Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
On Thu, Oct 15, 2020 at 07:15:18AM +0200, Mauro Carvalho Chehab wrote: > I guess there might be some misunderstanding here. My fault. The plan > is to have zero doc warnings for 5.10[1]. I'd be glad to help and convert all the documentation under my maintainership to .txt files for you.
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
Hi Paul, Em Wed, 14 Oct 2020 11:57:20 -0700 "Paul E. McKenney" escreveu: > On Tue, Oct 13, 2020 at 06:58:40PM -0700, Paul E. McKenney wrote: > > On Tue, Oct 13, 2020 at 12:38:36PM -0400, Alan Stern wrote: > > > On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: > > > > On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > > > > > - The sysfs.txt file was converted to ReST and renamed; > > > > > - The control-dependencies.txt is not at > > > > > Documentation/control-dependencies.txt. As it is at the > > > > > same dir as the README file, which mentions it, just > > > > > remove Documentation/. > > > > > > > > > > With that, ./scripts/documentation-file-ref-check script > > > > > is now happy again for files under tools/. > > > > > > > > > > Signed-off-by: Mauro Carvalho Chehab > > > > > > > > Queued for review and testing, likely target v5.11. > > > > > > Instead of changing the path in the README reference, shouldn't > > > tools/memory-model/control-dependencies.txt be moved to its proper > > > position in .../Documentation? > > > > You are of course quite right. My thought is to let Mauro go ahead, > > given his short deadline. I guess there might be some misunderstanding here. My fault. The plan is to have zero doc warnings for 5.10[1]. In order to get there, The patches for it were split on two series, both for 5.10: - The /80 series with patches that already applies on the top of master; - This /24 patch series, which depends on trees that weren't merged upstream yet (back on Oct, 13). Those applies on the top of next-20201013. I'm intending to submit later today (after next-20201015) a PR with patches from the /80 series. The remaining ones should be sent as a late pull request by the end of the merge window, if the patch that caused the issue gets merged for 5.10. That's the case of this patch. [1] With Sphinx < 3. Sphinx 3 and above brings some additional warnings that depends on a fix at the toolset. The fixup patches for Sphinx were proposed yesterday by the Sphinx maintainer of the C domain parser. More details can be seen here: https://github.com/sphinx-doc/sphinx/pull/8313 > > We can then make this "git mv" change once > > v5.10-rc1 comes out, given that it should have Mauro's patches. I have > > added a reminder to my calendar. > > Except that I cannot find a commit where control-dependencies.txt is > in tools/memory-model. And this file is not yet in mainline, but > only in -rcu and -next. In both places, it is here: > > tools/memory-model/Documentation/control-dependencies.txt > > Mauro, to what commit in what tree are you applying this patch? This is against next-20201013. The specific commit adding README and control-dependencies.txt is this one: commit d34a972f67252457158122e5ba7a0ce5ece62067 Author: Paul E. McKenney AuthorDate: Tue Aug 11 11:27:33 2020 -0700 Commit: Paul E. McKenney CommitDate: Sun Oct 4 17:21:31 2020 -0700 tools/memory-model: Document categories of ordering primitives The Linux kernel has a number of categories of ordering primitives, which are recorded in the LKMM implementation and hinted at by cheatsheet.txt. But there is no overview of these categories, and such an overview is needed in order to understand multithreaded LKMM litmus tests. This commit therefore adds an ordering.txt as well as extracting a control-dependencies.txt from memory-barriers.txt. It also updates the README file. [ paulmck: Apply Akira Yokosawa file-placement feedback. ] [ paulmck: Apply Alan Stern feedback. ] Signed-off-by: Paul E. McKenney Btw, after re-checking the patch, I would drop this hunk: diff --git a/tools/memory-model/Documentation/README b/tools/memory-model/Documentation/README index 16177aaa9752..004969992bac 100644 --- a/tools/memory-model/Documentation/README +++ b/tools/memory-model/Documentation/README @@ -55,7 +55,7 @@ README Documentation/cheatsheet.txt Quick-reference guide to the Linux-kernel memory model. -Documentation/control-dependencies.txt +control-dependencies.txt A guide to preventing compiler optimizations from destroying your control dependencies. The ./scripts/documentation-file-ref-check doesn't complain about broken references for Documentation/ files outside the main docs dir. So, this hunk is not really needed to fix warnings with 5.10. Besides that, there are other references to those files: $ git grep Documentation tools/memory-model/Documentation/README tools/memory-model/Documentation/README:tools/memory-model/Documentation directory. It has been said that at tools/memory-model/Documentation/README:Documentation/cheatsheet.txt tools/memory-model/Documentation/README:Documentation/control-dependencies.txt tools/memory-model/Documentation/README:Documentation/explanation.txt
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
On Tue, Oct 13, 2020 at 06:58:40PM -0700, Paul E. McKenney wrote: > On Tue, Oct 13, 2020 at 12:38:36PM -0400, Alan Stern wrote: > > On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: > > > On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > > > > - The sysfs.txt file was converted to ReST and renamed; > > > > - The control-dependencies.txt is not at > > > > Documentation/control-dependencies.txt. As it is at the > > > > same dir as the README file, which mentions it, just > > > > remove Documentation/. > > > > > > > > With that, ./scripts/documentation-file-ref-check script > > > > is now happy again for files under tools/. > > > > > > > > Signed-off-by: Mauro Carvalho Chehab > > > > > > Queued for review and testing, likely target v5.11. > > > > Instead of changing the path in the README reference, shouldn't > > tools/memory-model/control-dependencies.txt be moved to its proper > > position in .../Documentation? > > You are of course quite right. My thought is to let Mauro go ahead, > given his short deadline. We can then make this "git mv" change once > v5.10-rc1 comes out, given that it should have Mauro's patches. I have > added a reminder to my calendar. Except that I cannot find a commit where control-dependencies.txt is in tools/memory-model. And this file is not yet in mainline, but only in -rcu and -next. In both places, it is here: tools/memory-model/Documentation/control-dependencies.txt Mauro, to what commit in what tree are you applying this patch? Thanx, Paul
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
Em Wed, 14 Oct 2020 23:14:00 +0900 Akira Yokosawa escreveu: > On Wed, 14 Oct 2020 09:56:03 +0200, Mauro Carvalho Chehab wrote: > > Em Tue, 13 Oct 2020 18:58:40 -0700 > > "Paul E. McKenney" escreveu: > > > >> On Tue, Oct 13, 2020 at 12:38:36PM -0400, Alan Stern wrote: > >>> On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: > On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > > > - The sysfs.txt file was converted to ReST and renamed; > > - The control-dependencies.txt is not at > > Documentation/control-dependencies.txt. As it is at the > > same dir as the README file, which mentions it, just > > remove Documentation/. > > > > With that, ./scripts/documentation-file-ref-check script > > is now happy again for files under tools/. > > > > Signed-off-by: Mauro Carvalho Chehab > > Queued for review and testing, likely target v5.11. > >>> > >>> Instead of changing the path in the README reference, shouldn't > >>> tools/memory-model/control-dependencies.txt be moved to its proper > >>> position in .../Documentation? > >> > >> You are of course quite right. My thought is to let Mauro go ahead, > >> given his short deadline. We can then make this "git mv" change once > >> v5.10-rc1 comes out, given that it should have Mauro's patches. I have > >> added a reminder to my calendar. > > > > Sounds like a plan to me. > > > > > > If it helps on 5.11 plans, converting this file to ReST format is quite > > trivial: it just needs to use "::" for C/asm code literal blocks, and > > to replace "(*) " by something that matches ReST syntax for lists, > > like "(#) " or just "* ": > > > > > > https://docutils.sourceforge.io/docs/ref/rst/restructuredtext.html#bullet-lists > > > > See enclosed. > > I'm afraid conversion of LKMM documents to ReST is unlikely to happen > any time soon. > It should wait until such time comes when the auto markup tools become > clever enough and .rst files looks exactly the same as plain .txt files. > > Am I asking too much? :-) > > Thanks, Akira Yes :-) $ git log --author aki...@gmail.com Documentation/sphinx $ The auto markup tools don't write themselves alone. Someone needs to write them and test if no regressions will happen with the existing documents. - That's said, I suspect that one of the hardest things for something like that to be possible is to be able to distinguish something like: (some text) From something like: /* some C code snippet or bash script, or other literal block */ So, at least "::" (or some other markup replacing it) is needed. If you have some bright idea about how to implement it, feel free to contribute with patches. Thanks, Mauro
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
On Wed, 14 Oct 2020 09:56:03 +0200, Mauro Carvalho Chehab wrote: > Em Tue, 13 Oct 2020 18:58:40 -0700 > "Paul E. McKenney" escreveu: > >> On Tue, Oct 13, 2020 at 12:38:36PM -0400, Alan Stern wrote: >>> On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > - The sysfs.txt file was converted to ReST and renamed; > - The control-dependencies.txt is not at > Documentation/control-dependencies.txt. As it is at the > same dir as the README file, which mentions it, just > remove Documentation/. > > With that, ./scripts/documentation-file-ref-check script > is now happy again for files under tools/. > > Signed-off-by: Mauro Carvalho Chehab Queued for review and testing, likely target v5.11. >>> >>> Instead of changing the path in the README reference, shouldn't >>> tools/memory-model/control-dependencies.txt be moved to its proper >>> position in .../Documentation? >> >> You are of course quite right. My thought is to let Mauro go ahead, >> given his short deadline. We can then make this "git mv" change once >> v5.10-rc1 comes out, given that it should have Mauro's patches. I have >> added a reminder to my calendar. > > Sounds like a plan to me. > > > If it helps on 5.11 plans, converting this file to ReST format is quite > trivial: it just needs to use "::" for C/asm code literal blocks, and > to replace "(*) " by something that matches ReST syntax for lists, > like "(#) " or just "* ": > > > https://docutils.sourceforge.io/docs/ref/rst/restructuredtext.html#bullet-lists > > See enclosed. I'm afraid conversion of LKMM documents to ReST is unlikely to happen any time soon. It should wait until such time comes when the auto markup tools become clever enough and .rst files looks exactly the same as plain .txt files. Am I asking too much? :-) Thanks, Akira > > Thanks, > Mauro > > [PATCH] convert control-dependencies.rst to ReST > [snip]
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
Em Tue, 13 Oct 2020 18:58:40 -0700 "Paul E. McKenney" escreveu: > On Tue, Oct 13, 2020 at 12:38:36PM -0400, Alan Stern wrote: > > On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: > > > On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > > > > - The sysfs.txt file was converted to ReST and renamed; > > > > - The control-dependencies.txt is not at > > > > Documentation/control-dependencies.txt. As it is at the > > > > same dir as the README file, which mentions it, just > > > > remove Documentation/. > > > > > > > > With that, ./scripts/documentation-file-ref-check script > > > > is now happy again for files under tools/. > > > > > > > > Signed-off-by: Mauro Carvalho Chehab > > > > > > Queued for review and testing, likely target v5.11. > > > > Instead of changing the path in the README reference, shouldn't > > tools/memory-model/control-dependencies.txt be moved to its proper > > position in .../Documentation? > > You are of course quite right. My thought is to let Mauro go ahead, > given his short deadline. We can then make this "git mv" change once > v5.10-rc1 comes out, given that it should have Mauro's patches. I have > added a reminder to my calendar. Sounds like a plan to me. If it helps on 5.11 plans, converting this file to ReST format is quite trivial: it just needs to use "::" for C/asm code literal blocks, and to replace "(*) " by something that matches ReST syntax for lists, like "(#) " or just "* ": https://docutils.sourceforge.io/docs/ref/rst/restructuredtext.html#bullet-lists See enclosed. Thanks, Mauro [PATCH] convert control-dependencies.rst to ReST - Mark literal blocks as such; - Use a numbered list at the summary. Signed-off-by: Mauro Carvalho Chehab diff --git a/tools/memory-model/Documentation/control-dependencies.rst b/tools/memory-model/Documentation/control-dependencies.rst index 366520cac937..52dc6a5bc173 100644 --- a/tools/memory-model/Documentation/control-dependencies.rst +++ b/tools/memory-model/Documentation/control-dependencies.rst @@ -7,7 +7,7 @@ the compiler's ignorance from breaking your code. A load-load control dependency requires a full read memory barrier, not simply a data dependency barrier to make it work correctly. Consider the -following bit of code: +following bit of code:: q = READ_ONCE(a); if (q) { @@ -19,7 +19,7 @@ This will not have the desired effect because there is no actual data dependency, but rather a control dependency that the CPU may short-circuit by attempting to predict the outcome in advance, so that other CPUs see the load from b as having happened before the load from a. In such a -case what's actually required is: +case what's actually required is:: q = READ_ONCE(a); if (q) { @@ -28,7 +28,7 @@ case what's actually required is: } However, stores are not speculated. This means that ordering -is- provided -for load-store control dependencies, as in the following example: +for load-store control dependencies, as in the following example:: q = READ_ONCE(a); if (q) { @@ -45,7 +45,7 @@ or, worse yet, convert the store into a check followed by a store. Worse yet, if the compiler is able to prove (say) that the value of variable "a" is always non-zero, it would be well within its rights to optimize the original example by eliminating the "if" statement -as follows: +as follows:: q = a; b = 1; /* BUG: Compiler and CPU can both reorder!!! */ @@ -53,7 +53,7 @@ as follows: So don't leave out either the READ_ONCE() or the WRITE_ONCE(). It is tempting to try to enforce ordering on identical stores on both -branches of the "if" statement as follows: +branches of the "if" statement as follows:: q = READ_ONCE(a); if (q) { @@ -67,7 +67,7 @@ branches of the "if" statement as follows: } Unfortunately, current compilers will transform this as follows at high -optimization levels: +optimization levels:: q = READ_ONCE(a); barrier(); @@ -85,7 +85,7 @@ Now there is no conditional between the load from "a" and the store to The conditional is absolutely required, and must be present in the assembly code even after all compiler optimizations have been applied. Therefore, if you need ordering in this example, you need explicit -memory barriers, for example, smp_store_release(): +memory barriers, for example, smp_store_release():: q = READ_ONCE(a); if (q) { @@ -97,7 +97,7 @@ memory barriers, for example, smp_store_release(): } In contrast, without explicit memory barriers, two-legged-if control -ordering is guaranteed only when the stores differ, for example: +ordering is guaranteed only when the stores differ, for example:: q = READ_ONCE(a); if (q) { @@ -113,7 +113,7 @@ proving the value of "a". In addition, you need to be careful what you do with the local variable "q", otherwise the
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
On Tue, Oct 13, 2020 at 12:38:36PM -0400, Alan Stern wrote: > On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: > > On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > > > - The sysfs.txt file was converted to ReST and renamed; > > > - The control-dependencies.txt is not at > > > Documentation/control-dependencies.txt. As it is at the > > > same dir as the README file, which mentions it, just > > > remove Documentation/. > > > > > > With that, ./scripts/documentation-file-ref-check script > > > is now happy again for files under tools/. > > > > > > Signed-off-by: Mauro Carvalho Chehab > > > > Queued for review and testing, likely target v5.11. > > Instead of changing the path in the README reference, shouldn't > tools/memory-model/control-dependencies.txt be moved to its proper > position in .../Documentation? You are of course quite right. My thought is to let Mauro go ahead, given his short deadline. We can then make this "git mv" change once v5.10-rc1 comes out, given that it should have Mauro's patches. I have added a reminder to my calendar. Thanx, Paul
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
On Tue, Oct 13, 2020 at 09:33:54AM -0700, Paul E. McKenney wrote: > On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > > - The sysfs.txt file was converted to ReST and renamed; > > - The control-dependencies.txt is not at > > Documentation/control-dependencies.txt. As it is at the > > same dir as the README file, which mentions it, just > > remove Documentation/. > > > > With that, ./scripts/documentation-file-ref-check script > > is now happy again for files under tools/. > > > > Signed-off-by: Mauro Carvalho Chehab > > Queued for review and testing, likely target v5.11. Instead of changing the path in the README reference, shouldn't tools/memory-model/control-dependencies.txt be moved to its proper position in .../Documentation? Alan Stern
Re: [PATCH v2 02/24] tools: docs: memory-model: fix references for some files
On Tue, Oct 13, 2020 at 02:14:29PM +0200, Mauro Carvalho Chehab wrote: > - The sysfs.txt file was converted to ReST and renamed; > - The control-dependencies.txt is not at > Documentation/control-dependencies.txt. As it is at the > same dir as the README file, which mentions it, just > remove Documentation/. > > With that, ./scripts/documentation-file-ref-check script > is now happy again for files under tools/. > > Signed-off-by: Mauro Carvalho Chehab Queued for review and testing, likely target v5.11. Thanx, Paul > --- > tools/memory-model/Documentation/README | 2 +- > tools/memory-model/Documentation/ordering.txt | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/tools/memory-model/Documentation/README > b/tools/memory-model/Documentation/README > index 16177aaa9752..004969992bac 100644 > --- a/tools/memory-model/Documentation/README > +++ b/tools/memory-model/Documentation/README > @@ -55,7 +55,7 @@ README > Documentation/cheatsheet.txt > Quick-reference guide to the Linux-kernel memory model. > > -Documentation/control-dependencies.txt > +control-dependencies.txt > A guide to preventing compiler optimizations from destroying > your control dependencies. > > diff --git a/tools/memory-model/Documentation/ordering.txt > b/tools/memory-model/Documentation/ordering.txt > index 3d020bed8585..629b19ae64a6 100644 > --- a/tools/memory-model/Documentation/ordering.txt > +++ b/tools/memory-model/Documentation/ordering.txt > @@ -346,7 +346,7 @@ o Accessing RCU-protected pointers via rcu_dereference() > > If there is any significant processing of the pointer value > between the rcu_dereference() that returned it and a later > - dereference(), please read Documentation/RCU/rcu_dereference.txt. > + dereference(), please read Documentation/RCU/rcu_dereference.rst. > > It can also be quite helpful to review uses in the Linux kernel. > > -- > 2.26.2 >
[PATCH v2 02/24] tools: docs: memory-model: fix references for some files
- The sysfs.txt file was converted to ReST and renamed; - The control-dependencies.txt is not at Documentation/control-dependencies.txt. As it is at the same dir as the README file, which mentions it, just remove Documentation/. With that, ./scripts/documentation-file-ref-check script is now happy again for files under tools/. Signed-off-by: Mauro Carvalho Chehab --- tools/memory-model/Documentation/README | 2 +- tools/memory-model/Documentation/ordering.txt | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/memory-model/Documentation/README b/tools/memory-model/Documentation/README index 16177aaa9752..004969992bac 100644 --- a/tools/memory-model/Documentation/README +++ b/tools/memory-model/Documentation/README @@ -55,7 +55,7 @@ README Documentation/cheatsheet.txt Quick-reference guide to the Linux-kernel memory model. -Documentation/control-dependencies.txt +control-dependencies.txt A guide to preventing compiler optimizations from destroying your control dependencies. diff --git a/tools/memory-model/Documentation/ordering.txt b/tools/memory-model/Documentation/ordering.txt index 3d020bed8585..629b19ae64a6 100644 --- a/tools/memory-model/Documentation/ordering.txt +++ b/tools/memory-model/Documentation/ordering.txt @@ -346,7 +346,7 @@ o Accessing RCU-protected pointers via rcu_dereference() If there is any significant processing of the pointer value between the rcu_dereference() that returned it and a later - dereference(), please read Documentation/RCU/rcu_dereference.txt. + dereference(), please read Documentation/RCU/rcu_dereference.rst. It can also be quite helpful to review uses in the Linux kernel. -- 2.26.2