From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH] i2c: I2C_ELEKTOR should depend on HAS_IOPORT Date: Tue, 16 Aug 2011 21:42:06 +0200 Message-ID: <4991605.b7BYCfYE3S@wuerfel> References: <1312802446-19154-1-git-send-email-geert@linux-m68k.org> <20110816195033.014f302b@endymion.delvare> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7Bit Return-path: In-Reply-To: Sender: linux-i2c-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Geert Uytterhoeven Cc: Jean Delvare , Hans Berglund , Jonas Bonn , linux-i2c-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-i2c@vger.kernel.org On Tuesday 16 August 2011 21:14:42 Geert Uytterhoeven wrote: > >> config I2C_ELEKTOR > >> tristate "Elektor ISA card" > >> - depends on ISA && BROKEN_ON_SMP > >> + depends on ISA && HAS_IOPORT && BROKEN_ON_SMP > >> select I2C_ALGOPCF > >> help > >> This supports the PCF8584 ISA bus I2C adapter. Say Y if you own > > > > How can ISA be set if HAS_IOPORT is not? This shouldn't be allowed. > > It's always been like that. With CONFIG_ISA, we have inb() and friends. But not > ioport_map(), which was introduced more recently. Right. BTW, I'd really like to change that once I've picked up and fixed the use of inb() in 8250.c. I think we should change CONFIG_HAS_IOPORT to mean that inb(), ioport_resource and /proc/ioports are available, while a new CONFIG_HAS_IOPORT_MAP can then replace the current CONFIG_HAS_IOPORT symbol. Arnd