netdev - Re: [PATCH v2 01/20] ethernet: alteon: convert tasklets to use new tasklet

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:   Wed, 09 Sep 2020 14:33:24 -0700 (PDT)
From:   David Miller <davem@...emloft.net>
To:     allen.lkml@...il.com
Cc:     jes@...ined-monkey.org, kuba@...nel.org, dougmill@...ux.ibm.com,
        cooldavid@...ldavid.org, mlindner@...vell.com,
        stephen@...workplumber.org, borisp@...lanox.com,
        netdev@...r.kernel.org, romain.perier@...il.com
Subject: Re: [PATCH v2 01/20] ethernet: alteon: convert tasklets to use new
 tasklet_setup() API

From: Allen <allen.lkml@...il.com>
Date: Thu, 10 Sep 2020 00:06:47 +0530

>>
>> > @@ -1562,10 +1562,11 @@ static void ace_watchdog(struct net_device *data, unsigned int txqueue)
>> >  }
>> >
>> >
>> > -static void ace_tasklet(unsigned long arg)
>> > +static void ace_tasklet(struct tasklet_struct *t)
>> >  {
>> > -     struct net_device *dev = (struct net_device *) arg;
>> > -     struct ace_private *ap = netdev_priv(dev);
>> > +     struct ace_private *ap = from_tasklet(ap, t, ace_tasklet);
>> > +     struct net_device *dev = (struct net_device *)((char *)ap -
>> > +                             ALIGN(sizeof(struct net_device), NETDEV_ALIGN));
>> >       int cur_size;
>> >
>>
>> I don't see this is as an improvement.  The 'dev' assignment looks so
>> incredibly fragile and exposes so many internal details about netdev
>> object allocation, alignment, and layout.
>>
>> Who is going to find and fix this if someone changes how netdev object
>> allocation works?
>>
> 
> Thanks for pointing it out. I'll see if I can fix it to keep it simple.

Just add a backpointer to the netdev from the netdev_priv() if you
absolutely have too.

>> I don't want to apply this, it sets a very bad precedent.  The existing
>> code is so much cleaner and easier to understand and audit.
> 
> Will you pick the rest of the patches or would they have to wait till
> this one is
> fixed.

I never pick up a partial series, ever.  So yes you will have to fix these
two patches up and resubmit the entire thing.