2005-10-30 01:16:54 +00:00
|
|
|
/*
|
2011-12-21 22:48:43 +00:00
|
|
|
* Memory subsystem support
|
2005-10-30 01:16:54 +00:00
|
|
|
*
|
|
|
|
* Written by Matt Tolentino <matthew.e.tolentino@intel.com>
|
|
|
|
* Dave Hansen <haveblue@us.ibm.com>
|
|
|
|
*
|
|
|
|
* This file provides the necessary infrastructure to represent
|
|
|
|
* a SPARSEMEM-memory-model system's physical memory in /sysfs.
|
|
|
|
* All arch-independent code that assumes MEMORY_HOTPLUG requires
|
|
|
|
* SPARSEMEM should be contained here, or in mm/memory_hotplug.c.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/topology.h>
|
2006-01-11 20:17:46 +00:00
|
|
|
#include <linux/capability.h>
|
2005-10-30 01:16:54 +00:00
|
|
|
#include <linux/device.h>
|
|
|
|
#include <linux/memory.h>
|
|
|
|
#include <linux/kobject.h>
|
|
|
|
#include <linux/memory_hotplug.h>
|
|
|
|
#include <linux/mm.h>
|
2008-02-05 07:35:47 +00:00
|
|
|
#include <linux/mutex.h>
|
2008-10-19 03:27:12 +00:00
|
|
|
#include <linux/stat.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
|
|
|
#include <linux/slab.h>
|
2008-10-19 03:27:12 +00:00
|
|
|
|
2011-07-26 23:09:06 +00:00
|
|
|
#include <linux/atomic.h>
|
2005-10-30 01:16:54 +00:00
|
|
|
#include <asm/uaccess.h>
|
|
|
|
|
2010-10-19 17:45:24 +00:00
|
|
|
static DEFINE_MUTEX(mem_sysfs_mutex);
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
#define MEMORY_CLASS_NAME "memory"
|
2011-01-20 16:43:34 +00:00
|
|
|
|
|
|
|
static int sections_per_block;
|
|
|
|
|
|
|
|
static inline int base_memory_block_id(int section_nr)
|
|
|
|
{
|
|
|
|
return section_nr / sections_per_block;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
static struct bus_type memory_subsys = {
|
2007-12-20 01:09:39 +00:00
|
|
|
.name = MEMORY_CLASS_NAME,
|
2011-12-21 22:48:43 +00:00
|
|
|
.dev_name = MEMORY_CLASS_NAME,
|
2005-10-30 01:16:54 +00:00
|
|
|
};
|
|
|
|
|
[PATCH] Notifier chain update: API changes
The kernel's implementation of notifier chains is unsafe. There is no
protection against entries being added to or removed from a chain while the
chain is in use. The issues were discussed in this thread:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2
We noticed that notifier chains in the kernel fall into two basic usage
classes:
"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;
"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.
We decided to codify this distinction and make it part of the API. Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name). New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain. The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.
With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed. For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections. (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)
There are some limitations, which should not be too hard to live with. For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem. Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain. (This did happen in a couple of places and the code
had to be changed to avoid it.)
Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization. Instead we use RCU. The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.
Here is the list of chains that we adjusted and their classifications. None
of them use the raw API, so for the moment it is only a placeholder.
ATOMIC CHAINS
-------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain
BLOCKING CHAINS
---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain
It's possible that some of these classifications are wrong. If they are,
please let us know or submit a patch to fix them. Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)
The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.
[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 09:16:30 +00:00
|
|
|
static BLOCKING_NOTIFIER_HEAD(memory_chain);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2006-01-06 08:10:35 +00:00
|
|
|
int register_memory_notifier(struct notifier_block *nb)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
[PATCH] Notifier chain update: API changes
The kernel's implementation of notifier chains is unsafe. There is no
protection against entries being added to or removed from a chain while the
chain is in use. The issues were discussed in this thread:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2
We noticed that notifier chains in the kernel fall into two basic usage
classes:
"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;
"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.
We decided to codify this distinction and make it part of the API. Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name). New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain. The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.
With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed. For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections. (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)
There are some limitations, which should not be too hard to live with. For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem. Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain. (This did happen in a couple of places and the code
had to be changed to avoid it.)
Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization. Instead we use RCU. The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.
Here is the list of chains that we adjusted and their classifications. None
of them use the raw API, so for the moment it is only a placeholder.
ATOMIC CHAINS
-------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain
BLOCKING CHAINS
---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain
It's possible that some of these classifications are wrong. If they are,
please let us know or submit a patch to fix them. Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)
The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.
[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 09:16:30 +00:00
|
|
|
return blocking_notifier_chain_register(&memory_chain, nb);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
2008-05-07 12:43:01 +00:00
|
|
|
EXPORT_SYMBOL(register_memory_notifier);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2006-01-06 08:10:35 +00:00
|
|
|
void unregister_memory_notifier(struct notifier_block *nb)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
[PATCH] Notifier chain update: API changes
The kernel's implementation of notifier chains is unsafe. There is no
protection against entries being added to or removed from a chain while the
chain is in use. The issues were discussed in this thread:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2
We noticed that notifier chains in the kernel fall into two basic usage
classes:
"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;
"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.
We decided to codify this distinction and make it part of the API. Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name). New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain. The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.
With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed. For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections. (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)
There are some limitations, which should not be too hard to live with. For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem. Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain. (This did happen in a couple of places and the code
had to be changed to avoid it.)
Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization. Instead we use RCU. The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.
Here is the list of chains that we adjusted and their classifications. None
of them use the raw API, so for the moment it is only a placeholder.
ATOMIC CHAINS
-------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain
BLOCKING CHAINS
---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain
It's possible that some of these classifications are wrong. If they are,
please let us know or submit a patch to fix them. Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)
The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.
[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 09:16:30 +00:00
|
|
|
blocking_notifier_chain_unregister(&memory_chain, nb);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
2008-05-07 12:43:01 +00:00
|
|
|
EXPORT_SYMBOL(unregister_memory_notifier);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2009-12-17 14:44:38 +00:00
|
|
|
static ATOMIC_NOTIFIER_HEAD(memory_isolate_chain);
|
|
|
|
|
|
|
|
int register_memory_isolate_notifier(struct notifier_block *nb)
|
|
|
|
{
|
|
|
|
return atomic_notifier_chain_register(&memory_isolate_chain, nb);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(register_memory_isolate_notifier);
|
|
|
|
|
|
|
|
void unregister_memory_isolate_notifier(struct notifier_block *nb)
|
|
|
|
{
|
|
|
|
atomic_notifier_chain_unregister(&memory_isolate_chain, nb);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(unregister_memory_isolate_notifier);
|
|
|
|
|
2012-12-12 00:00:44 +00:00
|
|
|
static void memory_block_release(struct device *dev)
|
|
|
|
{
|
|
|
|
struct memory_block *mem = container_of(dev, struct memory_block, dev);
|
|
|
|
|
|
|
|
kfree(mem);
|
|
|
|
}
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* register_memory - Setup a sysfs device for a memory block
|
|
|
|
*/
|
2008-02-11 17:23:18 +00:00
|
|
|
static
|
2011-01-20 16:43:34 +00:00
|
|
|
int register_memory(struct memory_block *memory)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
int error;
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
memory->dev.bus = &memory_subsys;
|
|
|
|
memory->dev.id = memory->start_section_nr / sections_per_block;
|
2012-12-12 00:00:44 +00:00
|
|
|
memory->dev.release = memory_block_release;
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
error = device_register(&memory->dev);
|
2005-10-30 01:16:54 +00:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
unsigned long __weak memory_block_size_bytes(void)
|
|
|
|
{
|
|
|
|
return MIN_MEMORY_BLOCK_SIZE;
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long get_memory_block_size(void)
|
|
|
|
{
|
|
|
|
unsigned long block_sz;
|
|
|
|
|
|
|
|
block_sz = memory_block_size_bytes();
|
|
|
|
|
|
|
|
/* Validate blk_sz is a power of 2 and not less than section size */
|
|
|
|
if ((block_sz & (block_sz - 1)) || (block_sz < MIN_MEMORY_BLOCK_SIZE)) {
|
|
|
|
WARN_ON(1);
|
|
|
|
block_sz = MIN_MEMORY_BLOCK_SIZE;
|
|
|
|
}
|
|
|
|
|
|
|
|
return block_sz;
|
|
|
|
}
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* use this as the physical section index that this memsection
|
|
|
|
* uses.
|
|
|
|
*/
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
static ssize_t show_mem_start_phys_index(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
struct memory_block *mem =
|
2011-12-21 22:48:43 +00:00
|
|
|
container_of(dev, struct memory_block, dev);
|
2011-01-20 16:44:29 +00:00
|
|
|
unsigned long phys_index;
|
|
|
|
|
|
|
|
phys_index = mem->start_section_nr / sections_per_block;
|
|
|
|
return sprintf(buf, "%08lx\n", phys_index);
|
|
|
|
}
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
static ssize_t show_mem_end_phys_index(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2011-01-20 16:44:29 +00:00
|
|
|
{
|
|
|
|
struct memory_block *mem =
|
2011-12-21 22:48:43 +00:00
|
|
|
container_of(dev, struct memory_block, dev);
|
2011-01-20 16:44:29 +00:00
|
|
|
unsigned long phys_index;
|
|
|
|
|
|
|
|
phys_index = mem->end_section_nr / sections_per_block;
|
|
|
|
return sprintf(buf, "%08lx\n", phys_index);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
|
2008-07-24 04:28:19 +00:00
|
|
|
/*
|
|
|
|
* Show whether the section of memory is likely to be hot-removable
|
|
|
|
*/
|
2011-12-21 22:48:43 +00:00
|
|
|
static ssize_t show_mem_removable(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2008-07-24 04:28:19 +00:00
|
|
|
{
|
2011-01-20 16:43:34 +00:00
|
|
|
unsigned long i, pfn;
|
|
|
|
int ret = 1;
|
2008-07-24 04:28:19 +00:00
|
|
|
struct memory_block *mem =
|
2011-12-21 22:48:43 +00:00
|
|
|
container_of(dev, struct memory_block, dev);
|
2008-07-24 04:28:19 +00:00
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
for (i = 0; i < sections_per_block; i++) {
|
2011-01-20 16:44:29 +00:00
|
|
|
pfn = section_nr_to_pfn(mem->start_section_nr + i);
|
2011-01-20 16:43:34 +00:00
|
|
|
ret &= is_mem_section_removable(pfn, PAGES_PER_SECTION);
|
|
|
|
}
|
|
|
|
|
2008-07-24 04:28:19 +00:00
|
|
|
return sprintf(buf, "%d\n", ret);
|
|
|
|
}
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* online, offline, going offline, etc.
|
|
|
|
*/
|
2011-12-21 22:48:43 +00:00
|
|
|
static ssize_t show_mem_state(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
struct memory_block *mem =
|
2011-12-21 22:48:43 +00:00
|
|
|
container_of(dev, struct memory_block, dev);
|
2005-10-30 01:16:54 +00:00
|
|
|
ssize_t len = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We can probably put these states in a nice little array
|
|
|
|
* so that they're not open-coded
|
|
|
|
*/
|
|
|
|
switch (mem->state) {
|
|
|
|
case MEM_ONLINE:
|
|
|
|
len = sprintf(buf, "online\n");
|
|
|
|
break;
|
|
|
|
case MEM_OFFLINE:
|
|
|
|
len = sprintf(buf, "offline\n");
|
|
|
|
break;
|
|
|
|
case MEM_GOING_OFFLINE:
|
|
|
|
len = sprintf(buf, "going-offline\n");
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
len = sprintf(buf, "ERROR-UNKNOWN-%ld\n",
|
|
|
|
mem->state);
|
|
|
|
WARN_ON(1);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return len;
|
|
|
|
}
|
|
|
|
|
2007-10-21 23:41:36 +00:00
|
|
|
int memory_notify(unsigned long val, void *v)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
[PATCH] Notifier chain update: API changes
The kernel's implementation of notifier chains is unsafe. There is no
protection against entries being added to or removed from a chain while the
chain is in use. The issues were discussed in this thread:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2
We noticed that notifier chains in the kernel fall into two basic usage
classes:
"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;
"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.
We decided to codify this distinction and make it part of the API. Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name). New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain. The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.
With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed. For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections. (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)
There are some limitations, which should not be too hard to live with. For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem. Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain. (This did happen in a couple of places and the code
had to be changed to avoid it.)
Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization. Instead we use RCU. The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.
Here is the list of chains that we adjusted and their classifications. None
of them use the raw API, so for the moment it is only a placeholder.
ATOMIC CHAINS
-------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain
BLOCKING CHAINS
---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain
It's possible that some of these classifications are wrong. If they are,
please let us know or submit a patch to fix them. Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)
The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.
[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 09:16:30 +00:00
|
|
|
return blocking_notifier_call_chain(&memory_chain, val, v);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
|
2009-12-17 14:44:38 +00:00
|
|
|
int memory_isolate_notify(unsigned long val, void *v)
|
|
|
|
{
|
|
|
|
return atomic_notifier_call_chain(&memory_isolate_chain, val, v);
|
|
|
|
}
|
|
|
|
|
2011-10-17 14:38:20 +00:00
|
|
|
/*
|
|
|
|
* The probe routines leave the pages reserved, just as the bootmem code does.
|
|
|
|
* Make sure they're still that way.
|
|
|
|
*/
|
2013-04-29 22:08:40 +00:00
|
|
|
static bool pages_correctly_reserved(unsigned long start_pfn)
|
2011-10-17 14:38:20 +00:00
|
|
|
{
|
|
|
|
int i, j;
|
|
|
|
struct page *page;
|
|
|
|
unsigned long pfn = start_pfn;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* memmap between sections is not contiguous except with
|
|
|
|
* SPARSEMEM_VMEMMAP. We lookup the page once per section
|
|
|
|
* and assume memmap is contiguous within each section
|
|
|
|
*/
|
|
|
|
for (i = 0; i < sections_per_block; i++, pfn += PAGES_PER_SECTION) {
|
|
|
|
if (WARN_ON_ONCE(!pfn_valid(pfn)))
|
|
|
|
return false;
|
|
|
|
page = pfn_to_page(pfn);
|
|
|
|
|
|
|
|
for (j = 0; j < PAGES_PER_SECTION; j++) {
|
|
|
|
if (PageReserved(page + j))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
printk(KERN_WARNING "section number %ld page number %d "
|
|
|
|
"not reserved, was it already online?\n",
|
|
|
|
pfn_to_section_nr(pfn), j);
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* MEMORY_HOTPLUG depends on SPARSEMEM in mm/Kconfig, so it is
|
|
|
|
* OK to have direct references to sparsemem variables in here.
|
|
|
|
*/
|
|
|
|
static int
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
memory_block_action(unsigned long phys_index, unsigned long action, int online_type)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
2012-10-08 23:33:58 +00:00
|
|
|
unsigned long start_pfn;
|
2011-05-11 07:25:14 +00:00
|
|
|
unsigned long nr_pages = PAGES_PER_SECTION * sections_per_block;
|
2011-10-18 21:00:57 +00:00
|
|
|
struct page *first_page;
|
2005-10-30 01:16:54 +00:00
|
|
|
int ret;
|
|
|
|
|
2011-10-18 21:00:57 +00:00
|
|
|
first_page = pfn_to_page(phys_index << PFN_SECTION_SHIFT);
|
2012-10-08 23:33:58 +00:00
|
|
|
start_pfn = page_to_pfn(first_page);
|
2011-10-18 21:00:57 +00:00
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
switch (action) {
|
|
|
|
case MEM_ONLINE:
|
2013-04-29 22:08:40 +00:00
|
|
|
if (!pages_correctly_reserved(start_pfn))
|
2011-10-17 14:38:20 +00:00
|
|
|
return -EBUSY;
|
|
|
|
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
ret = online_pages(start_pfn, nr_pages, online_type);
|
2005-10-30 01:16:54 +00:00
|
|
|
break;
|
|
|
|
case MEM_OFFLINE:
|
2012-10-08 23:33:58 +00:00
|
|
|
ret = offline_pages(start_pfn, nr_pages);
|
2005-10-30 01:16:54 +00:00
|
|
|
break;
|
|
|
|
default:
|
2011-01-20 16:43:34 +00:00
|
|
|
WARN(1, KERN_WARNING "%s(%ld, %ld) unknown action: "
|
|
|
|
"%ld\n", __func__, phys_index, action, action);
|
2005-10-30 01:16:54 +00:00
|
|
|
ret = -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2012-10-08 23:34:01 +00:00
|
|
|
static int __memory_block_change_state(struct memory_block *mem,
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
unsigned long to_state, unsigned long from_state_req,
|
|
|
|
int online_type)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
2011-10-18 21:00:57 +00:00
|
|
|
int ret = 0;
|
2011-01-20 16:43:34 +00:00
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
if (mem->state != from_state_req) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
if (to_state == MEM_OFFLINE)
|
|
|
|
mem->state = MEM_GOING_OFFLINE;
|
|
|
|
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
ret = memory_block_action(mem->start_section_nr, to_state, online_type);
|
2011-01-20 16:43:34 +00:00
|
|
|
|
2012-01-13 01:20:23 +00:00
|
|
|
if (ret) {
|
2011-01-20 16:43:34 +00:00
|
|
|
mem->state = from_state_req;
|
2012-01-13 01:20:23 +00:00
|
|
|
goto out;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2012-01-13 01:20:23 +00:00
|
|
|
mem->state = to_state;
|
|
|
|
switch (mem->state) {
|
|
|
|
case MEM_OFFLINE:
|
|
|
|
kobject_uevent(&mem->dev.kobj, KOBJ_OFFLINE);
|
|
|
|
break;
|
|
|
|
case MEM_ONLINE:
|
|
|
|
kobject_uevent(&mem->dev.kobj, KOBJ_ONLINE);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
out:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2012-10-08 23:34:01 +00:00
|
|
|
static int memory_block_change_state(struct memory_block *mem,
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
unsigned long to_state, unsigned long from_state_req,
|
|
|
|
int online_type)
|
2012-10-08 23:34:01 +00:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
mutex_lock(&mem->state_mutex);
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
ret = __memory_block_change_state(mem, to_state, from_state_req,
|
|
|
|
online_type);
|
2012-10-08 23:34:01 +00:00
|
|
|
mutex_unlock(&mem->state_mutex);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
static ssize_t
|
2011-12-21 22:48:43 +00:00
|
|
|
store_mem_state(struct device *dev,
|
|
|
|
struct device_attribute *attr, const char *buf, size_t count)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
struct memory_block *mem;
|
|
|
|
int ret = -EINVAL;
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
mem = container_of(dev, struct memory_block, dev);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
if (!strncmp(buf, "online_kernel", min_t(int, count, 13)))
|
|
|
|
ret = memory_block_change_state(mem, MEM_ONLINE,
|
|
|
|
MEM_OFFLINE, ONLINE_KERNEL);
|
|
|
|
else if (!strncmp(buf, "online_movable", min_t(int, count, 14)))
|
|
|
|
ret = memory_block_change_state(mem, MEM_ONLINE,
|
|
|
|
MEM_OFFLINE, ONLINE_MOVABLE);
|
|
|
|
else if (!strncmp(buf, "online", min_t(int, count, 6)))
|
|
|
|
ret = memory_block_change_state(mem, MEM_ONLINE,
|
|
|
|
MEM_OFFLINE, ONLINE_KEEP);
|
|
|
|
else if(!strncmp(buf, "offline", min_t(int, count, 7)))
|
|
|
|
ret = memory_block_change_state(mem, MEM_OFFLINE,
|
|
|
|
MEM_ONLINE, -1);
|
2011-01-20 16:43:34 +00:00
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
return count;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* phys_device is a bad name for this. What I really want
|
|
|
|
* is a way to differentiate between memory ranges that
|
|
|
|
* are part of physical devices that constitute
|
|
|
|
* a complete removable unit or fru.
|
|
|
|
* i.e. do these ranges belong to the same physical device,
|
|
|
|
* s.t. if I offline all of these sections I can then
|
|
|
|
* remove the physical device?
|
|
|
|
*/
|
2011-12-21 22:48:43 +00:00
|
|
|
static ssize_t show_phys_device(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
struct memory_block *mem =
|
2011-12-21 22:48:43 +00:00
|
|
|
container_of(dev, struct memory_block, dev);
|
2005-10-30 01:16:54 +00:00
|
|
|
return sprintf(buf, "%d\n", mem->phys_device);
|
|
|
|
}
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
|
|
|
|
static DEVICE_ATTR(end_phys_index, 0444, show_mem_end_phys_index, NULL);
|
|
|
|
static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
|
|
|
|
static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
|
|
|
|
static DEVICE_ATTR(removable, 0444, show_mem_removable, NULL);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
|
|
|
#define mem_create_simple_file(mem, attr_name) \
|
2011-12-21 22:48:43 +00:00
|
|
|
device_create_file(&mem->dev, &dev_attr_##attr_name)
|
2005-10-30 01:16:54 +00:00
|
|
|
#define mem_remove_simple_file(mem, attr_name) \
|
2011-12-21 22:48:43 +00:00
|
|
|
device_remove_file(&mem->dev, &dev_attr_##attr_name)
|
2005-10-30 01:16:54 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Block size attribute stuff
|
|
|
|
*/
|
|
|
|
static ssize_t
|
2011-12-21 22:48:43 +00:00
|
|
|
print_block_size(struct device *dev, struct device_attribute *attr,
|
2010-01-05 11:48:06 +00:00
|
|
|
char *buf)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
2011-01-20 16:43:34 +00:00
|
|
|
return sprintf(buf, "%lx\n", get_memory_block_size());
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
static DEVICE_ATTR(block_size_bytes, 0444, print_block_size, NULL);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
|
|
|
static int block_size_init(void)
|
|
|
|
{
|
2011-12-21 22:48:43 +00:00
|
|
|
return device_create_file(memory_subsys.dev_root,
|
|
|
|
&dev_attr_block_size_bytes);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Some architectures will have custom drivers to do this, and
|
|
|
|
* will not need to do it from userspace. The fake hot-add code
|
|
|
|
* as well as ppc64 will do all of their discovery in userspace
|
|
|
|
* and will require this interface.
|
|
|
|
*/
|
|
|
|
#ifdef CONFIG_ARCH_MEMORY_PROBE
|
|
|
|
static ssize_t
|
2011-12-21 22:48:43 +00:00
|
|
|
memory_probe_store(struct device *dev, struct device_attribute *attr,
|
2010-01-05 11:48:07 +00:00
|
|
|
const char *buf, size_t count)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
u64 phys_addr;
|
2006-06-27 09:53:30 +00:00
|
|
|
int nid;
|
2011-01-31 16:55:23 +00:00
|
|
|
int i, ret;
|
2011-09-14 20:26:15 +00:00
|
|
|
unsigned long pages_per_block = PAGES_PER_SECTION * sections_per_block;
|
2005-10-30 01:16:54 +00:00
|
|
|
|
|
|
|
phys_addr = simple_strtoull(buf, NULL, 0);
|
|
|
|
|
2011-09-14 20:26:15 +00:00
|
|
|
if (phys_addr & ((pages_per_block << PAGE_SHIFT) - 1))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2011-01-31 16:55:23 +00:00
|
|
|
for (i = 0; i < sections_per_block; i++) {
|
|
|
|
nid = memory_add_physaddr_to_nid(phys_addr);
|
|
|
|
ret = add_memory(nid, phys_addr,
|
|
|
|
PAGES_PER_SECTION << PAGE_SHIFT);
|
|
|
|
if (ret)
|
2011-03-24 06:16:18 +00:00
|
|
|
goto out;
|
2011-01-31 16:55:23 +00:00
|
|
|
|
|
|
|
phys_addr += MIN_MEMORY_BLOCK_SIZE;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-03-24 06:16:18 +00:00
|
|
|
ret = count;
|
|
|
|
out:
|
|
|
|
return ret;
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
2011-12-21 22:48:43 +00:00
|
|
|
static DEVICE_ATTR(probe, S_IWUSR, NULL, memory_probe_store);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
|
|
|
static int memory_probe_init(void)
|
|
|
|
{
|
2011-12-21 22:48:43 +00:00
|
|
|
return device_create_file(memory_subsys.dev_root, &dev_attr_probe);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
#else
|
2006-12-07 04:37:29 +00:00
|
|
|
static inline int memory_probe_init(void)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
#endif
|
|
|
|
|
2009-12-16 11:20:00 +00:00
|
|
|
#ifdef CONFIG_MEMORY_FAILURE
|
|
|
|
/*
|
|
|
|
* Support for offlining pages of memory
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Soft offline a page */
|
|
|
|
static ssize_t
|
2011-12-21 22:48:43 +00:00
|
|
|
store_soft_offline_page(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
2010-01-05 11:48:07 +00:00
|
|
|
const char *buf, size_t count)
|
2009-12-16 11:20:00 +00:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
u64 pfn;
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
if (strict_strtoull(buf, 0, &pfn) < 0)
|
|
|
|
return -EINVAL;
|
|
|
|
pfn >>= PAGE_SHIFT;
|
|
|
|
if (!pfn_valid(pfn))
|
|
|
|
return -ENXIO;
|
|
|
|
ret = soft_offline_page(pfn_to_page(pfn), 0);
|
|
|
|
return ret == 0 ? count : ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Forcibly offline a page, including killing processes. */
|
|
|
|
static ssize_t
|
2011-12-21 22:48:43 +00:00
|
|
|
store_hard_offline_page(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
2010-01-05 11:48:07 +00:00
|
|
|
const char *buf, size_t count)
|
2009-12-16 11:20:00 +00:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
u64 pfn;
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
if (strict_strtoull(buf, 0, &pfn) < 0)
|
|
|
|
return -EINVAL;
|
|
|
|
pfn >>= PAGE_SHIFT;
|
2011-12-15 18:48:12 +00:00
|
|
|
ret = memory_failure(pfn, 0, 0);
|
2009-12-16 11:20:00 +00:00
|
|
|
return ret ? ret : count;
|
|
|
|
}
|
|
|
|
|
2013-02-18 19:09:03 +00:00
|
|
|
static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
|
|
|
|
static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
|
2009-12-16 11:20:00 +00:00
|
|
|
|
|
|
|
static __init int memory_fail_init(void)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
err = device_create_file(memory_subsys.dev_root,
|
|
|
|
&dev_attr_soft_offline_page);
|
2009-12-16 11:20:00 +00:00
|
|
|
if (!err)
|
2011-12-21 22:48:43 +00:00
|
|
|
err = device_create_file(memory_subsys.dev_root,
|
|
|
|
&dev_attr_hard_offline_page);
|
2009-12-16 11:20:00 +00:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
static inline int memory_fail_init(void)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* Note that phys_device is optional. It is here to allow for
|
|
|
|
* differentiation between which *physical* devices each
|
|
|
|
* section belongs to...
|
|
|
|
*/
|
2010-03-15 04:35:03 +00:00
|
|
|
int __weak arch_get_memory_phys_device(unsigned long start_pfn)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
/*
|
|
|
|
* A reference for the returned object is held and the reference for the
|
|
|
|
* hinted object is released.
|
|
|
|
*/
|
2010-09-29 19:00:55 +00:00
|
|
|
struct memory_block *find_memory_block_hinted(struct mem_section *section,
|
|
|
|
struct memory_block *hint)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
2011-01-20 16:43:34 +00:00
|
|
|
int block_id = base_memory_block_id(__section_nr(section));
|
2011-12-21 22:48:43 +00:00
|
|
|
struct device *hintdev = hint ? &hint->dev : NULL;
|
|
|
|
struct device *dev;
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
dev = subsys_find_device_by_id(&memory_subsys, block_id, hintdev);
|
|
|
|
if (hint)
|
|
|
|
put_device(&hint->dev);
|
|
|
|
if (!dev)
|
2005-10-30 01:16:54 +00:00
|
|
|
return NULL;
|
2011-12-21 22:48:43 +00:00
|
|
|
return container_of(dev, struct memory_block, dev);
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
|
2010-09-29 19:00:55 +00:00
|
|
|
/*
|
|
|
|
* For now, we have a linear search to go find the appropriate
|
|
|
|
* memory_block corresponding to a particular phys_index. If
|
|
|
|
* this gets to be a real problem, we can always use a radix
|
|
|
|
* tree or something here.
|
|
|
|
*
|
2011-12-21 22:48:43 +00:00
|
|
|
* This could be made generic for all device subsystems.
|
2010-09-29 19:00:55 +00:00
|
|
|
*/
|
|
|
|
struct memory_block *find_memory_block(struct mem_section *section)
|
|
|
|
{
|
|
|
|
return find_memory_block_hinted(section, NULL);
|
|
|
|
}
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
static int init_memory_block(struct memory_block **memory,
|
|
|
|
struct mem_section *section, unsigned long state)
|
2010-10-19 17:44:20 +00:00
|
|
|
{
|
2011-01-20 16:43:34 +00:00
|
|
|
struct memory_block *mem;
|
2010-10-19 17:44:20 +00:00
|
|
|
unsigned long start_pfn;
|
2011-01-20 16:43:34 +00:00
|
|
|
int scn_nr;
|
2010-10-19 17:44:20 +00:00
|
|
|
int ret = 0;
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
mem = kzalloc(sizeof(*mem), GFP_KERNEL);
|
2010-10-19 17:44:20 +00:00
|
|
|
if (!mem)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
scn_nr = __section_nr(section);
|
2011-01-20 16:44:29 +00:00
|
|
|
mem->start_section_nr =
|
|
|
|
base_memory_block_id(scn_nr) * sections_per_block;
|
|
|
|
mem->end_section_nr = mem->start_section_nr + sections_per_block - 1;
|
2010-10-19 17:44:20 +00:00
|
|
|
mem->state = state;
|
2010-10-19 17:46:19 +00:00
|
|
|
mem->section_count++;
|
2010-10-19 17:44:20 +00:00
|
|
|
mutex_init(&mem->state_mutex);
|
2011-01-20 16:44:29 +00:00
|
|
|
start_pfn = section_nr_to_pfn(mem->start_section_nr);
|
2010-10-19 17:44:20 +00:00
|
|
|
mem->phys_device = arch_get_memory_phys_device(start_pfn);
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
ret = register_memory(mem);
|
2010-10-19 17:44:20 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = mem_create_simple_file(mem, phys_index);
|
2011-01-20 16:44:29 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = mem_create_simple_file(mem, end_phys_index);
|
2010-10-19 17:44:20 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = mem_create_simple_file(mem, state);
|
|
|
|
if (!ret)
|
|
|
|
ret = mem_create_simple_file(mem, phys_device);
|
|
|
|
if (!ret)
|
|
|
|
ret = mem_create_simple_file(mem, removable);
|
2011-01-20 16:43:34 +00:00
|
|
|
|
|
|
|
*memory = mem;
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int add_memory_section(int nid, struct mem_section *section,
|
2012-01-30 21:57:12 +00:00
|
|
|
struct memory_block **mem_p,
|
2011-01-20 16:43:34 +00:00
|
|
|
unsigned long state, enum mem_add_context context)
|
|
|
|
{
|
2012-01-30 21:57:12 +00:00
|
|
|
struct memory_block *mem = NULL;
|
|
|
|
int scn_nr = __section_nr(section);
|
2011-01-20 16:43:34 +00:00
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
mutex_lock(&mem_sysfs_mutex);
|
|
|
|
|
2012-01-30 21:57:12 +00:00
|
|
|
if (context == BOOT) {
|
|
|
|
/* same memory block ? */
|
|
|
|
if (mem_p && *mem_p)
|
|
|
|
if (scn_nr >= (*mem_p)->start_section_nr &&
|
|
|
|
scn_nr <= (*mem_p)->end_section_nr) {
|
|
|
|
mem = *mem_p;
|
|
|
|
kobject_get(&mem->dev.kobj);
|
|
|
|
}
|
|
|
|
} else
|
|
|
|
mem = find_memory_block(section);
|
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
if (mem) {
|
|
|
|
mem->section_count++;
|
2011-12-21 22:48:43 +00:00
|
|
|
kobject_put(&mem->dev.kobj);
|
2012-01-30 21:57:12 +00:00
|
|
|
} else {
|
2011-01-20 16:43:34 +00:00
|
|
|
ret = init_memory_block(&mem, section, state);
|
2012-01-30 21:57:12 +00:00
|
|
|
/* store memory_block pointer for next loop */
|
|
|
|
if (!ret && context == BOOT)
|
|
|
|
if (mem_p)
|
|
|
|
*mem_p = mem;
|
|
|
|
}
|
2011-01-20 16:43:34 +00:00
|
|
|
|
2010-10-19 17:44:20 +00:00
|
|
|
if (!ret) {
|
2011-01-20 16:43:34 +00:00
|
|
|
if (context == HOTPLUG &&
|
|
|
|
mem->section_count == sections_per_block)
|
2010-10-19 17:44:20 +00:00
|
|
|
ret = register_mem_sect_under_node(mem, nid);
|
|
|
|
}
|
|
|
|
|
2010-10-19 17:45:24 +00:00
|
|
|
mutex_unlock(&mem_sysfs_mutex);
|
2010-10-19 17:44:20 +00:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2013-04-29 22:08:22 +00:00
|
|
|
/*
|
|
|
|
* need an interface for the VM to add new memory regions,
|
|
|
|
* but without onlining it.
|
|
|
|
*/
|
|
|
|
int register_new_memory(int nid, struct mem_section *section)
|
|
|
|
{
|
|
|
|
return add_memory_section(nid, section, NULL, MEM_OFFLINE, HOTPLUG);
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_MEMORY_HOTREMOVE
|
|
|
|
static void
|
|
|
|
unregister_memory(struct memory_block *memory)
|
|
|
|
{
|
|
|
|
BUG_ON(memory->dev.bus != &memory_subsys);
|
|
|
|
|
|
|
|
/* drop the ref. we got in remove_memory_block() */
|
|
|
|
kobject_put(&memory->dev.kobj);
|
|
|
|
device_unregister(&memory->dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int remove_memory_block(unsigned long node_id,
|
|
|
|
struct mem_section *section, int phys_device)
|
2005-10-30 01:16:54 +00:00
|
|
|
{
|
|
|
|
struct memory_block *mem;
|
|
|
|
|
2010-10-19 17:45:24 +00:00
|
|
|
mutex_lock(&mem_sysfs_mutex);
|
2005-10-30 01:16:54 +00:00
|
|
|
mem = find_memory_block(section);
|
2011-01-20 16:44:29 +00:00
|
|
|
unregister_mem_sect_under_nodes(mem, __section_nr(section));
|
2010-10-19 17:46:19 +00:00
|
|
|
|
|
|
|
mem->section_count--;
|
|
|
|
if (mem->section_count == 0) {
|
|
|
|
mem_remove_simple_file(mem, phys_index);
|
2011-01-20 16:44:29 +00:00
|
|
|
mem_remove_simple_file(mem, end_phys_index);
|
2010-10-19 17:46:19 +00:00
|
|
|
mem_remove_simple_file(mem, state);
|
|
|
|
mem_remove_simple_file(mem, phys_device);
|
|
|
|
mem_remove_simple_file(mem, removable);
|
2011-01-20 16:43:34 +00:00
|
|
|
unregister_memory(mem);
|
|
|
|
} else
|
2011-12-21 22:48:43 +00:00
|
|
|
kobject_put(&mem->dev.kobj);
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2010-10-19 17:45:24 +00:00
|
|
|
mutex_unlock(&mem_sysfs_mutex);
|
2005-10-30 01:16:54 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int unregister_memory_section(struct mem_section *section)
|
|
|
|
{
|
2007-10-16 08:24:11 +00:00
|
|
|
if (!present_section(section))
|
2005-10-30 01:16:54 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
return remove_memory_block(0, section, 0);
|
|
|
|
}
|
2013-04-29 22:08:22 +00:00
|
|
|
#endif /* CONFIG_MEMORY_HOTREMOVE */
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2012-10-08 23:34:01 +00:00
|
|
|
/*
|
|
|
|
* offline one memory block. If the memory block has been offlined, do nothing.
|
|
|
|
*/
|
|
|
|
int offline_memory_block(struct memory_block *mem)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
mutex_lock(&mem->state_mutex);
|
|
|
|
if (mem->state != MEM_OFFLINE)
|
mm, memory-hotplug: dynamic configure movable memory and portion memory
Add online_movable and online_kernel for logic memory hotplug. This is
the dynamic version of "movablecore" & "kernelcore".
We have the same reason to introduce it as to introduce "movablecore" &
"kernelcore". It has the same motive as "movablecore" & "kernelcore", but
it is dynamic/running-time:
o We can configure memory as kernelcore or movablecore after boot.
Userspace workload is increased, we need more hugepage, we can't use
"online_movable" to add memory and allow the system use more
THP(transparent-huge-page), vice-verse when kernel workload is increase.
Also help for virtualization to dynamic configure host/guest's memory,
to save/(reduce waste) memory.
Memory capacity on Demand
o When a new node is physically online after boot, we need to use
"online_movable" or "online_kernel" to configure/portion it as we
expected when we logic-online it.
This configuration also helps for physically-memory-migrate.
o all benefit as the same as existed "movablecore" & "kernelcore".
o Preparing for movable-node, which is very important for power-saving,
hardware partitioning and high-available-system(hardware fault
management).
(Note, we don't introduce movable-node here.)
Action behavior:
When a memoryblock/memorysection is onlined by "online_movable", the kernel
will not have directly reference to the page of the memoryblock,
thus we can remove that memory any time when needed.
When it is online by "online_kernel", the kernel can use it.
When it is online by "online", the zone type doesn't changed.
Current constraints:
Only the memoryblock which is adjacent to the ZONE_MOVABLE
can be online from ZONE_NORMAL to ZONE_MOVABLE.
[akpm@linux-foundation.org: use min_t, cleanups]
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-12 00:03:16 +00:00
|
|
|
ret = __memory_block_change_state(mem, MEM_OFFLINE, MEM_ONLINE, -1);
|
2012-10-08 23:34:01 +00:00
|
|
|
mutex_unlock(&mem->state_mutex);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2013-02-23 00:32:52 +00:00
|
|
|
/* return true if the memory block is offlined, otherwise, return false */
|
|
|
|
bool is_memblock_offlined(struct memory_block *mem)
|
|
|
|
{
|
|
|
|
return mem->state == MEM_OFFLINE;
|
|
|
|
}
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* Initialize the sysfs support for memory devices...
|
|
|
|
*/
|
|
|
|
int __init memory_dev_init(void)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
int ret;
|
2006-12-07 04:37:29 +00:00
|
|
|
int err;
|
2011-01-20 16:43:34 +00:00
|
|
|
unsigned long block_sz;
|
2012-01-30 21:57:12 +00:00
|
|
|
struct memory_block *mem = NULL;
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-12-21 22:48:43 +00:00
|
|
|
ret = subsys_system_register(&memory_subsys, NULL);
|
2006-12-07 04:37:29 +00:00
|
|
|
if (ret)
|
|
|
|
goto out;
|
2005-10-30 01:16:54 +00:00
|
|
|
|
2011-01-20 16:43:34 +00:00
|
|
|
block_sz = get_memory_block_size();
|
|
|
|
sections_per_block = block_sz / MIN_MEMORY_BLOCK_SIZE;
|
|
|
|
|
2005-10-30 01:16:54 +00:00
|
|
|
/*
|
|
|
|
* Create entries for memory sections that were found
|
|
|
|
* during boot and have been initialized
|
|
|
|
*/
|
|
|
|
for (i = 0; i < NR_MEM_SECTIONS; i++) {
|
2007-10-16 08:24:11 +00:00
|
|
|
if (!present_section_nr(i))
|
2005-10-30 01:16:54 +00:00
|
|
|
continue;
|
2012-01-30 21:57:12 +00:00
|
|
|
/* don't need to reuse memory_block if only one per block */
|
|
|
|
err = add_memory_section(0, __nr_to_section(i),
|
|
|
|
(sections_per_block == 1) ? NULL : &mem,
|
|
|
|
MEM_ONLINE,
|
2011-01-20 16:43:34 +00:00
|
|
|
BOOT);
|
2006-12-07 04:37:29 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = err;
|
2005-10-30 01:16:54 +00:00
|
|
|
}
|
|
|
|
|
2006-12-07 04:37:29 +00:00
|
|
|
err = memory_probe_init();
|
2009-12-16 11:20:00 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = err;
|
|
|
|
err = memory_fail_init();
|
2006-12-07 04:37:29 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = err;
|
|
|
|
err = block_size_init();
|
|
|
|
if (!ret)
|
|
|
|
ret = err;
|
|
|
|
out:
|
|
|
|
if (ret)
|
2008-03-05 00:41:05 +00:00
|
|
|
printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
|
2005-10-30 01:16:54 +00:00
|
|
|
return ret;
|
|
|
|
}
|