On 2016/07/21 at 22:46, Juri Lelli wrote:
> On 21/07/16 15:36, Juri Lelli wrote:
>> On 21/07/16 15:21, Juri Lelli wrote:
>>> Hi,
>>>
>>> On 18/07/16 21:37, Xunlei Pang wrote:
>>>> On 2016/07/18 at 21:04, Juri Lelli wrote:
>>>>> On 1
PV kernel
>>>
>>> Changes in V3:
>>> - Revise comments, description, and symbol names
>>>
>>> Changes in V2:
>>> - Replace smp_send_stop() call with crash_kexec version which
>>> saves cpu states and cleans up VMX/SVM
>>> - Dr
o the very same update at tasks wakeup time anyway. By doing
> so, we don't need to worry about a potential PI donor anymore, as rt_
> mutex_setprio() takes care of that already for us.
>
> Cc: Ingo Molnar
> Cc: Peter Zijlstra
> Cc: Steven Rostedt
> Cc: Luca Abeni
>
On 2016/08/24 at 16:20, Dave Young wrote:
> On 08/23/16 at 06:11pm, Yinghai Lu wrote:
>> On Wed, Aug 17, 2016 at 1:20 AM, Dave Young wrote:
>>> On 08/17/16 at 09:50am, Xunlei Pang wrote:
>>>> "/sys/kernel/kexec_crash_size" only handles crashk_res, it
>
On 06/22/2017 at 01:44 AM, Michael Holzheu wrote:
> Am Fri, 9 Jun 2017 10:17:05 +0800
> schrieb Xunlei Pang :
>
>> S390 KEXEC_NOTE_BYTES is not used by note_buf_t as before, which
>> is now defined as follows:
>> typedef u32 note_buf_t[CRASH_CORE_NOTE_BYTES/4
On 04/19/2017 at 05:21 AM, Tom Lendacky wrote:
> Provide support so that kexec can be used to boot a kernel when SME is
> enabled.
>
> Support is needed to allocate pages for kexec without encryption. This
> is needed in order to be able to reboot in the kernel in the same manner
> as originally b
On 05/26/2017 at 10:49 AM, Dave Young wrote:
> Ccing Xunlei he is reading the patches see what need to be done for
> kdump. There should still be several places to handle to make kdump work.
>
> On 05/18/17 at 07:01pm, Borislav Petkov wrote:
>> On Tue, Apr 18, 2017 at 04:22:12PM -0500, Tom Lendacky
From: Xunlei Pang
Since we are using cpuidle_driver::safe_state_index directly as the
target state index, it is better to add the sanity check at the point
of registering the driver.
Signed-off-by: Xunlei Pang
---
v2->v3:
Move the code to a new cpuidle_coupled_state_verify() depending
Hi Rafael,
On 3 September 2015 at 09:35, Rafael J. Wysocki wrote:
> On Monday, August 31, 2015 11:34:05 AM Xunlei Pang wrote:
>> From: Xunlei Pang
>>
>> Since we are using cpuidle_driver::safe_state_index directly as the
>> target state index, it is better to add th
On 6/21/18 1:01 AM, bseg...@google.com wrote:
> Xunlei Pang writes:
>
>> I noticed the group constantly got throttled even it consumed
>> low cpu usage, this caused some jitters on the response time
>> to some of our business containers enabling cpu quota.
>>
&g
On 6/21/18 4:08 PM, Peter Zijlstra wrote:
> On Thu, Jun 21, 2018 at 11:56:56AM +0800, Xunlei Pang wrote:
>>>> Fixes: 51f2176d74ac ("sched/fair: Fix unlocked reads of some
>>>> cfs_b->quota/period")
>>>> Cc: Ben Segall
>>>
>>&g
last parse in cputime_adjust(), and accumulate the
corresponding results calculated into prev_cputime. A new field
of task_cputime type is added in structure prev_cputime to record
previous task_cputime so that we can get the elapsed time deltas.
Signed-off-by: Xunlei Pang
---
include/linux/sched.h
ase drop us a note to
> help improve the system]
>
> url:
> https://github.com/0day-ci/linux/commits/Xunlei-Pang/sched-cputime-Ensure-correct-utime-and-stime-proportion/20180622-153720
> reproduce: make htmldocs
>
> All warnings (new ones prefixed by >>):
>
;s a problem, as expires_seq will get synced in
assign_cfs_rq_runtime().
Thanks,
Xunlei
>
> Fixes: 512ac999d275 ("sched/fair: Fix bandwidth timer clock drift condition")
> Cc: Xunlei Pang
> Cc: Ben Segall
> Cc: Linus Torvalds
> Cc: Peter Zijlstra
> Cc: Thomas
On 7/31/18 1:55 AM, Cong Wang wrote:
> On Sun, Jul 29, 2018 at 10:29 PM Xunlei Pang wrote:
>>
>> Hi Cong,
>>
>> On 7/28/18 8:24 AM, Cong Wang wrote:
>>> Each time we sync cfs_rq->runtime_expires with cfs_b->runtime_expires,
>>> we should sync i
On 8/1/18 4:55 AM, Cong Wang wrote:
> On Tue, Jul 31, 2018 at 10:13 AM wrote:
>>
>> Xunlei Pang writes:
>>
>>> On 7/31/18 1:55 AM, Cong Wang wrote:
>>>> On Sun, Jul 29, 2018 at 10:29 PM Xunlei Pang
>>>> wrote:
>>>>>
>&g
On 7/23/18 5:21 PM, Peter Zijlstra wrote:
> On Tue, Jul 17, 2018 at 12:08:36PM +0800, Xunlei Pang wrote:
>> The trace data corresponds to the last sample period:
>> trace entry 1:
>> cat-20755 [022] d... 1370.106496: cputime_adjust: task
>> tick-bas
if positive after throttling.
Fixes: df8eac8cafce ("sched/deadline: Throttle a constrained deadline task
activated after the deadline)
Acked-by: Daniel Bristot de Oliveira
Signed-off-by: Xunlei Pang
---
kernel/sched/deadline.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/kern
errun_block: 0
dl.nr_underrun_yield: 0
Very large "dl.nr_underrun_sched" hints it's very likely that
there is some underlying scheduling issue.
Note that we don't use CONFIG_SCHED_DEBUG as the accounting
added has little overhead(also
case of missed
deadlines")
Cc: Luca Abeni
Signed-off-by: Xunlei Pang
---
kernel/sched/deadline.c | 9 +++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index d3d291e..5691149 100644
--- a/kernel/sched/deadline.c
+++ b/ke
On 05/11/2017 at 09:38 AM, Xunlei Pang wrote:
> On 05/10/2017 at 09:36 PM, Steven Rostedt wrote:
>> On Wed, 10 May 2017 21:03:37 +0800
>> Xunlei Pang wrote:
>>
>>> When a contrained task is throttled by dl_check_constrained_dl(),
>>> it may carry the remai
On 05/12/2017 at 01:57 PM, luca abeni wrote:
> Hi again,
>
> (sorry for the previous email; I replied from gmail and I did not
> realize I was sending it in html).
>
>
> On Fri, 12 May 2017 11:32:08 +0800
> Xunlei Pang wrote:
>
>> dl_runtime_exceeded() only
On 05/12/2017 at 03:01 PM, luca abeni wrote:
> Hi,
>
> On Fri, 12 May 2017 14:53:33 +0800
> Xunlei Pang wrote:
>
>> On 05/12/2017 at 01:57 PM, luca abeni wrote:
>>> Hi again,
>>>
>>> (sorry for the previous email; I replied from gmail an
On 05/13/2017 at 04:58 AM, luca abeni wrote:
> On Fri, 12 May 2017 15:19:55 +0800
> Xunlei Pang wrote:
> [...]
>>>> "As seen, enforcing that the total utilization is smaller than M
>>>> does not guarantee that global EDF schedules the tasks without
>
hen/Andrew,
Sorry for the trouble.
The following patch will fix it, do you want to me to send it out separately?
or help merge it into
fc7d2b44367f ("powerpc/fadump: use the correct VMCOREINFO_NOTE_SIZE for phdr")
directly?
Thanks,
Xunlei
===
Fr
On 05/12/2017 at 11:32 AM, Xunlei Pang wrote:
> When a contrained task is throttled by dl_check_constrained_dl(),
> it may carry the remaining positive runtime, as a result when
> dl_task_timer() fires and calls replenish_dl_entity(), it will
> not be replenished correctly due to
MCOREINFO_BYTES.
Andrew will help merge it into commit fc7d2b44367f ("powerpc/fadump:
use the correct VMCOREINFO_NOTE_SIZE for phdr") before sending it to
stable.
Signed-off-by: Xunlei Pang
---
kernel/crash_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel
Hi Roman,
On 2018/12/4 AM 2:00, Roman Gushchin wrote:
> On Mon, Dec 03, 2018 at 04:01:17PM +0800, Xunlei Pang wrote:
>> When usage exceeds min, min usage should be min other than 0.
>> Apply the same for low.
>>
>> Signed-off-by: Xunlei Pang
>> ---
>> mm
When usage exceeds min, min usage should be min other than 0.
Apply the same for low.
Signed-off-by: Xunlei Pang
---
mm/page_counter.c | 12 ++--
1 file changed, 2 insertions(+), 10 deletions(-)
diff --git a/mm/page_counter.c b/mm/page_counter.c
index de31470655f6..75d53f15f040 100644
igned-off-by: Xunlei Pang
---
mm/vmscan.c | 8
1 file changed, 8 insertions(+)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 62ac0c488624..3d412eb91f73 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3531,6 +3531,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int
classzon
this part of usages to be reclaimed.
Signed-off-by: Xunlei Pang
---
include/linux/memcontrol.h | 7 +--
mm/memcontrol.c| 9 +++--
mm/vmscan.c| 17 +++--
3 files changed, 27 insertions(+), 6 deletions(-)
diff --git a/include/linux/memcontrol.h b
On 2018/12/3 下午7:54, Michal Hocko wrote:
> On Mon 03-12-18 16:01:17, Xunlei Pang wrote:
>> When usage exceeds min, min usage should be min other than 0.
>> Apply the same for low.
>
> Why? What is the actual problem.
children_min_usage tracks the total children usages un
On 2018/12/3 下午7:56, Michal Hocko wrote:
> On Mon 03-12-18 16:01:18, Xunlei Pang wrote:
>> There may be cgroup memory overcommitment, it will become
>> even common in the future.
>>
>> Let's enable kswapd to reclaim low-protected memory in case
>> of memory p
On 2018/12/4 AM 1:22, Michal Hocko wrote:
> On Mon 03-12-18 23:20:31, Xunlei Pang wrote:
>> On 2018/12/3 下午7:56, Michal Hocko wrote:
>>> On Mon 03-12-18 16:01:18, Xunlei Pang wrote:
>>>> There may be cgroup memory overcommitment, it will become
>>>>
On 2018/12/3 PM 7:57, Michal Hocko wrote:
> On Mon 03-12-18 16:01:19, Xunlei Pang wrote:
>> When memcgs get reclaimed after its usage exceeds min, some
>> usages below the min may also be reclaimed in the current
>> implementation, the amount is considerably large dur
On 2018/12/4 PM 3:25, Michal Hocko wrote:
> On Tue 04-12-18 10:40:29, Xunlei Pang wrote:
>> On 2018/12/4 AM 1:22, Michal Hocko wrote:
>>> On Mon 03-12-18 23:20:31, Xunlei Pang wrote:
>>>> On 2018/12/3 下午7:56, Michal Hocko wrote:
>>>>> On Mon 03-12-18
Hi Peter,
On 7/5/18 9:21 PM, Xunlei Pang wrote:
> On 7/5/18 6:46 PM, Peter Zijlstra wrote:
>> On Wed, Jun 27, 2018 at 08:22:42PM +0800, Xunlei Pang wrote:
>>> tick-based whole utime is utime_0, tick-based whole stime
>>> is stime_0, scheduler time is rtime_0.
>>
time type field is added in prev_cputime to record
previous task_cputime so that we can get the elapsed times as the accurate
ratio.
Signed-off-by: Xunlei Pang
---
v1->v2:
- Rewrite the changelog.
include/linux/sched.h | 34
include/linux/sched/cputi
Hi Peter,
On 7/9/18 6:48 PM, Peter Zijlstra wrote:
> On Mon, Jul 09, 2018 at 01:52:38PM +0800, Xunlei Pang wrote:
>> Please see the enclosure for the reproducer cputime_adjust.tgz
>
> No, I'm not going to reverse engineer something if you cannot even
> explain what the pr
On 6/22/18 3:15 PM, Xunlei Pang wrote:
> We use per-cgroup cpu usage statistics similar to "cgroup rstat",
> and encountered a problem that user and sys usages are wrongly
> split sometimes.
>
> Run tasks with some random run-sleep pattern for a long time, and
> when t
On 6/26/18 11:49 PM, Peter Zijlstra wrote:
> On Tue, Jun 26, 2018 at 08:19:49PM +0800, Xunlei Pang wrote:
>> On 6/22/18 3:15 PM, Xunlei Pang wrote:
>>> We use per-cgroup cpu usage statistics similar to "cgroup rstat",
>>> and encountered a problem that user
is re-armed.
The global expiration should be advanced accordingly when the
bandwidth period timer is restarted.
Signed-off-by: Xunlei Pang
---
kernel/sched/fair.c | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 9f
following patch.
Fixes: 51f2176d74ac ("sched/fair: Fix unlocked reads of some
cfs_b->quota/period")
Cc: Ben Segall
Signed-off-by: Xunlei Pang
---
kernel/sched/fair.c | 14 --
kernel/sched/sched.h | 6 --
2 files changed, 12 insertions(+), 8 deletions(-)
diff --gi
On 6/19/18 2:44 AM, bseg...@google.com wrote:
> Xunlei Pang writes:
>
>> The current condition to judge clock drift in expire_cfs_rq_runtime()
>> is wrong, the two runtime_expires are actually the same when clock
>> drift happens, so this condtion can never hit
On 6/19/18 2:58 AM, bseg...@google.com wrote:
> Xunlei Pang writes:
>
>> I noticed the group frequently got throttled even it consumed
>> low cpu usage, this caused some jitters on the response time
>> to some of our business containers enabling cpu quota.
>>
On 6/19/18 12:36 PM, Cong Wang wrote:
> On Mon, Jun 18, 2018 at 2:16 AM, Xunlei Pang wrote:
>> I noticed the group frequently got throttled even it consumed
>> low cpu usage, this caused some jitters on the response time
>> to some of our business containers enabling cpu q
On 6/20/18 1:49 AM, bseg...@google.com wrote:
> Xunlei Pang writes:
>
>> On 6/19/18 2:58 AM, bseg...@google.com wrote:
>>> Xunlei Pang writes:
>>>
>>>> I noticed the group frequently got throttled even it consumed
>>>> low cpu usage, this ca
date the global expiration in start_cfs_bandwidth() to avoid frequent
expire_cfs_rq_runtime() calls once a new period begins.
Signed-off-by: Xunlei Pang
---
kernel/sched/fair.c | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair
76d74ac ("sched/fair: Fix unlocked reads of some
cfs_b->quota/period")
Cc: Ben Segall
Signed-off-by: Xunlei Pang
---
kernel/sched/fair.c | 14 --
kernel/sched/sched.h | 6 --
2 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/kernel/sched/
On 6/20/18 2:06 PM, Xunlei Pang wrote:
> On 6/20/18 1:49 AM, bseg...@google.com wrote:
>> Xunlei Pang writes:
>>
>>> On 6/19/18 2:58 AM, bseg...@google.com wrote:
>>>> Xunlei Pang writes:
>>>>
>>>>> I noticed the group frequently
On 7/2/18 11:21 PM, Tejun Heo wrote:
> Hello, Peter.
>
> On Tue, Jun 26, 2018 at 05:49:08PM +0200, Peter Zijlstra wrote:
>> Well, no, because the Changelog is incomprehensible and the patch
>> doesn't really have useful comments, so I'll have to reverse engineer
>> the entire thing, and I've just
On 7/5/18 6:46 PM, Peter Zijlstra wrote:
> On Wed, Jun 27, 2018 at 08:22:42PM +0800, Xunlei Pang wrote:
>> tick-based whole utime is utime_0, tick-based whole stime
>> is stime_0, scheduler time is rtime_0.
>
>> For a long time, the process runs mainly in userspace wi
On 7/17/18 1:41 AM, Ingo Molnar wrote:
>
> * Peter Zijlstra wrote:
>
>> On Sun, Jul 15, 2018 at 04:36:17PM -0700, tip-bot for Xunlei Pang wrote:
>>> Commit-ID: 8d4c00dc38a8aa30dae8402955e55e7b34e74bc8
>>> Gitweb:
>>> https://git.kernel.org/ti
ll disfunctional due to
timer2 with the earliest expires in the timerqueue.
So, rtc drivers must set this flag if they don't support UIE.
Signed-off-by: Xunlei Pang
---
drivers/rtc/rtc-ab8500.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/rtc/rtc-ab8500.c b/drivers/rtc/rtc-ab8
pires time, so the alarm irq will never be programmed.
When such failures happen, this patch will retry __rtc_set_alarm(), if
still can't program the alarm time, it will remove current rtc_timer from
timerqueue and fetch next one, thus preventing it from affecting other
rtc timers.
Signed-off-
Hi Thomas,
On 23 January 2015 at 05:07, Thomas Gleixner wrote:
> On Thu, 22 Jan 2015, Xunlei Pang wrote:
>> When doing timekeeping_resume(), if the nonstop clocksource
>> wraps back, "cycle_delta" will miss the wrap time.
>>
>> It's hard to de
On 23 January 2015 at 02:30, John Stultz wrote:
> On Thu, Jan 22, 2015 at 4:01 AM, Xunlei Pang wrote:
>> If a system does not provide a persistent_clock(), the time
>> will be updated on resume by rtc_resume(). With the addition
>> of the non-stop clocksources for suspend
Hi Peter, Juri,
Could you please give some comments on these 5 patches?
Thanks for your time.
Regards,
Xunlei
On 19 January 2015 at 12:49, Xunlei Pang wrote:
> Currently, cpudl::free_cpus contains all cpus during init, see
> cpudl_init(). When calling cpudl_find(), we have to add rd->
From: Xunlei Pang
rtc_class_ops's set_alarm() shouldn't deal with the alarm date,
as this is handled in the rtc core.
See rtc_dev_ioctl()'s RTC_ALM_SET and RTC_WKALM_SET cases.
Signed-off-by: Xunlei Pang
---
drivers/rtc/rtc-mxc.c | 22 --
1 file change
From: Xunlei Pang
Change alpha_rtc_set_mmss() and remote_set_mmss() to use
rtc_class_ops's set_mmss64().
Signed-off-by: Xunlei Pang
---
arch/alpha/kernel/rtc.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/arch/alpha/kernel/rtc.c b/arch/alpha/kernel/rtc.c
From: Xunlei Pang
As part of addressing "y2038 problem" for in-kernel uses, this
patch adds safe get_seconds64() using time64_t.
After this patch, get_seconds() is deprecated and all its call sites
will be fixed using get_seconds64(), after that it can be removed.
Signed-off-by: X
From: Xunlei Pang
This driver has a number of y2038/y2106 issues.
This patch resolves them by:
- Repalce get_seconds() with get_seconds64()
- Replace rtc_time_to_tm() with rtc_time64_to_tm()
- Change test_rtc_set_mmss() to use rtc_class_ops's set_mmss64()
After this patch, this driver s
From: Xunlei Pang
This driver has a number of y2038/y2106 issues.
This patch resolves them by:
- Replace rtc_time_to_tm() with rtc_time64_to_tm()
- Replace rtc_tm_to_time() with rtc_tm_to_time64()
- Change mxc_rtc_set_mmss() to use rtc_class_ops's set_mmss64()
After this patch, the d
From: Xunlei Pang
This patch series relies on a former patchset[1], it provides y2038/y2106
safe rtc_class_ops.set_mmss64(), and converts some possible users of set_mmss()
to use set_mmss64(), in the hope that making these users(i.e. rtc drivers)
y2038/y2106 safe.
The first version of this
From: Xunlei Pang
Currently the rtc_class_op's set_mmss() function takes a 32bit second
value (on 32bit systems), which is problematic for dates past y2038.
This patch provides a safe version named set_mmss64() using y2038 safe
time64_t.
After this patch, set_mmss() is deprecated and al
From: Xunlei Pang
We want to convert mxc_rtc_set_mmss() to use rtc_class_ops's set_mmss64(),
but it uses get_alarm_or_time()/set_alarm_or_time() internal interfaces
which are y2038 unsafe.
So here as a separate patch, it converts these two internal interfaces
of "mxc" to use s
From: Xunlei Pang
This driver has a number of y2038/y2106 issues.
This patch resolves them by:
- Replace rtc_tm_to_time() with rtc_tm_to_time64()
- Replace rtc_time_to_tm() with rtc_time64_to_tm()
- Change ab3100_rtc_set_mmss() to use rtc_class_ops's set_mmss64()
After this patch, the d
From: Xunlei Pang
This driver has a number of y2038/y2106 issues.
This patch resolves them by:
- Replace rtc_time_to_tm() with rtc_time64_to_tm()
- Change mc13xxx_rtc_set_mmss() to use rtc_class_ops's set_mmss64()
After this patch, the driver should not have any remaining
y2038/y2106 i
On 14 January 2015 at 04:42, Thomas Gleixner wrote:
> On Tue, 13 Jan 2015, Xunlei Pang wrote:
>
>> From: Xunlei Pang
>>
>> As part of addressing "y2038 problem" for in-kernel uses, this
>> patch adds safe get_seconds64() using time64_t.
>>
>> A
On 14 January 2015 at 00:19, Alessandro Zummo wrote:
> On Tue, 13 Jan 2015 23:44:51 +0800
> Xunlei Pang wrote:
>
>> This patch resolves them by:
>> - Repalce get_seconds() with get_seconds64()
>> - Replace rtc_time_to_tm() with rtc_time64_to_tm()
>> -
Cc Uwe Kleine-Koenig
On 13 January 2015 at 23:44, Xunlei Pang wrote:
> From: Xunlei Pang
>
> This driver has a number of y2038/y2106 issues.
>
> This patch resolves them by:
> - Replace rtc_time_to_tm() with rtc_time64_to_tm()
> - Change mc13xxx_rtc_set_mmss() to use rtc_c
Cc Fabio Estevam
On 13 January 2015 at 23:44, Xunlei Pang wrote:
> From: Xunlei Pang
>
> rtc_class_ops's set_alarm() shouldn't deal with the alarm date,
> as this is handled in the rtc core.
>
> See rtc_dev_ioctl()'s RTC_ALM_SET and RTC_WKALM_SET case
On 14 January 2015 at 23:03, Alessandro Zummo wrote:
> On Wed, 14 Jan 2015 22:56:15 +0800
> Xunlei Pang wrote:
>
>> We want to do away with set_mmss() eventually, time64 is the final version,
>> so I think this would be ok.
>
> We'll get rid of set_mmss() from th
On 14 January 2015 at 23:39, Alessandro Zummo wrote:
> On Wed, 14 Jan 2015 23:26:37 +0800
> Xunlei Pang wrote:
>
>> But on the other hand, we will have no test for set_mmss64(),
>> because adding the set_mmss64() will make set_mmss() dysfunctional.
>
> add a modu
t use of percpu local_cpu_mask
to save an extra mask allocation, so always passing a non-NULL
lowest_mask to cpupri_find().
Signed-off-by: Xunlei Pang
---
kernel/sched/core.c | 2 ++
kernel/sched/cpupri.c | 22 +-
kernel/sched/cpupri.h | 1 +
kernel/sched/rt.c | 9
onto a dying cpu) which
rq_offline_dl() can't handle timely, then it can be handled
through _cpu_down()->...->multi_cpu_stop()->migration_call()
->migrate_tasks(), preventing the task from hanging on the
dead cpu.
Signed-off-by: Xunlei Pang
---
kernel/sched/cpudeadline.c | 3 +--
In check_preempt_equal_dl(), cpudl_find() is called with a NULL
later_mask, thus cpudl_find() here doesn't check cpudl::free_cpus
at all.
This patch takles this issue by always passing a non-NULL later_mask
to cpudl_find(), thereby fixing this issue.
Signed-off-by: Xunlei Pang
---
kernel/
e interation of
sched_domain() in passing.
Signed-off-by: Xunlei Pang
---
kernel/sched/rt.c | 17 +
1 file changed, 13 insertions(+), 4 deletions(-)
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index d28cfa4..e6a42e6 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1535,6 +1
_rq().
This patch adds cpudl_set_freecpu() and cpudl_clear_freecpu() for
changing cpudl::free_cpus when doing rq_online_dl()/rq_offline_dl(),
so we can avoid the rd->span operation when calling cpudl_find()
in find_later_rq().
Signed-off-by: Xunlei Pang
---
kernel/sched/cpudeadline.
Hi Steven,
On 28 January 2015 at 07:04, Steven Rostedt wrote:
> On Tue, 27 Jan 2015 13:58:59 +0100
> Peter Zijlstra wrote:
>
>
>> Not a bad idea, Cc'ed Steve who likes to look after the RT bits,
>> excessive quoting for him.
>
> Yep I do.
>
>
>> > index ade2958..48c9576 100644
>> > --- a/kernel/
Hi Peter,
On 28 January 2015 at 00:47, Peter Zijlstra wrote:
> On Mon, Jan 19, 2015 at 04:49:38AM +0000, Xunlei Pang wrote:
>> In check_preempt_equal_dl(), cpudl_find() is called with a NULL
>> later_mask, thus cpudl_find() here doesn't check cpudl::free_cpus
>> at al
From: Xunlei Pang
isl1208_i2c_set_alarm() uses deprecated rtc_tm_to_time(),
which will overflow in year 2106 on 32-bit machines.
This patch solves this by:
- Replacing rtc_tm_to_time() with rtc_tm_to_time64()
Cc: Herbert Valerio Riedel
Signed-off-by: Xunlei Pang
---
drivers/rtc/rtc-isl1208
From: Xunlei Pang
pcf8563_rtc_set_alarm() uses deprecated rtc_tm_to_time()
and rtc_time_to_tm(), which will overflow in year 2106
on 32-bit machines.
This patch solves this by:
- Replacing rtc_time_to_tm() with rtc_time64_to_tm()
- Replacing rtc_tm_to_time() with rtc_tm_to_time64()
Signed
From: Xunlei Pang
sunxi_rtc_setalarm() uses deprecated rtc_tm_to_time(),
which will overflow in year 2106 on 32-bit machines.
This patch solves this by:
- Replacing rtc_tm_to_time() with rtc_tm_to_time64()
Also remove the unnecessary initial zeroing of some
local variables in
From: Xunlei Pang
The driver uses deprecated rtc_tm_to_time() and rtc_time_to_tm(),
which will overflow in year 2106 on 32-bit machines.
This patch solves this by:
- Replacing rtc_tm_to_time() with rtc_tm_to_time64()
- Replacing rtc_time_to_tm() with rtc_time64_to_tm()
Cc: Russell King
From: Xunlei Pang
The driver uses deprecated rtc_tm_to_time() and rtc_time_to_tm(),
which will overflow in year 2106 on 32-bit machines.
This patch solves this by:
- Replacing rtc_tm_to_time() with rtc_tm_to_time64()
- Replacing rtc_time_to_tm() with rtc_time64_to_tm()
Cc: CIH
Cc: Nicolas
On 27 March 2015 at 23:28, Steven Rostedt wrote:
> On Mon, 9 Mar 2015 15:32:27 +0800
> Xunlei Pang wrote:
>
>> From: Xunlei Pang
>>
>> Currently, SMP RT scheduler has some trouble in dealing with
>> equal prio cases.
>>
>> For example, in check_p
On 27 March 2015 at 23:30, Peter Zijlstra wrote:
> On Wed, Mar 18, 2015 at 02:31:02PM +0800, Xunlei Pang wrote:
>> From: Xunlei Pang
>>
>> In load_balance(), some members of lb_env will be assigned with
>> new values in LBF_DST_PINNED case. But lb_env::flags may still
Vincent,
On 27 March 2015 at 23:37, Vincent Guittot wrote:
> On 27 March 2015 at 16:12, Xunlei Pang wrote:
>>> +static int get_cpu_usage(int cpu)
>>> +{
>>> + unsigned long usage = cpu_rq(cpu)->cfs.utilization_load_avg;
>>> + uns
Hi Vincent,
On 27 March 2015 at 23:59, Vincent Guittot wrote:
> On 27 March 2015 at 15:52, Xunlei Pang wrote:
>> Hi Vincent,
>>
>> On 27 February 2015 at 23:54, Vincent Guittot
>> wrote:
>>> /**
>>> @@ -6432,18 +6435,19 @@ static inline void updat
Hi Vincent,
On 1 April 2015 at 17:06, Vincent Guittot wrote:
> On 1 April 2015 at 05:37, Xunlei Pang wrote:
>> Hi Vincent,
>>
>> On 27 March 2015 at 23:59, Vincent Guittot
>> wrote:
>>> On 27 March 2015 at 15:52, Xunlei Pang wrote:
>>>> Hi
From: Xunlei Pang
If there're multiple nodes with the same prio as @node, currently
plist_add() will add @node behind all of them. Now we need to add
@node before all of these nodes for SMP RT scheduler.
This patch adds a common __plist_add() for adding @node before or
after existing nodes
From: Xunlei Pang
We may suffer from extra rt overload rq due to the affinity,
so when the affinity of any runnable rt task is changed, we
should check to trigger balancing, otherwise it will cause
some unnecessary delayed real-time response. Unfortunately,
current RT global scheduler does
From: Xunlei Pang
In check_preempt_equal_prio_common(), it requeues "next" ahead
in the "run queue" and want to push current away. But when doing
the actual pushing, if the system state changes, the pushing may
fail as a result.
In this case, p finally becomes the new curr
From: Xunlei Pang
When p is queued, there may be other tasks already queued at the
same priority in the "run queue", so we should peek the most front
one to do the equal priority preemption.
This patch modifies check_preempt_equal_prio() and provides new
check_preempt_equal_prio_comm
From: Xunlei Pang
There may be some non-migratable tasks queued in the "run queue"
with the same priority as current which is FIFO and migratable,
so at each tick we can check and try to push current away and
give these tasks a chance of running(we don't do this for tasks
qu
From: Xunlei Pang
We know, there are two main queues each cpu for RT scheduler:
Let's call them "run queue" and "pushable queue" respectively.
For RT tasks, the scheduler uses "plist" to manage the pushable queue,
so when there are multiple tasks queued at th
From: Xunlei Pang
1. TWO FIFO QUEUES
We know, there are two main queues each cpu for RT SMP scheduler. Let's
name them "run queue" and "pushable queue" respectively.
"run queue" is from the intra-cpu's perspective, while "pushable queue" is
From: Xunlei Pang
In check_preempt_equal_prio(), when p is queued, there may be other
tasks already queued at the same priority in the "run queue", so we
should peek the most front one to do the preemption, not the p.
This patch modified it and moved the preemption job to a new func
From: Xunlei Pang
If there're multiple nodes with the same prio as @node, currently
plist_add() will add @node behind all of them. Now we need to add
@node before all of these nodes for SMP RT scheduler.
This patch adds a common __plist_add() for adding @node before or
after existing nodes
1 - 100 of 553 matches
Mail list logo
