Introduction - enhancment of support for larger rsize/wsize

diff .prev/fs/nfsd/nfs2acl.c ./fs/nfsd/nfs2acl.c
@@ -241,7 +241,7 @@ static int nfsaclsvc_encode_getaclres(st
 
 	rqstp->rq_res.page_len = w;
 	while (w > 0) {
-		if (!svc_take_res_page(rqstp))
+		if (!rqstp->rq_respages[rqstp->rq_resused++])
 			return 0;
 		w -= PAGE_SIZE;
 	}

diff .prev/fs/nfsd/nfs3acl.c ./fs/nfsd/nfs3acl.c
@@ -185,7 +185,7 @@ static int nfs3svc_encode_getaclres(stru
 
 		rqstp->rq_res.page_len = w;
 		while (w > 0) {
-			if (!svc_take_res_page(rqstp))
+			if (!rqstp->rq_respages[rqstp->rq_resused++])
 				return 0;
 			w -= PAGE_SIZE;
 		}

diff .prev/fs/nfsd/nfs3xdr.c ./fs/nfsd/nfs3xdr.c
@@ -343,8 +343,7 @@ nfs3svc_decode_readargs(struct svc_rqst 
 	/* set up the kvec */
 	v=0;
 	while (len > 0) {
-		pn = rqstp->rq_resused;
-		svc_take_page(rqstp);
+		pn = rqstp->rq_resused++;
 		args->vec[v].iov_base = page_address(rqstp->rq_respages[pn]);
 		args->vec[v].iov_len = len < PAGE_SIZE? len : PAGE_SIZE;
 		len -= args->vec[v].iov_len;

@@ -382,7 +381,7 @@ nfs3svc_decode_writeargs(struct svc_rqst
 	while (len > args->vec[v].iov_len) {
 		len -= args->vec[v].iov_len;
 		v++;
-		args->vec[v].iov_base = page_address(rqstp->rq_argpages[v]);
+		args->vec[v].iov_base = page_address(rqstp->rq_pages[v]);
 		args->vec[v].iov_len = PAGE_SIZE;
 	}
 	args->vec[v].iov_len = len;

@@ -446,11 +445,11 @@ nfs3svc_decode_symlinkargs(struct svc_rq
 	 * This page appears in the rq_res.pages list, but as pages_len is always
 	 * 0, it won't get in the way
 	 */
-	svc_take_page(rqstp);
 	len = ntohl(*p++);
 	if (len == 0 || len > NFS3_MAXPATHLEN || len >= PAGE_SIZE)
 		return 0;
-	args->tname = new = page_address(rqstp->rq_respages[rqstp->rq_resused-1]);
+	args->tname = new =
+		page_address(rqstp->rq_respages[rqstp->rq_resused++]);
 	args->tlen = len;
 	/* first copy and check from the first page */
 	old = (char*)p;

@@ -522,8 +521,8 @@ nfs3svc_decode_readlinkargs(struct svc_r
 {
 	if (!(p = decode_fh(p, &args->fh)))
 		return 0;
-	svc_take_page(rqstp);
-	args->buffer = page_address(rqstp->rq_respages[rqstp->rq_resused-1]);
+	args->buffer =
+		page_address(rqstp->rq_respages[rqstp->rq_resused++]);
 
 	return xdr_argsize_check(rqstp, p);
 }

@@ -554,8 +553,8 @@ nfs3svc_decode_readdirargs(struct svc_rq
 	if (args->count > PAGE_SIZE)
 		args->count = PAGE_SIZE;
 
-	svc_take_page(rqstp);
-	args->buffer = page_address(rqstp->rq_respages[rqstp->rq_resused-1]);
+	args->buffer =
+		page_address(rqstp->rq_respages[rqstp->rq_resused++]);
 
 	return xdr_argsize_check(rqstp, p);
 }

@@ -578,8 +577,7 @@ nfs3svc_decode_readdirplusargs(struct sv
 	args->count = len;
 
 	while (len > 0) {
-		pn = rqstp->rq_resused;
-		svc_take_page(rqstp);
+		pn = rqstp->rq_resused++;
 		if (!args->buffer)
 			args->buffer = page_address(rqstp->rq_respages[pn]);
 		len -= PAGE_SIZE;

@@ -668,7 +666,6 @@ nfs3svc_encode_readlinkres(struct svc_rq
 		rqstp->rq_res.page_len = resp->len;
 		if (resp->len & 3) {
 			/* need to pad the tail */
-			rqstp->rq_restailpage = 0;
 			rqstp->rq_res.tail[0].iov_base = p;
 			*p = 0;
 			rqstp->rq_res.tail[0].iov_len = 4 - (resp->len&3);

@@ -693,7 +690,6 @@ nfs3svc_encode_readres(struct svc_rqst *
 		rqstp->rq_res.page_len = resp->count;
 		if (resp->count & 3) {
 			/* need to pad the tail */
-			rqstp->rq_restailpage = 0;
 			rqstp->rq_res.tail[0].iov_base = p;
 			*p = 0;
 			rqstp->rq_res.tail[0].iov_len = 4 - (resp->count & 3);

@@ -768,7 +764,6 @@ nfs3svc_encode_readdirres(struct svc_rqs
 		rqstp->rq_res.page_len = (resp->count) << 2;
 
 		/* add the 'tail' to the end of the 'head' page - page 0. */
-		rqstp->rq_restailpage = 0;
 		rqstp->rq_res.tail[0].iov_base = p;
 		*p++ = 0;		/* no more entries */
 		*p++ = htonl(resp->common.err == nfserr_eof);

diff .prev/fs/nfsd/nfs4xdr.c ./fs/nfsd/nfs4xdr.c
@@ -2040,7 +2040,8 @@ nfsd4_encode_open_downgrade(struct nfsd4
 }
 
 static int
-nfsd4_encode_read(struct nfsd4_compoundres *resp, int nfserr, struct nfsd4_read *read)
+nfsd4_encode_read(struct nfsd4_compoundres *resp, int nfserr,
+		  struct nfsd4_read *read)
 {
 	u32 eof;
 	int v, pn;

@@ -2062,10 +2063,11 @@ nfsd4_encode_read(struct nfsd4_compoundr
 	len = maxcount;
 	v = 0;
 	while (len > 0) {
-		pn = resp->rqstp->rq_resused;
-		svc_take_page(resp->rqstp);
-		read->rd_iov[v].iov_base = page_address(resp->rqstp->rq_respages[pn]);
-		read->rd_iov[v].iov_len = len < PAGE_SIZE ? len : PAGE_SIZE;
+		pn = resp->rqstp->rq_resused++;
+		read->rd_iov[v].iov_base =
+			page_address(resp->rqstp->rq_respages[pn]);
+		read->rd_iov[v].iov_len =
+			len < PAGE_SIZE ? len : PAGE_SIZE;
 		v++;
 		len -= PAGE_SIZE;
 	}

@@ -2079,7 +2081,8 @@ nfsd4_encode_read(struct nfsd4_compoundr
 		nfserr = nfserr_inval;
 	if (nfserr)
 		return nfserr;
-	eof = (read->rd_offset + maxcount >= read->rd_fhp->fh_dentry->d_inode->i_size);
+	eof = (read->rd_offset + maxcount >=
+	       read->rd_fhp->fh_dentry->d_inode->i_size);
 
 	WRITE32(eof);
 	WRITE32(maxcount);

@@ -2089,7 +2092,6 @@ nfsd4_encode_read(struct nfsd4_compoundr
 	resp->xbuf->page_len = maxcount;
 
 	/* Use rest of head for padding and remaining ops: */
-	resp->rqstp->rq_restailpage = 0;
 	resp->xbuf->tail[0].iov_base = p;
 	resp->xbuf->tail[0].iov_len = 0;
 	if (maxcount&3) {

@@ -2114,8 +2116,7 @@ nfsd4_encode_readlink(struct nfsd4_compo
 	if (resp->xbuf->page_len)
 		return nfserr_resource;
 
-	svc_take_page(resp->rqstp);
-	page = page_address(resp->rqstp->rq_respages[resp->rqstp->rq_resused-1]);
+	page = page_address(resp->rqstp->rq_respages[resp->rqstp->rq_resused++]);
 
 	maxcount = PAGE_SIZE;
 	RESERVE_SPACE(4);

@@ -2139,7 +2140,6 @@ nfsd4_encode_readlink(struct nfsd4_compo
 	resp->xbuf->page_len = maxcount;
 
 	/* Use rest of head for padding and remaining ops: */
-	resp->rqstp->rq_restailpage = 0;
 	resp->xbuf->tail[0].iov_base = p;
 	resp->xbuf->tail[0].iov_len = 0;
 	if (maxcount&3) {

@@ -2190,8 +2190,7 @@ nfsd4_encode_readdir(struct nfsd4_compou
 		goto err_no_verf;
 	}
 
-	svc_take_page(resp->rqstp);
-	page = page_address(resp->rqstp->rq_respages[resp->rqstp->rq_resused-1]);
+	page = page_address(resp->rqstp->rq_respages[resp->rqstp->rq_resused++]);
 	readdir->common.err = 0;
 	readdir->buflen = maxcount;
 	readdir->buffer = page;

@@ -2216,10 +2215,10 @@ nfsd4_encode_readdir(struct nfsd4_compou
 	p = readdir->buffer;
 	*p++ = 0;	/* no more entries */
 	*p++ = htonl(readdir->common.err == nfserr_eof);
-	resp->xbuf->page_len = ((char*)p) - (char*)page_address(resp->rqstp->rq_respages[resp->rqstp->rq_resused-1]);
+	resp->xbuf->page_len = ((char*)p) - (char*)page_address(
+		resp->rqstp->rq_respages[resp->rqstp->rq_resused-1]);
 
 	/* Use rest of head for padding and remaining ops: */
-	resp->rqstp->rq_restailpage = 0;
 	resp->xbuf->tail[0].iov_base = tailbase;
 	resp->xbuf->tail[0].iov_len = 0;
 	resp->p = resp->xbuf->tail[0].iov_base;

diff .prev/fs/nfsd/nfsxdr.c ./fs/nfsd/nfsxdr.c
@@ -262,8 +262,7 @@ nfssvc_decode_readargs(struct svc_rqst *
 	 */
 	v=0;
 	while (len > 0) {
-		pn=rqstp->rq_resused;
-		svc_take_page(rqstp);
+		pn = rqstp->rq_resused++;
 		args->vec[v].iov_base = page_address(rqstp->rq_respages[pn]);
 		args->vec[v].iov_len = len < PAGE_SIZE?len:PAGE_SIZE;
 		len -= args->vec[v].iov_len;

@@ -295,7 +294,7 @@ nfssvc_decode_writeargs(struct svc_rqst 
 	while (len > args->vec[v].iov_len) {
 		len -= args->vec[v].iov_len;
 		v++;
-		args->vec[v].iov_base = page_address(rqstp->rq_argpages[v]);
+		args->vec[v].iov_base = page_address(rqstp->rq_pages[v]);
 		args->vec[v].iov_len = PAGE_SIZE;
 	}
 	args->vec[v].iov_len = len;

@@ -333,8 +332,7 @@ nfssvc_decode_readlinkargs(struct svc_rq
 {
 	if (!(p = decode_fh(p, &args->fh)))
 		return 0;
-	svc_take_page(rqstp);
-	args->buffer = page_address(rqstp->rq_respages[rqstp->rq_resused-1]);
+	args->buffer = page_address(rqstp->rq_respages[rqstp->rq_resused++]);
 
 	return xdr_argsize_check(rqstp, p);
 }

@@ -375,8 +373,7 @@ nfssvc_decode_readdirargs(struct svc_rqs
 	if (args->count > PAGE_SIZE)
 		args->count = PAGE_SIZE;
 
-	svc_take_page(rqstp);
-	args->buffer = page_address(rqstp->rq_respages[rqstp->rq_resused-1]);
+	args->buffer = page_address(rqstp->rq_respages[rqstp->rq_resused++]);
 
 	return xdr_argsize_check(rqstp, p);
 }

@@ -416,7 +413,6 @@ nfssvc_encode_readlinkres(struct svc_rqs
 	rqstp->rq_res.page_len = resp->len;
 	if (resp->len & 3) {
 		/* need to pad the tail */
-		rqstp->rq_restailpage = 0;
 		rqstp->rq_res.tail[0].iov_base = p;
 		*p = 0;
 		rqstp->rq_res.tail[0].iov_len = 4 - (resp->len&3);

@@ -436,7 +432,6 @@ nfssvc_encode_readres(struct svc_rqst *r
 	rqstp->rq_res.page_len = resp->count;
 	if (resp->count & 3) {
 		/* need to pad the tail */
-		rqstp->rq_restailpage = 0;
 		rqstp->rq_res.tail[0].iov_base = p;
 		*p = 0;
 		rqstp->rq_res.tail[0].iov_len = 4 - (resp->count&3);

diff .prev/fs/nfsd/vfs.c ./fs/nfsd/vfs.c
@@ -791,22 +791,26 @@ nfsd_read_actor(read_descriptor_t *desc,
 {
 	unsigned long count = desc->count;
 	struct svc_rqst *rqstp = desc->arg.data;
+	struct page **pp = rqstp->rq_respages + rqstp->rq_resused;
 
 	if (size > count)
 		size = count;
 
 	if (rqstp->rq_res.page_len == 0) {
 		get_page(page);
-		rqstp->rq_respages[rqstp->rq_resused++] = page;
+		put_page(*pp);
+		*pp = page;
+		rqstp->rq_resused++;
 		rqstp->rq_res.page_base = offset;
 		rqstp->rq_res.page_len = size;
-	} else if (page != rqstp->rq_respages[rqstp->rq_resused-1]) {
+	} else if (page != pp[-1]) {
 		get_page(page);
-		rqstp->rq_respages[rqstp->rq_resused++] = page;
+		put_page(*pp);
+		*pp = page;
+		rqstp->rq_resused++;
 		rqstp->rq_res.page_len += size;
-	} else {
+	} else
 		rqstp->rq_res.page_len += size;
-	}
 
 	desc->count = count - size;
 	desc->written += size;

@@ -840,7 +844,7 @@ nfsd_vfs_read(struct svc_rqst *rqstp, st
 		file->f_ra = ra->p_ra;
 
 	if (file->f_op->sendfile && rqstp->rq_sendfile_ok) {
-		svc_pushback_unused_pages(rqstp);
+		rqstp->rq_resused = 1;
 		err = file->f_op->sendfile(file, &offset, *count,
 						 nfsd_read_actor, rqstp);
 	} else {

92 lines not on this page · raw message

diff .prev/net/sunrpc/auth_gss/svcauth_gss.c ./net/sunrpc/auth_gss/svcauth_gss.c
@@ -1191,7 +1191,6 @@ svcauth_gss_wrap_resp_integ(struct svc_r
 		resbuf->tail[0].iov_base = resbuf->head[0].iov_base
 						+ resbuf->head[0].iov_len;
 		resbuf->tail[0].iov_len = 0;
-		rqstp->rq_restailpage = 0;
 		resv = &resbuf->tail[0];
 	} else {
 		resv = &resbuf->tail[0];

@@ -1240,7 +1239,7 @@ svcauth_gss_wrap_resp_priv(struct svc_rq
 	inpages = resbuf->pages;
 	/* XXX: Would be better to write some xdr helper functions for
 	 * nfs{2,3,4}xdr.c that place the data right, instead of copying: */
-	if (resbuf->tail[0].iov_base && rqstp->rq_restailpage == 0) {
+	if (resbuf->tail[0].iov_base) {
 		BUG_ON(resbuf->tail[0].iov_base >= resbuf->head[0].iov_base
 							+ PAGE_SIZE);
 		BUG_ON(resbuf->tail[0].iov_base < resbuf->head[0].iov_base);

@@ -1258,7 +1257,6 @@ svcauth_gss_wrap_resp_priv(struct svc_rq
 		resbuf->tail[0].iov_base = resbuf->head[0].iov_base
 			+ resbuf->head[0].iov_len + RPC_MAX_AUTH_SIZE;
 		resbuf->tail[0].iov_len = 0;
-		rqstp->rq_restailpage = 0;
 	}
 	if (gss_wrap(gsd->rsci->mechctx, offset, resbuf, inpages))
 		return -ENOMEM;

diff .prev/net/sunrpc/svc.c ./net/sunrpc/svc.c
@@ -408,18 +408,15 @@ svc_init_buffer(struct svc_rqst *rqstp, 
 	if (size > RPCSVC_MAXPAYLOAD)
 		size = RPCSVC_MAXPAYLOAD;
 	pages = 2 + (size+ PAGE_SIZE -1) / PAGE_SIZE;
-	rqstp->rq_argused = 0;
-	rqstp->rq_resused = 0;
 	arghi = 0;
 	BUG_ON(pages > RPCSVC_MAXPAGES);
 	while (pages) {
 		struct page *p = alloc_page(GFP_KERNEL);
 		if (!p)
 			break;
-		rqstp->rq_argpages[arghi++] = p;
+		rqstp->rq_pages[arghi++] = p;
 		pages--;
 	}
-	rqstp->rq_arghi = arghi;
 	return ! pages;
 }
 

@@ -429,14 +426,10 @@ svc_init_buffer(struct svc_rqst *rqstp, 
 static void
 svc_release_buffer(struct svc_rqst *rqstp)
 {
-	while (rqstp->rq_arghi)
-		put_page(rqstp->rq_argpages[--rqstp->rq_arghi]);
-	while (rqstp->rq_resused) {
-		if (rqstp->rq_respages[--rqstp->rq_resused] == NULL)
-			continue;
-		put_page(rqstp->rq_respages[rqstp->rq_resused]);
-	}
-	rqstp->rq_argused = 0;
+	int i;
+	for (i=0; i<ARRAY_SIZE(rqstp->rq_pages); i++)
+		if (rqstp->rq_pages[i])
+			put_page(rqstp->rq_pages[i]);
 }
 
 /*

@@ -698,10 +691,10 @@ svc_process(struct svc_rqst *rqstp)
 	/* setup response xdr_buf.
 	 * Initially it has just one page 
 	 */
-	svc_take_page(rqstp); /* must succeed */
+	rqstp->rq_resused = 1;
 	resv->iov_base = page_address(rqstp->rq_respages[0]);
 	resv->iov_len = 0;
-	rqstp->rq_res.pages = rqstp->rq_respages+1;
+	rqstp->rq_res.pages = rqstp->rq_respages + 1;
 	rqstp->rq_res.len = 0;
 	rqstp->rq_res.page_base = 0;
 	rqstp->rq_res.page_len = 0;

diff .prev/net/sunrpc/svcsock.c ./net/sunrpc/svcsock.c
@@ -313,7 +313,7 @@ svc_sock_release(struct svc_rqst *rqstp)
 
 	svc_release_skb(rqstp);
 
-	svc_free_allpages(rqstp);
+	svc_free_res_pages(rqstp);
 	rqstp->rq_res.page_len = 0;
 	rqstp->rq_res.page_base = 0;
 

@@ -412,7 +412,8 @@ svc_sendto(struct svc_rqst *rqstp, struc
 	/* send head */
 	if (slen == xdr->head[0].iov_len)
 		flags = 0;
-	len = sock->ops->sendpage(sock, rqstp->rq_respages[0], 0, xdr->head[0].iov_len, flags);
+	len = sock->ops->sendpage(sock, rqstp->rq_respages[0], 0,
+				  xdr->head[0].iov_len, flags);
 	if (len != xdr->head[0].iov_len)
 		goto out;
 	slen -= xdr->head[0].iov_len;

@@ -437,8 +438,9 @@ svc_sendto(struct svc_rqst *rqstp, struc
 	}
 	/* send tail */
 	if (xdr->tail[0].iov_len) {
-		result = sock->ops->sendpage(sock, rqstp->rq_respages[rqstp->rq_restailpage], 
-					     ((unsigned long)xdr->tail[0].iov_base)& (PAGE_SIZE-1),
+		result = sock->ops->sendpage(sock, rqstp->rq_respages[0],
+					     ((unsigned long)xdr->tail[0].iov_base)
+					        & (PAGE_SIZE-1),
 					     xdr->tail[0].iov_len, 0);
 
 		if (result > 0)

@@ -708,9 +710,11 @@ svc_udp_recvfrom(struct svc_rqst *rqstp)
 	if (len <= rqstp->rq_arg.head[0].iov_len) {
 		rqstp->rq_arg.head[0].iov_len = len;
 		rqstp->rq_arg.page_len = 0;
+		rqstp->rq_respages = rqstp->rq_pages+1;
 	} else {
 		rqstp->rq_arg.page_len = len - rqstp->rq_arg.head[0].iov_len;
-		rqstp->rq_argused += (rqstp->rq_arg.page_len + PAGE_SIZE - 1)/ PAGE_SIZE;
+		rqstp->rq_respages = rqstp->rq_pages + 1 +
+			(rqstp->rq_arg.page_len + PAGE_SIZE - 1)/ PAGE_SIZE;
 	}
 
 	if (serv->sv_stats)

@@ -1060,11 +1064,12 @@ svc_tcp_recvfrom(struct svc_rqst *rqstp)
 	vlen = PAGE_SIZE;
 	pnum = 1;
 	while (vlen < len) {
-		vec[pnum].iov_base = page_address(rqstp->rq_argpages[rqstp->rq_argused++]);
+		vec[pnum].iov_base = page_address(rqstp->rq_pages[pnum]);
 		vec[pnum].iov_len = PAGE_SIZE;
 		pnum++;
 		vlen += PAGE_SIZE;
 	}
+	rqstp->rq_respages = &rqstp->rq_pages[pnum];
 
 	/* Now receive data */
 	len = svc_recvfrom(rqstp, vec, pnum, len);

@@ -1216,7 +1221,7 @@ svc_recv(struct svc_rqst *rqstp, long ti
 	struct svc_sock		*svsk =NULL;
 	struct svc_serv		*serv = rqstp->rq_server;
 	struct svc_pool		*pool = rqstp->rq_pool;
-	int			len;
+	int			len, i;
 	int 			pages;
 	struct xdr_buf		*arg;
 	DECLARE_WAITQUEUE(wait, current);

@@ -1233,27 +1238,22 @@ svc_recv(struct svc_rqst *rqstp, long ti
 			"svc_recv: service %p, wait queue active!\n",
 			 rqstp);
 
-	/* Initialize the buffers */
-	/* first reclaim pages that were moved to response list */
-	svc_pushback_allpages(rqstp);
 
 	/* now allocate needed pages.  If we get a failure, sleep briefly */
 	pages = 2 + (serv->sv_bufsz + PAGE_SIZE -1) / PAGE_SIZE;
-	while (rqstp->rq_arghi < pages) {
-		struct page *p = alloc_page(GFP_KERNEL);
-		if (!p) {
-			schedule_timeout_uninterruptible(msecs_to_jiffies(500));
-			continue;
+	for (i=0; i < pages ; i++)
+		while (rqstp->rq_pages[i] == NULL) {
+			struct page *p = alloc_page(GFP_KERNEL);
+			if (!p)
+				schedule_timeout_uninterruptible(msecs_to_jiffies(500));
+			rqstp->rq_pages[i] = p;
 		}
-		rqstp->rq_argpages[rqstp->rq_arghi++] = p;
-	}
 
 	/* Make arg->head point to first page and arg->pages point to rest */
 	arg = &rqstp->rq_arg;
-	arg->head[0].iov_base = page_address(rqstp->rq_argpages[0]);
+	arg->head[0].iov_base = page_address(rqstp->rq_pages[0]);
 	arg->head[0].iov_len = PAGE_SIZE;
-	rqstp->rq_argused = 1;
-	arg->pages = rqstp->rq_argpages + 1;
+	arg->pages = rqstp->rq_pages + 1;
 	arg->page_base = 0;
 	/* save at least one page for response */
 	arg->page_len = (pages-2)*PAGE_SIZE;

diff .prev/fs/nfsd/nfs3proc.c ./fs/nfsd/nfs3proc.c
@@ -180,7 +180,7 @@ nfsd3_proc_read(struct svc_rqst *rqstp, 
 	fh_copy(&resp->fh, &argp->fh);
 	nfserr = nfsd_read(rqstp, &resp->fh, NULL,
 				  argp->offset,
-			   	  argp->vec, argp->vlen,
+			   	  rqstp->rq_vec, argp->vlen,
 				  &resp->count);
 	if (nfserr == 0) {
 		struct inode	*inode = resp->fh.fh_dentry->d_inode;

@@ -210,7 +210,7 @@ nfsd3_proc_write(struct svc_rqst *rqstp,
 	resp->committed = argp->stable;
 	nfserr = nfsd_write(rqstp, &resp->fh, NULL,
 				   argp->offset,
-				   argp->vec, argp->vlen,
+				   rqstp->rq_vec, argp->vlen,
 				   argp->len,
 				   &resp->committed);
 	resp->count = argp->count;

diff .prev/fs/nfsd/nfs3xdr.c ./fs/nfsd/nfs3xdr.c
@@ -344,9 +344,9 @@ nfs3svc_decode_readargs(struct svc_rqst 
 	v=0;
 	while (len > 0) {
 		pn = rqstp->rq_resused++;
-		args->vec[v].iov_base = page_address(rqstp->rq_respages[pn]);
-		args->vec[v].iov_len = len < PAGE_SIZE? len : PAGE_SIZE;
-		len -= args->vec[v].iov_len;
+		rqstp->rq_vec[v].iov_base = page_address(rqstp->rq_respages[pn]);
+		rqstp->rq_vec[v].iov_len = len < PAGE_SIZE? len : PAGE_SIZE;
+		len -= rqstp->rq_vec[v].iov_len;
 		v++;
 	}
 	args->vlen = v;

@@ -372,22 +372,22 @@ nfs3svc_decode_writeargs(struct svc_rqst
 	    rqstp->rq_arg.len - hdr < len)
 		return 0;
 
-	args->vec[0].iov_base = (void*)p;
-	args->vec[0].iov_len = rqstp->rq_arg.head[0].iov_len - hdr;
+	rqstp->rq_vec[0].iov_base = (void*)p;
+	rqstp->rq_vec[0].iov_len = rqstp->rq_arg.head[0].iov_len - hdr;
 
 	if (len > NFSSVC_MAXBLKSIZE)
 		len = NFSSVC_MAXBLKSIZE;
 	v=  0;
-	while (len > args->vec[v].iov_len) {
-		len -= args->vec[v].iov_len;
+	while (len > rqstp->rq_vec[v].iov_len) {
+		len -= rqstp->rq_vec[v].iov_len;
 		v++;
-		args->vec[v].iov_base = page_address(rqstp->rq_pages[v]);
-		args->vec[v].iov_len = PAGE_SIZE;
+		rqstp->rq_vec[v].iov_base = page_address(rqstp->rq_pages[v]);
+		rqstp->rq_vec[v].iov_len = PAGE_SIZE;
 	}
-	args->vec[v].iov_len = len;
+	rqstp->rq_vec[v].iov_len = len;
 	args->vlen = v+1;
 
-	return args->count == args->len && args->vec[0].iov_len > 0;
+	return args->count == args->len && rqstp->rq_vec[0].iov_len > 0;
 }
 
 int

diff .prev/fs/nfsd/nfs4proc.c ./fs/nfsd/nfs4proc.c
@@ -646,7 +646,7 @@ nfsd4_write(struct svc_rqst *rqstp, stru
 	*p++ = nfssvc_boot.tv_usec;
 
 	status =  nfsd_write(rqstp, current_fh, filp, write->wr_offset,
-			write->wr_vec, write->wr_vlen, write->wr_buflen,
+			rqstp->rq_vec, write->wr_vlen, write->wr_buflen,
 			&write->wr_how_written);
 	if (filp)
 		fput(filp);

diff .prev/fs/nfsd/nfs4xdr.c ./fs/nfsd/nfs4xdr.c
@@ -927,26 +927,26 @@ nfsd4_decode_write(struct nfsd4_compound
 		printk(KERN_NOTICE "xdr error! (%s:%d)\n", __FILE__, __LINE__); 
 		goto xdr_error;
 	}
-	write->wr_vec[0].iov_base = p;
-	write->wr_vec[0].iov_len = avail;
+	argp->rqstp->rq_vec[0].iov_base = p;
+	argp->rqstp->rq_vec[0].iov_len = avail;
 	v = 0;
 	len = write->wr_buflen;
-	while (len > write->wr_vec[v].iov_len) {
-		len -= write->wr_vec[v].iov_len;
+	while (len > argp->rqstp->rq_vec[v].iov_len) {
+		len -= argp->rqstp->rq_vec[v].iov_len;
 		v++;
-		write->wr_vec[v].iov_base = page_address(argp->pagelist[0]);
+		argp->rqstp->rq_vec[v].iov_base = page_address(argp->pagelist[0]);
 		argp->pagelist++;
 		if (argp->pagelen >= PAGE_SIZE) {
-			write->wr_vec[v].iov_len = PAGE_SIZE;
+			argp->rqstp->rq_vec[v].iov_len = PAGE_SIZE;
 			argp->pagelen -= PAGE_SIZE;
 		} else {
-			write->wr_vec[v].iov_len = argp->pagelen;
+			argp->rqstp->rq_vec[v].iov_len = argp->pagelen;
 			argp->pagelen -= len;
 		}
 	}
-	argp->end = (u32*) (write->wr_vec[v].iov_base + write->wr_vec[v].iov_len);
-	argp->p = (u32*)  (write->wr_vec[v].iov_base + (XDR_QUADLEN(len) << 2));
-	write->wr_vec[v].iov_len = len;
+	argp->end = (u32*) (argp->rqstp->rq_vec[v].iov_base + argp->rqstp->rq_vec[v].iov_len);
+	argp->p = (u32*)  (argp->rqstp->rq_vec[v].iov_base + (XDR_QUADLEN(len) << 2));
+	argp->rqstp->rq_vec[v].iov_len = len;
 	write->wr_vlen = v+1;
 
 	DECODE_TAIL;

@@ -2064,9 +2064,9 @@ nfsd4_encode_read(struct nfsd4_compoundr
 	v = 0;
 	while (len > 0) {
 		pn = resp->rqstp->rq_resused++;
-		read->rd_iov[v].iov_base =
+		resp->rqstp->rq_vec[v].iov_base =
 			page_address(resp->rqstp->rq_respages[pn]);
-		read->rd_iov[v].iov_len =
+		resp->rqstp->rq_vec[v].iov_len =
 			len < PAGE_SIZE ? len : PAGE_SIZE;
 		v++;
 		len -= PAGE_SIZE;

@@ -2074,7 +2074,7 @@ nfsd4_encode_read(struct nfsd4_compoundr
 	read->rd_vlen = v;
 
 	nfserr = nfsd_read(read->rd_rqstp, read->rd_fhp, read->rd_filp,
-			read->rd_offset, read->rd_iov, read->rd_vlen,
+			read->rd_offset, resp->rqstp->rq_vec, read->rd_vlen,
 			&maxcount);
 
 	if (nfserr == nfserr_symlink)

diff .prev/fs/nfsd/nfsproc.c ./fs/nfsd/nfsproc.c
@@ -159,7 +159,7 @@ nfsd_proc_read(struct svc_rqst *rqstp, s
 	resp->count = argp->count;
 	nfserr = nfsd_read(rqstp, fh_copy(&resp->fh, &argp->fh), NULL,
 				  argp->offset,
-			   	  argp->vec, argp->vlen,
+			   	  rqstp->rq_vec, argp->vlen,
 				  &resp->count);
 
 	if (nfserr) return nfserr;

@@ -185,7 +185,7 @@ nfsd_proc_write(struct svc_rqst *rqstp, 
 
 	nfserr = nfsd_write(rqstp, fh_copy(&resp->fh, &argp->fh), NULL,
 				   argp->offset,
-				   argp->vec, argp->vlen,
+				   rqstp->rq_vec, argp->vlen,
 				   argp->len,
 				   &stable);
 	return nfsd_return_attrs(nfserr, resp);

diff .prev/fs/nfsd/nfsxdr.c ./fs/nfsd/nfsxdr.c
@@ -263,9 +263,9 @@ nfssvc_decode_readargs(struct svc_rqst *
 	v=0;
 	while (len > 0) {
 		pn = rqstp->rq_resused++;
-		args->vec[v].iov_base = page_address(rqstp->rq_respages[pn]);
-		args->vec[v].iov_len = len < PAGE_SIZE?len:PAGE_SIZE;
-		len -= args->vec[v].iov_len;
+		rqstp->rq_vec[v].iov_base = page_address(rqstp->rq_respages[pn]);
+		rqstp->rq_vec[v].iov_len = len < PAGE_SIZE?len:PAGE_SIZE;
+		len -= rqstp->rq_vec[v].iov_len;
 		v++;
 	}
 	args->vlen = v;

@@ -285,21 +285,21 @@ nfssvc_decode_writeargs(struct svc_rqst 
 	args->offset = ntohl(*p++);	/* offset */
 	p++;				/* totalcount */
 	len = args->len = ntohl(*p++);
-	args->vec[0].iov_base = (void*)p;
-	args->vec[0].iov_len = rqstp->rq_arg.head[0].iov_len -
+	rqstp->rq_vec[0].iov_base = (void*)p;
+	rqstp->rq_vec[0].iov_len = rqstp->rq_arg.head[0].iov_len -
 				(((void*)p) - rqstp->rq_arg.head[0].iov_base);
 	if (len > NFSSVC_MAXBLKSIZE)
 		len = NFSSVC_MAXBLKSIZE;
 	v = 0;
-	while (len > args->vec[v].iov_len) {
-		len -= args->vec[v].iov_len;
+	while (len > rqstp->rq_vec[v].iov_len) {
+		len -= rqstp->rq_vec[v].iov_len;
 		v++;
-		args->vec[v].iov_base = page_address(rqstp->rq_pages[v]);
-		args->vec[v].iov_len = PAGE_SIZE;
+		rqstp->rq_vec[v].iov_base = page_address(rqstp->rq_pages[v]);
+		rqstp->rq_vec[v].iov_len = PAGE_SIZE;
 	}
-	args->vec[v].iov_len = len;
+	rqstp->rq_vec[v].iov_len = len;
 	args->vlen = v+1;
-	return args->vec[0].iov_len > 0;
+	return rqstp->rq_vec[0].iov_len > 0;
 }
 
 int

diff .prev/include/linux/nfsd/xdr.h ./include/linux/nfsd/xdr.h
@@ -30,7 +30,6 @@ struct nfsd_readargs {
 	struct svc_fh		fh;
 	__u32			offset;
 	__u32			count;
-	struct kvec		vec[RPCSVC_MAXPAGES];
 	int			vlen;
 };
 

@@ -38,7 +37,6 @@ struct nfsd_writeargs {
 	svc_fh			fh;
 	__u32			offset;
 	int			len;
-	struct kvec		vec[RPCSVC_MAXPAGES];
 	int			vlen;
 };
 

diff .prev/include/linux/nfsd/xdr3.h ./include/linux/nfsd/xdr3.h
@@ -33,7 +33,6 @@ struct nfsd3_readargs {
 	struct svc_fh		fh;
 	__u64			offset;
 	__u32			count;
-	struct kvec		vec[RPCSVC_MAXPAGES];
 	int			vlen;
 };
 

@@ -43,7 +42,6 @@ struct nfsd3_writeargs {
 	__u32			count;
 	int			stable;
 	__u32			len;
-	struct kvec		vec[RPCSVC_MAXPAGES];
 	int			vlen;
 };
 

diff .prev/include/linux/nfsd/xdr4.h ./include/linux/nfsd/xdr4.h
@@ -241,7 +241,6 @@ struct nfsd4_read {
 	stateid_t	rd_stateid;         /* request */
 	u64		rd_offset;          /* request */
 	u32		rd_length;          /* request */
-	struct kvec	rd_iov[RPCSVC_MAXPAGES];
 	int		rd_vlen;
 	struct file     *rd_filp;
 	

@@ -326,7 +325,6 @@ struct nfsd4_write {
 	u64		wr_offset;          /* request */
 	u32		wr_stable_how;      /* request */
 	u32		wr_buflen;          /* request */
-	struct kvec	wr_vec[RPCSVC_MAXPAGES]; /* request */
 	int		wr_vlen;
 
 	u32		wr_bytes_written;   /* response */

diff .prev/include/linux/sunrpc/svc.h ./include/linux/sunrpc/svc.h
@@ -175,6 +175,8 @@ struct svc_rqst {
 	struct page *		*rq_respages;	/* points into rq_pages */
 	int			rq_resused;	/* number of pages used for result */
 
+	struct kvec		rq_vec[RPCSVC_MAXPAGES]; /* generally useful.. */
+
 	u32			rq_xid;		/* transmission id */
 	u32			rq_prog;	/* program number */
 	u32			rq_vers;	/* program version */

diff .prev/net/sunrpc/svcsock.c ./net/sunrpc/svcsock.c
@@ -962,7 +962,7 @@ svc_tcp_recvfrom(struct svc_rqst *rqstp)
 	struct svc_sock	*svsk = rqstp->rq_sock;
 	struct svc_serv	*serv = svsk->sk_server;
 	int		len;
-	struct kvec vec[RPCSVC_MAXPAGES];
+	struct kvec *vec;
 	int pnum, vlen;
 
 	dprintk("svc: tcp_recv %p data %d conn %d close %d\n",

@@ -1060,6 +1060,7 @@ svc_tcp_recvfrom(struct svc_rqst *rqstp)
 	len = svsk->sk_reclen;
 	set_bit(SK_DATA, &svsk->sk_flags);
 
+	vec = rqstp->rq_vec;
 	vec[0] = rqstp->rq_arg.head[0];
 	vlen = PAGE_SIZE;
 	pnum = 1;

diff .prev/fs/nfsd/nfs3proc.c ./fs/nfsd/nfs3proc.c
@@ -160,6 +160,7 @@ nfsd3_proc_read(struct svc_rqst *rqstp, 
 				        struct nfsd3_readres  *resp)
 {
 	int	nfserr;
+	u32	max_blocksize = svc_max_payload(rqstp);
 
 	dprintk("nfsd: READ(3) %s %lu bytes at %lu\n",
 				SVCFH_fmt(&argp->fh),

@@ -172,8 +173,8 @@ nfsd3_proc_read(struct svc_rqst *rqstp, 
 	 */
 
 	resp->count = argp->count;
-	if (NFSSVC_MAXBLKSIZE < resp->count)
-		resp->count = NFSSVC_MAXBLKSIZE;
+	if (max_blocksize < resp->count)
+		resp->count = max_blocksize;
 
 	svc_reserve(rqstp, ((1 + NFS3_POST_OP_ATTR_WORDS + 3)<<2) + resp->count +4);
 

@@ -538,15 +539,16 @@ nfsd3_proc_fsinfo(struct svc_rqst * rqst
 					   struct nfsd3_fsinfores *resp)
 {
 	int	nfserr;
+	u32	max_blocksize = svc_max_payload(rqstp);
 
 	dprintk("nfsd: FSINFO(3)   %s\n",
 				SVCFH_fmt(&argp->fh));
 
-	resp->f_rtmax  = NFSSVC_MAXBLKSIZE;
-	resp->f_rtpref = NFSSVC_MAXBLKSIZE;
+	resp->f_rtmax  = max_blocksize;
+	resp->f_rtpref = max_blocksize;
 	resp->f_rtmult = PAGE_SIZE;
-	resp->f_wtmax  = NFSSVC_MAXBLKSIZE;
-	resp->f_wtpref = NFSSVC_MAXBLKSIZE;
+	resp->f_wtmax  = max_blocksize;
+	resp->f_wtpref = max_blocksize;
 	resp->f_wtmult = PAGE_SIZE;
 	resp->f_dtpref = PAGE_SIZE;
 	resp->f_maxfilesize = ~(u32) 0;

diff .prev/fs/nfsd/nfs3xdr.c ./fs/nfsd/nfs3xdr.c
@@ -330,6 +330,7 @@ nfs3svc_decode_readargs(struct svc_rqst 
 {
 	unsigned int len;
 	int v,pn;
+	u32 max_blocksize = svc_max_payload(rqstp);
 
 	if (!(p = decode_fh(p, &args->fh))
 	 || !(p = xdr_decode_hyper(p, &args->offset)))

@@ -337,8 +338,8 @@ nfs3svc_decode_readargs(struct svc_rqst 
 
 	len = args->count = ntohl(*p++);
 
-	if (len > NFSSVC_MAXBLKSIZE)
-		len = NFSSVC_MAXBLKSIZE;
+	if (len > max_blocksize)
+		len = max_blocksize;
 
 	/* set up the kvec */
 	v=0;

@@ -358,6 +359,7 @@ nfs3svc_decode_writeargs(struct svc_rqst
 					struct nfsd3_writeargs *args)
 {
 	unsigned int len, v, hdr;
+	u32 max_blocksize = svc_max_payload(rqstp);
 
 	if (!(p = decode_fh(p, &args->fh))
 	 || !(p = xdr_decode_hyper(p, &args->offset)))

@@ -375,8 +377,8 @@ nfs3svc_decode_writeargs(struct svc_rqst
 	rqstp->rq_vec[0].iov_base = (void*)p;
 	rqstp->rq_vec[0].iov_len = rqstp->rq_arg.head[0].iov_len - hdr;
 
-	if (len > NFSSVC_MAXBLKSIZE)
-		len = NFSSVC_MAXBLKSIZE;
+	if (len > max_blocksize)
+		len = max_blocksize;
 	v=  0;
 	while (len > rqstp->rq_vec[v].iov_len) {
 		len -= rqstp->rq_vec[v].iov_len;

@@ -564,6 +566,7 @@ nfs3svc_decode_readdirplusargs(struct sv
 					struct nfsd3_readdirargs *args)
 {
 	int len, pn;
+	u32 max_blocksize = svc_max_payload(rqstp);
 
 	if (!(p = decode_fh(p, &args->fh)))
 		return 0;

@@ -572,7 +575,7 @@ nfs3svc_decode_readdirplusargs(struct sv
 	args->dircount = ntohl(*p++);
 	args->count    = ntohl(*p++);
 
-	len = (args->count > NFSSVC_MAXBLKSIZE) ? NFSSVC_MAXBLKSIZE :
+	len = (args->count > max_blocksize) ? max_blocksize :
 						  args->count;
 	args->count = len;
 

diff .prev/fs/nfsd/nfs4xdr.c ./fs/nfsd/nfs4xdr.c
@@ -1537,12 +1537,12 @@ out_acl:
 	if (bmval0 & FATTR4_WORD0_MAXREAD) {
 		if ((buflen -= 8) < 0)
 			goto out_resource;
-		WRITE64((u64) NFSSVC_MAXBLKSIZE);
+		WRITE64((u64) svc_max_payload(rqstp));
 	}
 	if (bmval0 & FATTR4_WORD0_MAXWRITE) {
 		if ((buflen -= 8) < 0)
 			goto out_resource;
-		WRITE64((u64) NFSSVC_MAXBLKSIZE);
+		WRITE64((u64) svc_max_payload(rqstp));
 	}
 	if (bmval1 & FATTR4_WORD1_MODE) {
 		if ((buflen -= 4) < 0)

@@ -2056,7 +2056,7 @@ nfsd4_encode_read(struct nfsd4_compoundr
 
 	RESERVE_SPACE(8); /* eof flag and byte count */
 
-	maxcount = NFSSVC_MAXBLKSIZE;
+	maxcount = svc_max_payload(resp->rqstp);
 	if (maxcount > read->rd_length)
 		maxcount = read->rd_length;
 

diff .prev/fs/nfsd/nfsproc.c ./fs/nfsd/nfsproc.c
@@ -146,13 +146,13 @@ nfsd_proc_read(struct svc_rqst *rqstp, s
 	 * status, 17 words for fattr, and 1 word for the byte count.
 	 */
 
-	if (NFSSVC_MAXBLKSIZE < argp->count) {
+	if (NFSSVC_MAXBLKSIZE_V2 < argp->count) {
 		printk(KERN_NOTICE
 			"oversized read request from %u.%u.%u.%u:%d (%d bytes)\n",
 				NIPQUAD(rqstp->rq_addr.sin_addr.s_addr),
 				ntohs(rqstp->rq_addr.sin_port),
 				argp->count);
-		argp->count = NFSSVC_MAXBLKSIZE;
+		argp->count = NFSSVC_MAXBLKSIZE_V2;
 	}
 	svc_reserve(rqstp, (19<<2) + argp->count + 4);
 

@@ -553,7 +553,7 @@ static struct svc_procedure		nfsd_proced
   PROC(none,	 void,		void,		none,		RC_NOCACHE, ST),
   PROC(lookup,	 diropargs,	diropres,	fhandle,	RC_NOCACHE, ST+FH+AT),
   PROC(readlink, readlinkargs,	readlinkres,	none,		RC_NOCACHE, ST+1+NFS_MAXPATHLEN/4),
-  PROC(read,	 readargs,	readres,	fhandle,	RC_NOCACHE, ST+AT+1+NFSSVC_MAXBLKSIZE/4),
+  PROC(read,	 readargs,	readres,	fhandle,	RC_NOCACHE, ST+AT+1+NFSSVC_MAXBLKSIZE_V2/4),
   PROC(none,	 void,		void,		none,		RC_NOCACHE, ST),
   PROC(write,	 writeargs,	attrstat,	fhandle,	RC_REPLBUFF, ST+AT),
   PROC(create,	 createargs,	diropres,	fhandle,	RC_REPLBUFF, ST+FH+AT),

diff .prev/fs/nfsd/nfsxdr.c ./fs/nfsd/nfsxdr.c
@@ -254,8 +254,8 @@ nfssvc_decode_readargs(struct svc_rqst *
 	len = args->count     = ntohl(*p++);
 	p++; /* totalcount - unused */
 
-	if (len > NFSSVC_MAXBLKSIZE)
-		len = NFSSVC_MAXBLKSIZE;
+	if (len > NFSSVC_MAXBLKSIZE_V2)
+		len = NFSSVC_MAXBLKSIZE_V2;
 
 	/* set up somewhere to store response.
 	 * We take pages, put them on reslist and include in iovec

@@ -288,8 +288,8 @@ nfssvc_decode_writeargs(struct svc_rqst 
 	rqstp->rq_vec[0].iov_base = (void*)p;
 	rqstp->rq_vec[0].iov_len = rqstp->rq_arg.head[0].iov_len -
 				(((void*)p) - rqstp->rq_arg.head[0].iov_base);
-	if (len > NFSSVC_MAXBLKSIZE)
-		len = NFSSVC_MAXBLKSIZE;
+	if (len > NFSSVC_MAXBLKSIZE_V2)
+		len = NFSSVC_MAXBLKSIZE_V2;
 	v = 0;
 	while (len > rqstp->rq_vec[v].iov_len) {
 		len -= rqstp->rq_vec[v].iov_len;

@@ -458,7 +458,7 @@ nfssvc_encode_statfsres(struct svc_rqst 
 {
 	struct kstatfs	*stat = &resp->stats;
 
-	*p++ = htonl(NFSSVC_MAXBLKSIZE);	/* max transfer size */
+	*p++ = htonl(NFSSVC_MAXBLKSIZE_V2);	/* max transfer size */
 	*p++ = htonl(stat->f_bsize);
 	*p++ = htonl(stat->f_blocks);
 	*p++ = htonl(stat->f_bfree);

diff .prev/include/linux/nfsd/const.h ./include/linux/nfsd/const.h
@@ -13,6 +13,7 @@
 #include <linux/nfs2.h>
 #include <linux/nfs3.h>
 #include <linux/nfs4.h>
+#include <linux/sunrpc/msg_prot.h>
 
 /*
  * Maximum protocol version supported by knfsd

@@ -23,6 +24,8 @@
  * Maximum blocksize supported by daemon currently at 32K
  */
 #define NFSSVC_MAXBLKSIZE	(32*1024)
+/* NFSv2 is limited by the protocol specification, see RFC 1094 */
+#define NFSSVC_MAXBLKSIZE_V2	(8*1024)
 
 #ifdef __KERNEL__
 

@@ -30,7 +33,17 @@
 # define NFS_SUPER_MAGIC	0x6969
 #endif
 
-#define NFSD_BUFSIZE		(1024 + NFSSVC_MAXBLKSIZE)
+/*
+ * Largest number of bytes we need to allocate for an NFS
+ * call or reply.  Used to control buffer sizes.  We use
+ * the length of v3 WRITE, READDIR and READDIR replies
+ * which are an RPC header, up to 26 XDR units of reply
+ * data, and some page data.
+ *
+ * Note that accuracy here doesn't matter too much as the
+ * size is rounded up to a page size when allocating space.
+ */
+#define NFSD_BUFSIZE		((RPC_MAX_HEADER_WITH_AUTH+26)*XDR_UNIT + NFSSVC_MAXBLKSIZE)
 
 #ifdef CONFIG_NFSD_V4
 # define NFSSVC_XDRSIZE		NFS4_SVC_XDRSIZE

diff .prev/include/linux/sunrpc/auth.h ./include/linux/sunrpc/auth.h
@@ -20,9 +20,6 @@
 /* size of the nodename buffer */
 #define UNX_MAXNODENAME	32
 
-/* Maximum size (in bytes) of an rpc credential or verifier */
-#define RPC_MAX_AUTH_SIZE (400)
-
 /* Work around the lack of a VFS credential */
 struct auth_cred {
 	uid_t	uid;

diff .prev/include/linux/sunrpc/msg_prot.h ./include/linux/sunrpc/msg_prot.h
@@ -11,6 +11,9 @@
 
 #define RPC_VERSION 2
 
+/* size of an XDR encoding unit in bytes, i.e. 32bit */
+#define XDR_UNIT	(4)
+
 /* spec defines authentication flavor as an unsigned 32 bit integer */
 typedef u32	rpc_authflavor_t;
 

@@ -34,6 +37,9 @@ enum rpc_auth_flavors {
 	RPC_AUTH_GSS_SPKMP = 390011,
 };
 
+/* Maximum size (in bytes) of an rpc credential or verifier */
+#define RPC_MAX_AUTH_SIZE (400)
+
 enum rpc_msg_type {
 	RPC_CALL = 0,
 	RPC_REPLY = 1

@@ -101,5 +107,39 @@ typedef u32	rpc_fraghdr;
 #define	RPC_FRAGMENT_SIZE_MASK		(~RPC_LAST_STREAM_FRAGMENT)
 #define	RPC_MAX_FRAGMENT_SIZE		((1U << 31) - 1)
 
+/*
+ * RPC call and reply header size as number of 32bit words (verifier
+ * size computed separately, see below)
+ */
+#define RPC_CALLHDRSIZE		(6)
+#define RPC_REPHDRSIZE		(4)
+
+
+/*
+ * Maximum RPC header size, including authentication,
+ * as number of 32bit words (see RFCs 1831, 1832).
+ *
+ *	xid			    1 xdr unit = 4 bytes
+ *	mtype			    1
+ *	rpc_version		    1
+ *	program			    1
+ *	prog_version		    1
+ *	procedure		    1
+ *	cred {
+ *	    flavor		    1
+ *	    length		    1
+ *	    body<RPC_MAX_AUTH_SIZE> 100 xdr units = 400 bytes
+ *	}
+ *	verf {
+ *	    flavor		    1
+ *	    length		    1
+ *	    body<RPC_MAX_AUTH_SIZE> 100 xdr units = 400 bytes
+ *	}
+ *	TOTAL			    210 xdr units = 840 bytes
+ */
+#define RPC_MAX_HEADER_WITH_AUTH \
+	(RPC_CALLHDRSIZE + 2*(2+RPC_MAX_AUTH_SIZE/4))
+
+
 #endif /* __KERNEL__ */
 #endif /* _LINUX_SUNRPC_MSGPROT_H_ */

diff .prev/include/linux/sunrpc/svc.h ./include/linux/sunrpc/svc.h
@@ -13,6 +13,7 @@
 #include <linux/in.h>
 #include <linux/sunrpc/types.h>
 #include <linux/sunrpc/xdr.h>
+#include <linux/sunrpc/auth.h>
 #include <linux/sunrpc/svcauth.h>
 #include <linux/wait.h>
 #include <linux/mm.h>

@@ -95,8 +96,28 @@ static inline void svc_get(struct svc_se
  * Maximum payload size supported by a kernel RPC server.
  * This is use to determine the max number of pages nfsd is
  * willing to return in a single READ operation.
- */
-#define RPCSVC_MAXPAYLOAD	(64*1024u)
+ *
+ * These happen to all be powers of 2, which is not strictly
+ * necessary but helps enforce the real limitation, which is
+ * that they should be multiples of PAGE_CACHE_SIZE.
+ *
+ * For UDP transports, a block plus NFS,RPC, and UDP headers
+ * has to fit into the IP datagram limit of 64K.  The largest
+ * feasible number for all known page sizes is probably 48K,
+ * but we choose 32K here.  This is the same as the historical
+ * Linux limit; someone who cares more about NFS/UDP performance
+ * can test a larger number.
+ *
+ * For TCP transports we have more freedom.  A size of 1MB is
+ * chosen to match the client limit.  Other OSes are known to
+ * have larger limits, but those numbers are probably beyond
+ * the point of diminishing returns.
+ */
+#define RPCSVC_MAXPAYLOAD	(1*1024*1024u)
+#define RPCSVC_MAXPAYLOAD_TCP	RPCSVC_MAXPAYLOAD
+#define RPCSVC_MAXPAYLOAD_UDP	(32*1024u)
+
+extern u32 svc_max_payload(const struct svc_rqst *rqstp);
 
 /*
  * RPC Requsts and replies are stored in one or more pages.

diff .prev/include/linux/sunrpc/xprt.h ./include/linux/sunrpc/xprt.h
@@ -14,6 +14,7 @@
 #include <linux/in.h>
 #include <linux/sunrpc/sched.h>
 #include <linux/sunrpc/xdr.h>
+#include <linux/sunrpc/msg_prot.h>
 
 extern unsigned int xprt_udp_slot_table_entries;
 extern unsigned int xprt_tcp_slot_table_entries;

@@ -23,13 +24,6 @@ extern unsigned int xprt_tcp_slot_table_
 #define RPC_MAX_SLOT_TABLE	(128U)
 
 /*
- * RPC call and reply header size as number of 32bit words (verifier
- * size computed separately)
- */
-#define RPC_CALLHDRSIZE		6
-#define RPC_REPHDRSIZE		4
-
-/*
  * Parameters for choosing a free port
  */
 extern unsigned int xprt_min_resvport;

diff .prev/net/sunrpc/svc.c ./net/sunrpc/svc.c
@@ -910,3 +910,18 @@ err_bad:
 	svc_putu32(resv, rpc_stat);
 	goto sendit;
 }
+
+/*
+ * Return (transport-specific) limit on the rpc payload.
+ */
+u32 svc_max_payload(const struct svc_rqst *rqstp)
+{
+	int max = RPCSVC_MAXPAYLOAD_TCP;
+
+	if (rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
+		max = RPCSVC_MAXPAYLOAD_UDP;
+	if (rqstp->rq_server->sv_bufsz < max)
+		max = rqstp->rq_server->sv_bufsz;
+	return max;
+}
+EXPORT_SYMBOL_GPL(svc_max_payload);

diff .prev/fs/nfsd/nfsctl.c ./fs/nfsd/nfsctl.c
@@ -57,6 +57,7 @@ enum {
 	NFSD_Pool_Threads,
 	NFSD_Versions,
 	NFSD_Ports,
+	NFSD_MaxBlkSize,
 	/*
 	 * The below MUST come last.  Otherwise we leave a hole in nfsd_files[]
 	 * with !CONFIG_NFSD_V4 and simple_fill_super() goes oops

@@ -82,6 +83,7 @@ static ssize_t write_threads(struct file
 static ssize_t write_pool_threads(struct file *file, char *buf, size_t size);
 static ssize_t write_versions(struct file *file, char *buf, size_t size);
 static ssize_t write_ports(struct file *file, char *buf, size_t size);
+static ssize_t write_maxblksize(struct file *file, char *buf, size_t size);
 #ifdef CONFIG_NFSD_V4
 static ssize_t write_leasetime(struct file *file, char *buf, size_t size);
 static ssize_t write_recoverydir(struct file *file, char *buf, size_t size);

@@ -100,6 +102,7 @@ static ssize_t (*write_op[])(struct file
 	[NFSD_Pool_Threads] = write_pool_threads,
 	[NFSD_Versions] = write_versions,
 	[NFSD_Ports] = write_ports,
+	[NFSD_MaxBlkSize] = write_maxblksize,
 #ifdef CONFIG_NFSD_V4
 	[NFSD_Leasetime] = write_leasetime,
 	[NFSD_RecoveryDir] = write_recoverydir,

@@ -553,6 +556,27 @@ static ssize_t write_ports(struct file *
 	return -EINVAL;
 }
 
+int nfsd_max_blksize;
+
+static ssize_t write_maxblksize(struct file *file, char *buf, size_t size)
+{
+	char *mesg = buf;
+	if (size > 0) {
+		int bsize;
+		int rv = get_int(&mesg, &bsize);
+		if (rv)
+			return rv;
+		if (bsize < 1024)
+			bsize = 1024;
+		if (bsize > NFSSVC_MAXBLKSIZE)
+			bsize = NFSSVC_MAXBLKSIZE;
+		if (nfsd_serv && nfsd_serv->sv_nrthreads)
+			return -EBUSY;
+		nfsd_max_blksize = bsize;
+	}
+	return sprintf(buf, "%d\n", nfsd_max_blksize);
+}
+
 #ifdef CONFIG_NFSD_V4
 extern time_t nfs4_leasetime(void);
 

@@ -618,6 +642,7 @@ static int nfsd_fill_super(struct super_
 		[NFSD_Pool_Threads] = {"pool_threads", &transaction_ops, S_IWUSR|S_IRUSR},
 		[NFSD_Versions] = {"versions", &transaction_ops, S_IWUSR|S_IRUSR},
 		[NFSD_Ports] = {"portlist", &transaction_ops, S_IWUSR|S_IRUGO},
+		[NFSD_MaxBlkSize] = {"max_block_size", &transaction_ops, S_IWUSR|S_IRUGO},
 #ifdef CONFIG_NFSD_V4
 		[NFSD_Leasetime] = {"nfsv4leasetime", &transaction_ops, S_IWUSR|S_IRUSR},
 		[NFSD_RecoveryDir] = {"nfsv4recoverydir", &transaction_ops, S_IWUSR|S_IRUSR},

diff .prev/fs/nfsd/nfssvc.c ./fs/nfsd/nfssvc.c
@@ -198,9 +198,21 @@ int nfsd_create_serv(void)
 		unlock_kernel();
 		return 0;
 	}
+	if (nfsd_max_blksize == 0) {
+		/* choose a suitable default */
+		struct sysinfo i;
+		si_meminfo(&i);
+		if (i.totalram > (1<<(30-PAGE_SHIFT)))
+			nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
+		else if (i.totalram > (1<<(24-PAGE_SHIFT)))
+			nfsd_max_blksize = 128*1024;
+		else
+			nfsd_max_blksize = 32*1024;
+	}
 
 	atomic_set(&nfsd_busy, 0);
-	nfsd_serv = svc_create_pooled(&nfsd_program, NFSD_BUFSIZE,
+	nfsd_serv = svc_create_pooled(&nfsd_program,
+				      NFSD_BUFSIZE - NFSSVC_MAXBLKSIZE + nfsd_max_blksize,
 				      nfsd_last_thread,
 				      nfsd, SIG_NOCLEAN, THIS_MODULE);
 	if (nfsd_serv == NULL)

diff .prev/include/linux/nfsd/const.h ./include/linux/nfsd/const.h
@@ -21,9 +21,9 @@
 #define NFSSVC_MAXVERS		3
 
 /*
- * Maximum blocksize supported by daemon currently at 32K
+ * Maximum blocksizes supported by daemon under various circumstances.
  */
-#define NFSSVC_MAXBLKSIZE	(32*1024)
+#define NFSSVC_MAXBLKSIZE	RPCSVC_MAXPAYLOAD
 /* NFSv2 is limited by the protocol specification, see RFC 1094 */
 #define NFSSVC_MAXBLKSIZE_V2	(8*1024)
 

diff .prev/include/linux/nfsd/nfsd.h ./include/linux/nfsd/nfsd.h
@@ -145,6 +145,7 @@ int nfsd_vers(int vers, enum vers_op cha
 void nfsd_reset_versions(void);
 int nfsd_create_serv(void);
 
+extern int nfsd_max_blksize;
 
 /* 
  * NFSv4 State