| Message ID | 20171114103138.9931-1-mark.rutland@arm.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | perf mmap: Convert ACCESS_ONCE() to READ_ONCE() | expand |
On Tue, Nov 14, 2017 at 10:31:38AM +0000, Mark Rutland wrote: > Recently there was a treewide conversion of ACCESS_ONCE() to > {READ,WRITE}_ONCE(), but a new use was introduced concurrently by > commit: > > 1695849735752d2a ("perf mmap: Move perf_mmap and methods to separate mmap.[ch] files") > > Let's convert this over to READ_ONCE() so that we can remove the > ACCESS_ONCE() definitions in subsequent patches. > > Signed-off-by: Mark Rutland <mark.rutland@arm.com> > Cc: Arnaldo Carvalho de Melo <acme@redhat.com> > Cc: Ingo Molnar <mingo@redhat.com> > Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > Cc: Peter Zijlstra <peterz@infradead.org> Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Woo-hoo! Good to see that your Coccinelle script has already done its magic! ;-) Thanx, Paul > --- > tools/perf/util/mmap.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > Hi, > > Would it be possible for this to be taken as a fixup in an upcoming rc? That > way we should be able to remove the ACCESS_ONCE() definitions prior to v4.15. > > Thanks, > Mark. > > diff --git a/tools/perf/util/mmap.h b/tools/perf/util/mmap.h > index efd78b827b05..3a5cb5a6e94a 100644 > --- a/tools/perf/util/mmap.h > +++ b/tools/perf/util/mmap.h > @@ -70,7 +70,7 @@ void perf_mmap__read_catchup(struct perf_mmap *md); > static inline u64 perf_mmap__read_head(struct perf_mmap *mm) > { > struct perf_event_mmap_page *pc = mm->base; > - u64 head = ACCESS_ONCE(pc->data_head); > + u64 head = READ_ONCE(pc->data_head); > rmb(); > return head; > } > -- > 2.11.0 >
On Tue, 2017-11-14 at 14:08 -0800, Paul E. McKenney wrote: > On Tue, Nov 14, 2017 at 10:31:38AM +0000, Mark Rutland wrote: > > Recently there was a treewide conversion of ACCESS_ONCE() to > > {READ,WRITE}_ONCE(), but a new use was introduced concurrently by > > commit: > > > > 1695849735752d2a ("perf mmap: Move perf_mmap and methods to separate mmap.[ch] files") > > > > Let's convert this over to READ_ONCE() so that we can remove the > > ACCESS_ONCE() definitions in subsequent patches. > > > > Signed-off-by: Mark Rutland <mark.rutland@arm.com> > > Cc: Arnaldo Carvalho de Melo <acme@redhat.com> > > Cc: Ingo Molnar <mingo@redhat.com> > > Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > Cc: Peter Zijlstra <peterz@infradead.org> > > Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > Woo-hoo! Good to see that your Coccinelle script has already done > its magic! ;-) Might be nice to remove all the other references too $ git grep -w ACCESS_ONCE Documentation/RCU/RTFP.txt: ACCESS_ONCE(). include/linux/compiler.h: * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the include/linux/compiler.h: * WRITE_ONCE or ACCESS_ONCE() in different C statements. include/linux/compiler.h: * In contrast to ACCESS_ONCE these two macros will also work on aggregate include/linux/compiler.h: * is also forbidden from reordering successive instances of ACCESS_ONCE(), include/linux/compiler.h: * ACCESS_ONCE() in different C statements. include/linux/compiler.h: * ACCESS_ONCE will only work on scalar types. For union types, ACCESS_ONCE include/linux/compiler.h: * The major use cases of ACCESS_ONCE used to be (1) Mediating communication include/linux/compiler.h:#define ACCESS_ONCE(x) (*__ACCESS_ONCE(x)) scripts/checkpatch.pl:# whine about ACCESS_ONCE scripts/checkpatch.pl: "Prefer WRITE_ONCE(<FOO>, <BAR>) over ACCESS_ONCE(<FOO>) = <BAR>\n" . $herecurr) && scripts/checkpatch.pl: "Prefer READ_ONCE(<FOO>) over ACCESS_ONCE(<FOO>)\n" . $herecurr) && tools/include/linux/compiler.h:#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x)) tools/include/linux/compiler.h: * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the tools/include/linux/compiler.h: * WRITE_ONCE or ACCESS_ONCE() in different C statements. tools/include/linux/compiler.h: * In contrast to ACCESS_ONCE these two macros will also work on aggregate tools/perf/util/mmap.h: u64 head = ACCESS_ONCE(pc->data_head); --- include/linux/compiler.h | 45 ++++++++++-------------------------------- scripts/checkpatch.pl | 22 --------------------- tools/include/linux/compiler.h | 19 ++++++++---------- 3 files changed, 18 insertions(+), 68 deletions(-) diff --git a/include/linux/compiler.h b/include/linux/compiler.h index 3672353a0acd..f729154dae9b 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -215,17 +215,17 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s /* * Prevent the compiler from merging or refetching reads or writes. The * compiler is also forbidden from reordering successive instances of - * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the - * compiler is aware of some particular ordering. One way to make the - * compiler aware of ordering is to put the two invocations of READ_ONCE, - * WRITE_ONCE or ACCESS_ONCE() in different C statements. + * READ_ONCE and WRITE_ONCE (see below), but only when the compiler is + * aware of some particular ordering. One way to make the compiler aware + * of ordering is to put the two invocations of READ_ONCE and WRITE_ONCE + * in different C statements. * - * In contrast to ACCESS_ONCE these two macros will also work on aggregate - * data types like structs or unions. If the size of the accessed data - * type exceeds the word size of the machine (e.g., 32 bits or 64 bits) - * READ_ONCE() and WRITE_ONCE() will fall back to memcpy(). There's at - * least two memcpy()s: one for the __builtin_memcpy() and then one for - * the macro doing the copy of variable - '__u' allocated on the stack. + * These two macros will work on aggregate data types like structs or unions. + * If the size of the accessed data type exceeds the word size of the machine + * (e.g., 32 bits or 64 bits) READ_ONCE() and WRITE_ONCE() will fall back to + * memcpy(). There are at least two memcpy()s: one for the __builtin_memcpy() + * and then one for the macro doing the copy of variable - '__u' allocated on + * the stack. * * Their two major use cases are: (1) Mediating communication between * process-level code and irq/NMI handlers, all running on the same CPU, @@ -322,29 +322,4 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s compiletime_assert(__native_word(t), \ "Need native word sized stores/loads for atomicity.") -/* - * Prevent the compiler from merging or refetching accesses. The compiler - * is also forbidden from reordering successive instances of ACCESS_ONCE(), - * but only when the compiler is aware of some particular ordering. One way - * to make the compiler aware of ordering is to put the two invocations of - * ACCESS_ONCE() in different C statements. - * - * ACCESS_ONCE will only work on scalar types. For union types, ACCESS_ONCE - * on a union member will work as long as the size of the member matches the - * size of the union and the size is smaller than word size. - * - * The major use cases of ACCESS_ONCE used to be (1) Mediating communication - * between process-level code and irq/NMI handlers, all running on the same CPU, - * and (2) Ensuring that the compiler does not fold, spindle, or otherwise - * mutilate accesses that either do not require ordering or that interact - * with an explicit memory barrier or atomic instruction that provides the - * required ordering. - * - * If possible use READ_ONCE()/WRITE_ONCE() instead. - */ -#define __ACCESS_ONCE(x) ({ \ - __maybe_unused typeof(x) __var = (__force typeof(x)) 0; \ - (volatile typeof(x) *)&(x); }) -#define ACCESS_ONCE(x) (*__ACCESS_ONCE(x)) - #endif /* __LINUX_COMPILER_H */ diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl index 8b80bac055e4..fffe1e5895a2 100755 --- a/scripts/checkpatch.pl +++ b/scripts/checkpatch.pl @@ -6242,28 +6242,6 @@ sub process { } } -# whine about ACCESS_ONCE - if ($^V && $^V ge 5.10.0 && - $line =~ /\bACCESS_ONCE\s*$balanced_parens\s*(=(?!=))?\s*($FuncArg)?/) { - my $par = $1; - my $eq = $2; - my $fun = $3; - $par =~ s/^\(\s*(.*)\s*\)$/$1/; - if (defined($eq)) { - if (WARN("PREFER_WRITE_ONCE", - "Prefer WRITE_ONCE(<FOO>, <BAR>) over ACCESS_ONCE(<FOO>) = <BAR>\n" . $herecurr) && - $fix) { - $fixed[$fixlinenr] =~ s/\bACCESS_ONCE\s*\(\s*\Q$par\E\s*\)\s*$eq\s*\Q$fun\E/WRITE_ONCE($par, $fun)/; - } - } else { - if (WARN("PREFER_READ_ONCE", - "Prefer READ_ONCE(<FOO>) over ACCESS_ONCE(<FOO>)\n" . $herecurr) && - $fix) { - $fixed[$fixlinenr] =~ s/\bACCESS_ONCE\s*\(\s*\Q$par\E\s*\)/READ_ONCE($par)/; - } - } - } - # check for mutex_trylock_recursive usage if ($line =~ /mutex_trylock_recursive/) { ERROR("LOCKING", diff --git a/tools/include/linux/compiler.h b/tools/include/linux/compiler.h index 07fd03c74a77..cb77706af769 100644 --- a/tools/include/linux/compiler.h +++ b/tools/include/linux/compiler.h @@ -84,8 +84,6 @@ #define uninitialized_var(x) x = *(&(x)) -#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x)) - #include <linux/types.h> /* @@ -135,16 +133,15 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s /* * Prevent the compiler from merging or refetching reads or writes. The * compiler is also forbidden from reordering successive instances of - * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the - * compiler is aware of some particular ordering. One way to make the - * compiler aware of ordering is to put the two invocations of READ_ONCE, - * WRITE_ONCE or ACCESS_ONCE() in different C statements. + * READ_ONCE and WRITE_ONCE, but only when the compiler is aware of some + * particular ordering. One way to make the compiler aware of ordering is + * to put the two invocations of READ_ONCE or WRITE_ONCE in different C + * statements. * - * In contrast to ACCESS_ONCE these two macros will also work on aggregate - * data types like structs or unions. If the size of the accessed data - * type exceeds the word size of the machine (e.g., 32 bits or 64 bits) - * READ_ONCE() and WRITE_ONCE() will fall back to memcpy and print a - * compile-time warning. + * These two macros will also work on aggregate data types like structs or + * unions. If the size of the accessed data type exceeds the word size of the + * machine (e.g., 32 bits or 64 bits) READ_ONCE() and WRITE_ONCE() will fall + * back to memcpy and print a compile-time warning. * * Their two major use cases are: (1) Mediating communication between * process-level code and irq/NMI handlers, all running on the same CPU,
On Tue, Nov 14, 2017 at 02:21:56PM -0800, Joe Perches wrote: > On Tue, 2017-11-14 at 14:08 -0800, Paul E. McKenney wrote: > > On Tue, Nov 14, 2017 at 10:31:38AM +0000, Mark Rutland wrote: > > > Recently there was a treewide conversion of ACCESS_ONCE() to > > > {READ,WRITE}_ONCE(), but a new use was introduced concurrently by > > > commit: > > > > > > 1695849735752d2a ("perf mmap: Move perf_mmap and methods to separate mmap.[ch] files") > > > > > > Let's convert this over to READ_ONCE() so that we can remove the > > > ACCESS_ONCE() definitions in subsequent patches. > > > > > > Signed-off-by: Mark Rutland <mark.rutland@arm.com> > > > Cc: Arnaldo Carvalho de Melo <acme@redhat.com> > > > Cc: Ingo Molnar <mingo@redhat.com> > > > Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > > Cc: Peter Zijlstra <peterz@infradead.org> > > > > Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > > > Woo-hoo! Good to see that your Coccinelle script has already done > > its magic! ;-) > > Might be nice to remove all the other references too Agreed, and that is exactly what Mark was referring to in his "remove the ACCESS_ONCE() definitions in subsequent patches". ;-) Thanx, Paul > $ git grep -w ACCESS_ONCE > Documentation/RCU/RTFP.txt: ACCESS_ONCE(). > include/linux/compiler.h: * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the > include/linux/compiler.h: * WRITE_ONCE or ACCESS_ONCE() in different C statements. > include/linux/compiler.h: * In contrast to ACCESS_ONCE these two macros will also work on aggregate > include/linux/compiler.h: * is also forbidden from reordering successive instances of ACCESS_ONCE(), > include/linux/compiler.h: * ACCESS_ONCE() in different C statements. > include/linux/compiler.h: * ACCESS_ONCE will only work on scalar types. For union types, ACCESS_ONCE > include/linux/compiler.h: * The major use cases of ACCESS_ONCE used to be (1) Mediating communication > include/linux/compiler.h:#define ACCESS_ONCE(x) (*__ACCESS_ONCE(x)) > scripts/checkpatch.pl:# whine about ACCESS_ONCE > scripts/checkpatch.pl: "Prefer WRITE_ONCE(<FOO>, <BAR>) over ACCESS_ONCE(<FOO>) = <BAR>\n" . $herecurr) && > scripts/checkpatch.pl: "Prefer READ_ONCE(<FOO>) over ACCESS_ONCE(<FOO>)\n" . $herecurr) && > tools/include/linux/compiler.h:#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x)) > tools/include/linux/compiler.h: * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the > tools/include/linux/compiler.h: * WRITE_ONCE or ACCESS_ONCE() in different C statements. > tools/include/linux/compiler.h: * In contrast to ACCESS_ONCE these two macros will also work on aggregate > tools/perf/util/mmap.h: u64 head = ACCESS_ONCE(pc->data_head); > > --- > include/linux/compiler.h | 45 ++++++++++-------------------------------- > scripts/checkpatch.pl | 22 --------------------- > tools/include/linux/compiler.h | 19 ++++++++---------- > 3 files changed, 18 insertions(+), 68 deletions(-) > > diff --git a/include/linux/compiler.h b/include/linux/compiler.h > index 3672353a0acd..f729154dae9b 100644 > --- a/include/linux/compiler.h > +++ b/include/linux/compiler.h > @@ -215,17 +215,17 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s > /* > * Prevent the compiler from merging or refetching reads or writes. The > * compiler is also forbidden from reordering successive instances of > - * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the > - * compiler is aware of some particular ordering. One way to make the > - * compiler aware of ordering is to put the two invocations of READ_ONCE, > - * WRITE_ONCE or ACCESS_ONCE() in different C statements. > + * READ_ONCE and WRITE_ONCE (see below), but only when the compiler is > + * aware of some particular ordering. One way to make the compiler aware > + * of ordering is to put the two invocations of READ_ONCE and WRITE_ONCE > + * in different C statements. > * > - * In contrast to ACCESS_ONCE these two macros will also work on aggregate > - * data types like structs or unions. If the size of the accessed data > - * type exceeds the word size of the machine (e.g., 32 bits or 64 bits) > - * READ_ONCE() and WRITE_ONCE() will fall back to memcpy(). There's at > - * least two memcpy()s: one for the __builtin_memcpy() and then one for > - * the macro doing the copy of variable - '__u' allocated on the stack. > + * These two macros will work on aggregate data types like structs or unions. > + * If the size of the accessed data type exceeds the word size of the machine > + * (e.g., 32 bits or 64 bits) READ_ONCE() and WRITE_ONCE() will fall back to > + * memcpy(). There are at least two memcpy()s: one for the __builtin_memcpy() > + * and then one for the macro doing the copy of variable - '__u' allocated on > + * the stack. > * > * Their two major use cases are: (1) Mediating communication between > * process-level code and irq/NMI handlers, all running on the same CPU, > @@ -322,29 +322,4 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s > compiletime_assert(__native_word(t), \ > "Need native word sized stores/loads for atomicity.") > > -/* > - * Prevent the compiler from merging or refetching accesses. The compiler > - * is also forbidden from reordering successive instances of ACCESS_ONCE(), > - * but only when the compiler is aware of some particular ordering. One way > - * to make the compiler aware of ordering is to put the two invocations of > - * ACCESS_ONCE() in different C statements. > - * > - * ACCESS_ONCE will only work on scalar types. For union types, ACCESS_ONCE > - * on a union member will work as long as the size of the member matches the > - * size of the union and the size is smaller than word size. > - * > - * The major use cases of ACCESS_ONCE used to be (1) Mediating communication > - * between process-level code and irq/NMI handlers, all running on the same CPU, > - * and (2) Ensuring that the compiler does not fold, spindle, or otherwise > - * mutilate accesses that either do not require ordering or that interact > - * with an explicit memory barrier or atomic instruction that provides the > - * required ordering. > - * > - * If possible use READ_ONCE()/WRITE_ONCE() instead. > - */ > -#define __ACCESS_ONCE(x) ({ \ > - __maybe_unused typeof(x) __var = (__force typeof(x)) 0; \ > - (volatile typeof(x) *)&(x); }) > -#define ACCESS_ONCE(x) (*__ACCESS_ONCE(x)) > - > #endif /* __LINUX_COMPILER_H */ > diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl > index 8b80bac055e4..fffe1e5895a2 100755 > --- a/scripts/checkpatch.pl > +++ b/scripts/checkpatch.pl > @@ -6242,28 +6242,6 @@ sub process { > } > } > > -# whine about ACCESS_ONCE > - if ($^V && $^V ge 5.10.0 && > - $line =~ /\bACCESS_ONCE\s*$balanced_parens\s*(=(?!=))?\s*($FuncArg)?/) { > - my $par = $1; > - my $eq = $2; > - my $fun = $3; > - $par =~ s/^\(\s*(.*)\s*\)$/$1/; > - if (defined($eq)) { > - if (WARN("PREFER_WRITE_ONCE", > - "Prefer WRITE_ONCE(<FOO>, <BAR>) over ACCESS_ONCE(<FOO>) = <BAR>\n" . $herecurr) && > - $fix) { > - $fixed[$fixlinenr] =~ s/\bACCESS_ONCE\s*\(\s*\Q$par\E\s*\)\s*$eq\s*\Q$fun\E/WRITE_ONCE($par, $fun)/; > - } > - } else { > - if (WARN("PREFER_READ_ONCE", > - "Prefer READ_ONCE(<FOO>) over ACCESS_ONCE(<FOO>)\n" . $herecurr) && > - $fix) { > - $fixed[$fixlinenr] =~ s/\bACCESS_ONCE\s*\(\s*\Q$par\E\s*\)/READ_ONCE($par)/; > - } > - } > - } > - > # check for mutex_trylock_recursive usage > if ($line =~ /mutex_trylock_recursive/) { > ERROR("LOCKING", > diff --git a/tools/include/linux/compiler.h b/tools/include/linux/compiler.h > index 07fd03c74a77..cb77706af769 100644 > --- a/tools/include/linux/compiler.h > +++ b/tools/include/linux/compiler.h > @@ -84,8 +84,6 @@ > > #define uninitialized_var(x) x = *(&(x)) > > -#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x)) > - > #include <linux/types.h> > > /* > @@ -135,16 +133,15 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s > /* > * Prevent the compiler from merging or refetching reads or writes. The > * compiler is also forbidden from reordering successive instances of > - * READ_ONCE, WRITE_ONCE and ACCESS_ONCE (see below), but only when the > - * compiler is aware of some particular ordering. One way to make the > - * compiler aware of ordering is to put the two invocations of READ_ONCE, > - * WRITE_ONCE or ACCESS_ONCE() in different C statements. > + * READ_ONCE and WRITE_ONCE, but only when the compiler is aware of some > + * particular ordering. One way to make the compiler aware of ordering is > + * to put the two invocations of READ_ONCE or WRITE_ONCE in different C > + * statements. > * > - * In contrast to ACCESS_ONCE these two macros will also work on aggregate > - * data types like structs or unions. If the size of the accessed data > - * type exceeds the word size of the machine (e.g., 32 bits or 64 bits) > - * READ_ONCE() and WRITE_ONCE() will fall back to memcpy and print a > - * compile-time warning. > + * These two macros will also work on aggregate data types like structs or > + * unions. If the size of the accessed data type exceeds the word size of the > + * machine (e.g., 32 bits or 64 bits) READ_ONCE() and WRITE_ONCE() will fall > + * back to memcpy and print a compile-time warning. > * > * Their two major use cases are: (1) Mediating communication between > * process-level code and irq/NMI handlers, all running on the same CPU, >
diff --git a/tools/perf/util/mmap.h b/tools/perf/util/mmap.h index efd78b827b05..3a5cb5a6e94a 100644 --- a/tools/perf/util/mmap.h +++ b/tools/perf/util/mmap.h @@ -70,7 +70,7 @@ void perf_mmap__read_catchup(struct perf_mmap *md); static inline u64 perf_mmap__read_head(struct perf_mmap *mm) { struct perf_event_mmap_page *pc = mm->base; - u64 head = ACCESS_ONCE(pc->data_head); + u64 head = READ_ONCE(pc->data_head); rmb(); return head; }
Recently there was a treewide conversion of ACCESS_ONCE() to {READ,WRITE}_ONCE(), but a new use was introduced concurrently by commit: 1695849735752d2a ("perf mmap: Move perf_mmap and methods to separate mmap.[ch] files") Let's convert this over to READ_ONCE() so that we can remove the ACCESS_ONCE() definitions in subsequent patches. Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> --- tools/perf/util/mmap.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) Hi, Would it be possible for this to be taken as a fixup in an upcoming rc? That way we should be able to remove the ACCESS_ONCE() definitions prior to v4.15. Thanks, Mark. -- 2.11.0