From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757836Ab3JNW3S (ORCPT ); Mon, 14 Oct 2013 18:29:18 -0400 Received: from merlin.infradead.org ([205.233.59.134]:42393 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756707Ab3JNW3Q (ORCPT ); Mon, 14 Oct 2013 18:29:16 -0400 Message-ID: <525C6F9F.3030801@infradead.org> Date: Mon, 14 Oct 2013 15:26:39 -0700 From: Randy Dunlap User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.0 MIME-Version: 1.0 To: Greg Kroah-Hartman CC: kbuild test robot , kbuild-all@01.org, "devel@driverdev.osuosl.org" , LKML , Daniel Drake , Jens Frederich , Jon Nettleton Subject: Re: [PATCH v2] staging/olpc_dcon: fix kconfig to fix build errors References: <524a3b76.oejnG6gfqBFdWbl/%fengguang.wu@intel.com> <525AF4B8.8030005@infradead.org> <20131014163332.GA10225@kroah.com> In-Reply-To: <20131014163332.GA10225@kroah.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 10/14/13 09:33, Greg Kroah-Hartman wrote: > On Sun, Oct 13, 2013 at 12:30:00PM -0700, Randy Dunlap wrote: >> From: Randy Dunlap >> >> Fix build errors when GPIO_CS5535=m and FB_OLPC_DCON=y >> by preventing that kconfig combination. >> >> These build errors are caused by having a kconfig bool symbol >> (FB_OLPC_DCON_1) that depend on a tristate symbol (GPIO_CS5535), >> but when the tristate symbol is =m, the bool symbol is =y. >> >> drivers/built-in.o: In function `dcon_read_status_xo_1': >> olpc_dcon_xo_1.c:(.text+0x359531): undefined reference to `cs5535_gpio_set' >> drivers/built-in.o: In function `dcon_wiggle_xo_1': >> olpc_dcon_xo_1.c:(.text+0x35959f): undefined reference to `cs5535_gpio_set' >> olpc_dcon_xo_1.c:(.text+0x359610): undefined reference to `cs5535_gpio_clear' >> drivers/built-in.o:olpc_dcon_xo_1.c:(.text+0x3596a1): more undefined references to `cs5535_gpio_clear' follow >> drivers/built-in.o: In function `dcon_wiggle_xo_1': >> olpc_dcon_xo_1.c:(.text+0x359708): undefined reference to `cs5535_gpio_set' >> drivers/built-in.o: In function `dcon_init_xo_1': >> olpc_dcon_xo_1.c:(.text+0x35989b): undefined reference to `cs5535_gpio_clear' >> olpc_dcon_xo_1.c:(.text+0x3598b5): undefined reference to `cs5535_gpio_isset' >> olpc_dcon_xo_1.c:(.text+0x359963): undefined reference to `cs5535_gpio_setup_event' >> olpc_dcon_xo_1.c:(.text+0x359980): undefined reference to `cs5535_gpio_set_irq' >> olpc_dcon_xo_1.c:(.text+0x359a36): undefined reference to `cs5535_gpio_set' >> >> However, adding GPIO_CS5535 to the Kconfig dependencies also creates >> a kconfig recursive dependency error on powerpc: >> drivers/i2c/Kconfig:5:error: recursive dependency detected! >> drivers/i2c/Kconfig:5: symbol I2C is selected by FB_OLPC_DCON >> drivers/staging/olpc_dcon/Kconfig:1: symbol FB_OLPC_DCON depends on GPIO_CS5535 >> drivers/gpio/Kconfig:577: symbol GPIO_CS5535 depends on GPIOLIB >> drivers/gpio/Kconfig:38: symbol GPIOLIB is selected by ARCH_REQUIRE_GPIOLIB >> drivers/gpio/Kconfig:23: symbol ARCH_REQUIRE_GPIOLIB is selected by MCU_MPC8349EMITX >> arch/powerpc/platforms/Kconfig:351: symbol MCU_MPC8349EMITX depends on I2C >> >> This is due to FB_OLPC_DCON selecting I2C instead of depending on it, >> so change the select to a dependency. >> >> Signed-off-by: Randy Dunlap >> Cc: Daniel Drake >> Cc: Jens Frederich >> Cc: Jon Nettleton >> --- >> drivers/staging/olpc_dcon/Kconfig | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> --- lnx-312-rc4.orig/drivers/staging/olpc_dcon/Kconfig >> +++ lnx-312-rc4/drivers/staging/olpc_dcon/Kconfig >> @@ -1,7 +1,8 @@ >> config FB_OLPC_DCON >> tristate "One Laptop Per Child Display CONtroller support" >> depends on OLPC && FB >> - select I2C >> + depends on I2C >> + depends on (GPIO_CS5535 || GPIO_CS5535=n) > > > This fails to apply to my staging-next branch of staging.git due to a > previous patch from you to try to fix this same issue (or maybe a > different one, I don't know...) You had asked if the previous patch should be reverted and I replied (very belatedly) Yes. However, I'll be glad to make a new patch against linux-next instead of mainline. -- ~Randy