Re: [flashrom] [PATCH] Unsignify lengths and addresses in chip functions and structs

[Stefan Tauner]

On Wed, 23 Nov 2011 00:19:32 +0100
Carl-Daniel Hailfinger  wrote:
> Am 22.11.2011 23:09 schrieb Stefan Tauner:
> > should i commit the attached version?
>
> Please do, the ack is still valid.

thanks, r1470
> I wonder whether we should enable the sign warnings once the cleanups
> and remaining conversions are in. The big problem might be platform
> (OS/library version) dependent changes which we don't know of, but OTOH
> we won't know until we try it.

i have been playing around with -Wsign-conversion for this patch... and
at least on my gcc version it is way over-sensitive, for example
ichspi.c: In function ‘ich7_run_opcode’:
ichspi.c:742: warning: negative integer implicitly converted to unsigned type

that is:
    temp32 = REGREAD32(ICH7_REG_SPIA) & ~0x00FFFFFF;

together with -Werror -Wsign-conversion will probably explode somewhere
even if we try hard to get it right on dev platforms.