[PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Heiko Carstens]

v2:
As requested by Greg KH: avoid negative logic and use uppercase defines.

v1:
Create die and cluster cpu topology sysfs attributes only if an
architecture makes uses of it, instead of creating them always for all
architectures with bogus default values.
Also change the book and drawer cpu topology macros so they match all
all other topology macros.

v1: https://lore.kernel.org/lkml/20211128212221.1069726-1-hca@linux.ibm.com/

Heiko Carstens (3):
  topology/sysfs: export die attributes only if an architectures has support
  topology/sysfs: export cluster attributes only if an architectures has support
  topology/sysfs: rework book and drawer topology ifdefery

 Documentation/admin-guide/cputopology.rst | 33 +++++++++++------------
 drivers/base/topology.c                   | 28 ++++++++++++++-----
 include/linux/topology.h                  | 25 +++++++++++++++++
 3 files changed, 62 insertions(+), 24 deletions(-)

-- 
2.32.0

[PATCH v2 1/3] topology/sysfs: export die attributes only if an architectures has support

[Heiko Carstens]

The die_id and die_cpus topology sysfs attributes have been added with
commit 0e344d8c709f ("cpu/topology: Export die_id") and commit
2e4c54dac7b3 ("topology: Create core_cpus and die_cpus sysfs attributes").

While they are currently only used and useful for x86 they are still
present with bogus default values for all architectures. Instead of
enforcing such new sysfs attributes to all architectures, make them
only optional visible if an architecture opts in by defining both the
topology_die_id and topology_die_cpumask attributes.

This is similar to what was done when the book and drawer topology
levels were introduced: avoid useless and therefore confusing sysfs
attributes for architectures which cannot make use of them.

This should not break any existing applications, since this is a
rather new interface and applications should be able to handle also
older kernel versions without such attributes - besides that they
contain only useful information for x86.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 Documentation/admin-guide/cputopology.rst | 3 +++
 drivers/base/topology.c                   | 8 ++++++++
 include/linux/topology.h                  | 4 ++++
 3 files changed, 15 insertions(+)

diff --git a/Documentation/admin-guide/cputopology.rst b/Documentation/admin-guide/cputopology.rst
index 6b62e182baf4..c68d07533c45 100644
--- a/Documentation/admin-guide/cputopology.rst
+++ b/Documentation/admin-guide/cputopology.rst
@@ -11,6 +11,9 @@ Architecture-neutral, drivers/base/topology.c, exports these attributes.
 However, the book and drawer related sysfs files will only be created if
 CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are selected, respectively.
 
+The die hierarchy related sysfs files will only be created if an architecture
+provides the related macros as described below.
+
 CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are currently only used on s390,
 where they reflect the cpu and cache hierarchy.
 
diff --git a/drivers/base/topology.c b/drivers/base/topology.c
index 8f2b641d0b8c..f079a55793ec 100644
--- a/drivers/base/topology.c
+++ b/drivers/base/topology.c
@@ -45,8 +45,10 @@ static ssize_t name##_list_read(struct file *file, struct kobject *kobj,	\
 define_id_show_func(physical_package_id);
 static DEVICE_ATTR_RO(physical_package_id);
 
+#ifdef TOPOLOGY_DIE_SYSFS
 define_id_show_func(die_id);
 static DEVICE_ATTR_RO(die_id);
+#endif
 
 define_id_show_func(cluster_id);
 static DEVICE_ATTR_RO(cluster_id);
@@ -70,9 +72,11 @@ define_siblings_read_func(cluster_cpus, cluster_cpumask);
 static BIN_ATTR_RO(cluster_cpus, 0);
 static BIN_ATTR_RO(cluster_cpus_list, 0);
 
+#ifdef TOPOLOGY_DIE_SYSFS
 define_siblings_read_func(die_cpus, die_cpumask);
 static BIN_ATTR_RO(die_cpus, 0);
 static BIN_ATTR_RO(die_cpus_list, 0);
+#endif
 
 define_siblings_read_func(package_cpus, core_cpumask);
 static BIN_ATTR_RO(package_cpus, 0);
@@ -103,8 +107,10 @@ static struct bin_attribute *bin_attrs[] = {
 	&bin_attr_core_siblings_list,
 	&bin_attr_cluster_cpus,
 	&bin_attr_cluster_cpus_list,
+#ifdef TOPOLOGY_DIE_SYSFS
 	&bin_attr_die_cpus,
 	&bin_attr_die_cpus_list,
+#endif
 	&bin_attr_package_cpus,
 	&bin_attr_package_cpus_list,
 #ifdef CONFIG_SCHED_BOOK
@@ -120,7 +126,9 @@ static struct bin_attribute *bin_attrs[] = {
 
 static struct attribute *default_attrs[] = {
 	&dev_attr_physical_package_id.attr,
+#ifdef TOPOLOGY_DIE_SYSFS
 	&dev_attr_die_id.attr,
+#endif
 	&dev_attr_cluster_id.attr,
 	&dev_attr_core_id.attr,
 #ifdef CONFIG_SCHED_BOOK
diff --git a/include/linux/topology.h b/include/linux/topology.h
index 0b3704ad13c8..8d1bdae76230 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -180,6 +180,10 @@ static inline int cpu_to_mem(int cpu)
 
 #endif	/* [!]CONFIG_HAVE_MEMORYLESS_NODES */
 
+#if defined(topology_die_id) && defined(topology_die_cpumask)
+#define TOPOLOGY_DIE_SYSFS
+#endif
+
 #ifndef topology_physical_package_id
 #define topology_physical_package_id(cpu)	((void)(cpu), -1)
 #endif
-- 
2.32.0

[PATCH v2 2/3] topology/sysfs: export cluster attributes only if an architectures has support

[Heiko Carstens]

The cluster_id and cluster_cpus topology sysfs attributes have been
added with commit c5e22feffdd7 ("topology: Represent clusters of CPUs
within a die").

They are currently only used for x86, arm64, and riscv (via generic
arch topology), however they are still present with bogus default
values for all other architectures. Instead of enforcing such new
sysfs attributes to all architectures, make them only optional visible
if an architecture opts in by defining both the topology_cluster_id
and topology_cluster_cpumask attributes.

This is similar to what was done when the book and drawer topology
levels were introduced: avoid useless and therefore confusing sysfs
attributes for architectures which cannot make use of them.

This should not break any existing applications, since this is a
new interface introduced with the v5.16 merge window.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 Documentation/admin-guide/cputopology.rst | 4 ++--
 drivers/base/topology.c                   | 8 ++++++++
 include/linux/topology.h                  | 3 +++
 3 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/Documentation/admin-guide/cputopology.rst b/Documentation/admin-guide/cputopology.rst
index c68d07533c45..ad2238b41439 100644
--- a/Documentation/admin-guide/cputopology.rst
+++ b/Documentation/admin-guide/cputopology.rst
@@ -11,8 +11,8 @@ Architecture-neutral, drivers/base/topology.c, exports these attributes.
 However, the book and drawer related sysfs files will only be created if
 CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are selected, respectively.
 
-The die hierarchy related sysfs files will only be created if an architecture
-provides the related macros as described below.
+The die and cluster hierarchy related sysfs files will only be created if an
+architecture provides the related macros as described below.
 
 CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are currently only used on s390,
 where they reflect the cpu and cache hierarchy.
diff --git a/drivers/base/topology.c b/drivers/base/topology.c
index f079a55793ec..9d049724e4b4 100644
--- a/drivers/base/topology.c
+++ b/drivers/base/topology.c
@@ -50,8 +50,10 @@ define_id_show_func(die_id);
 static DEVICE_ATTR_RO(die_id);
 #endif
 
+#ifdef TOPOLOGY_CLUSTER_SYSFS
 define_id_show_func(cluster_id);
 static DEVICE_ATTR_RO(cluster_id);
+#endif
 
 define_id_show_func(core_id);
 static DEVICE_ATTR_RO(core_id);
@@ -68,9 +70,11 @@ define_siblings_read_func(core_siblings, core_cpumask);
 static BIN_ATTR_RO(core_siblings, 0);
 static BIN_ATTR_RO(core_siblings_list, 0);
 
+#ifdef TOPOLOGY_CLUSTER_SYSFS
 define_siblings_read_func(cluster_cpus, cluster_cpumask);
 static BIN_ATTR_RO(cluster_cpus, 0);
 static BIN_ATTR_RO(cluster_cpus_list, 0);
+#endif
 
 #ifdef TOPOLOGY_DIE_SYSFS
 define_siblings_read_func(die_cpus, die_cpumask);
@@ -105,8 +109,10 @@ static struct bin_attribute *bin_attrs[] = {
 	&bin_attr_thread_siblings_list,
 	&bin_attr_core_siblings,
 	&bin_attr_core_siblings_list,
+#ifdef TOPOLOGY_CLUSTER_SYSFS
 	&bin_attr_cluster_cpus,
 	&bin_attr_cluster_cpus_list,
+#endif
 #ifdef TOPOLOGY_DIE_SYSFS
 	&bin_attr_die_cpus,
 	&bin_attr_die_cpus_list,
@@ -129,7 +135,9 @@ static struct attribute *default_attrs[] = {
 #ifdef TOPOLOGY_DIE_SYSFS
 	&dev_attr_die_id.attr,
 #endif
+#ifdef TOPOLOGY_CLUSTER_SYSFS
 	&dev_attr_cluster_id.attr,
+#endif
 	&dev_attr_core_id.attr,
 #ifdef CONFIG_SCHED_BOOK
 	&dev_attr_book_id.attr,
diff --git a/include/linux/topology.h b/include/linux/topology.h
index 8d1bdae76230..d52be69037db 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -183,6 +183,9 @@ static inline int cpu_to_mem(int cpu)
 #if defined(topology_die_id) && defined(topology_die_cpumask)
 #define TOPOLOGY_DIE_SYSFS
 #endif
+#if defined(topology_cluster_id) && defined(topology_cluster_cpumask)
+#define TOPOLOGY_CLUSTER_SYSFS
+#endif
 
 #ifndef topology_physical_package_id
 #define topology_physical_package_id(cpu)	((void)(cpu), -1)
-- 
2.32.0

[PATCH v2 3/3] topology/sysfs: rework book and drawer topology ifdefery

[Heiko Carstens]

Provide default defines for the topology_book_[id|cpumask] and
topology_drawer_[id|cpumask] macros just like for each other topology
level.
This way all topology levels are handled in a similar way. Still the
the book and drawer levels are only used on s390, and also the sysfs
attributes are only created on s390. However other architectures may
opt in if wanted.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 Documentation/admin-guide/cputopology.rst | 36 ++++++++++-------------
 drivers/base/topology.c                   | 12 ++++----
 include/linux/topology.h                  | 18 ++++++++++++
 3 files changed, 39 insertions(+), 27 deletions(-)

diff --git a/Documentation/admin-guide/cputopology.rst b/Documentation/admin-guide/cputopology.rst
index ad2238b41439..677ba1c2c820 100644
--- a/Documentation/admin-guide/cputopology.rst
+++ b/Documentation/admin-guide/cputopology.rst
@@ -8,14 +8,9 @@ to /proc/cpuinfo output of some architectures. They reside in
 Documentation/ABI/stable/sysfs-devices-system-cpu.
 
 Architecture-neutral, drivers/base/topology.c, exports these attributes.
-However, the book and drawer related sysfs files will only be created if
-CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are selected, respectively.
-
-The die and cluster hierarchy related sysfs files will only be created if an
-architecture provides the related macros as described below.
-
-CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are currently only used on s390,
-where they reflect the cpu and cache hierarchy.
+However the die, cluster, book, and drawer hierarchy related sysfs files will
+only be created if an architecture provides the related macros as described
+below.
 
 For an architecture to support this feature, it must define some of
 these macros in include/asm-XXX/topology.h::
@@ -42,19 +37,18 @@ To be consistent on all architectures, include/linux/topology.h
 provides default definitions for any of the above macros that are
 not defined by include/asm-XXX/topology.h:
 
-1) topology_physical_package_id: -1
-2) topology_die_id: -1
-3) topology_cluster_id: -1
-4) topology_core_id: 0
-5) topology_sibling_cpumask: just the given CPU
-6) topology_core_cpumask: just the given CPU
-7) topology_cluster_cpumask: just the given CPU
-8) topology_die_cpumask: just the given CPU
-
-For architectures that don't support books (CONFIG_SCHED_BOOK) there are no
-default definitions for topology_book_id() and topology_book_cpumask().
-For architectures that don't support drawers (CONFIG_SCHED_DRAWER) there are
-no default definitions for topology_drawer_id() and topology_drawer_cpumask().
+ 1) topology_physical_package_id: -1
+ 2) topology_die_id: -1
+ 3) topology_cluster_id: -1
+ 4) topology_core_id: 0
+ 5) topology_book_id: -1
+ 6) topology_drawer_id: -1
+ 7) topology_sibling_cpumask: just the given CPU
+ 8) topology_core_cpumask: just the given CPU
+ 9) topology_cluster_cpumask: just the given CPU
+10) topology_die_cpumask: just the given CPU
+11) topology_book_cpumask:  just the given CPU
+12) topology_drawer_cpumask: just the given CPU
 
 Additionally, CPU topology information is provided under
 /sys/devices/system/cpu and includes these files.  The internal
diff --git a/drivers/base/topology.c b/drivers/base/topology.c
index 9d049724e4b4..fc24e89f9592 100644
--- a/drivers/base/topology.c
+++ b/drivers/base/topology.c
@@ -86,7 +86,7 @@ define_siblings_read_func(package_cpus, core_cpumask);
 static BIN_ATTR_RO(package_cpus, 0);
 static BIN_ATTR_RO(package_cpus_list, 0);
 
-#ifdef CONFIG_SCHED_BOOK
+#ifdef TOPOLOGY_BOOK_SYSFS
 define_id_show_func(book_id);
 static DEVICE_ATTR_RO(book_id);
 define_siblings_read_func(book_siblings, book_cpumask);
@@ -94,7 +94,7 @@ static BIN_ATTR_RO(book_siblings, 0);
 static BIN_ATTR_RO(book_siblings_list, 0);
 #endif
 
-#ifdef CONFIG_SCHED_DRAWER
+#ifdef TOPOLOGY_DRAWER_SYSFS
 define_id_show_func(drawer_id);
 static DEVICE_ATTR_RO(drawer_id);
 define_siblings_read_func(drawer_siblings, drawer_cpumask);
@@ -119,11 +119,11 @@ static struct bin_attribute *bin_attrs[] = {
 #endif
 	&bin_attr_package_cpus,
 	&bin_attr_package_cpus_list,
-#ifdef CONFIG_SCHED_BOOK
+#ifdef TOPOLOGY_BOOK_SYSFS
 	&bin_attr_book_siblings,
 	&bin_attr_book_siblings_list,
 #endif
-#ifdef CONFIG_SCHED_DRAWER
+#ifdef TOPOLOGY_DRAWER_SYSFS
 	&bin_attr_drawer_siblings,
 	&bin_attr_drawer_siblings_list,
 #endif
@@ -139,10 +139,10 @@ static struct attribute *default_attrs[] = {
 	&dev_attr_cluster_id.attr,
 #endif
 	&dev_attr_core_id.attr,
-#ifdef CONFIG_SCHED_BOOK
+#ifdef TOPOLOGY_BOOK_SYSFS
 	&dev_attr_book_id.attr,
 #endif
-#ifdef CONFIG_SCHED_DRAWER
+#ifdef TOPOLOGY_DRAWER_SYSFS
 	&dev_attr_drawer_id.attr,
 #endif
 	NULL
diff --git a/include/linux/topology.h b/include/linux/topology.h
index d52be69037db..a6e201758ae9 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -186,6 +186,12 @@ static inline int cpu_to_mem(int cpu)
 #if defined(topology_cluster_id) && defined(topology_cluster_cpumask)
 #define TOPOLOGY_CLUSTER_SYSFS
 #endif
+#if defined(topology_book_id) && defined(topology_book_cpumask)
+#define TOPOLOGY_BOOK_SYSFS
+#endif
+#if defined(topology_drawer_id) && defined(topology_drawer_cpumask)
+#define TOPOLOGY_DRAWER_SYSFS
+#endif
 
 #ifndef topology_physical_package_id
 #define topology_physical_package_id(cpu)	((void)(cpu), -1)
@@ -199,6 +205,12 @@ static inline int cpu_to_mem(int cpu)
 #ifndef topology_core_id
 #define topology_core_id(cpu)			((void)(cpu), 0)
 #endif
+#ifndef topology_book_id
+#define topology_book_id(cpu)			((void)(cpu), -1)
+#endif
+#ifndef topology_drawer_id
+#define topology_drawer_id(cpu)			((void)(cpu), -1)
+#endif
 #ifndef topology_sibling_cpumask
 #define topology_sibling_cpumask(cpu)		cpumask_of(cpu)
 #endif
@@ -211,6 +223,12 @@ static inline int cpu_to_mem(int cpu)
 #ifndef topology_die_cpumask
 #define topology_die_cpumask(cpu)		cpumask_of(cpu)
 #endif
+#ifndef topology_book_cpumask
+#define topology_book_cpumask(cpu)		cpumask_of(cpu)
+#endif
+#ifndef topology_drawer_cpumask
+#define topology_drawer_cpumask(cpu)		cpumask_of(cpu)
+#endif
 
 #if defined(CONFIG_SCHED_SMT) && !defined(cpu_smt_mask)
 static inline const struct cpumask *cpu_smt_mask(int cpu)
-- 
2.32.0

Re: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Peter Zijlstra]

On Mon, Nov 29, 2021 at 02:03:06PM +0100, Heiko Carstens wrote:
> v2:
> As requested by Greg KH: avoid negative logic and use uppercase defines.
> 
> v1:
> Create die and cluster cpu topology sysfs attributes only if an
> architecture makes uses of it, instead of creating them always for all
> architectures with bogus default values.
> Also change the book and drawer cpu topology macros so they match all
> all other topology macros.
> 
> v1: https://lore.kernel.org/lkml/20211128212221.1069726-1-hca@linux.ibm.com/
> 
> Heiko Carstens (3):
>   topology/sysfs: export die attributes only if an architectures has support
>   topology/sysfs: export cluster attributes only if an architectures has support
>   topology/sysfs: rework book and drawer topology ifdefery
> 
>  Documentation/admin-guide/cputopology.rst | 33 +++++++++++------------
>  drivers/base/topology.c                   | 28 ++++++++++++++-----
>  include/linux/topology.h                  | 25 +++++++++++++++++
>  3 files changed, 62 insertions(+), 24 deletions(-)

Seems entirely reasonable to me,

Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>

Re: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Heiko Carstens]

On Mon, Nov 29, 2021 at 03:03:18PM +0100, Peter Zijlstra wrote:
> On Mon, Nov 29, 2021 at 02:03:06PM +0100, Heiko Carstens wrote:
> > v2:
> > As requested by Greg KH: avoid negative logic and use uppercase defines.
> > 
> > v1:
> > Create die and cluster cpu topology sysfs attributes only if an
> > architecture makes uses of it, instead of creating them always for all
> > architectures with bogus default values.
> > Also change the book and drawer cpu topology macros so they match all
> > all other topology macros.
> > 
> > v1: https://lore.kernel.org/lkml/20211128212221.1069726-1-hca@linux.ibm.com/
> > 
> > Heiko Carstens (3):
> >   topology/sysfs: export die attributes only if an architectures has support
> >   topology/sysfs: export cluster attributes only if an architectures has support
> >   topology/sysfs: rework book and drawer topology ifdefery
> > 
> >  Documentation/admin-guide/cputopology.rst | 33 +++++++++++------------
> >  drivers/base/topology.c                   | 28 ++++++++++++++-----
> >  include/linux/topology.h                  | 25 +++++++++++++++++
> >  3 files changed, 62 insertions(+), 24 deletions(-)
> 
> Seems entirely reasonable to me,
> 
> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>

Since there seem to be no objections, who should pick this up?
I'd assume the s390 tree would not be appropriate.

Andrew, Greg, Peter?

Re: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Greg Kroah-Hartman]

On Thu, Dec 02, 2021 at 03:22:05PM +0100, Heiko Carstens wrote:
> On Mon, Nov 29, 2021 at 03:03:18PM +0100, Peter Zijlstra wrote:
> > On Mon, Nov 29, 2021 at 02:03:06PM +0100, Heiko Carstens wrote:
> > > v2:
> > > As requested by Greg KH: avoid negative logic and use uppercase defines.
> > > 
> > > v1:
> > > Create die and cluster cpu topology sysfs attributes only if an
> > > architecture makes uses of it, instead of creating them always for all
> > > architectures with bogus default values.
> > > Also change the book and drawer cpu topology macros so they match all
> > > all other topology macros.
> > > 
> > > v1: https://lore.kernel.org/lkml/20211128212221.1069726-1-hca@linux.ibm.com/
> > > 
> > > Heiko Carstens (3):
> > >   topology/sysfs: export die attributes only if an architectures has support
> > >   topology/sysfs: export cluster attributes only if an architectures has support
> > >   topology/sysfs: rework book and drawer topology ifdefery
> > > 
> > >  Documentation/admin-guide/cputopology.rst | 33 +++++++++++------------
> > >  drivers/base/topology.c                   | 28 ++++++++++++++-----
> > >  include/linux/topology.h                  | 25 +++++++++++++++++
> > >  3 files changed, 62 insertions(+), 24 deletions(-)
> > 
> > Seems entirely reasonable to me,
> > 
> > Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> 
> Since there seem to be no objections, who should pick this up?
> I'd assume the s390 tree would not be appropriate.
> 
> Andrew, Greg, Peter?

I will take it, thanks.

greg k-h

RE: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Song Bao Hua (Barry Song)]

> -----Original Message-----
> From: Greg Kroah-Hartman [mailto:gregkh@linuxfoundation.org]
> Sent: Friday, December 3, 2021 3:27 AM
> To: Heiko Carstens <hca@linux.ibm.com>
> Cc: Peter Zijlstra <peterz@infradead.org>; Rafael J . Wysocki
> <rafael@kernel.org>; Andrew Morton <akpm@linux-foundation.org>;
> linux-kernel@vger.kernel.org; Jonathan Cameron <jonathan.cameron@huawei.com>;
> Len Brown <len.brown@intel.com>; Thomas Richter <tmricht@linux.ibm.com>; Ian
> Rogers <irogers@google.com>
> Subject: Re: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes
> 
> On Thu, Dec 02, 2021 at 03:22:05PM +0100, Heiko Carstens wrote:
> > On Mon, Nov 29, 2021 at 03:03:18PM +0100, Peter Zijlstra wrote:
> > > On Mon, Nov 29, 2021 at 02:03:06PM +0100, Heiko Carstens wrote:
> > > > v2:
> > > > As requested by Greg KH: avoid negative logic and use uppercase defines.
> > > >
> > > > v1:
> > > > Create die and cluster cpu topology sysfs attributes only if an
> > > > architecture makes uses of it, instead of creating them always for all
> > > > architectures with bogus default values.
> > > > Also change the book and drawer cpu topology macros so they match all
> > > > all other topology macros.
> > > >
> > > > v1:
> https://lore.kernel.org/lkml/20211128212221.1069726-1-hca@linux.ibm.com/
> > > >
> > > > Heiko Carstens (3):
> > > >   topology/sysfs: export die attributes only if an architectures has
> support
> > > >   topology/sysfs: export cluster attributes only if an architectures has
> support
> > > >   topology/sysfs: rework book and drawer topology ifdefery
> > > >
> > > >  Documentation/admin-guide/cputopology.rst | 33 +++++++++++------------
> > > >  drivers/base/topology.c                   | 28 ++++++++++++++-----
> > > >  include/linux/topology.h                  | 25 +++++++++++++++++
> > > >  3 files changed, 62 insertions(+), 24 deletions(-)
> > >
> > > Seems entirely reasonable to me,
> > >
> > > Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> >
> > Since there seem to be no objections, who should pick this up?
> > I'd assume the s390 tree would not be appropriate.
> >
> > Andrew, Greg, Peter?
> 
> I will take it, thanks.

Could you give me one minute?

+Brice

I'd like to hear some feedbacks from hwloc if there
is a chance to break userspace and if userspace depends
on the existence of sysfs even though the topology
doesn't exist.

If no, I feel it is safe to take.

> 
> greg k-h

Thanks
Barry

Re: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Brice Goglin]

[-- Attachment #1.1: Type: text/plain, Size: 547 bytes --]

Le 04/12/2021 à 10:07, Song Bao Hua (Barry Song) a écrit :
>
> Could you give me one minute?
>
> +Brice
>
> I'd like to hear some feedbacks from hwloc if there
> is a chance to break userspace and if userspace depends
> on the existence of sysfs even though the topology
> doesn't exist.
>
> If no, I feel it is safe to take.
>

Hello

If the question is whether hwloc *requires* cluster/die/book/drawer 
sysfs files to exist, then the answer is no. We have to support old 
kernels without those files anyway.

Brice




[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 840 bytes --]

RE: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes

[Song Bao Hua (Barry Song)]

> -----Original Message-----
> From: Brice Goglin [mailto:Brice.Goglin@inria.fr]
> Sent: Saturday, December 4, 2021 10:48 PM
> To: Song Bao Hua (Barry Song) <song.bao.hua@hisilicon.com>; Greg Kroah-Hartman
> <gregkh@linuxfoundation.org>; Heiko Carstens <hca@linux.ibm.com>
> Cc: Peter Zijlstra <peterz@infradead.org>; Rafael J . Wysocki
> <rafael@kernel.org>; Andrew Morton <akpm@linux-foundation.org>;
> linux-kernel@vger.kernel.org; Jonathan Cameron <jonathan.cameron@huawei.com>;
> Len Brown <len.brown@intel.com>; Thomas Richter <tmricht@linux.ibm.com>; Ian
> Rogers <irogers@google.com>
> Subject: Re: [PATCH v2 0/3] topology/sysfs: only export used sysfs attributes
> 
> Le 04/12/2021 à 10:07, Song Bao Hua (Barry Song) a écrit :
> >
> > Could you give me one minute?
> >
> > +Brice
> >
> > I'd like to hear some feedbacks from hwloc if there
> > is a chance to break userspace and if userspace depends
> > on the existence of sysfs even though the topology
> > doesn't exist.
> >
> > If no, I feel it is safe to take.
> >
> 
> Hello
> 
> If the question is whether hwloc *requires* cluster/die/book/drawer
> sysfs files to exist, then the answer is no. We have to support old
> kernels without those files anyway.

Thanks for clarification, Brice. I saw hwloc is reading sysfs
drawer, book and cluster. But since it doesn't assume their
existence in sysfs, I have more confidence on this patchset
now.

Greg, please ignore my comment then :-)

> 
> Brice
> 
> 

Thanks
Barry