2014-04-04 10:01:01

by Jeff Layton

Subject: [PATCH] nfs: don't pass non-NULL-terminated string to pr_notice()

There is no guarantee that the strings in the nfs_cache_array will be
NULL-terminated. In the event that we end up hitting a readdir loop, we
need to ensure that we pass the warning message a properly-terminated
string.

Reported-by: Lachlan McIlroy <[email protected]>
Signed-off-by: Jeff Layton <[email protected]>
---
fs/nfs/dir.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 5c0b6ecc3a88..4689b125f9fe 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -304,12 +304,13 @@ int nfs_readdir_search_for_cookie(struct nfs_cache_array *array, nfs_readdir_des
if (ctx->duped > 0
&& ctx->dup_cookie == *desc->dir_cookie) {
if (printk_ratelimit()) {
+ char *name = kstrndup(array->array[i].string.name, array->array[i].string.len, GFP_KERNEL);
+
pr_notice("NFS: directory %pD2 contains a readdir loop."
"Please contact your server vendor. "
"The file: %s has duplicate cookie %llu\n",
- desc->file,
- array->array[i].string.name,
- *desc->dir_cookie);
+ desc->file, name, *desc->dir_cookie);
+ kfree(name);
}
status = -ELOOP;
goto out;
--
1.9.0

2014-04-05 12:46:04

by Jeff Layton

Subject: [PATCH v2] nfs: pass string length to pr_notice message about readdir loops

There is no guarantee that the strings in the nfs_cache_array will be
NULL-terminated. In the event that we end up hitting a readdir loop, we
need to ensure that we pass the warning message the length of the
string.

Reported-by: Lachlan McIlroy <[email protected]>
Signed-off-by: Jeff Layton <[email protected]>
---
fs/nfs/dir.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index e6f8e2c97653..0013d5fc1aed 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -306,10 +306,9 @@ int nfs_readdir_search_for_cookie(struct nfs_cache_array *array, nfs_readdir_des
if (printk_ratelimit()) {
pr_notice("NFS: directory %pD2 contains a readdir loop."
"Please contact your server vendor. "
- "The file: %s has duplicate cookie %llu\n",
- desc->file,
- array->array[i].string.name,
- *desc->dir_cookie);
+ "The file: %.*s has duplicate cookie %llu\n",
+ desc->file, array->array[i].string.len,
+ array->array[i].string.name, *desc->dir_cookie);
}
status = -ELOOP;
goto out;
--
1.9.0

2014-04-04 13:51:16

by Trond Myklebust

Subject: Re: [PATCH] nfs: don't pass non-NULL-terminated string to pr_notice()

On Apr 4, 2014, at 6:00, Jeff Layton <[email protected]> wrote:

> There is no guarantee that the strings in the nfs_cache_array will be
> NULL-terminated. In the event that we end up hitting a readdir loop, we
> need to ensure that we pass the warning message a properly-terminated
> string.
>
> Reported-by: Lachlan McIlroy <[email protected]>
> Signed-off-by: Jeff Layton <[email protected]>
> ---
> fs/nfs/dir.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
> index 5c0b6ecc3a88..4689b125f9fe 100644
> --- a/fs/nfs/dir.c
> +++ b/fs/nfs/dir.c
> @@ -304,12 +304,13 @@ int nfs_readdir_search_for_cookie(struct nfs_cache_array *array, nfs_readdir_des
> if (ctx->duped > 0
> && ctx->dup_cookie == *desc->dir_cookie) {
> if (printk_ratelimit()) {
> + char *name = kstrndup(array->array[i].string.name, array->array[i].string.len, GFP_KERNEL);
> +
> pr_notice("NFS: directory %pD2 contains a readdir loop."
> "Please contact your server vendor. "
> "The file: %s has duplicate cookie %llu\n",
> - desc->file,
> - array->array[i].string.name,
> - *desc->dir_cookie);
> + desc->file, name, *desc->dir_cookie);
> + kfree(name);
> }

Umm? Any reason why we couldn?t just use ?%.*s? ?

_________________________________
Trond Myklebust
Linux NFS client maintainer, PrimaryData
[email protected]

2014-04-04 14:01:27

by Jeff Layton

Subject: Re: [PATCH] nfs: don't pass non-NULL-terminated string to pr_notice()

On Fri, 4 Apr 2014 09:51:11 -0400
Trond Myklebust <[email protected]> wrote:

>
> On Apr 4, 2014, at 6:00, Jeff Layton <[email protected]> wrote:
>
> > There is no guarantee that the strings in the nfs_cache_array will be
> > NULL-terminated. In the event that we end up hitting a readdir loop, we
> > need to ensure that we pass the warning message a properly-terminated
> > string.
> >
> > Reported-by: Lachlan McIlroy <[email protected]>
> > Signed-off-by: Jeff Layton <[email protected]>
> > ---
> > fs/nfs/dir.c | 7 ++++---
> > 1 file changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
> > index 5c0b6ecc3a88..4689b125f9fe 100644
> > --- a/fs/nfs/dir.c
> > +++ b/fs/nfs/dir.c
> > @@ -304,12 +304,13 @@ int nfs_readdir_search_for_cookie(struct nfs_cache_array *array, nfs_readdir_des
> > if (ctx->duped > 0
> > && ctx->dup_cookie == *desc->dir_cookie) {
> > if (printk_ratelimit()) {
> > + char *name = kstrndup(array->array[i].string.name, array->array[i].string.len, GFP_KERNEL);
> > +
> > pr_notice("NFS: directory %pD2 contains a readdir loop."
> > "Please contact your server vendor. "
> > "The file: %s has duplicate cookie %llu\n",
> > - desc->file,
> > - array->array[i].string.name,
> > - *desc->dir_cookie);
> > + desc->file, name, *desc->dir_cookie);
> > + kfree(name);
> > }
>
> Umm? Any reason why we couldn?t just use ?%.*s? ?
>

No reason at all. I had never run across that...

I'll respin and repost -- thanks!
--
Jeff Layton <[email protected]>