linux-kernel - Re: [PATCH] debugobjects: Move printk out of db lock critical sections

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: <20181213090026.GA7278@jagdpanzerIV>
Date:   Thu, 13 Dec 2018 18:00:27 +0900
From:   Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>
To:     Dmitry Safonov <0x7f454c46@...il.com>
Cc:     sergey.senozhatsky.work@...il.com,
        Dmitry Safonov <dima@...sta.com>, longman@...hat.com,
        Thomas Gleixner <tglx@...utronix.de>,
        Andrew Morton <akpm@...ux-foundation.org>,
        open list <linux-kernel@...r.kernel.org>,
        Ingo Molnar <mingo@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        yang.shi@...ux.alibaba.com, Arnd Bergmann <arnd@...db.de>
Subject: Re: [PATCH] debugobjects: Move printk out of db lock critical
 sections

On (12/13/18 05:54), Dmitry Safonov wrote:
> > I thought about it for a second, but couldn't figure out if this race
> > was real.
> >
> > E.g. uart case - if there are two paths which concurrently free and
> > access debug object, then the same race condition should exist for
> > the xmit.buf page. debug object's lifespan should be the same as the
> > lifespan of, ummm, whatever the thing that debug object is attached
> > to? I'm surely missing something here.
> 
> I think the point of debugobjects is to reduce the damage caused
> by an improper life-time of objects.
> 
> IIUC:
> 1. alloc_page()
> 2. dynamically allocate a workqueue on a new page
> 3. schedule some work
> 4. until the work has run free_page()

Hmm, if we free() that page then __debug_check_no_obj_freed() should ->fixup
all ODEBUG_STATE_ACTIVE objects first. Maybe we are talking about different
scenarios.

	-ss