linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Samuel Mendoza-Jonas <sam@mendozajonas.com>
To: Joel Stanley <joel@jms.id.au>,
	Kumar Thangavel <kumarthangavel.hcl@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>,
	linux-aspeed <linux-aspeed@lists.ozlabs.org>,
	Networking <netdev@vger.kernel.org>,
	OpenBMC Maillist <openbmc@lists.ozlabs.org>,
	Amithash Prasad <amithash@fb.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	velumanit@hcl.com, patrickw3@fb.com
Subject: Re: [PATCH] net: ncsi: Adding padding bytes in the payload
Date: Tue, 12 Oct 2021 19:46:28 -0700	[thread overview]
Message-ID: <dd6df3ee84110649d901f041883c3d44c0798cda.camel@mendozajonas.com> (raw)
In-Reply-To: <CACPK8XcJudWoKgXORvRzGAbtBwHm3a56RULriVABfERZgNgt9w@mail.gmail.com>

On Tue, 2021-10-12 at 22:44 +0000, Joel Stanley wrote:
> On Tue, 12 Oct 2021 at 06:23, Kumar Thangavel
> <kumarthangavel.hcl@gmail.com> wrote:
> > 
> > Update NC-SI command handler (both standard and OEM) to take into
> > account of payload paddings in allocating skb (in case of payload
> > size is not 32-bit aligned).
> > 
> > The checksum field follows payload field, without taking payload
> > padding into account can cause checksum being truncated, leading to
> > dropped packets.
> 
> Can you help us review this by pointing out where this is described in
> the NCSI spec?
> 
> We've been running this code for a number of years now and I wonder
> why this hasn't been a problem so far.

I'm assuming this is referencing section 8.2.2.2:
If the payload is present and does not end on a 32-bit boundary, one to
three padding bytes equal to 0x00 shall be present to align the
checksum field to a 32-bit boundary.

But I'm also surprised this hasn't caused issues so far if we've been
getting it wrong. Is there an example that reproduces the issue?

Cheers,
Sam

> 
> Cheers,
> 
> Joel
> 
> > 
> > Signed-off-by: Kumar Thangavel <thangavel.k@hcl.com>
> > 
> > ---
> >  net/ncsi/ncsi-cmd.c | 21 +++++++++++++++++----
> >  1 file changed, 17 insertions(+), 4 deletions(-)
> > 
> > diff --git a/net/ncsi/ncsi-cmd.c b/net/ncsi/ncsi-cmd.c
> > index ba9ae482141b..4625fc935603 100644
> > --- a/net/ncsi/ncsi-cmd.c
> > +++ b/net/ncsi/ncsi-cmd.c
> > @@ -214,11 +214,19 @@ static int ncsi_cmd_handler_oem(struct sk_buff
> > *skb,
> >         struct ncsi_cmd_oem_pkt *cmd;
> >         unsigned int len;
> > 
> > +       /* NC-SI spec requires payload to be padded with 0
> > +        * to 32-bit boundary before the checksum field.
> > +        * Ensure the padding bytes are accounted for in
> > +        * skb allocation
> > +        */
> > +
> > +       unsigned short payload = ALIGN(nca->payload, 4);
> > +
> >         len = sizeof(struct ncsi_cmd_pkt_hdr) + 4;
> > -       if (nca->payload < 26)
> > +       if (payload < 26)
> >                 len += 26;
> >         else
> > -               len += nca->payload;
> > +               len += payload;
> > 
> >         cmd = skb_put_zero(skb, len);
> >         memcpy(&cmd->mfr_id, nca->data, nca->payload);
> > @@ -272,6 +280,7 @@ static struct ncsi_request
> > *ncsi_alloc_command(struct ncsi_cmd_arg *nca)
> >         struct net_device *dev = nd->dev;
> >         int hlen = LL_RESERVED_SPACE(dev);
> >         int tlen = dev->needed_tailroom;
> > +       int payload;
> >         int len = hlen + tlen;
> >         struct sk_buff *skb;
> >         struct ncsi_request *nr;
> > @@ -281,14 +290,18 @@ static struct ncsi_request
> > *ncsi_alloc_command(struct ncsi_cmd_arg *nca)
> >                 return NULL;
> > 
> >         /* NCSI command packet has 16-bytes header, payload, 4 bytes
> > checksum.
> > +        * Payload needs padding so that the checksum field follwoing
> > payload is
> > +        * aligned to 32bit boundary.
> >          * The packet needs padding if its payload is less than 26
> > bytes to
> >          * meet 64 bytes minimal ethernet frame length.
> >          */
> >         len += sizeof(struct ncsi_cmd_pkt_hdr) + 4;
> > -       if (nca->payload < 26)
> > +
> > +       payload = ALIGN(nca->payload, 4);
> > +       if (payload < 26)
> >                 len += 26;
> >         else
> > -               len += nca->payload;
> > +               len += payload;
> > 
> >         /* Allocate skb */
> >         skb = alloc_skb(len, GFP_ATOMIC);
> > --
> > 2.17.1
> > 



      reply	other threads:[~2021-10-13  2:46 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-12  6:22 [PATCH] net: ncsi: Adding padding bytes in the payload Kumar Thangavel
2021-10-12  8:02 ` Paul Menzel
2021-10-12 22:44 ` Joel Stanley
2021-10-13  2:46   ` Samuel Mendoza-Jonas [this message]

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=dd6df3ee84110649d901f041883c3d44c0798cda.camel@mendozajonas.com \
    --to=sam@mendozajonas.com \
    --cc=amithash@fb.com \
    --cc=davem@davemloft.net \
    --cc=joel@jms.id.au \
    --cc=kuba@kernel.org \
    --cc=kumarthangavel.hcl@gmail.com \
    --cc=linux-aspeed@lists.ozlabs.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=openbmc@lists.ozlabs.org \
    --cc=patrickw3@fb.com \
    --cc=velumanit@hcl.com \
    /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).