Add a testcase for types of kprobe event. This checks
kprobe event can accept and correctly expressed the
arguments typed as s32, u32, x32 and bitfield.
Here is the test result.
-
# ./ftracetest test.d/kprobe/kprobe_args_type.tc
=== Ftrace unit tests ===
[1] Kprobes event arguments wit
Reset ftrace to initial state before running each test.
This fixes some test cases to enable tracing before starting
trace test. This can avoid false-positive failure when
previous testcase fails while disabling tracing.
Signed-off-by: Masami Hiramatsu
Suggested-by: Steven Rostedt
---
tools/tes
Introduce TMPDIR variable which is removed after each test
is done, so that the test script can put their temporary
files in that.
Signed-off-by: Masami Hiramatsu
---
tools/testing/selftests/ftrace/ftracetest |2 ++
1 file changed, 2 insertions(+)
diff --git a/tools/testing/selftests/ftrace
Add function filter glob matching test case.
This checks whether the kernel supports glob matching
(front match, end match, middle match, side match,
character class and '?').
Here is the test result.
-
./ftracetest test.d/ftrace/func-filter-glob.tc
=== Ftrace unit tests ===
[1] ftrace
Hide ftracetest result log directory from git.
Signed-off-by: Masami Hiramatsu
---
tools/testing/selftests/ftrace/.gitignore |1 +
1 file changed, 1 insertion(+)
create mode 100644 tools/testing/selftests/ftrace/.gitignore
diff --git a/tools/testing/selftests/ftrace/.gitignore
b/tools/tes
Hi,
Here is the 2nd version of the series of patches to
add full glob matching support in ftrace and update
ftracetest to test the glob matching and hexadecimal
types.
Previous version is here:
https://lkml.org/lkml/2016/10/19/7
I've update ftracetest to initialize ftrace before
each testcase a
Use glob_match() to support flexible glob wildcards (*,?)
and character classes ([) for ftrace.
Since the full glob matching is slower than the current
partial matching routines(*pat, pat*, *pat*), this leaves
those routines and just add MATCH_GLOB for complex glob
expression.
e.g.
[root@loca
Hi Linus,
It's probably a bug fix that unveils the link errors.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: 2a26d99b251b8625d27aed14e97fc10707a3a81f
commit: 2527ecc9195e9c66252af24c4689e8a67cd4ccb9 gpio: Fix OF build problem on
UM
date: 2 months a
ase=auto for
convenience) to record what (public, well-known) commit your patch series was
built on]
[Check https://git-scm.com/docs/git-format-patch for more information]
url:
https://github.com/0day-ci/linux/commits/Tushar-Dave/sparc-Enable-sun4v-hypervisor-PCI-IOMMU-v2-APIs-and-ATU/20161
On Sat, Oct 29, 2016 at 05:12:30PM +0100, Al Viro wrote:
> Eww... IOW, as soon as we'd submitted an async iocb, nobody can so much as
> look at struct file *or* iocb, right? Or underlying inode, or any fs-private
> data structures attached to it...
Yeah.
> I certainly agree that it's a bug, but
On Sat, Oct 29, 2016 at 10:47:58AM -0700, Linus Torvalds wrote:
> On Sat, Oct 29, 2016 at 8:20 AM, Christoph Hellwig wrote:
> >
> > We can't as that would not fix the use after free (at least for the lockdep
> > case - otherwise the call is a no-op). Once iter_op returns aio_complete
> > might ha
Hi Paul,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: 2a26d99b251b8625d27aed14e97fc10707a3a81f
commit: cf2a5e0bb4c66e8c43caf9f1be93a1bd7fd07b17 MIPS: Support generating
Flattened Image Trees (.itb)
date: 3 week
On Sun, 30 Oct 2016 10:06:20 +0800
Chen-Yu Tsai wrote:
> >> Yes, I know that the burst size is always a power of 2.
> >> The best way to check it would be to change the {src,dts}_maxburst to a
> >> bitmap of the possible bursts as 0x0d for 1,4 and 8 bytes. But this
> >> asks for a lot of changes.
Hi Joel,
[auto build test ERROR on linus/master]
[also build test ERROR on v4.9-rc3 next-20161028]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
[Suggest to use git(>=2.9.0) format-patch --base= (or --base=auto for
convenience) to record what
Hi Pete,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: 2a26d99b251b8625d27aed14e97fc10707a3a81f
commit: d0b73b488c55df905ea8faaad079f8535629ed26 xtensa: Add config files for
Diamond 233L - Rev C processor variant
Hi Dave,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: 2a26d99b251b8625d27aed14e97fc10707a3a81f
commit: 85d1225ec066b2ef46fbd0ed1bae78ae1f3e6c91 drm/i915: Introduce & use new
lightweight SGL iterators
date: 5 mo
Hi Dave,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: 2a26d99b251b8625d27aed14e97fc10707a3a81f
commit: ab9d1e4f7b0217948a3b35a64178602ab30ff45d Merge branch
'xfs-misc-fixes-4.6-3' into for-next
date: 8 months a
Hi Will,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: a909d3e636995ba7c349e2ca5dbb528154d4ac30
commit: da48d094ce5d7c7dcdad9011648a81c42fd1c2ef Kconfig: remove
HAVE_LATENCYTOP_SUPPORT
date: 10 months ago
config
Sorry, the warning is still there.
I'm not sure adding sched_annotate_sleep() does anything, since it's
defined as (in case CONFIG_DEBUG_ATOMIC_SLEEP is not set):
# define sched_annotate_sleep() do { } while (0)
On Sat, Oct 29, 2016 at 8:05 PM, Eric Dumazet wrote:
> On Sat, 2016-10-29 at 19:59 +
> I think we should do those two things completely independently.
> We need to do something now to preserve the current interfaces
> for the glibc changes that are coming soon [1], and Deepa's
> patches do that (though I now realize the changelog doesn't
> mention the requirement).
I'll update the
'blk_mq_alloc_request()' returns an error pointer in case of error, not
NULL. So test it with IS_ERR.
Fixes: fd8383fd88a2 ("nbd: convert to blkmq")
Signed-off-by: Christophe JAILLET
---
drivers/block/nbd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/block/nbd.c
> btw, where did you post the libevdev patch? I haven't seen it anywhere I'm
> subscribed to.
The libevdev patch was posted to input-to...@lists.freedesktop.org :
https://www.mail-archive.com/y2038@lists.linaro.org/msg01824.html
-Deepa
On 10/29/2016 10:10 AM, Wei Yongjun wrote:
From: Wei Yongjun
Fixes the following sparse warning:
drivers/block/brd.c:411:15: warning:
symbol 'rd_size' was not declared. Should it be static?
We just tried this, it's used in arm:
arch/arm/kernel/atags_parse.c
static int __init parse_tag_ram
On Sat, 29 Oct 2016, Eric Wheeler wrote:
Kent, Davidlohr, does 3312845 need to land in 4.9 for some reason?
No, not particularly. I can do my stuff whenever this is ends up being merged.
Thanks,
Davidlohr
Hi Paul,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: a909d3e636995ba7c349e2ca5dbb528154d4ac30
commit: cf2a5e0bb4c66e8c43caf9f1be93a1bd7fd07b17 MIPS: Support generating
Flattened Image Trees (.itb)
date: 3 week
> Il giorno 29 ott 2016, alle ore 16:12, Jens Axboe ha
> scritto:
>
> On 10/28/2016 11:38 PM, Paolo Valente wrote:
>>
>>> Il giorno 26 ott 2016, alle ore 18:12, Jens Axboe ha
>>> scritto:
>>>
>>> On 10/26/2016 10:04 AM, Paolo Valente wrote:
> Il giorno 26 ott 2016, alle ore 17:32,
Hi Vincent,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: a909d3e636995ba7c349e2ca5dbb528154d4ac30
commit: 3194c6870158e305dac2af52f83681e9cb67280f NFC: nfcmrvl: add firmware
download support
date: 1 year ago
co
From: Stefan Richter
Date: Sat, 29 Oct 2016 22:16:58 +0200
> The maximum unicast datagram size /without/ link fragmentation is
> 4096 - 4 = 4092 (max IEEE 1394 async payload size at >= S800 bus speed,
> minus unfragmented encapssulation header). Max broadcast datagram size
> without fragmentatio
Hi Guenter,
First bad commit (maybe != root cause):
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
head: a909d3e636995ba7c349e2ca5dbb528154d4ac30
commit: f9f3f864b5e8c09d7837d8980edba4ad52969819 cris: Fix section mismatches
in architecture startup code
date:
Hi Joel,
[auto build test ERROR on linus/master]
[also build test ERROR on v4.9-rc3 next-20161028]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
[Suggest to use git(>=2.9.0) format-patch --base= (or --base=auto for
convenience) to record what
Sorry, forgot to CC: the lists on this one, doh!
--- Begin Message ---
Lots of fixes, mostly drivers as is usually the case.
1) Don't treat zero DMA address as invalid in vmxnet3, from Alexey
Khoroshilov.
2) Fix element timeouts in netfilter's nft_dynset, from Anders
K. Pedersen.
3) Don'
Hi Joel,
[auto build test ERROR on linus/master]
[also build test ERROR on v4.9-rc3 next-20161028]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
[Suggest to use git(>=2.9.0) format-patch --base= (or --base=auto for
convenience) to record what
Hi,
On Fri, Oct 28, 2016 at 1:51 AM, Maxime Ripard
wrote:
> On Sun, Oct 23, 2016 at 06:31:07PM +0200, Jean-Francois Moine wrote:
>> On Tue, 4 Oct 2016 18:55:54 +0200
>> Maxime Ripard wrote:
>>
>> > On Tue, Oct 04, 2016 at 12:40:11PM +0200, Jean-Francois Moine wrote:
>> > > On Tue, 4 Oct 2016 11
On Thu, 27 Oct 2016, Jens Axboe wrote:
> On 10/27/2016 05:27 PM, Eric Wheeler wrote:
> > Hi Jens,
> >
> > Please pull this v4.9-rc2 based series of bcache updates for v4.9-rc3:
> > (You may disregard the previous -rc1-based request.)
> >
> > git pull https://bitbucket.org/ewheelerinc/linux.git
>
The Kconfig currently controlling compilation of this code is:
drivers/mfd/Kconfig:config AB8500_GPADC
drivers/mfd/Kconfig:bool "ST-Ericsson AB8500 GPADC driver"
...meaning that it currently is not being built as a module by anyone.
Lets remove the modular code that is essentially orphaned,
The Kconfig currently controlling compilation of this code is:
drivers/mfd/Kconfig:config AB8500_DEBUG
drivers/mfd/Kconfig: bool "Enable debug info via debugfs"
...meaning that it currently is not being built as a module by anyone.
Lets remove the modular code that is essentially orphaned,
The Kconfig currently controlling compilation of this code is:
drivers/mfd/Kconfig:config AB3100_CORE
drivers/mfd/Kconfig:bool "ST-Ericsson AB3100 Mixed Signal Circuit core
functions"
...meaning that it currently is not being built as a module by anyone.
Lets remove the modular code that is
The Kconfig currently controlling compilation of this code is:
drivers/mfd/Kconfig:config ABX500_CORE
drivers/mfd/Kconfig:bool "ST-Ericsson ABX500 Mixed Signal Circuit register
functions"
...meaning that it currently is not being built as a module by anyone.
Lets remove the couple traces of
Nothing new here ; just another instance where we had modular remove
and exit functions -- but the Makefile/Kconfig limited the use case to
being always built-in; hence the code is useless and needs removing.
So we remove the dead code and the misleading hints of modularity from
these drivers, hop
The Kconfig currently controlling compilation of this code is:
drivers/mfd/Kconfig:config AB8500_CORE
drivers/mfd/Kconfig:bool "ST-Ericsson AB8500 Mixed Signal Power Management
chip"
...meaning that it currently is not being built as a module by anyone.
Lets remove the modular code that is
The Kconfig currently controlling compilation of this code is:
drivers/mfd/Kconfig:config AB8500_CORE
drivers/mfd/Kconfig:bool "ST-Ericsson AB8500 Mixed Signal Power Management
chip"
...meaning that it currently is not being built as a module by anyone.
Lets remove the couple traces of modu
Hi Paul,
[auto build test ERROR on linus/master]
[also build test ERROR on v4.9-rc3 next-20161028]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
[Suggest to use git(>=2.9.0) format-patch --base= (or --base=auto for
convenience) to record what
CMT name has superseded CQM in Intel's documentation.
Rename cpufeatures. Next patches in this series will use the CMT name.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/include/asm/cpufeatures.h | 14 +++---
arch/x86/kernel/cpu/common.c | 10 +-
2 files changed, 12
Completely remove previous version of CQM + MBM driver to ease
review of new version (this patch series).
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/Makefile |2 +-
arch/x86/events/intel/cqm.c| 1766
include/linux/cpuhotplug.
New flag for building drivers for Intel RDT Monitoring
(currently CMT, future: MBM).
This driver may be converted to module once the hooks in
perf cgroup are discussed.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/Kconfig | 12
1 file changed, 12 insertions(+)
diff --git a/
Per-package locks potentially reduce the contention when compared to the
system-wide approach of the previous CQM/CMT driver.
Lockdep needs lock_class_key's to be statically initialized and/or use
nesting, but nesting is currently hard-coded for up to 8 levels and it's
fragile to depend on lockdep
First part of intel_cmt_event_read. Error conditions and placeholder
for oncoming chunks.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 25 +
1 file changed, 25 insertions(+)
diff --git a/arch/x86/events/intel/cmt.c b/arch/x86/events/intel/cmt.
Add root for monr hierarchy and auxiliary functions for locking.
Also, add support for attaching CPU and tasks events to monr hierarchy.
As of this patch, both types of events always use the root monr (this
will change when cgroups are introduced later in this series).
More details in code's comm
Introduce struct pkg_data to store per CPU package locks and data for
new CMT driver.
Each pkg_data is initialiazed/terminated on demand when the first/last CPU
in its package goes online/offline.
More details in code's comments.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel
A pmonr is the per-package component of a monr. This patch only adds
initialization/destruction of pmonrs. Future patches explain their
usage and add functionality.
CPU hotplug is supported by initializing/terminating all pmonrs in
monr hierarchy when first/last CPU in package goes online/offline.
uflags allow users to signal special behavior for a pmonr. This patch
series introduces two uflags that provide new behavior and are relevant
to users:
1) CMT_UF_NOLAZY_RMID: signal that rmids must be reserved immediately.
2) CMT_UF_NOSTEAL_RMID: rmids cannot be stolen.
A monr mantains one fie
Add pmu, llc_occupancy event attributes, and functions for event
initialization.
Empty pmu functions will be filled in future patches in this series.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 106
1 file changed, 106 in
Events use monrs to monitor CMT events. Events that share their
monitoring target (same thread or cgroup) share a monr.
This patch introduces monrs and adds support for monr creation/destruction.
An event's associated monr is referenced by event->cmt_monr (introduced
in previous patch).
monr->mo
A Resource Monitoring ID (RMID) is a hardware ID used to track cache
occupancy and memory bandwidth. The rmids are a per-package resource
and only one can be programmed at a time per logical CPU.
This patch series creates per-package rmids pools and (default)
lazy allocation of rmids (only reserve
Add attributes in hw_perf_event that are required by CMT and,
in the future, MBM.
Signed-off-by: David Carrillo-Cisneros
---
include/linux/perf_event.h | 6 ++
1 file changed, 6 insertions(+)
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 345ec20..0202b32 100644
This uflag allows user to specify that a rmid must be allocated at monr's
initialization or fail otherwise.
For this to work we split __pmonr_apply_uflags into reserve and apply
modes. The reserve mode will try to reserve a free rmid, and if successful,
the apply mode can proceed using the rmid pr
Encapsulate updates for PQR_ASSOC msr's RMID and CLOSID into
intel_rdt_common. Use new interface in Intel CMT.
Change RDT common code to build for both CONFIG_INTEL_RDT_A and
CONFIG_INTEL_RDT_M.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 6 ++
arch
First part of cgroup support for CMT.
A monr's position in monrs hierarchy depends on the position of it's
target cgroup or thread in the cgroup hierarchy.
(See code comments for details).
A monr that monitors a cgroup keeps a reference to in monr->monr_cgroup
and its used in future patches to ad
A pmonr uses a state machine to keep track of its rmids and their
hierarchical dependency with other pmonrs in the same package.
This patch introduces the first two states in such state machine.
It also adds pmonr_rmids: a word-size container to atomically access
a pmonr's sched and read rmids.
M
Start/stop cgroup monitoring for cgroups when intel_cmt device is
started/stopped.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 99 +
1 file changed, 99 insertions(+)
diff --git a/arch/x86/events/intel/cmt.c b/arch/x86/even
Reading CMT/MBM task events in intel_cmt poses a challenge since it
requires to read from multiple sockets (usually accomplished with
an IPI) while called with interruptions disabled.
The current usptream driver avoids the problematic read with
interruptions disabled by not making a dummy perf_eve
This hook allows architecture specific code to be called right after
perf_events' context switch but before the scheduler lock is released.
It will serve two uses in this patch series:
1) Calls CMT's cgroup context switch code that update the current RMID
when no perf event is active (in conti
This is the first patch for cgroup support in this series.
Adds a new field to perf_cgroup that is used by intel_cmt pmu to
associate a monr with a perf_cgroup instance.
Signed-off-by: David Carrillo-Cisneros
---
include/linux/perf_event.h | 4 +++-
kernel/events/core.c | 2 ++
2 files ch
The new pmu::event_terminate allows a PMU to access an event's cgroup
before the event is torn down.
Used in CMT to dettach the cgroup from a monr before perf clears
event->cgrp.
Signed-off-by: David Carrillo-Cisneros
---
include/linux/perf_event.h | 6 ++
kernel/events/core.c | 4 +++
The change from event_destroy to event_terminate allows to check
if the event is a event cgroup during event destuction, before
generic code clears event->cgrp.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff
Now that the pmonr state machine and pqr_common are in place, add
pmonr_update_sched_rmid to find the adequate rmid to use. With it,
complete the body of PMUs functions that start/stop, add/del events.
A pmonr in Unused state tries to allocate a free rmid the first time one
of its monitored thread
A llc_occupancy read for a cgroup event must read llc_occupancy for all
monrs in or below the event's monr.
The cgroup's monr's pmonr must have a valid rmid for the read to be
meaningful. Descendant pmonrs that do not have a valid read_rmid are
skipped since their occupancy is already included in
Add remaining states for pmonr's state machine:
- Active: A pmonr that is actively used.
- Dep_Idle: A pmonr that failed to obtain a rmid. It "borrows" its rmid
from its lowest monitored (Active in same pkgd) ancestor in the
monr hierarchy.
- Dep_Dirty: A pmonr that was Active but has
Expose the attribute intel_cmt.cont_monitoring to perf cgroups using
the newly introduced hook PERF_CGROUP_ARCH_CGRP_SUBSYS_ATTS.
The format of new attribute is a semi-colon separated list of per-package
hexadecimal flags with missing an empty string assigned to zero.
echo "1;2" > g1/intel_cmt.
Introduce two new PERF_EV_CAP_READ capabilities to save unnecessary IPIs.
Since PMU hw keeps track of rmids at all times, both capabilities in this
patch allow to read events even when inactive.
These capabilities also remove the need to read the value of an event on
pmu->stop (already baked in in
On 10/29/2016 06:49 AM, Greg Kroah-Hartman wrote:
This is the start of the stable review cycle for the 4.4.29 release.
There are 51 patches in this series, all will be posted as a response
to this one. If anyone has any issues with these being applied, please
let me know.
Responses should be ma
Use newly added flag.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 8
1 file changed, 8 insertions(+)
diff --git a/arch/x86/events/intel/cmt.c b/arch/x86/events/intel/cmt.c
index 614b2f4..194038b 100644
--- a/arch/x86/events/intel/cmt.c
+++ b/arch/x86/events
The generic code handles cgroup hierarchy by adding to the PMU the events
of all the ancestor cgroups of the cgroup to read.
This approach is incompatible with the CMT hw that only allows one rmid
per virtual core at a time. CMT's PMU work-arounds this limitation by
internally maintaining the hiera
Add a new enumeration to store monr's internal state information.
This new monr->flags contrasts with monr->uflag that stores user
provided flags.
The first flags is CMT_MONR_ZOMBIE, used to signal that the monr is no
longer valid, yet, must be kept around until its released by its last
user. This
Use newly introduced architecture specific hooks to update monr hierarchy
when a cgroup goes online/offline.
Add cmt_initialized_key and use it to avoid using the hooks when the
intel_cmt is not active.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 36 ++
Allow architectures to define additional attributes for the perf cgroup.
Signed-off-by: David Carrillo-Cisneros
---
include/linux/perf_event.h | 4
kernel/events/core.c | 2 ++
2 files changed, 6 insertions(+)
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index
On 10/29/2016 06:48 AM, Greg Kroah-Hartman wrote:
This is the start of the stable review cycle for the 4.8.6 release.
There are 125 patches in this series, all will be posted as a response
to this one. If anyone has any issues with these being applied, please
let me know.
Responses should be ma
Use new flags in CMT pmu.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/arch/x86/events/intel/cmt.c b/arch/x86/events/intel/cmt.c
index 275d128..614b2f4 100644
--- a/arch/x86/events/intel/cmt.c
+++ b/arch/x86/e
In order to support intel_cmt for cgroups with no perf_event, this driver
calls intel_pqr_ctx_switch during a task context switch, after perf_event
has add/del all events (if any), using the newly introduced
finish_arch_pre_lock_switch() .
If pmu->add sets a rmids, the next rmid is "marked" using
Allow user to specify that rmids used by a cgroup/event cannot be stolen.
The pmonrs that are marked with CMT_UF_NOSTEAL_RMID will never lose a valid
rmid once they have receive one.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 10 +-
arch/x86/events/intel/cm
Expose max_recycle_threshold as a configurable sysfs attribute of
intel_cmt pmu. max_recycle_threshold determines the maximum dirty
threshold for rmids with no zero occupancy.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 57
Try to activate monrs at a __cmt_min_progress_rate rate.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 274 +++-
1 file changed, 273 insertions(+), 1 deletion(-)
diff --git a/arch/x86/events/intel/cmt.c b/arch/x86/events/intel/c
Schedule rotation work each pmu->hrtimer_interval_ms ms.
Default to CMT_DEFAULT_ROTATION_PERIOD period.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 98 -
arch/x86/events/intel/cmt.h | 2 +
2 files changed, 98 insertions(+)
The hooks allows architectures to extend the behavior of the perf cgroup.
In this patch series, the hooks are used to allow intel_cmt to follow
changes in the perf_cgroup hierarchy.
Signed-off-by: David Carrillo-Cisneros
---
include/linux/perf_event.h | 11 +++
kernel/events/core.c
To make rmid rotation more dependable, this patch series introduces
rotation Service Level Objectives (SLOs) that are described in
code's documentation.
This patch introduces cmt_{pre,min}_mon_slice SLOs that protects from
bogus values when a rmid has not been available since the beginning of
moni
New PMUs, such as CMT's, do not guarantee that a read will succeed even
if pmu::add was successful.
In the generic code, this patch adds an int error return and completes the
error checking path up to perf_read().
In CMT's PMU, it adds proper error handling of hw read failure errors.
In other PMU
Add rmid rotation code to steal an rmid whenever not enough
pmonrs are being reactivated.
More details in code's comments.
Signed-off-by: David Carrillo-Cisneros
---
arch/x86/events/intel/cmt.c | 149 ++--
1 file changed, 144 insertions(+), 5 deletions(-)
From: Stephane Eranian
When an event is in error state, read() returns 0
instead of sizeof() buffer. In certain modes, such
as interval printing, ignoring the 0 return value
may cause bogus count deltas to be computed and
thus invalid results printed.
this patch fixes this problem by modifying r
Add debug fs to intel_cmt to help maintenance. It shows the following
human readable snapshots of the internals of the CMT driver:
- hrchy: A per monr view of the monr hierarchy.
- pkgs: A per-package view of all online struct pkg_data.
- rmids: A per-package view of occupancy and state of al
A package wide event can return a valid read even if it has not run in a
specific cpu, this does not fit well with the assumption that run == 0
is equivalent to a .
To fix the problem, this patch defines special error values for val,
run and ena (~0ULL), and use them to signal read errors, allowin
This series introduces the next iteration of kernel support for the
Cache Monitoring Technology or CMT (formerly Cache QoS Monitoring, CQM)
available in Intel Xeon processors.
Documentation has replaced the Intel CQM name with Intel CMT.
This version renames all code to CMT.
It is rebased at tip
On Fri, Oct 28, 2016 at 5:08 AM, Eric Dumazet wrote:
> On Thu, 2016-10-27 at 12:54 -0500, Thomas Falcon wrote:
>> On 10/27/2016 10:26 AM, Eric Dumazet wrote:
>> > On Wed, 2016-10-26 at 11:09 +1100, Jon Maxwell wrote:
>> >> We recently encountered a bug where a few customers using ibmveth on the
>>
None of the Kconfigs for any of these drivers are tristate,
meaning that they currently are not being built as a module by anyone.
Lets remove the modular code that is essentially orphaned, so that
when reading the drivers there is no doubt they are builtin-only. All
drivers get essentially the s
From: Alexander Usyskin
A mei client driver on the mei client bus can call disconnect function on
already internal disconnected client. A client can disconnect internally,
for example, during link reset or upon FW request. Those are legitimate
flows and we should not log an error message, hence
From: Alexander Usyskin
The HBM protocol version is negotiated during the setup phase, then settled
on a highest possible common version of the driver and the firmware.
The sysfs API advertises both negotiated and driver supported versions
in the device attributes.
Signed-off-by: Alexander Usysk
From: Kyyaly M.
Sent: 30 October 2016 00:12
To: Kyyaly M.
Subject: RE: Is This Email Still Working.
Transaction offer contact me---> jonathan_symond...@outlook.com for
details.
On 10/29/2016 07:49 AM, Greg Kroah-Hartman wrote:
> This is the start of the stable review cycle for the 4.4.29 release.
> There are 51 patches in this series, all will be posted as a response
> to this one. If anyone has any issues with these being applied, please
> let me know.
>
> Responses sh
On 10/29/2016 07:48 AM, Greg Kroah-Hartman wrote:
> This is the start of the stable review cycle for the 4.8.6 release.
> There are 125 patches in this series, all will be posted as a response
> to this one. If anyone has any issues with these being applied, please
> let me know.
>
> Responses sh
On 29 October 2016 22:45:41 BST, Yury Norov wrote:
>On Sat, Oct 29, 2016 at 11:02:40PM +0200, Arnd Bergmann wrote:
>> On Saturday, October 22, 2016 3:14:04 PM CEST Yury Norov wrote:
>> > The newer prlimit64 syscall provides all the functionality provided
>by
>> > the getrlimit and setrlimit syscal
Hi Paul,
[auto build test ERROR on linus/master]
[also build test ERROR on v4.9-rc3 next-20161028]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
[Suggest to use git(>=2.9.0) format-patch --base= (or --base=auto for
convenience) to record what
1 - 100 of 501 matches
Mail list logo
