| Message ID | 20200624101754.169612-3-kjain@linux.ibm.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Add cpu hotplug support for powerpc/perf/hv-24x7 | expand |
| Context | Check | Description |
|---|---|---|
| snowpatch_ozlabs/apply_patch | success | Successfully applied on branch powerpc/merge (3ddb24ac766e1b028462b27aeae05c2bf17c12c5) |
| snowpatch_ozlabs/build-ppc64le | success | Build succeeded |
| snowpatch_ozlabs/build-ppc64be | success | Build succeeded |
| snowpatch_ozlabs/build-ppc64e | success | Build succeeded |
| snowpatch_ozlabs/build-pmac32 | success | Build succeeded |
| snowpatch_ozlabs/checkpatch | warning | total: 0 errors, 4 warnings, 0 checks, 80 lines checked |
| snowpatch_ozlabs/needsstable | success | Patch has no Fixes tags |
Hi Kajol, On Wed, Jun 24, 2020 at 03:47:54PM +0530, Kajol Jain wrote: > Patch here adds a cpumask attr to hv_24x7 pmu along with ABI documentation. > > command:# cat /sys/devices/hv_24x7/cpumask > 0 Since this sysfs interface is read-only, and the user cannot change the CPU which will be making the HCALLs to obtain the 24x7 counts, does the user even need to know if currently CPU X is the one which is going to make HCALLs to retrive the 24x7 counts ? Does it help in any kind of trouble-shooting ? It would have made sense if the interface was read-write, since a user can set this to a CPU which is not running user applications. This would help in minimising jitter on those active CPUs running the user applications. > > Signed-off-by: Kajol Jain <kjain@linux.ibm.com> > --- > .../sysfs-bus-event_source-devices-hv_24x7 | 7 ++++ > arch/powerpc/perf/hv-24x7.c | 36 +++++++++++++++++-- > 2 files changed, 41 insertions(+), 2 deletions(-) > > diff --git a/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 b/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 > index e8698afcd952..f9dd3755b049 100644 > --- a/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 > +++ b/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 > @@ -43,6 +43,13 @@ Description: read only > This sysfs interface exposes the number of cores per chip > present in the system. > > +What: /sys/devices/hv_24x7/cpumask > +Date: June 2020 > +Contact: Linux on PowerPC Developer List <linuxppc-dev@lists.ozlabs.org> > +Description: read only > + This sysfs file exposes the cpumask which is designated to make > + HCALLs to retrive hv-24x7 pmu event counter data. > + > What: /sys/bus/event_source/devices/hv_24x7/event_descs/<event-name> > Date: February 2014 > Contact: Linux on PowerPC Developer List <linuxppc-dev@lists.ozlabs.org> > diff --git a/arch/powerpc/perf/hv-24x7.c b/arch/powerpc/perf/hv-24x7.c > index ce4739e2b407..3c699612d29f 100644 > --- a/arch/powerpc/perf/hv-24x7.c > +++ b/arch/powerpc/perf/hv-24x7.c > @@ -448,6 +448,12 @@ static ssize_t device_show_string(struct device *dev, > return sprintf(buf, "%s\n", (char *)d->var); > } > > +static ssize_t cpumask_get_attr(struct device *dev, > + struct device_attribute *attr, char *buf) > +{ > + return cpumap_print_to_pagebuf(true, buf, &hv_24x7_cpumask); > +} > + > static ssize_t sockets_show(struct device *dev, > struct device_attribute *attr, char *buf) > { > @@ -1116,6 +1122,17 @@ static DEVICE_ATTR_RO(sockets); > static DEVICE_ATTR_RO(chipspersocket); > static DEVICE_ATTR_RO(coresperchip); > > +static DEVICE_ATTR(cpumask, S_IRUGO, cpumask_get_attr, NULL); > + > +static struct attribute *cpumask_attrs[] = { > + &dev_attr_cpumask.attr, > + NULL, > +}; > + > +static struct attribute_group cpumask_attr_group = { > + .attrs = cpumask_attrs, > +}; > + > static struct bin_attribute *if_bin_attrs[] = { > &bin_attr_catalog, > NULL, > @@ -1143,6 +1160,11 @@ static const struct attribute_group *attr_groups[] = { > &event_desc_group, > &event_long_desc_group, > &if_group, > + /* > + * This NULL is a placeholder for the cpumask attr which will update > + * onlyif cpuhotplug registration is successful > + */ > + NULL, > NULL, > }; > > @@ -1683,7 +1705,7 @@ static int hv_24x7_cpu_hotplug_init(void) > > static int hv_24x7_init(void) > { > - int r; > + int r, i = -1; > unsigned long hret; > struct hv_perf_caps caps; > > @@ -1727,8 +1749,18 @@ static int hv_24x7_init(void) > > /* init cpuhotplug */ > r = hv_24x7_cpu_hotplug_init(); > - if (r) > + if (r) { > pr_err("hv_24x7: CPU hotplug init failed\n"); > + } else { > + /* > + * Cpu hotplug init is successful, add the > + * cpumask file as part of pmu attr group and > + * assign it to very first NULL location. > + */ > + while (attr_groups[++i]) > + /* nothing */; > + attr_groups[i] = &cpumask_attr_group; > + } > > r = perf_pmu_register(&h_24x7_pmu, h_24x7_pmu.name, -1); > if (r) > -- > 2.18.2 >
On 6/24/20 4:26 PM, Gautham R Shenoy wrote: > Hi Kajol, > > On Wed, Jun 24, 2020 at 03:47:54PM +0530, Kajol Jain wrote: >> Patch here adds a cpumask attr to hv_24x7 pmu along with ABI documentation. >> >> command:# cat /sys/devices/hv_24x7/cpumask >> 0 > Since this sysfs interface is read-only, and the user cannot change > the CPU which will be making the HCALLs to obtain the 24x7 counts, > does the user even need to know if currently CPU X is the one which is > going to make HCALLs to retrive the 24x7 counts ? Does it help in any > kind of trouble-shooting ? Primary use to expose the cpumask is for the perf tool. Which has the capability to parse the driver sysfs folder and understand the cpumask file. Having cpumask file will reduce the number of perf commandline parameters (will avoid "-C" option in the perf tool command line). I can also notify the user which is the current cpu used to retrieve the counter data. > It would have made sense if the interface was read-write, since a user > can set this to a CPU which is not running user applications. This > would help in minimising jitter on those active CPUs running the user > applications. With cpumask backed by hotplug notifiers, enabling user write access to it will complicate the code with more additional check. CPU will come to play only if the user request for counter data. If not, then there will be no HCALLs made using the CPU. Maddy > > >> Signed-off-by: Kajol Jain <kjain@linux.ibm.com> >> --- >> .../sysfs-bus-event_source-devices-hv_24x7 | 7 ++++ >> arch/powerpc/perf/hv-24x7.c | 36 +++++++++++++++++-- >> 2 files changed, 41 insertions(+), 2 deletions(-) >> >> diff --git a/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 b/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 >> index e8698afcd952..f9dd3755b049 100644 >> --- a/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 >> +++ b/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 >> @@ -43,6 +43,13 @@ Description: read only >> This sysfs interface exposes the number of cores per chip >> present in the system. >> >> +What: /sys/devices/hv_24x7/cpumask >> +Date: June 2020 >> +Contact: Linux on PowerPC Developer List <linuxppc-dev@lists.ozlabs.org> >> +Description: read only >> + This sysfs file exposes the cpumask which is designated to make >> + HCALLs to retrive hv-24x7 pmu event counter data. >> + >> What: /sys/bus/event_source/devices/hv_24x7/event_descs/<event-name> >> Date: February 2014 >> Contact: Linux on PowerPC Developer List <linuxppc-dev@lists.ozlabs.org> >> diff --git a/arch/powerpc/perf/hv-24x7.c b/arch/powerpc/perf/hv-24x7.c >> index ce4739e2b407..3c699612d29f 100644 >> --- a/arch/powerpc/perf/hv-24x7.c >> +++ b/arch/powerpc/perf/hv-24x7.c >> @@ -448,6 +448,12 @@ static ssize_t device_show_string(struct device *dev, >> return sprintf(buf, "%s\n", (char *)d->var); >> } >> >> +static ssize_t cpumask_get_attr(struct device *dev, >> + struct device_attribute *attr, char *buf) >> +{ >> + return cpumap_print_to_pagebuf(true, buf, &hv_24x7_cpumask); >> +} >> + >> static ssize_t sockets_show(struct device *dev, >> struct device_attribute *attr, char *buf) >> { >> @@ -1116,6 +1122,17 @@ static DEVICE_ATTR_RO(sockets); >> static DEVICE_ATTR_RO(chipspersocket); >> static DEVICE_ATTR_RO(coresperchip); >> >> +static DEVICE_ATTR(cpumask, S_IRUGO, cpumask_get_attr, NULL); >> + >> +static struct attribute *cpumask_attrs[] = { >> + &dev_attr_cpumask.attr, >> + NULL, >> +}; >> + >> +static struct attribute_group cpumask_attr_group = { >> + .attrs = cpumask_attrs, >> +}; >> + >> static struct bin_attribute *if_bin_attrs[] = { >> &bin_attr_catalog, >> NULL, >> @@ -1143,6 +1160,11 @@ static const struct attribute_group *attr_groups[] = { >> &event_desc_group, >> &event_long_desc_group, >> &if_group, >> + /* >> + * This NULL is a placeholder for the cpumask attr which will update >> + * onlyif cpuhotplug registration is successful >> + */ >> + NULL, >> NULL, >> }; >> >> @@ -1683,7 +1705,7 @@ static int hv_24x7_cpu_hotplug_init(void) >> >> static int hv_24x7_init(void) >> { >> - int r; >> + int r, i = -1; >> unsigned long hret; >> struct hv_perf_caps caps; >> >> @@ -1727,8 +1749,18 @@ static int hv_24x7_init(void) >> >> /* init cpuhotplug */ >> r = hv_24x7_cpu_hotplug_init(); >> - if (r) >> + if (r) { >> pr_err("hv_24x7: CPU hotplug init failed\n"); >> + } else { >> + /* >> + * Cpu hotplug init is successful, add the >> + * cpumask file as part of pmu attr group and >> + * assign it to very first NULL location. >> + */ >> + while (attr_groups[++i]) >> + /* nothing */; >> + attr_groups[i] = &cpumask_attr_group; >> + } >> >> r = perf_pmu_register(&h_24x7_pmu, h_24x7_pmu.name, -1); >> if (r) >> -- >> 2.18.2 >>
On Wed, Jun 24, 2020 at 05:58:31PM +0530, Madhavan Srinivasan wrote: > > > On 6/24/20 4:26 PM, Gautham R Shenoy wrote: > >Hi Kajol, > > > >On Wed, Jun 24, 2020 at 03:47:54PM +0530, Kajol Jain wrote: > >>Patch here adds a cpumask attr to hv_24x7 pmu along with ABI documentation. > >> > >>command:# cat /sys/devices/hv_24x7/cpumask > >>0 > >Since this sysfs interface is read-only, and the user cannot change > >the CPU which will be making the HCALLs to obtain the 24x7 counts, > >does the user even need to know if currently CPU X is the one which is > >going to make HCALLs to retrive the 24x7 counts ? Does it help in any > >kind of trouble-shooting ? > Primary use to expose the cpumask is for the perf tool. > Which has the capability to parse the driver sysfs folder > and understand the cpumask file. Having cpumask > file will reduce the number of perf commandline > parameters (will avoid "-C" option in the perf tool > command line). I can also notify the user which is > the current cpu used to retrieve the counter data. Fair enough. Can we include this in the patch description ? > > >It would have made sense if the interface was read-write, since a user > >can set this to a CPU which is not running user applications. This > >would help in minimising jitter on those active CPUs running the user > >applications. > > With cpumask backed by hotplug > notifiers, enabling user write access to it will > complicate the code with more additional check. > CPU will come to play only if the user request for > counter data. If not, then there will be no HCALLs made > using the CPU. Well, I was wondering if you could make the interface writable because I couldn't think of the use of a read-only interface. With the perf-use case you have provided, I guess it makes sense. I am ok with it being a read-only interface. > > Maddy -- Thanks and Regards gautham.
On 6/26/20 1:15 PM, Gautham R Shenoy wrote: > On Wed, Jun 24, 2020 at 05:58:31PM +0530, Madhavan Srinivasan wrote: >> >> >> On 6/24/20 4:26 PM, Gautham R Shenoy wrote: >>> Hi Kajol, >>> >>> On Wed, Jun 24, 2020 at 03:47:54PM +0530, Kajol Jain wrote: >>>> Patch here adds a cpumask attr to hv_24x7 pmu along with ABI documentation. >>>> >>>> command:# cat /sys/devices/hv_24x7/cpumask >>>> 0 >>> Since this sysfs interface is read-only, and the user cannot change >>> the CPU which will be making the HCALLs to obtain the 24x7 counts, >>> does the user even need to know if currently CPU X is the one which is >>> going to make HCALLs to retrive the 24x7 counts ? Does it help in any >>> kind of trouble-shooting ? >> Primary use to expose the cpumask is for the perf tool. >> Which has the capability to parse the driver sysfs folder >> and understand the cpumask file. Having cpumask >> file will reduce the number of perf commandline >> parameters (will avoid "-C" option in the perf tool >> command line). I can also notify the user which is >> the current cpu used to retrieve the counter data. > > Fair enough. Can we include this in the patch description ? Sure will update in next version of patchset. Thanks, Kajol Jain > >> >>> It would have made sense if the interface was read-write, since a user >>> can set this to a CPU which is not running user applications. This >>> would help in minimising jitter on those active CPUs running the user >>> applications. >> >> With cpumask backed by hotplug >> notifiers, enabling user write access to it will >> complicate the code with more additional check. >> CPU will come to play only if the user request for >> counter data. If not, then there will be no HCALLs made >> using the CPU. > > Well, I was wondering if you could make the interface writable because > I couldn't think of the use of a read-only interface. With the > perf-use case you have provided, I guess it makes sense. I am ok with > it being a read-only interface. > >> >> Maddy > > -- > Thanks and Regards > gautham. >
diff --git a/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 b/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 index e8698afcd952..f9dd3755b049 100644 --- a/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 +++ b/Documentation/ABI/testing/sysfs-bus-event_source-devices-hv_24x7 @@ -43,6 +43,13 @@ Description: read only This sysfs interface exposes the number of cores per chip present in the system. +What: /sys/devices/hv_24x7/cpumask +Date: June 2020 +Contact: Linux on PowerPC Developer List <linuxppc-dev@lists.ozlabs.org> +Description: read only + This sysfs file exposes the cpumask which is designated to make + HCALLs to retrive hv-24x7 pmu event counter data. + What: /sys/bus/event_source/devices/hv_24x7/event_descs/<event-name> Date: February 2014 Contact: Linux on PowerPC Developer List <linuxppc-dev@lists.ozlabs.org> diff --git a/arch/powerpc/perf/hv-24x7.c b/arch/powerpc/perf/hv-24x7.c index ce4739e2b407..3c699612d29f 100644 --- a/arch/powerpc/perf/hv-24x7.c +++ b/arch/powerpc/perf/hv-24x7.c @@ -448,6 +448,12 @@ static ssize_t device_show_string(struct device *dev, return sprintf(buf, "%s\n", (char *)d->var); } +static ssize_t cpumask_get_attr(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return cpumap_print_to_pagebuf(true, buf, &hv_24x7_cpumask); +} + static ssize_t sockets_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -1116,6 +1122,17 @@ static DEVICE_ATTR_RO(sockets); static DEVICE_ATTR_RO(chipspersocket); static DEVICE_ATTR_RO(coresperchip); +static DEVICE_ATTR(cpumask, S_IRUGO, cpumask_get_attr, NULL); + +static struct attribute *cpumask_attrs[] = { + &dev_attr_cpumask.attr, + NULL, +}; + +static struct attribute_group cpumask_attr_group = { + .attrs = cpumask_attrs, +}; + static struct bin_attribute *if_bin_attrs[] = { &bin_attr_catalog, NULL, @@ -1143,6 +1160,11 @@ static const struct attribute_group *attr_groups[] = { &event_desc_group, &event_long_desc_group, &if_group, + /* + * This NULL is a placeholder for the cpumask attr which will update + * onlyif cpuhotplug registration is successful + */ + NULL, NULL, }; @@ -1683,7 +1705,7 @@ static int hv_24x7_cpu_hotplug_init(void) static int hv_24x7_init(void) { - int r; + int r, i = -1; unsigned long hret; struct hv_perf_caps caps; @@ -1727,8 +1749,18 @@ static int hv_24x7_init(void) /* init cpuhotplug */ r = hv_24x7_cpu_hotplug_init(); - if (r) + if (r) { pr_err("hv_24x7: CPU hotplug init failed\n"); + } else { + /* + * Cpu hotplug init is successful, add the + * cpumask file as part of pmu attr group and + * assign it to very first NULL location. + */ + while (attr_groups[++i]) + /* nothing */; + attr_groups[i] = &cpumask_attr_group; + } r = perf_pmu_register(&h_24x7_pmu, h_24x7_pmu.name, -1); if (r)
Patch here adds a cpumask attr to hv_24x7 pmu along with ABI documentation. command:# cat /sys/devices/hv_24x7/cpumask 0 Signed-off-by: Kajol Jain <kjain@linux.ibm.com> --- .../sysfs-bus-event_source-devices-hv_24x7 | 7 ++++ arch/powerpc/perf/hv-24x7.c | 36 +++++++++++++++++-- 2 files changed, 41 insertions(+), 2 deletions(-)