From rob at landley.net Mon Sep 2 09:31:23 2013 From: rob at landley.net (Rob Landley) Date: Mon, 02 Sep 2013 04:31:23 -0500 Subject: [Aboriginal] Kernel patching failure with 1.2.4.1 In-Reply-To: <5220348B.1040202@gentoo.org> (from patrick@gentoo.org on Fri Aug 30 00:58:35 2013) References: <5210463E.8070407@gentoo.org> <5220348B.1040202@gentoo.org> Message-ID: <1378114283.2737.154@driftwood> On 08/30/2013 12:58:35 AM, Patrick Lauer wrote: > On 08/18/2013 11:57 AM, Patrick Lauer wrote: > > Trying to build 1.2.4.1 fails when fetching/patching kernel: > > > patching file init/Kconfig > > Hunk #1 succeeded at 1797 (offset 131 lines). > > Applying > > > /home/me/code/aboriginal-1.2.4.1/sources/patches/linux-noperl-timeconst.patch > > The next patch would create the file kernel/Makefile, > > which already exists! Assume -R? [n] > > Apply anyway? [n] > > Skipping patch. > > 1 out of 1 hunk ignored > > patching file kernel/timeconst.bc > > patching file kernel/Makefile > > Hunk #1 succeeded at 126 (offset 1 line). > > > > Exiting due to errors (host download linux) > > > > Exiting due to errors ( ) > > This is due to a change in behaviour with patch - version <=2.6.1 work > fine, >=2.7 fail. And _this_ is why I build host-tools.sh to isolate the build from crappy random gnu regressions. (Alas, there's a chicken and egg problem in protecting download.sh or host-tools.sh from this stuff.) Where do I get a test system that displays this problem? (I don't have anything with patch 2.7 installed. Should I build a linux from scratch, or is there something I can install into a qemu image by just fetching an iso...?) Rob 1378114283.0 From patrick at gentoo.org Mon Sep 2 11:23:08 2013 From: patrick at gentoo.org (Patrick Lauer) Date: Mon, 02 Sep 2013 19:23:08 +0800 Subject: [Aboriginal] Kernel patching failure with 1.2.4.1 In-Reply-To: <1378114283.2737.154@driftwood> References: <5210463E.8070407@gentoo.org> <5220348B.1040202@gentoo.org> <1378114283.2737.154@driftwood> Message-ID: <5224751C.1030907@gentoo.org> On 09/02/2013 05:31 PM, Rob Landley wrote: > On 08/30/2013 12:58:35 AM, Patrick Lauer wrote: >> On 08/18/2013 11:57 AM, Patrick Lauer wrote: >> > Trying to build 1.2.4.1 fails when fetching/patching kernel: >> >> > patching file init/Kconfig >> > Hunk #1 succeeded at 1797 (offset 131 lines). >> > Applying >> > >> /home/me/code/aboriginal-1.2.4.1/sources/patches/linux-noperl-timeconst.patch >> >> > The next patch would create the file kernel/Makefile, >> > which already exists! Assume -R? [n] >> > Apply anyway? [n] >> > Skipping patch. >> > 1 out of 1 hunk ignored >> > patching file kernel/timeconst.bc >> > patching file kernel/Makefile >> > Hunk #1 succeeded at 126 (offset 1 line). >> > >> > Exiting due to errors (host download linux) >> > >> > Exiting due to errors ( ) >> >> This is due to a change in behaviour with patch - version <=2.6.1 work >> fine, >=2.7 fail. > > And _this_ is why I build host-tools.sh to isolate the build from crappy > random gnu regressions. (Alas, there's a chicken and egg problem in > protecting download.sh or host-tools.sh from this stuff.) Then you'd end up at some point carrying your own statically-built everything, only relying on some archiver like cpio/tar/zip for the initial unpacking > > Where do I get a test system that displays this problem? (I don't have > anything with patch 2.7 installed. Should I build a linux from scratch, > or is there something I can install into a qemu image by just fetching > an iso...?) I upgraded patch on my server, so you can reproduce it there. > > Rob 1378120988.0 From rob at landley.net Sat Sep 7 01:31:31 2013 From: rob at landley.net (Rob Landley) Date: Fri, 06 Sep 2013 20:31:31 -0500 Subject: [Aboriginal] Kernel patching failure with 1.2.4.1 In-Reply-To: <5224751C.1030907@gentoo.org> (from patrick@gentoo.org on Mon Sep 2 06:23:08 2013) Message-ID: <1378517491.1985.15@driftwood> On 09/02/2013 06:23:08 AM, Patrick Lauer wrote: > On 09/02/2013 05:31 PM, Rob Landley wrote: > > On 08/30/2013 12:58:35 AM, Patrick Lauer wrote: > >> On 08/18/2013 11:57 AM, Patrick Lauer wrote: > >> > Trying to build 1.2.4.1 fails when fetching/patching kernel: > >> > >> > patching file init/Kconfig > >> > Hunk #1 succeeded at 1797 (offset 131 lines). > >> > Applying > >> > > >> > /home/me/code/aboriginal-1.2.4.1/sources/patches/linux-noperl-timeconst.patch > >> > >> > The next patch would create the file kernel/Makefile, > >> > which already exists! Assume -R? [n] > >> > Apply anyway? [n] > >> > Skipping patch. > >> > 1 out of 1 hunk ignored > >> > patching file kernel/timeconst.bc > >> > patching file kernel/Makefile > >> > Hunk #1 succeeded at 126 (offset 1 line). > >> > > >> > Exiting due to errors (host download linux) > >> > > >> > Exiting due to errors ( ) > >> > >> This is due to a change in behaviour with patch - version <=2.6.1 > work > >> fine, >=2.7 fail. > > > > And _this_ is why I build host-tools.sh to isolate the build from > crappy > > random gnu regressions. (Alas, there's a chicken and egg problem in > > protecting download.sh or host-tools.sh from this stuff.) > > Then you'd end up at some point carrying your own statically-built > everything, only relying on some archiver like cpio/tar/zip for the > initial unpacking Except I'm not assuming the build host is x86. :) No, we've gotta start somewhere, and the current host-tools.sh is as generic as I can get. > > Where do I get a test system that displays this problem? (I don't > have > > anything with patch 2.7 installed. Should I build a linux from > scratch, > > or is there something I can install into a qemu image by just > fetching > > an iso...?) > > I upgraded patch on my server, so you can reproduce it there. I just did a fresh EXTRACT_ALL=1 there and it worked. Admittedly it worked against 3.11 instead of 3.10, but the patch in question created a new file... Does this still reproduce for you? Rob 1378517491.0 From patrick at gentoo.org Sat Sep 7 03:46:19 2013 From: patrick at gentoo.org (Patrick Lauer) Date: Sat, 07 Sep 2013 11:46:19 +0800 Subject: [Aboriginal] Kernel patching failure with 1.2.4.1 In-Reply-To: <1378517491.1985.15@driftwood> References: <1378517491.1985.15@driftwood> Message-ID: <522AA18B.2070008@gentoo.org> On 09/07/2013 09:31 AM, Rob Landley wrote: > On 09/02/2013 06:23:08 AM, Patrick Lauer wrote: >> On 09/02/2013 05:31 PM, Rob Landley wrote: >> > On 08/30/2013 12:58:35 AM, Patrick Lauer wrote: >> >> On 08/18/2013 11:57 AM, Patrick Lauer wrote: >> >> > Trying to build 1.2.4.1 fails when fetching/patching kernel: >> >> >> >> > patching file init/Kconfig >> >> > Hunk #1 succeeded at 1797 (offset 131 lines). >> >> > Applying >> >> > >> >> >> /home/me/code/aboriginal-1.2.4.1/sources/patches/linux-noperl-timeconst.patch >> >> >> >> >> > The next patch would create the file kernel/Makefile, >> >> > which already exists! Assume -R? [n] >> >> > Apply anyway? [n] >> >> > Skipping patch. >> >> > 1 out of 1 hunk ignored >> >> > patching file kernel/timeconst.bc >> >> > patching file kernel/Makefile >> >> > Hunk #1 succeeded at 126 (offset 1 line). >> >> > >> >> > Exiting due to errors (host download linux) >> >> > >> >> > Exiting due to errors ( ) >> >> >> >> This is due to a change in behaviour with patch - version <=2.6.1 work >> >> fine, >=2.7 fail. >> > >> > Where do I get a test system that displays this problem? (I don't have >> > anything with patch 2.7 installed. Should I build a linux from scratch, >> > or is there something I can install into a qemu image by just fetching >> > an iso...?) >> >> I upgraded patch on my server, so you can reproduce it there. > > I just did a fresh EXTRACT_ALL=1 there and it worked. > > Admittedly it worked against 3.11 instead of 3.10, but the patch in > question created a new file... > > Does this still reproduce for you? Yes, on my desktop at home I can reliably make the patching fail with patch-2.7.1 and make it work with patch-2.6.1. I've tested with a previous aboriginal release (1.2.3 I think) which extracts and builds fine, it is only the 1.2.4.1 release that shows the problem. 1378525579.0 From alain.toussaint at securivm.ca Mon Sep 9 02:24:28 2013 From: alain.toussaint at securivm.ca (Alain Toussaint) Date: Sun, 8 Sep 2013 22:24:28 -0400 Subject: [Aboriginal] building Chromium OS images on aboriginal Message-ID: Hello :) I have my previous computer, an AMD zacate E-350 box with 8GB of ram (plenty overkill for aboriginal) and a 500GB hard drive which will serve as netatalk box to 3 mac laptop and will also be used for some development tasks which I hope to use GNUstep along Xcode on my mac and finally, building the occasional Chromium OS image for my brother's netbook. As far as Chromium goes, it build in a chroot and is self contained with actually, a Gentoo development environment and it only need sudo and chroot from the host system. At the moment, the computer is getting its disk wiped out with dd from a gentoo minimal disk and I thought I'd use the root filesystem offered on the landley.net website and build a kernel and lilo (along with its dependancy, grub is too finicky and I'm old school) to have a basic system. Given Chromium OS dependancies, will I be able to install aboriginal on my system? I foresee only installing ncurses to have menu when building the linux kernel. Thanks Alain 1378693468.0 From alain.toussaint at securivm.ca Mon Sep 9 03:45:34 2013 From: alain.toussaint at securivm.ca (Alain Toussaint) Date: Sun, 8 Sep 2013 23:45:34 -0400 Subject: [Aboriginal] ash prompt disappear when using tab completion Message-ID: <8DA9C110-49E7-458C-B7A6-300F2B6FF28E@securivm.ca> Hello, I've been installing my system over ssh from my mac to the aboriginal system and the prompt as well as characters disappear behind the cursor when using tab completion or erasing characters. I did not test if it applies to bash 2.04 on aboriginal but on the bash from the install cd, it does not happen. Alain 1378698334.0 From rob at landley.net Mon Sep 9 17:52:22 2013 From: rob at landley.net (Rob Landley) Date: Mon, 09 Sep 2013 12:52:22 -0500 Subject: [Aboriginal] ash prompt disappear when using tab completion In-Reply-To: <8DA9C110-49E7-458C-B7A6-300F2B6FF28E@securivm.ca> (from alain.toussaint@securivm.ca on Sun Sep 8 22:45:34 2013) References: <8DA9C110-49E7-458C-B7A6-300F2B6FF28E@securivm.ca> Message-ID: <1378749142.8385.16@driftwood> On 09/08/2013 10:45:34 PM, Alain Toussaint wrote: > Hello, > > I've been installing my system over ssh from my mac to the aboriginal > system > and the prompt as well as characters disappear behind the cursor when > using > tab completion or erasing characters. You mean characters to the right of the cursor aren't redrawn correctly? Sounds like a busybox bug. > I did not test if it applies to bash > 2.04 on aboriginal but on the bash from the install cd, it does not > happen. The bash 2.04 isn't built with curses support, and thus may not have command line history at all. (It's really intended for scripting, ash is the interactive one.) The linux from scratch images have a much newer version of bash built with ncurses support, but I dunno if I uploaded prebuilt ones of those last release. I mean to do them for the pending release, though. (Not quite sure when that'll be, I have to get on a plane thursday to spend the weekend at ohio linuxfest...) Rob 1378749142.0 From rob at landley.net Mon Sep 9 10:18:55 2013 From: rob at landley.net (Rob Landley) Date: Mon, 09 Sep 2013 05:18:55 -0500 Subject: [Aboriginal] building Chromium OS images on aboriginal In-Reply-To: (from alain.toussaint@securivm.ca on Sun Sep 8 21:24:28 2013) Message-ID: <1378721935.8385.8@driftwood> On 09/08/2013 09:24:28 PM, Alain Toussaint wrote: > Hello :) > > I have my previous computer, an AMD zacate E-350 box with 8GB of ram > (plenty overkill for aboriginal) and a 500GB hard drive which will > serve as netatalk box to 3 mac laptop and will also be used for some > development tasks which I hope to use GNUstep along Xcode on my mac > and finally, building the occasional Chromium OS image for my > brother's > netbook. As far as Chromium goes, it build in a chroot and is self > contained with actually, a Gentoo development environment and it only > need sudo and chroot from the host system. > > At the moment, the computer is getting its disk wiped out with dd from > a gentoo minimal disk and I thought I'd use the root filesystem > offered > on the landley.net website and build a kernel and lilo (along with its > dependancy, grub is too finicky and I'm old school) to have a basic > system. Cool. > Given Chromium OS dependancies, will I be able to install aboriginal > on > my system? I foresee only installing ncurses to have menu when > building > the linux kernel. Install aboriginal, yes. (It's just extracting a tarball.) Then building chromium under the result: I don't know? Can you build chromium under Linux From Scratch? Because you can build linux from scratch under aboriginal, I even have an automated build control image to do it: http://landley.net/aboriginal/control-images So you can bootstrap your way up to an arbitrarily more complex environment starting from the aboriginal linux root filesystem. Building distros is hard because they all assume they're building under themselves (even gentoo), but if you can work out how to build chromium under LFS and document it I'd love to make an automated control image to do that, and then regression test it with each new release. :) The above LFS is 6.8 and I'm updating to 7.4 as we speak. (This is delaying the new release a bit, current source control includes the 3.11 kernel and builds fine but I'm trying to get a new toybox release in there and LFS 7.4 too.) So those package versions might be a bit stale, but should still work. Rob 1378721935.0 From alain.toussaint at securivm.ca Tue Sep 10 02:21:28 2013 From: alain.toussaint at securivm.ca (Alain Toussaint) Date: Mon, 9 Sep 2013 22:21:28 -0400 Subject: [Aboriginal] building Chromium OS images on aboriginal In-Reply-To: <1378721935.8385.8@driftwood> References: <1378721935.8385.8@driftwood> Message-ID: <83648095-0551-4499-A64F-C541E4B870F8@securivm.ca> > Install aboriginal, yes. (It's just extracting a tarball.) > > Then building chromium under the result: I don't know? Can you build chromium > under Linux From Scratch? Because you can build linux from scratch under > aboriginal, I even have an automated build control image to do it: I suspect there would be no problems at all to build Chromium OS (hereafter Chr.OS) under LFS. The worst that can happen is if Chr.OS' cros_sdk command call chroot with argument not supported by LFS chroot (or the aboriginal one if I decide to test it under aboriginal). > > http://landley.net/aboriginal/control-images > > So you can bootstrap your way up to an arbitrarily more complex environment > starting from the aboriginal linux root filesystem. Building distros is hard > because they all assume they're building under themselves (even gentoo), but > if you can work out how to build chromium under LFS and document it I'd love > to make an automated control image to do that, and then regression test it > with each new release. :) I can't make promises right now because of my school load (I'm doing a certificate in canadian law to joint to my neuroscience unfinished bachelor and next year, I'll be starting a nursing course) but I can dedicate both my zacate box for that and a day of hacking to that goal (probably Sunday) along with 30 minutes here and there over the rest of the week > The above LFS is 6.8 and I'm updating to 7.4 as we speak. (This is delaying > the new release a bit, current source control includes the 3.11 kernel and > builds fine but I'm trying to get a new toybox release in there and LFS 7.4 > too.) So those package versions might be a bit stale, but should still work. > > Rob For the moment, I'll do a quick test of Chr.OS on aboriginal and see what barf and later on, I'll build LFS. Sound like a good plan? Alain 1378779688.0 From rob at landley.net Thu Sep 12 19:19:08 2013 From: rob at landley.net (Rob Landley) Date: Thu, 12 Sep 2013 14:19:08 -0500 Subject: [Aboriginal] Kernel patching failure with 1.2.4.1 In-Reply-To: <522AA18B.2070008@gentoo.org> (from patrick@gentoo.org on Fri Sep 6 22:46:19 2013) References: <1378517491.1985.15@driftwood> <522AA18B.2070008@gentoo.org> Message-ID: <1379013548.1974.18@driftwood> On 09/06/2013 10:46:19 PM, Patrick Lauer wrote: > >> I upgraded patch on my server, so you can reproduce it there. Query: can _you_ reproduce it there? > > I just did a fresh EXTRACT_ALL=1 there and it worked. > > > > Admittedly it worked against 3.11 instead of 3.10, but the patch in > > question created a new file... > > > > Does this still reproduce for you? > > Yes, on my desktop at home I can reliably make the patching fail with > patch-2.7.1 and make it work with patch-2.6.1. I've tested with a > previous aboriginal release (1.2.3 I think) which extracts and builds > fine, it is only the 1.2.4.1 release that shows the problem. I still can't reproduce this, not even on your server. Rob 1379013548.0 From rob at landley.net Thu Sep 12 16:26:40 2013 From: rob at landley.net (Rob Landley) Date: Thu, 12 Sep 2013 11:26:40 -0500 Subject: [Aboriginal] Cutting releases. Message-ID: <1379003200.1974.16@driftwood> I spent this morning (thursday the 12th) cutting a toybox and an aboriginal linux release, but it probably wont' get uploaded until tuesday due to Ohio Linuxfest. All the targets built fine (using the new grep and unmount even), and built dropbear and strace, and one of them built the whole of linux from scratch 6.8. But the qemu guys broke the mips networking (1.5.0 works, 1.6.0 does not), and I wasted enough time bisecting it that I had to leave for the airport before writing up proper release notes. (Giving a talk saturday, might even be recorded). I'm taking the netbook with me, not the big laptop I built all the targets on, so I can't finish it on the way and upload it there. Now grinding through the ~5000 back emails in my inbox, queueing up replies (like this one) to go out next time there's net that isn't strange about port 25... Rob 1379003200.0 From rob at landley.net Mon Sep 23 15:02:16 2013 From: rob at landley.net (Rob Landley) Date: Mon, 23 Sep 2013 10:02:16 -0500 Subject: [Aboriginal] Pondering license change GPL->MIT. Message-ID: <1379948536.1974.52@driftwood> So a year or two back I switched toybox from GPL to BSD after doing some serious historical auditing. My recent Ohio LinuxFest talk (the rise and fall of copyleft) may eventually have recordings up. But the real trigger was this thread that wandered by recently on the buildroot list: http://lists.busybox.net/pipermail/buildroot/2013-September/078330.html (I don't hugely follow that list but I created it abusing my root powers on the busybox/uClibc server to kick the traffic off the uClibc list, and then never unsubscribed from it...) And they're wasting huge amounts of time going on about how the GPL applies to a bunch of makefiles and shell scripts that don't directly produce binary output. What does the GPL mean when there's no binary version of the code to distribute separately from a source version? (How can you be on the hook to distribute source if you never distributed binaries because there's no binary version?) Aboriginal is a bunch of bash scripts. It's gplv2 because when I started this project GPL was The License. Now there's no such thing as "The GPL" (the Linux kernel and Samba can't share code, even though both are gpl and implement 2 ends of the same protocol). So I'm thinking of switching to the "zero clause BSD" toybox license, which looks enough like BSD to reassure people but is actually effectively public domain: http://landley.net/toybox/license.html I need to go through the source control history and find third party changes I've merged, but there actually aren't that many. (I've taken ideas from people, but mostly not code.) $ hg log -v | grep '^user:' | sort -u user: Alessio Igor Bogani user: Alessio Igor Bogani user: Eric S. Raymond user: landley at driftwood user: Marc Andre Tanner user: Mark Miller user: Rob Landley Not a big list... The largest lump of gplv2 code in the project is ccwrap, which I'm in the process of rewriting anyway. (Yeah, been stalled for a while. Working on it...) So I just thought I'd throw that out there. Shouldn't make any difference to anyone but me, but "GPL for a bunch of bash scripts" has struck me as funny for years... Rob 1379948536.0 From onefang at gmail.com Tue Sep 24 00:12:20 2013 From: onefang at gmail.com (David Seikel) Date: Tue, 24 Sep 2013 10:12:20 +1000 Subject: [Aboriginal] Pondering license change GPL->MIT. In-Reply-To: <1379948536.1974.52@driftwood> References: <1379948536.1974.52@driftwood> Message-ID: <20130924101220.65764046.onefang@gmail.com> On Mon, 23 Sep 2013 10:02:16 -0500 Rob Landley wrote: > So a year or two back I switched toybox from GPL to BSD after doing > some serious historical auditing. My recent Ohio LinuxFest talk (the > rise and fall of copyleft) may eventually have recordings up. But > the real trigger was this thread that wandered by recently on the > buildroot list: > > > http://lists.busybox.net/pipermail/buildroot/2013-September/078330.html > > (I don't hugely follow that list but I created it abusing my root > powers on the busybox/uClibc server to kick the traffic off the > uClibc list, and then never unsubscribed from it...) > > And they're wasting huge amounts of time going on about how the GPL > applies to a bunch of makefiles and shell scripts that don't > directly produce binary output. What does the GPL mean when there's > no binary version of the code to distribute separately from a source > version? (How can you be on the hook to distribute source if you > never distributed binaries because there's no binary version?) > > Aboriginal is a bunch of bash scripts. It's gplv2 because when I > started this project GPL was The License. Now there's no such thing > as "The GPL" (the Linux kernel and Samba can't share code, even > though both are gpl and implement 2 ends of the same protocol). > > So I'm thinking of switching to the "zero clause BSD" toybox > license, which looks enough like BSD to reassure people but is > actually effectively public domain: > > http://landley.net/toybox/license.html > > I need to go through the source control history and find third party > changes I've merged, but there actually aren't that many. (I've > taken ideas from people, but mostly not code.) > > $ hg log -v | grep '^user:' | sort -u > user: Alessio Igor Bogani > user: Alessio Igor Bogani > user: Eric S. Raymond > user: landley at driftwood > user: Marc Andre Tanner > user: Mark Miller > user: Rob Landley > > Not a big list... > > The largest lump of gplv2 code in the project is ccwrap, which I'm > in the process of rewriting anyway. (Yeah, been stalled for a while. > Working on it...) > > So I just thought I'd throw that out there. Shouldn't make any > difference to anyone but me, but "GPL for a bunch of bash scripts" > has struck me as funny for years... I'm not on that list, though you do have some "source code" from me (sources/targets/i486), and maybe a tiny patch or two. I agree with this change. -- A big old stinking pile of genius that no one wants coz there are too many silver coated monkeys in the world. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: not available URL: From rob at landley.net Tue Sep 24 02:04:46 2013 From: rob at landley.net (Rob Landley) Date: Mon, 23 Sep 2013 21:04:46 -0500 Subject: [Aboriginal] Pondering license change GPL->MIT. In-Reply-To: <20130924101220.65764046.onefang@gmail.com> (from onefang@gmail.com on Mon Sep 23 19:12:20 2013) References: <1379948536.1974.52@driftwood> <20130924101220.65764046.onefang@gmail.com> Message-ID: <1379988286.1974.67@driftwood> On 09/23/2013 07:12:20 PM, David Seikel wrote: > On Mon, 23 Sep 2013 10:02:16 -0500 Rob Landley > wrote: > > So I just thought I'd throw that out there. Shouldn't make any > > difference to anyone but me, but "GPL for a bunch of bash scripts" > > has struck me as funny for years... > > I'm not on that list, though you do have some "source code" from me > (sources/targets/i486), and maybe a tiny patch or two. I agree with > this change. Didn't say it was an exhaustive list. (Hence needing to search more. :) I plan to look at every commit, but there's 1000 of them. (I made it up to #36 before getting distracted...) Thanks, Rob 1379988286.0 From linux at roeck-us.net Sun Sep 22 05:30:15 2013 From: linux at roeck-us.net (Guenter Roeck) Date: Sat, 21 Sep 2013 22:30:15 -0700 Subject: [Aboriginal] [PATCH] sources/functions.sh: Fix ext2/ext3 build failures Message-ID: <1379827815-10151-1-git-send-email-linux@roeck-us.net> Signed-off-by: Guenter Roeck --- This patch should apply on top of the 1.2.5 release. It was generated using git format-patch from a git repo, so hg commands might not work. sources/functions.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sources/functions.sh b/sources/functions.sh index 4f700d2..36b2ab0 100755 --- a/sources/functions.sh +++ b/sources/functions.sh @@ -405,7 +405,7 @@ image_filesystem() BLOCKS=$[1024*(($(du -m -s "$1" | awk '{print $1}')*12)/10)] [ $BLOCKS -lt 4096 ] && BLOCKS=4096 - FILE="$.$SYSIMAGE_TYPE" + FILE="$2.$SYSIMAGE_TYPE" echo "/dev d 755 0 0 - - - - -" > "$WORK/devs" && echo "/dev/console c 640 0 0 5 1 0 0 -" >> "$WORK/devs" && @@ -423,7 +423,7 @@ image_filesystem() resize2fs "$FILE" ${SYSIMAGE_HDA_MEGS}M || dienow fi - tune2fs -c 0 -i 0 $([$SYS_IMAGE_TYPE == ext3] && echo -j) "$FILE" || dienow + tune2fs -c 0 -i 0 $([ "$SYS_IMAGE_TYPE" = "ext3" ] && echo -j) "$FILE" || dienow echo $SYSIMAGE_TYPE generated elif [ "$SYSIMAGE_TYPE" == "squashfs" ] -- 1.7.9.7 1379827815.0 From rob at landley.net Tue Sep 24 02:24:08 2013 From: rob at landley.net (Rob Landley) Date: Mon, 23 Sep 2013 21:24:08 -0500 Subject: [Aboriginal] new binutils needed for arm in 3.12-rc1 In-Reply-To: <20130923235917.GA30967@amd.pavel.ucw.cz> (from pavel@ucw.cz on Mon Sep 23 18:59:17 2013) References: <20130919093002.GA23806@amd.pavel.ucw.cz> <20130919093627.GB24866@mudshark.cambridge.arm.com> <20130923235917.GA30967@amd.pavel.ucw.cz> Message-ID: <1379989448.1974.71@driftwood> Teriffic. Anybody want to chime in on linux-kernel about this? Rob On 09/23/2013 06:59:17 PM, Pavel Machek wrote: > During 3.12-rc, Will Deacon introduced code into arch/arm that > requires binutils 2.22. > > Signed-off-by: Pavel Machek > > --- > > Or that changes should be reverted. I have updated my buildsystem on > main > machine, but ... it seems that debian-cross repository does not > contain new enough bintuils for kernel now. We are talking > > 62cbbc42e0019aff6310259f275ae812463f8836 > 6af396a6b6c698eb3834184518fc9a59bc22c817 > 73a6fdc48bf52e93c26874dc8c0f0f8d5585a809 > 6abdd491698a27f7df04a32ca12cc453810e4396 > > diff --git a/Documentation/Changes b/Documentation/Changes > index b175808..0f8deaf 100644 > --- a/Documentation/Changes > +++ b/Documentation/Changes > @@ -23,7 +23,7 @@ you probably needn't concern yourself with > isdn4k-utils. > > o Gnu C 3.2 # gcc --version > o Gnu make 3.80 # make --version > -o binutils 2.12 # ld -v > +o binutils 2.22 # ld -v > o util-linux 2.10o # fdformat > --version > o module-init-tools 0.9.10 # depmod -V > o e2fsprogs 1.41.4 # e2fsck -V > > -- > (english) http://www.livejournal.com/~pavelmachek > (cesky, pictures) > http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html > -- > To unsubscribe from this list: send the line "unsubscribe > linux-kernel" in > the body of a message to majordomo at vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ > 1379989448.0 From alain.toussaint at securivm.ca Tue Sep 24 02:39:15 2013 From: alain.toussaint at securivm.ca (Alain Toussaint) Date: Mon, 23 Sep 2013 22:39:15 -0400 Subject: [Aboriginal] new binutils needed for arm in 3.12-rc1 In-Reply-To: <1379989448.1974.71@driftwood> References: <20130919093002.GA23806@amd.pavel.ucw.cz> <20130919093627.GB24866@mudshark.cambridge.arm.com> <20130923235917.GA30967@amd.pavel.ucw.cz> <1379989448.1974.71@driftwood> Message-ID: <74AAC67C-9750-4F0A-8ED8-3B269769A761@securivm.ca> Ugh? From aboriginal: (x86_64:1) /home # ld -V GNU ld (GNU Binutils) 2.17.50.20070703 Supported emulations: elf_x86_64 elf_i386 i386linux While on zacate (gentoo multilib stage 3 tarball + qemu for a full range of target): alain at zacate ~ $ ld -V GNU ld (GNU Binutils) 2.23.1 Supported emulations: elf_x86_64 elf32_x86_64 elf_i386 i386linux elf_l1om elf_k1om Is there any rationale that binutils is 2.17 on aboriginal? Alain Le 2013-09-23 ? 22:24, Rob Landley a ?crit : > Teriffic. > > Anybody want to chime in on linux-kernel about this? > > Rob > > On 09/23/2013 06:59:17 PM, Pavel Machek wrote: >> During 3.12-rc, Will Deacon introduced code into arch/arm that >> requires binutils 2.22. >> Signed-off-by: Pavel Machek >> --- >> Or that changes should be reverted. I have updated my buildsystem on main >> machine, but ... it seems that debian-cross repository does not >> contain new enough bintuils for kernel now. We are talking >> 62cbbc42e0019aff6310259f275ae812463f8836 >> 6af396a6b6c698eb3834184518fc9a59bc22c817 >> 73a6fdc48bf52e93c26874dc8c0f0f8d5585a809 >> 6abdd491698a27f7df04a32ca12cc453810e4396 >> diff --git a/Documentation/Changes b/Documentation/Changes >> index b175808..0f8deaf 100644 >> --- a/Documentation/Changes >> +++ b/Documentation/Changes >> @@ -23,7 +23,7 @@ you probably needn't concern yourself with isdn4k-utils. >> o Gnu C 3.2 # gcc --version >> o Gnu make 3.80 # make --version >> -o binutils 2.12 # ld -v >> +o binutils 2.22 # ld -v >> o util-linux 2.10o # fdformat --version >> o module-init-tools 0.9.10 # depmod -V >> o e2fsprogs 1.41.4 # e2fsck -V >> -- >> (english) http://www.livejournal.com/~pavelmachek >> (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in >> the body of a message to majordomo at vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/ > > _______________________________________________ > Aboriginal mailing list > Aboriginal at lists.landley.net > http://lists.landley.net/listinfo.cgi/aboriginal-landley.net 1379990355.0 From rob at landley.net Tue Sep 24 21:12:05 2013 From: rob at landley.net (Rob Landley) Date: Tue, 24 Sep 2013 16:12:05 -0500 Subject: [Aboriginal] new binutils needed for arm in 3.12-rc1 In-Reply-To: <74AAC67C-9750-4F0A-8ED8-3B269769A761@securivm.ca> (from alain.toussaint@securivm.ca on Mon Sep 23 21:39:15 2013) References: <20130919093002.GA23806@amd.pavel.ucw.cz> <20130919093627.GB24866@mudshark.cambridge.arm.com> <20130923235917.GA30967@amd.pavel.ucw.cz> <1379989448.1974.71@driftwood> <74AAC67C-9750-4F0A-8ED8-3B269769A761@securivm.ca> Message-ID: <1380057125.1974.72@driftwood> On 09/23/2013 09:39:15 PM, Alain Toussaint wrote: > Is there any rationale that binutils is 2.17 on aboriginal? Yes, it's the last gplv2 release. I don't distribute GPLv3 binaries. Long term plan is to migrate to another toolchain, but llvm isn't quite there yet and nothing else is close. Rob 1380057125.0 From alain.toussaint at securivm.ca Wed Sep 25 02:00:57 2013 From: alain.toussaint at securivm.ca (Alain Toussaint) Date: Tue, 24 Sep 2013 22:00:57 -0400 Subject: [Aboriginal] bootstrapping llvm Message-ID: <9B3BF685-CB5E-4CB3-A41E-65660EA86AAE@securivm.ca> Hello, For the next 3 month, I'll work toward bootstrapping llvm into aboriginal. I'll first dissect the sources of aboriginal's bash script which are responsible for creating a cross-compiler as well as a native one in aboriginal's qemu images and I also plan to register on the various llvm mailing list applicable to the project. Alain 1380074457.0 From maillist-aboriginal at barfooze.de Wed Sep 25 02:12:14 2013 From: maillist-aboriginal at barfooze.de (John Spencer) Date: Wed, 25 Sep 2013 04:12:14 +0200 Subject: [Aboriginal] new binutils needed for arm in 3.12-rc1 In-Reply-To: <1379989448.1974.71@driftwood> References: <20130919093002.GA23806@amd.pavel.ucw.cz> <20130919093627.GB24866@mudshark.cambridge.arm.com> <20130923235917.GA30967@amd.pavel.ucw.cz> <1379989448.1974.71@driftwood> Message-ID: <5242467E.70300@barfooze.de> On 09/24/2013 04:24 AM, Rob Landley wrote: > Teriffic. > > Anybody want to chime in on linux-kernel about this? i think the best fix is to change the offending commit to use portable .byte or .word syntax instead of the new armv7 mnemonic as Rich suggested in #musl. i'm quite sure the kernel guys would accept such a change. > > Rob > > On 09/23/2013 06:59:17 PM, Pavel Machek wrote: >> During 3.12-rc, Will Deacon introduced code into arch/arm that >> requires binutils 2.22. >> >> Signed-off-by: Pavel Machek >> >> --- >> >> Or that changes should be reverted. I have updated my buildsystem on main >> machine, but ... it seems that debian-cross repository does not >> contain new enough bintuils for kernel now. We are talking >> >> 62cbbc42e0019aff6310259f275ae812463f8836 >> 6af396a6b6c698eb3834184518fc9a59bc22c817 >> 73a6fdc48bf52e93c26874dc8c0f0f8d5585a809 >> 6abdd491698a27f7df04a32ca12cc453810e4396 >> >> diff --git a/Documentation/Changes b/Documentation/Changes >> index b175808..0f8deaf 100644 >> --- a/Documentation/Changes >> +++ b/Documentation/Changes >> @@ -23,7 +23,7 @@ you probably needn't concern yourself with >> isdn4k-utils. >> >> o Gnu C 3.2 # gcc --version >> o Gnu make 3.80 # make --version >> -o binutils 2.12 # ld -v >> +o binutils 2.22 # ld -v >> o util-linux 2.10o # fdformat --version >> o module-init-tools 0.9.10 # depmod -V >> o e2fsprogs 1.41.4 # e2fsck -V >> >> -- >> (english) http://www.livejournal.com/~pavelmachek >> (cesky, pictures) >> http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html >> -- >> To unsubscribe from this list: send the line "unsubscribe >> linux-kernel" in >> the body of a message to majordomo at vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/ >> 1380075134.0 From rob at landley.net Wed Sep 25 15:11:00 2013 From: rob at landley.net (Rob Landley) Date: Wed, 25 Sep 2013 10:11:00 -0500 Subject: [Aboriginal] bootstrapping llvm In-Reply-To: <9B3BF685-CB5E-4CB3-A41E-65660EA86AAE@securivm.ca> (from alain.toussaint@securivm.ca on Tue Sep 24 21:00:57 2013) References: <9B3BF685-CB5E-4CB3-A41E-65660EA86AAE@securivm.ca> Message-ID: <1380121860.1974.83@driftwood> On 09/24/2013 09:00:57 PM, Alain Toussaint wrote: > Hello, > > For the next 3 month, I'll work toward bootstrapping llvm into > aboriginal. > I'll first dissect the sources of aboriginal's bash script which are > responsible for creating a cross-compiler as well as a native one in > aboriginal's qemu images and I also plan to register on the various > llvm > mailing list applicable to the project. Yay! I sent you some links already. If you can build a cross compiler even without a libc and use that to build a kernel, that's a huge step up. (We can use the existing root filesystem images to test said kernel, then tackle libc and such as a second pass. The kernel build shouldn't use any external libraries for the target code. The host code uses the host compiler, of course.) If uClibc doesn't build out of the box, I should have musl in before too much longer. That's been built with llvm before. Rob 1380121860.0 From rob at landley.net Wed Sep 25 15:32:41 2013 From: rob at landley.net (Rob Landley) Date: Wed, 25 Sep 2013 10:32:41 -0500 Subject: [Aboriginal] new binutils needed for arm in 3.12-rc1 In-Reply-To: <5242467E.70300@barfooze.de> (from maillist-aboriginal@barfooze.de on Tue Sep 24 21:12:14 2013) References: <20130919093002.GA23806@amd.pavel.ucw.cz> <20130919093627.GB24866@mudshark.cambridge.arm.com> <20130923235917.GA30967@amd.pavel.ucw.cz> <1379989448.1974.71@driftwood> <5242467E.70300@barfooze.de> Message-ID: <1380123161.1974.86@driftwood> On 09/24/2013 09:12:14 PM, John Spencer wrote: > On 09/24/2013 04:24 AM, Rob Landley wrote: >> Teriffic. >> >> Anybody want to chime in on linux-kernel about this? > > i think the best fix is to change the offending commit to use portable > .byte or .word syntax instead of the new armv7 mnemonic as Rich > suggested > in #musl. i'm quite sure the kernel guys would accept such a change. I'm not familiar enough with arm assembly to do that (I haven't yet looked up what the new mnemonic is doing), but I'd love a patch. :) Thanks, Rob 1380123161.0 From alain.toussaint at securivm.ca Wed Sep 25 19:45:03 2013 From: alain.toussaint at securivm.ca (Alain Toussaint) Date: Wed, 25 Sep 2013 15:45:03 -0400 Subject: [Aboriginal] bootstrapping llvm In-Reply-To: <1380121860.1974.83@driftwood> References: <9B3BF685-CB5E-4CB3-A41E-65660EA86AAE@securivm.ca> <1380121860.1974.83@driftwood> Message-ID: <5B7A395C-7FB7-4CAB-B2E6-7B82FE1C1165@securivm.ca> Envoy? de mon iPhone > Le 2013-09-25 ? 11:11, Rob Landley a ?crit : > >> On 09/24/2013 09:00:57 PM, Alain Toussaint wrote: >> Hello, >> For the next 3 month, I'll work toward bootstrapping llvm into aboriginal. >> I'll first dissect the sources of aboriginal's bash script which are >> responsible for creating a cross-compiler as well as a native one in >> aboriginal's qemu images and I also plan to register on the various llvm >> mailing list applicable to the project. > > Yay! > > I sent you some links already. If you can build a cross compiler even without a libc and use that to build a kernel, that's a huge step up. (We can use the existing root filesystem images to test said kernel, then tackle libc and such as a second pass. The kernel build shouldn't use any external libraries for the target code. The host code uses the host compiler, of course.) > > If uClibc doesn't build out of the box, I should have musl in before too much longer. That's been built with llvm before. > > Rob At the moment, I have built llvm 3.3 under gentoo with all the targets enabled but tonight when I get back home, I'll try to build a cross-compiler in my homedir. Alain 1380138303.0 From maillist-aboriginal at barfooze.de Thu Sep 26 18:53:49 2013 From: maillist-aboriginal at barfooze.de (John Spencer) Date: Thu, 26 Sep 2013 20:53:49 +0200 Subject: [Aboriginal] release with kernel 3.11 and posix_spawn ? Message-ID: <524482BD.7090209@barfooze.de> did i miss the release announcement, or is the 3.11 release still in the works? i could make good use of it currently as i plan to use it for another sabotage release for all supported archs, but this time without posix_spawn headaches... --JS 1380221629.0