netdev - Re: [PATCH 01/10] x86/mm/numa: Remove uninitialized

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: <202006040728.8797FAA4@keescook>
Date:   Thu, 4 Jun 2020 07:34:54 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     linux-kernel@...r.kernel.org,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Miguel Ojeda <miguel.ojeda.sandonis@...il.com>,
        Alexander Potapenko <glider@...gle.com>,
        Joe Perches <joe@...ches.com>,
        Andy Whitcroft <apw@...onical.com>, x86@...nel.org,
        drbd-dev@...ts.linbit.com, linux-block@...r.kernel.org,
        b43-dev@...ts.infradead.org, netdev@...r.kernel.org,
        linux-wireless@...r.kernel.org, linux-ide@...r.kernel.org,
        linux-clk@...r.kernel.org, linux-spi@...r.kernel.org,
        linux-mm@...ck.org, clang-built-linux@...glegroups.com
Subject: Re: [PATCH 01/10] x86/mm/numa: Remove uninitialized_var() usage

On Thu, Jun 04, 2020 at 09:58:07AM +0200, Thomas Gleixner wrote:
> Kees Cook <keescook@...omium.org> writes:
> > -#ifdef NODE_NOT_IN_PAGE_FLAGS
> > -	pfn_align = node_map_pfn_alignment();
> > -	if (pfn_align && pfn_align < PAGES_PER_SECTION) {
> > -		printk(KERN_WARNING "Node alignment %LuMB < min %LuMB, rejecting NUMA config\n",
> > -		       PFN_PHYS(pfn_align) >> 20,
> > -		       PFN_PHYS(PAGES_PER_SECTION) >> 20);
> > -		return -EINVAL;
> > +	if (IS_ENABLED(NODE_NOT_IN_PAGE_FLAGS)) {
> 
> Hrm, clever ...
> 
> > +		unsigned long pfn_align = node_map_pfn_alignment();
> > +
> > +		if (pfn_align && pfn_align < PAGES_PER_SECTION) {
> > +			pr_warn("Node alignment %LuMB < min %LuMB, rejecting NUMA config\n",
> > +				PFN_PHYS(pfn_align) >> 20,
> > +				PFN_PHYS(PAGES_PER_SECTION) >> 20);
> > +			return -EINVAL;
> > +		}
> >  	}
> > -#endif
> >  	if (!numa_meminfo_cover_memory(mi))
> >  		return -EINVAL;
> >  
> > diff --git a/include/linux/page-flags-layout.h b/include/linux/page-flags-layout.h
> > index 71283739ffd2..1a4cdec2bd29 100644
> > --- a/include/linux/page-flags-layout.h
> > +++ b/include/linux/page-flags-layout.h
> > @@ -100,7 +100,7 @@
> >   * there.  This includes the case where there is no node, so it is implicit.
> >   */
> >  #if !(NODES_WIDTH > 0 || NODES_SHIFT == 0)
> > -#define NODE_NOT_IN_PAGE_FLAGS
> > +#define NODE_NOT_IN_PAGE_FLAGS 1
> 
> but if we ever lose the 1 then the above will silently compile the code
> within the IS_ENABLED() section out.

That's true, yes. I considered two other ways to do this:

1) smallest patch, but more #ifdef:

diff --git a/arch/x86/mm/numa.c b/arch/x86/mm/numa.c
index 59ba008504dc..fbf5231a3d35 100644
--- a/arch/x86/mm/numa.c
+++ b/arch/x86/mm/numa.c
@@ -541,7 +541,9 @@ static void __init numa_clear_kernel_node_hotplug(void)
 
 static int __init numa_register_memblks(struct numa_meminfo *mi)
 {
-	unsigned long uninitialized_var(pfn_align);
+#ifdef NODE_NOT_IN_PAGE_FLAGS
+	unsigned long pfn_align;
+#endif
 	int i, nid;
 
 	/* Account for nodes with cpus and no memory */

2) medium size, weird style:

diff --git a/arch/x86/mm/numa.c b/arch/x86/mm/numa.c
index 59ba008504dc..0df7ba9b21b2 100644
--- a/arch/x86/mm/numa.c
+++ b/arch/x86/mm/numa.c
@@ -541,7 +541,6 @@ static void __init numa_clear_kernel_node_hotplug(void)
 
 static int __init numa_register_memblks(struct numa_meminfo *mi)
 {
-	unsigned long uninitialized_var(pfn_align);
 	int i, nid;
 
 	/* Account for nodes with cpus and no memory */
@@ -570,12 +569,15 @@ static int __init numa_register_memblks(struct numa_meminfo *mi)
 	 * whether its granularity is fine enough.
 	 */
 #ifdef NODE_NOT_IN_PAGE_FLAGS
-	pfn_align = node_map_pfn_alignment();
-	if (pfn_align && pfn_align < PAGES_PER_SECTION) {
-		printk(KERN_WARNING "Node alignment %LuMB < min %LuMB, rejecting NUMA config\n",
-		       PFN_PHYS(pfn_align) >> 20,
-		       PFN_PHYS(PAGES_PER_SECTION) >> 20);
-		return -EINVAL;
+	{
+		unsigned long pfn_align = node_map_pfn_alignment();
+
+		if (pfn_align && pfn_align < PAGES_PER_SECTION) {
+			pr_warn("Node alignment %LuMB < min %LuMB, rejecting NUMA config\n",
+			       PFN_PHYS(pfn_align) >> 20,
+			       PFN_PHYS(PAGES_PER_SECTION) >> 20);
+			return -EINVAL;
+		}
 	}
 #endif
 	if (!numa_meminfo_cover_memory(mi))

and 3 is what I sent: biggest, but removes #ifdef

Any preference?

Thanks!

-- 
Kees Cook