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

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

On Wed, Jul 30, 2003 at 12:29:55PM +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.
> 
> I don't agree.  The stock kernel is a work in progress, and things get
> broken from time to time as a normal part of development.  Experienced
> users will realise that, and I wouldn't encourage inexperienced users
> to compile their own kernel from the stock trees anyway, because they
> could easily miss bugfixes, including data corruption and security
> ones, simply because they assume that they are in the mainline
> kernel.

Whether you like it or not:

Many non-kernel-hackers compile their own kernels.


Even if you wouldn't encourage them, there are enough situations where 
they can't choose:

It occurs often that a fix or support for some hardware is only in the
latest -pre or in the -ac tree.


You say "things get broken from time to time as a normal part of 
development". Ideally this should never happen in a stable series. We 
don't live in an ideal world, but we should try to be as close as 
possible to this goal.


> Compiling your own kernel from the stock kernel trees is still
> something that should be considered for experienced users only.
> 
> Besides, what's worse?  Possible data corruption or a bad impression?

Possible data corruption is worse, but completely disabling this driver 
is even better.

> > > >  * 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.
> 
> The name BROKEN_ON_SMP implies that if you don't set it, what's left
> is known to work on SMP.  In a lot of cases, it'll actually be
> untested on SMP.

Say which other drivers are completely broken on SMP without a fix 
available or in the near future and it's easy to add a BROKEN_ON_SMP.

As long as noone reports such a bug I assume a driver works.

> 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