Re: [2.6 patch] let broken drivers depend on BROKEN{,ON_SMP}

[Adrian Bunk <bunk@fs.tum.de>]

On Wed, Jul 30, 2003 at 10:11:17AM +0100, John Bradford wrote:
>...
> >  * Driver does not work, and is thus disabled. If it is not
> >    fixed in the near future, it will be considered to be
> >    OBSOLETE as well.
> >
> > 		CONFIG_BROKEN
> 
> Please do _NOT_ do this - there is a far more important and worthwhile
> reason to have a CONFIG_BROKEN than to simply save the few minutes of
> inconvenience that including a non-compiling option in a kernel build
> causes.
> 
> Imagine the situation where a driver such as a SCSI driver builds
> successfully, but it silently corrupts data under certain, (possibly
> rare), circumstances.
> 
> In that case, it's important to warn people that it's broken, because
> it's not necessarily obvious, and could case significant data loss.
> If something doesn't compile, it already gives you an error message.
> The only problem is a few minutes of wasted time.

You forget one important thing:
If a _user_ of a stable kernel notices "it doesn't even compile" this 
gives a very bad impression of the quality of the Linux kernel.

> >  * Driver works on uniprocessor but not on SMP and is thus
> >    disabled when compiling for SMP. It is assumed that the
> >    driver will be fixed for SMP if relevant.
> >
> > 		CONFIG_BROKEN_ON_SMP
> 
> Please _don't_ do this either.  It implies that if
> CONFIG_BROKEN_ON_SMP isn't set, then it's SMP safe - a lot of drivers
> will NOT have been tested on SMP, so it's a bad thing to assume that
> is the case.
>...

My patch adds BROKEN_ON_SMP only to drivers that don't compile, but if a 
driver causes e.g. data corruption on SMP I don't see a reason against 
letting it depend on BROKEN_ON_SMP.

> John.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed