Skip to contentSkip to conversation

udev races with 'arch_register_cpu' to write 1 to /sys/devices/system/cpu/cpu1/online (which is not yet created).

RFC Superseded Konrad Rzeszutek Wilk 3 participants 14 messages Apr 30, 2012 – May 14, 2013 1 message not in the archive Driver core

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw
Quoted 1 line · Konrad Rzeszutek Wilk

Hey Greg,

Grr.. Incorrectly added LKML to it. Sorry about that.

Quoted 24 lines

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

maintainer
#2 lore raw
In reply to a message not in the archive
Quoted 19 lines · Konrad Rzeszutek Wilk

Hey Greg,

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Quoted 5 lines

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

Yes.

maintainer
#3 lore raw
Quoted 20 lines · Greg KH
Quoted 19 lines · Konrad Rzeszutek Wilk

Hey Greg,

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Quoted 7 lines
Quoted 5 lines

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

Yes.

Oh, I imagine you want to know _how_ to do it too, right? (sorry, I couldn't resist...)

Make this a default attribute of the cpu device, and then it will be created by the driver core before the uevent is sent to userspace. That's what you are supposed to do in the first place, adding files "by hand" is wrong, for this very reason.

thanks,

greg k-h

#4 lore raw
Quoted 24 lines · Greg KH
Quoted 23 lines · Greg KH
Quoted 22 lines · Konrad Rzeszutek Wilk

Hey Greg,

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

Yes.

Oh, I imagine you want to know _how_ to do it too, right? (sorry, I couldn't resist...)

Heh.

Quoted 5 lines

Make this a default attribute of the cpu device, and then it will be created by the driver core before the uevent is sent to userspace. That's what you are supposed to do in the first place, adding files "by hand" is wrong, for this very reason.

OK, will prep up a patch shortly.

#5 lore raw

Re: udev races with 'arch_register_cpu' to write 1 to /sys/devices/system/cpu/cpu1/online (which is not yet created)

Quoted 33 lines · Greg KH
Quoted 24 lines · Greg KH
Quoted 23 lines · Konrad Rzeszutek Wilk
Quoted 22 lines

Hey Greg,

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

Yes.

Oh, I imagine you want to know _how_ to do it too, right? (sorry, I couldn't resist...)

Quoted 8 lines

Heh.

Quoted 5 lines

Make this a default attribute of the cpu device, and then it will be created by the driver core before the uevent is sent to userspace. That's what you are supposed to do in the first place, adding files "by hand" is wrong, for this very reason.

OK, will prep up a patch shortly.

Hello Konrad,

Is there a posted/accepted patch or idea was dropped?

#6 lore raw

Re: udev races with 'arch_register_cpu' to write 1 to /sys/devices/system/cpu/cpu1/online (which is not yet created)

Quoted 40 lines · Igor Mammedov
Quoted 40 lines · Greg KH
Quoted 24 lines · Greg KH
Quoted 23 lines · Konrad Rzeszutek Wilk
Quoted 22 lines

Hey Greg,

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

Yes.

Oh, I imagine you want to know _how_ to do it too, right? (sorry, I couldn't resist...)

Heh.

Quoted 5 lines

Make this a default attribute of the cpu device, and then it will be created by the driver core before the uevent is sent to userspace. That's what you are supposed to do in the first place, adding files "by hand" is wrong, for this very reason.

Quoted 5 lines
Quoted 1 line

OK, will prep up a patch shortly.

Hello Konrad,

Is there a posted/accepted patch or idea was dropped?

Hey Igor,

CC-ing here Chuck here who is looking at that. My recollection (and Chuck please correct me if I am wrong) is that the idea Greg outlined won't work very well. The reason is that making 'online' an attribute of 'struct dev' means that:
1) A lot of SysFS attributes that have no notion of online/offline (say ISA bus) will now have.
2) The default action item (so function) to do something based on writting/reading from 'online' will have to be overridden by the driver using it. Which means another race - we can create an SysFS attribute but the default points to something that does nothing.

Chuck, does that sound right?

Quoted 2 lines
maintainer
#7 lore raw

Re: udev races with 'arch_register_cpu' to write 1 to /sys/devices/system/cpu/cpu1/online (which is not yet created)

Quoted 47 lines · Konrad Rzeszutek Wilk
Quoted 40 lines · Igor Mammedov
Quoted 40 lines · Greg KH
Quoted 24 lines · Greg KH
Quoted 23 lines · Konrad Rzeszutek Wilk
Quoted 22 lines

Hey Greg,

Hoping you can help with some guidance on how to fix this.

The issue is with CPU hotplug is that when a CPU goes up it calls 'arch_register_cpu' which eventually calls register_cpu. That function does these two things:

251 error = device_register(&cpu->dev);
252 if (!error && cpu->hotpluggable)
253 register_cpu_control(cpu);

and the device_register creates a nice little SysFS directory:

/sys/devices/system/cpu/cpu2/ which at line 251 has the 'add' attribute
but no 'online' attribute. udev then tries to echo 1 to the 'online'
and it we get:
udevd-work[2421]: error opening ATTR{/sys/devices/system/cpu/cpu2/online} for writing: No such file or directory

Line 253 creates said 'online' and at that time udev [or the system admin] can write 1 to 'online' and the CPU goes up.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

So .. any thoughts? Is there some way to inhibit from uevent being sent until line 253 has run?

Yes.

Oh, I imagine you want to know _how_ to do it too, right? (sorry, I couldn't resist...)

Heh.

Quoted 5 lines

Make this a default attribute of the cpu device, and then it will be created by the driver core before the uevent is sent to userspace. That's what you are supposed to do in the first place, adding files "by hand" is wrong, for this very reason.

Quoted 5 lines
Quoted 1 line

OK, will prep up a patch shortly.

Hello Konrad,

Is there a posted/accepted patch or idea was dropped?

Quoted 7 lines

Hey Igor,

CC-ing here Chuck here who is looking at that. My recollection (and Chuck please correct me if I am wrong) is that the idea Greg outlined won't work very well. The reason is that making 'online' an attribute of 'struct dev' means that:
1) A lot of SysFS attributes that have no notion of online/offline (say ISA bus) will now have.

Then only create that attribute for busses that ask for it to be enabled.

Quoted 4 lines

2) The default action item (so function) to do something based on writting/reading from 'online' will have to be overridden by the driver using it. Which means another race - we can create an SysFS attribute but the default points to something that does nothing.

I don't understand this at all, how will there be a race exactly?

greg k-h

#8 lore raw

cover cpu: fix leak and udev race in register_cpu()

In reply to Konrad Rzeszutek Wilk

Here is a crude attempt fix race the way suggested by Greg, probably done wrong but hopefully in the right direction.

1. move "crash_notes" to static attributes to guarantee that it's destroyed with CPU on unregister.

2. fixes race between hotplugged CPU and onlining it via udev, described here
https://lkml.org/lkml/2012/4/30/193

Igor Mammedov (2):
cpu: fix crash_notes leak
cpu: make sure that cpu/online file created before KOBJ_ADD is
emitted

 drivers/base/cpu.c |   55 +++++++++++++++++++++++++++++++++++----------------
 1 files changed, 38 insertions(+), 17 deletions(-)
author Tested-by Reviewed-by on the v1 series
#9 lore raw

Re: [RFC 0/2] cpu: fix leak and udev race in register_cpu()

Quoted 2 lines · Igor Mammedov

Here is a crude attempt fix race the way suggested by Greg, probably done wrong but hopefully in the right direction.

Weird, I thought I had tried that at first but got tons of kobject warnings and such. But I think I tried to add it to kset instead of the one you did.

It fixes it for me so

and also (thought the git commit descriptions need a bit of work, but that is expected as an RFC patch):

Quoted 15 lines

1. move "crash_notes" to static attributes to guarantee that it's destroyed with CPU on unregister.

2. fixes race between hotplugged CPU and onlining it via udev, described here
https://lkml.org/lkml/2012/4/30/193

Igor Mammedov (2):
cpu: fix crash_notes leak
cpu: make sure that cpu/online file created before KOBJ_ADD is
emitted

 drivers/base/cpu.c |   55 +++++++++++++++++++++++++++++++++++----------------
 1 files changed, 38 insertions(+), 17 deletions(-)
maintainer
#10 lore raw

Re: [RFC 0/2] cpu: fix leak and udev race in register_cpu()

In reply to Igor Mammedov
Quoted 8 lines · Igor Mammedov

Here is a crude attempt fix race the way suggested by Greg, probably done wrong but hopefully in the right direction.

1. move "crash_notes" to static attributes to guarantee that it's destroyed with CPU on unregister.

2. fixes race between hotplugged CPU and onlining it via udev, described here
https://lkml.org/lkml/2012/4/30/193

Looks right to me, nice job.

greg k-h

#11 lore raw

PATCH 1/2 cpu: fix "crash_notes" leak in register_cpu()

In reply to Konrad Rzeszutek Wilk

"crash_notes" is dinamically created with device_create_file() but isn't anywhere deleted.
Define "crash_notes" statically via attribute groups so that device_register would create it automatically and file would be destroyed when CPU is destroyed.

---
 drivers/base/cpu.c |   21 +++++++++++++++++----
 1 files changed, 17 insertions(+), 4 deletions(-)
drivers/base/cpu.c +17 -4
index fb10728..02e4d73 100644
@@ -132,8 +132,24 @@ static ssize_t show_crash_notes(struct device *dev, struct device_attribute *att
 	return rc;
 }
 static DEVICE_ATTR(crash_notes, 0400, show_crash_notes, NULL);
+
+static struct attribute *crash_note_cpu_attrs[] = {
+	&dev_attr_crash_notes.attr,
+	NULL
+};
+
+static struct attribute_group crash_note_cpu_attr_group = {
+	.attrs = crash_note_cpu_attrs,
+};
 #endif
 
+static const struct attribute_group *common_cpu_attr_groups[] = {
+#ifdef CONFIG_KEXEC
+	&crash_note_cpu_attr_group,
+#endif
+	NULL
+};
+
 /*
  * Print cpu online, possible, present, and system maps
  */
@@ -248,6 +264,7 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 #ifdef CONFIG_ARCH_HAS_CPU_AUTOPROBE
 	cpu->dev.bus->uevent = arch_cpu_uevent;
 #endif
+	cpu->dev.groups = common_cpu_attr_groups;
 	error = device_register(&cpu->dev);
 	if (!error && cpu->hotpluggable)
 		register_cpu_control(cpu);
@@ -256,10 +273,6 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 	if (!error)
 		register_cpu_under_node(num, cpu_to_node(num));
 
-#ifdef CONFIG_KEXEC
-	if (!error)
-		error = device_create_file(&cpu->dev, &dev_attr_crash_notes);
-#endif
 	return error;
 }
 
Signature · 1 line
1.7.1
#12 lore raw

Re: [PATCH 1/2] cpu: fix "crash_notes" leak in register_cpu()

Quoted 1 line · Igor Mammedov

"crash_notes" is dinamically created with device_create_file() but

                   ^^^^^^^^^^^-dynamically
Quoted 4 lines

isn't anywhere deleted.
Define "crash_notes" statically via attribute groups so that device_register would create it automatically and file would be destroyed when CPU is destroyed.

I had to modify it a bit to work with v3.10-rc1 (It is missing the dev_attr_crash_notes_size), but otherwise it worked nicely.

Quoted 57 lines
---
 drivers/base/cpu.c |   21 +++++++++++++++++----
 1 files changed, 17 insertions(+), 4 deletions(-)
drivers/base/cpu.c +17 -4
index fb10728..02e4d73 100644
@@ -132,8 +132,24 @@ static ssize_t show_crash_notes(struct device *dev, struct device_attribute *att
 	return rc;
 }
 static DEVICE_ATTR(crash_notes, 0400, show_crash_notes, NULL);
+
+static struct attribute *crash_note_cpu_attrs[] = {
+	&dev_attr_crash_notes.attr,
+	NULL
+};
+
+static struct attribute_group crash_note_cpu_attr_group = {
+	.attrs = crash_note_cpu_attrs,
+};
 #endif
 
+static const struct attribute_group *common_cpu_attr_groups[] = {
+#ifdef CONFIG_KEXEC
+	&crash_note_cpu_attr_group,
+#endif
+	NULL
+};
+
 /*
  * Print cpu online, possible, present, and system maps
  */
@@ -248,6 +264,7 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 #ifdef CONFIG_ARCH_HAS_CPU_AUTOPROBE
 	cpu->dev.bus->uevent = arch_cpu_uevent;
 #endif
+	cpu->dev.groups = common_cpu_attr_groups;
 	error = device_register(&cpu->dev);
 	if (!error && cpu->hotpluggable)
 		register_cpu_control(cpu);
@@ -256,10 +273,6 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 	if (!error)
 		register_cpu_under_node(num, cpu_to_node(num));
 
-#ifdef CONFIG_KEXEC
-	if (!error)
-		error = device_create_file(&cpu->dev, &dev_attr_crash_notes);
-#endif
 	return error;
 }
 
Signature · 1 line
1.7.1
#13 lore raw

PATCH 2/2 cpu: make sure that cpu/online file created before KOBJ_ADD is emitted

In reply to Konrad Rzeszutek Wilk

Fixes race between udev and hotplugged CPU registration, described at
https://lkml.org/lkml/2012/4/30/198
by defining "online" attribute statically so that device_add() would create it before notifying udev about new CPU.

---
 drivers/base/cpu.c |   34 +++++++++++++++++++++-------------
 1 files changed, 21 insertions(+), 13 deletions(-)
drivers/base/cpu.c +21 -13
index 02e4d73..6578030 100644
@@ -64,18 +64,21 @@ static ssize_t __ref store_online(struct device *dev,
 }
 static DEVICE_ATTR(online, 0644, show_online, store_online);
 
-static void __cpuinit register_cpu_control(struct cpu *cpu)
-{
-	device_create_file(&cpu->dev, &dev_attr_online);
-}
+static struct attribute *hotplug_cpu_attrs[] = {
+	&dev_attr_online.attr,
+	NULL
+};
+
+static struct attribute_group hotplug_cpu_attr_group = {
+	.attrs = hotplug_cpu_attrs,
+};
+
 void unregister_cpu(struct cpu *cpu)
 {
 	int logical_cpu = cpu->dev.id;
 
 	unregister_cpu_under_node(logical_cpu, cpu_to_node(logical_cpu));
 
-	device_remove_file(&cpu->dev, &dev_attr_online);
-
 	device_unregister(&cpu->dev);
 	per_cpu(cpu_sys_devices, logical_cpu) = NULL;
 	return;
@@ -101,11 +104,6 @@ static ssize_t cpu_release_store(struct device *dev,
 static DEVICE_ATTR(probe, S_IWUSR, NULL, cpu_probe_store);
 static DEVICE_ATTR(release, S_IWUSR, NULL, cpu_release_store);
 #endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
-
-#else /* ... !CONFIG_HOTPLUG_CPU */
-static inline void register_cpu_control(struct cpu *cpu)
-{
-}
 #endif /* CONFIG_HOTPLUG_CPU */
 
 #ifdef CONFIG_KEXEC
@@ -150,6 +148,16 @@ static const struct attribute_group *common_cpu_attr_groups[] = {
 	NULL
 };
 
+static const struct attribute_group *hotplugable_cpu_attr_groups[] = {
+#ifdef CONFIG_KEXEC
+	&crash_note_cpu_attr_group,
+#endif
+#ifdef CONFIG_HOTPLUG_CPU
+	&hotplug_cpu_attr_group,
+#endif
+	NULL
+};
+
 /*
  * Print cpu online, possible, present, and system maps
  */
@@ -265,9 +273,9 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 	cpu->dev.bus->uevent = arch_cpu_uevent;
 #endif
 	cpu->dev.groups = common_cpu_attr_groups;
+	if (cpu->hotpluggable)
+		cpu->dev.groups = hotplugable_cpu_attr_groups;
 	error = device_register(&cpu->dev);
-	if (!error && cpu->hotpluggable)
-		register_cpu_control(cpu);
 	if (!error)
 		per_cpu(cpu_sys_devices, num) = &cpu->dev;
 	if (!error)
Signature · 1 line
1.7.1
#14 lore raw

Re: [PATCH 2/2] cpu: make sure that cpu/online file created before KOBJ_ADD is emitted

Quoted 2 lines · Igor Mammedov

Fixes race between udev and hotplugged CPU registration, described at
https://lkml.org/lkml/2012/4/30/198

It would be better if you copy-n-pasted the contents of that URL in here.

Quoted 85 lines

by defining "online" attribute statically so that device_add() would create it before notifying udev about new CPU.

---
 drivers/base/cpu.c |   34 +++++++++++++++++++++-------------
 1 files changed, 21 insertions(+), 13 deletions(-)
drivers/base/cpu.c +21 -13
index 02e4d73..6578030 100644
@@ -64,18 +64,21 @@ static ssize_t __ref store_online(struct device *dev,
 }
 static DEVICE_ATTR(online, 0644, show_online, store_online);
 
-static void __cpuinit register_cpu_control(struct cpu *cpu)
-{
-	device_create_file(&cpu->dev, &dev_attr_online);
-}
+static struct attribute *hotplug_cpu_attrs[] = {
+	&dev_attr_online.attr,
+	NULL
+};
+
+static struct attribute_group hotplug_cpu_attr_group = {
+	.attrs = hotplug_cpu_attrs,
+};
+
 void unregister_cpu(struct cpu *cpu)
 {
 	int logical_cpu = cpu->dev.id;
 
 	unregister_cpu_under_node(logical_cpu, cpu_to_node(logical_cpu));
 
-	device_remove_file(&cpu->dev, &dev_attr_online);
-
 	device_unregister(&cpu->dev);
 	per_cpu(cpu_sys_devices, logical_cpu) = NULL;
 	return;
@@ -101,11 +104,6 @@ static ssize_t cpu_release_store(struct device *dev,
 static DEVICE_ATTR(probe, S_IWUSR, NULL, cpu_probe_store);
 static DEVICE_ATTR(release, S_IWUSR, NULL, cpu_release_store);
 #endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
-
-#else /* ... !CONFIG_HOTPLUG_CPU */
-static inline void register_cpu_control(struct cpu *cpu)
-{
-}
 #endif /* CONFIG_HOTPLUG_CPU */
 
 #ifdef CONFIG_KEXEC
@@ -150,6 +148,16 @@ static const struct attribute_group *common_cpu_attr_groups[] = {
 	NULL
 };
 
+static const struct attribute_group *hotplugable_cpu_attr_groups[] = {
+#ifdef CONFIG_KEXEC
+	&crash_note_cpu_attr_group,
+#endif
+#ifdef CONFIG_HOTPLUG_CPU
+	&hotplug_cpu_attr_group,
+#endif
+	NULL
+};
+
 /*
  * Print cpu online, possible, present, and system maps
  */
@@ -265,9 +273,9 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 	cpu->dev.bus->uevent = arch_cpu_uevent;
 #endif
 	cpu->dev.groups = common_cpu_attr_groups;
+	if (cpu->hotpluggable)
+		cpu->dev.groups = hotplugable_cpu_attr_groups;
 	error = device_register(&cpu->dev);
-	if (!error && cpu->hotpluggable)
-		register_cpu_control(cpu);
 	if (!error)
 		per_cpu(cpu_sys_devices, num) = &cpu->dev;
 	if (!error)
Signature · 1 line
1.7.1