netdev - Re: [PATCH net-next v2 1/7] drivers: net: smc91x: Fix set but unused W=1 warning

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] [day] [month] [year] [list]

Date:   Fri, 6 Nov 2020 09:42:13 -0800
From:   Jakub Kicinski <kuba@...nel.org>
To:     David Laight <David.Laight@...LAB.COM>
Cc:     Andrew Lunn <andrew@...n.ch>, netdev <netdev@...r.kernel.org>,
        Nicolas Pitre <nico@...xnic.net>,
        "Lee Jones" <lee.jones@...aro.org>
Subject: Re: [PATCH net-next v2 1/7] drivers: net: smc91x: Fix set but
 unused W=1 warning

On Fri, 6 Nov 2020 08:48:47 +0000 David Laight wrote:
> > > +#define SMC_GET_PKT_HDR_STATUS(lp, status)				\
> > > +	do {								\
> > > +		if (SMC_32BIT(lp)) {					\
> > > +			unsigned int __val = SMC_inl(ioaddr, DATA_REG(lp)); \
> > > +			(status) = __val & 0xffff;			\
> > > +		} else {						\
> > > +			(status) = SMC_inw(ioaddr, DATA_REG(lp));	\
> > > +		}							\
> > > +	} while (0)  
> > 
> > This is the original/full macro:
> > 
> > #define SMC_GET_PKT_HDR(lp, status, length)				\
> > 	do {								\
> > 		if (SMC_32BIT(lp)) {				\
> > 			unsigned int __val = SMC_inl(ioaddr, DATA_REG(lp)); \
> > 			(status) = __val & 0xffff;			\
> > 			(length) = __val >> 16;				\
> > 		} else {						\
> > 			(status) = SMC_inw(ioaddr, DATA_REG(lp));	\
> > 			(length) = SMC_inw(ioaddr, DATA_REG(lp));	\
> > 		}							\
> > 	} while (0)
> > 
> > Note that it reads the same address twice in the else branch.
> > 
> > I'm 90% sure we can't remove the read here either so best treat it
> > like the ones in patch 3, right?  
> 
> One of the two SMC_inw() needs to use 'ioaddr + 2'.
> Probably the one for (length).
> 
> The code may also be buggy on BE systems.

More proof that this code is fragile.

Changing IO accesses is not acceptable in a "warning cleanup" patch,
unless it can be tested on real HW.

We can follow up on the issues you see separately, please.