From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============5506537712341246698==" MIME-Version: 1.0 From: Matthew Wilcox To: kbuild-all@lists.01.org Subject: Re: ARC compiler not available Date: Mon, 11 Nov 2019 07:27:58 -0800 Message-ID: <20191111152758.GA28550@bombadil.infradead.org> In-Reply-To: <20191110060005.GA12775@intel.com> List-Id: --===============5506537712341246698== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On Sun, Nov 10, 2019 at 02:00:05PM +0800, Philip Li wrote: > On Fri, Nov 08, 2019 at 07:49:48AM -0800, Matthew Wilcox wrote: > > On Fri, Nov 08, 2019 at 12:50:44PM +0800, kbuild test robot wrote: > > > tree: git://git.infradead.org/users/willy/linux-dax.git xarray > > > head: 498b1f3c48d9d74d7c2902d9d97a7612a5af869c > > > commit: 5a74ac4c4a97bd8b7dba054304d598e2a882fea6 [2/6] idr: Fix idr_g= et_next_ul race with idr_remove > > > config: arc-defconfig (attached as .config) > > > compiler: arc-elf-gcc (GCC) 7.4.0 > > > reproduce: > > > wget https://raw.githubusercontent.com/intel/lkp-tests/master= /sbin/make.cross -O ~/bin/make.cross > > > chmod +x ~/bin/make.cross > > > git checkout 5a74ac4c4a97bd8b7dba054304d598e2a882fea6 > > > # save the attached .config to linux build tree > > > GCC_VERSION=3D7.4.0 make.cross ARCH=3Darc = > > = > > That doesn't work: > > = > > $ GCC_VERSION=3D7.4.0 make.cross ARCH=3Darc O=3D.build-arc > > cd: received redirection to `https://download.01.org/0day-ci/cross-pack= age/' > > lftpget -c https://download.01.org/0day-ci/cross-package/./gcc-7.4.0-no= libc/x86_64-gcc-7.4.0-nolibc_sparc-linux.tar.xz > > tar Jxf gcc-7.4.0-nolibc/x86_64-gcc-7.4.0-nolibc_sparc-linux.tar.xz -C = /home/willy/0day > > No cross compiler for arc > > = > > Why is it downloading sparc binaries instead of arc binaries? > sorry, this is a bug in make.cross for arc compiler. Can you re-get the m= ake.cross file > and try again? It should be able download the right compiler Yes, this now works for me. I can't quite decipher the error messages from the assembler, but it seems like this is a bug in gcc-7.4.0 producing invalid assembler input. It's fixed in gcc-8.2.0. We could file a gcc bug about this, but I don't know if it's worthwhile. Maybe just stop testing with gcc-7.4.0 on arc? Thanks! --===============5506537712341246698==--