linux-kernel - Re: [PATCH v5 1/1] sched/uclamp: add SCHED_FLAG_UTIL_CLAMP

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <82f26263-a5cd-bd5c-52bd-37b98663a2e5@arm.com>
Date:   Thu, 12 Nov 2020 14:05:46 +0100
From:   Dietmar Eggemann <dietmar.eggemann@....com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Yun Hsiang <hsiang023167@...il.com>, linux-kernel@...r.kernel.org,
        qais.yousef@....com, patrick.bellasi@...bug.net,
        kernel test robot <lkp@...el.com>
Subject: Re: [PATCH v5 1/1] sched/uclamp: add SCHED_FLAG_UTIL_CLAMP_RESET flag
 to reset uclamp

On 11/11/2020 19:04, Peter Zijlstra wrote:
> On Wed, Nov 11, 2020 at 06:41:07PM +0100, Dietmar Eggemann wrote:
>> diff --git a/include/uapi/linux/sched/types.h b/include/uapi/linux/sched/types.h
>> index c852153ddb0d..b9165f17dddc 100644
>> --- a/include/uapi/linux/sched/types.h
>> +++ b/include/uapi/linux/sched/types.h
>> @@ -115,8 +115,8 @@ struct sched_attr {
>>  	__u64 sched_period;
>>  
>>  	/* Utilization hints */
>> -	__u32 sched_util_min;
>> -	__u32 sched_util_max;
>> +	__s32 sched_util_min;
>> +	__s32 sched_util_max;
> 
> So that's UAPI, not sure we can change the type here.

Yes, will remove this chunk. Not needed.

I probably should add some documentation there:

diff --git a/include/uapi/linux/sched/types.h
b/include/uapi/linux/sched/types.h
index c852153ddb0d..f2c4589d4dbf 100644
--- a/include/uapi/linux/sched/types.h
+++ b/include/uapi/linux/sched/types.h
@@ -96,6 +96,8 @@ struct sched_param {
  * on a CPU with a capacity big enough to fit the specified value.
  * A task with a max utilization value smaller than 1024 is more likely
  * scheduled on a CPU with no more capacity than the specified value.
+ *
+ * A task utilization boundary can be reset by setting the attribute to -1.
  */
 struct sched_attr {
        __u32 size;

[...]

>> +	if (attr->sched_flags & SCHED_FLAG_UTIL_CLAMP_MAX) {
>> +		util_max = attr->sched_util_max;
>> +
>> +		if (util_max < -1 || util_max > SCHED_CAPACITY_SCALE)
>> +			return -EINVAL;
>> +	}
> 
> Luckily we can write that range as a single branch like:
> 
> 	if (util_{min,max} + 1 > SCHED_CAPACITY_SCALE+1)
> 
> which assumes u32 :-)

Cool, will change it.

>> +
>> +	if (util_min != -1 && util_max != -1 && util_min > util_max)
>>  		return -EINVAL;
> 
> I think that will compile as is, otherwise write it like ~0u, which is
> the same bit pattern.

Yes, it compiles for me (arm64, gcc 9.2 and arm, gcc 8.3). Started a
0-Day build job to make sure.

Will do some more testing before sending out the updated version.