Re: [PATCH] Kconfig cleanup (PARPORT_PC dependencies)

[Ingo Molnar <mingo@kernel.org>]

* H. Peter Anvin <hpa@zytor.com> wrote:

> On 09/12/2013 11:53 AM, Mark Salter wrote:
> >  
> > diff --git a/drivers/parport/Kconfig b/drivers/parport/Kconfig
> > index 70694ce..fa005a9 100644
> > --- a/drivers/parport/Kconfig
> > +++ b/drivers/parport/Kconfig
> > @@ -35,9 +35,7 @@ if PARPORT
> >  
> >  config PARPORT_PC
> >  	tristate "PC-style hardware"
> > -	depends on (!SPARC64 || PCI) && !SPARC32 && !M32R && !FRV && !S390 && \
> > -		(!M68K || ISA) && !MN10300 && !AVR32 && !BLACKFIN && \
> > -		!XTENSA && !CRIS && !H8300
> > +	depends on ARCH_MAY_HAVE_PC_PARPORT
> >  
> >  	---help---
> >  	  You should say Y here if you have a PC-style parallel port. All
> > 
> 
> I'm somewhat astonished that this driver is that problematic.  It is 
> really just a very basic PIO interface, at least in the absence of 
> EPP/ECP.
> 
> It is, of course, also largely a dead interface at this point so it 
> probably doesn't matter.

So, my collection of 'randconfig kernel crashers' has this exclusion and 
comment:

 --- linux.orig/drivers/block/Kconfig
 +++ linux/drivers/block/Kconfig
 @@ -77,6 +77,11 @@ config GDROM
  config PARIDE
         tristate "Parallel port IDE device support"
         depends on PARPORT_PC
 +
 +       # the probe can hang during bootup on non-PARIDE boxes
 +       depends on BROKEN_BOOT_ALLOWED
 +       select BROKEN_BOOT if PARIDE = y
 +

So at least some uses of PARPORT_PC are unsafe. (might easily be a bug in 
PARIDE)

Thanks,

	Ingo