netdev - Re: [PATCH 1/1] net: usb: cdc_mbim: add flag FLAG_SEND

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: <87a7sggnch.fsf@miraculix.mork.no>
Date:   Thu, 31 May 2018 11:56:30 +0200
From:   Bjørn Mork <bjorn@...k.no>
To:     Daniele Palmas <dnlplm@...il.com>
Cc:     Oliver Neukum <oliver@...kum.org>, netdev@...r.kernel.org,
        linux-usb@...r.kernel.org
Subject: Re: [PATCH 1/1] net: usb: cdc_mbim: add flag FLAG_SEND_ZLP

Daniele Palmas <dnlplm@...il.com> writes:

> Testing Telit LM940 with ICMP packets > 14552 bytes revealed that
> the modem needs FLAG_SEND_ZLP to properly work, otherwise the cdc
> mbim data interface won't be anymore responsive.
>
> Signed-off-by: Daniele Palmas <dnlplm@...il.com>

Acked-by: Bjørn Mork <bjorn@...k.no>

Should have thought of this... I noticed your discussion, but couldn't
reproduce the issues myself.  This explains why.

Do you happen to know if the device announces larger buffers than the
driver wants to use, or if this happens with the max sized buffers too?

You can easily check these values by comparing dwNtbInMaxSize and
dwNtbOutMaxSize (device maximum values) with rx_max and tx_max
(neogtiated values) using e.g

 grep . /sys/class/net/wwan0/cdc_ncm/*

It has never been 100% clear to me whether we should send the ZLP by
default if we've negotiated a smaller than max buffer. But the ZLP ought
to be redundant in any case, since the device knows the negotiated
buffer size. So I do believe our current interpretation makes sense.

Not that it matters.  There are obviously more than enough device
implementations violating this requirement to make it completely
pointless.

Bjørn