Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
Hello, On Sun, 17 Feb 2008 19:54:16 +0100 Marcin Juszkiewicz <[EMAIL PROTECTED]> wrote: [] > > > | Personally I would go into 1st way if we are able to get working > > | system in ~400-500KB of initramfs (minimal Zaurus kernel for c7x0 > > was | ~700KB). > > > We would need only busybox and kexec-tools[1] in there, right? > > [1] does busybox have a kexec applet already? > > Paul methods can be used - I did not checked what he use as my c760 > is rather developer tool for use from time to time then each day use > PDA. IMHO, most promising way should be using klibc-linked utils (including its ash). Thomas Kunze works on that. I of course may imagine that will require patching for real-world usability (timeout for 'read' for example). But again, that should be more promising way than plucking busybox. Current numbers for initramfs-bootmenu-image: 820138 Angstrom-initramfs-bootmenu-image-uclibc-ipk-2007.12-r4-hx4700.rootfs.cpio.gz When bundled with the biggest linux-hh zImage (hx4700's due to pcmcia builtin): just dozen Kbs less than 2Mb which is pure luck as that's size of hx4700 kernel partition ;-). Inside: ~850Kb busybox, ~300k - NFS kernel modules. [] -- Best regards, Paul mailto:[EMAIL PROTECTED] ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
Dnia Sunday, 17 of February 2008, Koen Kooi napisał: > Marcin Juszkiewicz schreef: > | Basically there were two ideas according to Zaurus boot: > | > | 1. use kernel+initramfs in kernel partition and let it kexec > | /boot/zImage > > which leaves CMDLINE intact from /boot/zImage and thus behaves as a > non-broken angelboot, right? CMDLINE can be read from /boot/kernel_params (example name) file by initramfs and used for kexec-ed kernel. > | 2. use rescue flash partition to keep our boot+reflash system which > | will be executed by kernel from 1st point > > would be rescue filesystem be kernel+init(c)ramfs or a 'real' > filesystem? Whatever is possible. This needs someone who will try to re-use rescue partition (rootfs/kernel) for our use - remember that angelboot boots into rescue if 'OK' is hold on power on. > Would it include e.g. wireless-tools and parted? I'm certain ~ we can do > all that in less than 7MB, especially with uclibc+thumb We can add anything needed there. GNU Tar, Parted would be most important ones (think Spitz) but WiFi/BT support and bunch of drivers for networking would be nice to have. > (non-collie). Collie use simpler way of booting: 1. bootloader -> kernel 2. bootloader -> flasher 3. bootloader -> diagnostics There is no rescue kernel/rootfs like it is in PXA models. > | Personally I would go into 1st way if we are able to get working > | system in ~400-500KB of initramfs (minimal Zaurus kernel for c7x0 was > | ~700KB). > We would need only busybox and kexec-tools[1] in there, right? > [1] does busybox have a kexec applet already? Paul methods can be used - I did not checked what he use as my c760 is rather developer tool for use from time to time then each day use PDA. -- JID: hrw-jabber.org OpenEmbedded developer/consultant BSD is for people who love Unix. Linux is for people who hate Microsoft. ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Marcin Juszkiewicz schreef: | Dnia Sunday, 17 of February 2008, Koen Kooi napisał: |> * on angelboot (read: zaurus) machines the kernel partition holds |> stripped down kernel + initramfs that kexecs a second kernel (with more |> dirvers) with a bigger initramfs which kexecs a third kernel to boot |> into final userspace | |> Marcin: is that the same idea you had for angelboot machines, or will |> we repartition flash and put in our own rescue filesystem? | | Basically there were two ideas according to Zaurus boot: | | 1. use kernel+initramfs in kernel partition and let it kexec /boot/zImage which leaves CMDLINE intact from /boot/zImage and thus behaves as a non-broken angelboot, right? | 2. use rescue flash partition to keep our boot+reflash system which will |be executed by kernel from 1st point would be rescue filesystem be kernel+init(c)ramfs or a 'real' filesystem? Would it include e.g. wireless-tools and parted? I'm certain ~ we can do all that in less than 7MB, especially with uclibc+thumb (non-collie). | Personally I would go into 1st way if we are able to get working system in | ~400-500KB of initramfs (minimal Zaurus kernel for c7x0 was ~700KB). We would need only busybox and kexec-tools[1] in there, right? regards, Koen [1] does busybox have a kexec applet already? -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFHuEDlMkyGM64RGpERAujKAJ9rYr5U1OMJPDkNSeORF1EH7ZBW2QCbBE1Z P9vAybq54zJE1dhk2M4Wb8c= =4zz+ -END PGP SIGNATURE- ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
Dnia Sunday, 17 of February 2008, Koen Kooi napisał: > * on angelboot (read: zaurus) machines the kernel partition holds > stripped down kernel + initramfs that kexecs a second kernel (with more > dirvers) with a bigger initramfs which kexecs a third kernel to boot > into final userspace > Marcin: is that the same idea you had for angelboot machines, or will > we repartition flash and put in our own rescue filesystem? Basically there were two ideas according to Zaurus boot: 1. use kernel+initramfs in kernel partition and let it kexec /boot/zImage 2. use rescue flash partition to keep our boot+reflash system which will be executed by kernel from 1st point Personally I would go into 1st way if we are able to get working system in ~400-500KB of initramfs (minimal Zaurus kernel for c7x0 was ~700KB). -- JID: hrw-jabber.org OpenEmbedded developer/consultant We were talking about everything. It's called friendship. It's like therapy for poor people [WaT 2x18] ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
Koen Kooi wrote: > * on angelboot (read: zaurus) machines the kernel partition holds > stripped down kernel + initramfs that kexecs a second kernel (with more > dirvers) with a bigger initramfs which kexecs a third kernel to boot > into final userspace > I think we should have a stripped down kernel + initramfs thats kexecs the kernel to boot into userspace. When the klibc stuff is ready, we only need to disable networking the get the kernel small enough to get the initramfs into the partition too (I hope). So we would only gain the boot-from-nfs option if we use the solution you describe. Regards, Thomas ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Paul Sokolovsky schreef: | Well, I sent RFC on that: | http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2008-February/001776.html | So yes, this is 2nd stage kernel+initramfs solution, replacing a kernel | loaded by 1nd stage bootloader (be it haret, HH.org bootloader, | anything else) [some time ago I proposed to consider NAND bootstrap | loader as 0th stage, so I assume that numbering]. At the same time, | bootloading (actually, interactive selection) functionality is made to | be transparent and used only if no "root=" was specified on kernel | commandline. So actually, it's a drop-in replacement for production | kernel too, which fact RFC above emphasizes and exploits. Let me see if I understand it correctly: * on (non-hx4700) bootldr machines /boot/zImage will hold kernel + initramfs that kexecs a different kernel * on hx4700 the kernel partition holds kernel + initramfs that kexecs a different kernel (/boot/zImage?) * on angelboot (read: zaurus) machines the kernel partition holds stripped down kernel + initramfs that kexecs a second kernel (with more dirvers) with a bigger initramfs which kexecs a third kernel to boot into final userspace * we leave u-boot machines alone for now, since most u-boots are quite sensible nowadays. Marcin: is that the same idea you had for angelboot machines, or will we repartition flash and put in our own rescue filesystem? All in all this seems to becoming close to the debian-style initramfs installer we have wanted for angstrom for a long, long time. regards, Koen - -- [EMAIL PROTECTED] will go go away in december 2007, please use [EMAIL PROTECTED] instead. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFHuBlEMkyGM64RGpERAqhKAKCu3A4QlIJ5iAyXN0K54L+6hxcskACdE4Un 6NI6SLhUqVVjYTPNl5S4LJ8= =A3jj -END PGP SIGNATURE- ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
Hello, On Sun, 17 Feb 2008 11:17:28 +0100 Koen Kooi <[EMAIL PROTECTED]> wrote: [] > But |> only when new ones ready. I think that we should prioritize > package |> manager upgrades by all means, and use image upgrades only > when package |> manager upgrades don't work smoothly. > > It's not meant as an image upgrade, it's meant to improve the > important 'first impression' new users get from angstrom. Ok by me then. (Except that I have so much stuff unmerged to help with that and daily use even better ;-) ). [] > |> I'd like to bump to r5 right now, marking that linux-hh kernels now > |> support kexec command line. > > Sure, please go ahead with that. Done. > > | So, it would be Angstrom 2007.12-r5 and > |> kernel 2.6.21-hh20-r13. Ones the new bootloaders are out, all > before |> that would need to be nuked. (And all that because of > thinko in the |> arm-linux kernel and/or kexec command line support > patch, which |> bothers to put atags not where old kernels expect > them). > > What do you meant exactly with 'bootloader', is that a bootldr > replace, LAB, kernel+initramfs as first stage, kernel+initramfs as > second stage, something else? Well, I sent RFC on that: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2008-February/001776.html So yes, this is 2nd stage kernel+initramfs solution, replacing a kernel loaded by 1nd stage bootloader (be it haret, HH.org bootloader, anything else) [some time ago I proposed to consider NAND bootstrap loader as 0th stage, so I assume that numbering]. At the same time, bootloading (actually, interactive selection) functionality is made to be transparent and used only if no "root=" was specified on kernel commandline. So actually, it's a drop-in replacement for production kernel too, which fact RFC above emphasizes and exploits. (But I intend to add recognition and ignoring of botched commandlines bootldr passes, so this thing is going to solve lots of issues for us ;-) ). > > regards, > > Koen > > PS: altboot is broken because it doesn't handle timed-out reads, so > zaurus people can stop claiming there's a magical kernel version that > solved it; there isn't Well, at this time, initramfs-bootmenu-image doesn't support timeout at all. That's mostly because timeout implies ability to save last choice, and that in turn is machine-specific stuff which I wanted to keep away from initramfs-bootmenu-image until it will mature and procedure for its building will be settled. And even for specific machines, that requires good thought where to store settings. "If mtd map has 'kernel' partition, store settings in last few bytes of it"? How does that sound? ;-) [] -- Best regards, Paul mailto:[EMAIL PROTECTED] ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Paul Sokolovsky schreef: | Hello, | | On Sun, 17 Feb 2008 10:26:24 +0100 | Koen Kooi wrote: | | Hi, | | We've fixed a number of bugs in the stable branch and it would be nice | if users can benefit from that straight away, without having to use | the package manager. Up until this point there have been no big kernel | upgrades (e.g 2.6.x -> 2.6.x+1), only defconfig tweaks. | | So: | | Update release images with 2007.12-rX, where X >= 4 | | What do you think about this? | |> That's something I was thinking about as means to get out of the |> situation where both old bootloader and rootfs images for linux-hh |> devices are cross-incompatible with the new ones (to appear yet). But |> only when new ones ready. I think that we should prioritize package |> manager upgrades by all means, and use image upgrades only when package |> manager upgrades don't work smoothly. It's not meant as an image upgrade, it's meant to improve the important 'first impression' new users get from angstrom. |> On the other hand, r0 went with pretty annoying 100% CPU usage bug, |> which would be nice to exterminate completely. So, overall 0 points |> from me. | | If you think this is a good idea, do we | want to shoot for -r4 (after testing), or do mentors want to get some | more bugfixes in and aim for -r5? | |> I'd like to bump to r5 right now, marking that linux-hh kernels now |> support kexec command line. Sure, please go ahead with that. | So, it would be Angstrom 2007.12-r5 and |> kernel 2.6.21-hh20-r13. Ones the new bootloaders are out, all before |> that would need to be nuked. (And all that because of thinko in the |> arm-linux kernel and/or kexec command line support patch, which |> bothers to put atags not where old kernels expect them). What do you meant exactly with 'bootloader', is that a bootldr replace, LAB, kernel+initramfs as first stage, kernel+initramfs as second stage, something else? regards, Koen PS: altboot is broken because it doesn't handle timed-out reads, so zaurus people can stop claiming there's a magical kernel version that solved it; there isn't - -- [EMAIL PROTECTED] will go go away in december 2007, please use [EMAIL PROTECTED] instead. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFHuAm3MkyGM64RGpERAjGdAJ9QPX7busiMFb9JTeNEUbfvY5c4hQCcDSuM SzMqI/rLU4gIemdr3G7p3S8= =UDgW -END PGP SIGNATURE- ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] update release images with 2007.12-rX
Hello, On Sun, 17 Feb 2008 10:26:24 +0100 Koen Kooi <[EMAIL PROTECTED]> wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > Hi, > > We've fixed a number of bugs in the stable branch and it would be nice > if users can benefit from that straight away, without having to use > the package manager. Up until this point there have been no big kernel > upgrades (e.g 2.6.x -> 2.6.x+1), only defconfig tweaks. > > So: > > Update release images with 2007.12-rX, where X >= 4 > > What do you think about this? That's something I was thinking about as means to get out of the situation where both old bootloader and rootfs images for linux-hh devices are cross-incompatible with the new ones (to appear yet). But only when new ones ready. I think that we should prioritize package manager upgrades by all means, and use image upgrades only when package manager upgrades don't work smoothly. On the other hand, r0 went with pretty annoying 100% CPU usage bug, which would be nice to exterminate completely. So, overall 0 points from me. > If you think this is a good idea, do we > want to shoot for -r4 (after testing), or do mentors want to get some > more bugfixes in and aim for -r5? I'd like to bump to r5 right now, marking that linux-hh kernels now support kexec command line. So, it would be Angstrom 2007.12-r5 and kernel 2.6.21-hh20-r13. Ones the new bootloaders are out, all before that would need to be nuked. (And all that because of thinko in the arm-linux kernel and/or kexec command line support patch, which bothers to put atags not where old kernels expect them). > > regards, > > Koen > > - -- > [EMAIL PROTECTED] will go go away in december 2007, > please use [EMAIL PROTECTED] instead. > > -BEGIN PGP SIGNATURE- > Version: GnuPG v1.4.5 (Darwin) > > iD8DBQFHt/3AMkyGM64RGpERAsDBAJ4728mOXxV7udwK+D7nWIlHVXoCvQCfSlL2 > CoQt9kuPO7VW7Ck9MxNWwtg= > =c6mC > -END PGP SIGNATURE- > > ___ > Angstrom-distro-devel mailing list > Angstrom-distro-devel@linuxtogo.org > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel -- Best regards, Paul mailto:[EMAIL PROTECTED] ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel