![Alan Stern](/assets/img/avatar_default.png)
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>
451 lines
11 KiB
C
451 lines
11 KiB
C
/*
|
|
* drivers/base/memory.c - basic Memory class support
|
|
*
|
|
* 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/sysdev.h>
|
|
#include <linux/module.h>
|
|
#include <linux/init.h>
|
|
#include <linux/topology.h>
|
|
#include <linux/capability.h>
|
|
#include <linux/device.h>
|
|
#include <linux/memory.h>
|
|
#include <linux/kobject.h>
|
|
#include <linux/memory_hotplug.h>
|
|
#include <linux/mm.h>
|
|
#include <asm/atomic.h>
|
|
#include <asm/uaccess.h>
|
|
|
|
#define MEMORY_CLASS_NAME "memory"
|
|
|
|
static struct sysdev_class memory_sysdev_class = {
|
|
set_kset_name(MEMORY_CLASS_NAME),
|
|
};
|
|
|
|
static const char *memory_uevent_name(struct kset *kset, struct kobject *kobj)
|
|
{
|
|
return MEMORY_CLASS_NAME;
|
|
}
|
|
|
|
static int memory_uevent(struct kset *kset, struct kobject *kobj, char **envp,
|
|
int num_envp, char *buffer, int buffer_size)
|
|
{
|
|
int retval = 0;
|
|
|
|
return retval;
|
|
}
|
|
|
|
static struct kset_uevent_ops memory_uevent_ops = {
|
|
.name = memory_uevent_name,
|
|
.uevent = memory_uevent,
|
|
};
|
|
|
|
static BLOCKING_NOTIFIER_HEAD(memory_chain);
|
|
|
|
int register_memory_notifier(struct notifier_block *nb)
|
|
{
|
|
return blocking_notifier_chain_register(&memory_chain, nb);
|
|
}
|
|
|
|
void unregister_memory_notifier(struct notifier_block *nb)
|
|
{
|
|
blocking_notifier_chain_unregister(&memory_chain, nb);
|
|
}
|
|
|
|
/*
|
|
* register_memory - Setup a sysfs device for a memory block
|
|
*/
|
|
int register_memory(struct memory_block *memory, struct mem_section *section,
|
|
struct node *root)
|
|
{
|
|
int error;
|
|
|
|
memory->sysdev.cls = &memory_sysdev_class;
|
|
memory->sysdev.id = __section_nr(section);
|
|
|
|
error = sysdev_register(&memory->sysdev);
|
|
|
|
if (root && !error)
|
|
error = sysfs_create_link(&root->sysdev.kobj,
|
|
&memory->sysdev.kobj,
|
|
kobject_name(&memory->sysdev.kobj));
|
|
|
|
return error;
|
|
}
|
|
|
|
static void
|
|
unregister_memory(struct memory_block *memory, struct mem_section *section,
|
|
struct node *root)
|
|
{
|
|
BUG_ON(memory->sysdev.cls != &memory_sysdev_class);
|
|
BUG_ON(memory->sysdev.id != __section_nr(section));
|
|
|
|
sysdev_unregister(&memory->sysdev);
|
|
if (root)
|
|
sysfs_remove_link(&root->sysdev.kobj,
|
|
kobject_name(&memory->sysdev.kobj));
|
|
}
|
|
|
|
/*
|
|
* use this as the physical section index that this memsection
|
|
* uses.
|
|
*/
|
|
|
|
static ssize_t show_mem_phys_index(struct sys_device *dev, char *buf)
|
|
{
|
|
struct memory_block *mem =
|
|
container_of(dev, struct memory_block, sysdev);
|
|
return sprintf(buf, "%08lx\n", mem->phys_index);
|
|
}
|
|
|
|
/*
|
|
* online, offline, going offline, etc.
|
|
*/
|
|
static ssize_t show_mem_state(struct sys_device *dev, char *buf)
|
|
{
|
|
struct memory_block *mem =
|
|
container_of(dev, struct memory_block, sysdev);
|
|
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;
|
|
}
|
|
|
|
static inline int memory_notify(unsigned long val, void *v)
|
|
{
|
|
return blocking_notifier_call_chain(&memory_chain, val, v);
|
|
}
|
|
|
|
/*
|
|
* MEMORY_HOTPLUG depends on SPARSEMEM in mm/Kconfig, so it is
|
|
* OK to have direct references to sparsemem variables in here.
|
|
*/
|
|
static int
|
|
memory_block_action(struct memory_block *mem, unsigned long action)
|
|
{
|
|
int i;
|
|
unsigned long psection;
|
|
unsigned long start_pfn, start_paddr;
|
|
struct page *first_page;
|
|
int ret;
|
|
int old_state = mem->state;
|
|
|
|
psection = mem->phys_index;
|
|
first_page = pfn_to_page(psection << PFN_SECTION_SHIFT);
|
|
|
|
/*
|
|
* The probe routines leave the pages reserved, just
|
|
* as the bootmem code does. Make sure they're still
|
|
* that way.
|
|
*/
|
|
if (action == MEM_ONLINE) {
|
|
for (i = 0; i < PAGES_PER_SECTION; i++) {
|
|
if (PageReserved(first_page+i))
|
|
continue;
|
|
|
|
printk(KERN_WARNING "section number %ld page number %d "
|
|
"not reserved, was it already online? \n",
|
|
psection, i);
|
|
return -EBUSY;
|
|
}
|
|
}
|
|
|
|
switch (action) {
|
|
case MEM_ONLINE:
|
|
start_pfn = page_to_pfn(first_page);
|
|
ret = online_pages(start_pfn, PAGES_PER_SECTION);
|
|
break;
|
|
case MEM_OFFLINE:
|
|
mem->state = MEM_GOING_OFFLINE;
|
|
memory_notify(MEM_GOING_OFFLINE, NULL);
|
|
start_paddr = page_to_pfn(first_page) << PAGE_SHIFT;
|
|
ret = remove_memory(start_paddr,
|
|
PAGES_PER_SECTION << PAGE_SHIFT);
|
|
if (ret) {
|
|
mem->state = old_state;
|
|
break;
|
|
}
|
|
memory_notify(MEM_MAPPING_INVALID, NULL);
|
|
break;
|
|
default:
|
|
printk(KERN_WARNING "%s(%p, %ld) unknown action: %ld\n",
|
|
__FUNCTION__, mem, action, action);
|
|
WARN_ON(1);
|
|
ret = -EINVAL;
|
|
}
|
|
/*
|
|
* For now, only notify on successful memory operations
|
|
*/
|
|
if (!ret)
|
|
memory_notify(action, NULL);
|
|
|
|
return ret;
|
|
}
|
|
|
|
static int memory_block_change_state(struct memory_block *mem,
|
|
unsigned long to_state, unsigned long from_state_req)
|
|
{
|
|
int ret = 0;
|
|
down(&mem->state_sem);
|
|
|
|
if (mem->state != from_state_req) {
|
|
ret = -EINVAL;
|
|
goto out;
|
|
}
|
|
|
|
ret = memory_block_action(mem, to_state);
|
|
if (!ret)
|
|
mem->state = to_state;
|
|
|
|
out:
|
|
up(&mem->state_sem);
|
|
return ret;
|
|
}
|
|
|
|
static ssize_t
|
|
store_mem_state(struct sys_device *dev, const char *buf, size_t count)
|
|
{
|
|
struct memory_block *mem;
|
|
unsigned int phys_section_nr;
|
|
int ret = -EINVAL;
|
|
|
|
mem = container_of(dev, struct memory_block, sysdev);
|
|
phys_section_nr = mem->phys_index;
|
|
|
|
if (!valid_section_nr(phys_section_nr))
|
|
goto out;
|
|
|
|
if (!strncmp(buf, "online", min((int)count, 6)))
|
|
ret = memory_block_change_state(mem, MEM_ONLINE, MEM_OFFLINE);
|
|
else if(!strncmp(buf, "offline", min((int)count, 7)))
|
|
ret = memory_block_change_state(mem, MEM_OFFLINE, MEM_ONLINE);
|
|
out:
|
|
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?
|
|
*/
|
|
static ssize_t show_phys_device(struct sys_device *dev, char *buf)
|
|
{
|
|
struct memory_block *mem =
|
|
container_of(dev, struct memory_block, sysdev);
|
|
return sprintf(buf, "%d\n", mem->phys_device);
|
|
}
|
|
|
|
static SYSDEV_ATTR(phys_index, 0444, show_mem_phys_index, NULL);
|
|
static SYSDEV_ATTR(state, 0644, show_mem_state, store_mem_state);
|
|
static SYSDEV_ATTR(phys_device, 0444, show_phys_device, NULL);
|
|
|
|
#define mem_create_simple_file(mem, attr_name) \
|
|
sysdev_create_file(&mem->sysdev, &attr_##attr_name)
|
|
#define mem_remove_simple_file(mem, attr_name) \
|
|
sysdev_remove_file(&mem->sysdev, &attr_##attr_name)
|
|
|
|
/*
|
|
* Block size attribute stuff
|
|
*/
|
|
static ssize_t
|
|
print_block_size(struct class *class, char *buf)
|
|
{
|
|
return sprintf(buf, "%lx\n", (unsigned long)PAGES_PER_SECTION * PAGE_SIZE);
|
|
}
|
|
|
|
static CLASS_ATTR(block_size_bytes, 0444, print_block_size, NULL);
|
|
|
|
static int block_size_init(void)
|
|
{
|
|
sysfs_create_file(&memory_sysdev_class.kset.kobj,
|
|
&class_attr_block_size_bytes.attr);
|
|
return 0;
|
|
}
|
|
|
|
/*
|
|
* 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
|
|
memory_probe_store(struct class *class, const char *buf, size_t count)
|
|
{
|
|
u64 phys_addr;
|
|
int ret;
|
|
|
|
phys_addr = simple_strtoull(buf, NULL, 0);
|
|
|
|
ret = add_memory(phys_addr, PAGES_PER_SECTION << PAGE_SHIFT);
|
|
|
|
if (ret)
|
|
count = ret;
|
|
|
|
return count;
|
|
}
|
|
static CLASS_ATTR(probe, 0700, NULL, memory_probe_store);
|
|
|
|
static int memory_probe_init(void)
|
|
{
|
|
sysfs_create_file(&memory_sysdev_class.kset.kobj,
|
|
&class_attr_probe.attr);
|
|
return 0;
|
|
}
|
|
#else
|
|
#define memory_probe_init(...) do {} while (0)
|
|
#endif
|
|
|
|
/*
|
|
* Note that phys_device is optional. It is here to allow for
|
|
* differentiation between which *physical* devices each
|
|
* section belongs to...
|
|
*/
|
|
|
|
static int add_memory_block(unsigned long node_id, struct mem_section *section,
|
|
unsigned long state, int phys_device)
|
|
{
|
|
struct memory_block *mem = kzalloc(sizeof(*mem), GFP_KERNEL);
|
|
int ret = 0;
|
|
|
|
if (!mem)
|
|
return -ENOMEM;
|
|
|
|
mem->phys_index = __section_nr(section);
|
|
mem->state = state;
|
|
init_MUTEX(&mem->state_sem);
|
|
mem->phys_device = phys_device;
|
|
|
|
ret = register_memory(mem, section, NULL);
|
|
if (!ret)
|
|
ret = mem_create_simple_file(mem, phys_index);
|
|
if (!ret)
|
|
ret = mem_create_simple_file(mem, state);
|
|
if (!ret)
|
|
ret = mem_create_simple_file(mem, phys_device);
|
|
|
|
return ret;
|
|
}
|
|
|
|
/*
|
|
* 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.
|
|
*
|
|
* This could be made generic for all sysdev classes.
|
|
*/
|
|
static struct memory_block *find_memory_block(struct mem_section *section)
|
|
{
|
|
struct kobject *kobj;
|
|
struct sys_device *sysdev;
|
|
struct memory_block *mem;
|
|
char name[sizeof(MEMORY_CLASS_NAME) + 9 + 1];
|
|
|
|
/*
|
|
* This only works because we know that section == sysdev->id
|
|
* slightly redundant with sysdev_register()
|
|
*/
|
|
sprintf(&name[0], "%s%d", MEMORY_CLASS_NAME, __section_nr(section));
|
|
|
|
kobj = kset_find_obj(&memory_sysdev_class.kset, name);
|
|
if (!kobj)
|
|
return NULL;
|
|
|
|
sysdev = container_of(kobj, struct sys_device, kobj);
|
|
mem = container_of(sysdev, struct memory_block, sysdev);
|
|
|
|
return mem;
|
|
}
|
|
|
|
int remove_memory_block(unsigned long node_id, struct mem_section *section,
|
|
int phys_device)
|
|
{
|
|
struct memory_block *mem;
|
|
|
|
mem = find_memory_block(section);
|
|
mem_remove_simple_file(mem, phys_index);
|
|
mem_remove_simple_file(mem, state);
|
|
mem_remove_simple_file(mem, phys_device);
|
|
unregister_memory(mem, section, NULL);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/*
|
|
* need an interface for the VM to add new memory regions,
|
|
* but without onlining it.
|
|
*/
|
|
int register_new_memory(struct mem_section *section)
|
|
{
|
|
return add_memory_block(0, section, MEM_OFFLINE, 0);
|
|
}
|
|
|
|
int unregister_memory_section(struct mem_section *section)
|
|
{
|
|
if (!valid_section(section))
|
|
return -EINVAL;
|
|
|
|
return remove_memory_block(0, section, 0);
|
|
}
|
|
|
|
/*
|
|
* Initialize the sysfs support for memory devices...
|
|
*/
|
|
int __init memory_dev_init(void)
|
|
{
|
|
unsigned int i;
|
|
int ret;
|
|
|
|
memory_sysdev_class.kset.uevent_ops = &memory_uevent_ops;
|
|
ret = sysdev_class_register(&memory_sysdev_class);
|
|
|
|
/*
|
|
* Create entries for memory sections that were found
|
|
* during boot and have been initialized
|
|
*/
|
|
for (i = 0; i < NR_MEM_SECTIONS; i++) {
|
|
if (!valid_section_nr(i))
|
|
continue;
|
|
add_memory_block(0, __nr_to_section(i), MEM_ONLINE, 0);
|
|
}
|
|
|
|
memory_probe_init();
|
|
block_size_init();
|
|
|
|
return ret;
|
|
}
|