netdev - [PATCH -mm] sunrpc svc: Shut up bogus uninitialized variable warning

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]

Message-ID: <alpine.LFD.0.999.0709030137390.29617@enigma.security.iitk.ac.in>
Date:	Mon, 3 Sep 2007 01:39:04 +0530 (IST)
From:	Satyam Sharma <satyam@...radead.org>
To:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
cc:	Greg Banks <gnb@...bourne.sgi.com>,
	Olaf Kirch <olaf.kirch@...cle.com>,
	David Miller <davem@...emloft.net>, netdev@...r.kernel.org
Subject: [PATCH -mm] sunrpc svc: Shut up bogus uninitialized variable
 warning


net/sunrpc/svc.c: In function ‘__svc_create_thread’:
net/sunrpc/svc.c:550: warning: ‘oldmask.bits[0u]’ may be used uninitialized in this function

is a bogus warning, but gcc isn't smart enough to see why. We cannot just
reorganize the code in the function, because we want the set_cpus_allowed()
restore to happen only after the kernel_thread() is forked. Alas, we have
to use cpus_clear() to initialize oldmask instead to keep gcc happy.

Also add some comments to describe what's happening in the function.

Signed-off-by: Satyam Sharma <satyam@...radead.org>

---

 net/sunrpc/svc.c |    7 +++++++
 1 file changed, 7 insertions(+)

--- linux-2.6.23-rc4-mm1/net/sunrpc/svc.c‾fix	2007-09-02 22:55:25.000000000 +0530
+++ linux-2.6.23-rc4-mm1/net/sunrpc/svc.c	2007-09-02 23:03:45.000000000 +0530
@@ -568,17 +568,24 @@ __svc_create_thread(svc_thread_fn func, 
 	rqstp->rq_server = serv;
 	rqstp->rq_pool = pool;
 
+	/* Only to prevent gcc warning */
+	cpus_clear(oldmask);
+
+	/* Temporarily change CPU affinity before forking svc thread */
 	if (serv->sv_nrpools > 1)
 		have_oldmask = svc_pool_map_set_cpumask(pool->sp_id, &oldmask);
 
+	/* Will inherit current->cpus_allowed */
 	error = kernel_thread((int (*)(void *)) func, rqstp, 0);
 
+	/* Restore old cpus_allowed */
 	if (have_oldmask)
 		set_cpus_allowed(current, oldmask);
 
 	if (error < 0)
 		goto out_thread;
 	svc_sock_update_bufs(serv);
+
 	error = 0;
 out:
 	return error;