linux/fs/nfs/read.c
Chuck Lever dc59250c6e [PATCH] NFS: Introduce the use of inode->i_lock to protect fields in nfsi
Down the road we want to eliminate the use of the global kernel lock entirely
from the NFS client.  To do this, we need to protect the fields in the
nfs_inode structure adequately.  Start by serializing updates to the
"cache_validity" field.

Note this change addresses an SMP hang found by njw@osdl.org, where processes
deadlock because nfs_end_data_update and nfs_revalidate_mapping update the
"cache_validity" field without proper serialization.

Test plan:
 Millions of fsx ops on SMP clients.  Run Nick Wilson's breaknfs program on
 large SMP clients.

Signed-off-by: Chuck Lever <cel@netapp.com>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-08-18 12:53:57 -07:00

620 lines
15 KiB
C

/*
* 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
*
* We do an ugly hack here in order to return proper error codes to the
* user program when a read request failed: since generic_file_read
* only checks the return value of inode->i_op->readpage() which is always 0
* for async RPC, we set the error bit of the page to 1 when an error occurs,
* and make nfs_readpage transmit requests synchronously when encountering this.
* This is only a small problem, though, since we now retry all operations
* within the RPC code when root squashing is suspected.
*/
#include <linux/config.h>
#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/smp_lock.h>
#include <asm/system.h>
#define NFSDBG_FACILITY NFSDBG_PAGECACHE
static int nfs_pagein_one(struct list_head *, struct inode *);
static void nfs_readpage_result_partial(struct nfs_read_data *, int);
static void nfs_readpage_result_full(struct nfs_read_data *, int);
static kmem_cache_t *nfs_rdata_cachep;
mempool_t *nfs_rdata_mempool;
#define MIN_POOL_READ (32)
void nfs_readdata_release(struct rpc_task *task)
{
struct nfs_read_data *data = (struct nfs_read_data *)task->tk_calldata;
nfs_readdata_free(data);
}
static
unsigned int nfs_page_length(struct inode *inode, struct page *page)
{
loff_t i_size = i_size_read(inode);
unsigned long idx;
if (i_size <= 0)
return 0;
idx = (i_size - 1) >> PAGE_CACHE_SHIFT;
if (page->index > idx)
return 0;
if (page->index != idx)
return PAGE_CACHE_SIZE;
return 1 + ((i_size - 1) & (PAGE_CACHE_SIZE - 1));
}
static
int nfs_return_empty_page(struct page *page)
{
memclear_highpage_flush(page, 0, PAGE_CACHE_SIZE);
SetPageUptodate(page);
unlock_page(page);
return 0;
}
/*
* Read a page synchronously.
*/
static int nfs_readpage_sync(struct nfs_open_context *ctx, struct inode *inode,
struct page *page)
{
unsigned int rsize = NFS_SERVER(inode)->rsize;
unsigned int count = PAGE_CACHE_SIZE;
int result;
struct nfs_read_data *rdata;
rdata = nfs_readdata_alloc();
if (!rdata)
return -ENOMEM;
memset(rdata, 0, sizeof(*rdata));
rdata->flags = (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
rdata->cred = ctx->cred;
rdata->inode = inode;
INIT_LIST_HEAD(&rdata->pages);
rdata->args.fh = NFS_FH(inode);
rdata->args.context = ctx;
rdata->args.pages = &page;
rdata->args.pgbase = 0UL;
rdata->args.count = rsize;
rdata->res.fattr = &rdata->fattr;
dprintk("NFS: nfs_readpage_sync(%p)\n", page);
/*
* This works now because the socket layer never tries to DMA
* into this buffer directly.
*/
do {
if (count < rsize)
rdata->args.count = count;
rdata->res.count = rdata->args.count;
rdata->args.offset = page_offset(page) + rdata->args.pgbase;
dprintk("NFS: nfs_proc_read(%s, (%s/%Ld), %Lu, %u)\n",
NFS_SERVER(inode)->hostname,
inode->i_sb->s_id,
(long long)NFS_FILEID(inode),
(unsigned long long)rdata->args.pgbase,
rdata->args.count);
lock_kernel();
result = NFS_PROTO(inode)->read(rdata);
unlock_kernel();
/*
* Even if we had a partial success we can't mark the page
* cache valid.
*/
if (result < 0) {
if (result == -EISDIR)
result = -EINVAL;
goto io_error;
}
count -= result;
rdata->args.pgbase += result;
/* Note: result == 0 should only happen if we're caching
* a write that extends the file and punches a hole.
*/
if (rdata->res.eof != 0 || result == 0)
break;
} while (count);
spin_lock(&inode->i_lock);
NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ATIME;
spin_unlock(&inode->i_lock);
if (count)
memclear_highpage_flush(page, rdata->args.pgbase, count);
SetPageUptodate(page);
if (PageError(page))
ClearPageError(page);
result = 0;
io_error:
unlock_page(page);
nfs_readdata_free(rdata);
return result;
}
static int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
struct page *page)
{
LIST_HEAD(one_request);
struct nfs_page *new;
unsigned int len;
len = nfs_page_length(inode, page);
if (len == 0)
return nfs_return_empty_page(page);
new = nfs_create_request(ctx, inode, page, 0, len);
if (IS_ERR(new)) {
unlock_page(page);
return PTR_ERR(new);
}
if (len < PAGE_CACHE_SIZE)
memclear_highpage_flush(page, len, PAGE_CACHE_SIZE - len);
nfs_list_add_request(new, &one_request);
nfs_pagein_one(&one_request, inode);
return 0;
}
static void nfs_readpage_release(struct nfs_page *req)
{
unlock_page(req->wb_page);
nfs_clear_request(req);
nfs_release_request(req);
dprintk("NFS: read done (%s/%Ld %d@%Ld)\n",
req->wb_context->dentry->d_inode->i_sb->s_id,
(long long)NFS_FILEID(req->wb_context->dentry->d_inode),
req->wb_bytes,
(long long)req_offset(req));
}
/*
* Set up the NFS read request struct
*/
static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
unsigned int count, unsigned int offset)
{
struct inode *inode;
data->req = req;
data->inode = inode = req->wb_context->dentry->d_inode;
data->cred = req->wb_context->cred;
data->args.fh = NFS_FH(inode);
data->args.offset = req_offset(req) + offset;
data->args.pgbase = req->wb_pgbase + offset;
data->args.pages = data->pagevec;
data->args.count = count;
data->args.context = req->wb_context;
data->res.fattr = &data->fattr;
data->res.count = count;
data->res.eof = 0;
NFS_PROTO(inode)->read_setup(data);
data->task.tk_cookie = (unsigned long)inode;
data->task.tk_calldata = data;
/* Release requests */
data->task.tk_release = nfs_readdata_release;
dprintk("NFS: %4d initiated read call (req %s/%Ld, %u bytes @ offset %Lu)\n",
data->task.tk_pid,
inode->i_sb->s_id,
(long long)NFS_FILEID(inode),
count,
(unsigned long long)data->args.offset);
}
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);
SetPageError(req->wb_page);
nfs_readpage_release(req);
}
}
/*
* Start an async read operation
*/
static void nfs_execute_read(struct nfs_read_data *data)
{
struct rpc_clnt *clnt = NFS_CLIENT(data->inode);
sigset_t oldset;
rpc_clnt_sigmask(clnt, &oldset);
lock_kernel();
rpc_execute(&data->task);
unlock_kernel();
rpc_clnt_sigunmask(clnt, &oldset);
}
/*
* Generate multiple requests to fill a single page.
*
* We optimize to reduce the number of read operations on the wire. If we
* detect that we're reading a page, or an area of a page, that is past the
* end of file, we do not generate NFS read operations but just clear the
* parts of the page that would have come back zero from the server anyway.
*
* We rely on the cached value of i_size to make this determination; another
* client can fill pages on the server past our cached end-of-file, but we
* won't see the new data until our attribute cache is updated. This is more
* or less conventional NFS client behavior.
*/
static int nfs_pagein_multi(struct list_head *head, struct inode *inode)
{
struct nfs_page *req = nfs_list_entry(head->next);
struct page *page = req->wb_page;
struct nfs_read_data *data;
unsigned int rsize = NFS_SERVER(inode)->rsize;
unsigned int nbytes, offset;
int requests = 0;
LIST_HEAD(list);
nfs_list_remove_request(req);
nbytes = req->wb_bytes;
for(;;) {
data = nfs_readdata_alloc();
if (!data)
goto out_bad;
INIT_LIST_HEAD(&data->pages);
list_add(&data->pages, &list);
requests++;
if (nbytes <= rsize)
break;
nbytes -= rsize;
}
atomic_set(&req->wb_complete, requests);
ClearPageError(page);
offset = 0;
nbytes = req->wb_bytes;
do {
data = list_entry(list.next, struct nfs_read_data, pages);
list_del_init(&data->pages);
data->pagevec[0] = page;
data->complete = nfs_readpage_result_partial;
if (nbytes > rsize) {
nfs_read_rpcsetup(req, data, rsize, offset);
offset += rsize;
nbytes -= rsize;
} else {
nfs_read_rpcsetup(req, data, nbytes, offset);
nbytes = 0;
}
nfs_execute_read(data);
} while (nbytes != 0);
return 0;
out_bad:
while (!list_empty(&list)) {
data = list_entry(list.next, struct nfs_read_data, pages);
list_del(&data->pages);
nfs_readdata_free(data);
}
SetPageError(page);
nfs_readpage_release(req);
return -ENOMEM;
}
static int nfs_pagein_one(struct list_head *head, struct inode *inode)
{
struct nfs_page *req;
struct page **pages;
struct nfs_read_data *data;
unsigned int count;
if (NFS_SERVER(inode)->rsize < PAGE_CACHE_SIZE)
return nfs_pagein_multi(head, inode);
data = nfs_readdata_alloc();
if (!data)
goto out_bad;
INIT_LIST_HEAD(&data->pages);
pages = data->pagevec;
count = 0;
while (!list_empty(head)) {
req = nfs_list_entry(head->next);
nfs_list_remove_request(req);
nfs_list_add_request(req, &data->pages);
ClearPageError(req->wb_page);
*pages++ = req->wb_page;
count += req->wb_bytes;
}
req = nfs_list_entry(data->pages.next);
data->complete = nfs_readpage_result_full;
nfs_read_rpcsetup(req, data, count, 0);
nfs_execute_read(data);
return 0;
out_bad:
nfs_async_read_error(head);
return -ENOMEM;
}
static int
nfs_pagein_list(struct list_head *head, int rpages)
{
LIST_HEAD(one_request);
struct nfs_page *req;
int error = 0;
unsigned int pages = 0;
while (!list_empty(head)) {
pages += nfs_coalesce_requests(head, &one_request, rpages);
req = nfs_list_entry(one_request.next);
error = nfs_pagein_one(&one_request, req->wb_context->dentry->d_inode);
if (error < 0)
break;
}
if (error >= 0)
return pages;
nfs_async_read_error(head);
return error;
}
/*
* Handle a read reply that fills part of a page.
*/
static void nfs_readpage_result_partial(struct nfs_read_data *data, int status)
{
struct nfs_page *req = data->req;
struct page *page = req->wb_page;
if (status >= 0) {
unsigned int request = data->args.count;
unsigned int result = data->res.count;
if (result < request) {
memclear_highpage_flush(page,
data->args.pgbase + result,
request - result);
}
} else
SetPageError(page);
if (atomic_dec_and_test(&req->wb_complete)) {
if (!PageError(page))
SetPageUptodate(page);
nfs_readpage_release(req);
}
}
/*
* This is the callback from RPC telling us whether a reply was
* received or some error occurred (timeout or socket shutdown).
*/
static void nfs_readpage_result_full(struct nfs_read_data *data, int status)
{
unsigned int count = data->res.count;
while (!list_empty(&data->pages)) {
struct nfs_page *req = nfs_list_entry(data->pages.next);
struct page *page = req->wb_page;
nfs_list_remove_request(req);
if (status >= 0) {
if (count < PAGE_CACHE_SIZE) {
if (count < req->wb_bytes)
memclear_highpage_flush(page,
req->wb_pgbase + count,
req->wb_bytes - count);
count = 0;
} else
count -= PAGE_CACHE_SIZE;
SetPageUptodate(page);
} else
SetPageError(page);
nfs_readpage_release(req);
}
}
/*
* This is the callback from RPC telling us whether a reply was
* received or some error occurred (timeout or socket shutdown).
*/
void nfs_readpage_result(struct rpc_task *task)
{
struct nfs_read_data *data = (struct nfs_read_data *)task->tk_calldata;
struct nfs_readargs *argp = &data->args;
struct nfs_readres *resp = &data->res;
int status = task->tk_status;
dprintk("NFS: %4d nfs_readpage_result, (status %d)\n",
task->tk_pid, status);
/* Is this a short read? */
if (task->tk_status >= 0 && resp->count < argp->count && !resp->eof) {
/* Has the server at least made some progress? */
if (resp->count != 0) {
/* Yes, so retry the read at the end of the data */
argp->offset += resp->count;
argp->pgbase += resp->count;
argp->count -= resp->count;
rpc_restart_call(task);
return;
}
task->tk_status = -EIO;
}
spin_lock(&data->inode->i_lock);
NFS_I(data->inode)->cache_validity |= NFS_INO_INVALID_ATIME;
spin_unlock(&data->inode->i_lock);
data->complete(data, status);
}
/*
* 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->mapping->host;
int error;
dprintk("NFS: nfs_readpage (%p %ld@%lu)\n",
page, PAGE_CACHE_SIZE, page->index);
/*
* 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_error;
if (file == NULL) {
ctx = nfs_find_open_context(inode, FMODE_READ);
if (ctx == NULL)
return -EBADF;
} else
ctx = get_nfs_open_context((struct nfs_open_context *)
file->private_data);
if (!IS_SYNC(inode)) {
error = nfs_readpage_async(ctx, inode, page);
goto out;
}
error = nfs_readpage_sync(ctx, inode, page);
if (error < 0 && IS_SWAPFILE(inode))
printk("Aiee.. nfs swap-in of page failed!\n");
out:
put_nfs_open_context(ctx);
return error;
out_error:
unlock_page(page);
return error;
}
struct nfs_readdesc {
struct list_head *head;
struct nfs_open_context *ctx;
};
static int
readpage_async_filler(void *data, struct page *page)
{
struct nfs_readdesc *desc = (struct nfs_readdesc *)data;
struct inode *inode = page->mapping->host;
struct nfs_page *new;
unsigned int len;
nfs_wb_page(inode, page);
len = nfs_page_length(inode, page);
if (len == 0)
return nfs_return_empty_page(page);
new = nfs_create_request(desc->ctx, inode, page, 0, len);
if (IS_ERR(new)) {
SetPageError(page);
unlock_page(page);
return PTR_ERR(new);
}
if (len < PAGE_CACHE_SIZE)
memclear_highpage_flush(page, len, PAGE_CACHE_SIZE - len);
nfs_list_add_request(new, desc->head);
return 0;
}
int nfs_readpages(struct file *filp, struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
LIST_HEAD(head);
struct nfs_readdesc desc = {
.head = &head,
};
struct inode *inode = mapping->host;
struct nfs_server *server = NFS_SERVER(inode);
int ret;
dprintk("NFS: nfs_readpages (%s/%Ld %d)\n",
inode->i_sb->s_id,
(long long)NFS_FILEID(inode),
nr_pages);
if (filp == NULL) {
desc.ctx = nfs_find_open_context(inode, FMODE_READ);
if (desc.ctx == NULL)
return -EBADF;
} else
desc.ctx = get_nfs_open_context((struct nfs_open_context *)
filp->private_data);
ret = read_cache_pages(mapping, pages, readpage_async_filler, &desc);
if (!list_empty(&head)) {
int err = nfs_pagein_list(&head, server->rpages);
if (!ret)
ret = err;
}
put_nfs_open_context(desc.ctx);
return ret;
}
int nfs_init_readpagecache(void)
{
nfs_rdata_cachep = kmem_cache_create("nfs_read_data",
sizeof(struct nfs_read_data),
0, SLAB_HWCACHE_ALIGN,
NULL, NULL);
if (nfs_rdata_cachep == NULL)
return -ENOMEM;
nfs_rdata_mempool = mempool_create(MIN_POOL_READ,
mempool_alloc_slab,
mempool_free_slab,
nfs_rdata_cachep);
if (nfs_rdata_mempool == NULL)
return -ENOMEM;
return 0;
}
void nfs_destroy_readpagecache(void)
{
mempool_destroy(nfs_rdata_mempool);
if (kmem_cache_destroy(nfs_rdata_cachep))
printk(KERN_INFO "nfs_read_data: not all structures were freed\n");
}