Re: [PATCH v3 35/38] video: handle HAS_IOPORT dependencies

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Thu, Mar 23, 2023 at 03:17:38PM +0100, Niklas Schnelle wrote:
> On Wed, 2023-03-15 at 12:19 +0200, Ville Syrjälä wrote:
> > On Wed, Mar 15, 2023 at 09:16:50AM +0100, Geert Uytterhoeven wrote:
> > > Hi Niklas,
> > > 
> > > On Tue, Mar 14, 2023 at 1:13 PM Niklas Schnelle <schnelle@linux.ibm.com> wrote:
> > > > In a future patch HAS_IOPORT=n will result in inb()/outb() and friends
> > > > not being declared. We thus need to add HAS_IOPORT as dependency for
> > > > those drivers using them and guard inline code in headers.
> > > > 
> > > > Co-developed-by: Arnd Bergmann <arnd@kernel.org>
> > > > Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
> > > 
> > > Thanks for your patch!
> > > 
> > > > --- a/drivers/video/fbdev/Kconfig
> > > > +++ b/drivers/video/fbdev/Kconfig
> > > 
> > > > @@ -1284,7 +1285,7 @@ config FB_ATY128_BACKLIGHT
> > > > 
> > > >  config FB_ATY
> > > >         tristate "ATI Mach64 display support" if PCI || ATARI
> > > > -       depends on FB && !SPARC32
> > > > +       depends on FB && HAS_IOPORT && !SPARC32
> > > 
> > > On Atari, this works without ATARI_ROM_ISA, hence it must not depend
> > > on HAS_IOPORT.
> > > The only call to inb() is inside a section protected by #ifdef
> > > CONFIG_PCI. So:
> > 
> > That piece of code is a nop anyway. We immediately overwrite
> > clk_wr_offset with a hardcoded selection after the register reads.
> > So if you nuke that nop code then no IOPORT dependency required
> > at all.
> > 
> 
> I agree this "looks" like a nop but are we sure the inb() doesn't have
> side effects? 

Yes. It's just trying to check which PLL dividers/etc. are currently
used. In VGA mode it gets it from a the GENMO and in non-VGA mode from
CLOCK_CNTL. And then it says "screw that" and just uses index 3 instead.

Though I must say that mach64 GX seems to use that clk_wr_offset
very differently so I'm not sure the PCI+GX combo is even working
currently, assuming those even exist. I don't think I have anything
older than a PCI mach64 VT myself.

> (for reference drivers/video/fbdev/aty/aty/atyfb_base.c:
> atyfb_setup_generc() towards the end)
> 
> It does feel a bit out of scope for this series but if it's really a
> nop nuking it surely is the cleaner solution.
> 
> Thanks,
> Niklas

-- 
Ville Syrjälä
Intel