USB_EHCI_ROOT_HUB_TT adds about 10 lines of code and is safe even if
not needed, so it's never worth turning off.
USB_EHCI_TT_NEWSCHED was 'new' in 2006 and is now the default, so
hardly anyone will be testing the 'old' code now.
svn path=/dists/trunk/linux/; revision=20546
These devices need platform data (even for SDIO) and apparently are
only used with ARM boards at the moment.
svn path=/dists/trunk/linux/; revision=20543
At this point, Debian users should know that they may need to install
firmware from non-free. People using e.g. the r8169 driver may quite
reasonably choose not to install the associated firmware, either
because the driver doesn't actually request it for their chip or
because the driver can still work without it.
One thing we lose by doing this is a reminder that a firmware package
might also need to be upgraded, as a driver requires a newer version
of the firmware that has a different name. As an alternative, we
could compare the firmware file lists for old and new modules and only
warn about newly listed files that are missing. However, that would
also result in incorrect warnings for e.g. r8169 users, as that driver
may request a different file for each of the many chips it supports.
svn path=/dists/trunk/linux/; revision=20511
We warned about removal of the 'ramdisk' configuration variable
on upgrade to wheezy, and don't need to warn about it again - at
least not this prominently.
svn path=/dists/trunk/linux/; revision=20510
Use the same template syntax and implementation for maintainer
scripts, translations, etc. as we do for the control files. Define
the image-stem and initramfs variables to replace the old K and I
variables.
After this, debian/linux-* and debian/po/*.po are generated files (at
source preparation time) and should be ignored in svn.
Use debhelper to install the generated files at build time. This also
results in a redundant dependency on debconf (which we already have in
Pre-Depends), but this seems harmless.
svn path=/dists/trunk/linux/; revision=20508
Maintainer scripts generated by kernel-package pass an environment
variable $KERNEL_ARCH to hook scripts. This is undocumented but seems
to be the architecture string that the kernel will report
(e.g. through 'uname -m').
However, for the past few years our maintainer scripts have set it to
the source architecture name instead. Since no-one reported this bug,
I don't think anyone depends on it. codesearch didn't find any users
either. So remove it.
svn path=/dists/trunk/linux/; revision=20507
From: Thorsten Glaser <tg@mirbsd.de>
- disable some functionality due to size constraints
- switch to using initrd
- drop all flavours; build a generic image that runs on
multiple subarchitectores (old Amiga, Atari, Macintosh tested;
old BVME6000, MVME147, MVME16x, new Apollo, Q40/Q60 untested)
- use gcc-4.8
- add patch to make Atari EtherNAT actually selectable
svn path=/dists/sid/linux/; revision=20413
Ben Hutchings