2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* linux/fs/nfs/read.c
|
|
|
|
*
|
|
|
|
* Block I/O for NFS
|
|
|
|
*
|
|
|
|
* Partial copy of Linus' read cache modifications to fs/nfs/file.c
|
|
|
|
* modified for async RPC by okir@monad.swb.de
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/time.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/fcntl.h>
|
|
|
|
#include <linux/stat.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/pagemap.h>
|
|
|
|
#include <linux/sunrpc/clnt.h>
|
|
|
|
#include <linux/nfs_fs.h>
|
|
|
|
#include <linux/nfs_page.h>
|
2011-02-28 18:34:16 -07:00
|
|
|
#include <linux/module.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-04-01 07:22:25 -06:00
|
|
|
#include "nfs4_fs.h"
|
2006-12-04 22:35:38 -07:00
|
|
|
#include "internal.h"
|
2006-03-20 11:44:14 -07:00
|
|
|
#include "iostat.h"
|
2009-04-03 09:42:44 -06:00
|
|
|
#include "fscache.h"
|
2014-04-16 07:07:22 -06:00
|
|
|
#include "pnfs.h"
|
2006-03-20 11:44:14 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
#define NFSDBG_FACILITY NFSDBG_PAGECACHE
|
|
|
|
|
2012-04-20 12:47:48 -06:00
|
|
|
static const struct nfs_pgio_completion_ops nfs_async_read_completion_ops;
|
2014-05-06 07:12:30 -06:00
|
|
|
static const struct nfs_rw_ops nfs_rw_read_ops;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2006-12-06 21:33:20 -07:00
|
|
|
static struct kmem_cache *nfs_rdata_cachep;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2014-05-06 07:12:30 -06:00
|
|
|
static struct nfs_rw_header *nfs_readhdr_alloc(void)
|
2006-03-20 11:44:37 -07:00
|
|
|
{
|
2014-05-06 07:12:30 -06:00
|
|
|
return kmem_cache_zalloc(nfs_rdata_cachep, GFP_KERNEL);
|
2012-04-20 12:47:46 -06:00
|
|
|
}
|
|
|
|
|
2014-05-06 07:12:30 -06:00
|
|
|
static void nfs_readhdr_free(struct nfs_rw_header *rhdr)
|
2006-03-20 11:44:37 -07:00
|
|
|
{
|
2012-04-20 12:47:44 -06:00
|
|
|
kmem_cache_free(nfs_rdata_cachep, rhdr);
|
2006-03-20 11:44:37 -07:00
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
static
|
|
|
|
int nfs_return_empty_page(struct page *page)
|
|
|
|
{
|
2008-02-04 23:28:29 -07:00
|
|
|
zero_user(page, 0, PAGE_CACHE_SIZE);
|
2005-04-16 16:20:36 -06:00
|
|
|
SetPageUptodate(page);
|
|
|
|
unlock_page(page);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-06-20 13:53:47 -06:00
|
|
|
void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
|
2014-04-16 07:07:22 -06:00
|
|
|
struct inode *inode, bool force_mds,
|
2012-04-20 12:47:48 -06:00
|
|
|
const struct nfs_pgio_completion_ops *compl_ops)
|
2011-06-10 11:30:23 -06:00
|
|
|
{
|
2014-04-16 07:07:22 -06:00
|
|
|
struct nfs_server *server = NFS_SERVER(inode);
|
2014-05-06 07:12:40 -06:00
|
|
|
const struct nfs_pageio_ops *pg_ops = &nfs_pgio_rw_ops;
|
2014-04-16 07:07:22 -06:00
|
|
|
|
|
|
|
#ifdef CONFIG_NFS_V4_1
|
|
|
|
if (server->pnfs_curr_ld && !force_mds)
|
|
|
|
pg_ops = server->pnfs_curr_ld->pg_read_ops;
|
|
|
|
#endif
|
2014-05-06 07:12:30 -06:00
|
|
|
nfs_pageio_init(pgio, inode, pg_ops, compl_ops, &nfs_rw_read_ops,
|
|
|
|
server->rsize, 0);
|
2011-06-10 11:30:23 -06:00
|
|
|
}
|
2012-07-30 14:05:23 -06:00
|
|
|
EXPORT_SYMBOL_GPL(nfs_pageio_init_read);
|
2011-06-10 11:30:23 -06:00
|
|
|
|
2011-07-13 13:58:28 -06:00
|
|
|
void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio)
|
|
|
|
{
|
2014-05-06 07:12:40 -06:00
|
|
|
pgio->pg_ops = &nfs_pgio_rw_ops;
|
2011-07-13 13:58:28 -06:00
|
|
|
pgio->pg_bsize = NFS_SERVER(pgio->pg_inode)->rsize;
|
|
|
|
}
|
2011-07-13 13:59:57 -06:00
|
|
|
EXPORT_SYMBOL_GPL(nfs_pageio_reset_read_mds);
|
2011-07-13 13:58:28 -06:00
|
|
|
|
2009-04-03 09:42:44 -06:00
|
|
|
int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
|
|
|
|
struct page *page)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct nfs_page *new;
|
|
|
|
unsigned int len;
|
2011-03-03 08:13:48 -07:00
|
|
|
struct nfs_pageio_descriptor pgio;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2006-12-04 22:35:38 -07:00
|
|
|
len = nfs_page_length(page);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (len == 0)
|
|
|
|
return nfs_return_empty_page(page);
|
2014-05-15 09:56:45 -06:00
|
|
|
new = nfs_create_request(ctx, page, NULL, 0, len);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (IS_ERR(new)) {
|
|
|
|
unlock_page(page);
|
|
|
|
return PTR_ERR(new);
|
|
|
|
}
|
|
|
|
if (len < PAGE_CACHE_SIZE)
|
2008-02-04 23:28:29 -07:00
|
|
|
zero_user_segment(page, len, PAGE_CACHE_SIZE);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2014-04-16 07:07:22 -06:00
|
|
|
nfs_pageio_init_read(&pgio, inode, false,
|
|
|
|
&nfs_async_read_completion_ops);
|
2011-06-10 11:30:23 -06:00
|
|
|
nfs_pageio_add_request(&pgio, new);
|
2011-06-10 11:30:23 -06:00
|
|
|
nfs_pageio_complete(&pgio);
|
2012-05-24 11:13:24 -06:00
|
|
|
NFS_I(inode)->read_io += pgio.pg_bytes_written;
|
2005-04-16 16:20:36 -06:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void nfs_readpage_release(struct nfs_page *req)
|
|
|
|
{
|
2011-06-22 16:40:12 -06:00
|
|
|
struct inode *d_inode = req->wb_context->dentry->d_inode;
|
2009-04-03 09:42:45 -06:00
|
|
|
|
2014-05-15 09:56:46 -06:00
|
|
|
dprintk("NFS: read done (%s/%llu %d@%lld)\n", d_inode->i_sb->s_id,
|
|
|
|
(unsigned long long)NFS_FILEID(d_inode), req->wb_bytes,
|
|
|
|
(long long)req_offset(req));
|
2009-04-03 09:42:45 -06:00
|
|
|
|
2014-05-15 09:56:46 -06:00
|
|
|
if (nfs_page_group_sync_on_bit(req, PG_UNLOCKPAGE)) {
|
|
|
|
if (PageUptodate(req->wb_page))
|
|
|
|
nfs_readpage_to_fscache(d_inode, req->wb_page, 0);
|
|
|
|
|
|
|
|
unlock_page(req->wb_page);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2013-12-17 10:20:16 -07:00
|
|
|
dprintk("NFS: read done (%s/%Lu %d@%Ld)\n",
|
2011-06-22 16:40:12 -06:00
|
|
|
req->wb_context->dentry->d_inode->i_sb->s_id,
|
2013-12-17 10:20:16 -07:00
|
|
|
(unsigned long long)NFS_FILEID(req->wb_context->dentry->d_inode),
|
2005-04-16 16:20:36 -06:00
|
|
|
req->wb_bytes,
|
|
|
|
(long long)req_offset(req));
|
2005-09-22 22:44:28 -06:00
|
|
|
nfs_release_request(req);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2014-05-15 09:56:46 -06:00
|
|
|
static void nfs_page_group_set_uptodate(struct nfs_page *req)
|
|
|
|
{
|
|
|
|
if (nfs_page_group_sync_on_bit(req, PG_UPTODATE))
|
|
|
|
SetPageUptodate(req->wb_page);
|
|
|
|
}
|
|
|
|
|
2012-04-20 12:47:48 -06:00
|
|
|
static void nfs_read_completion(struct nfs_pgio_header *hdr)
|
2012-04-20 12:47:46 -06:00
|
|
|
{
|
|
|
|
unsigned long bytes = 0;
|
|
|
|
|
|
|
|
if (test_bit(NFS_IOHDR_REDO, &hdr->flags))
|
|
|
|
goto out;
|
2012-05-01 10:49:58 -06:00
|
|
|
while (!list_empty(&hdr->pages)) {
|
|
|
|
struct nfs_page *req = nfs_list_entry(hdr->pages.next);
|
|
|
|
struct page *page = req->wb_page;
|
2014-05-15 09:56:57 -06:00
|
|
|
unsigned long start = req->wb_pgbase;
|
|
|
|
unsigned long end = req->wb_pgbase + req->wb_bytes;
|
2012-05-01 10:49:58 -06:00
|
|
|
|
|
|
|
if (test_bit(NFS_IOHDR_EOF, &hdr->flags)) {
|
2014-05-15 09:56:57 -06:00
|
|
|
/* note: regions of the page not covered by a
|
|
|
|
* request are zeroed in nfs_readpage_async /
|
|
|
|
* readpage_async_filler */
|
|
|
|
if (bytes > hdr->good_bytes) {
|
|
|
|
/* nothing in this request was good, so zero
|
|
|
|
* the full extent of the request */
|
|
|
|
zero_user_segment(page, start, end);
|
|
|
|
|
|
|
|
} else if (hdr->good_bytes - bytes < req->wb_bytes) {
|
|
|
|
/* part of this request has good bytes, but
|
|
|
|
* not all. zero the bad bytes */
|
|
|
|
start += hdr->good_bytes - bytes;
|
|
|
|
WARN_ON(start < req->wb_pgbase);
|
|
|
|
zero_user_segment(page, start, end);
|
|
|
|
}
|
2012-04-20 12:47:46 -06:00
|
|
|
}
|
2012-05-01 10:49:58 -06:00
|
|
|
bytes += req->wb_bytes;
|
|
|
|
if (test_bit(NFS_IOHDR_ERROR, &hdr->flags)) {
|
2012-04-20 12:47:46 -06:00
|
|
|
if (bytes <= hdr->good_bytes)
|
2014-05-15 09:56:46 -06:00
|
|
|
nfs_page_group_set_uptodate(req);
|
2012-05-01 10:49:58 -06:00
|
|
|
} else
|
2014-05-15 09:56:46 -06:00
|
|
|
nfs_page_group_set_uptodate(req);
|
2012-05-01 10:49:58 -06:00
|
|
|
nfs_list_remove_request(req);
|
|
|
|
nfs_readpage_release(req);
|
2012-04-20 12:47:46 -06:00
|
|
|
}
|
|
|
|
out:
|
|
|
|
hdr->release(hdr);
|
|
|
|
}
|
|
|
|
|
2014-05-06 07:12:37 -06:00
|
|
|
static void nfs_initiate_read(struct nfs_pgio_data *data, struct rpc_message *msg,
|
|
|
|
struct rpc_task_setup *task_setup_data, int how)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2012-04-20 12:47:44 -06:00
|
|
|
struct inode *inode = data->header->inode;
|
2007-07-14 13:39:59 -06:00
|
|
|
int swap_flags = IS_SWAPFILE(inode) ? NFS_RPC_SWAPFLAGS : 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2014-05-06 07:12:37 -06:00
|
|
|
task_setup_data->flags |= swap_flags;
|
|
|
|
NFS_PROTO(inode)->read_setup(data, msg);
|
2011-02-28 18:34:16 -07:00
|
|
|
}
|
|
|
|
|
2012-04-20 12:47:48 -06:00
|
|
|
static void
|
2005-04-16 16:20:36 -06:00
|
|
|
nfs_async_read_error(struct list_head *head)
|
|
|
|
{
|
|
|
|
struct nfs_page *req;
|
|
|
|
|
|
|
|
while (!list_empty(head)) {
|
|
|
|
req = nfs_list_entry(head->next);
|
|
|
|
nfs_list_remove_request(req);
|
|
|
|
nfs_readpage_release(req);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-04-20 12:47:48 -06:00
|
|
|
static const struct nfs_pgio_completion_ops nfs_async_read_completion_ops = {
|
|
|
|
.error_cleanup = nfs_async_read_error,
|
|
|
|
.completion = nfs_read_completion,
|
|
|
|
};
|
|
|
|
|
2006-11-14 14:12:23 -07:00
|
|
|
/*
|
|
|
|
* This is the callback from RPC telling us whether a reply was
|
|
|
|
* received or some error occurred (timeout or socket shutdown).
|
|
|
|
*/
|
2014-05-06 07:12:32 -06:00
|
|
|
static int nfs_readpage_done(struct rpc_task *task, struct nfs_pgio_data *data,
|
|
|
|
struct inode *inode)
|
2006-11-14 14:12:23 -07:00
|
|
|
{
|
2014-05-06 07:12:32 -06:00
|
|
|
int status = NFS_PROTO(inode)->read_done(task, data);
|
2006-11-14 14:12:23 -07:00
|
|
|
if (status != 0)
|
|
|
|
return status;
|
|
|
|
|
2012-04-20 12:47:44 -06:00
|
|
|
nfs_add_stats(inode, NFSIOS_SERVERREADBYTES, data->res.count);
|
2006-11-14 14:12:23 -07:00
|
|
|
|
|
|
|
if (task->tk_status == -ESTALE) {
|
2012-04-20 12:47:44 -06:00
|
|
|
set_bit(NFS_INO_STALE, &NFS_I(inode)->flags);
|
|
|
|
nfs_mark_for_revalidate(inode);
|
2006-11-14 14:12:23 -07:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-05-06 07:12:26 -06:00
|
|
|
static void nfs_readpage_retry(struct rpc_task *task, struct nfs_pgio_data *data)
|
2006-11-14 14:12:23 -07:00
|
|
|
{
|
2014-05-06 07:12:24 -06:00
|
|
|
struct nfs_pgio_args *argp = &data->args;
|
2014-05-06 07:12:25 -06:00
|
|
|
struct nfs_pgio_res *resp = &data->res;
|
2006-11-14 14:12:23 -07:00
|
|
|
|
|
|
|
/* This is a short read! */
|
2012-04-20 12:47:44 -06:00
|
|
|
nfs_inc_stats(data->header->inode, NFSIOS_SHORTREAD);
|
2006-11-14 14:12:23 -07:00
|
|
|
/* Has the server at least made some progress? */
|
2012-04-20 12:47:46 -06:00
|
|
|
if (resp->count == 0) {
|
|
|
|
nfs_set_pgio_error(data->header, -EIO, argp->offset);
|
2009-12-05 17:32:19 -07:00
|
|
|
return;
|
2012-04-20 12:47:46 -06:00
|
|
|
}
|
2006-11-14 14:12:23 -07:00
|
|
|
/* Yes, so retry the read at the end of the data */
|
2011-02-28 18:34:20 -07:00
|
|
|
data->mds_offset += resp->count;
|
2006-11-14 14:12:23 -07:00
|
|
|
argp->offset += resp->count;
|
|
|
|
argp->pgbase += resp->count;
|
|
|
|
argp->count -= resp->count;
|
2011-10-19 13:17:29 -06:00
|
|
|
rpc_restart_call_prepare(task);
|
2006-11-14 14:12:23 -07:00
|
|
|
}
|
|
|
|
|
2014-05-06 07:12:32 -06:00
|
|
|
static void nfs_readpage_result(struct rpc_task *task, struct nfs_pgio_data *data)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2012-04-20 12:47:46 -06:00
|
|
|
struct nfs_pgio_header *hdr = data->header;
|
|
|
|
|
2014-05-06 07:12:32 -06:00
|
|
|
if (data->res.eof) {
|
2012-04-20 12:47:46 -06:00
|
|
|
loff_t bound;
|
|
|
|
|
|
|
|
bound = data->args.offset + data->res.count;
|
|
|
|
spin_lock(&hdr->lock);
|
|
|
|
if (bound < hdr->io_start + hdr->good_bytes) {
|
|
|
|
set_bit(NFS_IOHDR_EOF, &hdr->flags);
|
|
|
|
clear_bit(NFS_IOHDR_ERROR, &hdr->flags);
|
|
|
|
hdr->good_bytes = bound - hdr->io_start;
|
|
|
|
}
|
|
|
|
spin_unlock(&hdr->lock);
|
|
|
|
} else if (data->res.count != data->args.count)
|
|
|
|
nfs_readpage_retry(task, data);
|
2008-04-15 14:33:58 -06:00
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Read a page over NFS.
|
|
|
|
* We read the page synchronously in the following case:
|
|
|
|
* - The error flag is set for this page. This happens only when a
|
|
|
|
* previous async read operation failed.
|
|
|
|
*/
|
|
|
|
int nfs_readpage(struct file *file, struct page *page)
|
|
|
|
{
|
|
|
|
struct nfs_open_context *ctx;
|
2012-07-31 17:45:06 -06:00
|
|
|
struct inode *inode = page_file_mapping(page)->host;
|
2005-04-16 16:20:36 -06:00
|
|
|
int error;
|
|
|
|
|
|
|
|
dprintk("NFS: nfs_readpage (%p %ld@%lu)\n",
|
2012-07-31 17:45:06 -06:00
|
|
|
page, PAGE_CACHE_SIZE, page_file_index(page));
|
2006-03-20 11:44:14 -07:00
|
|
|
nfs_inc_stats(inode, NFSIOS_VFSREADPAGE);
|
|
|
|
nfs_add_stats(inode, NFSIOS_READPAGES, 1);
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Try to flush any pending writes to the file..
|
|
|
|
*
|
|
|
|
* NOTE! Because we own the page lock, there cannot
|
|
|
|
* be any new pending writes generated at this point
|
|
|
|
* for this page (other pages can be written to).
|
|
|
|
*/
|
|
|
|
error = nfs_wb_page(inode, page);
|
|
|
|
if (error)
|
2007-05-20 11:05:05 -06:00
|
|
|
goto out_unlock;
|
|
|
|
if (PageUptodate(page))
|
|
|
|
goto out_unlock;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2006-09-14 12:03:14 -06:00
|
|
|
error = -ESTALE;
|
|
|
|
if (NFS_STALE(inode))
|
2007-05-20 11:05:05 -06:00
|
|
|
goto out_unlock;
|
2006-09-14 12:03:14 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (file == NULL) {
|
2006-11-19 14:44:52 -07:00
|
|
|
error = -EBADF;
|
2005-11-04 13:33:38 -07:00
|
|
|
ctx = nfs_find_open_context(inode, NULL, FMODE_READ);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (ctx == NULL)
|
2007-05-20 11:05:05 -06:00
|
|
|
goto out_unlock;
|
2005-04-16 16:20:36 -06:00
|
|
|
} else
|
2007-08-10 15:44:32 -06:00
|
|
|
ctx = get_nfs_open_context(nfs_file_open_context(file));
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-04-03 09:42:44 -06:00
|
|
|
if (!IS_SYNC(inode)) {
|
|
|
|
error = nfs_readpage_from_fscache(ctx, inode, page);
|
|
|
|
if (error == 0)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2006-12-13 13:23:44 -07:00
|
|
|
error = nfs_readpage_async(ctx, inode, page);
|
|
|
|
|
2009-04-03 09:42:44 -06:00
|
|
|
out:
|
2005-04-16 16:20:36 -06:00
|
|
|
put_nfs_open_context(ctx);
|
|
|
|
return error;
|
2007-05-20 11:05:05 -06:00
|
|
|
out_unlock:
|
2005-04-16 16:20:36 -06:00
|
|
|
unlock_page(page);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct nfs_readdesc {
|
2007-04-02 16:48:28 -06:00
|
|
|
struct nfs_pageio_descriptor *pgio;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct nfs_open_context *ctx;
|
|
|
|
};
|
|
|
|
|
|
|
|
static int
|
|
|
|
readpage_async_filler(void *data, struct page *page)
|
|
|
|
{
|
|
|
|
struct nfs_readdesc *desc = (struct nfs_readdesc *)data;
|
|
|
|
struct nfs_page *new;
|
|
|
|
unsigned int len;
|
2007-05-20 11:05:05 -06:00
|
|
|
int error;
|
|
|
|
|
2006-12-04 22:35:38 -07:00
|
|
|
len = nfs_page_length(page);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (len == 0)
|
|
|
|
return nfs_return_empty_page(page);
|
2007-05-20 11:05:05 -06:00
|
|
|
|
2014-05-15 09:56:45 -06:00
|
|
|
new = nfs_create_request(desc->ctx, page, NULL, 0, len);
|
2007-05-20 11:05:05 -06:00
|
|
|
if (IS_ERR(new))
|
|
|
|
goto out_error;
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (len < PAGE_CACHE_SIZE)
|
2008-02-04 23:28:29 -07:00
|
|
|
zero_user_segment(page, len, PAGE_CACHE_SIZE);
|
nfs: don't ignore return value from nfs_pageio_add_request
Ignoring the return value from nfs_pageio_add_request can cause deadlocks.
In read path:
call nfs_pageio_add_request from readpage_async_filler
assume at this point that there are requests already in desc, that
can't be merged with the current request.
so nfs_pageio_doio is fired up to clear out desc.
assume something goes wrong in setting up the io, so desc->pg_error is set.
This causes nfs_pageio_add_request to return 0, *WITHOUT* adding the original
request.
BUT, since return code is ignored, readpage_async_filler assumes it has
been added, and does nothing further, leaving page locked.
do_generic_mapping_read will eventually call lock_page, resulting in deadlock
In write path:
page is marked dirty by generic_perform_write
nfs_writepages is called
call nfs_pageio_add_request from nfs_page_async_flush
assume at this point that there are requests already in desc, that
can't be merged with the current request.
so nfs_pageio_doio is fired up to clear out desc.
assume something goes wrong in setting up the io, so desc->pg_error is set.
This causes nfs_page_async_flush to return 0, *WITHOUT* adding the original
request, yet marking the request as locked (PG_BUSY) and in writeback,
clearing dirty marks.
The next time a write is done to the page, deadlock will result as
nfs_write_end calls nfs_update_request
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2008-03-19 09:24:39 -06:00
|
|
|
if (!nfs_pageio_add_request(desc->pgio, new)) {
|
|
|
|
error = desc->pgio->pg_error;
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
return 0;
|
2007-05-20 11:05:05 -06:00
|
|
|
out_error:
|
|
|
|
error = PTR_ERR(new);
|
|
|
|
out_unlock:
|
|
|
|
unlock_page(page);
|
|
|
|
return error;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
int nfs_readpages(struct file *filp, struct address_space *mapping,
|
|
|
|
struct list_head *pages, unsigned nr_pages)
|
|
|
|
{
|
2007-04-02 16:48:28 -06:00
|
|
|
struct nfs_pageio_descriptor pgio;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct nfs_readdesc desc = {
|
2007-04-02 16:48:28 -06:00
|
|
|
.pgio = &pgio,
|
2005-04-16 16:20:36 -06:00
|
|
|
};
|
|
|
|
struct inode *inode = mapping->host;
|
2007-04-02 16:48:28 -06:00
|
|
|
unsigned long npages;
|
2006-09-14 12:03:14 -06:00
|
|
|
int ret = -ESTALE;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2013-12-17 10:20:16 -07:00
|
|
|
dprintk("NFS: nfs_readpages (%s/%Lu %d)\n",
|
2005-04-16 16:20:36 -06:00
|
|
|
inode->i_sb->s_id,
|
2013-12-17 10:20:16 -07:00
|
|
|
(unsigned long long)NFS_FILEID(inode),
|
2005-04-16 16:20:36 -06:00
|
|
|
nr_pages);
|
2006-03-20 11:44:14 -07:00
|
|
|
nfs_inc_stats(inode, NFSIOS_VFSREADPAGES);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2006-09-14 12:03:14 -06:00
|
|
|
if (NFS_STALE(inode))
|
|
|
|
goto out;
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (filp == NULL) {
|
2005-11-04 13:33:38 -07:00
|
|
|
desc.ctx = nfs_find_open_context(inode, NULL, FMODE_READ);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (desc.ctx == NULL)
|
|
|
|
return -EBADF;
|
|
|
|
} else
|
2007-08-10 15:44:32 -06:00
|
|
|
desc.ctx = get_nfs_open_context(nfs_file_open_context(filp));
|
2009-04-03 09:42:44 -06:00
|
|
|
|
|
|
|
/* attempt to read as many of the pages as possible from the cache
|
|
|
|
* - this returns -ENOBUFS immediately if the cookie is negative
|
|
|
|
*/
|
|
|
|
ret = nfs_readpages_from_fscache(desc.ctx, inode, mapping,
|
|
|
|
pages, &nr_pages);
|
|
|
|
if (ret == 0)
|
|
|
|
goto read_complete; /* all pages were read */
|
|
|
|
|
2014-04-16 07:07:22 -06:00
|
|
|
nfs_pageio_init_read(&pgio, inode, false,
|
|
|
|
&nfs_async_read_completion_ops);
|
2007-04-02 16:48:28 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
ret = read_cache_pages(mapping, pages, readpage_async_filler, &desc);
|
2007-04-02 16:48:28 -06:00
|
|
|
|
|
|
|
nfs_pageio_complete(&pgio);
|
2012-05-24 11:13:24 -06:00
|
|
|
NFS_I(inode)->read_io += pgio.pg_bytes_written;
|
2007-04-02 16:48:28 -06:00
|
|
|
npages = (pgio.pg_bytes_written + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
|
|
|
|
nfs_add_stats(inode, NFSIOS_READPAGES, npages);
|
2009-04-03 09:42:44 -06:00
|
|
|
read_complete:
|
2005-04-16 16:20:36 -06:00
|
|
|
put_nfs_open_context(desc.ctx);
|
2006-09-14 12:03:14 -06:00
|
|
|
out:
|
2005-04-16 16:20:36 -06:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
NFS: Split fs/nfs/inode.c
As fs/nfs/inode.c is rather large, heterogenous and unwieldy, the attached
patch splits it up into a number of files:
(*) fs/nfs/inode.c
Strictly inode specific functions.
(*) fs/nfs/super.c
Superblock management functions for NFS and NFS4, normal access, clones
and referrals. The NFS4 superblock functions _could_ move out into a
separate conditionally compiled file, but it's probably not worth it as
there're so many common bits.
(*) fs/nfs/namespace.c
Some namespace-specific functions have been moved here.
(*) fs/nfs/nfs4namespace.c
NFS4-specific namespace functions (this could be merged into the previous
file). This file is conditionally compiled.
(*) fs/nfs/internal.h
Inter-file declarations, plus a few simple utility functions moved from
fs/nfs/inode.c.
Additionally, all the in-.c-file externs have been moved here, and those
files they were moved from now includes this file.
For the most part, the functions have not been changed, only some multiplexor
functions have changed significantly.
I've also:
(*) Added some extra banner comments above some functions.
(*) Rearranged the function order within the files to be more logical and
better grouped (IMO), though someone may prefer a different order.
(*) Reduced the number of #ifdefs in .c files.
(*) Added missing __init and __exit directives.
Signed-Off-By: David Howells <dhowells@redhat.com>
2006-06-09 07:34:33 -06:00
|
|
|
int __init nfs_init_readpagecache(void)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
nfs_rdata_cachep = kmem_cache_create("nfs_read_data",
|
2014-05-06 07:12:27 -06:00
|
|
|
sizeof(struct nfs_rw_header),
|
2005-04-16 16:20:36 -06:00
|
|
|
0, SLAB_HWCACHE_ALIGN,
|
2007-07-19 19:11:58 -06:00
|
|
|
NULL);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (nfs_rdata_cachep == NULL)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-06-27 13:59:15 -06:00
|
|
|
void nfs_destroy_readpagecache(void)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2006-09-27 02:49:40 -06:00
|
|
|
kmem_cache_destroy(nfs_rdata_cachep);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2014-05-06 07:12:30 -06:00
|
|
|
|
|
|
|
static const struct nfs_rw_ops nfs_rw_read_ops = {
|
2014-05-06 07:12:31 -06:00
|
|
|
.rw_mode = FMODE_READ,
|
2014-05-06 07:12:30 -06:00
|
|
|
.rw_alloc_header = nfs_readhdr_alloc,
|
|
|
|
.rw_free_header = nfs_readhdr_free,
|
2014-05-06 07:12:32 -06:00
|
|
|
.rw_done = nfs_readpage_done,
|
|
|
|
.rw_result = nfs_readpage_result,
|
2014-05-06 07:12:37 -06:00
|
|
|
.rw_initiate = nfs_initiate_read,
|
2014-05-06 07:12:30 -06:00
|
|
|
};
|