a pre-requisite for functioning O_DIRECT support for NFS. with this
patch: for each request, the client now stores the in-memory page offset
in a separate field from the file page offset. this allows direct I/O to
target user buffers at arbitrary offsets.
against 2.5.48.
diff -ruN 02-req_offset/fs/nfs/nfs3proc.c 03-wb_base/fs/nfs/nfs3proc.c
--- 02-req_offset/fs/nfs/nfs3proc.c Mon Nov 18 11:46:05 2002
+++ 03-wb_base/fs/nfs/nfs3proc.c Mon Nov 18 11:50:13 2002
@@ -709,7 +709,7 @@
req = nfs_list_entry(data->pages.next);
data->u.v3.args.fh = NFS_FH(inode);
data->u.v3.args.offset = req_offset(req);
- data->u.v3.args.pgbase = req->wb_offset;
+ data->u.v3.args.pgbase = req->wb_base;
data->u.v3.args.pages = data->pagevec;
data->u.v3.args.count = count;
data->u.v3.res.fattr = &data->fattr;
@@ -765,7 +765,7 @@
req = nfs_list_entry(data->pages.next);
data->u.v3.args.fh = NFS_FH(inode);
data->u.v3.args.offset = req_offset(req);
- data->u.v3.args.pgbase = req->wb_offset;
+ data->u.v3.args.pgbase = req->wb_base;
data->u.v3.args.count = count;
data->u.v3.args.stable = stable;
data->u.v3.args.pages = data->pagevec;
diff -ruN 02-req_offset/fs/nfs/nfs4proc.c 03-wb_base/fs/nfs/nfs4proc.c
--- 02-req_offset/fs/nfs/nfs4proc.c Mon Nov 18 11:46:05 2002
+++ 03-wb_base/fs/nfs/nfs4proc.c Mon Nov 18 11:50:13 2002
@@ -1380,7 +1380,7 @@
nfs4_setup_compound(cp, data->u.v4.ops, NFS_SERVER(inode), "read [async]");
nfs4_setup_putfh(cp, NFS_FH(inode));
nfs4_setup_read(cp, req_offset(req),
- count, data->pagevec, req->wb_offset,
+ count, data->pagevec, req->wb_base,
&data->u.v4.res_eof,
&data->u.v4.res_count);
@@ -1433,7 +1433,7 @@
nfs4_setup_compound(cp, data->u.v4.ops, NFS_SERVER(inode), "write [async]");
nfs4_setup_putfh(cp, NFS_FH(inode));
nfs4_setup_write(cp, req_offset(req),
- count, stable, data->pagevec, req->wb_offset,
+ count, stable, data->pagevec, req->wb_base,
&data->u.v4.res_count, &data->verf);
/* Set the initial flags for the task. */
diff -ruN 02-req_offset/fs/nfs/pagelist.c 03-wb_base/fs/nfs/pagelist.c
--- 02-req_offset/fs/nfs/pagelist.c Sun Nov 17 23:29:54 2002
+++ 03-wb_base/fs/nfs/pagelist.c Mon Nov 18 11:50:13 2002
@@ -91,6 +91,7 @@
req->wb_index = page->index;
page_cache_get(page);
req->wb_offset = offset;
+ req->wb_base = offset;
req->wb_bytes = count;
if (cred)
@@ -277,14 +278,13 @@
break;
if (req->wb_index != (prev->wb_index + 1))
break;
-
- if (req->wb_offset != 0)
+ if (req->wb_base != 0)
break;
}
nfs_list_remove_request(req);
nfs_list_add_request(req, dst);
npages++;
- if (req->wb_offset + req->wb_bytes != PAGE_CACHE_SIZE)
+ if (req->wb_base + req->wb_bytes != PAGE_CACHE_SIZE)
break;
if (npages >= nmax)
break;
diff -ruN 02-req_offset/fs/nfs/proc.c 03-wb_base/fs/nfs/proc.c
--- 02-req_offset/fs/nfs/proc.c Mon Nov 18 11:46:05 2002
+++ 03-wb_base/fs/nfs/proc.c Mon Nov 18 11:50:13 2002
@@ -544,7 +544,7 @@
req = nfs_list_entry(data->pages.next);
data->u.v3.args.fh = NFS_FH(inode);
data->u.v3.args.offset = req_offset(req);
- data->u.v3.args.pgbase = req->wb_offset;
+ data->u.v3.args.pgbase = req->wb_base;
data->u.v3.args.pages = data->pagevec;
data->u.v3.args.count = count;
data->u.v3.res.fattr = &data->fattr;
@@ -590,7 +590,7 @@
req = nfs_list_entry(data->pages.next);
data->u.v3.args.fh = NFS_FH(inode);
data->u.v3.args.offset = req_offset(req);
- data->u.v3.args.pgbase = req->wb_offset;
+ data->u.v3.args.pgbase = req->wb_base;
data->u.v3.args.count = count;
data->u.v3.args.stable = NFS_FILE_SYNC;
data->u.v3.args.pages = data->pagevec;
diff -ruN 02-req_offset/fs/nfs/read.c 03-wb_base/fs/nfs/read.c
--- 02-req_offset/fs/nfs/read.c Mon Nov 18 11:46:05 2002
+++ 03-wb_base/fs/nfs/read.c Mon Nov 18 11:50:13 2002
@@ -269,7 +269,7 @@
if (task->tk_status >= 0) {
if (count < PAGE_CACHE_SIZE) {
memclear_highpage_flush(page,
- req->wb_offset + count,
+ req->wb_base + count,
req->wb_bytes - count);
if (eof ||
diff -ruN 02-req_offset/fs/nfs/write.c 03-wb_base/fs/nfs/write.c
--- 02-req_offset/fs/nfs/write.c Mon Nov 18 11:46:05 2002
+++ 03-wb_base/fs/nfs/write.c Mon Nov 18 11:50:13 2002
@@ -586,6 +586,7 @@
/* Okay, the request matches. Update the region */
if (offset < req->wb_offset) {
req->wb_offset = offset;
+ req->wb_base = offset;
req->wb_bytes = rqend - req->wb_offset;
}
@@ -718,7 +719,7 @@
* Call the strategy routine so it can send out a bunch
* of requests.
*/
- if (req->wb_offset == 0 && req->wb_bytes == PAGE_CACHE_SIZE) {
+ if (req->wb_base == 0 && req->wb_bytes == PAGE_CACHE_SIZE) {
SetPageUptodate(page);
nfs_unlock_request(req);
nfs_strategy(inode);
diff -ruN 02-req_offset/include/linux/nfs_page.h 03-wb_base/include/linux/nfs_page.h
--- 02-req_offset/include/linux/nfs_page.h Sun Nov 17 23:29:22 2002
+++ 03-wb_base/include/linux/nfs_page.h Mon Nov 18 11:50:13 2002
@@ -31,7 +31,8 @@
struct page *wb_page; /* page to read in/write out */
wait_queue_head_t wb_wait; /* wait queue */
unsigned long wb_index; /* Offset within mapping */
- unsigned int wb_offset, /* Offset within page */
+ unsigned int wb_offset, /* Offset within file page */
+ wb_base, /* Offset within buffer page */
wb_bytes, /* Length of request */
wb_count; /* reference count */
unsigned long wb_flags;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Nov 23 2002 - 22:00:24 EST