linux-kernel - Re: [GIT pull] x86/urgent for 5.3-rc5

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, 25 Aug 2019 20:29:22 +0200
From:   Borislav Petkov <bp@...e.de>
To:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Pu Wen <puwen@...on.cn>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        Tom Lendacky <thomas.lendacky@....com>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        the arch/x86 maintainers <x86@...nel.org>
Subject: Re: [GIT pull] x86/urgent for 5.3-rc5

On Sun, Aug 25, 2019 at 10:37:47AM -0700, Linus Torvalds wrote:
> On Sun, Aug 25, 2019 at 10:30 AM Borislav Petkov <bp@...e.de> wrote:
> >
> > Should we do that somewhere in the early boot code by adding a WARN_ON()
> > or so and see who screams?
> 
> It might be a good idea, just to see if it ever happens (again).
> 
> It doesn't even have to be early boot. It's probably more important to
> let the user _know_, than it is to then disable the rdrand
> instruction.
> 
> Particularly since we might as well just do it in general, and in the
> general case we don't even know how to hide it in cpuid. So maybe just
> something like "read the rdrand value a few times, make sure it
> actually changes" at CPU bring-up (both boot and resume)
> 
> It sounds like a stupid test, but considering that AMD has had this
> particular bug now several times over at least three different
> generations, maybe it's not a stupid test after all.

My lazy, sticky Sunday brain could come up only with this:

---
diff --git a/arch/x86/kernel/cpu/rdrand.c b/arch/x86/kernel/cpu/rdrand.c
index 5c900f9527ff..0130a4f4f836 100644
--- a/arch/x86/kernel/cpu/rdrand.c
+++ b/arch/x86/kernel/cpu/rdrand.c
@@ -29,7 +29,8 @@ __setup("nordrand", x86_rdrand_setup);
 #ifdef CONFIG_ARCH_RANDOM
 void x86_init_rdrand(struct cpuinfo_x86 *c)
 {
-	unsigned long tmp;
+	unsigned int changed = 0;
+	unsigned long tmp, prev;
 	int i;
 
 	if (!cpu_has(c, X86_FEATURE_RDRAND))
@@ -42,5 +43,24 @@ void x86_init_rdrand(struct cpuinfo_x86 *c)
 			return;
 		}
 	}
+
+	/*
+	 * Stupid sanity-check whether RDRAND does *actually* generate
+	 * some at least random-looking data.
+	 */
+	prev = tmp;
+	for (i = 0; i < SANITY_CHECK_LOOPS; i++) {
+		if (rdrand_long(&tmp)) {
+			if (prev != tmp)
+				changed++;
+
+			prev = tmp;
+		}
+	}
+
+	if (!changed)
+		WARN(1,
+"RDRAND gives funky smelling output, might consider not using it by booting with \"nordrand\"");
+
 }
 #endif

---

> Who knows what the Chinese CPU's that use the AMD core do? Hygon?
> Whatever. Did they get the firmware fixes?

Pu Wen?

-- 
Regards/Gruss,
    Boris.

SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer, HRB 247165, AG München