Now that we're using the scaling property in the Intel driver I noticed
that the names were a bit confusing. I've corrected them according to
our discussion on IRC and the mailing list, though I've left out
potential new additions for a new scaling property with an integer (or
two) for the scaling factor. None of the drivers implement that today,
but if someone wants to do it, I think it could be done with the
addition of a single new type and a new property to describe the
scaling factor in the X and Y directions.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Just a DRM_MASTER flag is sufficient here, though maybe this call is
totally deprecated anyway (xf86-video-intel still calls it though).
(airlied: drop ioctl auth_magic as discussed on mailing list also)
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
If userspace destroys a framebuffer that is in use on a crtc,
don't just null it out, tear down the crtc properly so the
hw gets turned off.
Signed-off-by: Dave Airlie <airlied@redhat.com>
The fallback case wasn't getting executed properly if there
was no TV table, which my T42 M7 hasn't got.
Signed-off-by: Dave Airlie <airlied@redhat.com>
LVDS always requests RMX_FULL, we need to fix it so that doesn't happen
before we can enable LVDS on crtc 1.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Recycling turns out to be a bad idea! For most use cases, the
packet can not be reused: TCP packets are cloned. Even for the ideal
case of forwarding, it hurts performance because of CPU ping/pong.
On a multi-core system forwarding of 64 byte packets is worse
much worse: recycling = 24% forwarded vs no recycling = 42% forwarded
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Don't reference the list element in hardware transmit ring on transmit
completion. The list element is updated by hardware, therefore
it causes a cache miss. Do book keeping in software structure.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allocate and size transmit ring based on parameters. Saves excess
space and allows configuring larger rings for testing.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The code for list element error (which should only happen on hardware
errors) should be cleaner and safer. Gets rid of unused ring_size
argument, which makes next patch easier.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch saves elements on transmit ring by only updating the upper
64 bit address when it changes. With many workloads skb's are located
in same region, so it saves space.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the existing macros to show where DMA address is being broken
apart. This is cosmetic only.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
yellowfin_init_ring() needs to clean up if dev_alloc_skb() fails and
should pass an error status up to the caller. This also prevents an
buffer underrun if failure occurred in the first iteration.
yellowfin_open() which calls yellowfin_init_ring() should free its
requested irq upon failure.
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
I think arch/sparc/kernel/sys32.S has an incorrect splice definition:
SIGN2(sys32_splice, sys_splice, %o0, %o1)
The splice() prototype looks like :
long splice(int fd_in, loff_t *off_in, int fd_out,
loff_t *off_out, size_t len, unsigned int flags);
So I think we should have :
SIGN2(sys32_splice, sys_splice, %o0, %o2)
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
I misunderstood the meaning of __bitwise. In practice it makes sparse
warn about every use of mmds which is certainly not what we want.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
security: Fix prompt for LSM_MMAP_MIN_ADDR
security: Make LSM_MMAP_MIN_ADDR default match its help text.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu:
percpu: use the right flag for get_vm_area()
percpu, sparc64: fix sparse possible cpu map handling
init: set nr_cpu_ids before setup_per_cpu_areas()
If one were to run localmodconfig or localyesconfig without having
a .config already in the file, then the end of the process would give
a warning when it tries to move the old .config to .config.old.
This patch adds a test to check if .config exists and avoid the moves
if it does not.
[ Impact: remove warning after make localmodconfig ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
As Andi Kleen pointed out, most people would expect that the local .config
file to be based for a streamline config. This patch changes the order
of searching for a config file to consider the .config in the local
directory first.
Reported-by: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Many distros put their config in /boot/config-`uname -r`, add a check
for that right after /proc/config.gz
Reported-by: Alan Jenkins <sourcejedi.lkml@googlemail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Due to cut and paste error IKCONFIG was both set and cleared.
It was suppose to be IKCONFIG_PROC to be cleared.
Also cleaned up if nesting.
Reported-by: Alan Jenkins <sourcejedi.lkml@googlemail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Instead of using the .config in the local directory. This patch
changes streamline_config.pl to search various locations for a config.
Here's the list and order of search:
/proc/config.gz
/boot/vmlinuz-`uname -r`
vmlinux # local to the directory
/lib/modules/`uname -r`/kernel/kernel/configs.ko
kernel/configs.ko
kernel/configs.o
.config
Once it finds a file that contains a config (it checks if the binary
objects have configs first) it then uses it to create the .config
with minimum modules needed.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
If CONFIG_IKCONFIG is set but CONFIG_IKCONFIG_PROC is not, then
gcc will optimize the config.gz out, because nobody uses it.
This patch adds "__used" to the config.gz data to keep it around so that
code like extract-ikconfig can still find it.
[ Impact: allow extract-ikconfig to find config.gz ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
It would be nice to use extract-ikconfig to find the congfig.gz
in either vmlinux (not vmlinuz) or configs.ko.
This patch changes the script to also be able to read ELF files directly.
[ Impact: find config.gz in vmlinux and configs.ko ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Ingo Molnar suggested that the streamline_config.pl should enable
CONFIG_IKCONFIG to keep the current config in the kernel.
Then we can use scripts/extract-ikconfig to find the current
modules.
This patch changes streamline_config.pl to check if CONFIG_IKCONFIG
is not set, and if it is not, it enables it to be a module.
[ Impact: make current config options easier to find ]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
The streamline_config.pl finds all the configs that are needed to
compile the currently loaded modules. After it creates the .config
file, it tests to make sure all the configs that are needed were
set.
It only looks at the configs that are modules, it does not look
at the builtin configs. This causes unnecessary warnings about modules
not being covered.
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
If a config does not have a prompt, it must be selected.
streamline_config.pl keeps track of all configs that select other configs.
If a config that does not have a prompt needs to be set to enable a
current module, it will include all configs that select it.
Note, streamline_config.pl does not enable modules that are not already
enabled. It only keeps enabled those that were enabled and might be
needed to compile the current modules.
The code to find the selects of a config is after the code that
adds the depends. But if a config needed selects but had no dependencies,
it would not be set. Because the code would stop before getting to
the select.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
This adds the option localyesconfig to make. This is similar to
localmodconfig, but after it removes unnecessary modules it runs
sed -i s/=m/=y/
on the .config file. It then runs "make silentoldconfig" to fix any
wholes that were created by the conversion of modules to core.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Running the streamline_config.pl script manually can still be confusing
for some users. This patch adds the localmodconfig option. This will
automatically run streamline_config.pl on the current .config and
then run "make silentoldconfig" to fix any wholes that might have been
created.
$ make localmodconfig
This will remove any module configurations in .config that are not needed
to compile the modules that are loaded.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
streamline_config.pl is a very powerful tool. For those that install
a kernel to a new box using the config file from the distribution know that
it can take forever to compile the kernel.
Making a custom config file that will still boot your box, but bring
down the compile time of the kernel can be quit painful, and to ask
someone that reported a bug to do this can be a large burdon since that
person may not even know how to build a kernel.
This script will perform "lsmod" to find all the modules loaded on the
current running system. It will read all the Makefiles to map which
CONFIG enables a module. It will read the Kconfig files to find the
dependencies and selects that may be needed to support a CONFIG.
Finally, it reads the .config file and removes any module "=m" that is
not needed to enable the currently loaded modules. The output goes to
standard out.
Here's a way to run the script. From the Linux directory that holds
a distribution .config.
$ scripts/kconfig/streamline_config.pl arch/x86/Kconfig > config-sl
$ mv .config config-save
$ mv config-sl .config
$ make oldconfig
Now you have a .config that will still build all your modules, but also
take much less time to build the kernel.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
If node_load[] is cleared everytime build_zonelists() is
called,node_load[] will have no help to find the next node that should
appear in the given node's fallback list.
Because of the bug, zonelist's node_order is not calculated as expected.
This bug affects on big machine, which has asynmetric node distance.
[synmetric NUMA's node distance]
0 1 2
0 10 12 12
1 12 10 12
2 12 12 10
[asynmetric NUMA's node distance]
0 1 2
0 10 12 20
1 12 10 14
2 20 14 10
This (my bug) is very old but no one has reported this for a long time.
Maybe because the number of asynmetric NUMA is very small and they use
cpuset for customizing node memory allocation fallback.
[akpm@linux-foundation.org: fix CONFIG_NUMA=n build]
Signed-off-by: Bo Liu <bo-liu@hotmail.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since the changes to the bitbang driver, there is the possibility we will
be called with either the speed_hz or bpw values zero. We take these to
mean that the default values (8 bits per word, or maximum bus speed).
Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently the clock rate calculation may round as pleased, which means
that it is possible that we will round down and end up with a faster clock
rate than intended.
Change the calculation to use DIV_ROUND_UP() to ensure that we end up with
a clock rate either the same as or lower than the user requested one.
Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The commit 2ff05b2b (oom: move oom_adj value) moveed the oom_adj value to
the mm_struct. It was a very good first step for sanitize OOM.
However Paul Menage reported the commit makes regression to his job
scheduler. Current OOM logic can kill OOM_DISABLED process.
Why? His program has the code of similar to the following.
...
set_oom_adj(OOM_DISABLE); /* The job scheduler never killed by oom */
...
if (vfork() == 0) {
set_oom_adj(0); /* Invoked child can be killed */
execve("foo-bar-cmd");
}
....
vfork() parent and child are shared the same mm_struct. then above
set_oom_adj(0) doesn't only change oom_adj for vfork() child, it's also
change oom_adj for vfork() parent. Then, vfork() parent (job scheduler)
lost OOM immune and it was killed.
Actually, fork-setting-exec idiom is very frequently used in userland program.
We must not break this assumption.
Then, this patch revert commit 2ff05b2b and related commit.
Reverted commit list
---------------------
- commit 2ff05b2b4e (oom: move oom_adj value from task_struct to mm_struct)
- commit 4d8b9135c3 (oom: avoid unnecessary mm locking and scanning for OOM_DISABLE)
- commit 8123681022 (oom: only oom kill exiting tasks with attached memory)
- commit 933b787b57 (mm: copy over oom_adj value at fork time)
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 788084aba2 added the LSM_MMAP_MIN_ADDR
option, whose help text states "For most ia64, ppc64 and x86 users with lots
of address space a value of 65536 is reasonable and should cause no problems."
Which implies that it's default setting was typoed.
Signed-off-by: Dave Jones <davej@redhat.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
The auto-mic clean-up patches caused regressions on some ALC268 models
that have no proper input_mux but with "Input Source" mixer elements.
Such a combination results in Oops when accessed.
[A reason why set_capture_mixer() isn't used in patch_alc268() is that
ALC268 codec have HDA_OUTPUT direction for capture volumes unlike other
codecs. Thus it needs own definitions of capture elements.]
This patch fixes the issues:
- Add a capture mixer definition without input-source
- Use the new capture mixer appropriately
Signed-off-by: Takashi Iwai <tiwai@suse.de>
