linux-kernel - Re: [PATCH 1/4] net: fix uninitialised access in mii_nway

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: <Z_hC-9C7Bc2lPrig@qasdev.system>
Date: Thu, 10 Apr 2025 23:15:23 +0100
From: Qasim Ijaz <qasdev00@...il.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: andrew+netdev@...n.ch, davem@...emloft.net, edumazet@...gle.com,
	kuba@...nel.org, pabeni@...hat.com, horms@...nel.org,
	linux-usb@...r.kernel.org, netdev@...r.kernel.org,
	linux-kernel@...r.kernel.org,
	syzbot <syzbot+3361c2d6f78a3e0892f9@...kaller.appspotmail.com>
Subject: Re: [PATCH 1/4] net: fix uninitialised access in mii_nway_restart()

On Tue, Mar 25, 2025 at 06:33:07AM -0700, Jakub Kicinski wrote:
> On Wed, 19 Mar 2025 11:21:53 +0000 Qasim Ijaz wrote:
> > --- a/drivers/net/mii.c
> > +++ b/drivers/net/mii.c
> > @@ -464,6 +464,8 @@ int mii_nway_restart (struct mii_if_info *mii)
> >  
> >  	/* if autoneg is off, it's an error */
> >  	bmcr = mii->mdio_read(mii->dev, mii->phy_id, MII_BMCR);
> > +	if (bmcr < 0)
> > +		return bmcr;
> >  
> >  	if (bmcr & BMCR_ANENABLE) {
> >  		bmcr |= BMCR_ANRESTART;
> 
> We error check just one mdio_read() but there's a whole bunch of them
> in this file. What's the expected behavior then? Are all of them buggy?
>

Hi Jakub

Apologies for my delayed response, I had another look at this and I
think my patch may be off a bit. You are correct that there are multiple
mdio_read() calls and looking at the mii.c file we can see that calls to
functions like mdio_read (and a lot of others) dont check return values.

So in light of this I think a better patch would be to not edit the 
mii.c file at all and just make ch9200_mdio_read return 0 on     
error. This way if mdio_read fails and 0 is returned, the         
check for "bmcr & BMCR_ANENABLE" won't be triggered and mii_nway_restart
will just return 0 and end. If we return a negative on error it may
contain the exact bit the function checks.

Similiar to this patch:
<https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=c68b2c9eba38>

If this sounds good, should i send another patch series with all the
changes? 

> This patch should be split into core and driver parts.
> -- 
> pw-bot: cr