Re: [FIX]: ultra45 boot failing...
Version: 2.6.26-14 13:52 bwh davem: btw, the Debian bug with PCI enumeration by X vs kernel is supposed to be fixed, but perhaps you could verify that? http://bugs.debian.org/514418 22:38 davem bwh: yes that debian bug is fixed -- Ben Hutchings The program is absolutely right; therefore, the computer must be wrong. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Sun, May 24, 2009 at 02:49:25PM -0700, David Miller wrote: From: Julien Cristau jcris...@debian.org Date: Sun, 24 May 2009 15:52:20 +0200 I plan to revert it for lenny r2, and if time permits I'll try to make the xserver-xorg package generate an xorg.conf with Driver set to fbdev instead.. Indeed, that's likely to work much better. AFAICS this bug can be closed? Can any Sparc user confirm? Lenny's xorg switched to fbdev in 1:7.3+19/20, current Xorg from unstable shouldn't expose the problem any longer. Lenny's kernel has been fixed in 2.6.26-14. Current version is fine we well. Cheers, Moritz -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Wed, Feb 25, 2009 at 13:41:08 +0100, Julien Cristau wrote: On Mon, 2009-02-09 at 12:49 +0100, Julien Cristau wrote: On Mon, 2009-02-09 at 01:21 -0800, David Miller wrote: No, I would have said that if time is tight at least we can use fbdev as the Xorg driver for PCI devices on sparc until we have a better fix for Xorg. We can probably do that for r1. OK so here is a tentative patch to fall back to the fbdev driver on sparc. I'd appreciate if people could test it against lenny's xorg-server, and/or suggest better ways to do this. That patch doesn't work, because X craps itself if both a pci and a fb driver are loaded. I plan to revert it for lenny r2, and if time permits I'll try to make the xserver-xorg package generate an xorg.conf with Driver set to fbdev instead.. Cheers, Julien -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
From: Julien Cristau jcris...@debian.org Date: Sun, 24 May 2009 15:52:20 +0200 I plan to revert it for lenny r2, and if time permits I'll try to make the xserver-xorg package generate an xorg.conf with Driver set to fbdev instead.. Indeed, that's likely to work much better. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Wed, 2009-02-25 at 13:41 +0100, Julien Cristau wrote: On Mon, 2009-02-09 at 12:49 +0100, Julien Cristau wrote: On Mon, 2009-02-09 at 01:21 -0800, David Miller wrote: No, I would have said that if time is tight at least we can use fbdev as the Xorg driver for PCI devices on sparc until we have a better fix for Xorg. We can probably do that for r1. OK so here is a tentative patch to fall back to the fbdev driver on sparc. I'd appreciate if people could test it against lenny's xorg-server, and/or suggest better ways to do this. I uploaded this to stable-proposed-updates, sparc binaries should appear after the next mirror push. There's still time to tweak things before the point release, so testing reports from sparc X users to 488...@bugs.debian.org are welcome (whether things work or not). Make sure you have xserver-xorg-video-fbdev installed. Cheers, Julien -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
From: Julien Cristau jcris...@debian.org Date: Wed, 25 Feb 2009 13:41:08 +0100 On Mon, 2009-02-09 at 12:49 +0100, Julien Cristau wrote: On Mon, 2009-02-09 at 01:21 -0800, David Miller wrote: No, I would have said that if time is tight at least we can use fbdev as the Xorg driver for PCI devices on sparc until we have a better fix for Xorg. We can probably do that for r1. OK so here is a tentative patch to fall back to the fbdev driver on sparc. I'd appreciate if people could test it against lenny's xorg-server, and/or suggest better ways to do this. This looks good to me, thanks Julien. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Mon, 2009-02-09 at 12:49 +0100, Julien Cristau wrote: On Mon, 2009-02-09 at 01:21 -0800, David Miller wrote: No, I would have said that if time is tight at least we can use fbdev as the Xorg driver for PCI devices on sparc until we have a better fix for Xorg. We can probably do that for r1. OK so here is a tentative patch to fall back to the fbdev driver on sparc. I'd appreciate if people could test it against lenny's xorg-server, and/or suggest better ways to do this. Cheers, Julien 0001-Add-an-fbdev-screen-as-a-fallback-on-sparc.patch Description: application/mbox
Re: [FIX]: ultra45 boot failing... = Netra T1 200 too?
On Fri, 2009-02-06 at 23:11 -0800, David Miller wrote: The patch: debian/patches/bugfix/sparc/arch_pci_hostcontroller_workaround.patch causes ultra45 (and other PCI-Express based workstations) to hard reset when the PCI bus is initially scanned by the kernel. Please revert this patch from the debian kernel in Lenny and anywhere else it appears. The code in that patch creating dummy PCI root host controller nodes is wrong and does nothing other than cause trouble. If it fixes some problem with the X server for PCI devices on sparc64 that problem needs to be fixed some other way. On my Netra T1 200 the lenny (net) boot installer stopped just after detecting network hardware. I then tried the etch one which worked fine for network, mirror and stuff and at least up to disk selection, the machine has no disk yet (BTW is full NFS install possible?). May be it's the same problem? Anyone with a successful install of etch or lenny on a Netra T1 200? ~ # lspci -nn 00:00.0 Host bridge [0600]: Sun Microsystems Computer Corp. Ultra IIe [108e:a001] 00:01.0 PCI bridge [0604]: Sun Microsystems Computer Corp. Simba Advanced PCI Bridge [108e:5000] (rev 13) 00:01.1 PCI bridge [0604]: Sun Microsystems Computer Corp. Simba Advanced PCI Bridge [108e:5000] (rev 13) 01:03.0 Non-VGA unclassified device []: ALi Corporation M7101 Power Management Controller [PMU] [10b9:7101] 01:05.1 Ethernet controller [0200]: Sun Microsystems Computer Corp. RIO GEM [108e:1101] (rev 01) 01:05.3 USB Controller [0c03]: Sun Microsystems Computer Corp. RIO USB [108e:1103] (rev 01) 01:07.0 ISA bridge [0601]: ALi Corporation M1533/M1535 PCI to ISA Bridge [Aladdin IV/V/V+] [10b9:1533] 01:0c.0 Bridge [0680]: Sun Microsystems Computer Corp. RIO EBUS [108e:1100] (rev 01) 01:0c.1 Ethernet controller [0200]: Sun Microsystems Computer Corp. RIO GEM [108e:1101] (rev 01) 01:0c.3 USB Controller [0c03]: Sun Microsystems Computer Corp. RIO USB [108e:1103] (rev 01) 01:0d.0 IDE interface [0101]: ALi Corporation M5229 IDE [10b9:5229] (rev c3) 02:08.0 SCSI storage controller [0100]: LSI Logic / Symbios Logic 53C896/897 [1000:000b] (rev 07) 02:08.1 SCSI storage controller [0100]: LSI Logic / Symbios Logic 53C896/897 [1000:000b] (rev 07) Thanks in advance, Laurent -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing... = Netra T1 200 too?
From: Laurent GUERBY laur...@guerby.net Date: Wed, 18 Feb 2009 01:38:05 +0100 May be it's the same problem? Anyone with a successful install of etch or lenny on a Netra T1 200? No, different problem. You won't even get to the real console, the machine will just reset itself before booting does anything meaningful. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Sun, Feb 08, 2009 at 04:58:08PM -0800, David Miller wrote: So you're saying that X working is more important than machines actually booting at all? These priorities are wrong. When N (where N 0) users complain about dead X, and 0 users complain about not being able to boot, the priorities are fairly clear... -- 2. That which causes joy or happiness. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
From: Josip Rodin j...@debbugs.entuzijast.net Date: Mon, 9 Feb 2009 09:20:39 +0100 On Sun, Feb 08, 2009 at 04:58:08PM -0800, David Miller wrote: So you're saying that X working is more important than machines actually booting at all? These priorities are wrong. When N (where N 0) users complain about dead X, and 0 users complain about not being able to boot, the priorities are fairly clear... If their machine won't even boot into the installer they are unlikely to even make a report. You've really got it backwards. Users who hit this problem don't even have a chance to run your distribution at all. Furthermore the point remains that you put a change into the kernel that I would never have advocated had you presented the bug to me. I would have suggested ways to fix the X server and even worked on the patch. But since nobody contacted me about this, a broken change went into the kernel instead. And on top of it, you're going to sit here and argue with me about whether totally-unworking x-not-working. :-/ -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Mon, Feb 09, 2009 at 12:28:18AM -0800, David Miller wrote: So you're saying that X working is more important than machines actually booting at all? These priorities are wrong. When N (where N 0) users complain about dead X, and 0 users complain about not being able to boot, the priorities are fairly clear... If their machine won't even boot into the installer they are unlikely to even make a report. Nobody (before you that is) reported an installer failure on these machines, so the situation is still clear from our point of view - it's certainly not perfect or even good, but the system as a whole depends on user input. Furthermore the point remains that you put a change into the kernel that I would never have advocated had you presented the bug to me. I would have suggested ways to fix the X server and even worked on the patch. But since nobody contacted me about this, a broken change went into the kernel instead. That is true, someone should have contacted you (sparclinux list at least) about that. But then, it would have been completely your prerogative to respond to that simply by saying - DTRT and go upgrade X, patching old X is a waste of my time, and I guess nobody wanted to risk hearing that answer? :) -- 2. That which causes joy or happiness. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
From: Josip Rodin j...@debbugs.entuzijast.net Date: Mon, 9 Feb 2009 09:45:27 +0100 But then, it would have been completely your prerogative to respond to that simply by saying - DTRT and go upgrade X, patching old X is a waste of my time, and I guess nobody wanted to risk hearing that answer? :) No, I would have said that if time is tight at least we can use fbdev as the Xorg driver for PCI devices on sparc until we have a better fix for Xorg. And meanwhile I would have done an: apt-get source xserver-xorg and worked on a patch for you guys. I'm running this on my workstation so I actually give a shit for crying out loud. Really, you're just being rediculious. There were a thousand better ways to have handled this. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Mon, 2009-02-09 at 01:21 -0800, David Miller wrote: No, I would have said that if time is tight at least we can use fbdev as the Xorg driver for PCI devices on sparc until we have a better fix for Xorg. We can probably do that for r1. Cheers, Julien -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Breaking X.Org on sparc (was: Bug#514418: [FIX]: ultra45 boot failing...)
On 2009-02-08 00:07, Luk Claes wrote: I think it's best to delay that to r1. Can someone please provide a text for the release notes to describe the problem, TIA? Hi all: Please send me a paragraph of text ASAP. (Or file a bug with the proposed text against release-notes.) -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Breaking X.Org on sparc (was: Bug#514418: [FIX]: ultra45 boot failing...)
On Sun, Feb 08, 2009 at 03:49:22PM +0100, W. Martin Borgert wrote: On 2009-02-08 00:07, Luk Claes wrote: I think it's best to delay that to r1. Can someone please provide a text for the release notes to describe the problem, TIA? Hi all: Please send me a paragraph of text ASAP. (Or file a bug with the proposed text against release-notes.) Something like this: Lenny kernel fails to boot on Sparc workstations with PCI-Express Due to an unfortunate interaction of a kernel fix with PCI-Express subsystem, Lenny default kernel will fail to boot on Sparc workstations with PCI-Express slots, like Ultra 25 and Ultra 45. As this problem has been discovered very late in the release cycle, we were unable to include a fix in the original Lenny release, but we'll do our best to eliminate the problem for the first point release. Best regards, -- Jurij Smakov ju...@wooyd.org Key: http://www.wooyd.org/pgpkey/ KeyID: C99E03CC -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Breaking X.Org on sparc (was: Bug#514418: [FIX]: ultra45 boot failing...)
On 2009-02-08 17:33, Jurij Smakov wrote: Something like this: Added. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
On Sat, Feb 07, 2009 at 02:51:05AM -0800, David Miller wrote: From: Jurij Smakov ju...@wooyd.org Date: Sat, 7 Feb 2009 10:45:31 + Thanks for pointing it out, David. Please contact me in the future if you guys want to add non-trivial sparc specific changes. That's how you can keep stuff like this from getting broken. You can't even install on these boxes because of this bug. It sounds very Ubuntu-esque that you can't get a fix like this into the tree due to timing issues. This is Debian for crying out loud :-) Unfortunately, we are literally few days away from the release, and reverting this patch would break X for everyone on sparc, so it is probably better to keep it that way for now. Release team has agreed, that a better solution would be to get the fix (which might take a while to develop) into the first point release. We would really appreciate your insight into the way to properly address this issue. To give a little bit of background, this patch was supposed to fix http://bugs.debian.org/500358. Bug trail contains all the gory details, but the crux of the problem (as I understand it) is the following: the commit [0] into the upstream git broke X on Debian/sparc [1]. The version of the X.org by that time was already frozen on all architectures in anticipation of the upcoming release, so rolling in a new version, compatible with the kernel change, was not really an option. The patch in question simply reverts this commit, restoring the kernel interface (for the lack of a better word) to the state, understood by X.org which we have available in Lenny. As I said, we would really appreciate any suggestions on how to keep the Lenny version of X working on sparc, yet allow the boxes with PCI-Express to boot. I'm also adding some people who have helped to come up with the current solution and the newly opened bug to CC. Thanks. [0] http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=c26d3c0138970778fabe114df99dffb34a04b1d7 [1] http://marc.info/?t=12124785781r=1w=2 -- Jurij Smakov ju...@wooyd.org Key: http://www.wooyd.org/pgpkey/ KeyID: C99E03CC -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
From: Jurij Smakov ju...@wooyd.org Date: Sun, 8 Feb 2009 23:05:21 + To give a little bit of background, this patch was supposed to fix http://bugs.debian.org/500358. Bug trail contains all the gory details, but the crux of the problem (as I understand it) is the following: the commit [0] into the upstream git broke X on Debian/sparc [1]. The version of the X.org by that time was already frozen on all architectures in anticipation of the upcoming release, so rolling in a new version, compatible with the kernel change, was not really an option. The patch in question simply reverts this commit, restoring the kernel interface (for the lack of a better word) to the state, understood by X.org which we have available in Lenny. So you're saying that X working is more important than machines actually booting at all? These priorities are wrong. That kernel change got reverted upstream for a reason, it breaks things. If X is broken (which it is), you have to fix X. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
Forwarding to debian-kernel for consideration. On Fri, Feb 06, 2009 at 11:11:44PM -0800, David Miller wrote: The patch: debian/patches/bugfix/sparc/arch_pci_hostcontroller_workaround.patch causes ultra45 (and other PCI-Express based workstations) to hard reset when the PCI bus is initially scanned by the kernel. Please revert this patch from the debian kernel in Lenny and anywhere else it appears. The code in that patch creating dummy PCI root host controller nodes is wrong and does nothing other than cause trouble. If it fixes some problem with the X server for PCI devices on sparc64 that problem needs to be fixed some other way. Thank you. I believe that we are just a few days from Lenny release, so uploading a new kernel and rebuilding debian-installer might not be an option at this point, unfortunately. The best we can probably do is include the fixed kernel in the point release, but that's up to kernel team to decide. Thanks for pointing it out, David. -- Jurij Smakov ju...@wooyd.org Key: http://www.wooyd.org/pgpkey/ KeyID: C99E03CC -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
From: Jurij Smakov ju...@wooyd.org Date: Sat, 7 Feb 2009 10:45:31 + Thanks for pointing it out, David. Please contact me in the future if you guys want to add non-trivial sparc specific changes. That's how you can keep stuff like this from getting broken. You can't even install on these boxes because of this bug. It sounds very Ubuntu-esque that you can't get a fix like this into the tree due to timing issues. This is Debian for crying out loud :-) -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [FIX]: ultra45 boot failing...
Package: linux-2.6 Version: 2.6.26-13 Severity: grave I consider this RC. We'll break X.org either for r0 or r1. debian-sparc: Please provide someone who wants to take over sparc architecture maintenance of the Linux packages. On Sat, Feb 07, 2009 at 10:45:31AM +, Jurij Smakov wrote: On Fri, Feb 06, 2009 at 11:11:44PM -0800, David Miller wrote: The patch: debian/patches/bugfix/sparc/arch_pci_hostcontroller_workaround.patch causes ultra45 (and other PCI-Express based workstations) to hard reset when the PCI bus is initially scanned by the kernel. Please revert this patch from the debian kernel in Lenny and anywhere else it appears. The code in that patch creating dummy PCI root host controller nodes is wrong and does nothing other than cause trouble. If it fixes some problem with the X server for PCI devices on sparc64 that problem needs to be fixed some other way. Thank you. I believe that we are just a few days from Lenny release, so uploading a new kernel and rebuilding debian-installer might not be an option at this point, unfortunately. The best we can probably do is include the fixed kernel in the point release, but that's up to kernel team to decide. Bastian -- A princess should not be afraid -- not with a brave knight to protect her. -- McCoy, Shore Leave, stardate 3025.3 signature.asc Description: Digital signature
Re: Breaking X.Org on sparc (was: Bug#514418: [FIX]: ultra45 boot failing...)
Bastian Blank wrote: Hi We introduced a workaround by reverting a removal in the kernel to allow the current X.Org on sparc work. However this patch is not supportable and breaks all newer machines quite badly. This means that we have to back it out again and the only question is if we will do that for r0 or r1. I think it's best to delay that to r1. Can someone please provide a text for the release notes to describe the problem, TIA? Cheers Luk - Forwarded message from Bastian Blank wa...@debian.org - Package: linux-2.6 Version: 2.6.26-13 Severity: grave I consider this RC. We'll break X.org either for r0 or r1. debian-sparc: Please provide someone who wants to take over sparc architecture maintenance of the Linux packages. On Sat, Feb 07, 2009 at 10:45:31AM +, Jurij Smakov wrote: On Fri, Feb 06, 2009 at 11:11:44PM -0800, David Miller wrote: The patch: debian/patches/bugfix/sparc/arch_pci_hostcontroller_workaround.patch causes ultra45 (and other PCI-Express based workstations) to hard reset when the PCI bus is initially scanned by the kernel. Please revert this patch from the debian kernel in Lenny and anywhere else it appears. The code in that patch creating dummy PCI root host controller nodes is wrong and does nothing other than cause trouble. If it fixes some problem with the X server for PCI devices on sparc64 that problem needs to be fixed some other way. Thank you. I believe that we are just a few days from Lenny release, so uploading a new kernel and rebuilding debian-installer might not be an option at this point, unfortunately. The best we can probably do is include the fixed kernel in the point release, but that's up to kernel team to decide. - End forwarded message - -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
[FIX]: ultra45 boot failing...
The patch: debian/patches/bugfix/sparc/arch_pci_hostcontroller_workaround.patch causes ultra45 (and other PCI-Express based workstations) to hard reset when the PCI bus is initially scanned by the kernel. Please revert this patch from the debian kernel in Lenny and anywhere else it appears. The code in that patch creating dummy PCI root host controller nodes is wrong and does nothing other than cause trouble. If it fixes some problem with the X server for PCI devices on sparc64 that problem needs to be fixed some other way. Thank you. -- To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
