netdev - gianfar driver and GFAR_MQ

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

Message-ID: <ff918224-ce0c-9bdf-c5f9-932cb5d31e0d@prevas.dk>
Date:   Tue, 6 Apr 2021 11:56:15 +0200
From:   Rasmus Villemoes <rasmus.villemoes@...vas.dk>
To:     Claudiu Manoil <claudiu.manoil@....com>
Cc:     Network Development <netdev@...r.kernel.org>,
        Vladimir Oltean <vladimir.oltean@....com>,
        Ioana Ciornei <ioana.ciornei@....com>,
        Esben Haabendal <esben@...nix.com>
Subject: gianfar driver and GFAR_MQ_POLLING

Hi,

I noticed that gfar_of_init() has

	if (of_device_is_compatible(np, "fsl,etsec2")) {
		mode = MQ_MG_MODE;
		poll_mode = GFAR_SQ_POLLING;
	} else {
		mode = SQ_SG_MODE;
		poll_mode = GFAR_SQ_POLLING;
	}

i.e., poll_mode is always set to GFAR_SQ_POLLING, and I can't find
anywhere that GFAR_MQ_POLLING is used (except for comments). So it seems
that everything in the else branches of the "if (priv->poll_mode ==
GFAR_SQ_POLLING)"s, including the gfar_poll_rx and gfar_poll_tx
callbacks, is currently dead code.

I'm wondering if this is deliberate, and if so, if the dead code
could/should just be removed?

FWIW, some quick testing naively doing

        if (of_device_is_compatible(np, "fsl,etsec2")) {
                mode = MQ_MG_MODE;
-               poll_mode = GFAR_SQ_POLLING;
+               poll_mode = GFAR_MQ_POLLING;
        } else {

results in broken network - ping answers the first packet, but then
nothing after that, and ssh is completely broken (and if ssh is
attempted first, ping doesn't work at all). This is on a ls1021a-derived
board.

Rasmus