Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On 16 Aug 2007, at 12:39 pm, Thomas Evans wrote: i don't need any of the 2.6.23 functionality - I was just really trying the 2.6.23 kernel to make sure nothing major was broken with it. The only other thing I know of which is broken with it is both autofs and amd automounters; this is because of the new nosharedcache option in NFS mounts, which currently the userland mount.nfs and automounters do not know about. Nothing to do with Alphas per se, but a problem some might want to be aware of. Tim -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Hello, On Tue, Aug 21, 2007 at 07:03:52PM -0700, Steve Langasek wrote: On Tue, Aug 21, 2007 at 08:30:56PM -0500, Bob Tracy wrote: The SRM-HOWTO says netbooting with aboot is broken: there's a way to make a netbootable (using bootp) image using the kernel source tree code under arch/alpha/boot. As I said, though, I've not had any need for that functionality, and I haven't tried it. It's possible that the SRM-HOWTO included with the aboot source is out of date and therefore mistaken about netbooting being broken. Ah, that sounds like a bug in the SRM-HOWTO then, yes. Thanks to Helge, the aboot package includes a netabootwrap command that can be used to make If I gt a patch for the SRM-HOWTO (preferably after testing) I can include it both upstream and in the Debian package. Greetings Helge -- Dr. Helge Kreutzmann, Dipl.-Phys. [EMAIL PROTECTED] gpg signed mail preferred 64bit GNU powered http://www.itp.uni-hannover.de/~kreutzm Help keep free software libre: http://www.ffii.de/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Tried to rebuild on stable and got: gcc -I/home/sacha-utf/work/aboot-1.0~pre20040408/include -D__KERNEL__ -mcpu=ev4 -Os -Wall -fno-builtin -Wcast-align -mno-fp-regs -ffixed-8 -fno-builtin-printf -Os -Wall -mno-fp-regs -fno-builtin -I/home/sacha-utf/work/aboot-1.0~pre20040408/include -D__KERNEL__ -I../include -c -o isolib.o isolib.c In file included from isolib.c:10: /usr/include/sys/types.h:46: error: conflicting types for ‘loff_t’ /usr/include/linux/types.h:66: error: previous declaration of ‘loff_t’ was here /usr/include/sys/types.h:62: error: conflicting types for ‘dev_t’ /usr/include/linux/types.h:32: error: previous declaration of ‘dev_t’ was here In file included from /usr/include/sys/select.h:44, from /usr/include/sys/types.h:216, from isolib.c:10: /usr/include/time.h:119: error: redefinition of ‘struct timespec’ In file included from /usr/include/sys/select.h:46, from /usr/include/sys/types.h:216, from isolib.c:10: /usr/include/bits/time.h:70: error: redefinition of ‘struct timeval’ In file included from /usr/include/sys/types.h:216, from isolib.c:10: /usr/include/sys/select.h:78: error: conflicting types for ‘fd_set’ /usr/include/linux/types.h:31: error: previous declaration of ‘fd_set’ was here In file included from isolib.c:10: /usr/include/sys/types.h:231: error: conflicting types for ‘blkcnt_t’ /usr/include/linux/types.h:150: error: previous declaration of ‘blkcnt_t’ was here Anybody met something like this before? -- Alexander Kotelnikov Saint-Petersburg, Russia -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Fri, Aug 24, 2007 at 12:21:05AM +0400, Alexander Kotelnikov wrote: Tried to rebuild on stable and got: gcc -I/home/sacha-utf/work/aboot-1.0~pre20040408/include -D__KERNEL__ -mcpu=ev4 -Os -Wall -fno-builtin -Wcast-align -mno-fp-regs -ffixed-8 -fno-builtin-printf -Os -Wall -mno-fp-regs -fno-builtin -I/home/sacha-utf/work/aboot-1.0~pre20040408/include -D__KERNEL__ -I../include -c -o isolib.o isolib.c In file included from isolib.c:10: /usr/include/sys/types.h:46: error: conflicting types for ‘loff_t’ /usr/include/linux/types.h:66: error: previous declaration of ‘loff_t’ was here /usr/include/sys/types.h:62: error: conflicting types for ‘dev_t’ /usr/include/linux/types.h:32: error: previous declaration of ‘dev_t’ was here In file included from /usr/include/sys/select.h:44, from /usr/include/sys/types.h:216, from isolib.c:10: /usr/include/time.h:119: error: redefinition of ‘struct timespec’ In file included from /usr/include/sys/select.h:46, from /usr/include/sys/types.h:216, from isolib.c:10: /usr/include/bits/time.h:70: error: redefinition of ‘struct timeval’ In file included from /usr/include/sys/types.h:216, from isolib.c:10: /usr/include/sys/select.h:78: error: conflicting types for ‘fd_set’ /usr/include/linux/types.h:31: error: previous declaration of ‘fd_set’ was here In file included from isolib.c:10: /usr/include/sys/types.h:231: error: conflicting types for ‘blkcnt_t’ /usr/include/linux/types.h:150: error: previous declaration of ‘blkcnt_t’ was here Anybody met something like this before? I assume this is with 1.0~pre20040408-2? This looks like it's the result of my fixes to get the aboot-cross package to build successfully on i386; I didn't test this at all with older kernel header packages, and I'm not sure I care to try to fix this given that the above is a bug in that version of linux/types.h and working around it would be time-consuming. (Hmm, maybe I should have kept the build-dep on linux-libc-dev after all...) Suggestions to you on how to proceed: - grab the aboot-base package from unstable, which is arch: all and has no dependencies, and install it without worrying about the aboot package; the bootloader itself is contained in aboot-base, so that should be sufficient to get you a working 2.6.23 boot. - install linux-libc-dev from unstable, and build the package against that instead of the linux-kernel-headers included in etch. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Steve Langasek wrote: It's been uploaded to unstable now. Thanks again to Bob for the testing! Thanks Bob - finally got around to it yesterday - also worked for me with 2.6.23-rc3. I will try to test tonight or tomorrow - is there a .deb already, or should I build it? I went to the link looking for a .deb - which is why I backed off - figured someone would ask the questions that I was afraid to ask. Sorry, I tend not to distribute binary .debs directly because I don't want people to blindly trust unsigned .debs (from me or from anyone), and I don't want to be distributing signed changes files (the standard method for signing .debs) that could be uploaded straight to the Debian archive without my consent. (The latter actually wasn't an issue here, but meh, habits.) Well, I'm happy that it has all been tracked down and resolved. ...tom -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Steve Langasek wrote: Suggestions to you on how to proceed: (...) - install linux-libc-dev from unstable, and build the package against that instead of the linux-kernel-headers included in etch. That would have been my recommendation as well. I used linux-libc-dev 2.6.21-6 for my aboot build. -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 05:52:04PM -0400, Tom Evans wrote: Please help me understand - no patched version will go to experimental/unstable? It's been uploaded to unstable now. Thanks again to Bob for the testing! I will try to test tonight or tomorrow - is there a .deb already, or should I build it? I went to the link looking for a .deb - which is why I backed off - figured someone would ask the questions that I was afraid to ask. Sorry, I tend not to distribute binary .debs directly because I don't want people to blindly trust unsigned .debs (from me or from anyone), and I don't want to be distributing signed changes files (the standard method for signing .debs) that could be uploaded straight to the Debian archive without my consent. (The latter actually wasn't an issue here, but meh, habits.) Also, it was unclear to me what the difference between Richard's patch and the once provided. I realize it addresses issues other than the PT_NOTE issue - is that the sole reason that it is not used? I didn't use Richard's patch directly because it duplicated build fixes that had already been applied, and it included changes which I believe were both wrong and irrelevant to the PT_LOAD issue. On Tue, Aug 21, 2007 at 05:22:14PM -0500, Bob Tracy wrote: Steve's in a better position to explain himself, but my interpretation of recent postings is that nothing's going to experimental/unstable until *someone* screws up his courage and actually tries using a bootlx built from his patched source package. Well, it would've gone to unstable eventually, but if nobody else was able to test it, that probably wouldn't happen until I had a 2.6.23 kernel to test it with myself... -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
And here we have another patch http://groups.google.com/group/linux.kernel/browse_thread/thread/1abcbf6f1b5a7b91/ which one is going to experimental/unstable? On Thu, 16 Aug 2007 22:38:53 -0700 SL == Steve Langasek [EMAIL PROTECTED] wrote: SL SL On Thu, Aug 16, 2007 at 10:50:06PM -0500, Bob Tracy wrote: The aboot patch in the referenced l-k posting did the trick. My Alpha is up and running on 2.6.23-rc3 as I type this... SL I'm still willing to test an official updated aboot package when it's available, but the value added at this point is probably minimal if the Debian source mods (for packaging) are trivial. SL SL Please have a look at the latest package available from SL http://people.debian.org/~vorlon/aboot/. This does need tested, because SL the patch posted by Richard is muddled with unrelated changes that aren't SL needed for building in Debian, so I'm not 100% sure that I've grabbed all SL the bits required for the fix. -- Alexander Kotelnikov Saint-Petersburg, Russia -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 12:22:00PM +0400, Alexander Kotelnikov wrote: And here we have another patch http://groups.google.com/group/linux.kernel/browse_thread/thread/1abcbf6f1b5a7b91/ That's the same patch (and AFAICT, the same message). which one is going to experimental/unstable? Neither. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 12:22:00PM +0400, Alexander Kotelnikov wrote: And here we have another patch http://groups.google.com/group/linux.kernel/browse_thread/thread/1abcbf6f1 b5a7b91/ That's the same patch (and AFAICT, the same message). which one is going to experimental/unstable? Neither. And how will we boot 2.6.23 then??? The second question is: when we update aboot, how does the boot sectors on discs are rewritten on installation? Because a simple apt-get dist-upgrade would update the aboot binaries and install the new kernel. If the user tries to reboot, he will crash... There should be some automatism that tries to rewrite boot records on installation of the new aboot package! Uwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 11:08:54AM +0200, Uwe Schindler wrote: On Tue, Aug 21, 2007 at 12:22:00PM +0400, Alexander Kotelnikov wrote: And here we have another patch http://groups.google.com/group/linux.kernel/browse_thread/thread/1abcbf6f1 b5a7b91/ That's the same patch (and AFAICT, the same message). which one is going to experimental/unstable? Neither. And how will we boot 2.6.23 then??? How should I know? I asked for testing in http://lists.debian.org/debian-alpha/2007/08/msg00018.html, and apparently nobody who's concerned about 2.6.23 yet has seen fit to do so. The second question is: when we update aboot, how does the boot sectors on discs are rewritten on installation? Because a simple apt-get dist-upgrade would update the aboot binaries and install the new kernel. If the user tries to reboot, he will crash... There should be some automatism that tries to rewrite boot records on installation of the new aboot package! Patches welcome... -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Steve Langasek wrote: On Tue, Aug 21, 2007 at 11:08:54AM +0200, Uwe Schindler wrote: And how will we boot 2.6.23 then??? How should I know? I asked for testing in http://lists.debian.org/debian-alpha/2007/08/msg00018.html, and apparently nobody who's concerned about 2.6.23 yet has seen fit to do so. Huh??? I sent you the build logs: never got an ACK. Successfully installed the resulting binary .deb files (aboot-base and aboot). I'll actually try booting 2.6.23-rcX this evening with the new bootlx. The second question is: when we update aboot, how does the boot sectors on discs are rewritten on installation? Current answer: you do it manually. I guess the question is, is it safe to assume that /sbin/swriteboot -v -f1 /dev/sda /boot/bootlx is always the correct thing to do? I think the -f1 is safe, whether needed or not. Might want to sanity-check the device name for the first hard disk, and prompt the user Update boot sectors on /dev/??? (Y/n)? Beyond that, I don't know what hard disk boot scenarios make sense across the entire Alpha product line. -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 10:01:38 -0500, Bob Tracy wrote: Steve Langasek wrote: On Tue, Aug 21, 2007 at 11:08:54AM +0200, Uwe Schindler wrote: And how will we boot 2.6.23 then??? How should I know? I asked for testing in http://lists.debian.org/debian-alpha/2007/08/msg00018.html, and apparently nobody who's concerned about 2.6.23 yet has seen fit to do so. Huh??? I sent you the build logs: never got an ACK. http://lists.debian.org/msgid-search/[EMAIL PROTECTED] Cheers, Julien -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Previous time I tried not tested patch to aboot (not from Steve) it end up with major breackage: http://lists.debian.org/debian-alpha/2002/12/msg00074.html So what I am talking about? Steve, if you want somebody to test something you should assure him/her that it should work. Why don't you boot your alpha even with pre-2.6.23-rc1 kernel with these proposed changes? On Tue, 21 Aug 2007 03:16:59 -0700 SL == Steve Langasek [EMAIL PROTECTED] wrote: SL SL How should I know? I asked for testing in SL http://lists.debian.org/debian-alpha/2007/08/msg00018.html, and apparently SL nobody who's concerned about 2.6.23 yet has seen fit to do so. -- Alexander Kotelnikov Saint-Petersburg, Russia -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Julien Cristau wrote: On Tue, Aug 21, 2007 at 10:01:38 -0500, Bob Tracy wrote: Huh??? I sent you the build logs: never got an ACK. http://lists.debian.org/msgid-search/[EMAIL PROTECTED] Whoops! Thanks, Julien... I missed the post in all the dreck that passes for signal (as opposed to noise) in my in box these days :-(. As mentioned earlier, I'll try booting 2.6.23-rcX with the new bootlx this evening. Sorry for the mixup... I can also test against a Debian 2.6.18 kernel as far as the pre-2.6.23 case. Was the issue with using aboot for net booting addressed with this patch set? I've not had need of that functionality, but I can probably rig up a test if there's a reason to do so... -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Please help me understand - no patched version will go to experimental/unstable? I will try to test tonight or tomorrow - is there a .deb already, or should I build it? I went to the link looking for a .deb - which is why I backed off - figured someone would ask the questions that I was afraid to ask. Also, it was unclear to me what the difference between Richard's patch and the once provided. I realize it addresses issues other than the PT_NOTE issue - is that the sole reason that it is not used? Thanks, ...tom Steve Langasek wrote: On Tue, Aug 21, 2007 at 12:22:00PM +0400, Alexander Kotelnikov wrote: And here we have another patch http://groups.google.com/group/linux.kernel/browse_thread/thread/1abcbf6f1b5a7b91/ That's the same patch (and AFAICT, the same message). which one is going to experimental/unstable? Neither. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Tom Evans wrote: Please help me understand - no patched version will go to experimental/unstable? I will try to test tonight or tomorrow - is there a .deb already, or should I build it? I went to the link looking for a .deb - which is why I backed off - figured someone would ask the questions that I was afraid to ask. Also, it was unclear to me what the difference between Richard's patch and the once provided. I realize it addresses issues other than the PT_NOTE issue - is that the sole reason that it is not used? Steve's in a better position to explain himself, but my interpretation of recent postings is that nothing's going to experimental/unstable until *someone* screws up his courage and actually tries using a bootlx built from his patched source package. Yes, experimental and unstable both imply that a package in that category could conceivably result in hair growth on your palms and peace without honor, but the wider community has gotten spoiled and expects better quality than that. Consider yourself a qualified tester if you have an alternative method of booting your machine in the event the new bootlx is completely hosed :-). In all seriousness, the risk of damage to your system other than can't boot from the hard drive is minimal if you read the swriteboot documentation carefully and proceed with caution. By the original author's admission, swriteboot doesn't perform a lot of error checking, but as long as you aren't deliberately trying to screw yourself over, the program is pretty safe. It would probably be wise not to volunteer for testing if your machine is set up to boot multiple OSes. I feel pretty safe: I've got a bootable CD that should get the system up far enough to install a working bootlx if the new one doesn't work. -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 08:31:45PM +0400, Alexander Kotelnikov wrote: Previous time I tried not tested patch to aboot (not from Steve) it end up with major breackage: http://lists.debian.org/debian-alpha/2002/12/msg00074.html So what I am talking about? Steve, if you want somebody to test something you should assure him/her that it should work. Why don't you boot your alpha even with pre-2.6.23-rc1 kernel with these proposed changes? I don't know why you would think that I asked for testing of such a build without first verifying that it didn't introduce regressions for currently working 2.6.23 kernels. Yes, this still boots 2.6.22 just fine. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Tue, Aug 21, 2007 at 08:30:56PM -0500, Bob Tracy wrote: Steve Langasek wrote: On Tue, Aug 21, 2007 at 01:36:36PM -0500, Bob Tracy wrote: Was the issue with using aboot for net booting addressed with this patch set? I've not had need of that functionality, but I can probably rig up a test if there's a reason to do so... The patch I applied was strictly for fixing the PT_LOAD issue. I don't know what net booting issue you're referring to, but it's entirely possible that it's one that was fixed upstream prior to 1.0pre20040408. The SRM-HOWTO says netbooting with aboot is broken: there's a way to make a netbootable (using bootp) image using the kernel source tree code under arch/alpha/boot. As I said, though, I've not had any need for that functionality, and I haven't tried it. It's possible that the SRM-HOWTO included with the aboot source is out of date and therefore mistaken about netbooting being broken. Ah, that sounds like a bug in the SRM-HOWTO then, yes. Thanks to Helge, the aboot package includes a netabootwrap command that can be used to make netbootable images from a kernel and initrd; indeed, this is what's used to make all of the current netboot images for debian-installer, which is the primary mode I use for installer testing. However, I'm happy to report that your new aboot works just fine with 2.6.23-rc3. Cool, thanks for the test. I'll prepare an upload to unstable shortly. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Fri, 17 Aug 2007 17:05:34 -0700 SL == Steve Langasek [EMAIL PROTECTED] wrote: SL SL On Fri, Aug 17, 2007 at 09:59:09AM -0500, Bob Tracy wrote: Steve Langasek wrote: Please have a look at the latest package available from http://people.debian.org/~vorlon/aboot/. This does need tested, because the patch posted by Richard is muddled with unrelated changes that aren't needed for building in Debian, so I'm not 100% sure that I've grabbed all the bits required for the fix. SL ACK. Build log sent under separate cover... SL SL Ah, thanks, but I've built the package here as well so I know what the build SL output looks like. :) It's testing it with a 2.6.23rc kernel that we need, SL since I don't have time to build an rc kernel here. But have at least you tried this aboot with any other kernel? -- Alexander Kotelnikov Saint-Petersburg, Russia -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Steve Langasek wrote: Please have a look at the latest package available from http://people.debian.org/~vorlon/aboot/. This does need tested, because the patch posted by Richard is muddled with unrelated changes that aren't needed for building in Debian, so I'm not 100% sure that I've grabbed all the bits required for the fix. ACK. Build log sent under separate cover... -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Thu, Aug 16, 2007 at 10:50:06PM -0500, Bob Tracy wrote: The aboot patch in the referenced l-k posting did the trick. My Alpha is up and running on 2.6.23-rc3 as I type this... I'm still willing to test an official updated aboot package when it's available, but the value added at this point is probably minimal if the Debian source mods (for packaging) are trivial. I will start working on a new official deb in the next days, but since I no longer have console access I cannot test it, so I will need your feedback. For the moment please continue testing Steves version and report your results. Greetings Helge -- Dr. Helge Kreutzmann, Dipl.-Phys. [EMAIL PROTECTED] gpg signed mail preferred 64bit GNU powered http://www.itp.uni-hannover.de/~kreutzm Help keep free software libre: http://www.ffii.de/ signature.asc Description: Digital signature
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Fri, Aug 17, 2007 at 09:59:09AM -0500, Bob Tracy wrote: Steve Langasek wrote: Please have a look at the latest package available from http://people.debian.org/~vorlon/aboot/. This does need tested, because the patch posted by Richard is muddled with unrelated changes that aren't needed for building in Debian, so I'm not 100% sure that I've grabbed all the bits required for the fix. ACK. Build log sent under separate cover... Ah, thanks, but I've built the package here as well so I know what the build output looks like. :) It's testing it with a 2.6.23rc kernel that we need, since I don't have time to build an rc kernel here. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Tom Evans wrote: I was trying to boot 2.6.23-rc1 a few weeks ago. Asked about the newer binutils - thought that may have been the reason it failing. Glad you hit the kml instead and found an answer (and glad I wasn't the only person seeing a problem with the 2.6.23-rcX series). It wasn't a binutils issue, but that the stock aboot is misguided. I'd also be willing to test a patched aboot. If you can't wait for that, you can always use BOOTP, which bypasses aboot entirely... :-) Just: make boot make bootpzfile scp arch/alpha/boot/bootpzfile someplace:/tftpboot However, your config prolly would need to change, to build in most of the stuff that's now loaded from INITRD, although that can be accomplished as well via some trickery... ;-} Good luck. --Jay++ ...tom Bob Tracy wrote: I originally posted to linux-kernel w.r.t. 2.6.23-rcX not booting: Unfortunately, I can't say where the bug was introduced, as this is the first kernel I've tried on my Alpha since 2.6.22-rc7. Best guess is somewhere between .23-rc1 and .23-rc2, based on changes to files in arch/alpha/boot in that patch set. The problem happens early: aboot starts to load vmlinux.gz, and I get an unzip: invalid exec header error. There's an earlier error from aboot I can't quote exactly, but the template from aboot (bootlx) is: aboot: Can't load kernel. Memory at %lx - %lx (chunk %i) is %s The first %lx is 0. The last %lx is all fs. The chunk number is 1, and I *think* the %s is busy. The posted reply was: try http://lkml.org/lkml/2007/8/14/259 As it turns out, the problem was actually introduced with 2.6.23-rc1: 2.6.22 works fine. The l-k people say aboot needs a patch. Question to Steve L. et al: is a patch in the works? Would you like a motivated tester? :-) The patch supplied at the above URL doesn't apply cleanly to the current Debian aboot-0.9b-3 source tree. -- Jay A EstabrookHPTC - XC I B Hewlett-Packard Company - ZKO1-3/D-B.8 (603) 884-0301 110 Spit Brook Road, Nashua NH 03062 [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
i don't need any of the 2.6.23 functionality - I was just really trying the 2.6.23 kernel to make sure nothing major was broken with it. I *could* use bootp I suppose, but that would require a tftp service someplace - suppose that's not so hard. I will probably just wait for an update, or find the newer aboot sources, or figure our how to remove the offending PT_NOTE entries if necessary. I already boot without using an initrd image, so that's all kinda taken care of already I guess :). ...tom On Thu, 16 Aug 2007 10:41:54 -0400, Jay Estabrook [EMAIL PROTECTED] wrote: Tom Evans wrote: I was trying to boot 2.6.23-rc1 a few weeks ago. Asked about the newer binutils - thought that may have been the reason it failing. Glad you hit the kml instead and found an answer (and glad I wasn't the only person seeing a problem with the 2.6.23-rcX series). It wasn't a binutils issue, but that the stock aboot is misguided. I'd also be willing to test a patched aboot. If you can't wait for that, you can always use BOOTP, which bypasses aboot entirely... :-) Just: make boot make bootpzfile scp arch/alpha/boot/bootpzfile someplace:/tftpboot However, your config prolly would need to change, to build in most of the stuff that's now loaded from INITRD, although that can be accomplished as well via some trickery... ;-} Good luck. --Jay++ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Thomas Evans wrote: i don't need any of the 2.6.23 functionality - I was just really trying the 2.6.23 kernel to make sure nothing major was broken with it. Seems OK on the few machines I've tried: LX164, ES45, AS1200. I'm interested in the CFS stuff, see if it makes any difference in the feel of the desktop... I *could* use bootp I suppose, but that would require a tftp service someplace - suppose that's not so hard. Well, there's no longer a simple BOOTP server, it's all been rolled into the DHCP daemon, but it's not too hard to do, though it is complicated if there's another DHCP server on the network... :-\ I will probably just wait for an update, or find the newer aboot sources, or figure our how to remove the offending PT_NOTE entries if necessary. I'd look at the newer sources: http://aboot.sourceforge.net/tarballs/ I think the version that RTH made the patches against is there... I already boot without using an initrd image, so that's all kinda taken care of already I guess :). Indeed... :-) --Jay++ ...tom On Thu, 16 Aug 2007 10:41:54 -0400, Jay Estabrook [EMAIL PROTECTED] wrote: Tom Evans wrote: I was trying to boot 2.6.23-rc1 a few weeks ago. Asked about the newer binutils - thought that may have been the reason it failing. Glad you hit the kml instead and found an answer (and glad I wasn't the only person seeing a problem with the 2.6.23-rcX series). It wasn't a binutils issue, but that the stock aboot is misguided. I'd also be willing to test a patched aboot. If you can't wait for that, you can always use BOOTP, which bypasses aboot entirely... :-) Just: make boot make bootpzfile scp arch/alpha/boot/bootpzfile someplace:/tftpboot However, your config prolly would need to change, to build in most of the stuff that's now loaded from INITRD, although that can be accomplished as well via some trickery... ;-} Good luck. --Jay++ -- Jay A EstabrookHPTC - XC I B Hewlett-Packard Company - ZKO1-3/D-B.8 (603) 884-0301 110 Spit Brook Road, Nashua NH 03062 [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Jay Estabrook wrote: If you can't wait for that, you can always use BOOTP, which bypasses aboot entirely... :-) Unfortunately, that may be the only way to get past this problem in the near term :-(. Just for grins and giggles, I retrieved the source package for aboot-0.9b-3 and tried to build the current version, i.e., I wanted to verify I could build what I've got before messing with something as critical as a bootloader. To make a long story short, the build is broken, probably because the tool chain has changed significantly since the last time anyone looked at this program. The somewhat longer version of the story... Had to modify debian/rules to look for DEB_HOST_GNU_TYPE alpha-linux-gnu instead of alpha-linux just to get out of the starting blocks. The build is currently incredibly noisy: lots of warnings, most of which I have to assume are safe to ignore, but disconcerting nevertheless. The fatal error looks like this: (Basic environment is Debian Etch on a PWS 433au with gcc-4.1.2 (Debian 4.1.1-21), binutils 2.17.50.20070804, libc6.1_2.6-2, and linux-libc-dev_2.6.21-6). $ dpkg-buildpackage -rfakeroot (...) make[2]: Entering directory `/opt/downloads/aboot/aboot-0.9b/sdisklabel' gcc -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -g -I/opt/downloads/aboot/aboot-0.9b/include -I/usr/src/linux/include -fno-builtin-printf -D__KERNEL__ -mcpu=ev4 -Os -Wall -fno-builtin -Wcast-align -mno-fp-regs -g -O2 -I../include -Wall -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -g -I/opt/downloads/aboot/aboot-0.9b/include -I/usr/src/linux/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -g -I/opt/downloads/aboot/aboot-0.9b/include -I/usr/src/linux/include -c -o swriteboot.o swriteboot.c In file included from /usr/src/linux/include/linux/compiler-gcc4.h:4, from /usr/src/linux/include/linux/compiler.h:40, from /usr/src/linux/include/asm-generic/page.h:7, from /usr/src/linux/include/asm/page.h:97, from /usr/src/linux/include/asm/system.h:5, from swriteboot.c:9: /usr/src/linux/include/linux/compiler-gcc.h:41:1: warning: __attribute_pure__ redefined In file included from /usr/include/features.h:322, from /usr/include/stdio.h:28, from swriteboot.c:1: /usr/include/sys/cdefs.h:206:1: warning: this is the location of the previous definition In file included from /usr/src/linux/include/linux/compiler.h:40, from /usr/src/linux/include/asm-generic/page.h:7, from /usr/src/linux/include/asm/page.h:97, from /usr/src/linux/include/asm/system.h:5, from swriteboot.c:9: /usr/src/linux/include/linux/compiler-gcc4.h:16:1: warning: __attribute_used__ redefined In file included from /usr/include/features.h:322, from /usr/include/stdio.h:28, from swriteboot.c:1: /usr/include/sys/cdefs.h:215:1: warning: this is the location of the previous definition In file included from /usr/src/linux/include/linux/compiler.h:40, from /usr/src/linux/include/asm-generic/page.h:7, from /usr/src/linux/include/asm/page.h:97, from /usr/src/linux/include/asm/system.h:5, from swriteboot.c:9: /usr/src/linux/include/linux/compiler-gcc4.h:19:1: warning: __always_inline redefined In file included from /usr/include/features.h:322, from /usr/include/stdio.h:28, from swriteboot.c:1: /usr/include/sys/cdefs.h:277:1: warning: this is the location of the previous definition In file included from /usr/src/linux/include/linux/kernel.h:16, from /usr/src/linux/include/asm/system.h:50, from swriteboot.c:9: /usr/src/linux/include/linux/log2.h:52: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'is_power_of_2' In file included from /usr/src/linux/include/asm/system.h:50, from swriteboot.c:9: /usr/src/linux/include/linux/kernel.h:172: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'printk_timed_ratelimit' /usr/src/linux/include/linux/kernel.h:223: error: expected declaration specifiers or '...' before 'bool' /usr/src/linux/include/linux/kernel.h:226: error: expected declaration specifiers or '...' before 'bool' swriteboot.c: In function 'read_configured_partition': swriteboot.c:20: warning: cast increases required alignment of target type swriteboot.c: In function 'main': swriteboot.c:206: warning: cast increases required alignment of target type make[2]: *** [swriteboot.o] Error 1 make[2]: Leaving directory `/opt/downloads/aboot/aboot-0.9b/sdisklabel' make[1]: *** [sdisklabel/swriteboot] Error 2 make[1]: Leaving directory `/opt/downloads/aboot/aboot-0.9b' make: *** [build-aboot-stamp] Error 2 -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Wed, Aug 15, 2007 at 10:17:33AM -0500, Bob Tracy wrote: The posted reply was: try http://lkml.org/lkml/2007/8/14/259 As it turns out, the problem was actually introduced with 2.6.23-rc1: 2.6.22 works fine. The l-k people say aboot needs a patch. Question to Steve L. et al: is a patch in the works? Not until now... Would you like a motivated tester? :-) Yes; when we have a fixed package I won't be in a position to test it right away, so it would be helpful if you would test. On Thu, Aug 16, 2007 at 01:52:10PM -0500, Bob Tracy wrote: Jay Estabrook wrote: If you can't wait for that, you can always use BOOTP, which bypasses aboot entirely... :-) Unfortunately, that may be the only way to get past this problem in the near term :-(. Just for grins and giggles, I retrieved the source package for aboot-0.9b-3 and tried to build the current version, i.e., I wanted to verify I could build what I've got before messing with something as critical as a bootloader. To make a long story short, the build is broken, probably because the tool chain has changed significantly since the last time anyone looked at this program. Yes, that's bug #437296. Try http://people.debian.org/~vorlon/aboot/ instead as a starting point. The package builds with the current sid toolchain; I haven't yet applied Richard's patch, but will look at doing that soon. FWIW, the issue with 2.6.23rc booting deserves its own bug in the BTS. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Bob Tracy wrote: Jay Estabrook wrote: If you can't wait for that, you can always use BOOTP, which bypasses aboot entirely... :-) Unfortunately, that may be the only way to get past this problem in the near term :-(. Just for grins and giggles, I retrieved the source package for aboot-0.9b-3 and tried to build the current version, i.e., I wanted to verify I could build what I've got before messing with something as critical as a bootloader. To make a long story short, the build is broken, probably because the tool chain has changed significantly since the last time anyone looked at this program. A spot of good news to report. Jay provided a pointer to the SourceForge page where the latest code lives. Retrieved that, applied the patch for 2.6.23+, changed #include linux/config.h to #include linux/autoconf.h in lib/isolib.c, and everything built properly. I'm not in front of the machine, but I'll give the new bootloader a try later today. The instructions for installing bootlx were pretty straightforward, and the code seems to have been written where I expected it to end up after running swriteboot. I reckon as long as the new code works at least as well as the old, I won't have to scramble for a CD to regain boot capability :-). -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
Steve Langasek wrote: FWIW, the issue with 2.6.23rc booting deserves its own bug in the BTS. Just added: see #438431. Feel free to reclassify the severity if I understated it. -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
The aboot patch in the referenced l-k posting did the trick. My Alpha is up and running on 2.6.23-rc3 as I type this... I'm still willing to test an official updated aboot package when it's available, but the value added at this point is probably minimal if the Debian source mods (for packaging) are trivial. -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
On Thu, Aug 16, 2007 at 10:50:06PM -0500, Bob Tracy wrote: The aboot patch in the referenced l-k posting did the trick. My Alpha is up and running on 2.6.23-rc3 as I type this... I'm still willing to test an official updated aboot package when it's available, but the value added at this point is probably minimal if the Debian source mods (for packaging) are trivial. Please have a look at the latest package available from http://people.debian.org/~vorlon/aboot/. This does need tested, because the patch posted by Richard is muddled with unrelated changes that aren't needed for building in Debian, so I'm not 100% sure that I've grabbed all the bits required for the fix. Thanks, -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
I originally posted to linux-kernel w.r.t. 2.6.23-rcX not booting: Unfortunately, I can't say where the bug was introduced, as this is the first kernel I've tried on my Alpha since 2.6.22-rc7. Best guess is somewhere between .23-rc1 and .23-rc2, based on changes to files in arch/alpha/boot in that patch set. The problem happens early: aboot starts to load vmlinux.gz, and I get an unzip: invalid exec header error. There's an earlier error from aboot I can't quote exactly, but the template from aboot (bootlx) is: aboot: Can't load kernel. Memory at %lx - %lx (chunk %i) is %s The first %lx is 0. The last %lx is all fs. The chunk number is 1, and I *think* the %s is busy. The posted reply was: try http://lkml.org/lkml/2007/8/14/259 As it turns out, the problem was actually introduced with 2.6.23-rc1: 2.6.22 works fine. The l-k people say aboot needs a patch. Question to Steve L. et al: is a patch in the works? Would you like a motivated tester? :-) The patch supplied at the above URL doesn't apply cleanly to the current Debian aboot-0.9b-3 source tree. -- --- Bob Tracy | Eagles may soar, but weasels don't get [EMAIL PROTECTED]| sucked into jet engines. --Anon --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [BUG] 2.6.23-rc3 won't boot on alpha (fwd)
I was trying to boot 2.6.23-rc1 a few weeks ago. Asked about the newer binutils - thought that may have been the reason it failing. Glad you hit the kml instead and found an answer (and glad I wasn't the only person seeing a problem with the 2.6.23-rcX series). It wasn't a binutils issue, but that the stock aboot is misguided. I'd also be willing to test a patched aboot. ...tom Bob Tracy wrote: I originally posted to linux-kernel w.r.t. 2.6.23-rcX not booting: Unfortunately, I can't say where the bug was introduced, as this is the first kernel I've tried on my Alpha since 2.6.22-rc7. Best guess is somewhere between .23-rc1 and .23-rc2, based on changes to files in arch/alpha/boot in that patch set. The problem happens early: aboot starts to load vmlinux.gz, and I get an unzip: invalid exec header error. There's an earlier error from aboot I can't quote exactly, but the template from aboot (bootlx) is: aboot: Can't load kernel. Memory at %lx - %lx (chunk %i) is %s The first %lx is 0. The last %lx is all fs. The chunk number is 1, and I *think* the %s is busy. The posted reply was: try http://lkml.org/lkml/2007/8/14/259 As it turns out, the problem was actually introduced with 2.6.23-rc1: 2.6.22 works fine. The l-k people say aboot needs a patch. Question to Steve L. et al: is a patch in the works? Would you like a motivated tester? :-) The patch supplied at the above URL doesn't apply cleanly to the current Debian aboot-0.9b-3 source tree. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]