From: Yann E. MORIN Date: Mon, 2 Feb 2015 18:32:51 +0100 Subject: [Buildroot] [PATCH 3/4] toolchain: allow for stupid toolchains In-Reply-To: <54CFA351.2050408@openwide.fr> References: <1421951608-31233-1-git-send-email-ezequiel@vanguardiasur.com.ar> <1421951608-31233-4-git-send-email-ezequiel@vanguardiasur.com.ar> <54CFA351.2050408@openwide.fr> Message-ID: <20150202173251.GB17739@free.fr> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Ezequiel, Romain, All, On 2015-02-02 17:18 +0100, Romain Naour spake thusly: > Le 22/01/2015 19:33, Ezequiel Garcia a ?crit : > > From: Guido Mart?nez > > > > check_arm_abi builds a test C file to check that the toolchain is > > working correctly, with the output redirected to /dev/null. > > > > However, some toolchains (OSELAS 2013.12.2, for instance) foolishly > > append ".gdb" to the output filename for an intermediate file, causing > > an attempt to write to /dev/null.gdb, which obviously fails. > > > > Fix this by adding an option to skip this check, so it can be selected > > in the OSELAS toolchain. > > > > Signed-off-by: Guido Mart?nez > > --- > > toolchain/helpers.mk | 6 +++++- > > toolchain/toolchain-external/Config.in | 4 ++++ > > 2 files changed, 9 insertions(+), 1 deletion(-) > > > > diff --git a/toolchain/helpers.mk b/toolchain/helpers.mk > > index 3121da4..d008a4f 100644 > > --- a/toolchain/helpers.mk > > +++ b/toolchain/helpers.mk > > @@ -297,13 +297,17 @@ check_arm_abi = \ > > if ! echo $${EXT_TOOLCHAIN_TARGET} | grep -qE 'eabi(hf)?$$' ; then \ > > echo "External toolchain uses the unsuported OABI" ; \ > > exit 1 ; \ > > - fi ; \ > > + fi > > + > > +ifneq ($(BR2_ARM_TOOLCHAIN_SKIP_CHECK),y) > > +check_arm_abi += ; \ > > if ! echo 'int main(void) {}' | $${__CROSS_CC} -x c -o /dev/null - ; then \ > > abistr_$(BR2_ARM_EABI)='EABI'; \ > > abistr_$(BR2_ARM_EABIHF)='EABIhf'; \ > > echo "Incorrect ABI setting: $${abistr_y} selected, but toolchain is incompatible"; \ > > exit 1 ; \ > > fi > > +endif > > > > I looked at this issue and tried to avoid the use of /dev/null.gdb ... > In the Cortex-M toolchain, it seems to have a wrapper around the linker like > this one [1] which append the .gdb suffix. > > However, there is no error if "-Wl,-r" (Generate relocatable output) is added in > the gcc's command line but it's not really nice solution because this option > will be there for other toolchains. > > Yann proposed to use $BUILD_DIR/br-foo instead of /dev/null and add a comment > why we do that. Yeah, something like the following: diff --git a/toolchain/helpers.mk b/toolchain/helpers.mk index 3121da4..44696c0 100644 --- a/toolchain/helpers.mk +++ b/toolchain/helpers.mk @@ -298,12 +298,14 @@ check_arm_abi = \ echo "External toolchain uses the unsuported OABI" ; \ exit 1 ; \ fi ; \ - if ! echo 'int main(void) {}' | $${__CROSS_CC} -x c -o /dev/null - ; then \ + if ! echo 'int main(void) {}' | $${__CROSS_CC} -x c -o $(BUILD_DIR)/.br-temp-out - ; then \ + rm -f $(BUILD_DIR)/.br-temp-out*; \ abistr_$(BR2_ARM_EABI)='EABI'; \ abistr_$(BR2_ARM_EABIHF)='EABIhf'; \ echo "Incorrect ABI setting: $${abistr_y} selected, but toolchain is incompatible"; \ exit 1 ; \ - fi + fi; \ + rm -f $(BUILD_DIR)/.br-temp-out* # # Check that the external toolchain supports C++ (does not apply, copy-paste with tab-brokenness) Regards, Yann E. MORIN. -- .-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'