Linux NFS development
 help / color / mirror / Atom feed
* [PATCH 0/3] Should truncated READDIR replies return -EIO?
@ 2008-02-22 19:49 Jeff Layton
  2008-02-22 19:49 ` [PATCH 1/3] NFS: clean up short packet handling for NFSv2 readdir Jeff Layton
  0 siblings, 1 reply; 6+ messages in thread
From: Jeff Layton @ 2008-02-22 19:49 UTC (permalink / raw)
  To: Trond Myklebust; +Cc: linux-nfs, nfsv4


This patchset is a follow up to the discussion by the same title. I've
broken them out by NFS version. I've only really tested the NFSv2
version of this patch against the buggy server I have, but it does work
there as expected. The others have only had basic testing, but I didn't
see any obvious breakage there.

I don't see this patchset as particularly high priority, so 2.6.26 is
probably reasonable.

Signed-off-by: Jeff Layton <jlayton@redhat.com>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH 1/3] NFS: clean up short packet handling for NFSv2 readdir
  2008-02-22 19:49 [PATCH 0/3] Should truncated READDIR replies return -EIO? Jeff Layton
@ 2008-02-22 19:49 ` Jeff Layton
  2008-02-22 19:50   ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Jeff Layton
  0 siblings, 1 reply; 6+ messages in thread
From: Jeff Layton @ 2008-02-22 19:49 UTC (permalink / raw)
  To: Trond Myklebust; +Cc: linux-nfs, nfsv4

Currently, the NFS readdir decoders have a workaround for buggy servers
that send an empty readdir response with the EOF bit unset. If the
server sends a malformed response in some cases, this workaround kicks
in and just returns an empty response rather than returning a proper
error to the caller.

This patch does 3 things:

1) have malformed responses with no entries return error (-EIO)

2) preserve existing workaround for servers that send empty
   responses with the EOF marker unset.

3) Add some comments to clarify the logic in nfs_xdr_readdirres().

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/nfs/nfs2xdr.c |   37 ++++++++++++++++++++++++++++---------
 1 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c
index 1f7ea67..86a80b3 100644
--- a/fs/nfs/nfs2xdr.c
+++ b/fs/nfs/nfs2xdr.c
@@ -428,7 +428,7 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy)
 	size_t hdrlen;
 	unsigned int pglen, recvd;
 	u32 len;
-	int status, nr;
+	int status, nr = 0;
 	__be32 *end, *entry, *kaddr;
 
 	if ((status = ntohl(*p++)))
@@ -452,7 +452,12 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy)
 	kaddr = p = kmap_atomic(*page, KM_USER0);
 	end = (__be32 *)((char *)p + pglen);
 	entry = p;
-	for (nr = 0; *p++; nr++) {
+
+	/* Make sure the packet actually has a value_follows and EOF entry */
+	if ((entry + 1) > end)
+		goto short_pkt;
+
+	for (; *p++; nr++) {
 		if (p + 2 > end)
 			goto short_pkt;
 		p++; /* fileid */
@@ -467,18 +472,32 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy)
 			goto short_pkt;
 		entry = p;
 	}
-	if (!nr && (entry[0] != 0 || entry[1] == 0))
-		goto short_pkt;
+
+	/*
+	 * Apparently some server sends responses that are a valid size, but
+	 * contain no entries, and have value_follows==0 and EOF==0. For
+	 * those, just set the EOF marker.
+	 */
+	if (!nr && entry[1] == 0) {
+		dprintk("NFS: readdir reply truncated!\n");
+		entry[1] = 1;
+	}
  out:
 	kunmap_atomic(kaddr, KM_USER0);
 	return nr;
  short_pkt:
+	/*
+	 * When we get a short packet there are 2 possibilities. We can
+	 * return an error, or fix up the response to look like a valid
+	 * response and return what we have so far. If there are no
+	 * entries and the packet was short, then return -EIO. If there
+	 * are valid entries in the response, return them and pretend that
+	 * the call was successful, but incomplete. The caller can retry the
+	 * readdir starting at the last cookie.
+	 */
 	entry[0] = entry[1] = 0;
-	/* truncate listing ? */
-	if (!nr) {
-		dprintk("NFS: readdir reply truncated!\n");
-		entry[1] = 1;
-	}
+	if (!nr)
+		nr = -errno_NFSERR_IO;
 	goto out;
 err_unmap:
 	nr = -errno_NFSERR_IO;
-- 
1.5.4.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir
  2008-02-22 19:49 ` [PATCH 1/3] NFS: clean up short packet handling for NFSv2 readdir Jeff Layton
@ 2008-02-22 19:50   ` Jeff Layton
  2008-02-22 19:50     ` [PATCH 3/3] NFS: clean up short packet handling for NFSv4 readdir Jeff Layton
  2008-02-22 21:11     ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Chuck Lever
  0 siblings, 2 replies; 6+ messages in thread
From: Jeff Layton @ 2008-02-22 19:50 UTC (permalink / raw)
  To: Trond Myklebust; +Cc: linux-nfs, nfsv4

Currently, the NFS readdir decoders have a workaround for buggy servers
that send an empty readdir response with the EOF bit unset. If the
server sends a malformed response in some cases, this workaround kicks
in and just returns an empty response rather than returning a proper
error to the caller.

This patch does 3 things:

1) have malformed responses with no entries return error (-EIO)

2) preserve existing workaround for servers that send empty
   responses with the EOF marker unset.

3) Add some comments to clarify the logic in nfs3_xdr_readdirres().

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/nfs/nfs3xdr.c |   37 ++++++++++++++++++++++++++++---------
 1 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c
index 3917e2f..fb03048 100644
--- a/fs/nfs/nfs3xdr.c
+++ b/fs/nfs/nfs3xdr.c
@@ -508,7 +508,7 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
 	struct page **page;
 	size_t hdrlen;
 	u32 len, recvd, pglen;
-	int status, nr;
+	int status, nr = 0;
 	__be32 *entry, *end, *kaddr;
 
 	status = ntohl(*p++);
@@ -542,7 +542,12 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
 	kaddr = p = kmap_atomic(*page, KM_USER0);
 	end = (__be32 *)((char *)p + pglen);
 	entry = p;
-	for (nr = 0; *p++; nr++) {
+
+	/* Make sure the packet actually has a value_follows and EOF entry */
+	if ((entry + 1) > end)
+		goto short_pkt;
+
+	for (; *p++; nr++) {
 		if (p + 3 > end)
 			goto short_pkt;
 		p += 2;				/* inode # */
@@ -581,18 +586,32 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
 			goto short_pkt;
 		entry = p;
 	}
-	if (!nr && (entry[0] != 0 || entry[1] == 0))
-		goto short_pkt;
+
+	/*
+	 * Apparently some server sends responses that are a valid size, but
+	 * contain no entries, and have value_follows==0 and EOF==0. For
+	 * those, just set the EOF marker.
+	 */
+	if (!nr && entry[1] == 0) {
+		dprintk("NFS: readdir reply truncated!\n");
+		entry[1] = 1;
+	}
  out:
 	kunmap_atomic(kaddr, KM_USER0);
 	return nr;
  short_pkt:
+	/*
+	 * When we get a short packet there are 2 possibilities. We can
+	 * return an error, or fix up the response to look like a valid
+	 * response and return what we have so far. If there are no
+	 * entries and the packet was short, then return -EIO. If there
+	 * are valid entries in the response, return them and pretend that
+	 * the call was successful, but incomplete. The caller can retry the
+	 * readdir starting at the last cookie.
+	 */
 	entry[0] = entry[1] = 0;
-	/* truncate listing ? */
-	if (!nr) {
-		dprintk("NFS: readdir reply truncated!\n");
-		entry[1] = 1;
-	}
+	if (!nr)
+		nr = -errno_NFSERR_IO;
 	goto out;
 err_unmap:
 	nr = -errno_NFSERR_IO;
-- 
1.5.4.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* [PATCH 3/3] NFS: clean up short packet handling for NFSv4 readdir
  2008-02-22 19:50   ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Jeff Layton
@ 2008-02-22 19:50     ` Jeff Layton
  2008-02-22 21:11     ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Chuck Lever
  1 sibling, 0 replies; 6+ messages in thread
From: Jeff Layton @ 2008-02-22 19:50 UTC (permalink / raw)
  To: Trond Myklebust; +Cc: linux-nfs, nfsv4

Currently, the NFS readdir decoders have a workaround for buggy servers
that send an empty readdir response with the EOF bit unset. If the
server sends a malformed response in some cases, this workaround kicks
in and just returns an empty response rather than returning a proper
error to the caller.

This patch does 3 things:

1) have malformed responses with no entries return error (-EIO)

2) preserve existing workaround for servers that send empty
   responses with the EOF marker unset.

3) Add some comments to clarify the logic in decode_readdir().

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/nfs/nfs4xdr.c |   37 +++++++++++++++++++++++++++----------
 1 files changed, 27 insertions(+), 10 deletions(-)

diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index db1ed9c..6a203ba 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -3481,7 +3481,7 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 	size_t		hdrlen;
 	u32		recvd, pglen = rcvbuf->page_len;
 	__be32		*end, *entry, *p, *kaddr;
-	unsigned int	nr;
+	unsigned int	nr = 0;
 	int		status;
 
 	status = decode_op_hdr(xdr, OP_READDIR);
@@ -3505,7 +3505,12 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 	kaddr = p = kmap_atomic(page, KM_USER0);
 	end = p + ((pglen + readdir->pgbase) >> 2);
 	entry = p;
-	for (nr = 0; *p++; nr++) {
+
+	/* Make sure the packet actually has a value_follows and EOF entry */
+	if ((entry + 1) > end)
+		goto short_pkt;
+
+	for (; *p++; nr++) {
 		u32 len, attrlen, xlen;
 		if (end - p < 3)
 			goto short_pkt;
@@ -3532,20 +3537,32 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 		p += attrlen;		/* attributes */
 		entry = p;
 	}
-	if (!nr && (entry[0] != 0 || entry[1] == 0))
-		goto short_pkt;
+	/*
+	 * Apparently some server sends responses that are a valid size, but
+	 * contain no entries, and have value_follows==0 and EOF==0. For
+	 * those, just set the EOF marker.
+	 */
+	if (!nr && entry[1] == 0) {
+		dprintk("NFS: readdir reply truncated!\n");
+		entry[1] = 1;
+	}
 out:	
 	kunmap_atomic(kaddr, KM_USER0);
 	return 0;
 short_pkt:
+	/*
+	 * When we get a short packet there are 2 possibilities. We can
+	 * return an error, or fix up the response to look like a valid
+	 * response and return what we have so far. If there are no
+	 * entries and the packet was short, then return -EIO. If there
+	 * are valid entries in the response, return them and pretend that
+	 * the call was successful, but incomplete. The caller can retry the
+	 * readdir starting at the last cookie.
+	 */
 	dprintk("%s: short packet at entry %d\n", __FUNCTION__, nr);
 	entry[0] = entry[1] = 0;
-	/* truncate listing ? */
-	if (!nr) {
-		dprintk("NFS: readdir reply truncated!\n");
-		entry[1] = 1;
-	}
-	goto out;
+	if (nr)
+		goto out;
 err_unmap:
 	kunmap_atomic(kaddr, KM_USER0);
 	return -errno_NFSERR_IO;
-- 
1.5.4.1

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir
  2008-02-22 19:50   ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Jeff Layton
  2008-02-22 19:50     ` [PATCH 3/3] NFS: clean up short packet handling for NFSv4 readdir Jeff Layton
@ 2008-02-22 21:11     ` Chuck Lever
  2008-02-22 21:26       ` Jeff Layton
  1 sibling, 1 reply; 6+ messages in thread
From: Chuck Lever @ 2008-02-22 21:11 UTC (permalink / raw)
  To: Jeff Layton; +Cc: linux-nfs, nfsv4

[-- Attachment #1: Type: text/plain, Size: 3049 bytes --]

Hi Jeff-

For NFSv3, is the READDIRPLUS decoder affected as well?

Jeff Layton wrote:
> Currently, the NFS readdir decoders have a workaround for buggy servers
> that send an empty readdir response with the EOF bit unset. If the
> server sends a malformed response in some cases, this workaround kicks
> in and just returns an empty response rather than returning a proper
> error to the caller.
> 
> This patch does 3 things:
> 
> 1) have malformed responses with no entries return error (-EIO)
> 
> 2) preserve existing workaround for servers that send empty
>    responses with the EOF marker unset.
> 
> 3) Add some comments to clarify the logic in nfs3_xdr_readdirres().
> 
> Signed-off-by: Jeff Layton <jlayton@redhat.com>
> ---
>  fs/nfs/nfs3xdr.c |   37 ++++++++++++++++++++++++++++---------
>  1 files changed, 28 insertions(+), 9 deletions(-)
> 
> diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c
> index 3917e2f..fb03048 100644
> --- a/fs/nfs/nfs3xdr.c
> +++ b/fs/nfs/nfs3xdr.c
> @@ -508,7 +508,7 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
>  	struct page **page;
>  	size_t hdrlen;
>  	u32 len, recvd, pglen;
> -	int status, nr;
> +	int status, nr = 0;
>  	__be32 *entry, *end, *kaddr;
>  
>  	status = ntohl(*p++);
> @@ -542,7 +542,12 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
>  	kaddr = p = kmap_atomic(*page, KM_USER0);
>  	end = (__be32 *)((char *)p + pglen);
>  	entry = p;
> -	for (nr = 0; *p++; nr++) {
> +
> +	/* Make sure the packet actually has a value_follows and EOF entry */
> +	if ((entry + 1) > end)
> +		goto short_pkt;
> +
> +	for (; *p++; nr++) {
>  		if (p + 3 > end)
>  			goto short_pkt;
>  		p += 2;				/* inode # */
> @@ -581,18 +586,32 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
>  			goto short_pkt;
>  		entry = p;
>  	}
> -	if (!nr && (entry[0] != 0 || entry[1] == 0))
> -		goto short_pkt;
> +
> +	/*
> +	 * Apparently some server sends responses that are a valid size, but
> +	 * contain no entries, and have value_follows==0 and EOF==0. For
> +	 * those, just set the EOF marker.
> +	 */
> +	if (!nr && entry[1] == 0) {
> +		dprintk("NFS: readdir reply truncated!\n");
> +		entry[1] = 1;
> +	}
>   out:
>  	kunmap_atomic(kaddr, KM_USER0);
>  	return nr;
>   short_pkt:
> +	/*
> +	 * When we get a short packet there are 2 possibilities. We can
> +	 * return an error, or fix up the response to look like a valid
> +	 * response and return what we have so far. If there are no
> +	 * entries and the packet was short, then return -EIO. If there
> +	 * are valid entries in the response, return them and pretend that
> +	 * the call was successful, but incomplete. The caller can retry the
> +	 * readdir starting at the last cookie.
> +	 */
>  	entry[0] = entry[1] = 0;
> -	/* truncate listing ? */
> -	if (!nr) {
> -		dprintk("NFS: readdir reply truncated!\n");
> -		entry[1] = 1;
> -	}
> +	if (!nr)
> +		nr = -errno_NFSERR_IO;
>  	goto out;
>  err_unmap:
>  	nr = -errno_NFSERR_IO;

[-- Attachment #2: chuck_lever.vcf --]
[-- Type: text/x-vcard, Size: 259 bytes --]

begin:vcard
fn:Chuck Lever
n:Lever;Chuck
org:Oracle Corporation;Corporate Architecture: Linux Projects Group
adr:;;1015 Granger Avenue;Ann Arbor;MI;48104;USA
title:Principal Member of Staff
tel;work:+1 248 614 5091
x-mozilla-html:FALSE
version:2.1
end:vcard


[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

_______________________________________________
NFSv4 mailing list
NFSv4@linux-nfs.org
http://linux-nfs.org/cgi-bin/mailman/listinfo/nfsv4

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir
  2008-02-22 21:11     ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Chuck Lever
@ 2008-02-22 21:26       ` Jeff Layton
  0 siblings, 0 replies; 6+ messages in thread
From: Jeff Layton @ 2008-02-22 21:26 UTC (permalink / raw)
  To: chuck.lever; +Cc: linux-nfs, nfsv4

On Fri, 22 Feb 2008 16:11:12 -0500
Chuck Lever <chuck.lever@oracle.com> wrote:

> Hi Jeff-
> 
> For NFSv3, is the READDIRPLUS decoder affected as well?
> 

Yes. READDIR and READDIRPLUS use the same decoder routine, so this patch
should also make short READDIRPLUS packets return an error using the
same set of rules.

> Jeff Layton wrote:
> > Currently, the NFS readdir decoders have a workaround for buggy servers
> > that send an empty readdir response with the EOF bit unset. If the
> > server sends a malformed response in some cases, this workaround kicks
> > in and just returns an empty response rather than returning a proper
> > error to the caller.
> > 
> > This patch does 3 things:
> > 
> > 1) have malformed responses with no entries return error (-EIO)
> > 
> > 2) preserve existing workaround for servers that send empty
> >    responses with the EOF marker unset.
> > 
> > 3) Add some comments to clarify the logic in nfs3_xdr_readdirres().
> > 
> > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > ---
> >  fs/nfs/nfs3xdr.c |   37 ++++++++++++++++++++++++++++---------
> >  1 files changed, 28 insertions(+), 9 deletions(-)
> > 
> > diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c
> > index 3917e2f..fb03048 100644
> > --- a/fs/nfs/nfs3xdr.c
> > +++ b/fs/nfs/nfs3xdr.c
> > @@ -508,7 +508,7 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
> >  	struct page **page;
> >  	size_t hdrlen;
> >  	u32 len, recvd, pglen;
> > -	int status, nr;
> > +	int status, nr = 0;
> >  	__be32 *entry, *end, *kaddr;
> >  
> >  	status = ntohl(*p++);
> > @@ -542,7 +542,12 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
> >  	kaddr = p = kmap_atomic(*page, KM_USER0);
> >  	end = (__be32 *)((char *)p + pglen);
> >  	entry = p;
> > -	for (nr = 0; *p++; nr++) {
> > +
> > +	/* Make sure the packet actually has a value_follows and EOF entry */
> > +	if ((entry + 1) > end)
> > +		goto short_pkt;
> > +
> > +	for (; *p++; nr++) {
> >  		if (p + 3 > end)
> >  			goto short_pkt;
> >  		p += 2;				/* inode # */
> > @@ -581,18 +586,32 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res
> >  			goto short_pkt;
> >  		entry = p;
> >  	}
> > -	if (!nr && (entry[0] != 0 || entry[1] == 0))
> > -		goto short_pkt;
> > +
> > +	/*
> > +	 * Apparently some server sends responses that are a valid size, but
> > +	 * contain no entries, and have value_follows==0 and EOF==0. For
> > +	 * those, just set the EOF marker.
> > +	 */
> > +	if (!nr && entry[1] == 0) {
> > +		dprintk("NFS: readdir reply truncated!\n");
> > +		entry[1] = 1;
> > +	}
> >   out:
> >  	kunmap_atomic(kaddr, KM_USER0);
> >  	return nr;
> >   short_pkt:
> > +	/*
> > +	 * When we get a short packet there are 2 possibilities. We can
> > +	 * return an error, or fix up the response to look like a valid
> > +	 * response and return what we have so far. If there are no
> > +	 * entries and the packet was short, then return -EIO. If there
> > +	 * are valid entries in the response, return them and pretend that
> > +	 * the call was successful, but incomplete. The caller can retry the
> > +	 * readdir starting at the last cookie.
> > +	 */
> >  	entry[0] = entry[1] = 0;
> > -	/* truncate listing ? */
> > -	if (!nr) {
> > -		dprintk("NFS: readdir reply truncated!\n");
> > -		entry[1] = 1;
> > -	}
> > +	if (!nr)
> > +		nr = -errno_NFSERR_IO;
> >  	goto out;
> >  err_unmap:
> >  	nr = -errno_NFSERR_IO;


-- 
Jeff Layton <jlayton@redhat.com>

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2008-02-22 21:26 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-02-22 19:49 [PATCH 0/3] Should truncated READDIR replies return -EIO? Jeff Layton
2008-02-22 19:49 ` [PATCH 1/3] NFS: clean up short packet handling for NFSv2 readdir Jeff Layton
2008-02-22 19:50   ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Jeff Layton
2008-02-22 19:50     ` [PATCH 3/3] NFS: clean up short packet handling for NFSv4 readdir Jeff Layton
2008-02-22 21:11     ` [PATCH 2/3] NFS: clean up short packet handling for NFSv3 readdir Chuck Lever
2008-02-22 21:26       ` Jeff Layton

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox