 dostuff(); 496 dostuff();
513 } 497 }
514 498
515 /* This magic initialization needs a c 499 /* This magic initialization needs a comment. Maybe not? */
516 seed = MAGIC_CONSTANT; 500 seed = MAGIC_CONSTANT;
517 <<
518 Use C++ style, tail comments when documentin <<
519 achieve a more compact layout and better rea <<
520 <<
521 // eax <<
522 u32 x2apic_shift : 5, // Numbe <<
523 // for t <<
524 : 27; // Reser <<
525 // ebx <<
526 u32 num_processors : 16, // Numbe <<
527 : 16; // Reser <<
528 <<
529 versus:: <<
530 <<
531 /* eax */ <<
532 /* <<
533 * Number of bits to shift API <<
534 * at the next level <<
535 */ <<
536 u32 x2apic_shift : 5, <<
537 /* Reserved */ <<
538 : 27; <<
539 <<
540 /* ebx */ <<
541 /* Number of processors at cur <<
542 u32 num_processors : 16, <<
543 /* Reserved */ <<
544 : 16; <<
545 501
546 Comment the important things: 502 Comment the important things:
547 503
548 Comments should be added where the operation 504 Comments should be added where the operation is not obvious. Documenting
549 the obvious is just a distraction:: 505 the obvious is just a distraction::
550 506
551 /* Decrement refcount and check for ze 507 /* Decrement refcount and check for zero */
552 if (refcount_dec_and_test(&p->refcnt)) 508 if (refcount_dec_and_test(&p->refcnt)) {
553 do; 509 do;
554 lots; 510 lots;
555 of; 511 of;
556 magic; 512 magic;
557 things; 513 things;
558 } 514 }
559 515
560 Instead, comments should explain the non-obv 516 Instead, comments should explain the non-obvious details and document
561 constraints:: 517 constraints::
562 518
563 if (refcount_dec_and_test(&p->refcnt)) 519 if (refcount_dec_and_test(&p->refcnt)) {
564 /* 520 /*
565 * Really good explanation why 521 * Really good explanation why the magic things below
566 * need to be done, ordering a 522 * need to be done, ordering and locking constraints,
567 * etc.. 523 * etc..
568 */ 524 */
569 do; 525 do;
570 lots; 526 lots;
571 of; 527 of;
572 magic; 528 magic;
573 /* Needs to be the last operat 529 /* Needs to be the last operation because ... */
574 things; 530 things;
575 } 531 }
576 532
577 Function documentation comments: 533 Function documentation comments:
578 534
579 To document functions and their arguments pl 535 To document functions and their arguments please use kernel-doc format
580 and not free form comments:: 536 and not free form comments::
581 537
582 /** 538 /**
583 * magic_function - Do lots of magic s 539 * magic_function - Do lots of magic stuff
584 * @magic: Pointer to the magic d 540 * @magic: Pointer to the magic data to operate on
585 * @offset: Offset in the data arr 541 * @offset: Offset in the data array of @magic
586 * 542 *
587 * Deep explanation of mysterious thin 543 * Deep explanation of mysterious things done with @magic along
588 * with documentation of the return va 544 * with documentation of the return values.
589 * 545 *
590 * Note, that the argument descriptors 546 * Note, that the argument descriptors above are arranged
591 * in a tabular fashion. 547 * in a tabular fashion.
592 */ 548 */
593 549
594 This applies especially to globally visible 550 This applies especially to globally visible functions and inline
595 functions in public header files. It might b 551 functions in public header files. It might be overkill to use kernel-doc
596 format for every (static) function which nee 552 format for every (static) function which needs a tiny explanation. The
597 usage of descriptive function names often re 553 usage of descriptive function names often replaces these tiny comments.
598 Apply common sense as always. 554 Apply common sense as always.
599 555
600 556
601 Documenting locking requirements 557 Documenting locking requirements
602 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 558 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
603 Documenting locking requirements is a good t 559 Documenting locking requirements is a good thing, but comments are not
604 necessarily the best choice. Instead of writ 560 necessarily the best choice. Instead of writing::
605 561
606 /* Caller must hold foo->lock */ 562 /* Caller must hold foo->lock */
607 void func(struct foo *foo) 563 void func(struct foo *foo)
608 { 564 {
609 ... 565 ...
610 } 566 }
611 567
612 Please use:: 568 Please use::
613 569
614 void func(struct foo *foo) 570 void func(struct foo *foo)
615 { 571 {
616 lockdep_assert_held(&foo->lock 572 lockdep_assert_held(&foo->lock);
617 ... 573 ...
618 } 574 }
619 575
620 In PROVE_LOCKING kernels, lockdep_assert_hel 576 In PROVE_LOCKING kernels, lockdep_assert_held() emits a warning
621 if the caller doesn't hold the lock. Commen 577 if the caller doesn't hold the lock. Comments can't do that.
622 578
623 Bracket rules 579 Bracket rules
624 ^^^^^^^^^^^^^ 580 ^^^^^^^^^^^^^
625 581
626 Brackets should be omitted only if the stateme 582 Brackets should be omitted only if the statement which follows 'if', 'for',
627 'while' etc. is truly a single line:: 583 'while' etc. is truly a single line::
628 584
629 if (foo) 585 if (foo)
630 do_something(); 586 do_something();
631 587
632 The following is not considered to be a single 588 The following is not considered to be a single line statement even
633 though C does not require brackets:: 589 though C does not require brackets::
634 590
635 for (i = 0; i < end; i++) 591 for (i = 0; i < end; i++)
636 if (foo[i]) 592 if (foo[i])
637 do_something(foo[i]); 593 do_something(foo[i]);
638 594
639 Adding brackets around the outer loop enhances 595 Adding brackets around the outer loop enhances the reading flow::
640 596
641 for (i = 0; i < end; i++) { 597 for (i = 0; i < end; i++) {
642 if (foo[i]) 598 if (foo[i])
643 do_something(foo[i]); 599 do_something(foo[i]);
644 } 600 }
645 601
646 602
647 Variable declarations 603 Variable declarations
648 ^^^^^^^^^^^^^^^^^^^^^ 604 ^^^^^^^^^^^^^^^^^^^^^
649 605
650 The preferred ordering of variable declaration 606 The preferred ordering of variable declarations at the beginning of a
651 function is reverse fir tree order:: 607 function is reverse fir tree order::
652 608
653 struct long_struct_name *descriptive_n 609 struct long_struct_name *descriptive_name;
654 unsigned long foo, bar; 610 unsigned long foo, bar;
655 unsigned int tmp; 611 unsigned int tmp;
656 int ret; 612 int ret;
657 613
658 The above is faster to parse than the reverse 614 The above is faster to parse than the reverse ordering::
659 615
660 int ret; 616 int ret;
661 unsigned int tmp; 617 unsigned int tmp;
662 unsigned long foo, bar; 618 unsigned long foo, bar;
663 struct long_struct_name *descriptive_n 619 struct long_struct_name *descriptive_name;
664 620
665 And even more so than random ordering:: 621 And even more so than random ordering::
666 622
667 unsigned long foo, bar; 623 unsigned long foo, bar;
668 int ret; 624 int ret;
669 struct long_struct_name *descriptive_n 625 struct long_struct_name *descriptive_name;
670 unsigned int tmp; 626 unsigned int tmp;
671 627
672 Also please try to aggregate variables of the 628 Also please try to aggregate variables of the same type into a single
673 line. There is no point in wasting screen spac 629 line. There is no point in wasting screen space::
674 630
675 unsigned long a; 631 unsigned long a;
676 unsigned long b; 632 unsigned long b;
677 unsigned long c; 633 unsigned long c;
678 unsigned long d; 634 unsigned long d;
679 635
680 It's really sufficient to do:: 636 It's really sufficient to do::
681 637
682 unsigned long a, b, c, d; 638 unsigned long a, b, c, d;
683 639
684 Please also refrain from introducing line spli 640 Please also refrain from introducing line splits in variable declarations::
685 641
686 struct long_struct_name *descriptive_n 642 struct long_struct_name *descriptive_name = container_of(bar,
687 643 struct long_struct_name,
688 644 member);
689 struct foobar foo; 645 struct foobar foo;
690 646
691 It's way better to move the initialization to 647 It's way better to move the initialization to a separate line after the
692 declarations:: 648 declarations::
693 649
694 struct long_struct_name *descriptive_n 650 struct long_struct_name *descriptive_name;
695 struct foobar foo; 651 struct foobar foo;
696 652
697 descriptive_name = container_of(bar, s 653 descriptive_name = container_of(bar, struct long_struct_name, member);
698 654
699 655
700 Variable types 656 Variable types
701 ^^^^^^^^^^^^^^ 657 ^^^^^^^^^^^^^^
702 658
703 Please use the proper u8, u16, u32, u64 types 659 Please use the proper u8, u16, u32, u64 types for variables which are meant
704 to describe hardware or are used as arguments 660 to describe hardware or are used as arguments for functions which access
705 hardware. These types are clearly defining the 661 hardware. These types are clearly defining the bit width and avoid
706 truncation, expansion and 32/64-bit confusion. 662 truncation, expansion and 32/64-bit confusion.
707 663
708 u64 is also recommended in code which would be 664 u64 is also recommended in code which would become ambiguous for 32-bit
709 kernels when 'unsigned long' would be used ins 665 kernels when 'unsigned long' would be used instead. While in such
710 situations 'unsigned long long' could be used 666 situations 'unsigned long long' could be used as well, u64 is shorter
711 and also clearly shows that the operation is r 667 and also clearly shows that the operation is required to be 64 bits wide
712 independent of the target CPU. 668 independent of the target CPU.
713 669
714 Please use 'unsigned int' instead of 'unsigned 670 Please use 'unsigned int' instead of 'unsigned'.
715 671
716 672
717 Constants 673 Constants
718 ^^^^^^^^^ 674 ^^^^^^^^^
719 675
720 Please do not use literal (hexa)decimal number 676 Please do not use literal (hexa)decimal numbers in code or initializers.
721 Either use proper defines which have descripti 677 Either use proper defines which have descriptive names or consider using
722 an enum. 678 an enum.
723 679
724 680
725 Struct declarations and initializers 681 Struct declarations and initializers
726 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 682 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
727 683
728 Struct declarations should align the struct me 684 Struct declarations should align the struct member names in a tabular
729 fashion:: 685 fashion::
730 686
731 struct bar_order { 687 struct bar_order {
732 unsigned int guest_id; 688 unsigned int guest_id;
733 int ordered_item; 689 int ordered_item;
734 struct menu *menu; 690 struct menu *menu;
735 }; 691 };
736 692
737 Please avoid documenting struct members within 693 Please avoid documenting struct members within the declaration, because
738 this often results in strangely formatted comm 694 this often results in strangely formatted comments and the struct members
739 become obfuscated:: 695 become obfuscated::
740 696
741 struct bar_order { 697 struct bar_order {
742 unsigned int guest_id; /* U 698 unsigned int guest_id; /* Unique guest id */
743 int ordered_item; 699 int ordered_item;
744 /* Pointer to a menu instance 700 /* Pointer to a menu instance which contains all the drinks */
745 struct menu *menu; 701 struct menu *menu;
746 }; 702 };
747 703
748 Instead, please consider using the kernel-doc 704 Instead, please consider using the kernel-doc format in a comment preceding
749 the struct declaration, which is easier to rea 705 the struct declaration, which is easier to read and has the added advantage
750 of including the information in the kernel doc 706 of including the information in the kernel documentation, for example, as
751 follows:: 707 follows::
752 708
753 709
754 /** 710 /**
755 * struct bar_order - Description of a 711 * struct bar_order - Description of a bar order
756 * @guest_id: Unique guest i 712 * @guest_id: Unique guest id
757 * @ordered_item: The item numbe 713 * @ordered_item: The item number from the menu
758 * @menu: Pointer to the 714 * @menu: Pointer to the menu from which the item
759 * was ordered 715 * was ordered
760 * 716 *
761 * Supplementary information for using 717 * Supplementary information for using the struct.
762 * 718 *
763 * Note, that the struct member descri 719 * Note, that the struct member descriptors above are arranged
764 * in a tabular fashion. 720 * in a tabular fashion.
765 */ 721 */
766 struct bar_order { 722 struct bar_order {
767 unsigned int guest_id; 723 unsigned int guest_id;
768 int ordered_item; 724 int ordered_item;
769 struct menu *menu; 725 struct menu *menu;
770 }; 726 };
771 727
772 Static struct initializers must use C99 initia 728 Static struct initializers must use C99 initializers and should also be
773 aligned in a tabular fashion:: 729 aligned in a tabular fashion::
774 730
775 static struct foo statfoo = { 731 static struct foo statfoo = {
776 .a = 0, 732 .a = 0,
777 .plain_integer = CONSTANT_DEF 733 .plain_integer = CONSTANT_DEFINE_OR_ENUM,
778 .bar = &statbar, 734 .bar = &statbar,
779 }; 735 };
780 736
781 Note that while C99 syntax allows the omission 737 Note that while C99 syntax allows the omission of the final comma,
782 we recommend the use of a comma on the last li 738 we recommend the use of a comma on the last line because it makes
783 reordering and addition of new lines easier, a 739 reordering and addition of new lines easier, and makes such future
784 patches slightly easier to read as well. 740 patches slightly easier to read as well.
785 741
786 Line breaks 742 Line breaks
787 ^^^^^^^^^^^ 743 ^^^^^^^^^^^
788 744
789 Restricting line length to 80 characters makes 745 Restricting line length to 80 characters makes deeply indented code hard to
790 read. Consider breaking out code into helper 746 read. Consider breaking out code into helper functions to avoid excessive
791 line breaking. 747 line breaking.
792 748
793 The 80 character rule is not a strict rule, so 749 The 80 character rule is not a strict rule, so please use common sense when
794 breaking lines. Especially format strings shou 750 breaking lines. Especially format strings should never be broken up.
795 751
796 When splitting function declarations or functi 752 When splitting function declarations or function calls, then please align
797 the first argument in the second line with the 753 the first argument in the second line with the first argument in the first
798 line:: 754 line::
799 755
800 static int long_function_name(struct foobar 756 static int long_function_name(struct foobar *barfoo, unsigned int id,
801 unsigned int o 757 unsigned int offset)
802 { 758 {
803 759
804 if (!id) { 760 if (!id) {
805 ret = longer_function_name(bar 761 ret = longer_function_name(barfoo, DEFAULT_BARFOO_ID,
806 off 762 offset);
807 ... 763 ...
808 764
809 Namespaces 765 Namespaces
810 ^^^^^^^^^^ 766 ^^^^^^^^^^
811 767
812 Function/variable namespaces improve readabili 768 Function/variable namespaces improve readability and allow easy
813 grepping. These namespaces are string prefixes 769 grepping. These namespaces are string prefixes for globally visible
814 function and variable names, including inlines 770 function and variable names, including inlines. These prefixes should
815 combine the subsystem and the component name s 771 combine the subsystem and the component name such as 'x86_comp\_',
816 'sched\_', 'irq\_', and 'mutex\_'. 772 'sched\_', 'irq\_', and 'mutex\_'.
817 773
818 This also includes static file scope functions 774 This also includes static file scope functions that are immediately put
819 into globally visible driver templates - it's 775 into globally visible driver templates - it's useful for those symbols
820 to carry a good prefix as well, for backtrace 776 to carry a good prefix as well, for backtrace readability.
821 777
822 Namespace prefixes may be omitted for local st 778 Namespace prefixes may be omitted for local static functions and
823 variables. Truly local functions, only called 779 variables. Truly local functions, only called by other local functions,
824 can have shorter descriptive names - our prima 780 can have shorter descriptive names - our primary concern is greppability
825 and backtrace readability. 781 and backtrace readability.
826 782
827 Please note that 'xxx_vendor\_' and 'vendor_xx 783 Please note that 'xxx_vendor\_' and 'vendor_xxx_` prefixes are not
828 helpful for static functions in vendor-specifi 784 helpful for static functions in vendor-specific files. After all, it
829 is already clear that the code is vendor-speci 785 is already clear that the code is vendor-specific. In addition, vendor
830 names should only be for truly vendor-specific 786 names should only be for truly vendor-specific functionality.
831 787
832 As always apply common sense and aim for consi 788 As always apply common sense and aim for consistency and readability.
833 789
834 790
835 Commit notifications 791 Commit notifications
836 -------------------- 792 --------------------
837 793
838 The tip tree is monitored by a bot for new com 794 The tip tree is monitored by a bot for new commits. The bot sends an email
839 for each new commit to a dedicated mailing lis 795 for each new commit to a dedicated mailing list
840 (``linux-tip-commits@vger.kernel.org``) and Cc 796 (``linux-tip-commits@vger.kernel.org``) and Cc's all people who are
841 mentioned in one of the commit tags. It uses t 797 mentioned in one of the commit tags. It uses the email message ID from the
842 Link tag at the end of the tag list to set the 798 Link tag at the end of the tag list to set the In-Reply-To email header so
843 the message is properly threaded with the patc 799 the message is properly threaded with the patch submission email.
844 800
845 The tip maintainers and submaintainers try to 801 The tip maintainers and submaintainers try to reply to the submitter
846 when merging a patch, but they sometimes forge 802 when merging a patch, but they sometimes forget or it does not fit the
847 workflow of the moment. While the bot message 803 workflow of the moment. While the bot message is purely mechanical, it
848 also implies a 'Thank you! Applied.'. 804 also implies a 'Thank you! Applied.'.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.