[PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[Xiu Jianfeng]

Initializing top_cpuset.relax_domain_level and setting
CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
completed at the time of top_cpuset definition by compiler.

Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
---
 kernel/cgroup/cpuset.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index d8d3439eda4e..e70008a1d86a 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -369,8 +369,9 @@ static inline void notify_partition_change(struct cpuset *cs, int old_prs)
 
 static struct cpuset top_cpuset = {
 	.flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
-		  (1 << CS_MEM_EXCLUSIVE)),
+		  (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),
 	.partition_root_state = PRS_ROOT,
+	.relax_domain_level = -1,
 	.remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
 };
 
@@ -4309,8 +4310,6 @@ int __init cpuset_init(void)
 	nodes_setall(top_cpuset.effective_mems);
 
 	fmeter_init(&top_cpuset.fmeter);
-	set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags);
-	top_cpuset.relax_domain_level = -1;
 	INIT_LIST_HEAD(&remote_children);
 
 	BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
-- 
2.34.1

Re: [PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[Waiman Long]

On 4/20/24 05:46, Xiu Jianfeng wrote:
> Initializing top_cpuset.relax_domain_level and setting
> CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
> completed at the time of top_cpuset definition by compiler.
>
> Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
> ---
>   kernel/cgroup/cpuset.c | 5 ++---
>   1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index d8d3439eda4e..e70008a1d86a 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -369,8 +369,9 @@ static inline void notify_partition_change(struct cpuset *cs, int old_prs)
>   
>   static struct cpuset top_cpuset = {
>   	.flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
> -		  (1 << CS_MEM_EXCLUSIVE)),
> +		  (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),
>   	.partition_root_state = PRS_ROOT,
> +	.relax_domain_level = -1,
>   	.remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
>   };
>   
> @@ -4309,8 +4310,6 @@ int __init cpuset_init(void)
>   	nodes_setall(top_cpuset.effective_mems);
>   
>   	fmeter_init(&top_cpuset.fmeter);
> -	set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags);
> -	top_cpuset.relax_domain_level = -1;
>   	INIT_LIST_HEAD(&remote_children);
>   
>   	BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
Reviewed-by: Waiman Long <longman@redhat.com>

Re: [PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[Tejun Heo]

On Sat, Apr 20, 2024 at 09:46:16AM +0000, Xiu Jianfeng wrote:
> Initializing top_cpuset.relax_domain_level and setting
> CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
> completed at the time of top_cpuset definition by compiler.
> 
> Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>

Applied to cgroup/for-6.10.

Thanks.

-- 
tejun

Re: [PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[Klara Modin]

[-- Attachment #1: Type: text/plain, Size: 1453 bytes --]

Hi,

On 2024-04-20 11:46, Xiu Jianfeng wrote:
> Initializing top_cpuset.relax_domain_level and setting
> CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
> completed at the time of top_cpuset definition by compiler.
> 
> Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
> ---
>   kernel/cgroup/cpuset.c | 5 ++---
>   1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index d8d3439eda4e..e70008a1d86a 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -369,8 +369,9 @@ static inline void notify_partition_change(struct cpuset *cs, int old_prs)
>   
>   static struct cpuset top_cpuset = {
>   	.flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
> -		  (1 << CS_MEM_EXCLUSIVE)),
> +		  (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),

You dropped a '<' for the bitwise shift, this causes bad cpu utilization 
for me.

>   	.partition_root_state = PRS_ROOT,
> +	.relax_domain_level = -1,
>   	.remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
>   };
>   
> @@ -4309,8 +4310,6 @@ int __init cpuset_init(void)
>   	nodes_setall(top_cpuset.effective_mems);
>   
>   	fmeter_init(&top_cpuset.fmeter);
> -	set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags);
> -	top_cpuset.relax_domain_level = -1;
>   	INIT_LIST_HEAD(&remote_children);
>   
>   	BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));

Kind regards,
Klara Modin

[-- Attachment #2: bisect-bad-cpu-util --]
[-- Type: text/plain, Size: 4229 bytes --]

# bad: [a59668a9397e7245b26e9be85d23f242ff757ae8] Add linux-next specific files for 20240423
git bisect start 'next/master'
# status: waiting for good commit(s), bad commit known
# good: [a2c63a3f3d687ac4f63bf4ffa04d7458a2db350b] Merge tag 'bcachefs-2024-04-22' of https://evilpiepirate.org/git/bcachefs
git bisect good a2c63a3f3d687ac4f63bf4ffa04d7458a2db350b
# good: [8115df4dd94f98b35888e6e91681118649298142] Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next.git
git bisect good 8115df4dd94f98b35888e6e91681118649298142
# good: [3ad20cddaf2963acd454d6954f65abbd02f76b8e] Merge branch 'for-next' of git://git.kernel.dk/linux-block.git
git bisect good 3ad20cddaf2963acd454d6954f65abbd02f76b8e
# good: [adec11643114960ad457d0890ebc237cff4e485f] Merge branch 'tty-next' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git
git bisect good adec11643114960ad457d0890ebc237cff4e485f
# bad: [3d9862ccfb857aa54583f1f17953d87a013cd976] Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git
git bisect bad 3d9862ccfb857aa54583f1f17953d87a013cd976
# bad: [121418665b07afcbcff3e154f297c2f3bdcf5608] Merge branch 'gpio/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux.git
git bisect bad 121418665b07afcbcff3e154f297c2f3bdcf5608
# good: [24ddee0ff8c34256891323c92beeaa3bd55a0d30] Merge tag 'stable/vduse-virtio-net' into vhost
git bisect good 24ddee0ff8c34256891323c92beeaa3bd55a0d30
# good: [f92141e18c8b466027e226f3388de15b059b6f65] Merge patch series "convert SCSI to atomic queue limits, part 1 (v3)"
git bisect good f92141e18c8b466027e226f3388de15b059b6f65
# good: [8d1e84ab0176c2d2b49fd741d6609a021ecc1d01] gpio: regmap: Use -ENOTSUPP consistently
git bisect good 8d1e84ab0176c2d2b49fd741d6609a021ecc1d01
# good: [a6b974b40f942b3e51124de588383009f6a42d2d] drivers: remoteproc: xlnx: Add Versal and Versal-NET support
git bisect good a6b974b40f942b3e51124de588383009f6a42d2d
# good: [e99fcac055b3325283d6c5c61a117651fb147686] Merge branches 'rproc-next' and 'rpmsg-next' into for-next
git bisect good e99fcac055b3325283d6c5c61a117651fb147686
# bad: [6c07b00ba1ff48fa9be0d8e084a0503fb93d4b64] Merge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git
git bisect bad 6c07b00ba1ff48fa9be0d8e084a0503fb93d4b64
# bad: [d082aee133bf82bf52ffbbb93355cff6c4736862] Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git
git bisect bad d082aee133bf82bf52ffbbb93355cff6c4736862
# good: [1a65b0084768a0106ba6f37541f7216c447e49b2] Merge branch 'togreg' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git
git bisect good 1a65b0084768a0106ba6f37541f7216c447e49b2
# good: [0425cf50f5957e54852ea8f77baadae7823c0266] Merge branch 'spmi-next' of git://git.kernel.org/pub/scm/linux/kernel/git/sboyd/spmi.git
git bisect good 0425cf50f5957e54852ea8f77baadae7823c0266
# good: [1c9e16b73166bcc89b3d64674b60943ec6142c3e] staging: vc04_services: vchiq_arm: Split driver static and runtime data
git bisect good 1c9e16b73166bcc89b3d64674b60943ec6142c3e
# good: [98f2233a5c20ca567b2db1147278fd110681b9ed] dmaengine: pl08x: Use kcalloc() instead of kzalloc()
git bisect good 98f2233a5c20ca567b2db1147278fd110681b9ed
# bad: [8996f93fc3880d000a0a0cb40c724d5830e140fd] cgroup/cpuset: Statically initialize more members of top_cpuset
git bisect bad 8996f93fc3880d000a0a0cb40c724d5830e140fd
# good: [fc29e04ae1ad4c99422c0b8ae4b43cfe99c70429] cgroup/rstat: add cgroup_rstat_lock helpers and tracepoints
git bisect good fc29e04ae1ad4c99422c0b8ae4b43cfe99c70429
# good: [15a0b5fe1ad6fbecfa6517750718089e12ee8344] cgroup: don't call cgroup1_pidlist_destroy_all() for v2
git bisect good 15a0b5fe1ad6fbecfa6517750718089e12ee8344
# good: [f71bfbe1e281a707e7766eb0a59ba056dc0f29f9] cgroup, legacy_freezer: update comment for freezer_css_offline()
git bisect good f71bfbe1e281a707e7766eb0a59ba056dc0f29f9
# good: [19fc8a896565ecebb3951664fd0eeab0a80314a1] cgroup: Avoid unnecessary looping in cgroup_no_v1()
git bisect good 19fc8a896565ecebb3951664fd0eeab0a80314a1
# first bad commit: [8996f93fc3880d000a0a0cb40c724d5830e140fd] cgroup/cpuset: Statically initialize more members of top_cpuset

[-- Attachment #3: 0001-cgroup-cpuset-fix-bitwise-shift-in-top_cpuset-initia.patch --]
[-- Type: text/x-patch, Size: 1065 bytes --]

From dfc9b104bb62ac7d0546be445ce6add9e6d8c2fe Mon Sep 17 00:00:00 2001
From: Klara Modin <klarasmodin@gmail.com>
Date: Tue, 23 Apr 2024 21:00:56 +0200
Subject: [PATCH] cgroup/cpuset: fix bitwise shift in top_cpuset initialization

Fixes: 8996f93fc388 ("cgroup/cpuset: Statically initialize more members of top_cpuset")
Signed-off-by: Klara Modin <klarasmodin@gmail.com>
---
 kernel/cgroup/cpuset.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index e70008a1d86a..62d649d02f11 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -369,7 +369,7 @@ static inline void notify_partition_change(struct cpuset *cs, int old_prs)
 
 static struct cpuset top_cpuset = {
 	.flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
-		  (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),
+		  (1 << CS_MEM_EXCLUSIVE) | (1 << CS_SCHED_LOAD_BALANCE)),
 	.partition_root_state = PRS_ROOT,
 	.relax_domain_level = -1,
 	.remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
-- 
2.44.0

Re: [PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[Waiman Long]

On 4/23/24 15:21, Klara Modin wrote:
> Hi,
>
> On 2024-04-20 11:46, Xiu Jianfeng wrote:
>> Initializing top_cpuset.relax_domain_level and setting
>> CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
>> completed at the time of top_cpuset definition by compiler.
>>
>> Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
>> ---
>>   kernel/cgroup/cpuset.c | 5 ++---
>>   1 file changed, 2 insertions(+), 3 deletions(-)
>>
>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>> index d8d3439eda4e..e70008a1d86a 100644
>> --- a/kernel/cgroup/cpuset.c
>> +++ b/kernel/cgroup/cpuset.c
>> @@ -369,8 +369,9 @@ static inline void notify_partition_change(struct 
>> cpuset *cs, int old_prs)
>>     static struct cpuset top_cpuset = {
>>       .flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
>> -          (1 << CS_MEM_EXCLUSIVE)),
>> +          (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),
>
> You dropped a '<' for the bitwise shift, this causes bad cpu 
> utilization for me.

Oh, now load_balancing is disable by default in the root cgroup. That is 
bad. Will post a patch to fix it.

Thanks,
Longman



>
>>       .partition_root_state = PRS_ROOT,
>> +    .relax_domain_level = -1,
>>       .remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
>>   };
>>   @@ -4309,8 +4310,6 @@ int __init cpuset_init(void)
>>       nodes_setall(top_cpuset.effective_mems);
>>         fmeter_init(&top_cpuset.fmeter);
>> -    set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags);
>> -    top_cpuset.relax_domain_level = -1;
>>       INIT_LIST_HEAD(&remote_children);
>>         BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
>
> Kind regards,
> Klara Modin

Re: [PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[Klara Modin]

On 2024-04-23 21:50, Waiman Long wrote:
> 
> On 4/23/24 15:21, Klara Modin wrote:
>> Hi,
>>
>> On 2024-04-20 11:46, Xiu Jianfeng wrote:
>>> Initializing top_cpuset.relax_domain_level and setting
>>> CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
>>> completed at the time of top_cpuset definition by compiler.
>>>
>>> Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
>>> ---
>>>   kernel/cgroup/cpuset.c | 5 ++---
>>>   1 file changed, 2 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>>> index d8d3439eda4e..e70008a1d86a 100644
>>> --- a/kernel/cgroup/cpuset.c
>>> +++ b/kernel/cgroup/cpuset.c
>>> @@ -369,8 +369,9 @@ static inline void notify_partition_change(struct 
>>> cpuset *cs, int old_prs)
>>>     static struct cpuset top_cpuset = {
>>>       .flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
>>> -          (1 << CS_MEM_EXCLUSIVE)),
>>> +          (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),
>>
>> You dropped a '<' for the bitwise shift, this causes bad cpu 
>> utilization for me.
> 
> Oh, now load_balancing is disable by default in the root cgroup. That is 
> bad. Will post a patch to fix it.
> 
> Thanks,
> Longman
> 
I attached one in my previous message, though I'm still very new at this 
and may have missed something.

Regards,
Klara Modin

Re: [PATCH -next] cgroup/cpuset: Statically initialize more members of top_cpuset

[xiujianfeng]

On 2024/4/24 3:21, Klara Modin wrote:
> Hi,
> 
> On 2024-04-20 11:46, Xiu Jianfeng wrote:
>> Initializing top_cpuset.relax_domain_level and setting
>> CS_SCHED_LOAD_BALANCE to top_cpuset.flags in cpuset_init() could be
>> completed at the time of top_cpuset definition by compiler.
>>
>> Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
>> ---
>>   kernel/cgroup/cpuset.c | 5 ++---
>>   1 file changed, 2 insertions(+), 3 deletions(-)
>>
>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>> index d8d3439eda4e..e70008a1d86a 100644
>> --- a/kernel/cgroup/cpuset.c
>> +++ b/kernel/cgroup/cpuset.c
>> @@ -369,8 +369,9 @@ static inline void notify_partition_change(struct
>> cpuset *cs, int old_prs)
>>     static struct cpuset top_cpuset = {
>>       .flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) |
>> -          (1 << CS_MEM_EXCLUSIVE)),
>> +          (1 << CS_MEM_EXCLUSIVE) | (1 < CS_SCHED_LOAD_BALANCE)),
> 
> You dropped a '<' for the bitwise shift, this causes bad cpu utilization
> for me.

Oops, that's bad, I'm sorry for that!

> 
>>       .partition_root_state = PRS_ROOT,
>> +    .relax_domain_level = -1,
>>       .remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
>>   };
>>   @@ -4309,8 +4310,6 @@ int __init cpuset_init(void)
>>       nodes_setall(top_cpuset.effective_mems);
>>         fmeter_init(&top_cpuset.fmeter);
>> -    set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags);
>> -    top_cpuset.relax_domain_level = -1;
>>       INIT_LIST_HEAD(&remote_children);
>>         BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
> 
> Kind regards,
> Klara Modin