Re: ctcm: rename READ/WRITE defines to avoid redefinitions

From: Colin King (gmail)
Date: Mon Apr 25 2022 - 05:04:23 EST

On 25/04/2022 09:38, Alexandra Winter wrote:

On 24.04.22 20:58, Colin King (gmail) wrote:

static analysis with cppcheck detected a potential null pointer deference with the following commit:

commit 3c09e2647b5e1f1f9fd383971468823c2505e1b0
Author: Ursula Braun <ursula.braun@xxxxxxxxxx>
Date:   Thu Aug 12 01:58:28 2010 +0000

    ctcm: rename READ/WRITE defines to avoid redefinitions

The analysis is as follows:

drivers/s390/net/ctcm_sysfs.c:43:8: note: Assuming that condition 'priv' is not redundant
 if (!(priv && priv->channel[CTCM_READ] && ndev)) {
drivers/s390/net/ctcm_sysfs.c:42:9: note: Null pointer dereference
 ndev = priv->channel[CTCM_READ]->netdev;

The code in question is as follows:

        ndev = priv->channel[CTCM_READ]->netdev;

        ^^ priv may be null, as per check below but it is being dereferenced when assigning ndev

        if (!(priv && priv->channel[CTCM_READ] && ndev)) {
                CTCM_DBF_TEXT(SETUP, CTC_DBF_ERROR, "bfnondev");
                return -ENODEV;


Thank you very much for reporting this, we will provide a patch.

Thanks for working on a fix. Much appreciated.

Do you have any special requests for the Reported-by flag? Or is
Reported-by: Colin King (gmail) <colin.i.king@xxxxxxxxx>
fine with you?

Can I have:

Reported by: Colin Ian King <colin.i.king@xxxxxxxxx>

Thank you!


Kind regards