2016-05-18 23:15:08 +07:00
|
|
|
/*
|
|
|
|
* Copyright(c) 2016 Intel Corporation. All rights reserved.
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of version 2 of the GNU General Public License as
|
|
|
|
* published by the Free Software Foundation.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful, but
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* General Public License for more details.
|
|
|
|
*/
|
|
|
|
#include <linux/pagemap.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/device.h>
|
2016-07-25 11:55:45 +07:00
|
|
|
#include <linux/mount.h>
|
2016-05-18 23:15:08 +07:00
|
|
|
#include <linux/pfn_t.h>
|
2016-07-25 11:55:45 +07:00
|
|
|
#include <linux/hash.h>
|
2016-07-25 05:55:42 +07:00
|
|
|
#include <linux/cdev.h>
|
2016-05-18 23:15:08 +07:00
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/dax.h>
|
|
|
|
#include <linux/fs.h>
|
|
|
|
#include <linux/mm.h>
|
2016-08-07 06:05:06 +07:00
|
|
|
#include "dax.h"
|
2016-05-18 23:15:08 +07:00
|
|
|
|
2016-07-25 05:55:42 +07:00
|
|
|
static dev_t dax_devt;
|
2016-05-18 23:15:08 +07:00
|
|
|
static struct class *dax_class;
|
|
|
|
static DEFINE_IDA(dax_minor_ida);
|
2016-07-25 05:55:42 +07:00
|
|
|
static int nr_dax = CONFIG_NR_DEV_DAX;
|
|
|
|
module_param(nr_dax, int, S_IRUGO);
|
2016-07-25 11:55:45 +07:00
|
|
|
static struct vfsmount *dax_mnt;
|
|
|
|
static struct kmem_cache *dax_cache __read_mostly;
|
|
|
|
static struct super_block *dax_superblock __read_mostly;
|
2016-07-25 05:55:42 +07:00
|
|
|
MODULE_PARM_DESC(nr_dax, "max number of device-dax instances");
|
2016-05-18 23:15:08 +07:00
|
|
|
|
|
|
|
/**
|
|
|
|
* struct dax_region - mapping infrastructure for dax devices
|
|
|
|
* @id: kernel-wide unique region for a memory range
|
|
|
|
* @base: linear address corresponding to @res
|
|
|
|
* @kref: to pin while other agents have a need to do lookups
|
|
|
|
* @dev: parent device backing this region
|
|
|
|
* @align: allocation and mapping alignment for child dax devices
|
|
|
|
* @res: physical address range of the region
|
|
|
|
* @pfn_flags: identify whether the pfns are paged back or not
|
|
|
|
*/
|
|
|
|
struct dax_region {
|
|
|
|
int id;
|
|
|
|
struct ida ida;
|
|
|
|
void *base;
|
|
|
|
struct kref kref;
|
|
|
|
struct device *dev;
|
|
|
|
unsigned int align;
|
|
|
|
struct resource res;
|
|
|
|
unsigned long pfn_flags;
|
|
|
|
};
|
|
|
|
|
|
|
|
/**
|
|
|
|
* struct dax_dev - subdivision of a dax region
|
|
|
|
* @region - parent region
|
|
|
|
* @dev - device backing the character device
|
2016-07-25 05:55:42 +07:00
|
|
|
* @cdev - core chardev data
|
2016-05-15 02:20:44 +07:00
|
|
|
* @alive - !alive + rcu grace period == no new mappings can be established
|
2016-05-18 23:15:08 +07:00
|
|
|
* @id - child id in the region
|
|
|
|
* @num_resources - number of physical address extents in this device
|
|
|
|
* @res - array of physical address ranges
|
|
|
|
*/
|
|
|
|
struct dax_dev {
|
|
|
|
struct dax_region *region;
|
2016-07-25 11:55:45 +07:00
|
|
|
struct inode *inode;
|
2016-08-11 14:41:51 +07:00
|
|
|
struct device dev;
|
2016-07-25 05:55:42 +07:00
|
|
|
struct cdev cdev;
|
2016-05-15 02:20:44 +07:00
|
|
|
bool alive;
|
2016-05-18 23:15:08 +07:00
|
|
|
int id;
|
|
|
|
int num_resources;
|
|
|
|
struct resource res[0];
|
|
|
|
};
|
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
static struct inode *dax_alloc_inode(struct super_block *sb)
|
2016-05-18 23:15:08 +07:00
|
|
|
{
|
2016-07-25 11:55:45 +07:00
|
|
|
return kmem_cache_alloc(dax_cache, GFP_KERNEL);
|
|
|
|
}
|
2016-05-18 23:15:08 +07:00
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
static void dax_i_callback(struct rcu_head *head)
|
|
|
|
{
|
|
|
|
struct inode *inode = container_of(head, struct inode, i_rcu);
|
|
|
|
|
|
|
|
kmem_cache_free(dax_cache, inode);
|
2016-05-18 23:15:08 +07:00
|
|
|
}
|
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
static void dax_destroy_inode(struct inode *inode)
|
2016-05-18 23:15:08 +07:00
|
|
|
{
|
2016-07-25 11:55:45 +07:00
|
|
|
call_rcu(&inode->i_rcu, dax_i_callback);
|
2016-05-18 23:15:08 +07:00
|
|
|
}
|
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
static const struct super_operations dax_sops = {
|
|
|
|
.statfs = simple_statfs,
|
|
|
|
.alloc_inode = dax_alloc_inode,
|
|
|
|
.destroy_inode = dax_destroy_inode,
|
|
|
|
.drop_inode = generic_delete_inode,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct dentry *dax_mount(struct file_system_type *fs_type,
|
|
|
|
int flags, const char *dev_name, void *data)
|
2016-05-18 23:15:08 +07:00
|
|
|
{
|
2016-07-25 11:55:45 +07:00
|
|
|
return mount_pseudo(fs_type, "dax:", &dax_sops, NULL, DAXFS_MAGIC);
|
|
|
|
}
|
2016-05-18 23:15:08 +07:00
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
static struct file_system_type dax_type = {
|
|
|
|
.name = "dax",
|
|
|
|
.mount = dax_mount,
|
|
|
|
.kill_sb = kill_anon_super,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int dax_test(struct inode *inode, void *data)
|
|
|
|
{
|
|
|
|
return inode->i_cdev == data;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dax_set(struct inode *inode, void *data)
|
|
|
|
{
|
|
|
|
inode->i_cdev = data;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct inode *dax_inode_get(struct cdev *cdev, dev_t devt)
|
|
|
|
{
|
|
|
|
struct inode *inode;
|
|
|
|
|
|
|
|
inode = iget5_locked(dax_superblock, hash_32(devt + DAXFS_MAGIC, 31),
|
|
|
|
dax_test, dax_set, cdev);
|
|
|
|
|
|
|
|
if (!inode)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
if (inode->i_state & I_NEW) {
|
|
|
|
inode->i_mode = S_IFCHR;
|
|
|
|
inode->i_flags = S_DAX;
|
|
|
|
inode->i_rdev = devt;
|
|
|
|
mapping_set_gfp_mask(&inode->i_data, GFP_USER);
|
|
|
|
unlock_new_inode(inode);
|
|
|
|
}
|
|
|
|
return inode;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void init_once(void *inode)
|
|
|
|
{
|
|
|
|
inode_init_once(inode);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dax_inode_init(void)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
dax_cache = kmem_cache_create("dax_cache", sizeof(struct inode), 0,
|
|
|
|
(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
|
|
|
|
SLAB_MEM_SPREAD|SLAB_ACCOUNT),
|
|
|
|
init_once);
|
|
|
|
if (!dax_cache)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
rc = register_filesystem(&dax_type);
|
|
|
|
if (rc)
|
|
|
|
goto err_register_fs;
|
|
|
|
|
|
|
|
dax_mnt = kern_mount(&dax_type);
|
|
|
|
if (IS_ERR(dax_mnt)) {
|
|
|
|
rc = PTR_ERR(dax_mnt);
|
|
|
|
goto err_mount;
|
|
|
|
}
|
|
|
|
dax_superblock = dax_mnt->mnt_sb;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_mount:
|
|
|
|
unregister_filesystem(&dax_type);
|
|
|
|
err_register_fs:
|
|
|
|
kmem_cache_destroy(dax_cache);
|
|
|
|
|
|
|
|
return rc;
|
2016-05-18 23:15:08 +07:00
|
|
|
}
|
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
static void dax_inode_exit(void)
|
|
|
|
{
|
|
|
|
kern_unmount(dax_mnt);
|
|
|
|
unregister_filesystem(&dax_type);
|
|
|
|
kmem_cache_destroy(dax_cache);
|
|
|
|
}
|
|
|
|
|
2016-05-18 23:15:08 +07:00
|
|
|
static void dax_region_free(struct kref *kref)
|
|
|
|
{
|
|
|
|
struct dax_region *dax_region;
|
|
|
|
|
|
|
|
dax_region = container_of(kref, struct dax_region, kref);
|
|
|
|
kfree(dax_region);
|
|
|
|
}
|
|
|
|
|
|
|
|
void dax_region_put(struct dax_region *dax_region)
|
|
|
|
{
|
|
|
|
kref_put(&dax_region->kref, dax_region_free);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(dax_region_put);
|
|
|
|
|
|
|
|
struct dax_region *alloc_dax_region(struct device *parent, int region_id,
|
|
|
|
struct resource *res, unsigned int align, void *addr,
|
|
|
|
unsigned long pfn_flags)
|
|
|
|
{
|
|
|
|
struct dax_region *dax_region;
|
|
|
|
|
2016-07-20 06:17:58 +07:00
|
|
|
if (!IS_ALIGNED(res->start, align)
|
|
|
|
|| !IS_ALIGNED(resource_size(res), align))
|
|
|
|
return NULL;
|
2016-05-18 23:15:08 +07:00
|
|
|
|
2016-07-20 06:17:58 +07:00
|
|
|
dax_region = kzalloc(sizeof(*dax_region), GFP_KERNEL);
|
2016-05-18 23:15:08 +07:00
|
|
|
if (!dax_region)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
memcpy(&dax_region->res, res, sizeof(*res));
|
|
|
|
dax_region->pfn_flags = pfn_flags;
|
|
|
|
kref_init(&dax_region->kref);
|
|
|
|
dax_region->id = region_id;
|
|
|
|
ida_init(&dax_region->ida);
|
|
|
|
dax_region->align = align;
|
|
|
|
dax_region->dev = parent;
|
|
|
|
dax_region->base = addr;
|
|
|
|
|
|
|
|
return dax_region;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(alloc_dax_region);
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
static struct dax_dev *to_dax_dev(struct device *dev)
|
|
|
|
{
|
|
|
|
return container_of(dev, struct dax_dev, dev);
|
|
|
|
}
|
|
|
|
|
2016-05-18 23:15:08 +07:00
|
|
|
static ssize_t size_show(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
{
|
2016-08-11 14:41:51 +07:00
|
|
|
struct dax_dev *dax_dev = to_dax_dev(dev);
|
2016-05-18 23:15:08 +07:00
|
|
|
unsigned long long size = 0;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < dax_dev->num_resources; i++)
|
|
|
|
size += resource_size(&dax_dev->res[i]);
|
|
|
|
|
|
|
|
return sprintf(buf, "%llu\n", size);
|
|
|
|
}
|
|
|
|
static DEVICE_ATTR_RO(size);
|
|
|
|
|
|
|
|
static struct attribute *dax_device_attributes[] = {
|
|
|
|
&dev_attr_size.attr,
|
|
|
|
NULL,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group dax_device_attribute_group = {
|
|
|
|
.attrs = dax_device_attributes,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group *dax_attribute_groups[] = {
|
|
|
|
&dax_device_attribute_group,
|
|
|
|
NULL,
|
|
|
|
};
|
|
|
|
|
2016-05-15 02:20:44 +07:00
|
|
|
static int check_vma(struct dax_dev *dax_dev, struct vm_area_struct *vma,
|
|
|
|
const char *func)
|
|
|
|
{
|
|
|
|
struct dax_region *dax_region = dax_dev->region;
|
2016-08-11 14:41:51 +07:00
|
|
|
struct device *dev = &dax_dev->dev;
|
2016-05-15 02:20:44 +07:00
|
|
|
unsigned long mask;
|
|
|
|
|
|
|
|
if (!dax_dev->alive)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2016-11-17 00:00:38 +07:00
|
|
|
/* prevent private mappings from being established */
|
2016-12-07 08:03:35 +07:00
|
|
|
if ((vma->vm_flags & VM_MAYSHARE) != VM_MAYSHARE) {
|
2016-05-15 02:20:44 +07:00
|
|
|
dev_info(dev, "%s: %s: fail, attempted private mapping\n",
|
|
|
|
current->comm, func);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
mask = dax_region->align - 1;
|
|
|
|
if (vma->vm_start & mask || vma->vm_end & mask) {
|
|
|
|
dev_info(dev, "%s: %s: fail, unaligned vma (%#lx - %#lx, %#lx)\n",
|
|
|
|
current->comm, func, vma->vm_start, vma->vm_end,
|
|
|
|
mask);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if ((dax_region->pfn_flags & (PFN_DEV|PFN_MAP)) == PFN_DEV
|
|
|
|
&& (vma->vm_flags & VM_DONTCOPY) == 0) {
|
|
|
|
dev_info(dev, "%s: %s: fail, dax range requires MADV_DONTFORK\n",
|
|
|
|
current->comm, func);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!vma_is_dax(vma)) {
|
|
|
|
dev_info(dev, "%s: %s: fail, vma is not DAX capable\n",
|
|
|
|
current->comm, func);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static phys_addr_t pgoff_to_phys(struct dax_dev *dax_dev, pgoff_t pgoff,
|
|
|
|
unsigned long size)
|
|
|
|
{
|
|
|
|
struct resource *res;
|
|
|
|
phys_addr_t phys;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < dax_dev->num_resources; i++) {
|
|
|
|
res = &dax_dev->res[i];
|
|
|
|
phys = pgoff * PAGE_SIZE + res->start;
|
|
|
|
if (phys >= res->start && phys <= res->end)
|
|
|
|
break;
|
|
|
|
pgoff -= PHYS_PFN(resource_size(res));
|
|
|
|
}
|
|
|
|
|
|
|
|
if (i < dax_dev->num_resources) {
|
|
|
|
res = &dax_dev->res[i];
|
|
|
|
if (phys + size - 1 <= res->end)
|
|
|
|
return phys;
|
|
|
|
}
|
|
|
|
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __dax_dev_fault(struct dax_dev *dax_dev, struct vm_area_struct *vma,
|
|
|
|
struct vm_fault *vmf)
|
|
|
|
{
|
2016-08-11 14:41:51 +07:00
|
|
|
struct device *dev = &dax_dev->dev;
|
2016-05-15 02:20:44 +07:00
|
|
|
struct dax_region *dax_region;
|
|
|
|
int rc = VM_FAULT_SIGBUS;
|
|
|
|
phys_addr_t phys;
|
|
|
|
pfn_t pfn;
|
|
|
|
|
|
|
|
if (check_vma(dax_dev, vma, __func__))
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
dax_region = dax_dev->region;
|
|
|
|
if (dax_region->align > PAGE_SIZE) {
|
|
|
|
dev_dbg(dev, "%s: alignment > fault size\n", __func__);
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
}
|
|
|
|
|
|
|
|
phys = pgoff_to_phys(dax_dev, vmf->pgoff, PAGE_SIZE);
|
|
|
|
if (phys == -1) {
|
|
|
|
dev_dbg(dev, "%s: phys_to_pgoff(%#lx) failed\n", __func__,
|
|
|
|
vmf->pgoff);
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
}
|
|
|
|
|
|
|
|
pfn = phys_to_pfn_t(phys, dax_region->pfn_flags);
|
|
|
|
|
2016-12-15 06:07:01 +07:00
|
|
|
rc = vm_insert_mixed(vma, vmf->address, pfn);
|
2016-05-15 02:20:44 +07:00
|
|
|
|
|
|
|
if (rc == -ENOMEM)
|
|
|
|
return VM_FAULT_OOM;
|
|
|
|
if (rc < 0 && rc != -EBUSY)
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
return VM_FAULT_NOPAGE;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dax_dev_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
struct file *filp = vma->vm_file;
|
|
|
|
struct dax_dev *dax_dev = filp->private_data;
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
dev_dbg(&dax_dev->dev, "%s: %s: %s (%#lx - %#lx)\n", __func__,
|
2016-05-15 02:20:44 +07:00
|
|
|
current->comm, (vmf->flags & FAULT_FLAG_WRITE)
|
|
|
|
? "write" : "read", vma->vm_start, vma->vm_end);
|
|
|
|
rcu_read_lock();
|
|
|
|
rc = __dax_dev_fault(dax_dev, vma, vmf);
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __dax_dev_pmd_fault(struct dax_dev *dax_dev,
|
|
|
|
struct vm_area_struct *vma, unsigned long addr, pmd_t *pmd,
|
|
|
|
unsigned int flags)
|
|
|
|
{
|
|
|
|
unsigned long pmd_addr = addr & PMD_MASK;
|
2016-08-11 14:41:51 +07:00
|
|
|
struct device *dev = &dax_dev->dev;
|
2016-05-15 02:20:44 +07:00
|
|
|
struct dax_region *dax_region;
|
|
|
|
phys_addr_t phys;
|
|
|
|
pgoff_t pgoff;
|
|
|
|
pfn_t pfn;
|
|
|
|
|
|
|
|
if (check_vma(dax_dev, vma, __func__))
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
dax_region = dax_dev->region;
|
|
|
|
if (dax_region->align > PMD_SIZE) {
|
|
|
|
dev_dbg(dev, "%s: alignment > fault size\n", __func__);
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* dax pmd mappings require pfn_t_devmap() */
|
|
|
|
if ((dax_region->pfn_flags & (PFN_DEV|PFN_MAP)) != (PFN_DEV|PFN_MAP)) {
|
|
|
|
dev_dbg(dev, "%s: alignment > fault size\n", __func__);
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
}
|
|
|
|
|
|
|
|
pgoff = linear_page_index(vma, pmd_addr);
|
2016-09-04 00:36:00 +07:00
|
|
|
phys = pgoff_to_phys(dax_dev, pgoff, PMD_SIZE);
|
2016-05-15 02:20:44 +07:00
|
|
|
if (phys == -1) {
|
|
|
|
dev_dbg(dev, "%s: phys_to_pgoff(%#lx) failed\n", __func__,
|
|
|
|
pgoff);
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
}
|
|
|
|
|
|
|
|
pfn = phys_to_pfn_t(phys, dax_region->pfn_flags);
|
|
|
|
|
|
|
|
return vmf_insert_pfn_pmd(vma, addr, pmd, pfn,
|
|
|
|
flags & FAULT_FLAG_WRITE);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dax_dev_pmd_fault(struct vm_area_struct *vma, unsigned long addr,
|
|
|
|
pmd_t *pmd, unsigned int flags)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
struct file *filp = vma->vm_file;
|
|
|
|
struct dax_dev *dax_dev = filp->private_data;
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
dev_dbg(&dax_dev->dev, "%s: %s: %s (%#lx - %#lx)\n", __func__,
|
2016-05-15 02:20:44 +07:00
|
|
|
current->comm, (flags & FAULT_FLAG_WRITE)
|
|
|
|
? "write" : "read", vma->vm_start, vma->vm_end);
|
|
|
|
|
|
|
|
rcu_read_lock();
|
|
|
|
rc = __dax_dev_pmd_fault(dax_dev, vma, addr, pmd, flags);
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct vm_operations_struct dax_dev_vm_ops = {
|
|
|
|
.fault = dax_dev_fault,
|
|
|
|
.pmd_fault = dax_dev_pmd_fault,
|
|
|
|
};
|
|
|
|
|
2016-08-11 14:38:03 +07:00
|
|
|
static int dax_mmap(struct file *filp, struct vm_area_struct *vma)
|
2016-05-15 02:20:44 +07:00
|
|
|
{
|
|
|
|
struct dax_dev *dax_dev = filp->private_data;
|
|
|
|
int rc;
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
dev_dbg(&dax_dev->dev, "%s\n", __func__);
|
2016-05-15 02:20:44 +07:00
|
|
|
|
|
|
|
rc = check_vma(dax_dev, vma, __func__);
|
|
|
|
if (rc)
|
|
|
|
return rc;
|
|
|
|
|
|
|
|
vma->vm_ops = &dax_dev_vm_ops;
|
|
|
|
vma->vm_flags |= VM_MIXEDMAP | VM_HUGEPAGE;
|
|
|
|
return 0;
|
2016-08-07 22:23:56 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* return an unmapped area aligned to the dax region specified alignment */
|
2016-08-11 14:38:03 +07:00
|
|
|
static unsigned long dax_get_unmapped_area(struct file *filp,
|
2016-08-07 22:23:56 +07:00
|
|
|
unsigned long addr, unsigned long len, unsigned long pgoff,
|
|
|
|
unsigned long flags)
|
|
|
|
{
|
|
|
|
unsigned long off, off_end, off_align, len_align, addr_align, align;
|
|
|
|
struct dax_dev *dax_dev = filp ? filp->private_data : NULL;
|
|
|
|
struct dax_region *dax_region;
|
|
|
|
|
|
|
|
if (!dax_dev || addr)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
dax_region = dax_dev->region;
|
|
|
|
align = dax_region->align;
|
|
|
|
off = pgoff << PAGE_SHIFT;
|
|
|
|
off_end = off + len;
|
|
|
|
off_align = round_up(off, align);
|
|
|
|
|
|
|
|
if ((off_end <= off_align) || ((off_end - off_align) < align))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
len_align = len + align;
|
|
|
|
if ((off + len_align) < off)
|
|
|
|
goto out;
|
2016-05-15 02:20:44 +07:00
|
|
|
|
2016-08-07 22:23:56 +07:00
|
|
|
addr_align = current->mm->get_unmapped_area(filp, addr, len_align,
|
|
|
|
pgoff, flags);
|
|
|
|
if (!IS_ERR_VALUE(addr_align)) {
|
|
|
|
addr_align += (off - addr_align) & (align - 1);
|
|
|
|
return addr_align;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
return current->mm->get_unmapped_area(filp, addr, len, pgoff, flags);
|
|
|
|
}
|
|
|
|
|
2016-08-11 14:38:03 +07:00
|
|
|
static int dax_open(struct inode *inode, struct file *filp)
|
2016-08-07 22:23:56 +07:00
|
|
|
{
|
2016-07-25 05:55:42 +07:00
|
|
|
struct dax_dev *dax_dev;
|
2016-08-07 22:23:56 +07:00
|
|
|
|
2016-07-25 05:55:42 +07:00
|
|
|
dax_dev = container_of(inode->i_cdev, struct dax_dev, cdev);
|
|
|
|
dev_dbg(&dax_dev->dev, "%s\n", __func__);
|
2016-07-25 11:55:45 +07:00
|
|
|
inode->i_mapping = dax_dev->inode->i_mapping;
|
|
|
|
inode->i_mapping->host = dax_dev->inode;
|
|
|
|
filp->f_mapping = inode->i_mapping;
|
2016-08-11 14:41:51 +07:00
|
|
|
filp->private_data = dax_dev;
|
|
|
|
inode->i_flags = S_DAX;
|
2016-08-07 22:23:56 +07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2016-05-15 02:20:44 +07:00
|
|
|
|
2016-08-11 14:38:03 +07:00
|
|
|
static int dax_release(struct inode *inode, struct file *filp)
|
2016-08-07 22:23:56 +07:00
|
|
|
{
|
|
|
|
struct dax_dev *dax_dev = filp->private_data;
|
|
|
|
|
2016-07-25 05:55:42 +07:00
|
|
|
dev_dbg(&dax_dev->dev, "%s\n", __func__);
|
2016-08-07 22:23:56 +07:00
|
|
|
return 0;
|
2016-05-15 02:20:44 +07:00
|
|
|
}
|
|
|
|
|
2016-05-18 23:15:08 +07:00
|
|
|
static const struct file_operations dax_fops = {
|
|
|
|
.llseek = noop_llseek,
|
|
|
|
.owner = THIS_MODULE,
|
2016-08-11 14:38:03 +07:00
|
|
|
.open = dax_open,
|
|
|
|
.release = dax_release,
|
|
|
|
.get_unmapped_area = dax_get_unmapped_area,
|
|
|
|
.mmap = dax_mmap,
|
2016-05-18 23:15:08 +07:00
|
|
|
};
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
static void dax_dev_release(struct device *dev)
|
2016-08-07 22:23:56 +07:00
|
|
|
{
|
2016-08-11 14:41:51 +07:00
|
|
|
struct dax_dev *dax_dev = to_dax_dev(dev);
|
2016-08-07 22:23:56 +07:00
|
|
|
struct dax_region *dax_region = dax_dev->region;
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
ida_simple_remove(&dax_region->ida, dax_dev->id);
|
|
|
|
ida_simple_remove(&dax_minor_ida, MINOR(dev->devt));
|
|
|
|
dax_region_put(dax_region);
|
2016-07-25 11:55:45 +07:00
|
|
|
iput(dax_dev->inode);
|
2016-08-11 14:41:51 +07:00
|
|
|
kfree(dax_dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void unregister_dax_dev(void *dev)
|
|
|
|
{
|
|
|
|
struct dax_dev *dax_dev = to_dax_dev(dev);
|
2016-07-25 05:55:42 +07:00
|
|
|
struct cdev *cdev = &dax_dev->cdev;
|
2016-08-11 14:41:51 +07:00
|
|
|
|
2016-08-07 22:23:56 +07:00
|
|
|
dev_dbg(dev, "%s\n", __func__);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Note, rcu is not protecting the liveness of dax_dev, rcu is
|
|
|
|
* ensuring that any fault handlers that might have seen
|
|
|
|
* dax_dev->alive == true, have completed. Any fault handlers
|
|
|
|
* that start after synchronize_rcu() has started will abort
|
|
|
|
* upon seeing dax_dev->alive == false.
|
|
|
|
*/
|
|
|
|
dax_dev->alive = false;
|
|
|
|
synchronize_rcu();
|
2016-08-05 06:53:50 +07:00
|
|
|
unmap_mapping_range(dax_dev->inode->i_mapping, 0, 0, 1);
|
2016-07-25 05:55:42 +07:00
|
|
|
cdev_del(cdev);
|
2016-08-07 22:23:56 +07:00
|
|
|
device_unregister(dev);
|
|
|
|
}
|
|
|
|
|
2016-07-20 07:51:40 +07:00
|
|
|
struct dax_dev *devm_create_dax_dev(struct dax_region *dax_region,
|
|
|
|
struct resource *res, int count)
|
2016-08-07 22:23:56 +07:00
|
|
|
{
|
|
|
|
struct device *parent = dax_region->dev;
|
|
|
|
struct dax_dev *dax_dev;
|
2016-07-20 06:17:58 +07:00
|
|
|
int rc = 0, minor, i;
|
2016-08-07 22:23:56 +07:00
|
|
|
struct device *dev;
|
2016-07-25 05:55:42 +07:00
|
|
|
struct cdev *cdev;
|
2016-08-07 22:23:56 +07:00
|
|
|
dev_t dev_t;
|
|
|
|
|
|
|
|
dax_dev = kzalloc(sizeof(*dax_dev) + sizeof(*res) * count, GFP_KERNEL);
|
|
|
|
if (!dax_dev)
|
2016-07-20 07:51:40 +07:00
|
|
|
return ERR_PTR(-ENOMEM);
|
2016-08-07 22:23:56 +07:00
|
|
|
|
2016-07-20 06:17:58 +07:00
|
|
|
for (i = 0; i < count; i++) {
|
|
|
|
if (!IS_ALIGNED(res[i].start, dax_region->align)
|
|
|
|
|| !IS_ALIGNED(resource_size(&res[i]),
|
|
|
|
dax_region->align)) {
|
|
|
|
rc = -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
dax_dev->res[i].start = res[i].start;
|
|
|
|
dax_dev->res[i].end = res[i].end;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (i < count)
|
|
|
|
goto err_id;
|
|
|
|
|
2016-08-07 22:23:56 +07:00
|
|
|
dax_dev->id = ida_simple_get(&dax_region->ida, 0, 0, GFP_KERNEL);
|
|
|
|
if (dax_dev->id < 0) {
|
|
|
|
rc = dax_dev->id;
|
|
|
|
goto err_id;
|
|
|
|
}
|
|
|
|
|
|
|
|
minor = ida_simple_get(&dax_minor_ida, 0, 0, GFP_KERNEL);
|
|
|
|
if (minor < 0) {
|
|
|
|
rc = minor;
|
|
|
|
goto err_minor;
|
|
|
|
}
|
|
|
|
|
2016-09-08 20:53:28 +07:00
|
|
|
dev_t = MKDEV(MAJOR(dax_devt), minor);
|
|
|
|
dev = &dax_dev->dev;
|
2016-07-25 11:55:45 +07:00
|
|
|
dax_dev->inode = dax_inode_get(&dax_dev->cdev, dev_t);
|
|
|
|
if (!dax_dev->inode) {
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto err_inode;
|
|
|
|
}
|
|
|
|
|
2016-07-25 05:55:42 +07:00
|
|
|
/* device_initialize() so cdev can reference kobj parent */
|
2016-08-11 14:41:51 +07:00
|
|
|
device_initialize(dev);
|
2016-07-25 05:55:42 +07:00
|
|
|
|
|
|
|
cdev = &dax_dev->cdev;
|
|
|
|
cdev_init(cdev, &dax_fops);
|
|
|
|
cdev->owner = parent->driver->owner;
|
|
|
|
cdev->kobj.parent = &dev->kobj;
|
|
|
|
rc = cdev_add(&dax_dev->cdev, dev_t, 1);
|
|
|
|
if (rc)
|
|
|
|
goto err_cdev;
|
|
|
|
|
|
|
|
/* from here on we're committed to teardown via dax_dev_release() */
|
|
|
|
dax_dev->num_resources = count;
|
|
|
|
dax_dev->alive = true;
|
|
|
|
dax_dev->region = dax_region;
|
|
|
|
kref_get(&dax_region->kref);
|
|
|
|
|
2016-08-11 14:41:51 +07:00
|
|
|
dev->devt = dev_t;
|
|
|
|
dev->class = dax_class;
|
|
|
|
dev->parent = parent;
|
|
|
|
dev->groups = dax_attribute_groups;
|
|
|
|
dev->release = dax_dev_release;
|
|
|
|
dev_set_name(dev, "dax%d.%d", dax_region->id, dax_dev->id);
|
|
|
|
rc = device_add(dev);
|
|
|
|
if (rc) {
|
|
|
|
put_device(dev);
|
2016-07-20 07:51:40 +07:00
|
|
|
return ERR_PTR(rc);
|
2016-08-11 14:41:51 +07:00
|
|
|
}
|
2016-08-07 22:23:56 +07:00
|
|
|
|
2016-07-20 07:51:40 +07:00
|
|
|
rc = devm_add_action_or_reset(dax_region->dev, unregister_dax_dev, dev);
|
|
|
|
if (rc)
|
|
|
|
return ERR_PTR(rc);
|
|
|
|
|
|
|
|
return dax_dev;
|
2016-08-07 22:23:56 +07:00
|
|
|
|
2016-07-25 05:55:42 +07:00
|
|
|
err_cdev:
|
2016-07-25 11:55:45 +07:00
|
|
|
iput(dax_dev->inode);
|
|
|
|
err_inode:
|
2016-07-25 05:55:42 +07:00
|
|
|
ida_simple_remove(&dax_minor_ida, minor);
|
2016-08-07 22:23:56 +07:00
|
|
|
err_minor:
|
|
|
|
ida_simple_remove(&dax_region->ida, dax_dev->id);
|
|
|
|
err_id:
|
2016-08-11 14:41:51 +07:00
|
|
|
kfree(dax_dev);
|
2016-08-07 22:23:56 +07:00
|
|
|
|
2016-07-20 07:51:40 +07:00
|
|
|
return ERR_PTR(rc);
|
2016-08-07 22:23:56 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(devm_create_dax_dev);
|
|
|
|
|
2016-05-18 23:15:08 +07:00
|
|
|
static int __init dax_init(void)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
2016-07-25 11:55:45 +07:00
|
|
|
rc = dax_inode_init();
|
|
|
|
if (rc)
|
2016-05-18 23:15:08 +07:00
|
|
|
return rc;
|
2016-07-25 11:55:45 +07:00
|
|
|
|
2016-07-25 05:55:42 +07:00
|
|
|
nr_dax = max(nr_dax, 256);
|
|
|
|
rc = alloc_chrdev_region(&dax_devt, 0, nr_dax, "dax");
|
|
|
|
if (rc)
|
2016-07-25 11:55:45 +07:00
|
|
|
goto err_chrdev;
|
2016-05-18 23:15:08 +07:00
|
|
|
|
|
|
|
dax_class = class_create(THIS_MODULE, "dax");
|
|
|
|
if (IS_ERR(dax_class)) {
|
2016-07-25 11:55:45 +07:00
|
|
|
rc = PTR_ERR(dax_class);
|
|
|
|
goto err_class;
|
2016-05-18 23:15:08 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
2016-07-25 11:55:45 +07:00
|
|
|
|
|
|
|
err_class:
|
|
|
|
unregister_chrdev_region(dax_devt, nr_dax);
|
|
|
|
err_chrdev:
|
|
|
|
dax_inode_exit();
|
|
|
|
return rc;
|
2016-05-18 23:15:08 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static void __exit dax_exit(void)
|
|
|
|
{
|
|
|
|
class_destroy(dax_class);
|
2016-07-25 05:55:42 +07:00
|
|
|
unregister_chrdev_region(dax_devt, nr_dax);
|
2016-05-18 23:15:08 +07:00
|
|
|
ida_destroy(&dax_minor_ida);
|
2016-07-25 11:55:45 +07:00
|
|
|
dax_inode_exit();
|
2016-05-18 23:15:08 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
MODULE_AUTHOR("Intel Corporation");
|
|
|
|
MODULE_LICENSE("GPL v2");
|
|
|
|
subsys_initcall(dax_init);
|
|
|
|
module_exit(dax_exit);
|