2011-07-31 08:52:39 +08:00
|
|
|
/*
|
|
|
|
* linux/fs/nfs/blocklayout/blocklayout.c
|
|
|
|
*
|
|
|
|
* Module for the NFSv4.1 pNFS block layout driver.
|
|
|
|
*
|
|
|
|
* Copyright (c) 2006 The Regents of the University of Michigan.
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* Andy Adamson <andros@citi.umich.edu>
|
|
|
|
* Fred Isaman <iisaman@umich.edu>
|
|
|
|
*
|
|
|
|
* permission is granted to use, copy, create derivative works and
|
|
|
|
* redistribute this software and such derivative works for any purpose,
|
|
|
|
* so long as the name of the university of michigan is not used in
|
|
|
|
* any advertising or publicity pertaining to the use or distribution
|
|
|
|
* of this software without specific, written prior authorization. if
|
|
|
|
* the above copyright notice or any other identification of the
|
|
|
|
* university of michigan is included in any copy of any portion of
|
|
|
|
* this software, then the disclaimer below must also be included.
|
|
|
|
*
|
|
|
|
* this software is provided as is, without representation from the
|
|
|
|
* university of michigan as to its fitness for any purpose, and without
|
|
|
|
* warranty by the university of michigan of any kind, either express
|
|
|
|
* or implied, including without limitation the implied warranties of
|
|
|
|
* merchantability and fitness for a particular purpose. the regents
|
|
|
|
* of the university of michigan shall not be liable for any damages,
|
|
|
|
* including special, indirect, incidental, or consequential damages,
|
|
|
|
* with respect to any claim arising out or in connection with the use
|
|
|
|
* of the software, even if it has been or is hereafter advised of the
|
|
|
|
* possibility of such damages.
|
|
|
|
*/
|
2011-07-31 08:52:53 +08:00
|
|
|
|
2011-07-31 08:52:39 +08:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/init.h>
|
2011-07-31 08:52:42 +08:00
|
|
|
#include <linux/mount.h>
|
|
|
|
#include <linux/namei.h>
|
2011-07-31 08:52:53 +08:00
|
|
|
#include <linux/bio.h> /* struct bio */
|
2011-08-02 15:57:35 +08:00
|
|
|
#include <linux/prefetch.h>
|
2012-09-25 14:55:57 +08:00
|
|
|
#include <linux/pagevec.h>
|
2011-07-31 08:52:39 +08:00
|
|
|
|
2012-04-10 10:33:39 +08:00
|
|
|
#include "../pnfs.h"
|
2012-11-27 03:20:49 +08:00
|
|
|
#include "../nfs4session.h"
|
2012-04-10 10:33:39 +08:00
|
|
|
#include "../internal.h"
|
2011-07-31 08:52:39 +08:00
|
|
|
#include "blocklayout.h"
|
|
|
|
|
|
|
|
#define NFSDBG_FACILITY NFSDBG_PNFS_LD
|
|
|
|
|
|
|
|
MODULE_LICENSE("GPL");
|
|
|
|
MODULE_AUTHOR("Andy Adamson <andros@citi.umich.edu>");
|
|
|
|
MODULE_DESCRIPTION("The NFSv4.1 pNFS Block layout driver");
|
|
|
|
|
2014-09-10 23:23:34 +08:00
|
|
|
static bool is_hole(struct pnfs_block_extent *be)
|
2011-07-31 08:52:53 +08:00
|
|
|
{
|
2014-09-10 23:23:34 +08:00
|
|
|
switch (be->be_state) {
|
|
|
|
case PNFS_BLOCK_NONE_DATA:
|
|
|
|
return true;
|
|
|
|
case PNFS_BLOCK_INVALID_DATA:
|
|
|
|
return be->be_tag ? false : true;
|
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
2011-07-31 08:52:54 +08:00
|
|
|
}
|
|
|
|
|
2011-07-31 08:52:53 +08:00
|
|
|
/* The data we are handed might be spread across several bios. We need
|
|
|
|
* to track when the last one is finished.
|
|
|
|
*/
|
|
|
|
struct parallel_io {
|
|
|
|
struct kref refcnt;
|
2014-09-10 23:23:34 +08:00
|
|
|
void (*pnfs_callback) (void *data);
|
2011-07-31 08:52:53 +08:00
|
|
|
void *data;
|
|
|
|
};
|
|
|
|
|
|
|
|
static inline struct parallel_io *alloc_parallel(void *data)
|
|
|
|
{
|
|
|
|
struct parallel_io *rv;
|
|
|
|
|
|
|
|
rv = kmalloc(sizeof(*rv), GFP_NOFS);
|
|
|
|
if (rv) {
|
|
|
|
rv->data = data;
|
|
|
|
kref_init(&rv->refcnt);
|
|
|
|
}
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void get_parallel(struct parallel_io *p)
|
|
|
|
{
|
|
|
|
kref_get(&p->refcnt);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void destroy_parallel(struct kref *kref)
|
|
|
|
{
|
|
|
|
struct parallel_io *p = container_of(kref, struct parallel_io, refcnt);
|
|
|
|
|
|
|
|
dprintk("%s enter\n", __func__);
|
2014-09-10 23:23:34 +08:00
|
|
|
p->pnfs_callback(p->data);
|
2011-07-31 08:52:53 +08:00
|
|
|
kfree(p);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void put_parallel(struct parallel_io *p)
|
|
|
|
{
|
|
|
|
kref_put(&p->refcnt, destroy_parallel);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct bio *
|
|
|
|
bl_submit_bio(int rw, struct bio *bio)
|
|
|
|
{
|
|
|
|
if (bio) {
|
|
|
|
get_parallel(bio->bi_private);
|
|
|
|
dprintk("%s submitting %s bio %u@%llu\n", __func__,
|
2013-10-12 06:44:27 +08:00
|
|
|
rw == READ ? "read" : "write", bio->bi_iter.bi_size,
|
|
|
|
(unsigned long long)bio->bi_iter.bi_sector);
|
2011-07-31 08:52:53 +08:00
|
|
|
submit_bio(rw, bio);
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2014-09-11 08:37:27 +08:00
|
|
|
static struct bio *
|
|
|
|
bl_alloc_init_bio(int npg, struct block_device *bdev, sector_t disk_sector,
|
2015-07-20 21:29:37 +08:00
|
|
|
bio_end_io_t end_io, struct parallel_io *par)
|
2011-07-31 08:52:53 +08:00
|
|
|
{
|
|
|
|
struct bio *bio;
|
|
|
|
|
2012-01-12 23:18:48 +08:00
|
|
|
npg = min(npg, BIO_MAX_PAGES);
|
2011-07-31 08:52:53 +08:00
|
|
|
bio = bio_alloc(GFP_NOIO, npg);
|
2012-01-12 23:18:48 +08:00
|
|
|
if (!bio && (current->flags & PF_MEMALLOC)) {
|
|
|
|
while (!bio && (npg /= 2))
|
|
|
|
bio = bio_alloc(GFP_NOIO, npg);
|
|
|
|
}
|
2011-07-31 08:52:53 +08:00
|
|
|
|
2012-01-12 23:18:48 +08:00
|
|
|
if (bio) {
|
2014-09-11 08:37:27 +08:00
|
|
|
bio->bi_iter.bi_sector = disk_sector;
|
|
|
|
bio->bi_bdev = bdev;
|
2012-01-12 23:18:48 +08:00
|
|
|
bio->bi_end_io = end_io;
|
|
|
|
bio->bi_private = par;
|
|
|
|
}
|
2011-07-31 08:52:53 +08:00
|
|
|
return bio;
|
|
|
|
}
|
|
|
|
|
2014-09-11 08:37:27 +08:00
|
|
|
static struct bio *
|
|
|
|
do_add_page_to_bio(struct bio *bio, int npg, int rw, sector_t isect,
|
|
|
|
struct page *page, struct pnfs_block_dev_map *map,
|
2015-07-20 21:29:37 +08:00
|
|
|
struct pnfs_block_extent *be, bio_end_io_t end_io,
|
2014-09-11 08:37:27 +08:00
|
|
|
struct parallel_io *par, unsigned int offset, int *len)
|
2011-07-31 08:52:53 +08:00
|
|
|
{
|
2014-09-11 08:37:27 +08:00
|
|
|
struct pnfs_block_dev *dev =
|
|
|
|
container_of(be->be_device, struct pnfs_block_dev, node);
|
|
|
|
u64 disk_addr, end;
|
|
|
|
|
2012-08-24 00:27:51 +08:00
|
|
|
dprintk("%s: npg %d rw %d isect %llu offset %u len %d\n", __func__,
|
2014-09-11 08:37:27 +08:00
|
|
|
npg, rw, (unsigned long long)isect, offset, *len);
|
|
|
|
|
|
|
|
/* translate to device offset */
|
|
|
|
isect += be->be_v_offset;
|
|
|
|
isect -= be->be_f_offset;
|
|
|
|
|
|
|
|
/* translate to physical disk offset */
|
|
|
|
disk_addr = (u64)isect << SECTOR_SHIFT;
|
|
|
|
if (disk_addr < map->start || disk_addr >= map->start + map->len) {
|
|
|
|
if (!dev->map(dev, disk_addr, map))
|
|
|
|
return ERR_PTR(-EIO);
|
|
|
|
bio = bl_submit_bio(rw, bio);
|
|
|
|
}
|
|
|
|
disk_addr += map->disk_offset;
|
|
|
|
disk_addr -= map->start;
|
|
|
|
|
|
|
|
/* limit length to what the device mapping allows */
|
|
|
|
end = disk_addr + *len;
|
|
|
|
if (end >= map->start + map->len)
|
|
|
|
*len = map->start + map->len - disk_addr;
|
|
|
|
|
2011-07-31 08:52:53 +08:00
|
|
|
retry:
|
|
|
|
if (!bio) {
|
2014-09-11 08:37:27 +08:00
|
|
|
bio = bl_alloc_init_bio(npg, map->bdev,
|
|
|
|
disk_addr >> SECTOR_SHIFT, end_io, par);
|
2011-07-31 08:52:53 +08:00
|
|
|
if (!bio)
|
|
|
|
return ERR_PTR(-ENOMEM);
|
|
|
|
}
|
2014-09-11 08:37:27 +08:00
|
|
|
if (bio_add_page(bio, page, *len, offset) < *len) {
|
2011-07-31 08:52:53 +08:00
|
|
|
bio = bl_submit_bio(rw, bio);
|
|
|
|
goto retry;
|
|
|
|
}
|
|
|
|
return bio;
|
|
|
|
}
|
|
|
|
|
2015-07-20 21:29:37 +08:00
|
|
|
static void bl_end_io_read(struct bio *bio)
|
2011-07-31 08:52:53 +08:00
|
|
|
{
|
|
|
|
struct parallel_io *par = bio->bi_private;
|
|
|
|
|
2015-07-20 21:29:37 +08:00
|
|
|
if (bio->bi_error) {
|
2014-06-09 23:48:35 +08:00
|
|
|
struct nfs_pgio_header *header = par->data;
|
2012-04-21 02:47:44 +08:00
|
|
|
|
|
|
|
if (!header->pnfs_error)
|
|
|
|
header->pnfs_error = -EIO;
|
|
|
|
pnfs_set_lo_fail(header->lseg);
|
2011-07-31 08:52:53 +08:00
|
|
|
}
|
2014-09-10 23:23:33 +08:00
|
|
|
|
2011-07-31 08:52:53 +08:00
|
|
|
bio_put(bio);
|
|
|
|
put_parallel(par);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void bl_read_cleanup(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct rpc_task *task;
|
2014-06-09 23:48:35 +08:00
|
|
|
struct nfs_pgio_header *hdr;
|
2011-07-31 08:52:53 +08:00
|
|
|
dprintk("%s enter\n", __func__);
|
|
|
|
task = container_of(work, struct rpc_task, u.tk_work);
|
2014-06-09 23:48:35 +08:00
|
|
|
hdr = container_of(task, struct nfs_pgio_header, task);
|
|
|
|
pnfs_ld_read_done(hdr);
|
2011-07-31 08:52:53 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2014-09-10 23:23:34 +08:00
|
|
|
bl_end_par_io_read(void *data)
|
2011-07-31 08:52:53 +08:00
|
|
|
{
|
2014-06-09 23:48:35 +08:00
|
|
|
struct nfs_pgio_header *hdr = data;
|
2011-07-31 08:52:53 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
hdr->task.tk_status = hdr->pnfs_error;
|
|
|
|
INIT_WORK(&hdr->task.u.tk_work, bl_read_cleanup);
|
|
|
|
schedule_work(&hdr->task.u.tk_work);
|
2011-07-31 08:52:53 +08:00
|
|
|
}
|
|
|
|
|
2011-07-31 08:52:39 +08:00
|
|
|
static enum pnfs_try_status
|
2014-09-10 23:23:34 +08:00
|
|
|
bl_read_pagelist(struct nfs_pgio_header *header)
|
2011-07-31 08:52:39 +08:00
|
|
|
{
|
2014-09-10 23:23:34 +08:00
|
|
|
struct pnfs_block_layout *bl = BLK_LSEG2EXT(header->lseg);
|
2014-09-11 08:37:27 +08:00
|
|
|
struct pnfs_block_dev_map map = { .start = NFS4_MAX_UINT64 };
|
2011-07-31 08:52:53 +08:00
|
|
|
struct bio *bio = NULL;
|
2014-09-10 23:23:34 +08:00
|
|
|
struct pnfs_block_extent be;
|
2011-07-31 08:52:53 +08:00
|
|
|
sector_t isect, extent_length = 0;
|
|
|
|
struct parallel_io *par;
|
2014-09-10 23:23:34 +08:00
|
|
|
loff_t f_offset = header->args.offset;
|
|
|
|
size_t bytes_left = header->args.count;
|
nfs/blocklayout: Fix bad using of page offset in bl_read_pagelist
Blocklayout uses file offset for the read-back page's offset of first writing,
it's definitely wrong, it writes data to bad address of page that cause userspace
application segment fault. It must be the page base stored in header->args.pgbase.
Also, the pg_offset has no influence with isect and extent length.
Note: The offset of the non-first page is always zero.
Ps: A test program will segment fault at read() as,
#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
int main(int argc, char **argv)
{
char buf[2049];
char *filename = NULL;
int fd = -1;
if (argc < 2) {
printf("Usage: %s filename\n", argv[0]);
return 0;
}
filename = argv[1];
fd = open(filename, O_RDONLY | O_DIRECT);
if (fd < 0) {
printf("Open %s fail: %m\n", filename);
return 1;
}
lseek(fd, 2048, SEEK_SET);
if (read(fd, buf, sizeof(buf) - 1) != (sizeof(buf) - 1))
printf("Read 4096 bityes data from %s fail: %m\n", filename);
out:
close(fd);
return 0;
}
Signed-off-by: Kinglong Mee <kinglongmee@gmail.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-10-16 17:22:50 +08:00
|
|
|
unsigned int pg_offset = header->args.pgbase, pg_len;
|
2014-09-10 23:23:34 +08:00
|
|
|
struct page **pages = header->args.pages;
|
|
|
|
int pg_index = header->args.pgbase >> PAGE_CACHE_SHIFT;
|
2012-08-24 00:27:52 +08:00
|
|
|
const bool is_dio = (header->dreq != NULL);
|
2014-08-22 00:09:28 +08:00
|
|
|
struct blk_plug plug;
|
2014-09-10 23:23:34 +08:00
|
|
|
int i;
|
2011-07-31 08:52:53 +08:00
|
|
|
|
2012-03-21 02:12:46 +08:00
|
|
|
dprintk("%s enter nr_pages %u offset %lld count %u\n", __func__,
|
2014-09-10 23:23:34 +08:00
|
|
|
header->page_array.npages, f_offset,
|
|
|
|
(unsigned int)header->args.count);
|
2011-07-31 08:52:53 +08:00
|
|
|
|
2014-09-10 23:23:34 +08:00
|
|
|
par = alloc_parallel(header);
|
2011-07-31 08:52:53 +08:00
|
|
|
if (!par)
|
2014-09-10 23:23:34 +08:00
|
|
|
return PNFS_NOT_ATTEMPTED;
|
2011-07-31 08:52:53 +08:00
|
|
|
par->pnfs_callback = bl_end_par_io_read;
|
|
|
|
|
2014-08-22 00:09:28 +08:00
|
|
|
blk_start_plug(&plug);
|
|
|
|
|
2011-07-31 08:52:53 +08:00
|
|
|
isect = (sector_t) (f_offset >> SECTOR_SHIFT);
|
|
|
|
/* Code assumes extents are page-aligned */
|
2014-09-10 23:23:34 +08:00
|
|
|
for (i = pg_index; i < header->page_array.npages; i++) {
|
2014-08-22 00:09:29 +08:00
|
|
|
if (extent_length <= 0) {
|
2011-07-31 08:52:53 +08:00
|
|
|
/* We've used up the previous extent */
|
|
|
|
bio = bl_submit_bio(READ, bio);
|
2014-09-10 23:23:34 +08:00
|
|
|
|
2011-07-31 08:52:53 +08:00
|
|
|
/* Get the next one */
|
2014-09-10 23:23:34 +08:00
|
|
|
if (!ext_tree_lookup(bl, isect, &be, false)) {
|
2012-04-21 02:47:44 +08:00
|
|
|
header->pnfs_error = -EIO;
|
2011-07-31 08:52:53 +08:00
|
|
|
goto out;
|
|
|
|
}
|
2014-09-10 23:23:34 +08:00
|
|
|
extent_length = be.be_length - (isect - be.be_f_offset);
|
2011-07-31 08:52:53 +08:00
|
|
|
}
|
2012-08-24 00:27:52 +08:00
|
|
|
|
|
|
|
if (is_dio) {
|
|
|
|
if (pg_offset + bytes_left > PAGE_CACHE_SIZE)
|
|
|
|
pg_len = PAGE_CACHE_SIZE - pg_offset;
|
|
|
|
else
|
|
|
|
pg_len = bytes_left;
|
|
|
|
} else {
|
2014-09-10 23:23:32 +08:00
|
|
|
BUG_ON(pg_offset != 0);
|
2012-08-24 00:27:52 +08:00
|
|
|
pg_len = PAGE_CACHE_SIZE;
|
|
|
|
}
|
|
|
|
|
2014-09-10 23:23:34 +08:00
|
|
|
if (is_hole(&be)) {
|
2011-07-31 08:52:53 +08:00
|
|
|
bio = bl_submit_bio(READ, bio);
|
|
|
|
/* Fill hole w/ zeroes w/o accessing device */
|
|
|
|
dprintk("%s Zeroing page for hole\n", __func__);
|
2012-08-24 00:27:52 +08:00
|
|
|
zero_user_segment(pages[i], pg_offset, pg_len);
|
2014-09-11 08:37:27 +08:00
|
|
|
|
|
|
|
/* invalidate map */
|
|
|
|
map.start = NFS4_MAX_UINT64;
|
2011-07-31 08:52:53 +08:00
|
|
|
} else {
|
2014-06-09 23:48:34 +08:00
|
|
|
bio = do_add_page_to_bio(bio,
|
2014-09-10 23:23:34 +08:00
|
|
|
header->page_array.npages - i,
|
2012-04-21 02:47:45 +08:00
|
|
|
READ,
|
2014-09-11 08:37:27 +08:00
|
|
|
isect, pages[i], &map, &be,
|
2012-08-24 00:27:52 +08:00
|
|
|
bl_end_io_read, par,
|
2014-09-11 08:37:27 +08:00
|
|
|
pg_offset, &pg_len);
|
2011-07-31 08:52:53 +08:00
|
|
|
if (IS_ERR(bio)) {
|
2012-04-21 02:47:44 +08:00
|
|
|
header->pnfs_error = PTR_ERR(bio);
|
2011-09-23 09:50:16 +08:00
|
|
|
bio = NULL;
|
2011-07-31 08:52:53 +08:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
2012-08-24 00:27:52 +08:00
|
|
|
isect += (pg_len >> SECTOR_SHIFT);
|
2014-08-22 00:09:29 +08:00
|
|
|
extent_length -= (pg_len >> SECTOR_SHIFT);
|
2014-09-11 08:37:27 +08:00
|
|
|
f_offset += pg_len;
|
|
|
|
bytes_left -= pg_len;
|
nfs/blocklayout: Fix bad using of page offset in bl_read_pagelist
Blocklayout uses file offset for the read-back page's offset of first writing,
it's definitely wrong, it writes data to bad address of page that cause userspace
application segment fault. It must be the page base stored in header->args.pgbase.
Also, the pg_offset has no influence with isect and extent length.
Note: The offset of the non-first page is always zero.
Ps: A test program will segment fault at read() as,
#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
int main(int argc, char **argv)
{
char buf[2049];
char *filename = NULL;
int fd = -1;
if (argc < 2) {
printf("Usage: %s filename\n", argv[0]);
return 0;
}
filename = argv[1];
fd = open(filename, O_RDONLY | O_DIRECT);
if (fd < 0) {
printf("Open %s fail: %m\n", filename);
return 1;
}
lseek(fd, 2048, SEEK_SET);
if (read(fd, buf, sizeof(buf) - 1) != (sizeof(buf) - 1))
printf("Read 4096 bityes data from %s fail: %m\n", filename);
out:
close(fd);
return 0;
}
Signed-off-by: Kinglong Mee <kinglongmee@gmail.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-10-16 17:22:50 +08:00
|
|
|
pg_offset = 0;
|
2011-07-31 08:52:53 +08:00
|
|
|
}
|
2012-04-21 02:47:44 +08:00
|
|
|
if ((isect << SECTOR_SHIFT) >= header->inode->i_size) {
|
2014-09-10 23:23:34 +08:00
|
|
|
header->res.eof = 1;
|
|
|
|
header->res.count = header->inode->i_size - header->args.offset;
|
2011-07-31 08:52:53 +08:00
|
|
|
} else {
|
2014-09-10 23:23:34 +08:00
|
|
|
header->res.count = (isect << SECTOR_SHIFT) - header->args.offset;
|
2011-07-31 08:52:53 +08:00
|
|
|
}
|
|
|
|
out:
|
|
|
|
bl_submit_bio(READ, bio);
|
2014-08-22 00:09:28 +08:00
|
|
|
blk_finish_plug(&plug);
|
2011-07-31 08:52:53 +08:00
|
|
|
put_parallel(par);
|
|
|
|
return PNFS_ATTEMPTED;
|
2011-07-31 08:52:55 +08:00
|
|
|
}
|
|
|
|
|
2015-07-20 21:29:37 +08:00
|
|
|
static void bl_end_io_write(struct bio *bio)
|
2011-07-31 08:52:54 +08:00
|
|
|
{
|
|
|
|
struct parallel_io *par = bio->bi_private;
|
2014-06-09 23:48:35 +08:00
|
|
|
struct nfs_pgio_header *header = par->data;
|
2011-07-31 08:52:54 +08:00
|
|
|
|
2015-07-20 21:29:37 +08:00
|
|
|
if (bio->bi_error) {
|
2012-04-21 02:47:44 +08:00
|
|
|
if (!header->pnfs_error)
|
|
|
|
header->pnfs_error = -EIO;
|
|
|
|
pnfs_set_lo_fail(header->lseg);
|
2011-07-31 08:52:54 +08:00
|
|
|
}
|
|
|
|
bio_put(bio);
|
|
|
|
put_parallel(par);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Function scheduled for call during bl_end_par_io_write,
|
|
|
|
* it marks sectors as written and extends the commitlist.
|
|
|
|
*/
|
|
|
|
static void bl_write_cleanup(struct work_struct *work)
|
|
|
|
{
|
2014-09-10 23:23:34 +08:00
|
|
|
struct rpc_task *task = container_of(work, struct rpc_task, u.tk_work);
|
|
|
|
struct nfs_pgio_header *hdr =
|
|
|
|
container_of(task, struct nfs_pgio_header, task);
|
|
|
|
|
2011-07-31 08:52:54 +08:00
|
|
|
dprintk("%s enter\n", __func__);
|
2014-09-10 23:23:34 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
if (likely(!hdr->pnfs_error)) {
|
2014-09-10 23:23:34 +08:00
|
|
|
struct pnfs_block_layout *bl = BLK_LSEG2EXT(hdr->lseg);
|
|
|
|
u64 start = hdr->args.offset & (loff_t)PAGE_CACHE_MASK;
|
|
|
|
u64 end = (hdr->args.offset + hdr->args.count +
|
|
|
|
PAGE_CACHE_SIZE - 1) & (loff_t)PAGE_CACHE_MASK;
|
|
|
|
|
|
|
|
ext_tree_mark_written(bl, start >> SECTOR_SHIFT,
|
|
|
|
(end - start) >> SECTOR_SHIFT);
|
2011-07-31 08:52:55 +08:00
|
|
|
}
|
2014-09-10 23:23:34 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
pnfs_ld_write_done(hdr);
|
2011-07-31 08:52:54 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Called when last of bios associated with a bl_write_pagelist call finishes */
|
2014-09-10 23:23:34 +08:00
|
|
|
static void bl_end_par_io_write(void *data)
|
2011-07-31 08:52:54 +08:00
|
|
|
{
|
2014-06-09 23:48:35 +08:00
|
|
|
struct nfs_pgio_header *hdr = data;
|
2011-07-31 08:52:54 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
hdr->task.tk_status = hdr->pnfs_error;
|
2014-06-09 23:48:36 +08:00
|
|
|
hdr->verf.committed = NFS_FILE_SYNC;
|
2014-06-09 23:48:35 +08:00
|
|
|
INIT_WORK(&hdr->task.u.tk_work, bl_write_cleanup);
|
|
|
|
schedule_work(&hdr->task.u.tk_work);
|
2011-07-31 08:52:54 +08:00
|
|
|
}
|
|
|
|
|
2011-07-31 08:52:39 +08:00
|
|
|
static enum pnfs_try_status
|
2014-06-09 23:48:35 +08:00
|
|
|
bl_write_pagelist(struct nfs_pgio_header *header, int sync)
|
2011-07-31 08:52:39 +08:00
|
|
|
{
|
2014-09-10 23:23:34 +08:00
|
|
|
struct pnfs_block_layout *bl = BLK_LSEG2EXT(header->lseg);
|
2014-09-11 08:37:27 +08:00
|
|
|
struct pnfs_block_dev_map map = { .start = NFS4_MAX_UINT64 };
|
2011-07-31 08:52:54 +08:00
|
|
|
struct bio *bio = NULL;
|
2014-09-10 23:23:34 +08:00
|
|
|
struct pnfs_block_extent be;
|
2014-09-10 23:23:32 +08:00
|
|
|
sector_t isect, extent_length = 0;
|
2012-08-24 00:27:53 +08:00
|
|
|
struct parallel_io *par = NULL;
|
2014-06-09 23:48:35 +08:00
|
|
|
loff_t offset = header->args.offset;
|
|
|
|
size_t count = header->args.count;
|
|
|
|
struct page **pages = header->args.pages;
|
2014-10-21 19:32:10 +08:00
|
|
|
int pg_index = header->args.pgbase >> PAGE_CACHE_SHIFT;
|
2014-09-11 08:37:27 +08:00
|
|
|
unsigned int pg_len;
|
2014-08-22 00:09:28 +08:00
|
|
|
struct blk_plug plug;
|
2014-09-10 23:23:34 +08:00
|
|
|
int i;
|
2011-07-31 08:52:54 +08:00
|
|
|
|
|
|
|
dprintk("%s enter, %Zu@%lld\n", __func__, count, offset);
|
2012-08-24 00:27:53 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
/* At this point, header->page_aray is a (sequential) list of nfs_pages.
|
2011-07-31 08:52:56 +08:00
|
|
|
* We want to write each, and if there is an error set pnfs_error
|
|
|
|
* to have it redone using nfs.
|
2011-07-31 08:52:54 +08:00
|
|
|
*/
|
2014-06-09 23:48:35 +08:00
|
|
|
par = alloc_parallel(header);
|
2011-07-31 08:52:54 +08:00
|
|
|
if (!par)
|
2014-09-10 23:23:34 +08:00
|
|
|
return PNFS_NOT_ATTEMPTED;
|
2011-07-31 08:52:54 +08:00
|
|
|
par->pnfs_callback = bl_end_par_io_write;
|
|
|
|
|
2014-09-10 23:23:32 +08:00
|
|
|
blk_start_plug(&plug);
|
2011-07-31 08:52:56 +08:00
|
|
|
|
2014-09-10 23:23:32 +08:00
|
|
|
/* we always write out the whole page */
|
|
|
|
offset = offset & (loff_t)PAGE_CACHE_MASK;
|
|
|
|
isect = offset >> SECTOR_SHIFT;
|
2011-07-31 08:52:56 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
for (i = pg_index; i < header->page_array.npages; i++) {
|
2014-08-22 00:09:29 +08:00
|
|
|
if (extent_length <= 0) {
|
2011-07-31 08:52:54 +08:00
|
|
|
/* We've used up the previous extent */
|
|
|
|
bio = bl_submit_bio(WRITE, bio);
|
|
|
|
/* Get the next one */
|
2014-09-10 23:23:34 +08:00
|
|
|
if (!ext_tree_lookup(bl, isect, &be, true)) {
|
2012-04-21 02:47:44 +08:00
|
|
|
header->pnfs_error = -EINVAL;
|
2011-07-31 08:52:54 +08:00
|
|
|
goto out;
|
|
|
|
}
|
2012-08-24 00:27:51 +08:00
|
|
|
|
2014-09-10 23:23:34 +08:00
|
|
|
extent_length = be.be_length - (isect - be.be_f_offset);
|
2011-07-31 08:52:56 +08:00
|
|
|
}
|
2012-08-24 00:27:51 +08:00
|
|
|
|
2014-09-11 08:37:27 +08:00
|
|
|
pg_len = PAGE_CACHE_SIZE;
|
2014-06-09 23:48:35 +08:00
|
|
|
bio = do_add_page_to_bio(bio, header->page_array.npages - i,
|
2014-09-11 08:37:27 +08:00
|
|
|
WRITE, isect, pages[i], &map, &be,
|
2012-08-24 00:27:51 +08:00
|
|
|
bl_end_io_write, par,
|
2014-09-11 08:37:27 +08:00
|
|
|
0, &pg_len);
|
2011-07-31 08:52:56 +08:00
|
|
|
if (IS_ERR(bio)) {
|
2012-04-21 02:47:44 +08:00
|
|
|
header->pnfs_error = PTR_ERR(bio);
|
2011-09-23 09:50:16 +08:00
|
|
|
bio = NULL;
|
2011-07-31 08:52:56 +08:00
|
|
|
goto out;
|
2011-07-31 08:52:54 +08:00
|
|
|
}
|
2014-09-11 08:37:27 +08:00
|
|
|
|
|
|
|
offset += pg_len;
|
|
|
|
count -= pg_len;
|
|
|
|
isect += (pg_len >> SECTOR_SHIFT);
|
|
|
|
extent_length -= (pg_len >> SECTOR_SHIFT);
|
2011-07-31 08:52:54 +08:00
|
|
|
}
|
2011-07-31 08:52:56 +08:00
|
|
|
|
2014-06-09 23:48:35 +08:00
|
|
|
header->res.count = header->args.count;
|
2011-07-31 08:52:54 +08:00
|
|
|
out:
|
|
|
|
bl_submit_bio(WRITE, bio);
|
2014-08-22 00:09:28 +08:00
|
|
|
blk_finish_plug(&plug);
|
2011-07-31 08:52:54 +08:00
|
|
|
put_parallel(par);
|
|
|
|
return PNFS_ATTEMPTED;
|
2011-07-31 08:52:39 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void bl_free_layout_hdr(struct pnfs_layout_hdr *lo)
|
|
|
|
{
|
|
|
|
struct pnfs_block_layout *bl = BLK_LO2EXT(lo);
|
2014-09-10 23:23:34 +08:00
|
|
|
int err;
|
2011-07-31 08:52:39 +08:00
|
|
|
|
|
|
|
dprintk("%s enter\n", __func__);
|
2014-09-10 23:23:34 +08:00
|
|
|
|
|
|
|
err = ext_tree_remove(bl, true, 0, LLONG_MAX);
|
|
|
|
WARN_ON(err);
|
|
|
|
|
2011-07-31 08:52:39 +08:00
|
|
|
kfree(bl);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pnfs_layout_hdr *bl_alloc_layout_hdr(struct inode *inode,
|
|
|
|
gfp_t gfp_flags)
|
|
|
|
{
|
|
|
|
struct pnfs_block_layout *bl;
|
|
|
|
|
|
|
|
dprintk("%s enter\n", __func__);
|
|
|
|
bl = kzalloc(sizeof(*bl), gfp_flags);
|
|
|
|
if (!bl)
|
|
|
|
return NULL;
|
2014-09-10 23:23:34 +08:00
|
|
|
|
|
|
|
bl->bl_ext_rw = RB_ROOT;
|
|
|
|
bl->bl_ext_ro = RB_ROOT;
|
2011-07-31 08:52:39 +08:00
|
|
|
spin_lock_init(&bl->bl_ext_lock);
|
2014-09-10 23:23:34 +08:00
|
|
|
|
2011-07-31 08:52:39 +08:00
|
|
|
return &bl->bl_layout;
|
|
|
|
}
|
|
|
|
|
2011-07-31 08:52:44 +08:00
|
|
|
static void bl_free_lseg(struct pnfs_layout_segment *lseg)
|
2011-07-31 08:52:39 +08:00
|
|
|
{
|
2011-07-31 08:52:44 +08:00
|
|
|
dprintk("%s enter\n", __func__);
|
|
|
|
kfree(lseg);
|
2011-07-31 08:52:39 +08:00
|
|
|
}
|
|
|
|
|
2014-09-11 08:37:24 +08:00
|
|
|
/* Tracks info needed to ensure extents in layout obey constraints of spec */
|
|
|
|
struct layout_verification {
|
|
|
|
u32 mode; /* R or RW */
|
|
|
|
u64 start; /* Expected start of next non-COW extent */
|
|
|
|
u64 inval; /* Start of INVAL coverage */
|
|
|
|
u64 cowread; /* End of COW read coverage */
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Verify the extent meets the layout requirements of the pnfs-block draft,
|
|
|
|
* section 2.3.1.
|
|
|
|
*/
|
|
|
|
static int verify_extent(struct pnfs_block_extent *be,
|
|
|
|
struct layout_verification *lv)
|
|
|
|
{
|
|
|
|
if (lv->mode == IOMODE_READ) {
|
|
|
|
if (be->be_state == PNFS_BLOCK_READWRITE_DATA ||
|
|
|
|
be->be_state == PNFS_BLOCK_INVALID_DATA)
|
|
|
|
return -EIO;
|
|
|
|
if (be->be_f_offset != lv->start)
|
|
|
|
return -EIO;
|
|
|
|
lv->start += be->be_length;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
/* lv->mode == IOMODE_RW */
|
|
|
|
if (be->be_state == PNFS_BLOCK_READWRITE_DATA) {
|
|
|
|
if (be->be_f_offset != lv->start)
|
|
|
|
return -EIO;
|
|
|
|
if (lv->cowread > lv->start)
|
|
|
|
return -EIO;
|
|
|
|
lv->start += be->be_length;
|
|
|
|
lv->inval = lv->start;
|
|
|
|
return 0;
|
|
|
|
} else if (be->be_state == PNFS_BLOCK_INVALID_DATA) {
|
|
|
|
if (be->be_f_offset != lv->start)
|
|
|
|
return -EIO;
|
|
|
|
lv->start += be->be_length;
|
|
|
|
return 0;
|
|
|
|
} else if (be->be_state == PNFS_BLOCK_READ_DATA) {
|
|
|
|
if (be->be_f_offset > lv->start)
|
|
|
|
return -EIO;
|
|
|
|
if (be->be_f_offset < lv->inval)
|
|
|
|
return -EIO;
|
|
|
|
if (be->be_f_offset < lv->cowread)
|
|
|
|
return -EIO;
|
|
|
|
/* It looks like you might want to min this with lv->start,
|
|
|
|
* but you really don't.
|
|
|
|
*/
|
|
|
|
lv->inval = lv->inval + be->be_length;
|
|
|
|
lv->cowread = be->be_f_offset + be->be_length;
|
|
|
|
return 0;
|
|
|
|
} else
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int decode_sector_number(__be32 **rp, sector_t *sp)
|
|
|
|
{
|
|
|
|
uint64_t s;
|
|
|
|
|
|
|
|
*rp = xdr_decode_hyper(*rp, &s);
|
|
|
|
if (s & 0x1ff) {
|
|
|
|
printk(KERN_WARNING "NFS: %s: sector not aligned\n", __func__);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
*sp = s >> SECTOR_SHIFT;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2014-09-11 08:37:25 +08:00
|
|
|
bl_alloc_extent(struct xdr_stream *xdr, struct pnfs_layout_hdr *lo,
|
|
|
|
struct layout_verification *lv, struct list_head *extents,
|
|
|
|
gfp_t gfp_mask)
|
2014-09-11 08:37:24 +08:00
|
|
|
{
|
2014-09-11 08:37:25 +08:00
|
|
|
struct pnfs_block_extent *be;
|
|
|
|
struct nfs4_deviceid id;
|
|
|
|
int error;
|
2014-09-11 08:37:24 +08:00
|
|
|
__be32 *p;
|
2014-09-11 08:37:25 +08:00
|
|
|
|
|
|
|
p = xdr_inline_decode(xdr, 28 + NFS4_DEVICEID4_SIZE);
|
|
|
|
if (!p)
|
|
|
|
return -EIO;
|
|
|
|
|
|
|
|
be = kzalloc(sizeof(*be), GFP_NOFS);
|
|
|
|
if (!be)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
memcpy(&id, p, NFS4_DEVICEID4_SIZE);
|
|
|
|
p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE);
|
|
|
|
|
|
|
|
error = -EIO;
|
|
|
|
be->be_device = nfs4_find_get_deviceid(NFS_SERVER(lo->plh_inode), &id,
|
|
|
|
lo->plh_lc_cred, gfp_mask);
|
|
|
|
if (!be->be_device)
|
|
|
|
goto out_free_be;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The next three values are read in as bytes, but stored in the
|
|
|
|
* extent structure in 512-byte granularity.
|
|
|
|
*/
|
|
|
|
if (decode_sector_number(&p, &be->be_f_offset) < 0)
|
|
|
|
goto out_put_deviceid;
|
|
|
|
if (decode_sector_number(&p, &be->be_length) < 0)
|
|
|
|
goto out_put_deviceid;
|
|
|
|
if (decode_sector_number(&p, &be->be_v_offset) < 0)
|
|
|
|
goto out_put_deviceid;
|
|
|
|
be->be_state = be32_to_cpup(p++);
|
|
|
|
|
|
|
|
error = verify_extent(be, lv);
|
|
|
|
if (error) {
|
|
|
|
dprintk("%s: extent verification failed\n", __func__);
|
|
|
|
goto out_put_deviceid;
|
|
|
|
}
|
|
|
|
|
|
|
|
list_add_tail(&be->be_list, extents);
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
out_put_deviceid:
|
|
|
|
nfs4_put_deviceid_node(be->be_device);
|
|
|
|
out_free_be:
|
|
|
|
kfree(be);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pnfs_layout_segment *
|
|
|
|
bl_alloc_lseg(struct pnfs_layout_hdr *lo, struct nfs4_layoutget_res *lgr,
|
|
|
|
gfp_t gfp_mask)
|
|
|
|
{
|
2014-09-11 08:37:24 +08:00
|
|
|
struct layout_verification lv = {
|
|
|
|
.mode = lgr->range.iomode,
|
|
|
|
.start = lgr->range.offset >> SECTOR_SHIFT,
|
|
|
|
.inval = lgr->range.offset >> SECTOR_SHIFT,
|
|
|
|
.cowread = lgr->range.offset >> SECTOR_SHIFT,
|
|
|
|
};
|
2014-09-11 08:37:25 +08:00
|
|
|
struct pnfs_block_layout *bl = BLK_LO2EXT(lo);
|
|
|
|
struct pnfs_layout_segment *lseg;
|
|
|
|
struct xdr_buf buf;
|
|
|
|
struct xdr_stream xdr;
|
|
|
|
struct page *scratch;
|
|
|
|
int status, i;
|
|
|
|
uint32_t count;
|
|
|
|
__be32 *p;
|
2014-09-11 08:37:24 +08:00
|
|
|
LIST_HEAD(extents);
|
|
|
|
|
|
|
|
dprintk("---> %s\n", __func__);
|
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
lseg = kzalloc(sizeof(*lseg), gfp_mask);
|
|
|
|
if (!lseg)
|
|
|
|
return ERR_PTR(-ENOMEM);
|
|
|
|
|
|
|
|
status = -ENOMEM;
|
|
|
|
scratch = alloc_page(gfp_mask);
|
2014-09-11 08:37:24 +08:00
|
|
|
if (!scratch)
|
2014-09-11 08:37:25 +08:00
|
|
|
goto out;
|
2014-09-11 08:37:24 +08:00
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
xdr_init_decode_pages(&xdr, &buf,
|
|
|
|
lgr->layoutp->pages, lgr->layoutp->len);
|
|
|
|
xdr_set_scratch_buffer(&xdr, page_address(scratch), PAGE_SIZE);
|
2014-09-11 08:37:24 +08:00
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
status = -EIO;
|
|
|
|
p = xdr_inline_decode(&xdr, 4);
|
2014-09-11 08:37:24 +08:00
|
|
|
if (unlikely(!p))
|
2014-09-11 08:37:25 +08:00
|
|
|
goto out_free_scratch;
|
2014-09-11 08:37:24 +08:00
|
|
|
|
|
|
|
count = be32_to_cpup(p++);
|
2014-09-11 08:37:25 +08:00
|
|
|
dprintk("%s: number of extents %d\n", __func__, count);
|
2014-09-11 08:37:24 +08:00
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
/*
|
|
|
|
* Decode individual extents, putting them in temporary staging area
|
|
|
|
* until whole layout is decoded to make error recovery easier.
|
2014-09-11 08:37:24 +08:00
|
|
|
*/
|
|
|
|
for (i = 0; i < count; i++) {
|
2014-09-11 08:37:25 +08:00
|
|
|
status = bl_alloc_extent(&xdr, lo, &lv, &extents, gfp_mask);
|
|
|
|
if (status)
|
|
|
|
goto process_extents;
|
2014-09-11 08:37:24 +08:00
|
|
|
}
|
2014-09-11 08:37:25 +08:00
|
|
|
|
2014-09-11 08:37:24 +08:00
|
|
|
if (lgr->range.offset + lgr->range.length !=
|
|
|
|
lv.start << SECTOR_SHIFT) {
|
|
|
|
dprintk("%s Final length mismatch\n", __func__);
|
2014-09-11 08:37:25 +08:00
|
|
|
status = -EIO;
|
|
|
|
goto process_extents;
|
2014-09-11 08:37:24 +08:00
|
|
|
}
|
2014-09-11 08:37:25 +08:00
|
|
|
|
2014-09-11 08:37:24 +08:00
|
|
|
if (lv.start < lv.cowread) {
|
|
|
|
dprintk("%s Final uncovered COW extent\n", __func__);
|
2014-09-11 08:37:25 +08:00
|
|
|
status = -EIO;
|
2014-09-11 08:37:24 +08:00
|
|
|
}
|
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
process_extents:
|
2014-09-11 08:37:24 +08:00
|
|
|
while (!list_empty(&extents)) {
|
2014-09-11 08:37:25 +08:00
|
|
|
struct pnfs_block_extent *be =
|
|
|
|
list_first_entry(&extents, struct pnfs_block_extent,
|
|
|
|
be_list);
|
2014-09-11 08:37:24 +08:00
|
|
|
list_del(&be->be_list);
|
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
if (!status)
|
|
|
|
status = ext_tree_insert(bl, be);
|
2011-07-31 08:52:44 +08:00
|
|
|
|
2014-09-11 08:37:25 +08:00
|
|
|
if (status) {
|
|
|
|
nfs4_put_deviceid_node(be->be_device);
|
|
|
|
kfree(be);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
out_free_scratch:
|
|
|
|
__free_page(scratch);
|
|
|
|
out:
|
|
|
|
dprintk("%s returns %d\n", __func__, status);
|
2011-07-31 08:52:44 +08:00
|
|
|
if (status) {
|
|
|
|
kfree(lseg);
|
|
|
|
return ERR_PTR(status);
|
|
|
|
}
|
|
|
|
return lseg;
|
2011-07-31 08:52:39 +08:00
|
|
|
}
|
|
|
|
|
2014-09-10 23:23:35 +08:00
|
|
|
static void
|
|
|
|
bl_return_range(struct pnfs_layout_hdr *lo,
|
|
|
|
struct pnfs_layout_range *range)
|
|
|
|
{
|
|
|
|
struct pnfs_block_layout *bl = BLK_LO2EXT(lo);
|
|
|
|
sector_t offset = range->offset >> SECTOR_SHIFT, end;
|
|
|
|
|
|
|
|
if (range->offset % 8) {
|
|
|
|
dprintk("%s: offset %lld not block size aligned\n",
|
|
|
|
__func__, range->offset);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (range->length != NFS4_MAX_UINT64) {
|
|
|
|
if (range->length % 8) {
|
|
|
|
dprintk("%s: length %lld not block size aligned\n",
|
|
|
|
__func__, range->length);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
end = offset + (range->length >> SECTOR_SHIFT);
|
|
|
|
} else {
|
|
|
|
end = round_down(NFS4_MAX_UINT64, PAGE_SIZE);
|
|
|
|
}
|
|
|
|
|
2014-09-13 01:25:14 +08:00
|
|
|
ext_tree_remove(bl, range->iomode & IOMODE_RW, offset, end);
|
2014-09-10 23:23:35 +08:00
|
|
|
}
|
|
|
|
|
2014-09-11 08:36:30 +08:00
|
|
|
static int
|
|
|
|
bl_prepare_layoutcommit(struct nfs4_layoutcommit_args *arg)
|
2011-07-31 08:52:39 +08:00
|
|
|
{
|
2014-09-11 08:36:30 +08:00
|
|
|
return ext_tree_prepare_commit(arg);
|
2011-07-31 08:52:39 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
bl_cleanup_layoutcommit(struct nfs4_layoutcommit_data *lcdata)
|
|
|
|
{
|
2014-09-11 08:36:30 +08:00
|
|
|
ext_tree_mark_committed(&lcdata->args, lcdata->res.status);
|
2011-07-31 08:52:39 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
bl_set_layoutdriver(struct nfs_server *server, const struct nfs_fh *fh)
|
|
|
|
{
|
|
|
|
dprintk("%s enter\n", __func__);
|
2011-07-31 08:52:46 +08:00
|
|
|
|
|
|
|
if (server->pnfs_blksize == 0) {
|
|
|
|
dprintk("%s Server did not return blksize\n", __func__);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2014-08-22 00:09:26 +08:00
|
|
|
if (server->pnfs_blksize > PAGE_SIZE) {
|
|
|
|
printk(KERN_ERR "%s: pNFS blksize %d not supported.\n",
|
|
|
|
__func__, server->pnfs_blksize);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2014-09-11 08:36:31 +08:00
|
|
|
return 0;
|
2011-07-31 08:52:39 +08:00
|
|
|
}
|
|
|
|
|
2012-08-24 00:27:52 +08:00
|
|
|
static bool
|
2014-09-10 23:23:32 +08:00
|
|
|
is_aligned_req(struct nfs_pageio_descriptor *pgio,
|
|
|
|
struct nfs_page *req, unsigned int alignment)
|
2012-08-24 00:27:52 +08:00
|
|
|
{
|
2014-09-10 23:23:32 +08:00
|
|
|
/*
|
|
|
|
* Always accept buffered writes, higher layers take care of the
|
|
|
|
* right alignment.
|
|
|
|
*/
|
|
|
|
if (pgio->pg_dreq == NULL)
|
|
|
|
return true;
|
|
|
|
|
|
|
|
if (!IS_ALIGNED(req->wb_offset, alignment))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
if (IS_ALIGNED(req->wb_bytes, alignment))
|
|
|
|
return true;
|
|
|
|
|
|
|
|
if (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode)) {
|
|
|
|
/*
|
|
|
|
* If the write goes up to the inode size, just write
|
|
|
|
* the full page. Data past the inode size is
|
|
|
|
* guaranteed to be zeroed by the higher level client
|
|
|
|
* code, and this behaviour is mandated by RFC 5663
|
|
|
|
* section 2.3.2.
|
|
|
|
*/
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
2012-08-24 00:27:52 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
bl_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
|
|
|
|
{
|
2014-09-10 23:23:32 +08:00
|
|
|
if (!is_aligned_req(pgio, req, SECTOR_SIZE)) {
|
2012-08-24 00:27:52 +08:00
|
|
|
nfs_pageio_reset_read_mds(pgio);
|
2014-09-10 23:23:32 +08:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
pnfs_generic_pg_init_read(pgio, req);
|
2012-08-24 00:27:52 +08:00
|
|
|
}
|
|
|
|
|
2014-05-15 23:56:43 +08:00
|
|
|
/*
|
|
|
|
* Return 0 if @req cannot be coalesced into @pgio, otherwise return the number
|
|
|
|
* of bytes (maximum @req->wb_bytes) that can be coalesced.
|
|
|
|
*/
|
|
|
|
static size_t
|
2012-08-24 00:27:52 +08:00
|
|
|
bl_pg_test_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
|
|
|
|
struct nfs_page *req)
|
|
|
|
{
|
2014-09-10 23:23:32 +08:00
|
|
|
if (!is_aligned_req(pgio, req, SECTOR_SIZE))
|
2014-05-15 23:56:43 +08:00
|
|
|
return 0;
|
2012-08-24 00:27:52 +08:00
|
|
|
return pnfs_generic_pg_test(pgio, prev, req);
|
|
|
|
}
|
|
|
|
|
2012-09-25 14:55:57 +08:00
|
|
|
/*
|
|
|
|
* Return the number of contiguous bytes for a given inode
|
|
|
|
* starting at page frame idx.
|
|
|
|
*/
|
|
|
|
static u64 pnfs_num_cont_bytes(struct inode *inode, pgoff_t idx)
|
|
|
|
{
|
|
|
|
struct address_space *mapping = inode->i_mapping;
|
|
|
|
pgoff_t end;
|
|
|
|
|
|
|
|
/* Optimize common case that writes from 0 to end of file */
|
|
|
|
end = DIV_ROUND_UP(i_size_read(inode), PAGE_CACHE_SIZE);
|
2014-11-25 05:47:02 +08:00
|
|
|
if (end != inode->i_mapping->nrpages) {
|
2012-09-25 14:55:57 +08:00
|
|
|
rcu_read_lock();
|
2014-04-04 05:47:44 +08:00
|
|
|
end = page_cache_next_hole(mapping, idx + 1, ULONG_MAX);
|
2012-09-25 14:55:57 +08:00
|
|
|
rcu_read_unlock();
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!end)
|
|
|
|
return i_size_read(inode) - (idx << PAGE_CACHE_SHIFT);
|
|
|
|
else
|
|
|
|
return (end - idx) << PAGE_CACHE_SHIFT;
|
|
|
|
}
|
|
|
|
|
2012-10-02 23:29:14 +08:00
|
|
|
static void
|
2012-08-24 00:27:53 +08:00
|
|
|
bl_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
|
|
|
|
{
|
2014-09-10 23:23:32 +08:00
|
|
|
u64 wb_size;
|
|
|
|
|
|
|
|
if (!is_aligned_req(pgio, req, PAGE_SIZE)) {
|
2012-08-24 00:27:53 +08:00
|
|
|
nfs_pageio_reset_write_mds(pgio);
|
2014-09-10 23:23:32 +08:00
|
|
|
return;
|
2012-09-25 14:55:57 +08:00
|
|
|
}
|
2014-09-10 23:23:32 +08:00
|
|
|
|
|
|
|
if (pgio->pg_dreq == NULL)
|
|
|
|
wb_size = pnfs_num_cont_bytes(pgio->pg_inode,
|
|
|
|
req->wb_index);
|
|
|
|
else
|
|
|
|
wb_size = nfs_dreq_bytes_left(pgio->pg_dreq);
|
|
|
|
|
|
|
|
pnfs_generic_pg_init_write(pgio, req, wb_size);
|
2012-08-24 00:27:53 +08:00
|
|
|
}
|
|
|
|
|
2014-05-15 23:56:43 +08:00
|
|
|
/*
|
|
|
|
* Return 0 if @req cannot be coalesced into @pgio, otherwise return the number
|
|
|
|
* of bytes (maximum @req->wb_bytes) that can be coalesced.
|
|
|
|
*/
|
|
|
|
static size_t
|
2012-08-24 00:27:53 +08:00
|
|
|
bl_pg_test_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
|
|
|
|
struct nfs_page *req)
|
|
|
|
{
|
2014-09-10 23:23:32 +08:00
|
|
|
if (!is_aligned_req(pgio, req, PAGE_SIZE))
|
2014-05-15 23:56:43 +08:00
|
|
|
return 0;
|
2012-08-24 00:27:53 +08:00
|
|
|
return pnfs_generic_pg_test(pgio, prev, req);
|
|
|
|
}
|
|
|
|
|
2011-07-31 08:52:40 +08:00
|
|
|
static const struct nfs_pageio_ops bl_pg_read_ops = {
|
2012-08-24 00:27:52 +08:00
|
|
|
.pg_init = bl_pg_init_read,
|
|
|
|
.pg_test = bl_pg_test_read,
|
2011-07-31 08:52:40 +08:00
|
|
|
.pg_doio = pnfs_generic_pg_readpages,
|
2014-09-11 03:48:01 +08:00
|
|
|
.pg_cleanup = pnfs_generic_pg_cleanup,
|
2011-07-31 08:52:40 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static const struct nfs_pageio_ops bl_pg_write_ops = {
|
2012-08-24 00:27:53 +08:00
|
|
|
.pg_init = bl_pg_init_write,
|
|
|
|
.pg_test = bl_pg_test_write,
|
2011-07-31 08:52:40 +08:00
|
|
|
.pg_doio = pnfs_generic_pg_writepages,
|
2014-09-11 03:48:01 +08:00
|
|
|
.pg_cleanup = pnfs_generic_pg_cleanup,
|
2011-07-31 08:52:40 +08:00
|
|
|
};
|
|
|
|
|
2011-07-31 08:52:39 +08:00
|
|
|
static struct pnfs_layoutdriver_type blocklayout_type = {
|
|
|
|
.id = LAYOUT_BLOCK_VOLUME,
|
|
|
|
.name = "LAYOUT_BLOCK_VOLUME",
|
2013-02-04 21:15:02 +08:00
|
|
|
.owner = THIS_MODULE,
|
2014-09-10 23:23:36 +08:00
|
|
|
.flags = PNFS_LAYOUTRET_ON_SETATTR |
|
|
|
|
PNFS_READ_WHOLE_PAGE,
|
2011-07-31 08:52:39 +08:00
|
|
|
.read_pagelist = bl_read_pagelist,
|
|
|
|
.write_pagelist = bl_write_pagelist,
|
|
|
|
.alloc_layout_hdr = bl_alloc_layout_hdr,
|
|
|
|
.free_layout_hdr = bl_free_layout_hdr,
|
|
|
|
.alloc_lseg = bl_alloc_lseg,
|
|
|
|
.free_lseg = bl_free_lseg,
|
2014-09-10 23:23:35 +08:00
|
|
|
.return_range = bl_return_range,
|
2014-09-11 08:36:30 +08:00
|
|
|
.prepare_layoutcommit = bl_prepare_layoutcommit,
|
2011-07-31 08:52:39 +08:00
|
|
|
.cleanup_layoutcommit = bl_cleanup_layoutcommit,
|
|
|
|
.set_layoutdriver = bl_set_layoutdriver,
|
2014-09-03 12:28:00 +08:00
|
|
|
.alloc_deviceid_node = bl_alloc_deviceid_node,
|
|
|
|
.free_deviceid_node = bl_free_deviceid_node,
|
2011-07-31 08:52:40 +08:00
|
|
|
.pg_read_ops = &bl_pg_read_ops,
|
|
|
|
.pg_write_ops = &bl_pg_write_ops,
|
2015-03-26 02:14:42 +08:00
|
|
|
.sync = pnfs_generic_sync,
|
2011-07-31 08:52:39 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static int __init nfs4blocklayout_init(void)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
dprintk("%s: NFSv4 Block Layout Driver Registering...\n", __func__);
|
|
|
|
|
|
|
|
ret = pnfs_register_layoutdriver(&blocklayout_type);
|
2011-07-31 08:52:42 +08:00
|
|
|
if (ret)
|
|
|
|
goto out;
|
2014-09-11 08:37:26 +08:00
|
|
|
ret = bl_init_pipefs();
|
2012-01-10 21:04:32 +08:00
|
|
|
if (ret)
|
2014-09-11 08:37:26 +08:00
|
|
|
goto out_unregister;
|
|
|
|
return 0;
|
2011-07-31 08:52:42 +08:00
|
|
|
|
2014-09-11 08:37:26 +08:00
|
|
|
out_unregister:
|
2011-07-31 08:52:42 +08:00
|
|
|
pnfs_unregister_layoutdriver(&blocklayout_type);
|
2014-09-11 08:37:26 +08:00
|
|
|
out:
|
2011-07-31 08:52:39 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __exit nfs4blocklayout_exit(void)
|
|
|
|
{
|
|
|
|
dprintk("%s: NFSv4 Block Layout Driver Unregistering...\n",
|
|
|
|
__func__);
|
|
|
|
|
2014-09-11 08:37:26 +08:00
|
|
|
bl_cleanup_pipefs();
|
2011-07-31 08:52:39 +08:00
|
|
|
pnfs_unregister_layoutdriver(&blocklayout_type);
|
|
|
|
}
|
|
|
|
|
|
|
|
MODULE_ALIAS("nfs-layouttype4-3");
|
|
|
|
|
|
|
|
module_init(nfs4blocklayout_init);
|
|
|
|
module_exit(nfs4blocklayout_exit);
|