On Sun, Dec 02, 2012 at 03:17:27AM +, Ben Hutchings wrote:
> On Fri, 2012-11-30 at 10:54 -0800, Greg Kroah-Hartman wrote:
> > 3.4-stable review patch. If anyone has any objections, please let me know.
> >
> > --
> >
> > From: Dan Carpenter
> >
> > commit 5bc9ad774c063f6b419
Hi Benjamin,
> > I can think of two mechanisms that might be useful in finding a
> > way to achieve this cleanly: a) Let a driver return a value telling
> > whether to change input device, and b) Let a second driver have a go
> > at the same device report. Some return value or state could determin
Ping!!
> -Original Message-
> From: Sethi Varun-B16395
> Sent: Monday, November 26, 2012 10:55 AM
> To: joerg.roe...@amd.com
> Cc: Sethi Varun-B16395; io...@lists.linux-foundation.org; linuxppc-
> d...@lists.ozlabs.org; linux-kernel@vger.kernel.org; Wood Scott-B07421;
> Tabi Timur-B04825
>
On Sat, 2012-12-01 at 22:44 +0100, Ingo Molnar wrote:
> Should we use some other file for that - or no file at all and
> just emit a bootup printk for kernel hackers with a short
> attention span?
Or, whack the file and don't bother with a printk either. If it's in
your config, and your comman
On Mon, 26 Mar 2012 00:29:24 +0200 "Rafael J. Wysocki" wrote:
> Thanks for the confirmation.
>
> Below it goes again with a changelog and tags.
>
> I don't really think that SDIO does the right thing here overall, but that's
> all I can do to address the problem timely.
>
> Thanks,
> Rafael
commit a66f59ba2e994bf70274ef0513e24e0e7ae20c63
bq27x00_battery: Add support for BQ27425 chip
introduced 2 bug.
1/ 'chip' was set to BQ27425 unconditionally - breaking support for
other devices.
2/ BQ27425 does not support cycle count, how the code still tries to
get the cycle count f
From: Wei Yongjun
Add the missing unlock before return from function host1x_drm_init()
and host1x_drm_exit() in the error handling case.
Signed-off-by: Wei Yongjun
---
drivers/gpu/drm/tegra/host1x.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/tegra/host1x.c b/drivers/
"entries" is unsigned here, so it is never less than zero. In theory,
len could be less than offset so I have added a check for that.
Signed-off-by: Dan Carpenter
diff --git a/drivers/net/wireless/p54/eeprom.c
b/drivers/net/wireless/p54/eeprom.c
index 1ef1bfe..d43e374 100644
--- a/drivers/net/
This is a static checker fix. The problem is that we store the bits
from "uv_apicid_hibits" into "apicid" (the high 16 bits) but then we
shift it 16 bit to the left. "apicid" is an int so it wraps and we lose
them.
Signed-off-by: Dan Carpenter
---
v2: Style fix. Don't use ulong.
I don't have
On Sat, Dec 1, 2012 at 7:00 PM, Michael Cree wrote:
> On Tue, Oct 09, 2012 at 10:15:08AM +0100, David Howells wrote:
>> Can you merge the following branch into the alpha tree please.
>>
>> This is to complete part of the UAPI disintegration for which the preparatory
>> patches were pulled recently
On 01.12.2012 23:42, Dave Airlie wrote:
> Guys I think you guys might be overthniking things here.
>
> I know you have some sort of upstream/downstream split, but really in
> the upstream kernel, we don't care about that, so don't make it our
> problem.
I am not trying to make anything your probl
Unify struct ft1000_info between ft1000-usb and ft1000-pcmcia and move it to
common ft1000.h.
Signed-off-by: Ondrej Zary
---
drivers/staging/ft1000/ft1000-pcmcia/ft1000.h | 33 +-
drivers/staging/ft1000/ft1000-pcmcia/ft1000_hw.c | 21 ++-
drivers/staging/ft1000/ft1000-pcmcia/ft100
I've narrowed it down to sometime between v3.6.8 and v3.7-rc1. So none of the
v3.7 series will work with this GPU/panel.
I don't think it has anything to do with the dell modules since they haven't
changed.
Tom
- Original Message -
> From: "Tom St Denis"
> To: linux-kernel@vger.kern
On Sun, Dec 2, 2012 at 12:32 PM, Tom St Denis wrote:
> I've narrowed it down to sometime between v3.6.8 and v3.7-rc1. So none of
> the v3.7 series will work with this GPU/panel.
Hm, that's still a few hundred relevant commits intermingled with a
few thousand that likely don't matter ;-) Can you
Including a bunch more lists to the cc.
On Sun, Dec 2, 2012 at 12:42 PM, Daniel Vetter wrote:
> On Sun, Dec 2, 2012 at 12:32 PM, Tom St Denis
> wrote:
>> I've narrowed it down to sometime between v3.6.8 and v3.7-rc1. So none of
>> the v3.7 series will work with this GPU/panel.
>
> Hm, that's
From: Wei Yongjun
Use for_each_compatible_node() macro.
Signed-off-by: Wei Yongjun
---
arch/xtensa/platforms/xtfpga/setup.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/xtensa/platforms/xtfpga/setup.c
b/arch/xtensa/platforms/xtfpga/setup.c
index 71d61ca6..237c3
On Thu, Nov 29, 2012 at 06:00:00PM +0800, zwu.ker...@gmail.com wrote:
> From: Zhi Yong Wu
Please write a commit log to describe this patch, even though it is
quite simple and straightfoward
Otherwise, it looks good to me.
Reviewed-by: Zheng Liu
Regards,
On Tue, Nov 27, 2012 at 8:08 AM, Torsten Kaiser
wrote:
> On Tue, Nov 27, 2012 at 2:05 AM, NeilBrown wrote:
>> Can you test to see if this fixes it?
>
> Patch applied, I will try to get it stuck again.
> I don't have a reliable reproducers, but if the problem persists I
> will definitly report bac
On 12/01/2012 10:48 PM, Arnd Bergmann wrote:
I agree that is a concern, but for now, I'm mostly worried about
the kernel-to-user interface. If we can agree on a driver interface
that works for Xillybus as well as any of the others we know about,
we can start using that as the generic kernel FPGA
Hi, the problem in the quoted message still happens, shouldn't all of
ext4_inode_cache slab be emptied after "echo 3
> /proc/sys/vm/drop_caches"? In my case slab uses too much memory even
after all processes finish and system is in bad shape due to lack of
physical RAM. CC'ing tytso and Catalin Mar
From: Wei Yongjun
The variables node_page and page_offset are initialized but never used
otherwise, so remove those unused variables.
Signed-off-by: Wei Yongjun
---
fs/f2fs/file.c | 2 --
fs/f2fs/dir.c | 2 --
2 files changed, 4 deletions(-)
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index
Hi Eric,
Attached benchmark test-cwf.c(cc -o test-cwf test-cwf.c), the result
shows when last level cache(LLC) miss and CPU fetches data from
memory, critical word as first 64bit member in cache line has better
performance(costs 158290336 cycles ) than other positions(offset 0x10,
costs 164100732
On Thu, Nov 29, 2012 at 08:38:53PM -0700, Bjorn Helgaas wrote:
> That's essentially the patch at
> https://bugzilla.redhat.com/show_bug.cgi?id=757166#c16, which in my
> opinion is too ugly to consider. But fortunately, I'm not the
> maintainer for any IOMMU drivers.
There is a quirk infrastructur
On Mon, Nov 26, 2012 at 04:00:25PM -0600, Steven Kinney wrote:
> From: "Steven L. Kinney"
>
> Add Kernel configuration selection for AMD IOMMUv2 performance
> counters.
How can the counters be used?
Joerg
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
t
From: Wei Yongjun
The dereference to 'mfd' should be moved below the NULL test.
Signed-off-by: Wei Yongjun
---
drivers/mfd/sta2x11-mfd.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/mfd/sta2x11-mfd.c b/drivers/mfd/sta2x11-mfd.c
index 012a521..1ecbdfb 100644
--
From: Wei Yongjun
The dereference to 'mfd' should be moved below the NULL test.
Signed-off-by: Wei Yongjun
---
drivers/mfd/sta2x11-mfd.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/mfd/sta2x11-mfd.c b/drivers/mfd/sta2x11-mfd.c
index 012a521..1ecbdfb 100644
--
On Sunday, December 02, 2012 07:46:25 PM NeilBrown wrote:
> On Mon, 26 Mar 2012 00:29:24 +0200 "Rafael J. Wysocki" wrote:
>
>
> > Thanks for the confirmation.
> >
> > Below it goes again with a changelog and tags.
> >
> > I don't really think that SDIO does the right thing here overall, but th
Ok so on v3.7-rc1 [and I suspect up] I can manually control the brightness by
echoing to
/sys/class/backlight/intel_backlight/brightness
But I still can't use the buttons to control it (it only goes one tick down
from max and stops there). I'm using GNOME3 from a x86_64 unstable debian
insta
On Sunday 02 December 2012 11:36:09 Dan Carpenter wrote:
> "entries" is unsigned here, so it is never less than zero. In theory,
> len could be less than offset so I have added a check for that.
>
> Signed-off-by: Dan Carpenter
Acked-by: Christian Lamparter
> diff --git a/drivers/net/wireless/
On Fri, Nov 30, 2012 at 03:20:27PM -0200, Raphael S Carvalho wrote:
> Well, the below function reports the physical-address width supported
> by the processor.
> It does its work very well, though I found a detail which it doesn't
> handle at all.
>
> PS: The following function is not a patch.
> i
Hmm, we need to work out a good abstraction for this.
On Tue, Nov 20, 2012 at 07:24:56PM +0530, Varun Sethi wrote:
> Added the following domain attributes required by FSL PAMU driver:
> 1. Subwindows field added to the iommu domain geometry attribute.
Are the Subwindows mapped with full size or d
Hi Marek,
On Mon, Nov 26, 2012 at 11:57:19AM +0100, Marek Szyprowski wrote:
> I've took all the patches to the next-dma-debug branch in my tree, I sorry
> that You have to wait so long for it. My branch is based on Joerg's
> dma-debug branch and I've included it for testing in linux-next branch.
Hey, Oleg.
On Tue, Nov 27, 2012 at 07:41:17PM +0100, Oleg Nesterov wrote:
> > Oleg, do you remember? Why do we need
> > this?
>
> No, I forgot. And this code was changed after that, the fat comment in
> cpuset_cpus_allowed_fallback() tried to explain the code below which
> was removed.
>
> I am
From: Rafael J. Wysocki
Move some suspend-specific and hibernate-specific code from
acpi_sleep_init() into separate functions to get rid of explicit
#ifdefs in acpi_sleep_init(). Use pr_info() to start and pr_cont()
to continue printing the supported ACPI sleep states line.
Signed-off-by: Rafae
On Sun, Nov 25, 2012 at 01:17:13AM +0400, Cyrill Gorcunov wrote:
> Otherwise we are getting a nil dereference
>
> | [0.00] BUG: unable to handle kernel NULL pointer dereference at
>(null)
> | [0.00] IP: [] strcmp+0x10/0x30
>
> Signed-off-by: Cyrill Gorcunov
Applied
The of_device_id table is supposed to be zero-terminated.
Signed-off-by: Axel Lin
---
drivers/irqchip/irq-sunxi.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/irqchip/irq-sunxi.c b/drivers/irqchip/irq-sunxi.c
index eef41a4..10974fa 100644
--- a/drivers/irqchip
On Sun, Dec 2, 2012 at 2:49 PM, Tom St Denis wrote:
> Ok so on v3.7-rc1 [and I suspect up] I can manually control the brightness by
> echoing to
>
> /sys/class/backlight/intel_backlight/brightness
>
> But I still can't use the buttons to control it (it only goes one tick down
> from max and stop
> + * If a different line discipline is selected
> by mistake it
> + * will have valid memory for both arguments.
> + */
> + ldisc_receive_buf(tty, buf->data, buf->data,
> framesize)
* Rik van Riel wrote:
> On 12/01/2012 03:10 PM, Ingo Molnar wrote:
> >
> >Convert the struct anon_vma::mutex to an rwsem, which will help
> >in solving a page-migration scalability problem. (Addressed in
> >a separate patch.)
> >
> >The conversion is simple and straightforward: in every case
> >
On 11/23/2012 06:44 PM, Tang Chen wrote:
> From: Yasuaki Ishimatsu
>
> If system can create movable node which all memory of the
> node is allocated as ZONE_MOVABLE, setup_node_data() cannot
> allocate memory for the node's pg_data_t.
> So when memblock_alloc_nid() fails, setup_node_data() retrie
Joerg Roedel wrote:
> When you add implementation specific attributes please add some
> indication to the names that it is only for PAMU. DOMAIN_ATTR_STASH
> sounds too generic.
We were thinking that maybe this attribute could be useful to other IOMMUs
in the future. Stashing is not a concept th
On Fri, Nov 30, 2012 at 01:19:33PM +0800, Axel Lin wrote:
> Use DIV_ROUND_UP to ensure selected voltage won't less than min_uV due to
> integer truncation.
Applied, thanks.
signature.asc
Description: Digital signature
* Rik van Riel wrote:
> >+static inline void anon_vma_lock_read(struct anon_vma *anon_vma)
> >+{
> >+down_read(&anon_vma->root->rwsem);
> >+}
>
> I see you did not rename anon_vma_lock and anon_vma_unlock to
> anon_vma_lock_write and anon_vma_unlock_write.
>
> That could get confusing to
On Fri, Nov 30, 2012 at 04:52:49PM +0800, Axel Lin wrote:
> If pdata->base_voltage_uV is missing or the settings of pdata->base_voltage_uV
> and pdata->max_voltage_uV are out of range, TPS51632_VOLT_VSEL macro
> returns wrong vsel.
Applied, thanks.
signature.asc
Description: Digital signature
On Fri, Nov 30, 2012 at 08:39:53PM +0800, Axel Lin wrote:
> Signed-off-by: Axel Lin
Applied, thanks. Always CC maintainers.
signature.asc
Description: Digital signature
The of_device_id table is supposed to be zero-terminated.
Signed-off-by: Axel Lin
---
drivers/clocksource/sunxi_timer.c |1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/clocksource/sunxi_timer.c
b/drivers/clocksource/sunxi_timer.c
index 3c46434..3cd1bd3 100644
--- a/drivers/clocks
On Mon, Nov 19, 2012 at 12:12:29PM +0530, Shubhrajyoti D wrote:
> Fix the below warning
> drivers/spi/spi-omap2-mcspi.c:336:34: warning: symbol 'tx' shadows an earlier
> one
> drivers/spi/spi-omap2-mcspi.c:327:12: originally declared here
Applied, thanks.
signature.asc
Description: Digital sign
On Fri, Nov 23, 2012 at 01:25:24PM -0800, Andrey Smirnov wrote:
> Second version of the patch with incorporated suggestions form
> Mark(hopefully):
>https://lkml.org/lkml/2012/11/20/249
>
> Once again at this point this patch is for illustrative purposes. If
> the overall structure and nature
This fixes below build error:
CC arch/arm/mach-sunxi/sunxi.o
arch/arm/mach-sunxi/sunxi.c: In function 'sunxi_restart':
arch/arm/mach-sunxi/sunxi.c:55:3: error: implicit declaration of function
'mdelay' [-Werror=implicit-function-declaration]
cc1: some warnings being treated as errors
make[
Thanks Daniel so I'm taking back my bug report on the i915 and claiming this as
a bug against the acpi_video stack.
What sort of information would the ACPI folk like from me to help diagnose this?
Here's a dmesg dump of all ACPI messages.
http://pastebin.com/b4XKe5Fk
Tom
- Original Messag
Am 02.12.2012 11:44, schrieb Dan Carpenter:
> This is a static checker fix. The problem is that we store the bits
> from "uv_apicid_hibits" into "apicid" (the high 16 bits) but then we
> shift it 16 bit to the left. "apicid" is an int so it wraps and we lose
> them.
>
> Signed-off-by: Dan Carp
On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> This addes multiqueue support to virtio_net driver. In multiple queue modes,
> the
> driver expects the number of queue paris is equal to the number of vcpus. To
> eliminate the contention bettwen vcpus and virtqueues, per-cpu virtqueue
On Tue, Nov 27, 2012 at 06:16:00PM +0800, Jason Wang wrote:
> This patch implement the {set|get}_channels method of ethool to allow user to
> change the number of queues dymaically when the device is running. This would
> let the user to configure it on demand.
>
> Signed-off-by: Jason Wang
> ---
Don't try to NUMA-balance hard-bound tasks in vein. This
also makes it easier to compare hard-bound workloads against
NUMA-balanced workloads, because the NUMA code will
be completely inactive for those hard-bound tasks.
( Keep a debugging feature flag around: for development it
makes sense to o
NUMA-balancing, combined with NUMA-affine memory migration,
is a relatively long-term process (compared to the typical
time scale of scheduling) that takes time to establish and
converge - on the time scale of of several seconds or more.
Small, short-lived and don't have much of a NUMA placement
c
I've been testing wider workloads and here's two more small and
obvious patches rounding up numa/core behavior around the edges.
The NUMA code should now be pretty unintrusive to all but the
long-running, memory-intense workloads where it's expected to
make a (positive) difference.
Short-run work
On Sat, Dec 01, 2012 at 12:44:01PM +0100, Francois Romieu wrote:
> Wang YanQing :
> > + /*
> > +*This is a fix for BIOS forget to set
> > +*extend GigaMAC registers
> > +*Wang YanQing 12/1/2012
> > +*/
>
> This part will go into the changelog.
I think brevity comment in code is
inputbox code don't support use Left/Right to change input
position in the input box, so it use continue to skip them
when input key is Left/Right, but use break for Up/Down, so
we can change selected button with Up/Down but we can't do it
with Left/Right key when input box selected.
This patch fi
On Sun, Dec 2, 2012 at 4:05 PM, Wei Yongjun wrote:
> From: Wei Yongjun
>
> Use for_each_compatible_node() macro.
>
> Signed-off-by: Wei Yongjun
Acked-by: Max Filippov
Queued at my xtensa-fixes-for-upstream branch.
--
Thanks.
-- Max
--
To unsubscribe from this list: send the line "unsubscrib
On Sat, Dec 01, 2012 at 07:07:48PM -0700, Shuah Khan wrote:
> On Fri, Nov 30, 2012 at 11:45 AM, Greg Kroah-Hartman
> wrote:
> > This is the start of the stable review cycle for the 3.0.54 release.
> > There are 37 patches in this series, all will be posted as a response
> > to this one. If anyone
On Sat, Dec 01, 2012 at 07:11:48PM -0700, Shuah Khan wrote:
> On Sat, Dec 1, 2012 at 9:50 AM, Greg Kroah-Hartman
> wrote:
> > On Sat, Dec 01, 2012 at 11:58:28PM +0900, Satoru Takeuchi wrote:
> >> At Fri, 30 Nov 2012 10:55:31 -0800,
> >> Greg Kroah-Hartman wrote:
> >> >
> >> > This is the start of
On Sun, 2012-12-02 at 21:25 +0800, Ling Ma wrote:
> Hi Eric,
>
> Attached benchmark test-cwf.c(cc -o test-cwf test-cwf.c), the result
> shows when last level cache(LLC) miss and CPU fetches data from
> memory, critical word as first 64bit member in cache line has better
> performance(costs 1582903
There is a fairly well written paper ;-) explaining how things are
supposed to work:
http://ols.fedoraproject.org/OLS/Reprints-2008/hallyn-reprint.pdf
The inheritable set is not intended to work the way you seem to want.
Naive inheritance like that is quite explicitly the opposite of what
was d
This is a static checker fix. The problem is that we store the bits
from "uv_apicid_hibits" into "apicid" (the high 16 bits) but then we
shift it 16 bit to the left. "apicid" is an int so it wraps and we lose
them.
I have also simplified uv_wakeup_secondary() a little based on a
suggestion.
Sig
If "buf" wasn't equal to LP5521_REG_R_CURR_DEFAULT the probe fails but
we still return zero. I've changed it to print an error message and
return -EINVAL.
Reported-by: Ben Hutchings
Signed-off-by: Dan Carpenter
diff --git a/drivers/leds/leds-lp5521.c b/drivers/leds/leds-lp5521.c
index 2bee6b9.
Hello,
so I tested v3.7-rc7-163-g3c46f3d and got a kernel panic.
Usually I am running Ubuntu on my Dell Latitutde E6520,
which usually works fine except some sound issues.
I did this:
git clone https://github.com/torvalds/linux.git
cd linux
make deb-pkg
cd ..
dpkg -i linux-headers-3.7.0-rc7+_3.7
On Sun, Dec 02, 2012 at 08:28:43PM +0300, Dan Carpenter wrote:
> This is a static checker fix. The problem is that we store the bits
> from "uv_apicid_hibits" into "apicid" (the high 16 bits) but then we
> shift it 16 bit to the left. "apicid" is an int so it wraps and we lose
> them.
>
> I have
On 11/30/2012 06:32 PM, Greg KH wrote:
> >>+static struct class *xillybus_class;
> >Why not just use the misc interface instead of your own class?
> When Xillybus is used, the whole system's mission is usually around
> it (e.g. it's a computer doing data acquisition throug
Wang, All,
On Sunday 02 December 2012 Wang YanQing wrote:
> inputbox code don't support use Left/Right to change input
> position in the input box, so it use continue to skip them
> when input key is Left/Right, but use break for Up/Down, so
> we can change selected button with Up/Down but we can'
So I tested vanilla v3.6 and v3.5 as well and this panic comes up.
Hence I guess my .config is screwed or the way I install the
kernel.
On 12/02/2012 06:30 PM, Stefan Beller wrote:
> Hello,
>
> so I tested v3.7-rc7-163-g3c46f3d and got a kernel panic.
> Usually I am running Ubuntu on my Dell Lati
After unifying struct ft1000_info, struct ft1000_device is now wrong name.
Rename it to ft1000_usb to match its semantics.
Signed-off-by: Ondrej Zary
---
drivers/staging/ft1000/ft1000-usb/ft1000_debug.c | 12 ++--
.../staging/ft1000/ft1000-usb/ft1000_download.c| 40 +-
drivers/
Wang, All,
On Saturday 01 December 2012 Wang YanQing wrote:
> If menuconfig have a Save button like alternative
> .config editors, xconfig, nconfig, etc.We will have
> a obvious benefit when use menuconfig just like
> when we use others, we can save our .config quickly
> and conveniently.
>
> Thi
On 12/02/2012 10:12 AM, Ingo Molnar wrote:
Rename anon_vma_[un]lock() => anon_vma_[un]lock_write(),
to make it clearer that it's an exclusive write-lock in
that case - suggested by Rik van Riel.
... close, but you forgot to actually rename the unlock function :)
diff --git a/include/linux/hu
On Sun, 2 Dec 2012 10:11:58 -0600
Paul Fulghum wrote:
> True, in this mode line disciplines other than N_HDLC would not be functional
> and N_HDLC ignores the flag buffer.
> This change won’t make other line disciplines useful, it will just prevent
> the case of a mistakenly selected line disci
On Sun, Dec 2, 2012 at 9:21 AM, Andrew G. Morgan wrote:
> There is a fairly well written paper ;-) explaining how things are
> supposed to work:
>
> http://ols.fedoraproject.org/OLS/Reprints-2008/hallyn-reprint.pdf
>
> The inheritable set is not intended to work the way you seem to want.
> Naive
Now that the code in numa/core is settling down nicely, I thought I'd
address a growing testing/contribution problem that has become a problem
over the last two weeks: the 'balacenuma' tree by Mel Gorman was drifting
away from numa/core (or, numa/core was drifting away from Mel's tree,
depending on
From: Mel Gorman
The compact_pages_moved and compact_pagemigrate_failed events
are convenient for determining if compaction is active and to
what degree migration is succeeding but it's at the wrong level.
Other users of migration may also want to know if migration is
working properly and this wi
From: Andrea Arcangeli
The objective of _PAGE_NUMA is to be able to trigger NUMA
hinting page faults to identify the per NUMA node working set of
the thread at runtime.
Arming the NUMA hinting page fault mechanism works similarly to
setting up a mprotect(PROT_NONE) virtual range: the present bit
From: Lee Schermerhorn
NOTE: I have not yet addressed by own review feedback of this
patch. At this point I'm trying to construct a baseline tree and will
apply
my own review feedback later and then fold it in.
This patch augments the MPOL_MF_LAZY feature by adding a "NOOP"
policy
From: Peter Zijlstra
Note: This was originally based on Peter's patch "mm/migrate:
Introduce migrate_misplaced_page()" but borrows extremely heavily from
Andrea's
"autonuma: memory follows CPU algorithm and task/mm_autonuma stats
collection". The end result is barely recogn
From: Mel Gorman
Note: Based on "mm/mpol: Use special PROT_NONE to migrate pages"
but sufficiently different that the signed-off-bys were dropped
Combine our previous _PAGE_NUMA, mpol_misplaced and
migrate_misplaced_page() pieces into an effective migrate on
fault scheme.
Note that (on x86)
From: Mel Gorman
The use of MPOL_NOOP and MPOL_MF_LAZY to allow an application to
explicitly request lazy migration is a good idea but the actual
API has not been well reviewed and once released we have to
support it. For now this patch prevents an application using the
services. This will need t
From: Peter Zijlstra
Introduce a per-page last_cpu field, fold this into the struct
page::flags field whenever possible.
The unlikely/rare 32bit NUMA configs will likely grow the page-frame.
[ Completely dropping 32bit support for CONFIG_NUMA_BALANCING would simplify
things, but it would also
From: Peter Zijlstra
Its a bit awkward but it was the least painful means of modifying the
queue selection. Used in a later patch to conditionally use a random
queue.
Signed-off-by: Peter Zijlstra
Cc: Paul Turner
Cc: Lee Schermerhorn
Cc: Christoph Lameter
Cc: Rik van Riel
Cc: Andrew Morton
From: Peter Zijlstra
This is probably a first: formal description of a complex high-level
computing problem, within the kernel source.
Signed-off-by: Peter Zijlstra
Cc: Linus Torvalds
Cc: Andrew Morton
Cc: "H. Peter Anvin"
Cc: Mike Galbraith
Cc: Rik van Riel
Cc: Hugh Dickins
[ Next step:
From: Peter Zijlstra
By accounting against the present PTEs, scanning speed reflects the
actual present (mapped) memory.
Suggested-by: Ingo Molnar
Signed-off-by: Peter Zijlstra
Cc: Linus Torvalds
Cc: Andrew Morton
Cc: Andrea Arcangeli
Cc: Rik van Riel
Cc: Mel Gorman
Cc: Hugh Dickins
Sign
Back off slowly from scanning, up to sysctl_sched_numa_scan_period_max
(1.6 seconds). Scan faster again if we were forced to switch to
another node.
This makes sure that workload in equilibrium don't get scanned as often
as workloads that are still converging.
Cc: Peter Zijlstra
Cc: Linus Torval
- break out of can_do_numa_run() earlier if we can make no progress
- don't flip between siblings that often
- turn on bidirectional fault balancing
- improve the flow in task_numa_work()
Cc: Peter Zijlstra
Cc: Linus Torvalds
Cc: Andrew Morton
Cc: Andrea Arcangeli
Cc: Rik van Riel
Cc: Mel
We will rely on the per CPU fault statistics and its
shared/private derivative even more in the future, so
stabilize this metric even better.
The staged updates introduced in commit:
sched: Introduce staged average NUMA faults
Already stabilized this key metric significantly, but in
real work
Interleave tasks that are 'shared' - i.e. whose memory access patterns
indicate that they are intensively sharing memory with other tasks.
If such a task ends up converging then it switches back into the lazy
node-local policy.
Build-Bug-Reported-by: Fengguang Wu
Cc: Peter Zijlstra
Cc: Linus To
This patch shows the power of the shared/private distinction: in
the shared tasks active balancing function (sched_update_ideal_cpu_shared())
we are able to to build a per (shared) task node mask of the nodes that
it and its buddies occupy at the moment.
Private tasks on the other hand are not aff
Zhouping Liu reported worse out-of-memory behavior with
CONFIG_NUMA_BALANCING=y, compared to the mainline kernel.
One reason for that change in behavior is that with typical
applications the mainline kernel allocates memory essentially
randomly, and leaves it where it was.
"Random" placement is n
Remove leftovers of the (now inactive) statistical NUMA scheduling code.
Cc: Linus Torvalds
Cc: Andrew Morton
Cc: Peter Zijlstra
Cc: Andrea Arcangeli
Cc: Rik van Riel
Cc: Mel Gorman
Cc: Hugh Dickins
Signed-off-by: Ingo Molnar
---
include/linux/sched.h | 2 -
kernel/sched/core.c |
Track strength of convergence, which is a value between 1 and 1024.
This will be used by the placement logic later on.
A strength value of 1024 means that the workload has fully
converged, all faults after the last scan period came from a
single node.
A value of 1024/nr_nodes means a totally spre
Change the adaptive memory policy code to take a majority of buddies
on a node into account. Previously, since this commit:
"sched: Track shared task's node groups and interleave their memory
allocations"
We'd include any node that has run a buddy in the past, which was too
aggressive and spre
Convert the struct anon_vma::mutex to an rwsem, which will help
in solving a page-migration scalability problem. (Addressed in
a separate patch.)
The conversion is simple and straightforward: in every case
where we mutex_lock()ed we'll now down_write().
Suggested-by: Linus Torvalds
Reviewed-by:
NUMA-balancing, combined with NUMA-affine memory migration,
is a relatively long-term process (compared to the typical
time scale of scheduling) that takes time to establish and
converge - on the time scale of of several seconds or more.
Small, short-lived and don't have much of a NUMA placement
c
Don't try to NUMA-balance hard-bound tasks in vein. This
also makes it easier to compare hard-bound workloads against
NUMA-balanced workloads, because the NUMA code will
be completely inactive for those hard-bound tasks.
( Keep a debugging feature flag around: for development it
makes sense to o
rmap_walk_anon() and try_to_unmap_anon() appears to be too
careful about locking the anon vma: while it needs protection
against anon vma list modifications, it does not need exclusive
access to the list itself.
Transforming this exclusive lock to a read-locked rwsem removes
a global lock from the
Mel Gorman reported that the NUMA code is system-time intense even
after a workload has converged.
To remedy this, turn sched_numa_scan_size into a range:
sched_numa_scan_size_min[default: 32 MB]
sched_numa_scan_size_max[default: 512 MB]
As workloads converge, so does thei
1 - 100 of 257 matches
Mail list logo
