From: Jeff Layton <jlayton@kernel.org>
To: chuck.lever@oracle.com
Cc: linux-nfs@vger.kernel.org, dcritch@redhat.com,
d.lesca@solinos.it, viro@zeniv.linux.org.uk
Subject: Re: [PATCH v2 2/2] sunrpc: add bounds checking to svc_rqst_replace_page
Date: Fri, 17 Mar 2023 13:51:49 -0400 [thread overview]
Message-ID: <c2fa806e20e55353ac0cfca2d4c1bfb0de6bbb5f.camel@kernel.org> (raw)
In-Reply-To: <20230317171309.73607-2-jlayton@kernel.org>
On Fri, 2023-03-17 at 13:13 -0400, Jeff Layton wrote:
> If rq_next_page ends up pointing outside the array, then we can corrupt
> memory when we go to change its value. Ensure that it hasn't strayed
> outside the array, and have svc_rqst_replace_page return -EIO without
> changing anything if it has.
>
> Fix up nfsd_splice_actor (the only caller) to handle this case by either
> returning an error or a short splice when this happens.
>
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
> fs/nfsd/vfs.c | 23 +++++++++++++++++------
> include/linux/sunrpc/svc.h | 2 +-
> net/sunrpc/svc.c | 14 +++++++++++++-
> 3 files changed, 31 insertions(+), 8 deletions(-)
>
> diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
> index 97b38b47c563..0ebd7a65a9f0 100644
> --- a/fs/nfsd/vfs.c
> +++ b/fs/nfsd/vfs.c
> @@ -939,6 +939,7 @@ nfsd_splice_actor(struct pipe_inode_info *pipe, struct pipe_buffer *buf,
> struct page *page = buf->page; // may be a compound one
> unsigned offset = buf->offset;
> struct page *last_page;
> + int ret = 0, consumed = 0;
>
> last_page = page + (offset + sd->len - 1) / PAGE_SIZE;
> for (page += offset / PAGE_SIZE; page <= last_page; page++) {
> @@ -946,13 +947,23 @@ nfsd_splice_actor(struct pipe_inode_info *pipe, struct pipe_buffer *buf,
> * Skip page replacement when extending the contents
> * of the current page.
> */
> - if (page != *(rqstp->rq_next_page - 1))
> - svc_rqst_replace_page(rqstp, page);
> + if (page != *(rqstp->rq_next_page - 1)) {
> + ret = svc_rqst_replace_page(rqstp, page);
> + if (ret)
> + break;
> + }
> + consumed += min_t(int,
> + PAGE_SIZE - offset_in_page(offset),
> + sd->len - consumed);
> + offset = 0;
> }
> - if (rqstp->rq_res.page_len == 0) // first call
> - rqstp->rq_res.page_base = offset % PAGE_SIZE;
> - rqstp->rq_res.page_len += sd->len;
> - return sd->len;
> + if (consumed) {
> + if (rqstp->rq_res.page_len == 0) // first call
> + rqstp->rq_res.page_base = offset % PAGE_SIZE;
Oops, this will get the page_base wrong if the first splice covers
multiple pages. I'll need to respin the offset handling.
> + rqstp->rq_res.page_len += consumed;
> + return consumed;
> + }
> + return ret;
> }
>
> static int nfsd_direct_splice_actor(struct pipe_inode_info *pipe,
> diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h
> index 877891536c2f..9ea52f143f49 100644
> --- a/include/linux/sunrpc/svc.h
> +++ b/include/linux/sunrpc/svc.h
> @@ -422,7 +422,7 @@ struct svc_serv *svc_create(struct svc_program *, unsigned int,
> int (*threadfn)(void *data));
> struct svc_rqst *svc_rqst_alloc(struct svc_serv *serv,
> struct svc_pool *pool, int node);
> -void svc_rqst_replace_page(struct svc_rqst *rqstp,
> +int svc_rqst_replace_page(struct svc_rqst *rqstp,
> struct page *page);
> void svc_rqst_free(struct svc_rqst *);
> void svc_exit_thread(struct svc_rqst *);
> diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
> index fea7ce8fba14..d624c02f09be 100644
> --- a/net/sunrpc/svc.c
> +++ b/net/sunrpc/svc.c
> @@ -843,8 +843,19 @@ EXPORT_SYMBOL_GPL(svc_set_num_threads);
> * When replacing a page in rq_pages, batch the release of the
> * replaced pages to avoid hammering the page allocator.
> */
> -void svc_rqst_replace_page(struct svc_rqst *rqstp, struct page *page)
> +int svc_rqst_replace_page(struct svc_rqst *rqstp, struct page *page)
> {
> + struct page **begin, **end;
> +
> + /*
> + * Bounds check: make sure rq_next_page points into the rq_respages
> + * part of the array.
> + */
> + begin = rqstp->rq_pages;
> + end = &rqstp->rq_pages[RPCSVC_MAXPAGES];
> + if (WARN_ON_ONCE(rqstp->rq_next_page < begin || rqstp->rq_next_page > end))
> + return -EIO;
> +
> if (*rqstp->rq_next_page) {
> if (!pagevec_space(&rqstp->rq_pvec))
> __pagevec_release(&rqstp->rq_pvec);
> @@ -853,6 +864,7 @@ void svc_rqst_replace_page(struct svc_rqst *rqstp, struct page *page)
>
> get_page(page);
> *(rqstp->rq_next_page++) = page;
> + return 0;
> }
> EXPORT_SYMBOL_GPL(svc_rqst_replace_page);
>
--
Jeff Layton <jlayton@kernel.org>
next prev parent reply other threads:[~2023-03-17 17:52 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-17 17:13 [PATCH v2 1/2] nfsd: don't replace page in rq_pages if it's a continuation of last page Jeff Layton
2023-03-17 17:13 ` [PATCH v2 2/2] sunrpc: add bounds checking to svc_rqst_replace_page Jeff Layton
2023-03-17 17:51 ` Jeff Layton [this message]
2023-03-17 17:51 ` Chuck Lever III
2023-03-17 18:04 ` Jeff Layton
2023-03-17 18:08 ` Chuck Lever III
2023-03-17 18:59 ` Jeff Layton
2023-03-17 20:55 ` Chuck Lever III
2023-03-17 22:10 ` Jeff Layton
2023-03-17 18:32 ` Chuck Lever III
2023-03-17 18:52 ` Jeff Layton
2023-03-17 17:33 ` [PATCH v2 1/2] nfsd: don't replace page in rq_pages if it's a continuation of last page Chuck Lever III
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=c2fa806e20e55353ac0cfca2d4c1bfb0de6bbb5f.camel@kernel.org \
--to=jlayton@kernel.org \
--cc=chuck.lever@oracle.com \
--cc=d.lesca@solinos.it \
--cc=dcritch@redhat.com \
--cc=linux-nfs@vger.kernel.org \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox