linux-kernel - Re: fuse uring / wake

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>] [day] [month] [year] [list]

Date:   Fri, 5 May 2023 13:10:26 +0000
From:   Bernd Schubert <bschubert@....com>
To:     Hillf Danton <hdanton@...a.com>
CC:     Peter Zijlstra <peterz@...radead.org>,
        Miklos Szeredi <miklos@...redi.hu>,
        K Prateek Nayak <kprateek.nayak@....com>,
        Andrei Vagin <avagin@...il.com>,
        "linux-mm@...ck.org" <linux-mm@...ck.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: fuse uring / wake_up on the same core

On 5/4/23 04:16, Hillf Danton wrote:
> 
>> +		    time_after(jiffies, p->seesaw_jiffies + 10),
>> +		    cpumask_test_cpu(cpu, p->cpus_ptr))
>> +			return cpu;

Above is a big typo, I don't even see on the first glance how this 
compiled at all.

This was supposed to be

if (p->seesaw_req && current->seesaw_proc &&
     time_after(jiffies, p->seesaw_jiffies + 10) &&
     cpumask_test_cpu(cpu, p->cpus_ptr))

Anyway, I now understand that the WF_TTWU flag is related to waitq - we 
don't need the timeout at all. But then if the main issue is about waitq 
migration, I don't understand yet why Andrei's WF_CURRENT_CPU is not 
sufficient. I'm going to investigate that next. Probably much easier to 
get that accepted that a rather fuse specific seesaw.

Thanks,
Bernd