linux-stable/fs/nfs/read.c

425 lines
11 KiB
C
Raw Normal View History

/*
* 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>
#include <linux/module.h>
#include "nfs4_fs.h"
#include "internal.h"
#include "iostat.h"
#include "fscache.h"
#include "pnfs.h"
#define NFSDBG_FACILITY NFSDBG_PAGECACHE
static const struct nfs_pgio_completion_ops nfs_async_read_completion_ops;
static const struct nfs_rw_ops nfs_rw_read_ops;
static struct kmem_cache *nfs_rdata_cachep;
static struct nfs_pgio_header *nfs_readhdr_alloc(void)
{
return kmem_cache_zalloc(nfs_rdata_cachep, GFP_KERNEL);
}
static void nfs_readhdr_free(struct nfs_pgio_header *rhdr)
{
kmem_cache_free(nfs_rdata_cachep, rhdr);
}
static
int nfs_return_empty_page(struct page *page)
{
Pagecache zeroing: zero_user_segment, zero_user_segments and zero_user Simplify page cache zeroing of segments of pages through 3 functions zero_user_segments(page, start1, end1, start2, end2) Zeros two segments of the page. It takes the position where to start and end the zeroing which avoids length calculations and makes code clearer. zero_user_segment(page, start, end) Same for a single segment. zero_user(page, start, length) Length variant for the case where we know the length. We remove the zero_user_page macro. Issues: 1. Its a macro. Inline functions are preferable. 2. The KM_USER0 macro is only defined for HIGHMEM. Having to treat this special case everywhere makes the code needlessly complex. The parameter for zeroing is always KM_USER0 except in one single case that we open code. Avoiding KM_USER0 makes a lot of code not having to be dealing with the special casing for HIGHMEM anymore. Dealing with kmap is only necessary for HIGHMEM configurations. In those configurations we use KM_USER0 like we do for a series of other functions defined in highmem.h. Since KM_USER0 is depends on HIGHMEM the existing zero_user_page function could not be a macro. zero_user_* functions introduced here can be be inline because that constant is not used when these functions are called. Also extract the flushing of the caches to be outside of the kmap. [akpm@linux-foundation.org: fix nfs and ntfs build] [akpm@linux-foundation.org: fix ntfs build some more] Signed-off-by: Christoph Lameter <clameter@sgi.com> Cc: Steven French <sfrench@us.ibm.com> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: <linux-ext4@vger.kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Cc: "J. Bruce Fields" <bfields@fieldses.org> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Mark Fasheh <mark.fasheh@oracle.com> Cc: David Chinner <dgc@sgi.com> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: Steven French <sfrench@us.ibm.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 06:28:29 +00:00
zero_user(page, 0, PAGE_CACHE_SIZE);
SetPageUptodate(page);
unlock_page(page);
return 0;
}
void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
struct inode *inode, bool force_mds,
const struct nfs_pgio_completion_ops *compl_ops)
{
struct nfs_server *server = NFS_SERVER(inode);
const struct nfs_pageio_ops *pg_ops = &nfs_pgio_rw_ops;
#ifdef CONFIG_NFS_V4_1
if (server->pnfs_curr_ld && !force_mds)
pg_ops = server->pnfs_curr_ld->pg_read_ops;
#endif
nfs_pageio_init(pgio, inode, pg_ops, compl_ops, &nfs_rw_read_ops,
server->rsize, 0);
}
EXPORT_SYMBOL_GPL(nfs_pageio_init_read);
void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio)
{
pgio->pg_ops = &nfs_pgio_rw_ops;
pgio->pg_bsize = NFS_SERVER(pgio->pg_inode)->rsize;
}
EXPORT_SYMBOL_GPL(nfs_pageio_reset_read_mds);
int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
struct page *page)
{
struct nfs_page *new;
unsigned int len;
struct nfs_pageio_descriptor pgio;
len = nfs_page_length(page);
if (len == 0)
return nfs_return_empty_page(page);
nfs: add support for multiple nfs reqs per page Add "page groups" - a circular list of nfs requests (struct nfs_page) that all reference the same page. This gives nfs read and write paths the ability to account for sub-page regions independently. This somewhat follows the design of struct buffer_head's sub-page accounting. Only "head" requests are ever added/removed from the inode list in the buffered write path. "head" and "sub" requests are treated the same through the read path and the rest of the write/commit path. Requests are given an extra reference across the life of the list. Page groups are never rejoined after being split. If the read/write request fails and the client falls back to another path (ie revert to MDS in PNFS case), the already split requests are pushed through the recoalescing code again, which may split them further and then coalesce them into properly sized requests on the wire. Fragmentation shouldn't be a problem with the current design, because we flush all requests in page group when a non-contiguous request is added, so the only time resplitting should occur is on a resend of a read or write. This patch lays the groundwork for sub-page splitting, but does not actually do any splitting. For now all page groups have one request as pg_test functions don't yet split pages. There are several related patches that are needed support multiple requests per page group. Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-05-15 15:56:45 +00:00
new = nfs_create_request(ctx, page, NULL, 0, len);
if (IS_ERR(new)) {
unlock_page(page);
return PTR_ERR(new);
}
if (len < PAGE_CACHE_SIZE)
Pagecache zeroing: zero_user_segment, zero_user_segments and zero_user Simplify page cache zeroing of segments of pages through 3 functions zero_user_segments(page, start1, end1, start2, end2) Zeros two segments of the page. It takes the position where to start and end the zeroing which avoids length calculations and makes code clearer. zero_user_segment(page, start, end) Same for a single segment. zero_user(page, start, length) Length variant for the case where we know the length. We remove the zero_user_page macro. Issues: 1. Its a macro. Inline functions are preferable. 2. The KM_USER0 macro is only defined for HIGHMEM. Having to treat this special case everywhere makes the code needlessly complex. The parameter for zeroing is always KM_USER0 except in one single case that we open code. Avoiding KM_USER0 makes a lot of code not having to be dealing with the special casing for HIGHMEM anymore. Dealing with kmap is only necessary for HIGHMEM configurations. In those configurations we use KM_USER0 like we do for a series of other functions defined in highmem.h. Since KM_USER0 is depends on HIGHMEM the existing zero_user_page function could not be a macro. zero_user_* functions introduced here can be be inline because that constant is not used when these functions are called. Also extract the flushing of the caches to be outside of the kmap. [akpm@linux-foundation.org: fix nfs and ntfs build] [akpm@linux-foundation.org: fix ntfs build some more] Signed-off-by: Christoph Lameter <clameter@sgi.com> Cc: Steven French <sfrench@us.ibm.com> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: <linux-ext4@vger.kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Cc: "J. Bruce Fields" <bfields@fieldses.org> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Mark Fasheh <mark.fasheh@oracle.com> Cc: David Chinner <dgc@sgi.com> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: Steven French <sfrench@us.ibm.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 06:28:29 +00:00
zero_user_segment(page, len, PAGE_CACHE_SIZE);
nfs_pageio_init_read(&pgio, inode, false,
&nfs_async_read_completion_ops);
nfs_pageio_add_request(&pgio, new);
nfs_pageio_complete(&pgio);
NFS_I(inode)->read_io += pgio.pg_bytes_written;
return 0;
}
static void nfs_readpage_release(struct nfs_page *req)
{
struct inode *d_inode = req->wb_context->dentry->d_inode;
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));
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);
}
nfs_release_request(req);
}
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);
}
static void nfs_read_completion(struct nfs_pgio_header *hdr)
{
unsigned long bytes = 0;
if (test_bit(NFS_IOHDR_REDO, &hdr->flags))
goto out;
while (!list_empty(&hdr->pages)) {
struct nfs_page *req = nfs_list_entry(hdr->pages.next);
struct page *page = req->wb_page;
unsigned long start = req->wb_pgbase;
unsigned long end = req->wb_pgbase + req->wb_bytes;
if (test_bit(NFS_IOHDR_EOF, &hdr->flags)) {
/* 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);
}
}
bytes += req->wb_bytes;
if (test_bit(NFS_IOHDR_ERROR, &hdr->flags)) {
if (bytes <= hdr->good_bytes)
nfs_page_group_set_uptodate(req);
} else
nfs_page_group_set_uptodate(req);
nfs_list_remove_request(req);
nfs_readpage_release(req);
}
out:
hdr->release(hdr);
}
static void nfs_initiate_read(struct nfs_pgio_header *hdr,
struct rpc_message *msg,
struct rpc_task_setup *task_setup_data, int how)
{
struct inode *inode = hdr->inode;
int swap_flags = IS_SWAPFILE(inode) ? NFS_RPC_SWAPFLAGS : 0;
task_setup_data->flags |= swap_flags;
NFS_PROTO(inode)->read_setup(hdr, msg);
}
static void
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);
}
}
static const struct nfs_pgio_completion_ops nfs_async_read_completion_ops = {
.error_cleanup = nfs_async_read_error,
.completion = nfs_read_completion,
};
/*
* This is the callback from RPC telling us whether a reply was
* received or some error occurred (timeout or socket shutdown).
*/
static int nfs_readpage_done(struct rpc_task *task,
struct nfs_pgio_header *hdr,
struct inode *inode)
{
int status = NFS_PROTO(inode)->read_done(task, hdr);
if (status != 0)
return status;
nfs_add_stats(inode, NFSIOS_SERVERREADBYTES, hdr->res.count);
if (task->tk_status == -ESTALE) {
set_bit(NFS_INO_STALE, &NFS_I(inode)->flags);
nfs_mark_for_revalidate(inode);
}
return 0;
}
static void nfs_readpage_retry(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
struct nfs_pgio_args *argp = &hdr->args;
struct nfs_pgio_res *resp = &hdr->res;
/* This is a short read! */
nfs_inc_stats(hdr->inode, NFSIOS_SHORTREAD);
/* Has the server at least made some progress? */
if (resp->count == 0) {
nfs_set_pgio_error(hdr, -EIO, argp->offset);
return;
}
/* Yes, so retry the read at the end of the hdr */
hdr->mds_offset += resp->count;
argp->offset += resp->count;
argp->pgbase += resp->count;
argp->count -= resp->count;
rpc_restart_call_prepare(task);
}
static void nfs_readpage_result(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (hdr->res.eof) {
loff_t bound;
bound = hdr->args.offset + hdr->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 (hdr->res.count != hdr->args.count)
nfs_readpage_retry(task, hdr);
}
/*
* 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;
struct inode *inode = page_file_mapping(page)->host;
int error;
dprintk("NFS: nfs_readpage (%p %ld@%lu)\n",
page, PAGE_CACHE_SIZE, page_file_index(page));
nfs_inc_stats(inode, NFSIOS_VFSREADPAGE);
nfs_inc_stats(inode, NFSIOS_READPAGES);
/*
* 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)
goto out_unlock;
if (PageUptodate(page))
goto out_unlock;
error = -ESTALE;
if (NFS_STALE(inode))
goto out_unlock;
if (file == NULL) {
error = -EBADF;
ctx = nfs_find_open_context(inode, NULL, FMODE_READ);
if (ctx == NULL)
goto out_unlock;
} else
ctx = get_nfs_open_context(nfs_file_open_context(file));
if (!IS_SYNC(inode)) {
error = nfs_readpage_from_fscache(ctx, inode, page);
if (error == 0)
goto out;
}
error = nfs_readpage_async(ctx, inode, page);
out:
put_nfs_open_context(ctx);
return error;
out_unlock:
unlock_page(page);
return error;
}
struct nfs_readdesc {
struct nfs_pageio_descriptor *pgio;
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;
int error;
len = nfs_page_length(page);
if (len == 0)
return nfs_return_empty_page(page);
nfs: add support for multiple nfs reqs per page Add "page groups" - a circular list of nfs requests (struct nfs_page) that all reference the same page. This gives nfs read and write paths the ability to account for sub-page regions independently. This somewhat follows the design of struct buffer_head's sub-page accounting. Only "head" requests are ever added/removed from the inode list in the buffered write path. "head" and "sub" requests are treated the same through the read path and the rest of the write/commit path. Requests are given an extra reference across the life of the list. Page groups are never rejoined after being split. If the read/write request fails and the client falls back to another path (ie revert to MDS in PNFS case), the already split requests are pushed through the recoalescing code again, which may split them further and then coalesce them into properly sized requests on the wire. Fragmentation shouldn't be a problem with the current design, because we flush all requests in page group when a non-contiguous request is added, so the only time resplitting should occur is on a resend of a read or write. This patch lays the groundwork for sub-page splitting, but does not actually do any splitting. For now all page groups have one request as pg_test functions don't yet split pages. There are several related patches that are needed support multiple requests per page group. Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-05-15 15:56:45 +00:00
new = nfs_create_request(desc->ctx, page, NULL, 0, len);
if (IS_ERR(new))
goto out_error;
if (len < PAGE_CACHE_SIZE)
Pagecache zeroing: zero_user_segment, zero_user_segments and zero_user Simplify page cache zeroing of segments of pages through 3 functions zero_user_segments(page, start1, end1, start2, end2) Zeros two segments of the page. It takes the position where to start and end the zeroing which avoids length calculations and makes code clearer. zero_user_segment(page, start, end) Same for a single segment. zero_user(page, start, length) Length variant for the case where we know the length. We remove the zero_user_page macro. Issues: 1. Its a macro. Inline functions are preferable. 2. The KM_USER0 macro is only defined for HIGHMEM. Having to treat this special case everywhere makes the code needlessly complex. The parameter for zeroing is always KM_USER0 except in one single case that we open code. Avoiding KM_USER0 makes a lot of code not having to be dealing with the special casing for HIGHMEM anymore. Dealing with kmap is only necessary for HIGHMEM configurations. In those configurations we use KM_USER0 like we do for a series of other functions defined in highmem.h. Since KM_USER0 is depends on HIGHMEM the existing zero_user_page function could not be a macro. zero_user_* functions introduced here can be be inline because that constant is not used when these functions are called. Also extract the flushing of the caches to be outside of the kmap. [akpm@linux-foundation.org: fix nfs and ntfs build] [akpm@linux-foundation.org: fix ntfs build some more] Signed-off-by: Christoph Lameter <clameter@sgi.com> Cc: Steven French <sfrench@us.ibm.com> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: <linux-ext4@vger.kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Cc: "J. Bruce Fields" <bfields@fieldses.org> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Mark Fasheh <mark.fasheh@oracle.com> Cc: David Chinner <dgc@sgi.com> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: Steven French <sfrench@us.ibm.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 06:28:29 +00:00
zero_user_segment(page, len, PAGE_CACHE_SIZE);
2008-03-19 15:24:39 +00:00
if (!nfs_pageio_add_request(desc->pgio, new)) {
error = desc->pgio->pg_error;
goto out_unlock;
}
return 0;
out_error:
error = PTR_ERR(new);
out_unlock:
unlock_page(page);
return error;
}
int nfs_readpages(struct file *filp, struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
struct nfs_pageio_descriptor pgio;
struct nfs_readdesc desc = {
.pgio = &pgio,
};
struct inode *inode = mapping->host;
unsigned long npages;
int ret = -ESTALE;
dprintk("NFS: nfs_readpages (%s/%Lu %d)\n",
inode->i_sb->s_id,
(unsigned long long)NFS_FILEID(inode),
nr_pages);
nfs_inc_stats(inode, NFSIOS_VFSREADPAGES);
if (NFS_STALE(inode))
goto out;
if (filp == NULL) {
desc.ctx = nfs_find_open_context(inode, NULL, FMODE_READ);
if (desc.ctx == NULL)
return -EBADF;
} else
desc.ctx = get_nfs_open_context(nfs_file_open_context(filp));
/* 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 */
nfs_pageio_init_read(&pgio, inode, false,
&nfs_async_read_completion_ops);
ret = read_cache_pages(mapping, pages, readpage_async_filler, &desc);
nfs_pageio_complete(&pgio);
NFS_I(inode)->read_io += pgio.pg_bytes_written;
npages = (pgio.pg_bytes_written + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
nfs_add_stats(inode, NFSIOS_READPAGES, npages);
read_complete:
put_nfs_open_context(desc.ctx);
out:
return ret;
}
int __init nfs_init_readpagecache(void)
{
nfs_rdata_cachep = kmem_cache_create("nfs_read_data",
sizeof(struct nfs_pgio_header),
0, SLAB_HWCACHE_ALIGN,
NULL);
if (nfs_rdata_cachep == NULL)
return -ENOMEM;
return 0;
}
void nfs_destroy_readpagecache(void)
{
kmem_cache_destroy(nfs_rdata_cachep);
}
static const struct nfs_rw_ops nfs_rw_read_ops = {
.rw_mode = FMODE_READ,
.rw_alloc_header = nfs_readhdr_alloc,
.rw_free_header = nfs_readhdr_free,
.rw_done = nfs_readpage_done,
.rw_result = nfs_readpage_result,
.rw_initiate = nfs_initiate_read,
};