status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Mon, Jun 28, 2004 at 04:10:32PM -0700, William Lee Irwin III wrote: > On Mon, Jun 28, 2004 at 06:57:40PM -0400, Andres Salomon wrote: > > Package: kernel-image-2.6.6-i386 > > Version: 2.6.6-2 > > Severity: grave > > Token RC bug to keep this out of testing. I'm not sure if the initrd > > issues have been solved or not (ie, #256214). To be on the safe side, > > I'm filing this now. When I get home tonight, I'll take a closer look > > at my laptop (which managed to trigger a problem similar to #256214), > > and resolve/reassign this bug if it turns out to be something else. > > Sounds good. We should move to the 2.6.7 debs ASAP so this should keep > the thing out of people's hands until it's removed from the archives. Any news on the 2.6.7 debs ? It is again something like 2-3 weeks now since they where uploaded, isn't it ? Also, what is your opinion on going with 2.6 over 2.4 on certain arches by default ? I noticed nobody seemed to care about my mail on the subject. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Mon, Jun 28, 2004 at 04:10:32PM -0700, William Lee Irwin III wrote: >> Sounds good. We should move to the 2.6.7 debs ASAP so this should keep >> the thing out of people's hands until it's removed from the archives. On Tue, Jun 29, 2004 at 08:46:04AM +0200, Sven Luther wrote: > Any news on the 2.6.7 debs ? It is again something like 2-3 weeks now > since they where uploaded, isn't it ? > Also, what is your opinion on going with 2.6 over 2.4 on certain arches > by default ? I noticed nobody seemed to care about my mail on the > subject. > Friendly, > Sven Luther It would be great to move to 2.6 in as many areas as possible. It's the new stable kernel, and 2.4 is in a deep freeze, so we would make the best use of maintenance effort that way. I think the only exceptions are architecture ports not supported and/or working in 2.6 but that need 2.4 or earlier to function. But this can only be a recommendation since I only have the i386 and alpha packages. -- wli
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Mon, Jun 28, 2004 at 11:55:19PM -0700, William Lee Irwin III wrote: > On Mon, Jun 28, 2004 at 04:10:32PM -0700, William Lee Irwin III wrote: > >> Sounds good. We should move to the 2.6.7 debs ASAP so this should keep > >> the thing out of people's hands until it's removed from the archives. > > On Tue, Jun 29, 2004 at 08:46:04AM +0200, Sven Luther wrote: > > Any news on the 2.6.7 debs ? It is again something like 2-3 weeks now > > since they where uploaded, isn't it ? > > Also, what is your opinion on going with 2.6 over 2.4 on certain arches > > by default ? I noticed nobody seemed to care about my mail on the > > subject. > > Friendly, > > Sven Luther > > It would be great to move to 2.6 in as many areas as possible. It's the > new stable kernel, and 2.4 is in a deep freeze, so we would make the > best use of maintenance effort that way. I think the only exceptions are > architecture ports not supported and/or working in 2.6 but that need 2.4 > or earlier to function. But this can only be a recommendation since I > only have the i386 and alpha packages. I wish to do the same for powerpc, but when i mentioned it here sunday evening, nobody responded at all. It has also an influence on how we threat bugs present in 2.4 and not in 2.6, Christoph wanted to close them or something such, and we can do this only if we are actively going to move to 2.6. On the other hand, if we want to move to 2.6, we need to take the decision, and make an active search for problems and solving them, including investigate about userland incompatibilities and such. That said, we are paralysed by the lack of action of the ftp-masters again, which is a pain. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Tue, Jun 29, 2004 at 09:40:15AM +0200, Sven Luther wrote: > On Mon, Jun 28, 2004 at 11:55:19PM -0700, William Lee Irwin III wrote: > > On Mon, Jun 28, 2004 at 04:10:32PM -0700, William Lee Irwin III wrote: > > >> Sounds good. We should move to the 2.6.7 debs ASAP so this should keep > > >> the thing out of people's hands until it's removed from the archives. > > > > On Tue, Jun 29, 2004 at 08:46:04AM +0200, Sven Luther wrote: > > > Any news on the 2.6.7 debs ? It is again something like 2-3 weeks now > > > since they where uploaded, isn't it ? > > > Also, what is your opinion on going with 2.6 over 2.4 on certain arches > > > by default ? I noticed nobody seemed to care about my mail on the > > > subject. > > > Friendly, > > > Sven Luther > > > > It would be great to move to 2.6 in as many areas as possible. It's the > > new stable kernel, and 2.4 is in a deep freeze, so we would make the > > best use of maintenance effort that way. I think the only exceptions are > > architecture ports not supported and/or working in 2.6 but that need 2.4 > > or earlier to function. But this can only be a recommendation since I > > only have the i386 and alpha packages. > > I wish to do the same for powerpc, but when i mentioned it here sunday > evening, nobody responded at all. It has also an influence on how we > threat bugs present in 2.4 and not in 2.6, Christoph wanted to close > them or something such, and we can do this only if we are actively going > to move to 2.6. AFAIK, there's one or two arches that haven't even got to 2.4 yet. I'm pushing for all 2.2 kernel packages to get the boot from the archive, but I think m68k has issues with 2.4. I think we're a long way away from universal 2.6. regards Andrew
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 09:15:03AM +1000, Andrew Pollock wrote: > AFAIK, there's one or two arches that haven't even got to 2.4 yet. > I'm pushing for all 2.2 kernel packages to get the boot from the > archive, but I think m68k has issues with 2.4. I think we're a long > way away from universal 2.6. > regards > Andrew I'm acutely aware of this and have gone out of my way to acquire several such machines. Unfortunately some of them have developed hardware problems over time and I need to repeat that process. -- wli
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 09:15:03AM +1000, Andrew Pollock wrote: > On Tue, Jun 29, 2004 at 09:40:15AM +0200, Sven Luther wrote: > > On Mon, Jun 28, 2004 at 11:55:19PM -0700, William Lee Irwin III wrote: > > > On Mon, Jun 28, 2004 at 04:10:32PM -0700, William Lee Irwin III wrote: > > > >> Sounds good. We should move to the 2.6.7 debs ASAP so this should keep > > > >> the thing out of people's hands until it's removed from the archives. > > > > > > On Tue, Jun 29, 2004 at 08:46:04AM +0200, Sven Luther wrote: > > > > Any news on the 2.6.7 debs ? It is again something like 2-3 weeks now > > > > since they where uploaded, isn't it ? > > > > Also, what is your opinion on going with 2.6 over 2.4 on certain arches > > > > by default ? I noticed nobody seemed to care about my mail on the > > > > subject. > > > > Friendly, > > > > Sven Luther > > > > > > It would be great to move to 2.6 in as many areas as possible. It's the > > > new stable kernel, and 2.4 is in a deep freeze, so we would make the > > > best use of maintenance effort that way. I think the only exceptions are > > > architecture ports not supported and/or working in 2.6 but that need 2.4 > > > or earlier to function. But this can only be a recommendation since I > > > only have the i386 and alpha packages. > > > > I wish to do the same for powerpc, but when i mentioned it here sunday > > evening, nobody responded at all. It has also an influence on how we > > threat bugs present in 2.4 and not in 2.6, Christoph wanted to close > > them or something such, and we can do this only if we are actively going > > to move to 2.6. > > AFAIK, there's one or two arches that haven't even got to 2.4 yet. I'm > pushing for all 2.2 kernel packages to get the boot from the archive, but I > think m68k has issues with 2.4. I think we're a long way away from universal > 2.6. Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). I believe now is the time to take that decision, it may even be too late already, given the sarge release schedule, and provided the GR doesn't finish in some catastrophic result for the sarge release. (Still a bit pissed at the syntactic GR proponent who slyly passed this when nobody was noticing) Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 07:45:25AM +0200, Sven Luther wrote: > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > I believe now is the time to take that decision, it may even be too late > already, given the sarge release schedule, and provided the GR doesn't > finish in some catastrophic result for the sarge release. Well, AIUI, d-i should be able to on a per-arch basis default to a 2.6 kernel. So we can have sarge release with a 2.6 kernel by default on selected architectures. -boot may correct me. > (Still a bit pissed at the syntactic GR proponent who slyly passed this > when nobody was noticing) Yes well, what's done is done. Learn from it. regards Andrew
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 03:59:10PM +1000, Andrew Pollock wrote: > On Wed, Jun 30, 2004 at 07:45:25AM +0200, Sven Luther wrote: > > > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > > > I believe now is the time to take that decision, it may even be too late > > already, given the sarge release schedule, and provided the GR doesn't > > finish in some catastrophic result for the sarge release. > > Well, AIUI, d-i should be able to on a per-arch basis default to a 2.6 > kernel. So we can have sarge release with a 2.6 kernel by default on > selected architectures. -boot may correct me. Well, the important thing is not so much -boot, but compatbility with the rest of userland, as well as upgrrades from woody with a 2.2 or 2.4 kernel to sarge with a 2.6 kernel. For example, i know that the XF86Config-4 file needs to be changed when using a ps2 mouse, since it was /dev/psaux previously, and is /dev/input/mice now. Breaking X during the upgrade is hardly acceptable if we are going to make 2.6 the default. > > (Still a bit pissed at the syntactic GR proponent who slyly passed this > > when nobody was noticing) > > Yes well, what's done is done. Learn from it. Yeah, never let your guard done, even if you are away, and hardly connected, and just got out of a month/year long GR over the non-free issue, even if they claimed it was syntactical changes only. I should have learned from the first tentative of Branden to get the non-free removal clause in on the sly. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 08:19:07AM +0200, Sven Luther wrote: > On Wed, Jun 30, 2004 at 03:59:10PM +1000, Andrew Pollock wrote: > > On Wed, Jun 30, 2004 at 07:45:25AM +0200, Sven Luther wrote: > > > > > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > > > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > > > > > I believe now is the time to take that decision, it may even be too late > > > already, given the sarge release schedule, and provided the GR doesn't > > > finish in some catastrophic result for the sarge release. > > > > Well, AIUI, d-i should be able to on a per-arch basis default to a 2.6 > > kernel. So we can have sarge release with a 2.6 kernel by default on > > selected architectures. -boot may correct me. > > Well, the important thing is not so much -boot, but compatbility with > the rest of userland, as well as upgrrades from woody with a 2.2 or 2.4 > kernel to sarge with a 2.6 kernel. > > For example, i know that the XF86Config-4 file needs to be changed when > using a ps2 mouse, since it was /dev/psaux previously, and is > /dev/input/mice now. Breaking X during the upgrade is hardly acceptable > if we are going to make 2.6 the default. Hmm. dist-upgrading won't install a new kernel-image package, and a fresh install with 2.6 is going to get the mouse device right, so I don't think this particular example is a showstopper. Sure, if you've got a pre-existing install, with XFree86, and you choose to install a 2.6 kernel-image package, you're in for a bit of mouse pain... Andrew
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 09:43:40PM +1000, Andrew Pollock wrote: > On Wed, Jun 30, 2004 at 08:19:07AM +0200, Sven Luther wrote: > > On Wed, Jun 30, 2004 at 03:59:10PM +1000, Andrew Pollock wrote: > > > On Wed, Jun 30, 2004 at 07:45:25AM +0200, Sven Luther wrote: > > > > > > > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > > > > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > > > > > > > I believe now is the time to take that decision, it may even be too late > > > > already, given the sarge release schedule, and provided the GR doesn't > > > > finish in some catastrophic result for the sarge release. > > > > > > Well, AIUI, d-i should be able to on a per-arch basis default to a 2.6 > > > kernel. So we can have sarge release with a 2.6 kernel by default on > > > selected architectures. -boot may correct me. > > > > Well, the important thing is not so much -boot, but compatbility with > > the rest of userland, as well as upgrrades from woody with a 2.2 or 2.4 > > kernel to sarge with a 2.6 kernel. > > > > For example, i know that the XF86Config-4 file needs to be changed when > > using a ps2 mouse, since it was /dev/psaux previously, and is > > /dev/input/mice now. Breaking X during the upgrade is hardly acceptable > > if we are going to make 2.6 the default. > > Hmm. dist-upgrading won't install a new kernel-image package, and a fresh > install with 2.6 is going to get the mouse device right, so I don't think > this particular example is a showstopper. Sure, if you've got a pre-existing > install, with XFree86, and you choose to install a 2.6 kernel-image package, > you're in for a bit of mouse pain... No, it should just work. But i hear there is a psaux workaround in the 2.6 package by Herbert. Is this still there, or did it get removed by Christoph and William, and if so, why ? Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
Sven Luther wrote: [snip] > > AFAIK, there's one or two arches that haven't even got to 2.4 yet. I'm > > pushing for all 2.2 kernel packages to get the boot from the archive, but I > > think m68k has issues with 2.4. I think we're a long way away from universal > > 2.6. > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). It would kill floppy boot images for i386 (2.6 is too big for it). Thiemo
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 04:06:35PM +0200, Thiemo Seufer wrote: > Sven Luther wrote: > [snip] > > > AFAIK, there's one or two arches that haven't even got to 2.4 yet. I'm > > > pushing for all 2.2 kernel packages to get the boot from the archive, but > > > I > > > think m68k has issues with 2.4. I think we're a long way away from > > > universal > > > 2.6. > > > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > It would kill floppy boot images for i386 (2.6 is too big for it). Err, what is the problem in having 2.4 floppy debian-installer, and then install the 2.6 kernel on the installed system. We also have this problem on powerpc, and another solution may be a separate kernel flavour for 2.6 debian-installer floppies. Altough we are only 300-400KB away of fiting on a floppy with miboot. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
Sven Luther wrote: [snip] > > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > > > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > > > It would kill floppy boot images for i386 (2.6 is too big for it). > > Err, what is the problem in having 2.4 floppy debian-installer, and then > install the 2.6 kernel on the installed system. That's not a problem, apart from diverging device support for e.g. SATA. But d-i uses the normal debian kernel. If you eliminate 2.4 i386, then there's no such kernel left for d-i. > We also have this > problem on powerpc, and another solution may be a separate kernel > flavour for 2.6 debian-installer floppies. Altough we are only 300-400KB > away of fiting on a floppy with miboot. IOW, it doesn't work there as well. Everything not needed for boot should already be in modules anyway. Thiemo
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 05:14:30PM +0200, Thiemo Seufer wrote: > Sven Luther wrote: > [snip] > > > > Yeah, but what about 2.6 for powerpc and x86 (and maybe some other who > > > > is ready) and 2.4 for the rest of it (and 2.2 for some m68k subarches). > > > > > > It would kill floppy boot images for i386 (2.6 is too big for it). > > > > Err, what is the problem in having 2.4 floppy debian-installer, and then > > install the 2.6 kernel on the installed system. > > That's not a problem, apart from diverging device support for e.g. SATA. > But d-i uses the normal debian kernel. If you eliminate 2.4 i386, then > there's no such kernel left for d-i. I am not proposing anything so drastic, just that 2.6 be made the defaultm with 2.4 as fallback for those who need it. > > We also have this > > problem on powerpc, and another solution may be a separate kernel > > flavour for 2.6 debian-installer floppies. Altough we are only 300-400KB > > away of fiting on a floppy with miboot. > > IOW, it doesn't work there as well. Everything not needed for boot > should already be in modules anyway. Ah, this is indeed worse than on powerpc, where there is ample possibilities to remove stuff still, at least for oldworld who don't need a whole bunch of stuff which is actually builtin. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
Sven Luther wrote: [snip] > > > Err, what is the problem in having 2.4 floppy debian-installer, and then > > > install the 2.6 kernel on the installed system. > > > > That's not a problem, apart from diverging device support for e.g. SATA. > > But d-i uses the normal debian kernel. If you eliminate 2.4 i386, then > > there's no such kernel left for d-i. The general idea, however, is to use the same kernel for d-i and for the installed system. Rationale: An installer which dies immediately is better than one which installs half of the new system and then dies on reboot. > I am not proposing anything so drastic, just that 2.6 be made the > defaultm with 2.4 as fallback for those who need it. AFAIK starting the installer with "linux26" or from 2.6 images should already have this effect. Since 2.4 is still much better tested, it's probably a good idea to keep it that way (at least as long as we have some hope for releasing sarge in the next months). Thiemo
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 06:22:51PM +0200, Thiemo Seufer wrote: > Sven Luther wrote: > [snip] > > > > Err, what is the problem in having 2.4 floppy debian-installer, and then > > > > install the 2.6 kernel on the installed system. > > > > > > That's not a problem, apart from diverging device support for e.g. SATA. > > > But d-i uses the normal debian kernel. If you eliminate 2.4 i386, then > > > there's no such kernel left for d-i. > > The general idea, however, is to use the same kernel for d-i and for > the installed system. Rationale: An installer which dies immediately > is better than one which installs half of the new system and then > dies on reboot. Yeah sure, but remember this will only affect those boxes without netboot or CD/DVD drives. > > I am not proposing anything so drastic, just that 2.6 be made the > > defaultm with 2.4 as fallback for those who need it. > > AFAIK starting the installer with "linux26" or from 2.6 images should > already have this effect. Since 2.4 is still much better tested, it's > probably a good idea to keep it that way (at least as long as we have > some hope for releasing sarge in the next months). Well, not really, since we are speaking about floppies, but you are right. i was suposing installing with 2.6 by default and having linux24 just in case, but this is a decision that is for the x86 mfolk to make. For powerpc i believe the decision is much more easy. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 08:19:07AM +0200, Sven Luther wrote: > For example, i know that the XF86Config-4 file needs to be changed when > using a ps2 mouse, since it was /dev/psaux previously, and is > /dev/input/mice now. Breaking X during the upgrade is hardly acceptable > if we are going to make 2.6 the default. The /dev/psaux situation seems to be a big mess. I added /dev/psaux upstream long ago so I wouldn't have to change my XF86Config for 2.4 vs 2.6. In the meantime X can have two different input devices specified and won't fail if one of them doesn't work but gives strange results when both work. And Debian has both on the default XF86Config. Currently the Debian kernel has another config option to have /dev/psaux support in the kernel but disabled by default. I really hate that hack and would just remove /dev/psaux from the Debian kernel as apparently our X packages don't need that transition-aid.
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Wed, Jun 30, 2004 at 08:19:07AM +0200, Sven Luther wrote: >> For example, i know that the XF86Config-4 file needs to be changed when >> using a ps2 mouse, since it was /dev/psaux previously, and is >> /dev/input/mice now. Breaking X during the upgrade is hardly acceptable >> if we are going to make 2.6 the default. On Sat, Jul 03, 2004 at 07:17:08PM +0200, Christoph Hellwig wrote: > The /dev/psaux situation seems to be a big mess. I added /dev/psaux > upstream long ago so I wouldn't have to change my XF86Config for 2.4 > vs 2.6. In the meantime X can have two different input devices > specified and won't fail if one of them doesn't work but gives strange > results when both work. And Debian has both on the default XF86Config. > Currently the Debian kernel has another config option to have /dev/psaux > support in the kernel but disabled by default. I really hate that hack > and would just remove /dev/psaux from the Debian kernel as apparently > our X packages don't need that transition-aid. Unless we've got a counterexample to its superfluity dropping it sounds like the way to go. The weird thing is Sven's going on about device nodes' names/locations. Sven, what's setting up your device nodes? -- wli
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 07:17:08PM +0200, Christoph Hellwig wrote: > On Wed, Jun 30, 2004 at 08:19:07AM +0200, Sven Luther wrote: > > For example, i know that the XF86Config-4 file needs to be changed when > > using a ps2 mouse, since it was /dev/psaux previously, and is > > /dev/input/mice now. Breaking X during the upgrade is hardly acceptable > > if we are going to make 2.6 the default. > > The /dev/psaux situation seems to be a big mess. I added /dev/psaux > upstream long ago so I wouldn't have to change my XF86Config for 2.4 > vs 2.6. In the meantime X can have two different input devices > specified and won't fail if one of them doesn't work but gives strange > results when both work. And Debian has both on the default XF86Config. I don't think so, the current debian package 4.3.0 generated XF86Config has psaux as corepointer, and input/mice sending coreevents. If the usb mouse is missing, no big problem, but if the ps2 one is mising, X refuses to start. > Currently the Debian kernel has another config option to have /dev/psaux > support in the kernel but disabled by default. I really hate that hack > and would just remove /dev/psaux from the Debian kernel as apparently > our X packages don't need that transition-aid. Yep, they need. I would suggest a solution as follows : 1) the XF86Config file is debconf managed : We query the database, inform the user with a low priority debconf question if it is set to psaux, and change it for him. This means a priority normal install will automatically make the change, so no user intervention is needed. Problem 1 : if you want to run 2.4 in parallel you are screwed, a solution would be to have the input/mice being the core pointer always, and the psaux sending core events. Problem 2 : kernel-package's scripts need to be debconfified. They need to be that anyway though in order to break debian-installer less. 2) The XF86Config file is not debconf managed : We parse the file in the post inst to check if psaux is the core pointer, and inform the user to fix it himself. After all, he is managing the file by hand and should know how to do this. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:54:03AM -0700, William Lee Irwin III wrote: > On Wed, Jun 30, 2004 at 08:19:07AM +0200, Sven Luther wrote: > >> For example, i know that the XF86Config-4 file needs to be changed when > >> using a ps2 mouse, since it was /dev/psaux previously, and is > >> /dev/input/mice now. Breaking X during the upgrade is hardly acceptable > >> if we are going to make 2.6 the default. > > On Sat, Jul 03, 2004 at 07:17:08PM +0200, Christoph Hellwig wrote: > > The /dev/psaux situation seems to be a big mess. I added /dev/psaux > > upstream long ago so I wouldn't have to change my XF86Config for 2.4 > > vs 2.6. In the meantime X can have two different input devices > > specified and won't fail if one of them doesn't work but gives strange > > results when both work. And Debian has both on the default XF86Config. > > Currently the Debian kernel has another config option to have /dev/psaux > > support in the kernel but disabled by default. I really hate that hack > > and would just remove /dev/psaux from the Debian kernel as apparently > > our X packages don't need that transition-aid. > > Unless we've got a counterexample to its superfluity dropping it sounds > like the way to go. The weird thing is Sven's going on about device nodes' > names/locations. Sven, what's setting up your device nodes? Huh ? They have always been there, i suppose they are generated by the package providing them, or by MAKEDEV. There may be other solution, like udev or devfs, they are all controversial, and you have to start from the principle that any guy upgrading from woody to sarge will have the traditional way of setting those. Willian, could you enlighten us of the possible future standard that is emerging for future kernels ? As well as how they will fit into this, especially given the sarge release schedule which is again on track ? Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:54:03AM -0700, William Lee Irwin III wrote: >> Unless we've got a counterexample to its superfluity dropping it sounds >> like the way to go. The weird thing is Sven's going on about device nodes' >> names/locations. Sven, what's setting up your device nodes? On Sat, Jul 03, 2004 at 10:46:00PM +0200, Sven Luther wrote: > Huh ? They have always been there, i suppose they are generated by the > package providing them, or by MAKEDEV. There may be other solution, like > udev or devfs, they are all controversial, and you have to start from > the principle that any guy upgrading from woody to sarge will have the > traditional way of setting those. > Willian, could you enlighten us of the possible future standard that is > emerging for future kernels ? As well as how they will fit into this, > especially given the sarge release schedule which is again on track ? There's no agenda behind this; I merely suspected it being a bad interaction with whatever's being used to manage device nodes, since from the above, it doesn't appear to be the kernel itself having trouble. -- wli
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 01:52:51PM -0700, William Lee Irwin III wrote: > On Sat, Jul 03, 2004 at 11:54:03AM -0700, William Lee Irwin III wrote: > >> Unless we've got a counterexample to its superfluity dropping it sounds > >> like the way to go. The weird thing is Sven's going on about device nodes' > >> names/locations. Sven, what's setting up your device nodes? > > On Sat, Jul 03, 2004 at 10:46:00PM +0200, Sven Luther wrote: > > Huh ? They have always been there, i suppose they are generated by the > > package providing them, or by MAKEDEV. There may be other solution, like > > udev or devfs, they are all controversial, and you have to start from > > the principle that any guy upgrading from woody to sarge will have the > > traditional way of setting those. > > Willian, could you enlighten us of the possible future standard that is > > emerging for future kernels ? As well as how they will fit into this, > > especially given the sarge release schedule which is again on track ? > > There's no agenda behind this; I merely suspected it being a bad > interaction with whatever's being used to manage device nodes, since > from the above, it doesn't appear to be the kernel itself having trouble. I don't understand you, or maybe you don't understand me. The real problem is with X, X has one and one only core pointer, which is set to psaux on most systems since most users have a ps2 mouse. IF the ps2 device is missing, then X refuses to start. This could maybe be solved by a psaux->input/mice link or something, maybe. X being the main app though, we have to care about it. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 01:52:51PM -0700, William Lee Irwin III wrote: >> There's no agenda behind this; I merely suspected it being a bad >> interaction with whatever's being used to manage device nodes, since >> from the above, it doesn't appear to be the kernel itself having trouble. On Sat, Jul 03, 2004 at 11:04:51PM +0200, Sven Luther wrote: > I don't understand you, or maybe you don't understand me. The real > problem is with X, X has one and one only core pointer, which is set to > psaux on most systems since most users have a ps2 mouse. IF the ps2 > device is missing, then X refuses to start. This could maybe be solved > by a psaux->input/mice link or something, maybe. > X being the main app though, we have to care about it. The symlink sounds like a good idea to me. I'm not sure why X is dead set on a given device node, but so be it. Who owns MKDEV? -- wli
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 10:43:20PM +0200, Sven Luther wrote: > > The /dev/psaux situation seems to be a big mess. I added /dev/psaux > > upstream long ago so I wouldn't have to change my XF86Config for 2.4 > > vs 2.6. In the meantime X can have two different input devices > > specified and won't fail if one of them doesn't work but gives strange > > results when both work. And Debian has both on the default XF86Config. > > I don't think so, the current debian package 4.3.0 generated XF86Config > has psaux as corepointer, and input/mice sending coreevents. If the usb > mouse is missing, no big problem, but if the ps2 one is mising, X > refuses to start. Aiih, crap. Okay, the situation is more complex than I thought. > > Currently the Debian kernel has another config option to have /dev/psaux > > support in the kernel but disabled by default. I really hate that hack > > and would just remove /dev/psaux from the Debian kernel as apparently > > our X packages don't need that transition-aid. > > Yep, they need. I would suggest a solution as follows : > > 1) the XF86Config file is debconf managed : We query the database, inform > the user with a low priority debconf question if it is set to psaux, > and change it for him. This means a priority normal install will > automatically make the change, so no user intervention is needed. > Problem 1 : if you want to run 2.4 in parallel you are screwed, a > solution would be to have the input/mice being the core pointer > always, and the psaux sending core events. > Problem 2 : kernel-package's scripts need to be debconfified. They > need to be that anyway though in order to break debian-installer less. > > 2) The XF86Config file is not debconf managed : We parse the file in the > post inst to check if psaux is the core pointer, and inform the user > to fix it himself. After all, he is managing the file by hand and > should know how to do this. I don't quite like either of those. I'll try to take a look at the X Code and/or talk to the X folks on whether we could fix X to autoprobe for a corepointer.
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 10:46:00PM +0200, Sven Luther wrote: > Willian, could you enlighten us of the possible future standard that is > emerging for future kernels ? As well as how they will fit into this, > especially given the sarge release schedule which is again on track ? What exactly does back on track mean? Any pointer that explains this a little more?
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 02:10:52PM -0700, William Lee Irwin III wrote: > On Sat, Jul 03, 2004 at 01:52:51PM -0700, William Lee Irwin III wrote: > >> There's no agenda behind this; I merely suspected it being a bad > >> interaction with whatever's being used to manage device nodes, since > >> from the above, it doesn't appear to be the kernel itself having trouble. > > On Sat, Jul 03, 2004 at 11:04:51PM +0200, Sven Luther wrote: > > I don't understand you, or maybe you don't understand me. The real > > problem is with X, X has one and one only core pointer, which is set to > > psaux on most systems since most users have a ps2 mouse. IF the ps2 > > device is missing, then X refuses to start. This could maybe be solved > > by a psaux->input/mice link or something, maybe. > > X being the main app though, we have to care about it. > > The symlink sounds like a good idea to me. I'm not sure why X is dead > set on a given device node, but so be it. Who owns MKDEV? dpkg -S /etc/MAKEDEV should reply to this, and well, X is X. I have cross posted the other post to debian-x, let's see what they have to say about this, and maybe reply to that mail for further discussion. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:16:43PM +0200, Christoph Hellwig wrote: > On Sat, Jul 03, 2004 at 10:46:00PM +0200, Sven Luther wrote: > > Willian, could you enlighten us of the possible future standard that is > > emerging for future kernels ? As well as how they will fit into this, > > especially given the sarge release schedule which is again on track ? > > What exactly does back on track mean? Any pointer that explains this > a little more? Well, there was a syntactic social contract GR which slyly passed the more important decision that all stuff in debian is now software, including documentation and data. This meant that all the GFDL software was now considered non-free, and should be removed from debian before the saarge release. The Release Manager claimed that this would take a long time, and that he could not iun concience make the release in the time he was planning with this. There was a new vote, which ended on july 2, in which we decided to waive the requirement of dropping all GNU documetnation until after the sarge release, so it is again possible to release sarge on the old schedule (or lack thereof), which also means that the sarge release will heppen rather sooner than later, hopefully this summer. For more info, read the debian-vote mailing lists archive. And welcome to debian politics :) Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:15:11PM +0200, Christoph Hellwig wrote: > On Sat, Jul 03, 2004 at 10:43:20PM +0200, Sven Luther wrote: > > > The /dev/psaux situation seems to be a big mess. I added /dev/psaux > > > upstream long ago so I wouldn't have to change my XF86Config for 2.4 > > > vs 2.6. In the meantime X can have two different input devices > > > specified and won't fail if one of them doesn't work but gives strange > > > results when both work. And Debian has both on the default XF86Config. > > > > I don't think so, the current debian package 4.3.0 generated XF86Config > > has psaux as corepointer, and input/mice sending coreevents. If the usb > > mouse is missing, no big problem, but if the ps2 one is mising, X > > refuses to start. > > Aiih, crap. Okay, the situation is more complex than I thought. > > > > Currently the Debian kernel has another config option to have /dev/psaux > > > support in the kernel but disabled by default. I really hate that hack > > > and would just remove /dev/psaux from the Debian kernel as apparently > > > our X packages don't need that transition-aid. > > > > Yep, they need. I would suggest a solution as follows : > > > > 1) the XF86Config file is debconf managed : We query the database, inform > > the user with a low priority debconf question if it is set to psaux, > > and change it for him. This means a priority normal install will > > automatically make the change, so no user intervention is needed. > > Problem 1 : if you want to run 2.4 in parallel you are screwed, a > > solution would be to have the input/mice being the core pointer > > always, and the psaux sending core events. > > Problem 2 : kernel-package's scripts need to be debconfified. They > > need to be that anyway though in order to break debian-installer less. > > > > 2) The XF86Config file is not debconf managed : We parse the file in the > > post inst to check if psaux is the core pointer, and inform the user > > to fix it himself. After all, he is managing the file by hand and > > should know how to do this. > > I don't quite like either of those. I'll try to take a look at the > X Code and/or talk to the X folks on whether we could fix X to autoprobe > for a corepointer. I would rather fix it in the installed XF86Config file, so we have to deal only one time with it, and be clear therefor after. Notice that the above where not two alternatives, but you had to do both. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:28:39PM +0200, Sven Luther wrote: > Well, there was a syntactic social contract GR which slyly passed the > more important decision that all stuff in debian is now software, > including documentation and data. This meant that all the GFDL software > was now considered non-free, and should be removed from debian before > the saarge release. The Release Manager claimed that this would take a > long time, and that he could not iun concience make the release in the > time he was planning with this. > > There was a new vote, which ended on july 2, in which we decided to > waive the requirement of dropping all GNU documetnation until after the > sarge release, so it is again possible to release sarge on the old > schedule (or lack thereof), which also means that the sarge release will > heppen rather sooner than later, hopefully this summer. > > For more info, read the debian-vote mailing lists archive. > > And welcome to debian politics :) Hmm, I wonder what this means for the firmware blob issues..
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:29:47PM +0200, Sven Luther wrote: > > I don't quite like either of those. I'll try to take a look at the > > X Code and/or talk to the X folks on whether we could fix X to autoprobe > > for a corepointer. > > I would rather fix it in the installed XF86Config file, so we have to > deal only one time with it, and be clear therefor after. Notice that the > above where not two alternatives, but you had to do both. Well, there's people dual-booting into both 2.4 and 2.6 kernels. I do that myself, although on macs we've used /dev/input/mice in 2.4 already anyway.
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:27:21PM +0200, Christoph Hellwig wrote: > On Sat, Jul 03, 2004 at 11:28:39PM +0200, Sven Luther wrote: > > Well, there was a syntactic social contract GR which slyly passed the > > more important decision that all stuff in debian is now software, > > including documentation and data. This meant that all the GFDL software > > was now considered non-free, and should be removed from debian before > > the saarge release. The Release Manager claimed that this would take a > > long time, and that he could not iun concience make the release in the > > time he was planning with this. > > > > There was a new vote, which ended on july 2, in which we decided to > > waive the requirement of dropping all GNU documetnation until after the > > sarge release, so it is again possible to release sarge on the old > > schedule (or lack thereof), which also means that the sarge release will > > heppen rather sooner than later, hopefully this summer. > > > > For more info, read the debian-vote mailing lists archive. > > > > And welcome to debian politics :) > > Hmm, I wonder what this means for the firmware blob issues.. Firmware is software, they can't be considered as data or documentation by any stretch of the imagination, so they still need to go, i think. Friendly, Svne Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:28:34PM +0200, Christoph Hellwig wrote: > On Sat, Jul 03, 2004 at 11:29:47PM +0200, Sven Luther wrote: > > > I don't quite like either of those. I'll try to take a look at the > > > X Code and/or talk to the X folks on whether we could fix X to autoprobe > > > for a corepointer. > > > > I would rather fix it in the installed XF86Config file, so we have to > > deal only one time with it, and be clear therefor after. Notice that the > > above where not two alternatives, but you had to do both. > > Well, there's people dual-booting into both 2.4 and 2.6 kernels. I do > that myself, although on macs we've used /dev/input/mice in 2.4 already > anyway. Becaue you had a usb keyboard. Also, this is one of the reason i want to kick 2.4 into oblivion on powerpc. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:36:02PM +0200, Sven Luther wrote: > > Well, there's people dual-booting into both 2.4 and 2.6 kernels. I do > > that myself, although on macs we've used /dev/input/mice in 2.4 already > > anyway. > > Becaue you had a usb keyboard. > > Also, this is one of the reason i want to kick 2.4 into oblivion on > powerpc. Actually it's a (fake-)adb keyboard. But adb already uses the generic input layer in 2.4 like usb, so it's the same reason.
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:34:37PM +0200, Christoph Hellwig wrote: > Actually it's a (fake-)adb keyboard. But adb already uses the generic > input layer in 2.4 like usb, so it's the same reason. s/keyboard/trackpad/ the former is true aswell, but completely irrelevant for this discussion ;-)
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
Christoph Hellwig wrote: [snip] > > For more info, read the debian-vote mailing lists archive. > > > > And welcome to debian politics :) > > Hmm, I wonder what this means for the firmware blob issues.. AFAICS every change is postponed until sarge releases, and then the whole circus starts again. Thiemo
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
Sven Luther wrote: [snip] > > > For more info, read the debian-vote mailing lists archive. > > > > > > And welcome to debian politics :) > > > > Hmm, I wonder what this means for the firmware blob issues.. > > Firmware is software, they can't be considered as data or documentation > by any stretch of the imagination, so they still need to go, i think. I can't see that. Properly licensed firmware was ok under the old release policy/SC interpretation, and the vote simply reverts to that one until sarge releases. Thiemo
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:45:24PM +0200, Thiemo Seufer wrote: > Sven Luther wrote: > [snip] > > > > For more info, read the debian-vote mailing lists archive. > > > > > > > > And welcome to debian politics :) > > > > > > Hmm, I wonder what this means for the firmware blob issues.. > > > > Firmware is software, they can't be considered as data or documentation > > by any stretch of the imagination, so they still need to go, i think. > > I can't see that. Properly licensed firmware was ok under the old > release policy/SC interpretation, and the vote simply reverts to > that one until sarge releases. Please tell me in how far you can consider firmware as something else than software ? It is a list of instructions or whatever, and is executed by a processor on the target device, is it not ? IF there is other reasons not to consider it as needing to be free, or if we are ready to make a compromise, then ok, but the firmware is not software clause should be really hard to defend, as opposed to RFC not being software, or fonts or whatever. Unless i misunderstand totally what firmware is. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, Jul 03, 2004 at 11:45:24PM +0200, Thiemo Seufer wrote: > Sven Luther wrote: > [snip] > > > > For more info, read the debian-vote mailing lists archive. > > > > And welcome to debian politics :) > > > Hmm, I wonder what this means for the firmware blob issues.. > > Firmware is software, they can't be considered as data or documentation > > by any stretch of the imagination, so they still need to go, i think. > I can't see that. Properly licensed firmware was ok under the old > release policy/SC interpretation, and the vote simply reverts to > that one until sarge releases. It is not that simple. The only thing that http://www.debian.org/vote/2004/vote_004 changed is that the DFSG is not applied to everything in Debian but only to "software" which allows us to ship non-free docs in main. The problem with the firmware is orthogonal to that it is the kernel's license - the kernel is GPL, therefore you need to provide Sourcecode for all parts (including the firmware) to be allowed to distribute it at all. Otherwise you are breaking the license. I do not like this but I am completely at loss to argue differently. cu andreas
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
* Andreas Metzler ([EMAIL PROTECTED]) [040704 10:55]: > The problem with the firmware is orthogonal to that it is the kernel's > license - the kernel is GPL, therefore you need to provide Sourcecode > for all parts (including the firmware) to be allowed to distribute it at > all. Otherwise you are breaking the license. Of course, this is an orthogonal issue. However, there are opinions that the binary blobs are just aggregation with the real kernel. If this is true, than it's no problem for us. Of course, IANAL, and till someone convinces the RM otherwise, binary blobs are not allowed directly in the kernel, including not in modules.[1] Cheers, Andi [1] http://lists.debian.org/debian-release/2004/04/msg00060.html -- http://home.arcor.de/andreas-barth/ PGP 1024/89FB5CE5 DC F1 85 6D A6 45 9C 0F 3B BE F1 D0 C5 D1 D9 0C
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Mon, Jul 05, 2004 at 12:19:19PM +0200, Michel Dänzer wrote: > On Sat, 2004-07-03 at 22:43 +0200, Sven Luther wrote: > > > > I would suggest a solution as follows : > > > > 1) the XF86Config file is debconf managed : We query the database, inform > > the user with a low priority debconf question if it is set to psaux, > > and change it for him. This means a priority normal install will > > automatically make the change, so no user intervention is needed. > > Problem 1 : if you want to run 2.4 in parallel you are screwed, a > > solution would be to have the input/mice being the core pointer > > always, and the psaux sending core events. > > Keep in mind that using both /dev/psaux and /dev/input/mice at the same > time with a 2.6 kernel will result in duplicate mouse events. Err, only if you enable /dev/psaux. in the actual 2.6.7 powerpc kernels, /dev/psaux is not even existent, so there is no way you can have it send mouse events. Which is also why the default XF86Config fails completely to launch the X server, since it is waiting for /dev/psaux core events. Friendly, Sven Luther
Re: status of 2.6.7 ? (Was Re: Bug#256763: kernel-image-2.6.6-i386: not ready for sarge just yet)
On Sat, 2004-07-03 at 22:43 +0200, Sven Luther wrote: > > I would suggest a solution as follows : > > 1) the XF86Config file is debconf managed : We query the database, inform > the user with a low priority debconf question if it is set to psaux, > and change it for him. This means a priority normal install will > automatically make the change, so no user intervention is needed. > Problem 1 : if you want to run 2.4 in parallel you are screwed, a > solution would be to have the input/mice being the core pointer > always, and the psaux sending core events. Keep in mind that using both /dev/psaux and /dev/input/mice at the same time with a 2.6 kernel will result in duplicate mouse events. -- Earthling Michel DÃnzer | Debian (powerpc), X and DRI developer Libre software enthusiast| http://svcs.affero.net/rm.php?r=daenzer