The kernel isn't able to initialize the system clock from a hardware clock
whose driver is a module, and we don't have an initramfs mechanism to do
that.
svn path=/dists/sid/linux/; revision=21929
The kernel isn't able to initialize the system clock from a hardware clock
whose drive is a module, and there is no initramfs mechanism to do that.
svn path=/dists/sid/linux/; revision=21928
KVM_BOOK3S_64_HV, KVM_BOOK3S_64_PR and KVM_XICS, but keep KVM_BOOK3S_64
disabled for now as it would trigger an ABI change (Partially addresses:
#761656).
usb/gadget: Enable USB_NET2280 as module
svn path=/dists/sid/linux/; revision=21849
Several patches in 3.17 cc'd to stable, all of which look security-
relevant though I don't think they have CVE IDs.
svn path=/dists/sid/linux/; revision=21835
We currently FTBFS on mips64el:
cannot read /tmp/linux/linux-3.16.2/debian/installer/mips64el/modules/mips64el-sb1-bcm91250a/mips64el-sb1-bcm91250a
...
cannot read /tmp/linux/linux-3.16.2/debian/installer/mips64el/modules/mips64el-loongson-3/mips64el-loongson-3
...
find: `debian/speakup-modules-3.16-1-loongson-3-di': No such file or directory
speakup-modules-3.16-1-loongson-3-di will be empty
The 'cannot read' warnings are due to a bug in kernel-wedge
copy-modules (fixed in version 2.90) that affects module list
directories that are symlinked. Instead of getting a list of files
and symlinks in the directory, it only finds the directory-symlink.
The bug was generally harmless because copy-modules also gets a
list of inter-package dependencies (without relying on the first
list), and this includes all package names. Except that
speakup-modules has 'Depends: kernel-image!', and as there is
no package definition for 'kernel-image!' this dependency is
not included in the second list and copy-modules does not see
'speakup-modules'.
mips64el seems to be the only architecture which has a symlink to a
module list directory including speakup-modules.
Delete the '!'.
svn path=/dists/sid/linux/; revision=21834
Currently we set KW_CHECK_NONFATAL=y when building for experimental or
an unreleased package. In theory people would be checking for
kernel-wedge errors in the build logs. In practice this meant we
didn't see them until they appear in unstable where they are fatal.
svn path=/dists/sid/linux/; revision=21830
wheezy's udev init script requires /sys/kernel/uevent_helper as an
indicator of hotplug support. We can't remove it until after the
jessie release.
svn path=/dists/sid/linux/; revision=21829
phy-xgene is needed for ahci_xgene not the Ethernet driver, so place in
sata-modules for now. While there move ahci_xgene from the toplevel list into
the arm64 specific list.
svn path=/dists/sid/linux/; revision=21828
Ben Hutchings
Ian Campbell
Aurelien Jarno
Bastian Blank