Re: linux-5.17-rc5/arch/mips/cavium-octeon/executive/octeon-model.c bug report

Re: linux-5.17-rc5/arch/mips/cavium-octeon/executive/octeon-model.c bug report

[Thomas Bogendoerfer]

On Tue, Feb 22, 2022 at 05:00:26PM +0000, David Binderman wrote:
> I just ran static analyser cppcheck over the linux kernel 5.17-rc5. It said:
> 
> linux-5.17-rc5/arch/mips/cavium-octeon/executive/octeon-model.c:359:11: style: Variable 'suffix' is reassigned a value before the old one has been used. [redundantAssignment]
> 
> Source code is
> 
>                 if (fus_dat2.cn66xx.nocrypto && fus_dat2.cn66xx.dorm_crypto)
>                         suffix = "AP";
>                 if (fus_dat2.cn66xx.nocrypto)
>                         suffix = "CP";
> 
> Maybe better code:
> 
>                 if (fus_dat2.cn66xx.nocrypto && fus_dat2.cn66xx.dorm_crypto)
>                         suffix = "AP";
>                 else if (fus_dat2.cn66xx.nocrypto)
>                         suffix = "CP";

IMHO this isn't better just different. As I don't have any Cavium docs
about their SoCs, it's hard to say what's the correct way to fix this.
There is the same pattern in CN61XX case

                if (fus_dat2.cn61xx.nocrypto && fus_dat2.cn61xx.dorm_crypto)
                        suffix = "AP";
                if (fus_dat2.cn61xx.nocrypto)
                        suffix = "CP";

which isn't present in the latest Cavium SDK. So maybe just removing

                if (fus_dat2.cn66xx.nocrypto && fus_dat2.cn66xx.dorm_crypto)
                        suffix = "AP";

is more appropriate.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]