linux_dsm_epyc7002/arch/powerpc/platforms/pseries/cmm.c
David Hildenbrand e352f576d3 powerpc/pseries/cmm: fix managed page counts when migrating pages between zones
Commit 63341ab037 (virtio-balloon: fix managed page counts when migrating
pages between zones) fixed a long existing BUG in the virtio-balloon
driver when pages would get migrated between zones.  I did not try to
reproduce on powerpc, but looking at the code, the same should apply to
powerpc/cmm ever since it started using the balloon compaction
infrastructure (luckily just recently).

In case we have to migrate a ballon page to a newpage of another zone, the
managed page count of both zones is wrong. Paired with memory offlining
(which will adjust the managed page count), we can trigger kernel crashes
and all kinds of different symptoms.

Fix it by properly adjusting the managed page count when migrating if
the zone changed.

We'll temporarily modify the totalram page count. If this ever becomes a
problem, we can fine tune by providing helpers that don't touch
the totalram pages (e.g., adjust_zone_managed_page_count()).

Fixes: fe030c9b85 ("powerpc/pseries/cmm: Implement balloon compaction")
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191216103058.4958-1-david@redhat.com
2019-12-16 23:15:16 +11:00

721 lines
17 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Collaborative memory management interface.
*
* Copyright (C) 2008 IBM Corporation
* Author(s): Brian King (brking@linux.vnet.ibm.com),
*/
#include <linux/ctype.h>
#include <linux/delay.h>
#include <linux/errno.h>
#include <linux/fs.h>
#include <linux/gfp.h>
#include <linux/kthread.h>
#include <linux/module.h>
#include <linux/oom.h>
#include <linux/reboot.h>
#include <linux/sched.h>
#include <linux/stringify.h>
#include <linux/swap.h>
#include <linux/device.h>
#include <linux/mount.h>
#include <linux/pseudo_fs.h>
#include <linux/magic.h>
#include <linux/balloon_compaction.h>
#include <asm/firmware.h>
#include <asm/hvcall.h>
#include <asm/mmu.h>
#include <asm/pgalloc.h>
#include <linux/uaccess.h>
#include <linux/memory.h>
#include <asm/plpar_wrappers.h>
#include "pseries.h"
#define CMM_DRIVER_VERSION "1.0.0"
#define CMM_DEFAULT_DELAY 1
#define CMM_HOTPLUG_DELAY 5
#define CMM_DEBUG 0
#define CMM_DISABLE 0
#define CMM_OOM_KB 1024
#define CMM_MIN_MEM_MB 256
#define KB2PAGES(_p) ((_p)>>(PAGE_SHIFT-10))
#define PAGES2KB(_p) ((_p)<<(PAGE_SHIFT-10))
#define CMM_MEM_HOTPLUG_PRI 1
static unsigned int delay = CMM_DEFAULT_DELAY;
static unsigned int hotplug_delay = CMM_HOTPLUG_DELAY;
static unsigned int oom_kb = CMM_OOM_KB;
static unsigned int cmm_debug = CMM_DEBUG;
static unsigned int cmm_disabled = CMM_DISABLE;
static unsigned long min_mem_mb = CMM_MIN_MEM_MB;
static bool __read_mostly simulate;
static unsigned long simulate_loan_target_kb;
static struct device cmm_dev;
MODULE_AUTHOR("Brian King <brking@linux.vnet.ibm.com>");
MODULE_DESCRIPTION("IBM System p Collaborative Memory Manager");
MODULE_LICENSE("GPL");
MODULE_VERSION(CMM_DRIVER_VERSION);
module_param_named(delay, delay, uint, 0644);
MODULE_PARM_DESC(delay, "Delay (in seconds) between polls to query hypervisor paging requests. "
"[Default=" __stringify(CMM_DEFAULT_DELAY) "]");
module_param_named(hotplug_delay, hotplug_delay, uint, 0644);
MODULE_PARM_DESC(hotplug_delay, "Delay (in seconds) after memory hotplug remove "
"before loaning resumes. "
"[Default=" __stringify(CMM_HOTPLUG_DELAY) "]");
module_param_named(oom_kb, oom_kb, uint, 0644);
MODULE_PARM_DESC(oom_kb, "Amount of memory in kb to free on OOM. "
"[Default=" __stringify(CMM_OOM_KB) "]");
module_param_named(min_mem_mb, min_mem_mb, ulong, 0644);
MODULE_PARM_DESC(min_mem_mb, "Minimum amount of memory (in MB) to not balloon. "
"[Default=" __stringify(CMM_MIN_MEM_MB) "]");
module_param_named(debug, cmm_debug, uint, 0644);
MODULE_PARM_DESC(debug, "Enable module debugging logging. Set to 1 to enable. "
"[Default=" __stringify(CMM_DEBUG) "]");
module_param_named(simulate, simulate, bool, 0444);
MODULE_PARM_DESC(simulate, "Enable simulation mode (no communication with hw).");
#define cmm_dbg(...) if (cmm_debug) { printk(KERN_INFO "cmm: "__VA_ARGS__); }
static atomic_long_t loaned_pages;
static unsigned long loaned_pages_target;
static unsigned long oom_freed_pages;
static DEFINE_MUTEX(hotplug_mutex);
static int hotplug_occurred; /* protected by the hotplug mutex */
static struct task_struct *cmm_thread_ptr;
static struct balloon_dev_info b_dev_info;
static long plpar_page_set_loaned(struct page *page)
{
const unsigned long vpa = page_to_phys(page);
unsigned long cmo_page_sz = cmo_get_page_size();
long rc = 0;
int i;
if (unlikely(simulate))
return 0;
for (i = 0; !rc && i < PAGE_SIZE; i += cmo_page_sz)
rc = plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_LOANED, vpa + i, 0);
for (i -= cmo_page_sz; rc && i != 0; i -= cmo_page_sz)
plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_ACTIVE,
vpa + i - cmo_page_sz, 0);
return rc;
}
static long plpar_page_set_active(struct page *page)
{
const unsigned long vpa = page_to_phys(page);
unsigned long cmo_page_sz = cmo_get_page_size();
long rc = 0;
int i;
if (unlikely(simulate))
return 0;
for (i = 0; !rc && i < PAGE_SIZE; i += cmo_page_sz)
rc = plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_ACTIVE, vpa + i, 0);
for (i -= cmo_page_sz; rc && i != 0; i -= cmo_page_sz)
plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_LOANED,
vpa + i - cmo_page_sz, 0);
return rc;
}
/**
* cmm_alloc_pages - Allocate pages and mark them as loaned
* @nr: number of pages to allocate
*
* Return value:
* number of pages requested to be allocated which were not
**/
static long cmm_alloc_pages(long nr)
{
struct page *page;
long rc;
cmm_dbg("Begin request for %ld pages\n", nr);
while (nr) {
/* Exit if a hotplug operation is in progress or occurred */
if (mutex_trylock(&hotplug_mutex)) {
if (hotplug_occurred) {
mutex_unlock(&hotplug_mutex);
break;
}
mutex_unlock(&hotplug_mutex);
} else {
break;
}
page = balloon_page_alloc();
if (!page)
break;
rc = plpar_page_set_loaned(page);
if (rc) {
pr_err("%s: Can not set page to loaned. rc=%ld\n", __func__, rc);
__free_page(page);
break;
}
balloon_page_enqueue(&b_dev_info, page);
atomic_long_inc(&loaned_pages);
adjust_managed_page_count(page, -1);
nr--;
}
cmm_dbg("End request with %ld pages unfulfilled\n", nr);
return nr;
}
/**
* cmm_free_pages - Free pages and mark them as active
* @nr: number of pages to free
*
* Return value:
* number of pages requested to be freed which were not
**/
static long cmm_free_pages(long nr)
{
struct page *page;
cmm_dbg("Begin free of %ld pages.\n", nr);
while (nr) {
page = balloon_page_dequeue(&b_dev_info);
if (!page)
break;
plpar_page_set_active(page);
adjust_managed_page_count(page, 1);
__free_page(page);
atomic_long_dec(&loaned_pages);
nr--;
}
cmm_dbg("End request with %ld pages unfulfilled\n", nr);
return nr;
}
/**
* cmm_oom_notify - OOM notifier
* @self: notifier block struct
* @dummy: not used
* @parm: returned - number of pages freed
*
* Return value:
* NOTIFY_OK
**/
static int cmm_oom_notify(struct notifier_block *self,
unsigned long dummy, void *parm)
{
unsigned long *freed = parm;
long nr = KB2PAGES(oom_kb);
cmm_dbg("OOM processing started\n");
nr = cmm_free_pages(nr);
loaned_pages_target = atomic_long_read(&loaned_pages);
*freed += KB2PAGES(oom_kb) - nr;
oom_freed_pages += KB2PAGES(oom_kb) - nr;
cmm_dbg("OOM processing complete\n");
return NOTIFY_OK;
}
/**
* cmm_get_mpp - Read memory performance parameters
*
* Makes hcall to query the current page loan request from the hypervisor.
*
* Return value:
* nothing
**/
static void cmm_get_mpp(void)
{
const long __loaned_pages = atomic_long_read(&loaned_pages);
const long total_pages = totalram_pages() + __loaned_pages;
int rc;
struct hvcall_mpp_data mpp_data;
signed long active_pages_target, page_loan_request, target;
signed long min_mem_pages = (min_mem_mb * 1024 * 1024) / PAGE_SIZE;
if (likely(!simulate)) {
rc = h_get_mpp(&mpp_data);
if (rc != H_SUCCESS)
return;
page_loan_request = div_s64((s64)mpp_data.loan_request,
PAGE_SIZE);
target = page_loan_request + __loaned_pages;
} else {
target = KB2PAGES(simulate_loan_target_kb);
page_loan_request = target - __loaned_pages;
}
if (target < 0 || total_pages < min_mem_pages)
target = 0;
if (target > oom_freed_pages)
target -= oom_freed_pages;
else
target = 0;
active_pages_target = total_pages - target;
if (min_mem_pages > active_pages_target)
target = total_pages - min_mem_pages;
if (target < 0)
target = 0;
loaned_pages_target = target;
cmm_dbg("delta = %ld, loaned = %lu, target = %lu, oom = %lu, totalram = %lu\n",
page_loan_request, __loaned_pages, loaned_pages_target,
oom_freed_pages, totalram_pages());
}
static struct notifier_block cmm_oom_nb = {
.notifier_call = cmm_oom_notify
};
/**
* cmm_thread - CMM task thread
* @dummy: not used
*
* Return value:
* 0
**/
static int cmm_thread(void *dummy)
{
unsigned long timeleft;
long __loaned_pages;
while (1) {
timeleft = msleep_interruptible(delay * 1000);
if (kthread_should_stop() || timeleft)
break;
if (mutex_trylock(&hotplug_mutex)) {
if (hotplug_occurred) {
hotplug_occurred = 0;
mutex_unlock(&hotplug_mutex);
cmm_dbg("Hotplug operation has occurred, "
"loaning activity suspended "
"for %d seconds.\n",
hotplug_delay);
timeleft = msleep_interruptible(hotplug_delay *
1000);
if (kthread_should_stop() || timeleft)
break;
continue;
}
mutex_unlock(&hotplug_mutex);
} else {
cmm_dbg("Hotplug operation in progress, activity "
"suspended\n");
continue;
}
cmm_get_mpp();
__loaned_pages = atomic_long_read(&loaned_pages);
if (loaned_pages_target > __loaned_pages) {
if (cmm_alloc_pages(loaned_pages_target - __loaned_pages))
loaned_pages_target = __loaned_pages;
} else if (loaned_pages_target < __loaned_pages)
cmm_free_pages(__loaned_pages - loaned_pages_target);
}
return 0;
}
#define CMM_SHOW(name, format, args...) \
static ssize_t show_##name(struct device *dev, \
struct device_attribute *attr, \
char *buf) \
{ \
return sprintf(buf, format, ##args); \
} \
static DEVICE_ATTR(name, 0444, show_##name, NULL)
CMM_SHOW(loaned_kb, "%lu\n", PAGES2KB(atomic_long_read(&loaned_pages)));
CMM_SHOW(loaned_target_kb, "%lu\n", PAGES2KB(loaned_pages_target));
static ssize_t show_oom_pages(struct device *dev,
struct device_attribute *attr, char *buf)
{
return sprintf(buf, "%lu\n", PAGES2KB(oom_freed_pages));
}
static ssize_t store_oom_pages(struct device *dev,
struct device_attribute *attr,
const char *buf, size_t count)
{
unsigned long val = simple_strtoul (buf, NULL, 10);
if (!capable(CAP_SYS_ADMIN))
return -EPERM;
if (val != 0)
return -EBADMSG;
oom_freed_pages = 0;
return count;
}
static DEVICE_ATTR(oom_freed_kb, 0644,
show_oom_pages, store_oom_pages);
static struct device_attribute *cmm_attrs[] = {
&dev_attr_loaned_kb,
&dev_attr_loaned_target_kb,
&dev_attr_oom_freed_kb,
};
static DEVICE_ULONG_ATTR(simulate_loan_target_kb, 0644,
simulate_loan_target_kb);
static struct bus_type cmm_subsys = {
.name = "cmm",
.dev_name = "cmm",
};
static void cmm_release_device(struct device *dev)
{
}
/**
* cmm_sysfs_register - Register with sysfs
*
* Return value:
* 0 on success / other on failure
**/
static int cmm_sysfs_register(struct device *dev)
{
int i, rc;
if ((rc = subsys_system_register(&cmm_subsys, NULL)))
return rc;
dev->id = 0;
dev->bus = &cmm_subsys;
dev->release = cmm_release_device;
if ((rc = device_register(dev)))
goto subsys_unregister;
for (i = 0; i < ARRAY_SIZE(cmm_attrs); i++) {
if ((rc = device_create_file(dev, cmm_attrs[i])))
goto fail;
}
if (!simulate)
return 0;
rc = device_create_file(dev, &dev_attr_simulate_loan_target_kb.attr);
if (rc)
goto fail;
return 0;
fail:
while (--i >= 0)
device_remove_file(dev, cmm_attrs[i]);
device_unregister(dev);
subsys_unregister:
bus_unregister(&cmm_subsys);
return rc;
}
/**
* cmm_unregister_sysfs - Unregister from sysfs
*
**/
static void cmm_unregister_sysfs(struct device *dev)
{
int i;
for (i = 0; i < ARRAY_SIZE(cmm_attrs); i++)
device_remove_file(dev, cmm_attrs[i]);
device_unregister(dev);
bus_unregister(&cmm_subsys);
}
/**
* cmm_reboot_notifier - Make sure pages are not still marked as "loaned"
*
**/
static int cmm_reboot_notifier(struct notifier_block *nb,
unsigned long action, void *unused)
{
if (action == SYS_RESTART) {
if (cmm_thread_ptr)
kthread_stop(cmm_thread_ptr);
cmm_thread_ptr = NULL;
cmm_free_pages(atomic_long_read(&loaned_pages));
}
return NOTIFY_DONE;
}
static struct notifier_block cmm_reboot_nb = {
.notifier_call = cmm_reboot_notifier,
};
/**
* cmm_memory_cb - Handle memory hotplug notifier calls
* @self: notifier block struct
* @action: action to take
* @arg: struct memory_notify data for handler
*
* Return value:
* NOTIFY_OK or notifier error based on subfunction return value
*
**/
static int cmm_memory_cb(struct notifier_block *self,
unsigned long action, void *arg)
{
int ret = 0;
switch (action) {
case MEM_GOING_OFFLINE:
mutex_lock(&hotplug_mutex);
hotplug_occurred = 1;
break;
case MEM_OFFLINE:
case MEM_CANCEL_OFFLINE:
mutex_unlock(&hotplug_mutex);
cmm_dbg("Memory offline operation complete.\n");
break;
case MEM_GOING_ONLINE:
case MEM_ONLINE:
case MEM_CANCEL_ONLINE:
break;
}
return notifier_from_errno(ret);
}
static struct notifier_block cmm_mem_nb = {
.notifier_call = cmm_memory_cb,
.priority = CMM_MEM_HOTPLUG_PRI
};
#ifdef CONFIG_BALLOON_COMPACTION
static struct vfsmount *balloon_mnt;
static int cmm_init_fs_context(struct fs_context *fc)
{
return init_pseudo(fc, PPC_CMM_MAGIC) ? 0 : -ENOMEM;
}
static struct file_system_type balloon_fs = {
.name = "ppc-cmm",
.init_fs_context = cmm_init_fs_context,
.kill_sb = kill_anon_super,
};
static int cmm_migratepage(struct balloon_dev_info *b_dev_info,
struct page *newpage, struct page *page,
enum migrate_mode mode)
{
unsigned long flags;
/*
* loan/"inflate" the newpage first.
*
* We might race against the cmm_thread who might discover after our
* loan request that another page is to be unloaned. However, once
* the cmm_thread runs again later, this error will automatically
* be corrected.
*/
if (plpar_page_set_loaned(newpage)) {
/* Unlikely, but possible. Tell the caller not to retry now. */
pr_err_ratelimited("%s: Cannot set page to loaned.", __func__);
return -EBUSY;
}
/* balloon page list reference */
get_page(newpage);
/*
* When we migrate a page to a different zone, we have to fixup the
* count of both involved zones as we adjusted the managed page count
* when inflating.
*/
if (page_zone(page) != page_zone(newpage)) {
adjust_managed_page_count(page, 1);
adjust_managed_page_count(newpage, -1);
}
spin_lock_irqsave(&b_dev_info->pages_lock, flags);
balloon_page_insert(b_dev_info, newpage);
balloon_page_delete(page);
b_dev_info->isolated_pages--;
spin_unlock_irqrestore(&b_dev_info->pages_lock, flags);
/*
* activate/"deflate" the old page. We ignore any errors just like the
* other callers.
*/
plpar_page_set_active(page);
/* balloon page list reference */
put_page(page);
return MIGRATEPAGE_SUCCESS;
}
static int cmm_balloon_compaction_init(void)
{
int rc;
balloon_devinfo_init(&b_dev_info);
b_dev_info.migratepage = cmm_migratepage;
balloon_mnt = kern_mount(&balloon_fs);
if (IS_ERR(balloon_mnt)) {
rc = PTR_ERR(balloon_mnt);
balloon_mnt = NULL;
return rc;
}
b_dev_info.inode = alloc_anon_inode(balloon_mnt->mnt_sb);
if (IS_ERR(b_dev_info.inode)) {
rc = PTR_ERR(b_dev_info.inode);
b_dev_info.inode = NULL;
kern_unmount(balloon_mnt);
balloon_mnt = NULL;
return rc;
}
b_dev_info.inode->i_mapping->a_ops = &balloon_aops;
return 0;
}
static void cmm_balloon_compaction_deinit(void)
{
if (b_dev_info.inode)
iput(b_dev_info.inode);
b_dev_info.inode = NULL;
kern_unmount(balloon_mnt);
balloon_mnt = NULL;
}
#else /* CONFIG_BALLOON_COMPACTION */
static int cmm_balloon_compaction_init(void)
{
return 0;
}
static void cmm_balloon_compaction_deinit(void)
{
}
#endif /* CONFIG_BALLOON_COMPACTION */
/**
* cmm_init - Module initialization
*
* Return value:
* 0 on success / other on failure
**/
static int cmm_init(void)
{
int rc;
if (!firmware_has_feature(FW_FEATURE_CMO) && !simulate)
return -EOPNOTSUPP;
rc = cmm_balloon_compaction_init();
if (rc)
return rc;
rc = register_oom_notifier(&cmm_oom_nb);
if (rc < 0)
goto out_balloon_compaction;
if ((rc = register_reboot_notifier(&cmm_reboot_nb)))
goto out_oom_notifier;
if ((rc = cmm_sysfs_register(&cmm_dev)))
goto out_reboot_notifier;
rc = register_memory_notifier(&cmm_mem_nb);
if (rc)
goto out_unregister_notifier;
if (cmm_disabled)
return 0;
cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread");
if (IS_ERR(cmm_thread_ptr)) {
rc = PTR_ERR(cmm_thread_ptr);
goto out_unregister_notifier;
}
return 0;
out_unregister_notifier:
unregister_memory_notifier(&cmm_mem_nb);
cmm_unregister_sysfs(&cmm_dev);
out_reboot_notifier:
unregister_reboot_notifier(&cmm_reboot_nb);
out_oom_notifier:
unregister_oom_notifier(&cmm_oom_nb);
out_balloon_compaction:
cmm_balloon_compaction_deinit();
return rc;
}
/**
* cmm_exit - Module exit
*
* Return value:
* nothing
**/
static void cmm_exit(void)
{
if (cmm_thread_ptr)
kthread_stop(cmm_thread_ptr);
unregister_oom_notifier(&cmm_oom_nb);
unregister_reboot_notifier(&cmm_reboot_nb);
unregister_memory_notifier(&cmm_mem_nb);
cmm_free_pages(atomic_long_read(&loaned_pages));
cmm_unregister_sysfs(&cmm_dev);
cmm_balloon_compaction_deinit();
}
/**
* cmm_set_disable - Disable/Enable CMM
*
* Return value:
* 0 on success / other on failure
**/
static int cmm_set_disable(const char *val, const struct kernel_param *kp)
{
int disable = simple_strtoul(val, NULL, 10);
if (disable != 0 && disable != 1)
return -EINVAL;
if (disable && !cmm_disabled) {
if (cmm_thread_ptr)
kthread_stop(cmm_thread_ptr);
cmm_thread_ptr = NULL;
cmm_free_pages(atomic_long_read(&loaned_pages));
} else if (!disable && cmm_disabled) {
cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread");
if (IS_ERR(cmm_thread_ptr))
return PTR_ERR(cmm_thread_ptr);
}
cmm_disabled = disable;
return 0;
}
module_param_call(disable, cmm_set_disable, param_get_uint,
&cmm_disabled, 0644);
MODULE_PARM_DESC(disable, "Disable CMM. Set to 1 to disable. "
"[Default=" __stringify(CMM_DISABLE) "]");
module_init(cmm_init);
module_exit(cmm_exit);