Replace 'powerpc/powernv: Add calls to support little endian host'
with the version committed upstream.
Drop the ABI-preserving changes.
svn path=/dists/trunk/linux/; revision=21452
Create a more complete 'debian/config/ppc64el/defines' file.
The ppc64el (little endian) config has most options common with
the ppc64 (big endian) config.
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
svn path=/dists/sid/linux/; revision=21423
This config has options for little-endian PowerPC64 systems.
It shares most options with big-endian PowerPC64 systems.
The differences are:
- choice: Endianness selection
Build a little endian kernel.
- choice: Page size
64k pages have benefits (performance et al) over 4k pages on
IBM POWER processors.
The Debian ppc64el port primarily runs on this sort of hardware
and chances are it will also run on hardware based on it (i.e.,
OpenPOWER) [1] [2].
- Maximum number of CPUs
This was increased to 2048 (following pseries_le_defconfig).
For the currently announced systems, the number of CPUs range
between 80-192 (1 or 2 processor module(s) * 10 or 12 cores
per module * 8 threads per core) [3]. This is enough to have
to diverge from CONFIG_NR_CPUS=32 in the other powerpc ports.
For future systems, it's likely larger ones will be announced.
The rationale: consider the announced systems are classified
as 'scale-out' and 'entry-level', plus larger ones have been
historically made available for addressing other markets; and
notice the largest POWER7 server has 1024 CPUs (threads) [4],
and that the threads-per-core doubled from POWER7 to POWER8.
So, the 2048 value is a reasonable 'max' in that projection.
Certainly it is greater than what would be required for most
systems, but I belive 'max' makes sense in that case, if we
are not looking for kernel rebuild and/with different config
for the larger systems (although I would be ok with flavours).
- choice: Default CPUFreq governor
As other architectures, we would prefer the default cpufreq
governor to be 'ondemand'.
The currently available cpufreq driver is for the PowerNV
(non-virtualized) platform, where all processors are available.
In that scenario, statically running at the highest frequency
(specially on idle processors) is not very desireable for the
hardware around (servers), and it is not unlikely for future
hardware (possibly non-servers) to benefit too, considering
that energy savings have been increasingly important on most
environments.
(Note: the powernv-cpufreq driver was introduced only in 3.15;
so, this option has no effect in 3.14; it is harmless.
I can put in patches for enabling this on 3.14 soon.)
- Apple PowerMac based machines
This is being disabled temporarily, until a patch makes upstream
(restricting it to 'depends on !CPU_LITTLE_ENDIAN').
This hardware line has no (known) support for little endian
mode currently, and disabling it has the useful effect of also
disabling a lot of config options which 'depends on PPC_PMAC',
thus saving tens of lines from changing config files.
It indeed has to be disabled because it's enabled by default
('depends on BOOK3S', 'default y'), so even changing it from
config files would not be sufficient.
[1] 'OpenPOWER Foundation Unveils First Innovations and Roadmap'
http://openpowerfoundation.org/press-releases/openpower-foundation-unveils-first-innovations-and-roadmap/
[2] 'POWER8 Reference Board now available for Development!'
http://openpowerfoundation.org/technical/related-links/
[3] 'IBM Power System S812L and S822L'
http://www-03.ibm.com/systems/power/hardware/s812l-s822l/specs.html
[4] 'IBM Power 795 server'
http://www-03.ibm.com/systems/power/hardware/795/perfdata.html
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
svn path=/dists/sid/linux/; revision=21422
Instead of overriding the global default in kernelarch-arm disable virtio on
those armel flavours which do not want it (which is all but vexpress). This
allows the armmp flavours to pickup the global default.
No change to any of the eventual .config files.
svn path=/dists/sid/linux/; revision=21412
This just explicitly sets the default (except for some flavours
where it is forced off) so it does not change anything for mipsel.
svn path=/dists/trunk/linux/; revision=21373
- Revert the struct net_device lockdep changes
- Revert the sock_diag_put_filterinfo() parameter change
- Revert the removal from struct scsi_target and hide the compatible
type change from genksyms
- Hide the change to struct nf_ct_ext from genksyms and limit its
effect to modules that actually use it
- Ignore the vsock_core_init() change
svn path=/dists/sid/linux/; revision=21370
Compiler changes have resulted in a different kernel module ABI.
[powerpc] Bump ABI to 1a as 3.14.2-1 was built with an older compiler.
svn path=/dists/sid/linux/; revision=21327
aufs: Update to aufs3.x-rcN-20140421
[rt] Disable until it is updated for 3.15 or later
Refresh/drop patches as appropriate; in particular:
- update filenames in fbdev patches
- drop the ARM sunxi backports, which came from 3.15
svn path=/dists/trunk/linux/; revision=21296
sparc/config is shared with sparc64, so move it to
kernelarch-sparc/config where it will be automatically used for both
of them.
sparc/config.sparc64 and sparc/config.sparc64-smp are also shared, so
move them to kernelarch-sparc/config-up and
kernelarch-sparc/config-smp respectively.
svn path=/dists/trunk/linux/; revision=21267
s390x/config is shared with s390, so move it to kernelarch-s390/config
where it will be automatically used for both of them.
s390x/config.s390x is also shared, so move it to
kernelarch-s390/config-arch-64.
svn path=/dists/trunk/linux/; revision=21266
powerpc/config is shared by powerpc, powerpcspe and ppc64, so move
it to kernelarch-powerpc/config where it will be automatically used
for all of them.
powerpc/config.powerpc64 is also used by ppc64, so move it to
kernelarch-powerpc/config-arch-64.
powerpcspe has only one flavour, so rename its config file to config.
svn path=/dists/trunk/linux/; revision=21265
The filename of the kernel image to be installed, and the stem of the
installed name, varies between architectures, so we define several
different rules to install it for different sets of architectures.
However the basic fact that we need to install this file in /boot does
not.
We also duplicate this name information in gencontrol.py and in
debian/config/{armel,armhf,sh4}/defines (used by buildcheck.py).
To address this:
* Define [image]install-stem and [build]image-file for each architecture
* Copy these settings to make-flags in gencontrol.py
* Copy [image]install-stem to the image-stem template variable in
gencontrol.py
* Replace the per-architecture rules with a single rule using those
make-flags
The per-architecture rules for ARM and PowerPC also installed DTB
and DTS files, respectively. Include those commands in the single
rule with appropriate conditions around them.
svn path=/dists/trunk/linux/; revision=21253
The image-file path could potentially vary between flavours but
currently doesn't. buildcheck.py works either way.
svn path=/dists/trunk/linux/; revision=21251
These were disabled for armel in 3.2.1-1 due to size concerns, but
the armel config (now in kernelarch-arm) is shared by armhf. Move
the overrides into a new armel-specific config.
svn path=/dists/trunk/linux/; revision=21231
It is inconsistent and potentially surprising that armhf uses
armel/config as well as armhf/config. Move the common config into a
new kernelarch-arm directory.
While we're at it, remove some redundant lines from both files.
svn path=/dists/trunk/linux/; revision=21221
CONFIG_IP6_NF_MATCH_HL was replaced by CONFIG_NETFILTER_XT_MATCH_HL
(years ago!) but kept around for compatibility. Set the latter
directly.
Similarly for CONFIG_IP6_NF_TARGET_HL and CONFIG_NETFILTER_XT_TARGET_HL
(which we already explicitly set).
CONFIG_NETPRIO_CGROUP was renamed to CONFIG_CGROUP_NET_PRIO with
no such care for compatibility.
CONFIG_VIDEO_EM28XX was split with some code now dependent on
CONFIG_VIDEO_EM28XX_V4L2 which doesn't default to y.
svn path=/dists/trunk/linux/; revision=21129
These have been promoted out of staging so hopefully they're really
portable now.
Keep them disabled on armel/{ixp4xx,orion5x} due to size constraints.
svn path=/dists/trunk/linux/; revision=21127
Drop one patch that went upstream and refresh the aufs and rt patches.
Work around or ignore various ABI changes as appropriate (not yet
tested whether this covers everything).
svn path=/dists/sid/linux/; revision=20974
bugfix/all/firmware_class-return-specific-errors-from-file-read.patch
and
bugfix/all/firmware_class-log-every-success-and-failure.patch needed
some substantial changes as similar but incomplete fixes have been
made upstream.
svn path=/dists/trunk/linux/; revision=20912
Drop/refresh patches as appropriate. Disable aufs for now.
Two of our longstanding patches finally went upstream.
CONFIG_MAGIC_SYSRQ_DEFAULT_MASK was renamed in the process so adjust
config accordingly.
svn path=/dists/trunk/linux/; revision=20909
Include the resulting module in the sata-modules udeb.
Add one recommended upstream fix which is in libata.git for v3.13.
svn path=/dists/trunk/linux/; revision=20847
1. Reset or remove values for config symbols that were only
configurable if EXPERT.
2. Set additional configurable symbols to their default values.
This should have no effect on the actual configuration, except for
sh4 where EXPERT is *always* set and so the settings in 1 were
still being followed.
svn path=/dists/trunk/linux/; revision=20825
These drivers used to be conditional on USB_PHY (which we disabled)
but now they select it. Disable them all and let individual configs
(and other symbols) enable the drivers.
svn path=/dists/trunk/linux/; revision=20788
Restrict creation of user namespaces to root (CAP_SYS_ADMIN) by default
(sysctl: kern.unprivileged_userns_clone)
svn path=/dists/trunk/linux/; revision=20773
Leave it enabled for armel/{iop32x,ixp4xx,kirkwood,orion5x} and m68k.
Drop the workaround patch for powerpc.
svn path=/dists/sid/linux/; revision=20732
Apparently no m68k platform has configurable HZ, cpufreq or system
power management support. The corresponding Kconfig files are not
included from arch/m68k/Kconfig, and there is no point in setting
their symbols.
svn path=/dists/trunk/linux/; revision=20605
These symbols were either (1) removed entirely or (2) merged or
renamed, and we already configure the other symbol.
svn path=/dists/trunk/linux/; revision=20604
drivers/pci/pcie/Kconfig is only included for specific architectures.
These two correctly do not include it, so don't bother to override
it or claim that any configuration change was made.
svn path=/dists/trunk/linux/; revision=20603
All that config cleanup has brought them back under the size limit again
... for now ... with gcc-4.7. (I don't have a gcc-4.8 cross-compiler
to check with.)
svn path=/dists/trunk/linux/; revision=20588
Enable ARCH_OMAP3, ARCH_OMAP4 instead of the (now automatic) ARCH_OMAP2PLUS.
Enable MAILBOX, OMAP2PLUS_MBOX instead of the removed OMAP_MBOX_FWK.
svn path=/dists/trunk/linux/; revision=20580
No Alpha, PA-RISC or SH4 system supports PCI Express.
The older Marvell SoCs supported by iop32x and ixp4xx don't, but the
newer SoCs do. ARM Versatile doesn't support it and I'm pretty sure
QEMU won't let you add it, but will leave versatile alone for now.
Most supported MIPS platforms don't, but Octeon does.
I don't think PowerPC SPE systems have either PCI or PCI Express, but
I won't touch that configuration now.
svn path=/dists/trunk/linux/; revision=20579
efi-pstore is now a separate module (dependent on efivars). We still
want it to be auto-loaded to enable crash dumps on EFI systems, so add
another module alias.
While we're at it, explicitly set EFI_VARS_PSTORE=m matching what the
actual configuration will be.
svn path=/dists/sid/linux/; revision=20577
This driver doesn't bind to any device IDs, and instead has a comment
saying that the serial_cs and hci_uart drivers should be used instead.
So there's not much point in building it.
svn path=/dists/trunk/linux/; revision=20559
Disable most platform drivers, SPI and I2C drivers at the top level.
Platform drivers should be selected by architecture and flavour
configurations, and generally are. SPI and I2C devices aren't easily
detectable and their drivers aren't auto-loaded, so again they should
usually be selected in specific configuration files and probed
according to board code or FDTs.
As exceptions, I2C hwmon devices may be probed by lm-sensors and many
media tuners include I2C devices which are probed with the help of the
higher-level device driver. I've tried to be conservative and also
left I2C iio, input, leds and misc devices alone for now.
Disable the regulator subsystem at the top level as only some
architectures will need it.
Disable MTD_NAND_PLATFORM, PDA_POWER and FB_S1D13XXX on x86, as these
don't appear likely to be used on any x86 system that could run our
generic kernel images.
svn path=/dists/trunk/linux/; revision=20556
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