Archaic wrote:
On Wed, Jun 29, 2005 at 03:39:46PM +0600, Alexander E. Patrakov wrote:
The book says:
<para>LFS Bootscripts make the assumption that both <quote>Support
for Host-side USB</quote> and <quote>USB device filesystem</quote>
have been compiled directly into the kernel, or that neither is
compiled at all.
In fact they also work if usbcore is mentioned in
/etc/sysconfig/modules.
This paragraph really bothers me anyway. Why exactly would the LFS
bootscripts have anything to do with usbfs?
Mounting it using /etc/fstab. If the usbcore module is not loaded /
compiled into the kernel, mountfs will fail. There was a suggestion to
trust /etc/hotplug/usb.rc to do that, but it fails for the following reason:
1) pci.rc loads uhci-hcd and ehci-hcd and then continues its work
2) The kernel emits _real_ USB hotplug events that are handled
asynchronously _before_ usb.rc has a chance to run. Thus, all
chmods/chowns/etc that may be present in userspace hotplug handlers
(such as those installed by SANE and gphoto2) fail because there is no
pseudofile in /proc/bus/usb to chmod/chown.
--
Alexander E. Patrakov
--
http://linuxfromscratch.org/mailman/listinfo/lfs-book
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
