netdev - Re: [PATCH 3/8] rhashtable: use cmpxchg() to protect ->future

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:   Sun, 06 May 2018 07:45:43 +1000
From:   NeilBrown <neilb@...e.com>
To:     Herbert Xu <herbert@...dor.apana.org.au>
Cc:     Thomas Graf <tgraf@...g.ch>, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/8] rhashtable: use cmpxchg() to protect ->future_tbl.

On Sat, May 05 2018, Herbert Xu wrote:

> On Fri, May 04, 2018 at 01:54:14PM +1000, NeilBrown wrote:
>> Rather than borrowing one of the bucket locks to
>> protect ->future_tbl updates, use cmpxchg().
>> This gives more freedom to change how bucket locking
>> is implemented.
>> 
>> Signed-off-by: NeilBrown <neilb@...e.com>
>
> This looks nice.
>
>> -	spin_unlock_bh(old_tbl->locks);
>> +	rcu_assign_pointer(tmp, new_tbl);
>
> Do we need this barrier since cmpxchg is supposed to provide memory
> barrier semantics?

It's hard to find documentation even for what cmpxchg() is meant do, let
alone what barriers is provides, but there does seem to be something
hidden in Documentation/core-api/atomic_ops.rst which suggests full
barrier semantics if the comparison succeeds.  I'll replace the
rcu_assign_pointer with a comment saying why it isn't needed.

Thanks,
NeilBrown

>
>> +	if (cmpxchg(&old_tbl->future_tbl, NULL, tmp) != NULL)
>> +		return -EEXIST;
>
> Thanks,
> -- 
> Email: Herbert Xu <herbert@...dor.apana.org.au>
> Home Page: http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Download attachment "signature.asc" of type "application/pgp-signature" (833 bytes)