netdev - Re: [PATCH net] qmi_wwan: Clone the skb when in pass-through mode

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]

Message-ID: <CAKfDRXgZeWCeGXhfukeeAGrHHUMtsHWRPEebUkZf07QCnU4CFw@mail.gmail.com>
Date:   Mon, 14 Jun 2021 17:49:16 +0200
From:   Kristian Evensen <kristian.evensen@...il.com>
To:     Bjørn Mork <bjorn@...k.no>
Cc:     Network Development <netdev@...r.kernel.org>,
        subashab@...eaurora.org
Subject: Re: [PATCH net] qmi_wwan: Clone the skb when in pass-through mode

Hi Bjørn,

On Mon, Jun 14, 2021 at 4:46 PM Bjørn Mork <bjorn@...k.no> wrote:
> Thanks for pointing this out.  But it still looks strange to me.  Why do
> we call netif_rx(skb) here instead of just returning 1 and leave that
> for usbnet_skb_return()?  With cloning we end up doing eth_type_trans()
> on the duplicate - is that wise?

Thanks for pointing this out. You are right and looking at the code
again, we should end up in usbnet_skb_return() if the call to netif_rx
is succesful. I do agree that calling netif_rx() is a bit strange,
considering that usbnet_skb_return() already calls this function.
Perhaps the issue is that we end up calling netif_rx() on the same skb
twice? rmnet also frees the skb when done with the de-aggregation.

I will do some more testing tomorrow and see if I can figure out
something. So far, the only thing I know is that if I try to perform a
more demanding transfer (like downloading a file) using qmi_wwan +
rmnet then I get a kernel oops immediatly.

Kristian