linux/Documentation/cpu-hotplug.txt
Linus Torvalds f991fae5c6 Power management and ACPI updates for 3.11-rc1
- Hotplug changes allowing device hot-removal operations to fail
   gracefully (instead of crashing the kernel) if they cannot be
   carried out completely.  From Rafael J Wysocki and Toshi Kani.
 
 - Freezer update from Colin Cross and Mandeep Singh Baines targeted
   at making the freezing of tasks a bit less heavy weight operation.
 
 - cpufreq resume fix from Srivatsa S Bhat for a regression introduced
   during the 3.10 cycle causing some cpufreq sysfs attributes to
   return wrong values to user space after resume.
 
 - New freqdomain_cpus sysfs attribute for the acpi-cpufreq driver to
   provide information previously available via related_cpus from
   Lan Tianyu.
 
 - cpufreq fixes and cleanups from Viresh Kumar, Jacob Shin,
   Heiko Stübner, Xiaoguang Chen, Ezequiel Garcia, Arnd Bergmann, and
   Tang Yuantian.
 
 - Fix for an ACPICA regression causing suspend/resume issues to
   appear on some systems introduced during the 3.4 development cycle
   from Lv Zheng.
 
 - ACPICA fixes and cleanups from Bob Moore, Tomasz Nowicki, Lv Zheng,
   Chao Guan, and Zhang Rui.
 
 - New cupidle driver for Xilinx Zynq processors from Michal Simek.
 
 - cpuidle fixes and cleanups from Daniel Lezcano.
 
 - Changes to make suspend/resume work correctly in Xen guests from
   Konrad Rzeszutek Wilk.
 
 - ACPI device power management fixes and cleanups from Fengguang Wu
   and Rafael J Wysocki.
 
 - ACPI documentation updates from Lv Zheng, Aaron Lu and Hanjun Guo.
 
 - Fix for the IA-64 issue that was the reason for reverting commit
   9f29ab1 and updates of the ACPI scan code from Rafael J Wysocki.
 
 - Mechanism for adding CMOS RTC address space handlers from Lan Tianyu
   (to allow some EC-related breakage to be fixed on some systems).
 
 - Spec-compliant implementation of acpi_os_get_timer() from
   Mika Westerberg.
 
 - Modification of do_acpi_find_child() to execute _STA in order to
   to avoid situations in which a pointer to a disabled device object
   is returned instead of an enabled one with the same _ADR value.
   From Jeff Wu.
 
 - Intel BayTrail PCH (Platform Controller Hub) support for the ACPI
   Intel Low-Power Subsystems (LPSS) driver and modificaions of that
   driver to work around a couple of known BIOS issues from
   Mika Westerberg and Heikki Krogerus.
 
 - EC driver fix from Vasiliy Kulikov to make it use get_user() and
   put_user() instead of dereferencing user space pointers blindly.
 
 - Assorted ACPI code cleanups from Bjorn Helgaas, Nicholas Mazzuca and
   Toshi Kani.
 
 - Modification of the "runtime idle" helper routine to take the return
   values of the callbacks executed by it into account and to call
   rpm_suspend() if they return 0, which allows some code bloat
   reduction to be done, from Rafael J Wysocki and Alan Stern.
 
 - New trace points for PM QoS from Sahara <keun-o.park@windriver.com>.
 
 - PM QoS documentation update from Lan Tianyu.
 
 - Assorted core PM code cleanups and changes from Bernie Thompson,
   Bjorn Helgaas, Julius Werner, and Shuah Khan.
 
 - New devfreq driver for the Exynos5-bus device from Abhilash Kesavan.
 
 - Minor devfreq cleanups, fixes and MAINTAINERS update from
   MyungJoo Ham, Abhilash Kesavan, Paul Bolle, Rajagopal Venkat, and
   Wei Yongjun.
 
 - OMAP Adaptive Voltage Scaling (AVS) SmartReflex voltage control
   driver updates from Andrii Tseglytskyi and Nishanth Menon.
 
 /
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.19 (GNU/Linux)
 
 iQIcBAABAgAGBQJR0ZNOAAoJEKhOf7ml8uNsDLYP/0EU4rmvw0TWTITfp6RS1KDE
 9GwBn96ZR4Q5bJd9gBCTPSqhHOYMqxWEUp99sn/M2wehG1pk/jw5LO56+2IhM3UZ
 g1HDcJ7te2nVT/iXsKiAGTVhU9Rk0aYwoVSknwk27qpIBGxW9w/s5tLX8pY3Q3Zq
 wL/7aTPjyL+PFFFEaxgH7qLqsl3DhbtYW5AriUBTkXout/tJ4eO1b7MNBncLDh8X
 VQ/0DNCKE95VEJfkO4rk9RKUyVp9GDn0i+HXCD/FS4IA5oYzePdVdNDmXf7g+swe
 CGlTZq8pB+oBpDiHl4lxzbNrKQjRNbGnDUkoRcWqn0nAw56xK+vmYnWJhW99gQ/I
 fKnvxeLca5po1aiqmC4VSJxZIatFZqLrZAI4dzoCLWY+bGeTnCKmj0/F8ytFnZA2
 8IuLLs7/dFOaHXV/pKmpg6FAlFa9CPxoqRFoyqb4M0GjEarADyalXUWsPtG+6xCp
 R/p0CISpwk+guKZR/qPhL7M654S7SHrPwd2DPF0KgGsvk+G2GhoB8EzvD8BVp98Z
 9siCGCdgKQfJQVI6R0k9aFmn/4gRQIAgyPhkhv9tqULUUkiaXki+/t8kPfnb8O/d
 zep+CA57E2G8MYLkDJfpFeKS7GpPD6TIdgFdGmOUC0Y6sl9iTdiw4yTx8O2JM37z
 rHBZfYGkJBrbGRu+Q1gs
 =VBBq
 -----END PGP SIGNATURE-----

Merge tag 'pm+acpi-3.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull power management and ACPI updates from Rafael Wysocki:
 "This time the total number of ACPI commits is slightly greater than
  the number of cpufreq commits, but Viresh Kumar (who works on cpufreq)
  remains the most active patch submitter.

  To me, the most significant change is the addition of offline/online
  device operations to the driver core (with the Greg's blessing) and
  the related modifications of the ACPI core hotplug code.  Next are the
  freezer updates from Colin Cross that should make the freezing of
  tasks a bit less heavy weight.

  We also have a couple of regression fixes, a number of fixes for
  issues that have not been identified as regressions, two new drivers
  and a bunch of cleanups all over.

  Highlights:

   - Hotplug changes to support graceful hot-removal failures.

     It sometimes is necessary to fail device hot-removal operations
     gracefully if they cannot be carried out completely.  For example,
     if memory from a memory module being hot-removed has been allocated
     for the kernel's own use and cannot be moved elsewhere, it's
     desirable to fail the hot-removal operation in a graceful way
     rather than to crash the kernel, but currenty a success or a kernel
     crash are the only possible outcomes of an attempted memory
     hot-removal.  Needless to say, that is not a very attractive
     alternative and it had to be addressed.

     However, in order to make it work for memory, I first had to make
     it work for CPUs and for this purpose I needed to modify the ACPI
     processor driver.  It's been split into two parts, a resident one
     handling the low-level initialization/cleanup and a modular one
     playing the actual driver's role (but it binds to the CPU system
     device objects rather than to the ACPI device objects representing
     processors).  That's been sort of like a live brain surgery on a
     patient who's riding a bike.

     So this is a little scary, but since we found and fixed a couple of
     regressions it caused to happen during the early linux-next testing
     (a month ago), nobody has complained.

     As a bonus we remove some duplicated ACPI hotplug code, because the
     ACPI-based CPU hotplug is now going to use the common ACPI hotplug
     code.

   - Lighter weight freezing of tasks.

     These changes from Colin Cross and Mandeep Singh Baines are
     targeted at making the freezing of tasks a bit less heavy weight
     operation.  They reduce the number of tasks woken up every time
     during the freezing, by using the observation that the freezer
     simply doesn't need to wake up some of them and wait for them all
     to call refrigerator().  The time needed for the freezer to decide
     to report a failure is reduced too.

     Also reintroduced is the check causing a lockdep warining to
     trigger when try_to_freeze() is called with locks held (which is
     generally unsafe and shouldn't happen).

   - cpufreq updates

     First off, a commit from Srivatsa S Bhat fixes a resume regression
     introduced during the 3.10 cycle causing some cpufreq sysfs
     attributes to return wrong values to user space after resume.  The
     fix is kind of fresh, but also it's pretty obvious once Srivatsa
     has identified the root cause.

     Second, we have a new freqdomain_cpus sysfs attribute for the
     acpi-cpufreq driver to provide information previously available via
     related_cpus.  From Lan Tianyu.

     Finally, we fix a number of issues, mostly related to the
     CPUFREQ_POSTCHANGE notifier and cpufreq Kconfig options and clean
     up some code.  The majority of changes from Viresh Kumar with bits
     from Jacob Shin, Heiko Stübner, Xiaoguang Chen, Ezequiel Garcia,
     Arnd Bergmann, and Tang Yuantian.

   - ACPICA update

     A usual bunch of updates from the ACPICA upstream.

     During the 3.4 cycle we introduced support for ACPI 5 extended
     sleep registers, but they are only supposed to be used if the
     HW-reduced mode bit is set in the FADT flags and the code attempted
     to use them without checking that bit.  That caused suspend/resume
     regressions to happen on some systems.  Fix from Lv Zheng causes
     those registers to be used only if the HW-reduced mode bit is set.

     Apart from this some other ACPICA bugs are fixed and code cleanups
     are made by Bob Moore, Tomasz Nowicki, Lv Zheng, Chao Guan, and
     Zhang Rui.

   - cpuidle updates

     New driver for Xilinx Zynq processors is added by Michal Simek.

     Multidriver support simplification, addition of some missing
     kerneldoc comments and Kconfig-related fixes come from Daniel
     Lezcano.

   - ACPI power management updates

     Changes to make suspend/resume work correctly in Xen guests from
     Konrad Rzeszutek Wilk, sparse warning fix from Fengguang Wu and
     cleanups and fixes of the ACPI device power state selection
     routine.

   - ACPI documentation updates

     Some previously missing pieces of ACPI documentation are added by
     Lv Zheng and Aaron Lu (hopefully, that will help people to
     uderstand how the ACPI subsystem works) and one outdated doc is
     updated by Hanjun Guo.

   - Assorted ACPI updates

     We finally nailed down the IA-64 issue that was the reason for
     reverting commit 9f29ab11dd ("ACPI / scan: do not match drivers
     against objects having scan handlers"), so we can fix it and move
     the ACPI scan handler check added to the ACPI video driver back to
     the core.

     A mechanism for adding CMOS RTC address space handlers is
     introduced by Lan Tianyu to allow some EC-related breakage to be
     fixed on some systems.

     A spec-compliant implementation of acpi_os_get_timer() is added by
     Mika Westerberg.

     The evaluation of _STA is added to do_acpi_find_child() to avoid
     situations in which a pointer to a disabled device object is
     returned instead of an enabled one with the same _ADR value.  From
     Jeff Wu.

     Intel BayTrail PCH (Platform Controller Hub) support is added to
     the ACPI driver for Intel Low-Power Subsystems (LPSS) and that
     driver is modified to work around a couple of known BIOS issues.
     Changes from Mika Westerberg and Heikki Krogerus.

     The EC driver is fixed by Vasiliy Kulikov to use get_user() and
     put_user() instead of dereferencing user space pointers blindly.

     Code cleanups are made by Bjorn Helgaas, Nicholas Mazzuca and Toshi
     Kani.

   - Assorted power management updates

     The "runtime idle" helper routine is changed to take the return
     values of the callbacks executed by it into account and to call
     rpm_suspend() if they return 0, which allows us to reduce the
     overall code bloat a bit (by dropping some code that's not
     necessary any more after that modification).

     The runtime PM documentation is updated by Alan Stern (to reflect
     the "runtime idle" behavior change).

     New trace points for PM QoS are added by Sahara
     (<keun-o.park@windriver.com>).

     PM QoS documentation is updated by Lan Tianyu.

     Code cleanups are made and minor issues are addressed by Bernie
     Thompson, Bjorn Helgaas, Julius Werner, and Shuah Khan.

   - devfreq updates

     New driver for the Exynos5-bus device from Abhilash Kesavan.

     Minor cleanups, fixes and MAINTAINERS update from MyungJoo Ham,
     Abhilash Kesavan, Paul Bolle, Rajagopal Venkat, and Wei Yongjun.

   - OMAP power management updates

     Adaptive Voltage Scaling (AVS) SmartReflex voltage control driver
     updates from Andrii Tseglytskyi and Nishanth Menon."

* tag 'pm+acpi-3.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (162 commits)
  cpufreq: Fix cpufreq regression after suspend/resume
  ACPI / PM: Fix possible NULL pointer deref in acpi_pm_device_sleep_state()
  PM / Sleep: Warn about system time after resume with pm_trace
  cpufreq: don't leave stale policy pointer in cdbs->cur_policy
  acpi-cpufreq: Add new sysfs attribute freqdomain_cpus
  cpufreq: make sure frequency transitions are serialized
  ACPI: implement acpi_os_get_timer() according the spec
  ACPI / EC: Add HP Folio 13 to ec_dmi_table in order to skip DSDT scan
  ACPI: Add CMOS RTC Operation Region handler support
  ACPI / processor: Drop unused variable from processor_perflib.c
  cpufreq: tegra: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: s3c64xx: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: omap: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: imx6q: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: exynos: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: dbx500: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: davinci: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: arm-big-little: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: powernow-k8: call CPUFREQ_POSTCHANGE notfier in error cases
  cpufreq: pcc: call CPUFREQ_POSTCHANGE notfier in error cases
  ...
2013-07-03 14:35:40 -07:00

408 lines
16 KiB
Plaintext

CPU hotplug Support in Linux(tm) Kernel
Maintainers:
CPU Hotplug Core:
Rusty Russell <rusty@rustcorp.com.au>
Srivatsa Vaddagiri <vatsa@in.ibm.com>
i386:
Zwane Mwaikambo <zwane@arm.linux.org.uk>
ppc64:
Nathan Lynch <nathanl@austin.ibm.com>
Joel Schopp <jschopp@austin.ibm.com>
ia64/x86_64:
Ashok Raj <ashok.raj@intel.com>
s390:
Heiko Carstens <heiko.carstens@de.ibm.com>
Authors: Ashok Raj <ashok.raj@intel.com>
Lots of feedback: Nathan Lynch <nathanl@austin.ibm.com>,
Joel Schopp <jschopp@austin.ibm.com>
Introduction
Modern advances in system architectures have introduced advanced error
reporting and correction capabilities in processors. CPU architectures permit
partitioning support, where compute resources of a single CPU could be made
available to virtual machine environments. There are couple OEMS that
support NUMA hardware which are hot pluggable as well, where physical
node insertion and removal require support for CPU hotplug.
Such advances require CPUs available to a kernel to be removed either for
provisioning reasons, or for RAS purposes to keep an offending CPU off
system execution path. Hence the need for CPU hotplug support in the
Linux kernel.
A more novel use of CPU-hotplug support is its use today in suspend
resume support for SMP. Dual-core and HT support makes even
a laptop run SMP kernels which didn't support these methods. SMP support
for suspend/resume is a work in progress.
General Stuff about CPU Hotplug
--------------------------------
Command Line Switches
---------------------
maxcpus=n Restrict boot time cpus to n. Say if you have 4 cpus, using
maxcpus=2 will only boot 2. You can choose to bring the
other cpus later online, read FAQ's for more info.
additional_cpus=n (*) Use this to limit hotpluggable cpus. This option sets
cpu_possible_mask = cpu_present_mask + additional_cpus
cede_offline={"off","on"} Use this option to disable/enable putting offlined
processors to an extended H_CEDE state on
supported pseries platforms.
If nothing is specified,
cede_offline is set to "on".
(*) Option valid only for following architectures
- ia64
ia64 uses the number of disabled local apics in ACPI tables MADT to
determine the number of potentially hot-pluggable cpus. The implementation
should only rely on this to count the # of cpus, but *MUST* not rely
on the apicid values in those tables for disabled apics. In the event
BIOS doesn't mark such hot-pluggable cpus as disabled entries, one could
use this parameter "additional_cpus=x" to represent those cpus in the
cpu_possible_mask.
possible_cpus=n [s390,x86_64] use this to set hotpluggable cpus.
This option sets possible_cpus bits in
cpu_possible_mask. Thus keeping the numbers of bits set
constant even if the machine gets rebooted.
CPU maps and such
-----------------
[More on cpumaps and primitive to manipulate, please check
include/linux/cpumask.h that has more descriptive text.]
cpu_possible_mask: Bitmap of possible CPUs that can ever be available in the
system. This is used to allocate some boot time memory for per_cpu variables
that aren't designed to grow/shrink as CPUs are made available or removed.
Once set during boot time discovery phase, the map is static, i.e no bits
are added or removed anytime. Trimming it accurately for your system needs
upfront can save some boot time memory. See below for how we use heuristics
in x86_64 case to keep this under check.
cpu_online_mask: Bitmap of all CPUs currently online. Its set in __cpu_up()
after a cpu is available for kernel scheduling and ready to receive
interrupts from devices. Its cleared when a cpu is brought down using
__cpu_disable(), before which all OS services including interrupts are
migrated to another target CPU.
cpu_present_mask: Bitmap of CPUs currently present in the system. Not all
of them may be online. When physical hotplug is processed by the relevant
subsystem (e.g ACPI) can change and new bit either be added or removed
from the map depending on the event is hot-add/hot-remove. There are currently
no locking rules as of now. Typical usage is to init topology during boot,
at which time hotplug is disabled.
You really dont need to manipulate any of the system cpu maps. They should
be read-only for most use. When setting up per-cpu resources almost always use
cpu_possible_mask/for_each_possible_cpu() to iterate.
Never use anything other than cpumask_t to represent bitmap of CPUs.
#include <linux/cpumask.h>
for_each_possible_cpu - Iterate over cpu_possible_mask
for_each_online_cpu - Iterate over cpu_online_mask
for_each_present_cpu - Iterate over cpu_present_mask
for_each_cpu_mask(x,mask) - Iterate over some random collection of cpu mask.
#include <linux/cpu.h>
get_online_cpus() and put_online_cpus():
The above calls are used to inhibit cpu hotplug operations. While the
cpu_hotplug.refcount is non zero, the cpu_online_mask will not change.
If you merely need to avoid cpus going away, you could also use
preempt_disable() and preempt_enable() for those sections.
Just remember the critical section cannot call any
function that can sleep or schedule this process away. The preempt_disable()
will work as long as stop_machine_run() is used to take a cpu down.
CPU Hotplug - Frequently Asked Questions.
Q: How to enable my kernel to support CPU hotplug?
A: When doing make defconfig, Enable CPU hotplug support
"Processor type and Features" -> Support for Hotpluggable CPUs
Make sure that you have CONFIG_SMP turned on as well.
You would need to enable CONFIG_HOTPLUG_CPU for SMP suspend/resume support
as well.
Q: What architectures support CPU hotplug?
A: As of 2.6.14, the following architectures support CPU hotplug.
i386 (Intel), ppc, ppc64, parisc, s390, ia64 and x86_64
Q: How to test if hotplug is supported on the newly built kernel?
A: You should now notice an entry in sysfs.
Check if sysfs is mounted, using the "mount" command. You should notice
an entry as shown below in the output.
....
none on /sys type sysfs (rw)
....
If this is not mounted, do the following.
#mkdir /sysfs
#mount -t sysfs sys /sys
Now you should see entries for all present cpu, the following is an example
in a 8-way system.
#pwd
#/sys/devices/system/cpu
#ls -l
total 0
drwxr-xr-x 10 root root 0 Sep 19 07:44 .
drwxr-xr-x 13 root root 0 Sep 19 07:45 ..
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu0
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu1
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu2
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu3
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu4
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu5
drwxr-xr-x 3 root root 0 Sep 19 07:44 cpu6
drwxr-xr-x 3 root root 0 Sep 19 07:48 cpu7
Under each directory you would find an "online" file which is the control
file to logically online/offline a processor.
Q: Does hot-add/hot-remove refer to physical add/remove of cpus?
A: The usage of hot-add/remove may not be very consistently used in the code.
CONFIG_HOTPLUG_CPU enables logical online/offline capability in the kernel.
To support physical addition/removal, one would need some BIOS hooks and
the platform should have something like an attention button in PCI hotplug.
CONFIG_ACPI_HOTPLUG_CPU enables ACPI support for physical add/remove of CPUs.
Q: How do i logically offline a CPU?
A: Do the following.
#echo 0 > /sys/devices/system/cpu/cpuX/online
Once the logical offline is successful, check
#cat /proc/interrupts
You should now not see the CPU that you removed. Also online file will report
the state as 0 when a cpu if offline and 1 when its online.
#To display the current cpu state.
#cat /sys/devices/system/cpu/cpuX/online
Q: Why can't i remove CPU0 on some systems?
A: Some architectures may have some special dependency on a certain CPU.
For e.g in IA64 platforms we have ability to sent platform interrupts to the
OS. a.k.a Corrected Platform Error Interrupts (CPEI). In current ACPI
specifications, we didn't have a way to change the target CPU. Hence if the
current ACPI version doesn't support such re-direction, we disable that CPU
by making it not-removable.
In such cases you will also notice that the online file is missing under cpu0.
Q: Is CPU0 removable on X86?
A: Yes. If kernel is compiled with CONFIG_BOOTPARAM_HOTPLUG_CPU0=y, CPU0 is
removable by default. Otherwise, CPU0 is also removable by kernel option
cpu0_hotplug.
But some features depend on CPU0. Two known dependencies are:
1. Resume from hibernate/suspend depends on CPU0. Hibernate/suspend will fail if
CPU0 is offline and you need to online CPU0 before hibernate/suspend can
continue.
2. PIC interrupts also depend on CPU0. CPU0 can't be removed if a PIC interrupt
is detected.
It's said poweroff/reboot may depend on CPU0 on some machines although I haven't
seen any poweroff/reboot failure so far after CPU0 is offline on a few tested
machines.
Please let me know if you know or see any other dependencies of CPU0.
If the dependencies are under your control, you can turn on CPU0 hotplug feature
either by CONFIG_BOOTPARAM_HOTPLUG_CPU0 or by kernel parameter cpu0_hotplug.
--Fenghua Yu <fenghua.yu@intel.com>
Q: How do i find out if a particular CPU is not removable?
A: Depending on the implementation, some architectures may show this by the
absence of the "online" file. This is done if it can be determined ahead of
time that this CPU cannot be removed.
In some situations, this can be a run time check, i.e if you try to remove the
last CPU, this will not be permitted. You can find such failures by
investigating the return value of the "echo" command.
Q: What happens when a CPU is being logically offlined?
A: The following happen, listed in no particular order :-)
- A notification is sent to in-kernel registered modules by sending an event
CPU_DOWN_PREPARE or CPU_DOWN_PREPARE_FROZEN, depending on whether or not the
CPU is being offlined while tasks are frozen due to a suspend operation in
progress
- All processes are migrated away from this outgoing CPU to new CPUs.
The new CPU is chosen from each process' current cpuset, which may be
a subset of all online CPUs.
- All interrupts targeted to this CPU is migrated to a new CPU
- timers/bottom half/task lets are also migrated to a new CPU
- Once all services are migrated, kernel calls an arch specific routine
__cpu_disable() to perform arch specific cleanup.
- Once this is successful, an event for successful cleanup is sent by an event
CPU_DEAD (or CPU_DEAD_FROZEN if tasks are frozen due to a suspend while the
CPU is being offlined).
"It is expected that each service cleans up when the CPU_DOWN_PREPARE
notifier is called, when CPU_DEAD is called its expected there is nothing
running on behalf of this CPU that was offlined"
Q: If i have some kernel code that needs to be aware of CPU arrival and
departure, how to i arrange for proper notification?
A: This is what you would need in your kernel code to receive notifications.
#include <linux/cpu.h>
static int __cpuinit foobar_cpu_callback(struct notifier_block *nfb,
unsigned long action, void *hcpu)
{
unsigned int cpu = (unsigned long)hcpu;
switch (action) {
case CPU_ONLINE:
case CPU_ONLINE_FROZEN:
foobar_online_action(cpu);
break;
case CPU_DEAD:
case CPU_DEAD_FROZEN:
foobar_dead_action(cpu);
break;
}
return NOTIFY_OK;
}
static struct notifier_block __cpuinitdata foobar_cpu_notifer =
{
.notifier_call = foobar_cpu_callback,
};
You need to call register_cpu_notifier() from your init function.
Init functions could be of two types:
1. early init (init function called when only the boot processor is online).
2. late init (init function called _after_ all the CPUs are online).
For the first case, you should add the following to your init function
register_cpu_notifier(&foobar_cpu_notifier);
For the second case, you should add the following to your init function
register_hotcpu_notifier(&foobar_cpu_notifier);
You can fail PREPARE notifiers if something doesn't work to prepare resources.
This will stop the activity and send a following CANCELED event back.
CPU_DEAD should not be failed, its just a goodness indication, but bad
things will happen if a notifier in path sent a BAD notify code.
Q: I don't see my action being called for all CPUs already up and running?
A: Yes, CPU notifiers are called only when new CPUs are on-lined or offlined.
If you need to perform some action for each cpu already in the system, then
for_each_online_cpu(i) {
foobar_cpu_callback(&foobar_cpu_notifier, CPU_UP_PREPARE, i);
foobar_cpu_callback(&foobar_cpu_notifier, CPU_ONLINE, i);
}
Q: If i would like to develop cpu hotplug support for a new architecture,
what do i need at a minimum?
A: The following are what is required for CPU hotplug infrastructure to work
correctly.
- Make sure you have an entry in Kconfig to enable CONFIG_HOTPLUG_CPU
- __cpu_up() - Arch interface to bring up a CPU
- __cpu_disable() - Arch interface to shutdown a CPU, no more interrupts
can be handled by the kernel after the routine
returns. Including local APIC timers etc are
shutdown.
- __cpu_die() - This actually supposed to ensure death of the CPU.
Actually look at some example code in other arch
that implement CPU hotplug. The processor is taken
down from the idle() loop for that specific
architecture. __cpu_die() typically waits for some
per_cpu state to be set, to ensure the processor
dead routine is called to be sure positively.
Q: I need to ensure that a particular cpu is not removed when there is some
work specific to this cpu is in progress.
A: There are two ways. If your code can be run in interrupt context, use
smp_call_function_single(), otherwise use work_on_cpu(). Note that
work_on_cpu() is slow, and can fail due to out of memory:
int my_func_on_cpu(int cpu)
{
int err;
get_online_cpus();
if (!cpu_online(cpu))
err = -EINVAL;
else
#if NEEDS_BLOCKING
err = work_on_cpu(cpu, __my_func_on_cpu, NULL);
#else
smp_call_function_single(cpu, __my_func_on_cpu, &err,
true);
#endif
put_online_cpus();
return err;
}
Q: How do we determine how many CPUs are available for hotplug.
A: There is no clear spec defined way from ACPI that can give us that
information today. Based on some input from Natalie of Unisys,
that the ACPI MADT (Multiple APIC Description Tables) marks those possible
CPUs in a system with disabled status.
Andi implemented some simple heuristics that count the number of disabled
CPUs in MADT as hotpluggable CPUS. In the case there are no disabled CPUS
we assume 1/2 the number of CPUs currently present can be hotplugged.
Caveat: ACPI MADT can only provide 256 entries in systems with only ACPI 2.0c
or earlier ACPI version supported, because the apicid field in MADT is only
8 bits. From ACPI 3.0, this limitation was removed since the apicid field
was extended to 32 bits with x2APIC introduced.
User Space Notification
Hotplug support for devices is common in Linux today. Its being used today to
support automatic configuration of network, usb and pci devices. A hotplug
event can be used to invoke an agent script to perform the configuration task.
You can add /etc/hotplug/cpu.agent to handle hotplug notification user space
scripts.
#!/bin/bash
# $Id: cpu.agent
# Kernel hotplug params include:
#ACTION=%s [online or offline]
#DEVPATH=%s
#
cd /etc/hotplug
. ./hotplug.functions
case $ACTION in
online)
echo `date` ":cpu.agent" add cpu >> /tmp/hotplug.txt
;;
offline)
echo `date` ":cpu.agent" remove cpu >>/tmp/hotplug.txt
;;
*)
debug_mesg CPU $ACTION event not supported
exit 1
;;
esac