Bug#292831: udev: udev prevents X from beeing started
Hamish Moffatt wrote: > I did an amd64 install last week from the (then) current install image I said it was fixed in sid (in particular, it's fixed in hw-detect 1.12); you've not given me enough information to narrow it down between one of perhaps 12 different images you could have used, and perhaps 5 ways you could have gotten a too old version. > I am yet to submit a report; my bad. Ok, looking forward to a proper report. -- see shy jo signature.asc Description: Digital signature
Bug#292831: udev: udev prevents X from beeing started
On Mon, Jan 31, 2005 at 03:46:50PM +1100, Hamish Moffatt wrote: > On Mon, Jan 31, 2005 at 05:31:03AM +0100, Joey Hess wrote: > > Marco d'Itri wrote: > > > My package works as designed, but let me know if you can design > > > something better. > > > > Oh, so it's udev that's responsible for what IIRC is a race that can > > cause X to not see the ps/2 mouse if the module is loaded as part of X's > > setup? Nice design. :-P > > > > FWIW, we have worked around this bug in d-i unstable for at least i386 > > and amd64 by always putting psmouse in /etc/modules. > > I did an amd64 install last week from the (then) current install image > and didn't end up with psmouse in /etc/modules; I added it by hand when > I found that udev was preventing X from starting. :-( I only commited it for amd64 to ddetect yesterday, and it's not uploaded yet. It already had it for i386 and ia64 (and some powerpc subarches). Kurt -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#292831: udev: udev prevents X from beeing started
Hi Marco, On Sunday, 30 Jan 2005, Marco d'Itri <[EMAIL PROTECTED]> wrote: > On Jan 30, Martin Zobel-Helas <[EMAIL PROTECTED]> wrote: > > > This breaks unrelated stuff here, so please fix it. > I'll wait for your patch. What about not enableing udev per default? Lets say, you would add a /etc/default/udev and add a UDEV_ENBALED switch, which is disabled per default. Add a remark about reading /usr/share/doc/udev/README.Debian before enableing UDEV_ENABLE. This would result in "the user did it himself" and not in "causes X problems by installation". Greetings Martin -- "If you ever want to get anywhere in politics, my boy, you're going to have to get a toehold in the public eye." signature.asc Description: Digital signature
Bug#292831: udev: udev prevents X from beeing started
On Jan 31, Joey Hess <[EMAIL PROTECTED]> wrote: > Oh, so it's udev that's responsible for what IIRC is a race that can > cause X to not see the ps/2 mouse if the module is loaded as part of X's > setup? Nice design. :-P This is the *upstream* design... The official word on this is that if you manually load a module then you need to wait for the device with a loop (which sucks if you do not know the device name, so you should use a dev.d script when possible). > FWIW, we have worked around this bug in d-i unstable for at least i386 > and amd64 by always putting psmouse in /etc/modules. Lack of serio hotplug support is a kernel bug, but I see that the needed bits are being merged. -- ciao, Marco signature.asc Description: Digital signature
Bug#292831: udev: udev prevents X from beeing started
On Jan 31, David Pashley <[EMAIL PROTECTED]> wrote: > Surely the solution is for hotplug/discover to load it during bootup. > Could hotplug use mdetect? No. The kernel should implement proper serio support (and maybe until it's done ship an aliases file which will make hotplug always load psmouse if a PS/2 port is present). -- ciao, Marco signature.asc Description: Digital signature
Bug#292831: udev: udev prevents X from beeing started
On Jan 31, 2005 at 04:46, Hamish Moffatt praised the llamas by saying: > On Mon, Jan 31, 2005 at 05:31:03AM +0100, Joey Hess wrote: > > Marco d'Itri wrote: > > > My package works as designed, but let me know if you can design > > > something better. > > > > Oh, so it's udev that's responsible for what IIRC is a race that can > > cause X to not see the ps/2 mouse if the module is loaded as part of X's > > setup? Nice design. :-P > > > > FWIW, we have worked around this bug in d-i unstable for at least i386 > > and amd64 by always putting psmouse in /etc/modules. > > I did an amd64 install last week from the (then) current install image > and didn't end up with psmouse in /etc/modules; I added it by hand when > I found that udev was preventing X from starting. :-( > > I am yet to submit a report; my bad. > Surely the solution is for hotplug/discover to load it during bootup. Could hotplug use mdetect? -- David Pashley [EMAIL PROTECTED] Nihil curo de ista tua stulta superstitione. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#292831: udev: udev prevents X from beeing started
On Mon, Jan 31, 2005 at 05:31:03AM +0100, Joey Hess wrote: > Marco d'Itri wrote: > > My package works as designed, but let me know if you can design > > something better. > > Oh, so it's udev that's responsible for what IIRC is a race that can > cause X to not see the ps/2 mouse if the module is loaded as part of X's > setup? Nice design. :-P > > FWIW, we have worked around this bug in d-i unstable for at least i386 > and amd64 by always putting psmouse in /etc/modules. I did an amd64 install last week from the (then) current install image and didn't end up with psmouse in /etc/modules; I added it by hand when I found that udev was preventing X from starting. :-( I am yet to submit a report; my bad. Hamish -- Hamish Moffatt VK3SB <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#292831: udev: udev prevents X from beeing started
Marco d'Itri wrote: > My package works as designed, but let me know if you can design > something better. Oh, so it's udev that's responsible for what IIRC is a race that can cause X to not see the ps/2 mouse if the module is loaded as part of X's setup? Nice design. :-P FWIW, we have worked around this bug in d-i unstable for at least i386 and amd64 by always putting psmouse in /etc/modules. -- see shy jo signature.asc Description: Digital signature
Bug#292831: udev: udev prevents X from beeing started
On Jan 30, Martin Zobel-Helas <[EMAIL PROTECTED]> wrote: > This breaks unrelated stuff here, so please fix it. I'll wait for your patch. > Every 0-8-15 user will do exactly the same as i did. It is your package > package breaking a perfectly running X here. My package works as designed, but let me know if you can design something better. -- ciao, Marco signature.asc Description: Digital signature
Bug#292831: udev: udev prevents X from beeing started
Hi Marco, On Sunday, 30 Jan 2005, Marco d'Itri <[EMAIL PROTECTED]> wrote: > On Jan 30, Martin Zobel-Helas <[EMAIL PROTECTED]> wrote: > > > installing udev results in X not able to start. > RTFM README.Debian. > As written, i installed sarge via linux26, than did a apt-get install x-window-system X runs perfect. Now i did apt-get install gnome which installs udev and will result in /dev/psaux being not present after a reboot. This breaks unrelated stuff here, so please fix it. Every 0-8-15 user will do exactly the same as i did. It is your package package breaking a perfectly running X here. Greetings Martin -- Renning's Maxim: Man is the highest animal. Man does the classifying. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#292831: udev: udev prevents X from beeing started
Package: udev Severity: serious Justification: Renders unrelated packages unusable Hi, installing udev results in X not able to start. I did a fresh install of my PC yesterday, using d-i2 (linux26). After that i upgraded to kernel-image-2.6.8-2 and installed and configured X. Rebooted, works. So i installed gnome and GDM. Starting GDM works. After rebooting, X is not beeing started complaining there is no CorePointer present. So i debugged and found, there is no /dev/psaux and /dev/input/mice present. Stopping udev via "/etc/init.d/udev stop" resolved the problem for me. As quite a lot of people will install X and gnome with kernel-2.6, this problem needs to be fixed before releasing sarge. Greetings. Martin -- I just know I'm a better manager when I have Joe DiMaggio in center field. -- Casey Stengel signature.asc Description: Digital signature