On Sun, Sep 30, 2012 at 10:53 AM, Jean Delvare [off-list ref] wrote:

On Sat, 29 Sep 2012 18:39:13 +0200, Geert Uytterhoeven wrote:

quoted

On Thu, Sep 27, 2012 at 1:51 PM, Heiko Carstens

quoted

From 24c68e72b6198b8412a8605fac7d9c90796502ac Mon Sep 17 00:00:00 2001
From: Heiko Carstens <redacted>
Date: Thu, 27 Sep 2012 13:37:14 +0200
Subject: [PATCH] i2c: Make I2C unavailable on s390 again

"i2c: Make I2C available on UML" removed the IOMEM dependency from I2C which
resulted in I2C being avaible on s390 and causing allmodconfig and allyesconfig
builds to fail:

drivers/misc/altera-stapl/altera-lpt.c: In function 'byteblaster_write':
drivers/misc/altera-stapl/altera-lpt.c:34:2: error: implicit declaration of function 'outb' [-Werror=implicit-function-declaration]

Adding a HAS_PORT dependency to that driver reveals that there are lots
of other build errors waiting for discovery because of this change since
s390 does not support generic hardirqs:

ERROR: "free_irq" [drivers/power/smb347-charger.ko] undefined!
ERROR: "free_irq" [drivers/power/sbs-battery.ko] undefined!
ERROR: "enable_irq" [drivers/power/max17042_battery.ko] undefined!
ERROR: "request_irq" [drivers/power/max17042_battery.ko] undefined!
ERROR: "free_irq" [drivers/power/max17042_battery.ko] undefined!
ERROR: "disable_irq" [drivers/power/max17042_battery.ko] undefined!
ERROR: "request_irq" [drivers/power/lp8727_charger.ko] undefined!
ERROR: "free_irq" [drivers/power/lp8727_charger.ko] undefined!
ERROR: "request_irq" [drivers/nfc/pn544_hci.ko] undefined!
ERROR: "free_irq" [drivers/nfc/pn544_hci.ko] undefined!
ERROR: "request_irq" [drivers/nfc/pn544.ko] undefined!
ERROR: "free_irq" [drivers/nfc/pn544.ko] undefined!
ERROR: "request_irq" [drivers/misc/fsa9480.ko] undefined!
ERROR: "free_irq" [drivers/misc/fsa9480.ko] undefined!
ERROR: "request_irq" [drivers/misc/bh1770glc.ko] undefined!
ERROR: "free_irq" [drivers/misc/bh1770glc.ko] undefined!
ERROR: "request_irq" [drivers/misc/apds990x.ko] undefined!
ERROR: "free_irq" [drivers/misc/apds990x.ko] undefined!

Since UML seems to need generic I2C support, but lot's of now available
drivers fail to compile on s390, just add a !S390 dependency.

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Peter Huewe <peterhuewe@gmx.de>
Cc: Jean Delvare <redacted>
Signed-off-by: Heiko Carstens <redacted>

Yeah, s390 is so special that it's the only architecture that can be used
in depends on !, so

Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>

quoted

---
 drivers/i2c/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/i2c/Kconfig b/drivers/i2c/Kconfig
index 17b35c1..2f8c76b 100644
--- a/drivers/i2c/Kconfig
+++ b/drivers/i2c/Kconfig

@@ -4,6 +4,7 @@

 menuconfig I2C
        tristate "I2C support"
+       depends on !S390
        select RT_MUTEXES
        ---help---
          I2C (pronounce: I-squared-C) is a slow serial bus protocol used in

I thought we were trying to move away from explicit arch exclusions,
but if S390 is special then fine with me. I've folded Heiko's fix into

"depends on GENERIC_HARDIRQS" is the alternative...

Hmm, I just noticed we already have lots of dependencies on GENERIC_HARDIRQS,
especially w.r.t I2C:

drivers/gpio/Kconfig:	depends on GPIO_MAX732X=y && GENERIC_HARDIRQS
drivers/iio/Kconfig:	depends on GENERIC_HARDIRQS
drivers/input/touchscreen/Kconfig:        depends on SPI_MASTER &&
GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on GENERIC_HARDIRQS && GPIOLIB && ARM
drivers/mfd/Kconfig:	depends on GENERIC_HARDIRQS && GPIOLIB && ARM
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on (I2C=y || SPI_MASTER=y) && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on SPI_MASTER && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on GENERIC_HARDIRQS && SPI_MASTER
drivers/mfd/Kconfig:	depends on GENERIC_HARDIRQS && ABX500_CORE &&
MFD_DB8500_PRCMU
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/mfd/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
drivers/net/wireless/p54/Kconfig:	depends on P54_COMMON && SPI_MASTER
&& GENERIC_HARDIRQS
drivers/net/wireless/ti/wl1251/Kconfig:	depends on MAC80211 &&
EXPERIMENTAL && GENERIC_HARDIRQS
drivers/net/wireless/ti/wlcore/Kconfig:	depends on WL_TI &&
GENERIC_HARDIRQS && MAC80211
drivers/power/Kconfig:	depends on GENERIC_HARDIRQS
drivers/pps/clients/Kconfig:	depends on PPS && GENERIC_HARDIRQS
drivers/w1/masters/Kconfig:	depends on W1 && GENERIC_HARDIRQS
kernel/irq/Kconfig:config GENERIC_HARDIRQS
lib/Kconfig.debug:	depends on DEBUG_KERNEL && GENERIC_HARDIRQS
net/Kconfig:	depends on RPS && GENERIC_HARDIRQS
sound/soc/samsung/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
sound/soc/samsung/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS
sound/soc/samsung/Kconfig:	depends on I2C=y && GENERIC_HARDIRQS

Peter's patch, as it is not committed yet. Result is at:

http://khali.linux-fr.org/devel/linux-3/jdelvare-i2c/i2c-make-I2C-available-on-UML.patch

That's what will be in the next linux-next.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds