netdev - Re: [PATCH net-next] vmxnet3: use correct tcp hdr length when packet is encapsulated

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20200807082300.27876448@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date:   Fri, 7 Aug 2020 08:23:00 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Ronak Doshi <doshir@...are.com>
Cc:     <netdev@...r.kernel.org>, "VMware, Inc." <pv-drivers@...are.com>,
        "David S. Miller" <davem@...emloft.net>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net-next] vmxnet3: use correct tcp hdr length when
 packet is encapsulated

On Thu, 6 Aug 2020 23:43:45 -0700 Ronak Doshi wrote:
> 'Commit dacce2be3312 ("vmxnet3: add geneve and vxlan tunnel offload
> support")' added support for encapsulation offload. However, while

nit: no need for the quotes around commit xzy ("")

> calculating tcp hdr length, it does not take into account if the
> packet is encapsulated or not.
> 
> This patch fixes this issue by using correct reference for inner
> tcp header.
> 
> Fixes: dacce2be3312 ("vmxnet3: add geneve and vxlan tunnel offload
> support")

Please don't wrap Fixes tags, let it overflow the line.

> Signed-off-by: Ronak Doshi <doshir@...are.com>
> Acked-by: Guolin Yang <gyang@...are.com>
> ---
>  drivers/net/vmxnet3/vmxnet3_drv.c | 3 ++-
>  drivers/net/vmxnet3/vmxnet3_int.h | 4 ++--
>  2 files changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c
> index ca395f9679d0..2818015324b8 100644
> --- a/drivers/net/vmxnet3/vmxnet3_drv.c
> +++ b/drivers/net/vmxnet3/vmxnet3_drv.c
> @@ -886,7 +886,8 @@ vmxnet3_parse_hdr(struct sk_buff *skb, struct vmxnet3_tx_queue *tq,
>  
>  			switch (protocol) {
>  			case IPPROTO_TCP:
> -				ctx->l4_hdr_size = tcp_hdrlen(skb);
> +				ctx->l4_hdr_size = skb->encapsulation ? inner_tcp_hdrlen(skb) :
> +						   tcp_hdrlen(skb);
>  				break;
>  			case IPPROTO_UDP:
>  				ctx->l4_hdr_size = sizeof(struct udphdr);
> diff --git a/drivers/net/vmxnet3/vmxnet3_int.h b/drivers/net/vmxnet3/vmxnet3_int.h
> index 5d2b062215a2..f99e3327a7b0 100644
> --- a/drivers/net/vmxnet3/vmxnet3_int.h
> +++ b/drivers/net/vmxnet3/vmxnet3_int.h
> @@ -69,12 +69,12 @@
>  /*
>   * Version numbers
>   */
> -#define VMXNET3_DRIVER_VERSION_STRING   "1.5.0.0-k"
> +#define VMXNET3_DRIVER_VERSION_STRING   "1.5.1.0-k"

Please don't do this.

We are working on removing all the driver version strings from the
kernel because upstream they are strictly inferior to the kernel
version alone.

For a fix like this it's especially bad to bump the version, as it
could cause a conflict during backporting. I'm not saying it will in
this case, but as a general rule fixes should be minimal in size.

>  /* Each byte of this 32-bit integer encodes a version number in
>   * VMXNET3_DRIVER_VERSION_STRING.
>   */
> -#define VMXNET3_DRIVER_VERSION_NUM      0x01050000
> +#define VMXNET3_DRIVER_VERSION_NUM      0x01050100
>  
>  #if defined(CONFIG_PCI_MSI)
>  	/* RSS only makes sense if MSI-X is supported. */