| Age | Commit message (Collapse) | Author |
|---|
|
Ensure the preempt-rt kernel has usb-mass-storage and boot-live support
in support of booting the hddimg images over USB.
Add the latencytop and profiling scc files to keep the default policy
consistent with the 'standard' kernel.
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
CC: Kishore Bodke <kishore.k.bodke@intel.com>
|
|
Ensure the preempt-rt kernel has usb-mass-storage and boot-live support
in support of booting the hddimg images over USB.
Add the latencytop and profiling scc files to keep the default policy
consistent with the 'standard' kernel.
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
CC: Kishore Bodke <kishore.k.bodke@intel.com>
|
|
Ensure the preempt-rt kernel has usb-mass-storage and boot-live support
in support of booting the hddimg images over USB.
Add the latencytop and profiling scc files to keep the default policy
consistent with the 'standard' kernel.
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
1/1 [
Author: Bruce Ashfield
Email: bruce.ashfield@windriver.com
Subject: preempt-rt: import 3.4.4-rt14
Date: Fri, 20 Jul 2012 12:53:28 -0400
Importing 3.4.4-rt14 from:
http://www.kernel.org/pub/linux/kernel/projects/rt/3.4/patch-3.4.4-rt14.patch
Only minor context adjustments were made.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
|
|
1/2 [
Author: Nitin A Kamble
Email: nitin.a.kamble@intel.com
Subject: yocto/emgd: emgd 1.14 driver
Date: Mon, 16 Jul 2012 07:14:38 -0700
The starting-point code that subsequent patches will modify. This is
a straight copy of the code in the emgd 1.14 emgd driver, specifically
IEMGD_HEAD_Linux/common/drm/emgd_drm.tgz from
LIN_IEMGD_1_14_GOLD_2443.tgz, the 'Linux Tar Ball' release downloaded
from http://edc.intel.com/Software/Downloads/EMGD/.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
]
2/2 [
Author: Nitin A Kamble
Email: nitin.a.kamble@intel.com
Subject: emgd/pvr: get it building with v3.4 kernel
Date: Mon, 16 Jul 2012 07:14:50 -0700
Add emgd config option (DRM_EGD) and modify Makefiles for in-tree
builds.
rebased to v3.4 code
Fix build errors caused by changes to kernel interfaces introduced by
linux 3.2 such as:
error: 'THIS_MODULE' undeclared here (not in a function)
due to a missing THIS_MODULE definition which can be found in linux/export.h.
It also fixes a ream of errors like:
error: expected ')' before string constant
due to missing MODULE_PARM_DESC definition which can be found in
linux/moduleparam.h, included by linux/module.h, needed also for
MODULE_AUTHOR, etc.
variable name "pitch" from drm_framebuffer structure has changed
to "pitches" array in v3.4 kernel soruces.
avoid errors like these:
emgd_fbcon.c:128:28: error: 'struct drm_framebuffer' has no member named 'pitch'
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
CONFIG_MTD_NAND_VERIFY_WRITE option: this works ok with JFFS2 fs but
leads to errors when using UBI.
http://www.linux-mtd.infradead.org/faq/ubi.html#L_subpage_verify_fail
http://lists.infradead.org/pipermail/linux-mtd/2011-February/034089.html
Reported-by: Andrea Adami <andrea.adami@gmail.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
CONFIG_KMEMCHECK depends on !FUNCTION_TRACER but
CONFIG_FUNCTION_TRACER is being turned on by a combination of other
configs which are set. In order to prevent this we need to explicitely
disable CONFIG_STACK_TRACER.
Signed-off-by: Mark Asselstine <mark.asselstine@windriver.com>
|
|
Add support for generic syscall entry/exit tracing.
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
|
|
1/1 [
Author: Bruce Ashfield
Email: bruce.ashfield@windriver.com
Subject: preempt-rt: import 3.4.4-rt13
Date: Sun, 1 Jul 2012 00:19:31 -0400
Importing 3.4.4-rt13 from:
http://www.kernel.org/pub/linux/kernel/projects/rt/3.4/patch-3.4.4-rt13.patch
The only changes were minor context adjustments.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
We've enabled (=m) INGRESS, but we should also explicitly
select ACT, since we have this in Kconfig:
config NET_SCH_INGRESS
tristate "Ingress Qdisc"
depends on NET_CLS_ACT
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Removed in mainline commit a77aea92010acf54ad785047234418d5d68772e2
-----------
cgroup: remove the ns_cgroup
The ns_cgroup is an annoying cgroup at the namespace / cgroup frontier and
leads to some problems:
* cgroup creation is out-of-control
* cgroup name can conflict when pids are looping
* it is not possible to have a single process handling a lot of
namespaces without falling in a exponential creation time
* we may want to create a namespace without creating a cgroup
The ns_cgroup was replaced by a compatibility flag 'clone_children',
where a newly created cgroup will copy the parent cgroup values.
The userspace has to manually create a cgroup and add a task to
the 'tasks' file.
This patch removes the ns_cgroup as suggested in the following thread:
https://lists.linux-foundation.org/pipermail/containers/2009-June/018616.html
[...]
-----------
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
The mti malta board description for a little endian config can't
be found by default unless it has a matching KMACHINE, or the
KMACHINE is set in a kernel recipe. In this case, it is easier to
just add qemumipsel in the KMACHINE list, so the recipe doesn't
need to do the mapping.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
This option no longer exists in the mainline kernel.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Enable virtio-console by default when virtio is requested by a BSP.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Posted to the RT list (April 15, 2012) here:
http://marc.info/?l=linux-rt-users&m=133452395925807&w=2
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
This is a stand-alone build/boot test automater. Since it is not version
dependent, we can draw in the v3.5 updates easily. Lots of examples on
how to use it were added in v3.5 which is really useful.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Straightforward backport from v3.5.
See http://lwn.net/Articles/496509/ for a nice description of it.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Make this a separate subdir of its own since it isn't strictly
related to cgroups, and we want a home for backports of new
net sched plugins.
Also fix the missing ".scc" extension on the cgroups include,
which isn't strictly required, but git grep can find it easier.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
The sys940 is failing when run through buildall. If we instrument the
failure case, we will see:
PASS:
---------------------
KMACHINE KTYPE KARCH META BRANCH SCC
atom-pc preempt-rt i386 atom-pc-preempt-rt-meta standard/preempt-rt/base atom-pc-preempt-rt
---------------------
FAIL:
---------------------
KMACHINE KTYPE KARCH META BRANCH SCC
sys940x standard i386 sys940x-standard-meta standard sys940x-standard
---------------------
Note in the above, the branch reference "standard" which does
not exist. It needs to be standard/base
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Both kernel types source cfg/fs/ext3 so there is no need for
anyone else to be duplicating the settings in a bsp or in
a ktype cfg file.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
The ipv6 support is no longer the experimental/academic project
that it was many years ago. People expect it to be there as
part of basic support. So make it built in, by aligning with
the current kernel.org defconfig.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Since there is no need for them to differ on something that
is not going to be impacted by RT patches.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
There is no reason for RT and standard to have their own personal
copies of bridge netfilter settings. Make them source a common file.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
There is no reason for RT and standard to have their own personal
copies of IPv6 netfilter settings. Make them source a common file.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
There is no reason for RT and standard to have their own personal
copies of IPv4 netfilter settings. Make them source a common file.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Since the RT scc derives itself from standard-nocfg.scc
it won't get the basic fs related options and similar
that are called out in standard.scc itself. Since there
is no reason to not align on these kinds of choices (like
ext3 etc) across standard and rt, the includes in rt should
largely mirror those used in standard, until there is a
concrete reason not to (i.e. feature X incompatible w/ RT).
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
So that it is alongside the other fs related content.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Take the handful of active lines in preempt-rt.cfg and migrate
them into preempt-rt-base.cfg. Then rename to lose the "-base"
part from the name, i.e.:
renamed: ktypes/preempt-rt/preempt-rt-base.cfg -> ktypes/preempt-rt/preempt-rt.cfg
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
The net_traffic.scc file was commented out "pending uprev" but
that really just applied to the blkio content, since the net one
is purely configurational and not containing patches.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
As seen here: http://sourceforge.net/projects/ioband/files/ the files
have been abandoned (last change in 2009) with the retooled solution
from the same author going upstream: http://lwn.net/Articles/413015/
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
This config/scc can be confusing if only looking at this
repo in isolation, since the scc is not included anywhere.
Add a comment so it is clear that the use case was external
and not included from an existing kernel type.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
These are the defaults for x86 and x86_64 when doing a defconfig
on those arch with the kernel.org tree. Ours might as well be
the same values.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
See 00-README file added in this commit for more details on
what it is, where it is documented and how it can be tested.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
1/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: Sched: Import base BFS-423 for v3.4
Date: Tue, 12 Jun 2012 14:11:21 -0700
Integrating and porting BFS to v3.4 from:
http://ck.kolivas.org/patches/bfs/3.4.0/3.4-sched-bfs-423.patch
[The Brain Fuck Scheduler v0.423 by Con Kolivas.
A single shared runqueue O(n) strict fairness earliest deadline first design.
Excellent throughput and latency for 1 to many CPUs on desktop and server
commodity hardware.
Not recommended for 4096 cpus.
Scalability is optimal when your workload is equal to the number of CPUs on
bfs. ie you should ONLY do make -j4 on quad core, -j2 on dual core and so on.
Features SCHED_IDLEPRIO and SCHED_ISO scheduling policies as well.
You do NOT need to use these policies for good performance, they are purely
optional for even better performance in extreme conditions.
To run something idleprio, use schedtool like so:
schedtool -D -e make -j4
To run something isoprio, use schedtool like so:
schedtool -I -e amarok
Includes accurate sub-tick accounting of tasks so userspace reported
cpu usage may be very different if you have very short lived tasks.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
2/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: mm minimal swappiness
Date: Thu, 14 Jun 2012 10:43:31 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[I've yet to find any desktop workload that benefits from anything higher than
a swappiness as low as possible. Unfortunately setting it to zero can rarely
lead to unexplained stalls in the VM so use 10.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
3/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: mm enable swaptoken only when swap full
Date: Thu, 14 Jun 2012 10:46:39 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[The swap token is only useful in conditions of swap thrash, and actually
worsens the common case by causing more swapping. Make it only have an effect
when swap is more than half full.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
4/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: mm drop swap cache aggressively
Date: Thu, 14 Jun 2012 10:48:18 -0700
Integrating and porting BFS from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[While it may be nice to have a copy of pages on swap once written there, the
more garbage we leave in the swapspace the slower any further writes and
reads to and from it are. Just free swapcache whenever we can.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
5/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: mm kswapd inherit prio 1
Date: Thu, 14 Jun 2012 10:50:20 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[When kswapd is awoken due to reclaim by a running task, set the priority of
kswapd to that of the calling task thus making memory reclaim cpu activity
affected by nice level.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
6/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: mm idleprio prio 1
Date: Thu, 14 Jun 2012 10:52:13 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[Set the effective priority of idleprio tasks to that of nice 19 tasks when
modifying vm reclaim behaviour.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
7/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: mm decrease default dirty ratio 1
Date: Thu, 14 Jun 2012 10:53:24 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[The default dirty ratio is chosen to be a compromise between throughput and
overall system latency. On a desktop, if an application writes to disk a lot,
that application should be the one to slow down rather than the desktop as a
whole. At higher dirty ratio settings, an application could write a lot to
disk and then happily use lots of CPU time after that while the rest of the
system is busy waiting on that naughty application's disk writes to complete
before anything else happening.
Lower ratios mean that the application that do a lot of disk writes end up
being responsible for their own actions and they're the ones that slow down
rather than the system in general.
This does decrease overall write throughput, but to the benefit of the latency
of the system as a whole with disk writes.
Other checks may well round this up to 10 still at startup.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
8/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: kconfig expose vmsplit option
Date: Thu, 14 Jun 2012 10:55:27 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[The options to alter the vmsplit to enable more lowmem are hidden behind the
expert option. Make it more exposed for -ck users and make the help menu
more explicit about what each option means.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
9/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: hz default 1000
Date: Thu, 14 Jun 2012 10:56:36 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[Set default HZ to 1000 which is what most desktop users should still be using.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
10/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: hz no default 250
Date: Thu, 14 Jun 2012 10:57:50 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[Make 250HZ not be the default to discourage desktop users from choosing this
option since 1000 will provide better latencies with only miniscule amounts
of extra overhead and power consumption.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
11/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: hz raise max
Date: Thu, 14 Jun 2012 10:59:11 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[There's some really badly broken software out there that is entirely
dependant on HZ for its maximum performance. Raise the maximum HZ value
to some higher and slightly unreasonable values up to some higher and
completely obscene values.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
12/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: preempt desktop tune
Date: Thu, 14 Jun 2012 11:00:14 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[Enable preempt by default and make people steer away from voluntary.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
13/13 [
Author: Yang Shi
Email: yang.shi@windriver.com
Subject: ck: ck2-version
Date: Thu, 14 Jun 2012 11:01:18 -0700
Integrating and porting from:
http://ck.kolivas.org/patches/3.0/3.4/3.4-ck2/patches/
[Add -ck version in inconspicuous place where it will merge relatively easily
with later kernel versions.
-ck]
Signed-off-by: Yang Shi <yang.shi@windriver.com>
]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
If we are in the cfg dir already, we don't need a leading
cfg prefix on the paths.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
This seems like a logical grouping to reduce the cluster
accumulation in cfg for another while.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
Things are starting to get a bit busy in the cfg dir. We may
not want to mirror every kernel path on a 1:1 basis, but at
least having a fs dir for filesystem cruft makes sense.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
In mainline defconfigs we have:
arch/x86/configs/i386_defconfig:# CONFIG_MTRR_SANITIZER is not set
arch/x86/configs/x86_64_defconfig:# CONFIG_MTRR_SANITIZER is not set
But the defaults given by the Kconfig stanza are "def_bool y" so
if we want to match the defconfigs, we need to call it out as off
explicitly.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
The efi.cfg calls out EFI and ACPI settings. Make sure it is in
both x86 and x86_64, and then delete all x86/x86_64 BSP specific
references directly to the efi.scc
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
Darren Hart
Bruce Ashfield
Nitin A Kamble
Mark Asselstine
Tom Zanussi
Paul Gortmaker