From: Jan Beulich <jbeulich@suse.com>
To: Juergen Gross <jgross@suse.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>,
Stefano Stabellini <sstabellini@kernel.org>,
"David S. Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>,
xen-devel@lists.xenproject.org, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 6/8] xen/netfront: don't read data from request on the ring page
Date: Mon, 17 May 2021 17:08:13 +0200 [thread overview]
Message-ID: <72cb5265-aca2-8717-6860-19b66ccf9191@suse.com> (raw)
In-Reply-To: <20210513100302.22027-7-jgross@suse.com>
On 13.05.2021 12:03, Juergen Gross wrote:
> In order to avoid a malicious backend being able to influence the local
> processing of a request build the request locally first and then copy
> it to the ring page. Any reading from the request needs to be done on
> the local instance.
"Any reading" isn't really true - you don't change xennet_make_one_txreq(),
yet that has a read-modify-write operation. Without that I would have
been inclined to ask whether ...
> --- a/drivers/net/xen-netfront.c
> +++ b/drivers/net/xen-netfront.c
> @@ -435,7 +435,8 @@ struct xennet_gnttab_make_txreq {
> struct netfront_queue *queue;
> struct sk_buff *skb;
> struct page *page;
> - struct xen_netif_tx_request *tx; /* Last request */
> + struct xen_netif_tx_request *tx; /* Last request on ring page */
> + struct xen_netif_tx_request tx_local; /* Last request local copy*/
... retaining the tx field here is a good idea.
> @@ -463,30 +464,27 @@ static void xennet_tx_setup_grant(unsigned long gfn, unsigned int offset,
> queue->grant_tx_page[id] = page;
> queue->grant_tx_ref[id] = ref;
>
> - tx->id = id;
> - tx->gref = ref;
> - tx->offset = offset;
> - tx->size = len;
> - tx->flags = 0;
> + info->tx_local.id = id;
> + info->tx_local.gref = ref;
> + info->tx_local.offset = offset;
> + info->tx_local.size = len;
> + info->tx_local.flags = 0;
> +
> + *tx = info->tx_local;
>
> info->tx = tx;
> - info->size += tx->size;
> + info->size += info->tx_local.size;
> }
>
> static struct xen_netif_tx_request *xennet_make_first_txreq(
> - struct netfront_queue *queue, struct sk_buff *skb,
> - struct page *page, unsigned int offset, unsigned int len)
> + struct xennet_gnttab_make_txreq *info,
> + unsigned int offset, unsigned int len)
> {
> - struct xennet_gnttab_make_txreq info = {
> - .queue = queue,
> - .skb = skb,
> - .page = page,
> - .size = 0,
> - };
> + info->size = 0;
>
> - gnttab_for_one_grant(page, offset, len, xennet_tx_setup_grant, &info);
> + gnttab_for_one_grant(info->page, offset, len, xennet_tx_setup_grant, info);
>
> - return info.tx;
> + return info->tx;
> }
Similarly this returning of a pointer into the ring looks at least
risky to me. At the very least it looks as if ...
> @@ -704,14 +699,16 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
> }
>
> /* First request for the linear area. */
> - first_tx = tx = xennet_make_first_txreq(queue, skb,
> - page, offset, len);
> + info.queue = queue;
> + info.skb = skb;
> + info.page = page;
> + first_tx = tx = xennet_make_first_txreq(&info, offset, len);
... you could avoid setting tx here; perhaps the local variable
could go away altogether, showing it's really just first_rx that
is still needed. It's odd that ...
> offset += tx->size;
... you don't change this one, when ...
> if (offset == PAGE_SIZE) {
> page++;
> offset = 0;
> }
> - len -= tx->size;
> + len -= info.tx_local.size;
... you do so here. Likely just an oversight.
Jan
next prev parent reply other threads:[~2021-05-17 15:08 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-13 10:02 [PATCH 0/8] xen: harden frontends against malicious backends Juergen Gross
2021-05-13 10:02 ` [PATCH 1/8] xen: sync include/xen/interface/io/ring.h with Xen's newest version Juergen Gross
2021-05-13 10:02 ` [PATCH 2/8] xen/blkfront: read response from backend only once Juergen Gross
2021-05-17 13:50 ` Jan Beulich
2021-05-13 10:02 ` [PATCH 3/8] xen/blkfront: don't take local copy of a request from the ring page Juergen Gross
2021-05-17 14:01 ` Jan Beulich
2021-05-17 14:11 ` Juergen Gross
2021-05-13 10:02 ` [PATCH 4/8] xen/blkfront: don't trust the backend response data blindly Juergen Gross
2021-05-17 14:11 ` Jan Beulich
2021-05-17 14:23 ` Juergen Gross
2021-05-17 15:12 ` Jan Beulich
2021-05-17 15:22 ` Juergen Gross
2021-05-17 15:33 ` Jan Beulich
2021-07-08 5:47 ` Juergen Gross
2021-07-08 6:37 ` Jan Beulich
2021-07-08 6:40 ` Juergen Gross
2021-07-08 6:52 ` Jan Beulich
2021-07-08 6:56 ` Juergen Gross
2021-05-13 10:02 ` [PATCH 5/8] xen/netfront: read response from backend only once Juergen Gross
2021-05-17 14:20 ` Jan Beulich
2021-05-17 14:24 ` Juergen Gross
2021-05-13 10:03 ` [PATCH 6/8] xen/netfront: don't read data from request on the ring page Juergen Gross
2021-05-17 15:08 ` Jan Beulich [this message]
2021-05-13 10:03 ` [PATCH 7/8] xen/netfront: don't trust the backend response data blindly Juergen Gross
2021-05-17 15:31 ` Jan Beulich
2021-05-13 10:03 ` [PATCH 8/8] xen/hvc: replace BUG_ON() with negative return value Juergen Gross
2021-05-13 10:16 ` Christophe Leroy
2021-05-13 10:20 ` Juergen Gross
2021-05-13 10:25 ` Greg Kroah-Hartman
2021-05-13 10:35 ` Juergen Gross
2021-05-21 10:43 ` [PATCH 0/8] xen: harden frontends against malicious backends Marek Marczykowski-Górecki
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=72cb5265-aca2-8717-6860-19b66ccf9191@suse.com \
--to=jbeulich@suse.com \
--cc=boris.ostrovsky@oracle.com \
--cc=davem@davemloft.net \
--cc=jgross@suse.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=sstabellini@kernel.org \
--cc=xen-devel@lists.xenproject.org \
/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;
as well as URLs for NNTP newsgroup(s).