netdev - Re: [PATCH v3 6/9] io_uring/cmd: Introduce SOCKET_URING_OP

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: Mon, 21 Aug 2023 02:09:26 -0700
From: Breno Leitao <leitao@...ian.org>
To: Gabriel Krisman Bertazi <krisman@...e.de>
Cc: sdf@...gle.com, axboe@...nel.dk, asml.silence@...il.com,
	willemdebruijn.kernel@...il.com, martin.lau@...ux.dev,
	bpf@...r.kernel.org, linux-kernel@...r.kernel.org,
	netdev@...r.kernel.org, io-uring@...r.kernel.org, kuba@...nel.org,
	pabeni@...hat.com
Subject: Re: [PATCH v3 6/9] io_uring/cmd: Introduce SOCKET_URING_OP_GETSOCKOPT

Hello Gabriel,

On Thu, Aug 17, 2023 at 02:38:46PM -0400, Gabriel Krisman Bertazi wrote:
> Breno Leitao <leitao@...ian.org> writes:
> 
> > +#if defined(CONFIG_NET)
> >  int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags)
> >  {
> >  	struct socket *sock = cmd->file->private_data;
> > @@ -189,8 +219,16 @@ int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags)
> >  		if (ret)
> >  			return ret;
> >  		return arg;
> > +	case SOCKET_URING_OP_GETSOCKOPT:
> > +		return io_uring_cmd_getsockopt(sock, cmd, issue_flags);
> >  	default:
> >  		return -EOPNOTSUPP;
> >  	}
> >  }
> > +#else
> > +int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags)
> > +{
> > +	return -EOPNOTSUPP;
> > +}
> > +#endif
> >  EXPORT_SYMBOL_GPL(io_uring_cmd_sock);
> 
> The CONFIG_NET guards are unrelated and need to go in a separate commit.
> Specially because it is not only gating getsockopt, but also the already
> merged SOCKET_URING_OP_SIOCINQ/_OP_SIOCOUTQ.

Well, so far, if CONFIG_NET is disable, and you call
io_uring_cmd_getsockopt, the callbacks will be called and returned
-EOPNOTSUPP.

With this new patch, it will eventually call sk_getsockopt which does
not exist in the CONFIG_NET=n world. That is why I have this
protection now. I.e, this `#if defined(CONFIG_NET)` is only necessary now,
since it is the first time this function (io_uring_cmd_sock) will call a
function that does not exist if CONFIG_NET is disabled.

I can split it in a different patch, if you think it makes a difference.