The GEM code is used by dsi, so it can't be part of
the main module. This exports all functions that are called
by other code, and puts it into the helper module.
Signed-off-by: Arnd Bergmann
---
drivers/gpu/drm/msm/Makefile | 3 +--
drivers/gpu/drm/msm/msm_gem.c | 29 +
Moving around the Makefile for DSI caused a link error with
the main driver referring to a builtin module that has no
exported symbols:
drivers/gpu/built-in.o: In function `msm_drm_register':
:(.init.text+0x654): undefined reference to `msm_dsi_register'
This changes the DSI symbol to a 'tristate
Moving around the Makefile for EDP caused a link error with
the main driver referring to a builtin module that has no
exported symbols:
drivers/gpu/built-in.o: In function `msm_drm_register':
:(.init.text+0x654): undefined reference to `msm_edp_register'
This changes the EDP symbol to a 'tristate
The msm driver currently has a single module_init() function but
consists of multiple modules that each register a platform driver.
This changes the msm_drm_register() function to not call the
other functions but instead rely on module load order and/or
link order to ensure that the other modules
On Mon, Feb 22, 2016 at 09:56:50AM +0100, Michał Kępień wrote:
> > > /*
> > > * Certain keys are flagged as KE_IGNORE. All of these are either
> > > * notifications (rather than requests for change) or are also sent
> > > @@ -513,6 +533,7 @@ static int __init dell_wmi_init(void)
> > > {
> > >
Refrain from defining default fw_cfg register offsets on
unsupported architectures -- throw an error instead. If
QEMU were to add fw_cfg support on additional architectures,
we should add them to the FW_CFG_SYSFS depends statement in
drivers/firmware/Kconfig, and provide default values for
register
Guennadi Liakhovetski writes:
> On Sun, 21 Feb 2016, Robert Jarzmik wrote:
> Please, have a look at
> http://git.linuxtv.org/gliakhovetski/v4l-dvb.git/log/?h=for-4.6-2
> If all is good there, no need for a v6
Thanks for fixing the *_dma_irq() mess, and sorry for that.
And I just tested your br
On Fri, Feb 19, 2016 at 08:01:17PM -0600, John Dahlstrom wrote:
> On Fri, 19 Feb 2016, Darren Hart wrote:
>
> >See stable_kernel_rules.txt for how to specify relevant versions. I'm curious
> >why there is a gap from 3.18 to 4.1.
> >
> >What you probably meant was:
> >
> >Cc: # 3.18.x
> >
> >If yo
On Mon, Feb 22, 2016 at 12:11:48PM -0700, Jason Gunthorpe wrote:
> On Mon, Feb 22, 2016 at 09:08:28PM +0200, Jarkko Sakkinen wrote:
> > On Mon, Feb 22, 2016 at 10:52:45AM -0700, Jason Gunthorpe wrote:
> > > On Mon, Feb 22, 2016 at 04:50:23PM +0200, Jarkko Sakkinen wrote:
> > >
> > > > I already pu
de up of characters
in (or not in) the set. This implementation differs from its glibc
counterpart in that it does not support character ranges (e.g., 'a-z' or
'0-9'), the hyphen '-' is *not* a special character, and the brackets
themselves cannot be matched.
Signed-off
On Mon, Feb 22, 2016 at 11:15:09AM -0800, Doug Anderson wrote:
> Note that historically I remember that Linus Torvalds has stated that
> there is no stable API within the Linux kernel and that forcing the
> in-kernel API to never change was bad for software development. I
> tracked down my memory
On Mon, Feb 22, 2016 at 10:32 AM, Juri Lelli wrote:
> On 19/02/16 23:14, Rafael J. Wysocki wrote:
>> On Friday, February 19, 2016 08:09:17 AM Juri Lelli wrote:
>> > Hi Rafael,
>> >
>> > On 18/02/16 21:22, Rafael J. Wysocki wrote:
>> > > On Mon, Feb 15, 2016 at 10:47 PM, Rafael J. Wysocki
>> > >
From: "Steven Rostedt (Red Hat)"
To have nanosecond output displayed in a more human readable format, its
nicer to convert it to a seconds format (XXX.Y). The problem is that
to do so, the numbers must be divided by NSEC_PER_SEC, and moded too. But as
these numbers are 64 bit, this can no
From: "Steven Rostedt (Red Hat)"
The sched_domain_sysctl setup is only enabled when SCHED_DEBUG is
configured. As debug.c is only compiled when SCHED_DEBUG is configured as
well, move the setup of sched_domain_sysctl into that file.
Note, the (un)register_sched_domain_sysctl() functions had to b
[ Nothing different since my v2 I posted, except hopefully people
will look at it this time :-) I also removed the RFC ]
I'm starting to play with SCHED_DEADLINE a bit and I'm able to cause
a bandwidth "leak". Then I realized there's no way to examine what bandwidths
are enabled on which CPUs. I
From: "Steven Rostedt (Red Hat)"
Playing with SCHED_DEADLINE and cpusets, I found that I was unable to create
new SCHED_DEADLINE tasks, with the error of EBUSY as if the bandwidth was
already used up. I then realized there wa no way to see what bandwidth is
used by the runqueues to debug the issu
On Mon, Feb 22, 2016 at 02:57:42PM +0300, Sergei Ianovich wrote:
Please don't send new patches in the middle of old threads, it makes it
hard to find what the current versions of things are.
>v2..v3 to fix Mark Brown review comments
>* drop dependencies in Kconfig
No, what I was asking y
From: "Steven Rostedt (Red Hat)"
As sched_feature is only created when SCHED_DEBUG is enabled, and the file
debug.c is only compiled when SCHED_DEBUG is enabled, it makes sense to move
sched_feature setup into that file and get rid of the #ifdef.
Signed-off-by: Steven Rostedt
---
kernel/sched/
On Mon, Feb 22, 2016 at 12:48:54PM -0500, Steven Rostedt wrote:
> > As it stands, the existing tracepoint have already been an ABI
> > trainwreck, why would I want to add more?
>
> Yes, this may become a type of ABI, but even the sched switch
> tracepoints haven't been that bad. Has it really pre
The fsl-dcu driver copies a drm_mode_config object to its
stack but then only accesses a single member (dpms_property)
once. The data structure is large enough to trigger a warning
about the amount of kernel stack being used:
drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_rgb.c: In function
'fsl_dcu_drm_con
Am 2016-02-22 um 20:10 schrieb Takashi Iwai:
> On Mon, 22 Feb 2016 19:58:18 +0100,
> Martin Kepplinger wrote:
>>
>> Am 2016-02-22 um 15:12 schrieb Takashi Iwai:
>>> On Mon, 22 Feb 2016 15:02:56 +0100,
>>> Martin Kepplinger wrote:
> And how about my questions in the previous mail? Does
> i9
On Mon, 22 Feb 2016, Pantelis Antoniou wrote:
> Hi Rob,
>
> > On Feb 22, 2016, at 04:55 , Rob Herring wrote:
> >
> > On Wed, Feb 17, 2016 at 11:41:25AM -0600, Alan Tull wrote:
> >> Add overlay callback functionality.
> >>
> >> When DT overlays are being added, some drivers/subsystems
> >> will
On Mon, Feb 22, 2016 at 10:42 AM, Juri Lelli wrote:
> Hi Rafael,
>
> On 19/02/16 23:26, Rafael J. Wysocki wrote:
>> On Friday, February 19, 2016 05:26:04 PM Juri Lelli wrote:
>> > Hi Srinivas,
[cut]
>> ---
>> From: Rafael J. Wysocki
>> Subject: [PATCH] cpufreq: Rework the scheduler hooks for tr
On Tue, Feb 23, 2016 at 1:33 AM, Paul E. McKenney
wrote:
> On Mon, Feb 22, 2016 at 11:16:39AM +, David Howells wrote:
>> SeongJae Park wrote:
>>
>> > From f7b5677790771599f418f1d95536935be971ae86 Mon Sep 17 00:00:00 2001
>> > From: SeongJae Park
>> > Date: Mon, 22 Feb 2016 19:26:18 +0900
>>
On Mon, Feb 22, 2016 at 11:52 AM, Peter Zijlstra wrote:
> On Fri, Feb 19, 2016 at 09:28:23AM -0800, Steve Muckle wrote:
>> On 02/19/2016 08:42 AM, Srinivas Pandruvada wrote:
>> > We did experiments using util/max in intel_pstate. For some benchmarks
>> > there were regression of 4 to 5%, for some
We must not use ubi_* log functions before the ubi_device
struct is initialized.
And while we are here, define a sane pr_fmt and add new lines to
existing pr_* calls.
Fixes: 32608703 ("UBI: Extend UBI layer debug/messaging capabilities")
Signed-off-by: Richard Weinberger
---
drivers/mtd/ubi/buil
On 02/22, dingu...@opensource.altera.com wrote:
> From: Dinh Nguyen
>
> There are some Arria10 clocks of type "altr,socfpga-a10-perip-clk" that can
> have multiple parents. Fix up the __socfpga_periph_init() to call
> of_clk_parent_fill() that will return the appropriate number of parents.
>
> A
On 02/22, Stephen Boyd wrote:
> On 02/22, dingu...@opensource.altera.com wrote:
> > From: Dinh Nguyen
> >
> > There are some Arria10 clocks of type "altr,socfpga-a10-perip-clk" that can
> > have multiple parents. Fix up the __socfpga_periph_init() to call
> > of_clk_parent_fill() that will return
The newly added generic crypto abstraction for file systems operates
on 'struct bio' objects, which do not exist when CONFIG_BLOCK is
disabled:
fs/crypto/crypto.c: In function 'fscrypt_zeroout_range':
fs/crypto/crypto.c:308:9: error: implicit declaration of function 'bio_alloc'
[-Werror=implicit-
On 22 February 2016 at 15:50, Alexandre Torgue
wrote:
> 2016-02-13 14:48 GMT+01:00 Joachim Eastwood :
>> On 3 February 2016 at 15:54, Alexandre TORGUE
>> wrote:
>>> + plat_dat->bsp_priv = dwmac;
>>> + plat_dat->init = stm32_dwmac_init;
>>> + plat_dat->exit = stm32_dwmac_exit;
>
With CONFIG_SMP and CONFIG_PREEMPT both disabled, gcc decides
to partially inline the get_state_failrec() function but cannot
figure out that means the failrec pointer is always valid
if the function returns success, which causes a harmless
warning:
fs/btrfs/extent_io.c: In function 'clean_io_fail
A recent bugfix changed pfn_t to always be 64-bit wide, but did not
change the code in pmem.c, which is now broken on 32-bit architectures
as reported by gcc:
In file included from ../drivers/nvdimm/pmem.c:28:0:
drivers/nvdimm/pmem.c: In function 'pmem_alloc':
include/linux/pfn_t.h:15:17: error: l
From: Dinh Nguyen
There are some Arria10 clocks of type "altr,socfpga-a10-perip-clk" that can
have multiple parents. Fix up the __socfpga_periph_init() to call
of_clk_parent_fill() that will return the appropriate number of parents.
Also, update __socfpga_gate_init() to call of_clk_parent_fill()
Currently we only per-cgroup per-queue data. This adds per-cgroup data
(cgroup weight). Changing the per-cgroup weight will change all
per-cgroup per-queue weight.
Signed-off-by: Shaohua Li
---
block/blk-throttle.c | 47 ++-
1 file changed, 46 insertio
The trace log is very helpful to track bug and performance issues.
Signed-off-by: Shaohua Li
---
block/blk-throttle.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 01ca04e..68e2598 100644
--- a/block/blk-throttle.c
+++ b/block/blk-
If a cgroup share gets shrinked but the cgroup hits its limit, the
shrink isn't optimal, we restore its original share.
Signed-off-by: Shaohua Li
---
block/blk-throttle.c | 13 -
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
If we don't dispatch enough requests, disk can't reach the max
bandwidth. As we don't know the max bandwidth, we over estimate the
bandwidth and dispatch more requests. This way the disk can reach the
max bandwidth slowly. The downside is this can introduce fairness issue,
but since we only over es
Example usage of the cgroup2 interface:
$echo "+io" > /sys/fs/cgroup/cgroup.subtree_control
set bandwidth based proporation mode
$echo "8:0 weight_bw" > /sys/fs/cgroup/io.throttle.mode_device
$mkdir /sys/fs/cgroup/test
set cgroup weight
$echo "8:0 200" > /sys/fs/cgroup/test/io.throttle.weight
$echo
Last patch detects wrong shrink for some cases, but not all. Let's have
an example. cg1 gets 20% share, cg2 gets 80% share.
disk max bps 100, cgroup can only dispatch 80 bps
cg1 bps = 20, cg2 bps = 80
This should be stable state, but we don't know about it. As we assign
extra bps to queue, we wil
We know total bandwidth of a disk and can calculate cgroup's bandwidth
percentage against disk bandwidth according to its weight. We can easily
calculate cgroup bandwidth.
Signed-off-by: Shaohua Li
---
block/blk-throttle.c | 160 +--
1 file changed
Hi,
Currently we have 2 iocontrollers. blk-throttling is bandwidth/iops based. CFQ
is weight based. It would be great there is a unified iocontroller for the two.
And blk-mq doesn't support ioscheduler, leaving blk-throttling the only option
for blk-mq. It's time to have a scalable iocontroller su
Move io cost related data to separate data structure and corresponding
code to separate functions. In the future, we will seek more advanced
algorithm to measure IO cost. The separation makes things easier.
Signed-off-by: Shaohua Li
---
block/blk-throttle.c | 230 +++-
cgroup's share could be over estimated. For example, cg1 gets 20% share,
cg2 gets 80% share.
1. disk max bps 200, cgroup has no limitation
cg1 bps == 40, cg2 bps = 160
2. disk max bps 200, cg2 can only dispatch 10 bps
cg1 bps == 190 (not 2.5), cg2 bps == 10
cg2 can't use its share, so we must adj
We don't have a tree to maintain active cgroups. If a cgroup is inactive
for some time, it should be excluded from bandwidth calculation.
Otherwise we will only assign partial bandwidth to active cgroups,
active cgroups will dispatch less IO, the estimated queue bandwidth
drops, which will cause ac
weight based blk-throttling can use the estimated performance to
calculate cgroup bandwidth/IOPS. That said we never can't get accurate
disk performance in a complex workload, a roughly estimation is enough.
One issue is workload might not send enough IO to make disk 100% busy.
The calculation sho
There is throttle.mode_device interface. By default blk-throttle is in
NONE mode. Setting original bps/iops limit wil change mode to THROTTLE
automatically, user doesn't need to configure the mode, which is for
backward compabitbility. To use proporation based throttle, user must
configure device t
For throttle, we have separate data to track READ/WRITE IO. For weight
based throttle, we combine read/write bandwidth and check combined
bandwidth. Separate throttle doesn't work, because it's possible one
cgroup only does READ and the other only does WRITE. This patch adds
an abstract to index da
On 02/22/2016 11:27 AM, Kees Cook wrote:
On Fri, Feb 19, 2016 at 3:07 PM, Laura Abbott wrote:
On 02/19/2016 02:19 PM, Kees Cook wrote:
On Fri, Feb 19, 2016 at 2:11 PM, Laura Abbott wrote:
On 02/19/2016 11:12 AM, Kees Cook wrote:
On Thu, Feb 18, 2016 at 5:15 PM, Laura Abbott
wrote:
Baremetal kernels clear .bss early in the boot. Since Xen PV guests don't
excecute that early code they should do it too.
(Since we introduce macros for specifying 32- and 64-bit registers we
can get rid of ifdefs in startup_xen())
Signed-off-by: Boris Ostrovsky
---
arch/x86/xen/xen-head.S | 29
On Mon 22 Feb 02:03 PST 2016, Srinivas Kandagatla wrote:
>
>
> On 22/02/16 05:32, Bjorn Andersson wrote:
> >On certain platforms (e.g. ARM64) the dma_ops needs to be explicitly set
> >to be able to do DMA allocations, so use the of_dma_configure() helper
> >to populate the dma properties and ass
On Wed, Jan 20, 2016 at 04:01:11PM -0600, Bjorn Helgaas wrote:
> On Sat, Jan 16, 2016 at 02:19:38PM -0800, Veal, Bryan E. wrote:
> > On Fri, Jan 15, 2016 at 03:49:02PM -0600, Bjorn Helgaas wrote:
> > > Even though you found this issue before posting the RFC code, I assume
> > > the issue is still r
Hi Dinh,
[auto build test WARNING on clk/clk-next]
[also build test WARNING on v4.5-rc5 next-20160222]
[if your patch is applied to the wrong git tree, please drop us a note to help
improving the system]
url:
https://github.com/0day-ci/linux/commits/dinguyen-opensource-altera-com/clk
On Fri, Feb 19, 2016 at 07:34:18PM -0600, John Dahlstrom wrote:
> Some Lenovo ideapad models lack a physical rfkill switch.
> On Lenovo models ideapad Y700 Touch-15ISK and ideapad Y700-15ISK,
> ideapad-laptop would wrongly report all radios as blocked by
> hardware which caused wireless network con
On 02/22/2016 03:50 PM, Stephen Boyd wrote:
>
> Oops, spoke too soon:
>
> drivers/clk/socfpga/clk-periph-a10.c:113:27: warning: incorrect type in
> assignment (different base types)
> drivers/clk/socfpga/clk-periph-a10.c:113:27:expected char const *const
> *[assigned] parent_names
> drivers
On 02/22, dingu...@opensource.altera.com wrote:
> From: Dinh Nguyen
>
> There are some Arria10 clocks of type "altr,socfpga-a10-perip-clk" that can
> have multiple parents. Fix up the __socfpga_periph_init() to call
> of_clk_parent_fill() that will return the appropriate number of parents.
>
> A
On Tue 23-02-16 08:08:14, Dave Chinner wrote:
> On Mon, Feb 22, 2016 at 02:04:35PM +0100, Jan Kara wrote:
> > On Mon 22-02-16 13:12:22, Peter Zijlstra wrote:
> > > On Mon, Feb 22, 2016 at 12:54:35PM +0100, Jan Kara wrote:
> > > > > Also, I think fsnotify_unmount_inodes() (as per mainline) is missin
The latest maintenance release Git v2.7.2 is now available at
the usual places.
The tarballs are found at:
https://www.kernel.org/pub/software/scm/git/
The following public repositories all have a copy of the 'v2.7.2'
tag and the 'maint' branch that the tag points at:
url = https://kernel
On 02/22/2016 12:48 PM, Yu-cheng Yu wrote:
> It should have been:
>
> xsave->header.xfeatures = xfeatures |
> (xsave->header.xfeatures & XFEATURE_MASK_SUPERVISOR);
>
> I'll fix it.
Can we break it out to make it more clear?
/*
* The state that came in from userspace was use
On Mon, Feb 22, 2016 at 1:58 PM, Arnd Bergmann wrote:
> A recent bugfix changed pfn_t to always be 64-bit wide, but did not
> change the code in pmem.c, which is now broken on 32-bit architectures
> as reported by gcc:
>
> In file included from ../drivers/nvdimm/pmem.c:28:0:
> drivers/nvdimm/pmem.
The only missing bit is to completely clear the hardware state on failure
exit. This is now a pretty simple exercise.
Undo the box->init_box() setup on all packages which have been initialized so
far.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c | 26 ++
Like uncore the rapl driver lacks error handling. It leaks memory and leaves
the hotplug notifier registered.
Add the proper error checks, cleanup the memory and register the hotplug
notifier only on success.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 29
Required to use it in modular perf drivers.
Signed-off-by: Thomas Gleixner
---
kernel/events/core.c |1 +
1 file changed, 1 insertion(+)
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -9335,6 +9335,7 @@ ssize_t perf_event_sysfs_show(struct dev
return 0;
}
+EXPORT_SYMBOL
RAPL is a per package facility and we already have a mechanism for a dedicated
per package reader. So there is no point to have multiple CPUs doing the
same. The current implementation actually starts two timers on two cpus if one
does:
perf stat -C1,2 -e -e power/energy-pkg
which ma
This lock is taken in hard interrupt context even on Preempt-RT. Make it raw
so RT does not have to patch it.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 20 ++--
1 file changed, 10 insertions(+), 10 deletions(-)
Index: b/arch/x86/kernel/cp
Add the necessary exit functions so it can be built as a module.
Signed-off-by: Thomas Gleixner
---
arch/x86/Kconfig.perf | 13 +
arch/x86/kernel/cpu/Makefile|4 +++-
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 22 +++---
Split out code from init into seperate functions. Tidy up the code and get rid
of pointless comments. I wish there would be comments for code which is not
obvious
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 61 ++--
1 file change
Store the pmu in event->pmu_private and use it instead of the per cpu
data. Preparatory step to get rid of the per cpu allocations. The usage sites
are the perf fast path, so we keep that even after the conversion to per
package storage as a cpu to package lookup involves 3 loads versus 1 with the
The Knights Landings support added the events and the detection case, but then
returns 0 without actually initializing the driver.
Fixes: 3a2a7797326a4 "perf/x86/intel/rapl: Add support for Knights Landing
(KNL)"
Signed-off-by: Thomas Gleixner
Cc: Dasaratharaman Chandramouli
---
arch/x86/kerne
The output is inconsistent. Use a proper pr_fmt prefix and split out the
advertisement into a seperate function.
Remove the WARN_ON in the failure case. It's pointless as we already know
where it failed.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 42 +
There is no point in having a quirk machinery for a single possible
function. Get rid of it and move the quirk to a place where it makes actually
sense.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 49 ++--
1 file changed, 19 insertio
No point in doing the same calculation over and over. Do it once in
rapl_check_hw_unit().
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_rapl.c | 39
1 file changed, 18 insertions(+), 21 deletions(-)
Index: b/arch/x86/kernel/cpu/perf_event
CQM is a strict per package facility. Use the proper cpumasks to lookup the
readers.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_cqm.c | 34 ++---
1 file changed, 12 insertions(+), 22 deletions(-)
Index: b/arch/x86/kernel/cpu/perf_event_inte
Now that we have a proper cleanup all over the place, it's simple to make this
a modular driver.
Signed-off-by: Thomas Gleixner
---
arch/x86/Kconfig |6 ++
arch/x86/Kconfig.perf | 16
arch/x86/kernel/cpu/Makefile
uncore is a per package facility, but the code tries to mimick a per cpu
facility with completely convoluted constructs.
Simplify the whole machinery by tracking per package information. While at it,
avoid the kfree/alloc dance when a cpu goes offline and online again. There is
no point in freeing
Almost every cpumask function is exported, just not the one I need to make the
intel uncore driver modular.
Signed-off-by: Thomas Gleixner
---
lib/cpumask.c |1 +
1 file changed, 1 insertion(+)
Index: b/lib/cpumask.c
===
--- a/
Andi wanted to do this before, but the patch fell down the cracks. Implement
it with the proper error handling.
Requested-by: Andi Kleen
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c | 34 +-
1 file changed, 18 insertions(+), 16 delet
For pmus which are not per cpu, but e.g. per package/socket, we want to be
able to store a reference to the underlying per package/socket facility in the
event at init time so we can avoid magic storage constructs in the pmu driver.
This allows us to get rid of the per cpu dance in the intel uncor
This driver lacks any form of proper error handling. If initialization fails
or hotplug prepare fails, it lets the facility with half initialized stuff
around.
Fix the state and memory leaks in a first step. As a second step we need to
undo the hardware state which is set via uncore_box_init() on
Store the pmu in event->pmu_private, so we can get rid of the per cpu data
storage. We keep it after converting to per package data, because a cpu to
package lookup will be 3 loads versus one and these usage sites are the perf
fast pathes.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/p
For per package oriented services we must be able to rely on the number of cpu
packages to be within bounds. Create a tracking facility, which
- calculates the number of possible packages depending on nr_cpu_ids after
boot
- makes sure that the package id is within the number of possible packag
This series addresses the following issues:
- Add proper error handling to uncore and rapl drivers
- Get rid of the pseudo per cpuness of these drivers and do a proper per
package storage
- Allow them to be modular
In order to do proper per package storage I added a facility which sanity
No users outside of this file.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c |2 +-
arch/x86/kernel/cpu/perf_event_intel_uncore.h |1 -
2 files changed, 1 insertion(+), 2 deletions(-)
Index: b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
=
On Sun, Feb 21, 2016 at 03:22:27PM +0100, Wolfram Sang wrote:
> We want the size of the struct, not of a pointer to it. To be future
> proof, just dereference the pointer to get the desired type.
>
> Signed-off-by: Wolfram Sang
Queued to testing. Thank you Wolfram.
> ---
>
> Compile tested onl
Cleanup the code a bit before reworking it completely.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c | 95 +-
1 file changed, 50 insertions(+), 45 deletions(-)
Index: b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
===
No point in doing partial rollbacks. Robustify uncore_exit_type() so it does
not dereference type->pmus unconditionally and remove all the partial rollback
hackery.
Preparatory patch for proper error handling.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c |
The perf core doesn't do anything related to hardware pmus on hotplug. Drop
the priority.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c |5 -
1 file changed, 5 deletions(-)
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c
+++ b/arch/x86/kernel/cpu/p
uncore_cpumask_init() is only ever called from intel_uncore_init() where the
mask is guaranteed to be empty.
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c |6 --
1 file changed, 6 deletions(-)
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c
+++ b/a
Tadeusz Struk wrote:
> I wonder if this should be merged with the crypto/rsa-pkcs1pad.c template
> that we already have. Looks like the two do the same padding now.
> Should we merge then and pass the hash param as a separate template param,
> e.g the public_key would allocate "pkcs1pad(rsa, sha1
The storage array is size limited, but misses a sanity check
Signed-off-by: Thomas Gleixner
---
arch/x86/kernel/cpu/perf_event_intel_uncore.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
=
When tearing down the boxes nothing undoes the hardware state which was setup
by box->init_box(). Add a box->exit_box() callback and implement it for the
uncores which have an init_box() callback.
This misses the cleanup in the error exit pathes, but I cannot be bothered to
implement it before cle
Mimi Zohar wrote:
> > (1) - (3) These are Tadeusz's RSA akcipher conversion.
>
> Up to here, IMA-appraisal works properly.
I don't have IMA set up anywhere.
David
From: Frank Zago
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: Frank Zago
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/13320
Reviewed-by: James Simmons
Reviewed-by:
From: John L. Hammond
Remove LUSTRE_LNET_PID (12354) and LUSTRE_SRV_LNET_PID (12345) from
the libcfs headers and replace their uses with a new macro
LNET_PID_LUSTRE (also 12345) in lnet/types.h.
Signed-off-by: John L. Hammond
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on:
From: Amir Shehata
This is the fifth patch of a set of patches that enables DLC.
This patch adds the new structures which will be used
in the IOCTL communication. It also added a set of
show operations to show buffers, networks, statistics
and peer information.
Signed-off-by: Amir Shehata
Int
This patch set fixes many of the LNet issues encounter run in
production environments. One of the long standing issues was
not being able to reconfigure LNet after initialization. Doing
so left it in a broken state. Several other issues are also
addressed in this patch set. Merged back into this pa
From: Doug Oucharek
In lnet_router_checker(), there are two LASSERTS. Neither protects
us from anything and one of them triggered for a customer crashing
the system unecessarily. This patch removes them.
Signed-off-by: Doug Oucharek
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7362
Rev
On Mon, 22 Feb 2016 22:30:17 +0100
Peter Zijlstra wrote:
> On Mon, Feb 22, 2016 at 12:48:54PM -0500, Steven Rostedt wrote:
>
> > > As it stands, the existing tracepoint have already been an ABI
> > > trainwreck, why would I want to add more?
> >
> > Yes, this may become a type of ABI, but eve
From: Bruno Faccini
This patch will avoid potential race, around socket sleepers
wait list, during acceptor thread termination and using
sk_callback_lock RW-Lock protection.
Signed-off-by: Bruno Faccini
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6476
Reviewed-on: http://review.whamclou
From: Chris Horn
lctl show_route and lctl route_list will output router aliveness
information via lnet_get_route(). lnet_get_route() should use the
lnet_is_route_alive() function, introduced in e8a1124
http://review.whamcloud.com/7857, to determine route aliveness.
Signed-off-by: Chris Horn
Int
On Mon, Feb 22, 2016 at 02:19:05PM -0800, Dave Hansen wrote:
> On 02/22/2016 12:48 PM, Yu-cheng Yu wrote:
> > It should have been:
> >
> > xsave->header.xfeatures = xfeatures |
> > (xsave->header.xfeatures & XFEATURE_MASK_SUPERVISOR);
> >
> > I'll fix it.
>
> Can we break it out
901 - 1000 of 1230 matches
Mail list logo
