Description of problem: Intalling the Alpha1 iso (DVD), both my keyboard and my mouse are dead: the install comes to the first screen ("Please choose a language to use"), but it is impossible to select a list item (keyboard arrows or mouse), nor to hit any button. Already before the language screen comes up, all the ALT-Fx buttons are insensitive. (Dell Optiplex 990, USB 3-button mouse, USB keyboard) Version-Release number of selected component (if applicable): Mageia 3 Alpha 1 How reproducible: For me 100%. Another tester reports that his keyboard worked, although the mouse did not work. But he installed on a laptop with an internal keyboard. Which would be consistent with the problem being with USB input devices Steps to Reproduce: 1. Launch install from the iso DVD 2. Wait until the language selection screen comes up 3. Try to select a language and hit a button.
Summary: Input devices are dead at begin of iso install => Input devices are dead at begin of iso install (Mageia 3 Alpha 1)
Most likely a missing hid module. Anyone with this problem, in mga 2 run "lspcidrake>lspcidrake.txt", then attach the output to this bug report, stating whether it's the mouse, keyboard, or both, that isn't working.
CC: (none) => davidwhodgins
It could be related to /proc/bus/usb no longer available ? see this thread today on mageia-dev ML: [Mageia-dev] dropping /proc/bus/usb https://www.mageia.org/pipermail/mageia-dev/2012-August/018351.html
CC: (none) => lmenut, mageia, mageia, thierry.vignaud
Very unlikely to be related to /proc/bus/usb I think. It's be deprecated for a long time and I doubt any input device stuff relies on it... I could be wrong tho' - the installer may make some strange use of it. So not 100% sure what's up here :s
(In reply to comment #0) > Another tester reports that his keyboard worked, although the > mouse did not work. But he installed on a laptop with an internal keyboard. > Which would be consistent with the problem being with USB input devices > correctly, that's me. I noticed that my mouse was bound to the xhci-controller instead of ehci. might that be the reason for?
CC: (none) => alfred.kretschmer
Depends on: (none) => 7130Target Milestone: --- => Mageia 3
Yeah this is likely related. I guess the boot iso doesn't include USB3 drivers properly (I'm not entirely sure how accurate this is as I know they were included on live CDs to allow booting from USB3 Flash Drives - but perhaps it does just mean something is missing for the installer in some capacity. TV will likely know better than me.
Created attachment 2694 [details] Output of lspcidrake (Mageia 2) Both keyboard and mouse do not work when trying to install Alpha1
Kernel-3.5 drop support for /proc/bus/usb which breaks ldetect Can you try boot.iso directly instead? eg from http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/distrib/cauldron/x86_64/install/images/ Does keyboard works with it? If no that means as ldetect cannot see USB modules, we may not load some modules.
CC: (none) => tmb
Created attachment 2695 [details] adapt ldetect to kernel-3.5+ This patch works on FC after running the following: mount -t debugfs none /sys/kernel/debug Can someone try it on Mga? Just: - unpack ldetect sources (from either SVN or the SRPM) - apply the patch (patch -p0 /where/it/was/downloaded/7248.diff) - run "make run" This will need we patch stage1 in order to mount it.
Status: NEW => ASSIGNEDAssignee: bugsquad => thierry.vignaud
> Can you try boot.iso directly instead? (Comment 7) Burned a CD with boot.iso from the url you gave and booted from it: cannot select an item from the menu for selecting the installation method - the keyboard is dead here also
i'have found the same bug than you tv. I'll compile ldetect and back in few minutes
CC: (none) => yves
The rescue system on the dvd is also affected by this
You can test with a netinstall, using : http://mageia.r0b0t.fr/people/hurdman/linux/Mageia/distrib/cauldron/x86_64 seems i ever got the error...
I just see the : http://svnweb.mageia.org/soft/drakx/trunk/mdk-stage1/init.c?revision=5449& view=markup i retry and back
i have re try after recompile the init into all.rdz and after made a boot.iso. It seems i'have got ever the same problem. Perhaps i have made an error, or there's something else to patch ;) http://mageia.r0b0t.fr/people/hurdman/linux/Mageia/distrib/cauldron/x86_64/install/images/bootcd.iso
good for me , with the last boot.iso and mirror ( from 13:29 ) test with distrib coffee
Whiteboard: (none) => 3alpha1
Created attachment 2697 [details] more changes for stage2 I've commited this in my git tree (in smaller commits of course). More changes were needed for stage2 I'll commit them into SVN tonight Note that the code that tries to guess the country code from USB keyboards is still broken (anyway, from what I remember, most keyboards reports bogus country, so ....=
The bug of ldetect vs kernel-3.5+ has been fixed
Status: ASSIGNED => RESOLVEDResolution: (none) => FIXEDSource RPM: (none) => ldetect
Depends on: 7130 => (none)
Re-opening. This is still current in the installer of DVD's 2nd build. No keyboard or mouse to perform the installation.
Status: RESOLVED => REOPENEDResolution: FIXED => (none)
Your issue is different (here it was about ldetect). We have another issue about mouse not working during the installer. Please comment on bug #7130 instead. And say there what kind of keyboard you use (USB, ...)
Status: REOPENED => RESOLVEDResolution: (none) => FIXED
We missed new hid-generic driver...
Re comment 19: this bug is on input devices being dead, see the problem description. Are we trying to solve a problem or playing chess? This said, it is not even a secondary question whether the problem is solved as bugzilla (a) or (b).
And they were several issues as well as several bug reports. We extensively talk about the ldetect one here so please do not mix the issues. This makes bug report unreadable. What's more I do not see the interest of such a troll one hour after I found & fix (in git) the issue...
Fix verified with virtualbox and a real physical machine that needed hid-generic.ko
Sorry, I only realize now that this bug is a duplicate of bug #7130 - difficult to find by a search