mm: Convert all PageMovable users to movable_operations

These drivers are rather uncomfortably hammered into the
address_space_operations hole.  They aren't filesystems and don't behave
like filesystems.  They just need their own movable_operations structure,
which we can point to directly from page->mapping.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
This commit is contained in:
Matthew Wilcox (Oracle) 2022-06-07 15:38:48 -04:00
Родитель 81218f80a7
Коммит 68f2736a85
17 изменённых файлов: 134 добавлений и 486 удалений

Просмотреть файл

@ -252,9 +252,7 @@ prototypes::
bool (*release_folio)(struct folio *, gfp_t); bool (*release_folio)(struct folio *, gfp_t);
void (*free_folio)(struct folio *); void (*free_folio)(struct folio *);
int (*direct_IO)(struct kiocb *, struct iov_iter *iter); int (*direct_IO)(struct kiocb *, struct iov_iter *iter);
bool (*isolate_page) (struct page *, isolate_mode_t);
int (*migratepage)(struct address_space *, struct page *, struct page *); int (*migratepage)(struct address_space *, struct page *, struct page *);
void (*putback_page) (struct page *);
int (*launder_folio)(struct folio *); int (*launder_folio)(struct folio *);
bool (*is_partially_uptodate)(struct folio *, size_t from, size_t count); bool (*is_partially_uptodate)(struct folio *, size_t from, size_t count);
int (*error_remove_page)(struct address_space *, struct page *); int (*error_remove_page)(struct address_space *, struct page *);
@ -280,9 +278,7 @@ invalidate_folio: yes exclusive
release_folio: yes release_folio: yes
free_folio: yes free_folio: yes
direct_IO: direct_IO:
isolate_page: yes
migratepage: yes (both) migratepage: yes (both)
putback_page: yes
launder_folio: yes launder_folio: yes
is_partially_uptodate: yes is_partially_uptodate: yes
error_remove_page: yes error_remove_page: yes

Просмотреть файл

@ -737,12 +737,8 @@ cache in your filesystem. The following members are defined:
bool (*release_folio)(struct folio *, gfp_t); bool (*release_folio)(struct folio *, gfp_t);
void (*free_folio)(struct folio *); void (*free_folio)(struct folio *);
ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter); ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter);
/* isolate a page for migration */
bool (*isolate_page) (struct page *, isolate_mode_t);
/* migrate the contents of a page to the specified target */ /* migrate the contents of a page to the specified target */
int (*migratepage) (struct page *, struct page *); int (*migratepage) (struct page *, struct page *);
/* put migration-failed page back to right list */
void (*putback_page) (struct page *);
int (*launder_folio) (struct folio *); int (*launder_folio) (struct folio *);
bool (*is_partially_uptodate) (struct folio *, size_t from, bool (*is_partially_uptodate) (struct folio *, size_t from,
@ -930,11 +926,6 @@ cache in your filesystem. The following members are defined:
data directly between the storage and the application's address data directly between the storage and the application's address
space. space.
``isolate_page``
Called by the VM when isolating a movable non-lru page. If page
is successfully isolated, VM marks the page as PG_isolated via
__SetPageIsolated.
``migrate_page`` ``migrate_page``
This is used to compact the physical memory usage. If the VM This is used to compact the physical memory usage. If the VM
wants to relocate a page (maybe off a memory card that is wants to relocate a page (maybe off a memory card that is
@ -942,9 +933,6 @@ cache in your filesystem. The following members are defined:
page to this function. migrate_page should transfer any private page to this function. migrate_page should transfer any private
data across and update any references that it has to the page. data across and update any references that it has to the page.
``putback_page``
Called by the VM when isolated page's migration fails.
``launder_folio`` ``launder_folio``
Called before freeing a folio - it writes back the dirty folio. Called before freeing a folio - it writes back the dirty folio.
To prevent redirtying the folio, it is kept locked during the To prevent redirtying the folio, it is kept locked during the

Просмотреть файл

@ -152,110 +152,15 @@ Steps:
Non-LRU page migration Non-LRU page migration
====================== ======================
Although migration originally aimed for reducing the latency of memory accesses Although migration originally aimed for reducing the latency of memory
for NUMA, compaction also uses migration to create high-order pages. accesses for NUMA, compaction also uses migration to create high-order
pages. For compaction purposes, it is also useful to be able to move
non-LRU pages, such as zsmalloc and virtio-balloon pages.
Current problem of the implementation is that it is designed to migrate only If a driver wants to make its pages movable, it should define a struct
*LRU* pages. However, there are potential non-LRU pages which can be migrated movable_operations. It then needs to call __SetPageMovable() on each
in drivers, for example, zsmalloc, virtio-balloon pages. page that it may be able to move. This uses the ``page->mapping`` field,
so this field is not available for the driver to use for other purposes.
For virtio-balloon pages, some parts of migration code path have been hooked
up and added virtio-balloon specific functions to intercept migration logics.
It's too specific to a driver so other drivers who want to make their pages
movable would have to add their own specific hooks in the migration path.
To overcome the problem, VM supports non-LRU page migration which provides
generic functions for non-LRU movable pages without driver specific hooks
in the migration path.
If a driver wants to make its pages movable, it should define three functions
which are function pointers of struct address_space_operations.
1. ``bool (*isolate_page) (struct page *page, isolate_mode_t mode);``
What VM expects from isolate_page() function of driver is to return *true*
if driver isolates the page successfully. On returning true, VM marks the page
as PG_isolated so concurrent isolation in several CPUs skip the page
for isolation. If a driver cannot isolate the page, it should return *false*.
Once page is successfully isolated, VM uses page.lru fields so driver
shouldn't expect to preserve values in those fields.
2. ``int (*migratepage) (struct address_space *mapping,``
| ``struct page *newpage, struct page *oldpage, enum migrate_mode);``
After isolation, VM calls migratepage() of driver with the isolated page.
The function of migratepage() is to move the contents of the old page to the
new page
and set up fields of struct page newpage. Keep in mind that you should
indicate to the VM the oldpage is no longer movable via __ClearPageMovable()
under page_lock if you migrated the oldpage successfully and returned
MIGRATEPAGE_SUCCESS. If driver cannot migrate the page at the moment, driver
can return -EAGAIN. On -EAGAIN, VM will retry page migration in a short time
because VM interprets -EAGAIN as "temporary migration failure". On returning
any error except -EAGAIN, VM will give up the page migration without
retrying.
Driver shouldn't touch the page.lru field while in the migratepage() function.
3. ``void (*putback_page)(struct page *);``
If migration fails on the isolated page, VM should return the isolated page
to the driver so VM calls the driver's putback_page() with the isolated page.
In this function, the driver should put the isolated page back into its own data
structure.
Non-LRU movable page flags
There are two page flags for supporting non-LRU movable page.
* PG_movable
Driver should use the function below to make page movable under page_lock::
void __SetPageMovable(struct page *page, struct address_space *mapping)
It needs argument of address_space for registering migration
family functions which will be called by VM. Exactly speaking,
PG_movable is not a real flag of struct page. Rather, VM
reuses the page->mapping's lower bits to represent it::
#define PAGE_MAPPING_MOVABLE 0x2
page->mapping = page->mapping | PAGE_MAPPING_MOVABLE;
so driver shouldn't access page->mapping directly. Instead, driver should
use page_mapping() which masks off the low two bits of page->mapping under
page lock so it can get the right struct address_space.
For testing of non-LRU movable pages, VM supports __PageMovable() function.
However, it doesn't guarantee to identify non-LRU movable pages because
the page->mapping field is unified with other variables in struct page.
If the driver releases the page after isolation by VM, page->mapping
doesn't have a stable value although it has PAGE_MAPPING_MOVABLE set
(look at __ClearPageMovable). But __PageMovable() is cheap to call whether
page is LRU or non-LRU movable once the page has been isolated because LRU
pages can never have PAGE_MAPPING_MOVABLE set in page->mapping. It is also
good for just peeking to test non-LRU movable pages before more expensive
checking with lock_page() in pfn scanning to select a victim.
For guaranteeing non-LRU movable page, VM provides PageMovable() function.
Unlike __PageMovable(), PageMovable() validates page->mapping and
mapping->a_ops->isolate_page under lock_page(). The lock_page() prevents
sudden destroying of page->mapping.
Drivers using __SetPageMovable() should clear the flag via
__ClearMovablePage() under page_lock() before the releasing the page.
* PG_isolated
To prevent concurrent isolation among several CPUs, VM marks isolated page
as PG_isolated under lock_page(). So if a CPU encounters PG_isolated
non-LRU movable page, it can skip it. Driver doesn't need to manipulate the
flag because VM will set/clear it automatically. Keep in mind that if the
driver sees a PG_isolated page, it means the page has been isolated by the
VM so it shouldn't touch the page.lru field.
The PG_isolated flag is aliased with the PG_reclaim flag so drivers
shouldn't use PG_isolated for its own purposes.
Monitoring Migration Monitoring Migration
===================== =====================
@ -286,3 +191,5 @@ THP_MIGRATION_FAIL and PGMIGRATE_FAIL to increase.
Christoph Lameter, May 8, 2006. Christoph Lameter, May 8, 2006.
Minchan Kim, Mar 28, 2016. Minchan Kim, Mar 28, 2016.
.. kernel-doc:: include/linux/migrate.h

Просмотреть файл

@ -19,9 +19,6 @@
#include <linux/stringify.h> #include <linux/stringify.h>
#include <linux/swap.h> #include <linux/swap.h>
#include <linux/device.h> #include <linux/device.h>
#include <linux/mount.h>
#include <linux/pseudo_fs.h>
#include <linux/magic.h>
#include <linux/balloon_compaction.h> #include <linux/balloon_compaction.h>
#include <asm/firmware.h> #include <asm/firmware.h>
#include <asm/hvcall.h> #include <asm/hvcall.h>
@ -500,19 +497,6 @@ static struct notifier_block cmm_mem_nb = {
}; };
#ifdef CONFIG_BALLOON_COMPACTION #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, static int cmm_migratepage(struct balloon_dev_info *b_dev_info,
struct page *newpage, struct page *page, struct page *newpage, struct page *page,
enum migrate_mode mode) enum migrate_mode mode)
@ -564,47 +548,13 @@ static int cmm_migratepage(struct balloon_dev_info *b_dev_info,
return MIGRATEPAGE_SUCCESS; return MIGRATEPAGE_SUCCESS;
} }
static int cmm_balloon_compaction_init(void) static void cmm_balloon_compaction_init(void)
{ {
int rc;
balloon_devinfo_init(&b_dev_info); balloon_devinfo_init(&b_dev_info);
b_dev_info.migratepage = cmm_migratepage; 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 */ #else /* CONFIG_BALLOON_COMPACTION */
static int cmm_balloon_compaction_init(void) static void cmm_balloon_compaction_init(void)
{
return 0;
}
static void cmm_balloon_compaction_deinit(void)
{ {
} }
#endif /* CONFIG_BALLOON_COMPACTION */ #endif /* CONFIG_BALLOON_COMPACTION */
@ -622,9 +572,7 @@ static int cmm_init(void)
if (!firmware_has_feature(FW_FEATURE_CMO) && !simulate) if (!firmware_has_feature(FW_FEATURE_CMO) && !simulate)
return -EOPNOTSUPP; return -EOPNOTSUPP;
rc = cmm_balloon_compaction_init(); cmm_balloon_compaction_init();
if (rc)
return rc;
rc = register_oom_notifier(&cmm_oom_nb); rc = register_oom_notifier(&cmm_oom_nb);
if (rc < 0) if (rc < 0)
@ -658,7 +606,6 @@ out_reboot_notifier:
out_oom_notifier: out_oom_notifier:
unregister_oom_notifier(&cmm_oom_nb); unregister_oom_notifier(&cmm_oom_nb);
out_balloon_compaction: out_balloon_compaction:
cmm_balloon_compaction_deinit();
return rc; return rc;
} }
@ -677,7 +624,6 @@ static void cmm_exit(void)
unregister_memory_notifier(&cmm_mem_nb); unregister_memory_notifier(&cmm_mem_nb);
cmm_free_pages(atomic_long_read(&loaned_pages)); cmm_free_pages(atomic_long_read(&loaned_pages));
cmm_unregister_sysfs(&cmm_dev); cmm_unregister_sysfs(&cmm_dev);
cmm_balloon_compaction_deinit();
} }
/** /**

Просмотреть файл

@ -29,8 +29,6 @@
#include <linux/rwsem.h> #include <linux/rwsem.h>
#include <linux/slab.h> #include <linux/slab.h>
#include <linux/spinlock.h> #include <linux/spinlock.h>
#include <linux/mount.h>
#include <linux/pseudo_fs.h>
#include <linux/balloon_compaction.h> #include <linux/balloon_compaction.h>
#include <linux/vmw_vmci_defs.h> #include <linux/vmw_vmci_defs.h>
#include <linux/vmw_vmci_api.h> #include <linux/vmw_vmci_api.h>
@ -1730,20 +1728,6 @@ static inline void vmballoon_debugfs_exit(struct vmballoon *b)
#ifdef CONFIG_BALLOON_COMPACTION #ifdef CONFIG_BALLOON_COMPACTION
static int vmballoon_init_fs_context(struct fs_context *fc)
{
return init_pseudo(fc, BALLOON_VMW_MAGIC) ? 0 : -ENOMEM;
}
static struct file_system_type vmballoon_fs = {
.name = "balloon-vmware",
.init_fs_context = vmballoon_init_fs_context,
.kill_sb = kill_anon_super,
};
static struct vfsmount *vmballoon_mnt;
/** /**
* vmballoon_migratepage() - migrates a balloon page. * vmballoon_migratepage() - migrates a balloon page.
* @b_dev_info: balloon device information descriptor. * @b_dev_info: balloon device information descriptor.
@ -1862,21 +1846,6 @@ out_unlock:
return ret; return ret;
} }
/**
* vmballoon_compaction_deinit() - removes compaction related data.
*
* @b: pointer to the balloon.
*/
static void vmballoon_compaction_deinit(struct vmballoon *b)
{
if (!IS_ERR(b->b_dev_info.inode))
iput(b->b_dev_info.inode);
b->b_dev_info.inode = NULL;
kern_unmount(vmballoon_mnt);
vmballoon_mnt = NULL;
}
/** /**
* vmballoon_compaction_init() - initialized compaction for the balloon. * vmballoon_compaction_init() - initialized compaction for the balloon.
* *
@ -1888,33 +1857,15 @@ static void vmballoon_compaction_deinit(struct vmballoon *b)
* *
* Return: zero on success or error code on failure. * Return: zero on success or error code on failure.
*/ */
static __init int vmballoon_compaction_init(struct vmballoon *b) static __init void vmballoon_compaction_init(struct vmballoon *b)
{ {
vmballoon_mnt = kern_mount(&vmballoon_fs);
if (IS_ERR(vmballoon_mnt))
return PTR_ERR(vmballoon_mnt);
b->b_dev_info.migratepage = vmballoon_migratepage; b->b_dev_info.migratepage = vmballoon_migratepage;
b->b_dev_info.inode = alloc_anon_inode(vmballoon_mnt->mnt_sb);
if (IS_ERR(b->b_dev_info.inode))
return PTR_ERR(b->b_dev_info.inode);
b->b_dev_info.inode->i_mapping->a_ops = &balloon_aops;
return 0;
} }
#else /* CONFIG_BALLOON_COMPACTION */ #else /* CONFIG_BALLOON_COMPACTION */
static inline void vmballoon_compaction_init(struct vmballoon *b)
static void vmballoon_compaction_deinit(struct vmballoon *b)
{ {
} }
static int vmballoon_compaction_init(struct vmballoon *b)
{
return 0;
}
#endif /* CONFIG_BALLOON_COMPACTION */ #endif /* CONFIG_BALLOON_COMPACTION */
static int __init vmballoon_init(void) static int __init vmballoon_init(void)
@ -1939,9 +1890,7 @@ static int __init vmballoon_init(void)
* balloon_devinfo_init() . * balloon_devinfo_init() .
*/ */
balloon_devinfo_init(&balloon.b_dev_info); balloon_devinfo_init(&balloon.b_dev_info);
error = vmballoon_compaction_init(&balloon); vmballoon_compaction_init(&balloon);
if (error)
goto fail;
INIT_LIST_HEAD(&balloon.huge_pages); INIT_LIST_HEAD(&balloon.huge_pages);
spin_lock_init(&balloon.comm_lock); spin_lock_init(&balloon.comm_lock);
@ -1958,7 +1907,6 @@ static int __init vmballoon_init(void)
return 0; return 0;
fail: fail:
vmballoon_unregister_shrinker(&balloon); vmballoon_unregister_shrinker(&balloon);
vmballoon_compaction_deinit(&balloon);
return error; return error;
} }
@ -1985,8 +1933,5 @@ static void __exit vmballoon_exit(void)
*/ */
vmballoon_send_start(&balloon, 0); vmballoon_send_start(&balloon, 0);
vmballoon_pop(&balloon); vmballoon_pop(&balloon);
/* Only once we popped the balloon, compaction can be deinit */
vmballoon_compaction_deinit(&balloon);
} }
module_exit(vmballoon_exit); module_exit(vmballoon_exit);

Просмотреть файл

@ -17,9 +17,6 @@
#include <linux/oom.h> #include <linux/oom.h>
#include <linux/wait.h> #include <linux/wait.h>
#include <linux/mm.h> #include <linux/mm.h>
#include <linux/mount.h>
#include <linux/magic.h>
#include <linux/pseudo_fs.h>
#include <linux/page_reporting.h> #include <linux/page_reporting.h>
/* /*
@ -42,10 +39,6 @@
(1 << (VIRTIO_BALLOON_HINT_BLOCK_ORDER + PAGE_SHIFT)) (1 << (VIRTIO_BALLOON_HINT_BLOCK_ORDER + PAGE_SHIFT))
#define VIRTIO_BALLOON_HINT_BLOCK_PAGES (1 << VIRTIO_BALLOON_HINT_BLOCK_ORDER) #define VIRTIO_BALLOON_HINT_BLOCK_PAGES (1 << VIRTIO_BALLOON_HINT_BLOCK_ORDER)
#ifdef CONFIG_BALLOON_COMPACTION
static struct vfsmount *balloon_mnt;
#endif
enum virtio_balloon_vq { enum virtio_balloon_vq {
VIRTIO_BALLOON_VQ_INFLATE, VIRTIO_BALLOON_VQ_INFLATE,
VIRTIO_BALLOON_VQ_DEFLATE, VIRTIO_BALLOON_VQ_DEFLATE,
@ -805,18 +798,6 @@ static int virtballoon_migratepage(struct balloon_dev_info *vb_dev_info,
return MIGRATEPAGE_SUCCESS; return MIGRATEPAGE_SUCCESS;
} }
static int balloon_init_fs_context(struct fs_context *fc)
{
return init_pseudo(fc, BALLOON_KVM_MAGIC) ? 0 : -ENOMEM;
}
static struct file_system_type balloon_fs = {
.name = "balloon-kvm",
.init_fs_context = balloon_init_fs_context,
.kill_sb = kill_anon_super,
};
#endif /* CONFIG_BALLOON_COMPACTION */ #endif /* CONFIG_BALLOON_COMPACTION */
static unsigned long shrink_free_pages(struct virtio_balloon *vb, static unsigned long shrink_free_pages(struct virtio_balloon *vb,
@ -909,19 +890,7 @@ static int virtballoon_probe(struct virtio_device *vdev)
goto out_free_vb; goto out_free_vb;
#ifdef CONFIG_BALLOON_COMPACTION #ifdef CONFIG_BALLOON_COMPACTION
balloon_mnt = kern_mount(&balloon_fs);
if (IS_ERR(balloon_mnt)) {
err = PTR_ERR(balloon_mnt);
goto out_del_vqs;
}
vb->vb_dev_info.migratepage = virtballoon_migratepage; vb->vb_dev_info.migratepage = virtballoon_migratepage;
vb->vb_dev_info.inode = alloc_anon_inode(balloon_mnt->mnt_sb);
if (IS_ERR(vb->vb_dev_info.inode)) {
err = PTR_ERR(vb->vb_dev_info.inode);
goto out_kern_unmount;
}
vb->vb_dev_info.inode->i_mapping->a_ops = &balloon_aops;
#endif #endif
if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_FREE_PAGE_HINT)) { if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_FREE_PAGE_HINT)) {
/* /*
@ -930,13 +899,13 @@ static int virtballoon_probe(struct virtio_device *vdev)
*/ */
if (virtqueue_get_vring_size(vb->free_page_vq) < 2) { if (virtqueue_get_vring_size(vb->free_page_vq) < 2) {
err = -ENOSPC; err = -ENOSPC;
goto out_iput; goto out_del_vqs;
} }
vb->balloon_wq = alloc_workqueue("balloon-wq", vb->balloon_wq = alloc_workqueue("balloon-wq",
WQ_FREEZABLE | WQ_CPU_INTENSIVE, 0); WQ_FREEZABLE | WQ_CPU_INTENSIVE, 0);
if (!vb->balloon_wq) { if (!vb->balloon_wq) {
err = -ENOMEM; err = -ENOMEM;
goto out_iput; goto out_del_vqs;
} }
INIT_WORK(&vb->report_free_page_work, report_free_page_func); INIT_WORK(&vb->report_free_page_work, report_free_page_func);
vb->cmd_id_received_cache = VIRTIO_BALLOON_CMD_ID_STOP; vb->cmd_id_received_cache = VIRTIO_BALLOON_CMD_ID_STOP;
@ -1030,13 +999,7 @@ out_unregister_shrinker:
out_del_balloon_wq: out_del_balloon_wq:
if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_FREE_PAGE_HINT)) if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_FREE_PAGE_HINT))
destroy_workqueue(vb->balloon_wq); destroy_workqueue(vb->balloon_wq);
out_iput:
#ifdef CONFIG_BALLOON_COMPACTION
iput(vb->vb_dev_info.inode);
out_kern_unmount:
kern_unmount(balloon_mnt);
out_del_vqs: out_del_vqs:
#endif
vdev->config->del_vqs(vdev); vdev->config->del_vqs(vdev);
out_free_vb: out_free_vb:
kfree(vb); kfree(vb);
@ -1083,12 +1046,6 @@ static void virtballoon_remove(struct virtio_device *vdev)
} }
remove_common(vb); remove_common(vb);
#ifdef CONFIG_BALLOON_COMPACTION
if (vb->vb_dev_info.inode)
iput(vb->vb_dev_info.inode);
kern_unmount(balloon_mnt);
#endif
kfree(vb); kfree(vb);
} }

Просмотреть файл

@ -57,7 +57,6 @@ struct balloon_dev_info {
struct list_head pages; /* Pages enqueued & handled to Host */ struct list_head pages; /* Pages enqueued & handled to Host */
int (*migratepage)(struct balloon_dev_info *, struct page *newpage, int (*migratepage)(struct balloon_dev_info *, struct page *newpage,
struct page *page, enum migrate_mode mode); struct page *page, enum migrate_mode mode);
struct inode *inode;
}; };
extern struct page *balloon_page_alloc(void); extern struct page *balloon_page_alloc(void);
@ -75,11 +74,10 @@ static inline void balloon_devinfo_init(struct balloon_dev_info *balloon)
spin_lock_init(&balloon->pages_lock); spin_lock_init(&balloon->pages_lock);
INIT_LIST_HEAD(&balloon->pages); INIT_LIST_HEAD(&balloon->pages);
balloon->migratepage = NULL; balloon->migratepage = NULL;
balloon->inode = NULL;
} }
#ifdef CONFIG_BALLOON_COMPACTION #ifdef CONFIG_BALLOON_COMPACTION
extern const struct address_space_operations balloon_aops; extern const struct movable_operations balloon_mops;
/* /*
* balloon_page_insert - insert a page into the balloon's page list and make * balloon_page_insert - insert a page into the balloon's page list and make
@ -94,7 +92,7 @@ static inline void balloon_page_insert(struct balloon_dev_info *balloon,
struct page *page) struct page *page)
{ {
__SetPageOffline(page); __SetPageOffline(page);
__SetPageMovable(page, balloon->inode->i_mapping); __SetPageMovable(page, &balloon_mops);
set_page_private(page, (unsigned long)balloon); set_page_private(page, (unsigned long)balloon);
list_add(&page->lru, &balloon->pages); list_add(&page->lru, &balloon->pages);
} }

Просмотреть файл

@ -367,8 +367,6 @@ struct address_space_operations {
*/ */
int (*migratepage) (struct address_space *, int (*migratepage) (struct address_space *,
struct page *, struct page *, enum migrate_mode); struct page *, struct page *, enum migrate_mode);
bool (*isolate_page)(struct page *, isolate_mode_t);
void (*putback_page)(struct page *);
int (*launder_folio)(struct folio *); int (*launder_folio)(struct folio *);
bool (*is_partially_uptodate) (struct folio *, size_t from, bool (*is_partially_uptodate) (struct folio *, size_t from,
size_t count); size_t count);

Просмотреть файл

@ -19,6 +19,43 @@ struct migration_target_control;
*/ */
#define MIGRATEPAGE_SUCCESS 0 #define MIGRATEPAGE_SUCCESS 0
/**
* struct movable_operations - Driver page migration
* @isolate_page:
* The VM calls this function to prepare the page to be moved. The page
* is locked and the driver should not unlock it. The driver should
* return ``true`` if the page is movable and ``false`` if it is not
* currently movable. After this function returns, the VM uses the
* page->lru field, so the driver must preserve any information which
* is usually stored here.
*
* @migrate_page:
* After isolation, the VM calls this function with the isolated
* @src page. The driver should copy the contents of the
* @src page to the @dst page and set up the fields of @dst page.
* Both pages are locked.
* If page migration is successful, the driver should call
* __ClearPageMovable(@src) and return MIGRATEPAGE_SUCCESS.
* If the driver cannot migrate the page at the moment, it can return
* -EAGAIN. The VM interprets this as a temporary migration failure and
* will retry it later. Any other error value is a permanent migration
* failure and migration will not be retried.
* The driver shouldn't touch the @src->lru field while in the
* migrate_page() function. It may write to @dst->lru.
*
* @putback_page:
* If migration fails on the isolated page, the VM informs the driver
* that the page is no longer a candidate for migration by calling
* this function. The driver should put the isolated page back into
* its own data structure.
*/
struct movable_operations {
bool (*isolate_page)(struct page *, isolate_mode_t);
int (*migrate_page)(struct page *dst, struct page *src,
enum migrate_mode);
void (*putback_page)(struct page *);
};
/* Defined in mm/debug.c: */ /* Defined in mm/debug.c: */
extern const char *migrate_reason_names[MR_TYPES]; extern const char *migrate_reason_names[MR_TYPES];
@ -91,13 +128,13 @@ static inline int next_demotion_node(int node)
#endif #endif
#ifdef CONFIG_COMPACTION #ifdef CONFIG_COMPACTION
extern int PageMovable(struct page *page); bool PageMovable(struct page *page);
extern void __SetPageMovable(struct page *page, struct address_space *mapping); void __SetPageMovable(struct page *page, const struct movable_operations *ops);
extern void __ClearPageMovable(struct page *page); void __ClearPageMovable(struct page *page);
#else #else
static inline int PageMovable(struct page *page) { return 0; } static inline bool PageMovable(struct page *page) { return false; }
static inline void __SetPageMovable(struct page *page, static inline void __SetPageMovable(struct page *page,
struct address_space *mapping) const struct movable_operations *ops)
{ {
} }
static inline void __ClearPageMovable(struct page *page) static inline void __ClearPageMovable(struct page *page)
@ -110,6 +147,15 @@ static inline bool folio_test_movable(struct folio *folio)
return PageMovable(&folio->page); return PageMovable(&folio->page);
} }
static inline
const struct movable_operations *page_movable_ops(struct page *page)
{
VM_BUG_ON(!__PageMovable(page));
return (const struct movable_operations *)
((unsigned long)page->mapping - PAGE_MAPPING_MOVABLE);
}
#ifdef CONFIG_NUMA_BALANCING #ifdef CONFIG_NUMA_BALANCING
extern int migrate_misplaced_page(struct page *page, extern int migrate_misplaced_page(struct page *page,
struct vm_area_struct *vma, int node); struct vm_area_struct *vma, int node);

Просмотреть файл

@ -639,7 +639,7 @@ __PAGEFLAG(Reported, reported, PF_NO_COMPOUND)
* structure which KSM associates with that merged page. See ksm.h. * structure which KSM associates with that merged page. See ksm.h.
* *
* PAGE_MAPPING_KSM without PAGE_MAPPING_ANON is used for non-lru movable * PAGE_MAPPING_KSM without PAGE_MAPPING_ANON is used for non-lru movable
* page and then page->mapping points a struct address_space. * page and then page->mapping points to a struct movable_operations.
* *
* Please note that, confusingly, "page_mapping" refers to the inode * Please note that, confusingly, "page_mapping" refers to the inode
* address_space which maps the page from disk; whereas "page_mapped" * address_space which maps the page from disk; whereas "page_mapped"

Просмотреть файл

@ -98,12 +98,8 @@
/* Since UDF 2.01 is ISO 13346 based... */ /* Since UDF 2.01 is ISO 13346 based... */
#define UDF_SUPER_MAGIC 0x15013346 #define UDF_SUPER_MAGIC 0x15013346
#define BALLOON_KVM_MAGIC 0x13661366
#define ZSMALLOC_MAGIC 0x58295829
#define DMA_BUF_MAGIC 0x444d4142 /* "DMAB" */ #define DMA_BUF_MAGIC 0x444d4142 /* "DMAB" */
#define DEVMEM_MAGIC 0x454d444d /* "DMEM" */ #define DEVMEM_MAGIC 0x454d444d /* "DMEM" */
#define Z3FOLD_MAGIC 0x33
#define PPC_CMM_MAGIC 0xc7571590
#define SECRETMEM_MAGIC 0x5345434d /* "SECM" */ #define SECRETMEM_MAGIC 0x5345434d /* "SECM" */
#endif /* __LINUX_MAGIC_H__ */ #endif /* __LINUX_MAGIC_H__ */

Просмотреть файл

@ -228,10 +228,8 @@ static void balloon_page_putback(struct page *page)
spin_unlock_irqrestore(&b_dev_info->pages_lock, flags); spin_unlock_irqrestore(&b_dev_info->pages_lock, flags);
} }
/* move_to_new_page() counterpart for a ballooned page */ /* move_to_new_page() counterpart for a ballooned page */
static int balloon_page_migrate(struct address_space *mapping, static int balloon_page_migrate(struct page *newpage, struct page *page,
struct page *newpage, struct page *page,
enum migrate_mode mode) enum migrate_mode mode)
{ {
struct balloon_dev_info *balloon = balloon_page_device(page); struct balloon_dev_info *balloon = balloon_page_device(page);
@ -250,11 +248,11 @@ static int balloon_page_migrate(struct address_space *mapping,
return balloon->migratepage(balloon, newpage, page, mode); return balloon->migratepage(balloon, newpage, page, mode);
} }
const struct address_space_operations balloon_aops = { const struct movable_operations balloon_mops = {
.migratepage = balloon_page_migrate, .migrate_page = balloon_page_migrate,
.isolate_page = balloon_page_isolate, .isolate_page = balloon_page_isolate,
.putback_page = balloon_page_putback, .putback_page = balloon_page_putback,
}; };
EXPORT_SYMBOL_GPL(balloon_aops); EXPORT_SYMBOL_GPL(balloon_mops);
#endif /* CONFIG_BALLOON_COMPACTION */ #endif /* CONFIG_BALLOON_COMPACTION */

Просмотреть файл

@ -110,28 +110,27 @@ static void split_map_pages(struct list_head *list)
} }
#ifdef CONFIG_COMPACTION #ifdef CONFIG_COMPACTION
bool PageMovable(struct page *page)
int PageMovable(struct page *page)
{ {
struct address_space *mapping; const struct movable_operations *mops;
VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page);
if (!__PageMovable(page)) if (!__PageMovable(page))
return 0; return false;
mapping = page_mapping(page); mops = page_movable_ops(page);
if (mapping && mapping->a_ops && mapping->a_ops->isolate_page) if (mops)
return 1; return true;
return 0; return false;
} }
EXPORT_SYMBOL(PageMovable); EXPORT_SYMBOL(PageMovable);
void __SetPageMovable(struct page *page, struct address_space *mapping) void __SetPageMovable(struct page *page, const struct movable_operations *mops)
{ {
VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page);
VM_BUG_ON_PAGE((unsigned long)mapping & PAGE_MAPPING_MOVABLE, page); VM_BUG_ON_PAGE((unsigned long)mops & PAGE_MAPPING_MOVABLE, page);
page->mapping = (void *)((unsigned long)mapping | PAGE_MAPPING_MOVABLE); page->mapping = (void *)((unsigned long)mops | PAGE_MAPPING_MOVABLE);
} }
EXPORT_SYMBOL(__SetPageMovable); EXPORT_SYMBOL(__SetPageMovable);
@ -139,12 +138,10 @@ void __ClearPageMovable(struct page *page)
{ {
VM_BUG_ON_PAGE(!PageMovable(page), page); VM_BUG_ON_PAGE(!PageMovable(page), page);
/* /*
* Clear registered address_space val with keeping PAGE_MAPPING_MOVABLE * This page still has the type of a movable page, but it's
* flag so that VM can catch up released page by driver after isolation. * actually not movable any more.
* With it, VM migration doesn't try to put it back.
*/ */
page->mapping = (void *)((unsigned long)page->mapping & page->mapping = (void *)PAGE_MAPPING_MOVABLE;
PAGE_MAPPING_MOVABLE);
} }
EXPORT_SYMBOL(__ClearPageMovable); EXPORT_SYMBOL(__ClearPageMovable);

Просмотреть файл

@ -59,7 +59,7 @@
int isolate_movable_page(struct page *page, isolate_mode_t mode) int isolate_movable_page(struct page *page, isolate_mode_t mode)
{ {
struct address_space *mapping; const struct movable_operations *mops;
/* /*
* Avoid burning cycles with pages that are yet under __free_pages(), * Avoid burning cycles with pages that are yet under __free_pages(),
@ -97,10 +97,10 @@ int isolate_movable_page(struct page *page, isolate_mode_t mode)
if (!PageMovable(page) || PageIsolated(page)) if (!PageMovable(page) || PageIsolated(page))
goto out_no_isolated; goto out_no_isolated;
mapping = page_mapping(page); mops = page_movable_ops(page);
VM_BUG_ON_PAGE(!mapping, page); VM_BUG_ON_PAGE(!mops, page);
if (!mapping->a_ops->isolate_page(page, mode)) if (!mops->isolate_page(page, mode))
goto out_no_isolated; goto out_no_isolated;
/* Driver shouldn't use PG_isolated bit of page->flags */ /* Driver shouldn't use PG_isolated bit of page->flags */
@ -120,10 +120,9 @@ out:
static void putback_movable_page(struct page *page) static void putback_movable_page(struct page *page)
{ {
struct address_space *mapping; const struct movable_operations *mops = page_movable_ops(page);
mapping = page_mapping(page); mops->putback_page(page);
mapping->a_ops->putback_page(page);
ClearPageIsolated(page); ClearPageIsolated(page);
} }
@ -846,16 +845,15 @@ static int fallback_migrate_page(struct address_space *mapping,
static int move_to_new_folio(struct folio *dst, struct folio *src, static int move_to_new_folio(struct folio *dst, struct folio *src,
enum migrate_mode mode) enum migrate_mode mode)
{ {
struct address_space *mapping;
int rc = -EAGAIN; int rc = -EAGAIN;
bool is_lru = !__PageMovable(&src->page); bool is_lru = !__PageMovable(&src->page);
VM_BUG_ON_FOLIO(!folio_test_locked(src), src); VM_BUG_ON_FOLIO(!folio_test_locked(src), src);
VM_BUG_ON_FOLIO(!folio_test_locked(dst), dst); VM_BUG_ON_FOLIO(!folio_test_locked(dst), dst);
mapping = folio_mapping(src);
if (likely(is_lru)) { if (likely(is_lru)) {
struct address_space *mapping = folio_mapping(src);
if (!mapping) if (!mapping)
rc = migrate_page(mapping, &dst->page, &src->page, mode); rc = migrate_page(mapping, &dst->page, &src->page, mode);
else if (mapping->a_ops->migratepage) else if (mapping->a_ops->migratepage)
@ -872,6 +870,8 @@ static int move_to_new_folio(struct folio *dst, struct folio *src,
rc = fallback_migrate_page(mapping, &dst->page, rc = fallback_migrate_page(mapping, &dst->page,
&src->page, mode); &src->page, mode);
} else { } else {
const struct movable_operations *mops;
/* /*
* In case of non-lru page, it could be released after * In case of non-lru page, it could be released after
* isolation step. In that case, we shouldn't try migration. * isolation step. In that case, we shouldn't try migration.
@ -883,8 +883,8 @@ static int move_to_new_folio(struct folio *dst, struct folio *src,
goto out; goto out;
} }
rc = mapping->a_ops->migratepage(mapping, &dst->page, mops = page_movable_ops(&src->page);
&src->page, mode); rc = mops->migrate_page(&dst->page, &src->page, mode);
WARN_ON_ONCE(rc == MIGRATEPAGE_SUCCESS && WARN_ON_ONCE(rc == MIGRATEPAGE_SUCCESS &&
!folio_test_isolated(src)); !folio_test_isolated(src));
} }

Просмотреть файл

@ -804,10 +804,10 @@ struct address_space *folio_mapping(struct folio *folio)
return swap_address_space(folio_swap_entry(folio)); return swap_address_space(folio_swap_entry(folio));
mapping = folio->mapping; mapping = folio->mapping;
if ((unsigned long)mapping & PAGE_MAPPING_ANON) if ((unsigned long)mapping & PAGE_MAPPING_FLAGS)
return NULL; return NULL;
return (void *)((unsigned long)mapping & ~PAGE_MAPPING_FLAGS); return mapping;
} }
EXPORT_SYMBOL(folio_mapping); EXPORT_SYMBOL(folio_mapping);

Просмотреть файл

@ -34,15 +34,11 @@
#include <linux/node.h> #include <linux/node.h>
#include <linux/compaction.h> #include <linux/compaction.h>
#include <linux/percpu.h> #include <linux/percpu.h>
#include <linux/mount.h>
#include <linux/pseudo_fs.h>
#include <linux/fs.h>
#include <linux/preempt.h> #include <linux/preempt.h>
#include <linux/workqueue.h> #include <linux/workqueue.h>
#include <linux/slab.h> #include <linux/slab.h>
#include <linux/spinlock.h> #include <linux/spinlock.h>
#include <linux/zpool.h> #include <linux/zpool.h>
#include <linux/magic.h>
#include <linux/kmemleak.h> #include <linux/kmemleak.h>
/* /*
@ -149,7 +145,6 @@ struct z3fold_header {
* @compact_wq: workqueue for page layout background optimization * @compact_wq: workqueue for page layout background optimization
* @release_wq: workqueue for safe page release * @release_wq: workqueue for safe page release
* @work: work_struct for safe page release * @work: work_struct for safe page release
* @inode: inode for z3fold pseudo filesystem
* *
* This structure is allocated at pool creation time and maintains metadata * This structure is allocated at pool creation time and maintains metadata
* pertaining to a particular z3fold pool. * pertaining to a particular z3fold pool.
@ -169,7 +164,6 @@ struct z3fold_pool {
struct workqueue_struct *compact_wq; struct workqueue_struct *compact_wq;
struct workqueue_struct *release_wq; struct workqueue_struct *release_wq;
struct work_struct work; struct work_struct work;
struct inode *inode;
}; };
/* /*
@ -334,54 +328,6 @@ static inline void free_handle(unsigned long handle, struct z3fold_header *zhdr)
} }
} }
static int z3fold_init_fs_context(struct fs_context *fc)
{
return init_pseudo(fc, Z3FOLD_MAGIC) ? 0 : -ENOMEM;
}
static struct file_system_type z3fold_fs = {
.name = "z3fold",
.init_fs_context = z3fold_init_fs_context,
.kill_sb = kill_anon_super,
};
static struct vfsmount *z3fold_mnt;
static int __init z3fold_mount(void)
{
int ret = 0;
z3fold_mnt = kern_mount(&z3fold_fs);
if (IS_ERR(z3fold_mnt))
ret = PTR_ERR(z3fold_mnt);
return ret;
}
static void z3fold_unmount(void)
{
kern_unmount(z3fold_mnt);
}
static const struct address_space_operations z3fold_aops;
static int z3fold_register_migration(struct z3fold_pool *pool)
{
pool->inode = alloc_anon_inode(z3fold_mnt->mnt_sb);
if (IS_ERR(pool->inode)) {
pool->inode = NULL;
return 1;
}
pool->inode->i_mapping->private_data = pool;
pool->inode->i_mapping->a_ops = &z3fold_aops;
return 0;
}
static void z3fold_unregister_migration(struct z3fold_pool *pool)
{
if (pool->inode)
iput(pool->inode);
}
/* Initializes the z3fold header of a newly allocated z3fold page */ /* Initializes the z3fold header of a newly allocated z3fold page */
static struct z3fold_header *init_z3fold_page(struct page *page, bool headless, static struct z3fold_header *init_z3fold_page(struct page *page, bool headless,
struct z3fold_pool *pool, gfp_t gfp) struct z3fold_pool *pool, gfp_t gfp)
@ -1002,14 +948,10 @@ static struct z3fold_pool *z3fold_create_pool(const char *name, gfp_t gfp,
pool->release_wq = create_singlethread_workqueue(pool->name); pool->release_wq = create_singlethread_workqueue(pool->name);
if (!pool->release_wq) if (!pool->release_wq)
goto out_wq; goto out_wq;
if (z3fold_register_migration(pool))
goto out_rwq;
INIT_WORK(&pool->work, free_pages_work); INIT_WORK(&pool->work, free_pages_work);
pool->ops = ops; pool->ops = ops;
return pool; return pool;
out_rwq:
destroy_workqueue(pool->release_wq);
out_wq: out_wq:
destroy_workqueue(pool->compact_wq); destroy_workqueue(pool->compact_wq);
out_unbuddied: out_unbuddied:
@ -1043,11 +985,12 @@ static void z3fold_destroy_pool(struct z3fold_pool *pool)
destroy_workqueue(pool->compact_wq); destroy_workqueue(pool->compact_wq);
destroy_workqueue(pool->release_wq); destroy_workqueue(pool->release_wq);
z3fold_unregister_migration(pool);
free_percpu(pool->unbuddied); free_percpu(pool->unbuddied);
kfree(pool); kfree(pool);
} }
static const struct movable_operations z3fold_mops;
/** /**
* z3fold_alloc() - allocates a region of a given size * z3fold_alloc() - allocates a region of a given size
* @pool: z3fold pool from which to allocate * @pool: z3fold pool from which to allocate
@ -1117,11 +1060,11 @@ retry:
} }
if (can_sleep) { if (can_sleep) {
lock_page(page); lock_page(page);
__SetPageMovable(page, pool->inode->i_mapping); __SetPageMovable(page, &z3fold_mops);
unlock_page(page); unlock_page(page);
} else { } else {
WARN_ON(!trylock_page(page)); WARN_ON(!trylock_page(page));
__SetPageMovable(page, pool->inode->i_mapping); __SetPageMovable(page, &z3fold_mops);
unlock_page(page); unlock_page(page);
} }
z3fold_page_lock(zhdr); z3fold_page_lock(zhdr);
@ -1554,12 +1497,11 @@ out:
return false; return false;
} }
static int z3fold_page_migrate(struct address_space *mapping, struct page *newpage, static int z3fold_page_migrate(struct page *newpage, struct page *page,
struct page *page, enum migrate_mode mode) enum migrate_mode mode)
{ {
struct z3fold_header *zhdr, *new_zhdr; struct z3fold_header *zhdr, *new_zhdr;
struct z3fold_pool *pool; struct z3fold_pool *pool;
struct address_space *new_mapping;
VM_BUG_ON_PAGE(!PageMovable(page), page); VM_BUG_ON_PAGE(!PageMovable(page), page);
VM_BUG_ON_PAGE(!PageIsolated(page), page); VM_BUG_ON_PAGE(!PageIsolated(page), page);
@ -1592,7 +1534,6 @@ static int z3fold_page_migrate(struct address_space *mapping, struct page *newpa
* so we only have to reinitialize it. * so we only have to reinitialize it.
*/ */
INIT_LIST_HEAD(&new_zhdr->buddy); INIT_LIST_HEAD(&new_zhdr->buddy);
new_mapping = page_mapping(page);
__ClearPageMovable(page); __ClearPageMovable(page);
get_page(newpage); get_page(newpage);
@ -1608,7 +1549,7 @@ static int z3fold_page_migrate(struct address_space *mapping, struct page *newpa
spin_lock(&pool->lock); spin_lock(&pool->lock);
list_add(&newpage->lru, &pool->lru); list_add(&newpage->lru, &pool->lru);
spin_unlock(&pool->lock); spin_unlock(&pool->lock);
__SetPageMovable(newpage, new_mapping); __SetPageMovable(newpage, &z3fold_mops);
z3fold_page_unlock(new_zhdr); z3fold_page_unlock(new_zhdr);
queue_work_on(new_zhdr->cpu, pool->compact_wq, &new_zhdr->work); queue_work_on(new_zhdr->cpu, pool->compact_wq, &new_zhdr->work);
@ -1642,9 +1583,9 @@ static void z3fold_page_putback(struct page *page)
z3fold_page_unlock(zhdr); z3fold_page_unlock(zhdr);
} }
static const struct address_space_operations z3fold_aops = { static const struct movable_operations z3fold_mops = {
.isolate_page = z3fold_page_isolate, .isolate_page = z3fold_page_isolate,
.migratepage = z3fold_page_migrate, .migrate_page = z3fold_page_migrate,
.putback_page = z3fold_page_putback, .putback_page = z3fold_page_putback,
}; };
@ -1746,17 +1687,11 @@ MODULE_ALIAS("zpool-z3fold");
static int __init init_z3fold(void) static int __init init_z3fold(void)
{ {
int ret;
/* /*
* Make sure the z3fold header is not larger than the page size and * Make sure the z3fold header is not larger than the page size and
* there has remaining spaces for its buddy. * there has remaining spaces for its buddy.
*/ */
BUILD_BUG_ON(ZHDR_SIZE_ALIGNED > PAGE_SIZE - CHUNK_SIZE); BUILD_BUG_ON(ZHDR_SIZE_ALIGNED > PAGE_SIZE - CHUNK_SIZE);
ret = z3fold_mount();
if (ret)
return ret;
zpool_register_driver(&z3fold_zpool_driver); zpool_register_driver(&z3fold_zpool_driver);
return 0; return 0;
@ -1764,7 +1699,6 @@ static int __init init_z3fold(void)
static void __exit exit_z3fold(void) static void __exit exit_z3fold(void)
{ {
z3fold_unmount();
zpool_unregister_driver(&z3fold_zpool_driver); zpool_unregister_driver(&z3fold_zpool_driver);
} }

Просмотреть файл

@ -41,7 +41,6 @@
#include <linux/module.h> #include <linux/module.h>
#include <linux/kernel.h> #include <linux/kernel.h>
#include <linux/sched.h> #include <linux/sched.h>
#include <linux/magic.h>
#include <linux/bitops.h> #include <linux/bitops.h>
#include <linux/errno.h> #include <linux/errno.h>
#include <linux/highmem.h> #include <linux/highmem.h>
@ -59,8 +58,6 @@
#include <linux/debugfs.h> #include <linux/debugfs.h>
#include <linux/zsmalloc.h> #include <linux/zsmalloc.h>
#include <linux/zpool.h> #include <linux/zpool.h>
#include <linux/mount.h>
#include <linux/pseudo_fs.h>
#include <linux/migrate.h> #include <linux/migrate.h>
#include <linux/wait.h> #include <linux/wait.h>
#include <linux/pagemap.h> #include <linux/pagemap.h>
@ -177,10 +174,6 @@ struct zs_size_stat {
static struct dentry *zs_stat_root; static struct dentry *zs_stat_root;
#endif #endif
#ifdef CONFIG_COMPACTION
static struct vfsmount *zsmalloc_mnt;
#endif
/* /*
* We assign a page to ZS_ALMOST_EMPTY fullness group when: * We assign a page to ZS_ALMOST_EMPTY fullness group when:
* n <= N / f, where * n <= N / f, where
@ -252,7 +245,6 @@ struct zs_pool {
struct dentry *stat_dentry; struct dentry *stat_dentry;
#endif #endif
#ifdef CONFIG_COMPACTION #ifdef CONFIG_COMPACTION
struct inode *inode;
struct work_struct free_work; struct work_struct free_work;
#endif #endif
/* protect page/zspage migration */ /* protect page/zspage migration */
@ -271,6 +263,7 @@ struct zspage {
unsigned int freeobj; unsigned int freeobj;
struct page *first_page; struct page *first_page;
struct list_head list; /* fullness list */ struct list_head list; /* fullness list */
struct zs_pool *pool;
#ifdef CONFIG_COMPACTION #ifdef CONFIG_COMPACTION
rwlock_t lock; rwlock_t lock;
#endif #endif
@ -295,8 +288,6 @@ static bool ZsHugePage(struct zspage *zspage)
} }
#ifdef CONFIG_COMPACTION #ifdef CONFIG_COMPACTION
static int zs_register_migration(struct zs_pool *pool);
static void zs_unregister_migration(struct zs_pool *pool);
static void migrate_lock_init(struct zspage *zspage); static void migrate_lock_init(struct zspage *zspage);
static void migrate_read_lock(struct zspage *zspage); static void migrate_read_lock(struct zspage *zspage);
static void migrate_read_unlock(struct zspage *zspage); static void migrate_read_unlock(struct zspage *zspage);
@ -307,10 +298,6 @@ static void kick_deferred_free(struct zs_pool *pool);
static void init_deferred_free(struct zs_pool *pool); static void init_deferred_free(struct zs_pool *pool);
static void SetZsPageMovable(struct zs_pool *pool, struct zspage *zspage); static void SetZsPageMovable(struct zs_pool *pool, struct zspage *zspage);
#else #else
static int zsmalloc_mount(void) { return 0; }
static void zsmalloc_unmount(void) {}
static int zs_register_migration(struct zs_pool *pool) { return 0; }
static void zs_unregister_migration(struct zs_pool *pool) {}
static void migrate_lock_init(struct zspage *zspage) {} static void migrate_lock_init(struct zspage *zspage) {}
static void migrate_read_lock(struct zspage *zspage) {} static void migrate_read_lock(struct zspage *zspage) {}
static void migrate_read_unlock(struct zspage *zspage) {} static void migrate_read_unlock(struct zspage *zspage) {}
@ -1083,6 +1070,7 @@ static struct zspage *alloc_zspage(struct zs_pool *pool,
create_page_chain(class, zspage, pages); create_page_chain(class, zspage, pages);
init_zspage(class, zspage); init_zspage(class, zspage);
zspage->pool = pool;
return zspage; return zspage;
} }
@ -1754,33 +1742,6 @@ static void lock_zspage(struct zspage *zspage)
migrate_read_unlock(zspage); migrate_read_unlock(zspage);
} }
static int zs_init_fs_context(struct fs_context *fc)
{
return init_pseudo(fc, ZSMALLOC_MAGIC) ? 0 : -ENOMEM;
}
static struct file_system_type zsmalloc_fs = {
.name = "zsmalloc",
.init_fs_context = zs_init_fs_context,
.kill_sb = kill_anon_super,
};
static int zsmalloc_mount(void)
{
int ret = 0;
zsmalloc_mnt = kern_mount(&zsmalloc_fs);
if (IS_ERR(zsmalloc_mnt))
ret = PTR_ERR(zsmalloc_mnt);
return ret;
}
static void zsmalloc_unmount(void)
{
kern_unmount(zsmalloc_mnt);
}
static void migrate_lock_init(struct zspage *zspage) static void migrate_lock_init(struct zspage *zspage)
{ {
rwlock_init(&zspage->lock); rwlock_init(&zspage->lock);
@ -1823,6 +1784,8 @@ static void dec_zspage_isolation(struct zspage *zspage)
zspage->isolated--; zspage->isolated--;
} }
static const struct movable_operations zsmalloc_mops;
static void replace_sub_page(struct size_class *class, struct zspage *zspage, static void replace_sub_page(struct size_class *class, struct zspage *zspage,
struct page *newpage, struct page *oldpage) struct page *newpage, struct page *oldpage)
{ {
@ -1843,7 +1806,7 @@ static void replace_sub_page(struct size_class *class, struct zspage *zspage,
set_first_obj_offset(newpage, get_first_obj_offset(oldpage)); set_first_obj_offset(newpage, get_first_obj_offset(oldpage));
if (unlikely(ZsHugePage(zspage))) if (unlikely(ZsHugePage(zspage)))
newpage->index = oldpage->index; newpage->index = oldpage->index;
__SetPageMovable(newpage, page_mapping(oldpage)); __SetPageMovable(newpage, &zsmalloc_mops);
} }
static bool zs_page_isolate(struct page *page, isolate_mode_t mode) static bool zs_page_isolate(struct page *page, isolate_mode_t mode)
@ -1865,8 +1828,8 @@ static bool zs_page_isolate(struct page *page, isolate_mode_t mode)
return true; return true;
} }
static int zs_page_migrate(struct address_space *mapping, struct page *newpage, static int zs_page_migrate(struct page *newpage, struct page *page,
struct page *page, enum migrate_mode mode) enum migrate_mode mode)
{ {
struct zs_pool *pool; struct zs_pool *pool;
struct size_class *class; struct size_class *class;
@ -1889,14 +1852,15 @@ static int zs_page_migrate(struct address_space *mapping, struct page *newpage,
VM_BUG_ON_PAGE(!PageMovable(page), page); VM_BUG_ON_PAGE(!PageMovable(page), page);
VM_BUG_ON_PAGE(!PageIsolated(page), page); VM_BUG_ON_PAGE(!PageIsolated(page), page);
pool = mapping->private_data; /* The page is locked, so this pointer must remain valid */
zspage = get_zspage(page);
pool = zspage->pool;
/* /*
* The pool migrate_lock protects the race between zpage migration * The pool migrate_lock protects the race between zpage migration
* and zs_free. * and zs_free.
*/ */
write_lock(&pool->migrate_lock); write_lock(&pool->migrate_lock);
zspage = get_zspage(page);
class = zspage_class(pool, zspage); class = zspage_class(pool, zspage);
/* /*
@ -1964,31 +1928,12 @@ static void zs_page_putback(struct page *page)
migrate_write_unlock(zspage); migrate_write_unlock(zspage);
} }
static const struct address_space_operations zsmalloc_aops = { static const struct movable_operations zsmalloc_mops = {
.isolate_page = zs_page_isolate, .isolate_page = zs_page_isolate,
.migratepage = zs_page_migrate, .migrate_page = zs_page_migrate,
.putback_page = zs_page_putback, .putback_page = zs_page_putback,
}; };
static int zs_register_migration(struct zs_pool *pool)
{
pool->inode = alloc_anon_inode(zsmalloc_mnt->mnt_sb);
if (IS_ERR(pool->inode)) {
pool->inode = NULL;
return 1;
}
pool->inode->i_mapping->private_data = pool;
pool->inode->i_mapping->a_ops = &zsmalloc_aops;
return 0;
}
static void zs_unregister_migration(struct zs_pool *pool)
{
flush_work(&pool->free_work);
iput(pool->inode);
}
/* /*
* Caller should hold page_lock of all pages in the zspage * Caller should hold page_lock of all pages in the zspage
* In here, we cannot use zspage meta data. * In here, we cannot use zspage meta data.
@ -2032,6 +1977,11 @@ static void kick_deferred_free(struct zs_pool *pool)
schedule_work(&pool->free_work); schedule_work(&pool->free_work);
} }
static void zs_flush_migration(struct zs_pool *pool)
{
flush_work(&pool->free_work);
}
static void init_deferred_free(struct zs_pool *pool) static void init_deferred_free(struct zs_pool *pool)
{ {
INIT_WORK(&pool->free_work, async_free_zspage); INIT_WORK(&pool->free_work, async_free_zspage);
@ -2043,10 +1993,12 @@ static void SetZsPageMovable(struct zs_pool *pool, struct zspage *zspage)
do { do {
WARN_ON(!trylock_page(page)); WARN_ON(!trylock_page(page));
__SetPageMovable(page, pool->inode->i_mapping); __SetPageMovable(page, &zsmalloc_mops);
unlock_page(page); unlock_page(page);
} while ((page = get_next_page(page)) != NULL); } while ((page = get_next_page(page)) != NULL);
} }
#else
static inline void zs_flush_migration(struct zs_pool *pool) { }
#endif #endif
/* /*
@ -2324,9 +2276,6 @@ struct zs_pool *zs_create_pool(const char *name)
/* debug only, don't abort if it fails */ /* debug only, don't abort if it fails */
zs_pool_stat_create(pool, name); zs_pool_stat_create(pool, name);
if (zs_register_migration(pool))
goto err;
/* /*
* Not critical since shrinker is only used to trigger internal * Not critical since shrinker is only used to trigger internal
* defragmentation of the pool which is pretty optional thing. If * defragmentation of the pool which is pretty optional thing. If
@ -2348,7 +2297,7 @@ void zs_destroy_pool(struct zs_pool *pool)
int i; int i;
zs_unregister_shrinker(pool); zs_unregister_shrinker(pool);
zs_unregister_migration(pool); zs_flush_migration(pool);
zs_pool_stat_destroy(pool); zs_pool_stat_destroy(pool);
for (i = 0; i < ZS_SIZE_CLASSES; i++) { for (i = 0; i < ZS_SIZE_CLASSES; i++) {
@ -2380,14 +2329,10 @@ static int __init zs_init(void)
{ {
int ret; int ret;
ret = zsmalloc_mount();
if (ret)
goto out;
ret = cpuhp_setup_state(CPUHP_MM_ZS_PREPARE, "mm/zsmalloc:prepare", ret = cpuhp_setup_state(CPUHP_MM_ZS_PREPARE, "mm/zsmalloc:prepare",
zs_cpu_prepare, zs_cpu_dead); zs_cpu_prepare, zs_cpu_dead);
if (ret) if (ret)
goto hp_setup_fail; goto out;
#ifdef CONFIG_ZPOOL #ifdef CONFIG_ZPOOL
zpool_register_driver(&zs_zpool_driver); zpool_register_driver(&zs_zpool_driver);
@ -2397,8 +2342,6 @@ static int __init zs_init(void)
return 0; return 0;
hp_setup_fail:
zsmalloc_unmount();
out: out:
return ret; return ret;
} }
@ -2408,7 +2351,6 @@ static void __exit zs_exit(void)
#ifdef CONFIG_ZPOOL #ifdef CONFIG_ZPOOL
zpool_unregister_driver(&zs_zpool_driver); zpool_unregister_driver(&zs_zpool_driver);
#endif #endif
zsmalloc_unmount();
cpuhp_remove_state(CPUHP_MM_ZS_PREPARE); cpuhp_remove_state(CPUHP_MM_ZS_PREPARE);
zs_stat_exit(); zs_stat_exit();