Description of problem: When start Fedora Core 3 installer and specifies the "nousbstorage" option, the full USB support is disabled, therefore FC3 cannot be installed to a machine which has only USB keyboard. The problem exist with FC4 as well. Version-Release number of selected component (if applicable): anaconda-10.1.0.2-1.x86_64 How reproducible: always Steps to Reproduce: 1. Get a machine with only USB keyboard 2. Start FC3 install with 'linux text nousbstorage' 3. Wait for anaconda to come up and load several drivers Actual results: The USB devices are disabled after the kernel boots Expected results: Only the USB storage devices should be disabled, not the whole USB support Additional info: Not only the 'nousbstorage' keyword disables all USB devices, but - I think - all ketwords begin with 'nousb'. I tried 'nousbasd' and 'nousbasdasd', both disabled the USB support. When used a PS/2 keyboard and specified any boot option begins with 'nousb', I experienced the following: From the shell the installer opens: # grep usb /proc/devices # grep usb /proc/filesystems # The /tmp/anaconda.log contains the following lines: ... failed to mount device usbfs: No such file or directory waiting for usb to become stable... ... I compiled the USB HCI drivers (ehci, ohci and uhci) to the kernel, it has the same problem: the full USB seems to be disabled. I added a new boot option to anaconda 'skipusbstorage' which does exactly the same as 'nousbstorage' in loader2/loader.c:parseCmdLineFlags(): sets the LOADER_FLAGS_NOUSBSTORAGE flag. After rebuilding anaconda packages and the distro and start the install with 'linux text skipusbstorage', everything is O.K., i.e. it has USB support and the USB storage module is not loaded. It seems that something before anaconda does a bad string match and disables the full USB support for all boot options begins with 'nousb'. Maybe the kernel.
Does this still occur with later releases? I'm not reproducing it in a quick test here.
I checked it with FC3 (dvd) and FC4 (cd) downloaded from an official Fedora mirror (wftp.tu-chemnitz.de). Did not check with the development tree. Used 2 machines to test. Always can reproduce on both configurations (USB keyboard and mouse): - HP xw8200 workstation - SUN Fire x4200
This is due to the kernel interpreting it and it still seems to happen with current (2.6.15-rc)
Adding Jeremy to keep him informed, though this is confirmed as kernel bug/feature/whatever.
Created attachment 122468 [details] Candidate #3 - replace __setup with __module_param_call This seems like that the only sensible solution. I definitely do not want to touch the implementation of __setup(), with all places doing __setup("foo=").
This is a mass-update to all currently open Fedora Core 3 kernel bugs. Fedora Core 3 support has transitioned to the Fedora Legacy project. Due to the limited resources of this project, typically only updates for new security issues are released. As this bug isn't security related, it has been migrated to a Fedora Core 4 bug. Please upgrade to this newer release, and test if this bug is still present there. This bug has been placed in NEEDINFO_REPORTER state. Due to the large volume of inactive bugs in bugzilla, if this bug is still in this state in two weeks time, it will be closed. Should this bug still be relevant after this period, the reporter can reopen the bug at any time. Any other users on the Cc: list of this bug can request that the bug be reopened by adding a comment to the bug. Thank you.
Still occurs when I tried a few weeks ago
The fix is in Linus's tree (2.6.15-git11), so I am waiting for Dave to import it in due time.
git11 was yesterdays build, so it should be there now for rawhide. FC4 will pick it up when it gets a .16 rebase.