linux-kernel - Re: [PATCH net 1/6] hinic: fix process of long length skb without frags

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

Date:   Mon, 16 Mar 2020 14:44:08 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Luo bin <luobin9@...wei.com>
Cc:     <davem@...emloft.net>, <linux-kernel@...r.kernel.org>,
        <netdev@...r.kernel.org>, <aviad.krawczyk@...wei.com>,
        <luoxianjun@...wei.com>, <cloud.wangxiaoyun@...wei.com>,
        <yin.yinshi@...wei.com>
Subject: Re: [PATCH net 1/6] hinic: fix process of long length skb without
 frags

On Mon, 16 Mar 2020 00:56:25 +0000 Luo bin wrote:
> -#define MIN_SKB_LEN                     17
> +#define MIN_SKB_LEN			17
> +#define HINIC_GSO_MAX_SIZE		65536

> +	if (unlikely(skb->len > HINIC_GSO_MAX_SIZE && nr_sges == 1)) {
> +		txq->txq_stats.frag_len_overflow++;
> +		goto skb_error;
> +	}

I don't think drivers should have to check this condition.

We have netdev->gso_max_size which should be initialized to 

include/linux/netdevice.h:#define GSO_MAX_SIZE          65536

in

net/core/dev.c: dev->gso_max_size = GSO_MAX_SIZE;

Please send a patch to pktgen to uphold the normal stack guarantees.