From alain at vocatus.pub Tue Feb 20 21:44:43 2018 From: alain at vocatus.pub (Alain Toussaint) Date: Tue, 20 Feb 2018 16:44:43 -0500 Subject: [mkroot] x86_64 x32 mkroot port status report Message-ID: <1519163083.1600.2.camel@vocatus.pub> Hello, Just reporting that I successfully built and booted up in both qemu and real hardware an x32[1] ABI port of mkroot along with musl-cross-make. [1] == not to be confused with a 32 bit i686 userspace; the particularity are 4 bytes pointers like i686 but there is full access to the 15 integer registers and 16 floating points registers like x86_64. On real hardware, it is necessary to pass the command init=/usr/bin/busybox sh (or /bin/busybox sh) to the bootloader but otherwise, there are no other requirements. The dropbear package does not compile and finally, I was able to install gcc 6.3.0 from mcm in native x32 configuration and it run. Alain From txt.file at txtfile.eu Tue Feb 20 22:48:00 2018 From: txt.file at txtfile.eu (txt.file) Date: Tue, 20 Feb 2018 22:48:00 +0000 Subject: [mkroot] x86_64 x32 mkroot port status report In-Reply-To: <1519163083.1600.2.camel@vocatus.pub> References: <1519163083.1600.2.camel@vocatus.pub> Message-ID: Congrats. Great to have x32 available. I like this architecture. Alain Toussaint: > Hello, > > Just reporting that I successfully built and booted up in both qemu and > real hardware an x32[1] ABI port of mkroot along with musl-cross-make. > > [1] == not to be confused with a 32 bit i686 userspace; the > particularity are 4 bytes pointers like i686 but there is full access > to the 15 integer registers and 16 floating points registers like > x86_64. > > On real hardware, it is necessary to pass the command > init=/usr/bin/busybox sh (or /bin/busybox sh) to the bootloader but > otherwise, there are no other requirements. > > The dropbear package does not compile and finally, I was able to > install gcc 6.3.0 from mcm in native x32 configuration and it run. > > Alain > > _______________________________________________ > mkroot mailing list > mkroot at lists.landley.net > http://lists.landley.net/listinfo.cgi/mkroot-landley.net > -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 833 bytes Desc: OpenPGP digital signature URL: From rob at landley.net Wed Feb 21 15:41:58 2018 From: rob at landley.net (Rob Landley) Date: Wed, 21 Feb 2018 09:41:58 -0600 Subject: [mkroot] x86_64 x32 mkroot port status report In-Reply-To: <1519163083.1600.2.camel@vocatus.pub> References: <1519163083.1600.2.camel@vocatus.pub> Message-ID: On 02/20/2018 03:44 PM, Alain Toussaint wrote: > Hello, > > Just reporting that I successfully built and booted up in both qemu and > real hardware an x32[1] ABI port of mkroot along with musl-cross-make. Cool! Got a patch? > [1] == not to be confused with a 32 bit i686 userspace; the > particularity are 4 bytes pointers like i686 but there is full access > to the 15 integer registers and 16 floating points registers like > x86_64. > > On real hardware, it is necessary to pass the command > init=/usr/bin/busybox sh (or /bin/busybox sh) to the bootloader but > otherwise, there are no other requirements. I have some local changes to handle chroot and real hardware better, but haven't checked them in yet. Last stable kernel release broke arm (perl requirement in the build) and x86_64 (libelf requirement in the build), so I should cut a release with the one before that and fix that after it's out. (THIS release added flex and bison as build requirements for kconfig, I need a bigger flamethrower...) I was trying to put together a release last weekend but testing toybox-git I found that logger doesn't build because of a musl bug (prioritynames and facilitynames are in the headers, but then the link fails), and I implemented a fix for that (inline the name tables) but am still testing/debugging it (new search loop didn't work on the first try and I haven't had a chance to sit down and work on it again since)... > The dropbear package does not compile and finally, I was able to > install gcc 6.3.0 from mcm in native x32 configuration and it run. Cool! Got a patch? Rob From alain at vocatus.pub Wed Feb 21 16:53:08 2018 From: alain at vocatus.pub (Alain Toussaint) Date: Wed, 21 Feb 2018 11:53:08 -0500 Subject: [mkroot] x86_64 x32 mkroot port status report In-Reply-To: References: <1519163083.1600.2.camel@vocatus.pub> Message-ID: <1519231988.1610.1.camel@vocatus.pub> > Cool! Got a patch? In preparation but you can expect it for tomorrow at the latest. Alain From alain at vocatus.pub Sun Feb 25 23:50:38 2018 From: alain at vocatus.pub (Alain Toussaint) Date: Sun, 25 Feb 2018 18:50:38 -0500 Subject: [mkroot] x86_64 x32 mkroot port status report In-Reply-To: <1519231988.1610.1.camel@vocatus.pub> References: <1519163083.1600.2.camel@vocatus.pub> <1519231988.1610.1.camel@vocatus.pub> Message-ID: <1519602638.9940.16.camel@vocatus.pub> Le mercredi 21 f?vrier 2018 ? 11:53 -0500, Alain Toussaint a ?crit?: > > Cool! Got a patch? > > > In preparation but you can expect it for tomorrow at the latest. > > Alain > I'm testing the first patch before sending but I must that mcm-buildall.sh is seriously taxing all my cores (4 + 4 hyperthreading). Alain From alain at vocatus.pub Mon Feb 26 14:05:09 2018 From: alain at vocatus.pub (Alain Toussaint) Date: Mon, 26 Feb 2018 09:05:09 -0500 Subject: [mkroot] x86_64 x32 mkroot port status report In-Reply-To: References: <1519163083.1600.2.camel@vocatus.pub> Message-ID: <1519653909.28422.3.camel@vocatus.pub> > Cool! Got a patch? > Preliminary patch to mcm-buildall.sh included. I have to do a test round with both x86_64 and x86_64 x32 enabled next a subsequent patch. Alain -------------- next part -------------- A non-text attachment was scrubbed... Name: mcm-buildall.x32.diff Type: text/x-patch Size: 598 bytes Desc: not available URL: From alain at vocatus.pub Tue Feb 27 20:02:09 2018 From: alain at vocatus.pub (Alain Toussaint) Date: Tue, 27 Feb 2018 15:02:09 -0500 Subject: [mkroot] x32 mcm-buildall.sh final patch Message-ID: <1519761729.26577.19.camel@vocatus.pub> Hi, Here's the final patch. Tested, work as intended for mcm-buildall.sh and build both variants of x86_64 ABI (64bits and x32). Minor issue with cross.sh is that it doesn't distinguishes between both ABI: testbuild at johnson:~/git/mkroot$ ./cross.sh? aarch64 armv5l armv7l armv7m armv7r i486 i686 microblaze mips mips64 mipsel powerpc powerpc64 powerpc64le s390x sh2eb sh4 x86_64 testbuild at johnson:~/git/mkroot$ ls -l mcm/ total 148 drwxr-xr-x 8 testbuild lfs 4096 Feb 26 14:07 aarch64-linux-musleabi-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 14:28 aarch64-linux-musleabi-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 11:01 armv5l-linux-musleabihf-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 11:26 armv5l-linux-musleabihf-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 11:49 armv7l-linux-musleabihf-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 12:13 armv7l-linux-musleabihf-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 12:36 armv7m-linux-musleabi-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 12:59 armv7m-linux-musleabi-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 13:22 armv7r-linux-musleabihf-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 13:46 armv7r-linux-musleabihf-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 14:50 i486-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 15:12 i486-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 09:34 i686-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 09:56 i686-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 20:12 microblaze-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 23:42 microblaze-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 18:52 mips-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 19:14 mips-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 27 00:03 mips64-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 27 00:28 mips64-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 18:07 mipsel-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 18:30 mipsel-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 19:34 powerpc-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 19:54 powerpc-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 27 00:48 powerpc64-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 27 01:08 powerpc64-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 27 01:29 powerpc64le-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 27 01:49 powerpc64le-linux-musl-native drwxr-xr-x 3 testbuild lfs 4096 Feb 26 15:55 s390x-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 15:32 sh2eb-linux-muslfdpic-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 15:54 sh2eb-linux-muslfdpic-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 10:16 sh4-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 10:37 sh4-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 16:36 x86_64-linux-musl-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 16:59 x86_64-linux-musl-native drwxr-xr-x 8 testbuild lfs 4096 Feb 26 17:22 x86_64-linux-muslx32-cross drwxr-xr-x 8 testbuild lfs 4096 Feb 26 17:46 x86_64-linux-muslx32-native Al -------------- next part -------------- A non-text attachment was scrubbed... Name: mcm-buildall.x32.final.diff Type: text/x-patch Size: 607 bytes Desc: not available URL: