linux-kernel - Re: [PATCH 03/30] thread_info: tif_need_resched() now takes resched

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]

Date: Tue, 20 Feb 2024 14:21:16 -0800
From: Ankur Arora <ankur.a.arora@...cle.com>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: Mark Rutland <mark.rutland@....com>,
        Ankur Arora
 <ankur.a.arora@...cle.com>, linux-kernel@...r.kernel.org,
        peterz@...radead.org, torvalds@...ux-foundation.org,
        paulmck@...nel.org, akpm@...ux-foundation.org, luto@...nel.org,
        bp@...en8.de, dave.hansen@...ux.intel.com, hpa@...or.com,
        mingo@...hat.com, juri.lelli@...hat.com, vincent.guittot@...aro.org,
        willy@...radead.org, mgorman@...e.de, jpoimboe@...nel.org,
        jgross@...e.com, andrew.cooper3@...rix.com, bristot@...nel.org,
        mathieu.desnoyers@...icios.com, geert@...ux-m68k.org,
        glaubitz@...sik.fu-berlin.de, anton.ivanov@...bridgegreys.com,
        mattst88@...il.com, krypton@...ich-teichert.org, rostedt@...dmis.org,
        David.Laight@...lab.com, richard@....at, mjguzik@...il.com,
        jon.grimm@....com, bharata@....com, raghavendra.kt@....com,
        boris.ostrovsky@...cle.com, konrad.wilk@...cle.com,
        Arnd Bergmann
 <arnd@...db.de>,
        "Rafael J. Wysocki" <rafael@...nel.org>
Subject: Re: [PATCH 03/30] thread_info: tif_need_resched() now takes
 resched_t as param


Thomas Gleixner <tglx@...utronix.de> writes:

> On Wed, Feb 14 2024 at 14:08, Mark Rutland wrote:
>> On Mon, Feb 12, 2024 at 09:55:27PM -0800, Ankur Arora wrote:
>>>
>>> -static __always_inline bool tif_need_resched(void)
>>> +static __always_inline bool __tif_need_resched(int nr_flag)
>>>  {
>>> -	return test_bit(TIF_NEED_RESCHED,
>>> -			(unsigned long *)(&current_thread_info()->flags));
>>> +	return test_bit(nr_flag,
>>> +		(unsigned long *)(&current_thread_info()->flags));
>>>  }
>>>
>>>  #endif /* _ASM_GENERIC_BITOPS_INSTRUMENTED_NON_ATOMIC_H */
>>>
>>> +static __always_inline bool tif_need_resched(resched_t rs)
>>> +{
>>> +	/*
>>> +	 * With !PREEMPT_AUTO tif_need_resched(NR_lazy) is defined
>>> +	 * as TIF_NEED_RESCHED (the TIF_NEED_RESCHED_LAZY flag is not
>>> +	 * defined). Return false in that case.
>>> +	 */
>>> +	if (IS_ENABLED(CONFIG_PREEMPT_AUTO) || rs == NR_now)
>>> +		return __tif_need_resched(tif_resched(rs));
>>> +	else
>>> +		return false;
>>> +}
>>
>> As above, I think this would be a bit simpler/clearer if we did:
>>
>> static __always_inline bool tif_need_resched_now(void)
>> {
>> 	return __tif_need_resched(TIF_NEED_RESCHED);
>> }
>>
>> static __always_inline bool tif_need_resched_lazy(void)
>> {
>> 	return IS_ENABLED(CONFIG_PREEMPT_AUTO) &&
>> 	        __tif_need_resched(TIF_NEED_RESCHED_LAZY);
>> }
>
> Yes please.

As I wrote to Mark in the sibling subthread, I think exposing
the lazy variants outside of the scheduler isn't really needed.

Non-scheduler/non-entry code only cares about checking if it needs
to do reschedule now. So, I think we can get away with just having:

    static __always_inline bool __tif_need_resched(resched_t rs)
    {
            /*
            * With !PREEMPT_AUTO tif_need_resched(NR_lazy) is defined
            * as TIF_NEED_RESCHED (the TIF_NEED_RESCHED_LAZY flag is not
            * defined). Return false in that case.
            */
            if (IS_ENABLED(CONFIG_PREEMPT_AUTO) || rs == NR_now)
                    return tif_need_resched_bitop(tif_resched(rs));
            else
                    return false;
    }

    static __always_inline bool tif_need_resched(void)
    {
            return __tif_need_resched(NR_now);
    }

(and similar for all the other interfaces.)

This way lazy and eager just becomes an implementation detail which
which seems to also match nicely to the point of PREEMPT_AUTO.

--
ankur