linux-kernel - Re: 4.15-rc6 PTI regression: L1 TLB mismatch MCE on Athlon64

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: <alpine.DEB.2.20.1801031407180.1957@nanos>
Date:   Wed, 3 Jan 2018 14:17:00 +0100 (CET)
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Borislav Petkov <bp@...en8.de>
cc:     Meelis Roos <mroos@...ux.ee>,
        Linux Kernel list <linux-kernel@...r.kernel.org>,
        x86@...nel.org, linux-edac@...r.kernel.org,
        Tom Lendacky <thomas.lendacky@....com>
Subject: Re: 4.15-rc6 PTI regression: L1 TLB mismatch MCE on Athlon64

On Wed, 3 Jan 2018, Borislav Petkov wrote:
> On Wed, Jan 03, 2018 at 11:16:48AM +0200, Meelis Roos wrote:
> > ---[ ESPfix Area ]---
> > 0xffffff0000000000-0xffffff1800000000          96G                               pud
> > 0xffffff1800000000-0xffffff1800009000          36K                               pte
> > 0xffffff1800009000-0xffffff180000a000           4K     ro                     NX pte
> > 0xffffff180000a000-0xffffff1800019000          60K                               pte
> > 0xffffff1800019000-0xffffff180001a000           4K     ro                     NX pte
> > 0xffffff180001a000-0xffffff1800029000          60K                               pte
> > 0xffffff1800029000-0xffffff180002a000           4K     ro                     NX pte
> > 0xffffff180002a000-0xffffff1800039000          60K                               pte
> > 0xffffff1800039000-0xffffff180003a000           4K     ro                     NX pte
> > 0xffffff180003a000-0xffffff1800049000          60K                               pte
> > 0xffffff1800049000-0xffffff180004a000           4K     ro                     NX pte
> > 0xffffff180004a000-0xffffff1800059000          60K                               pte
> > 0xffffff1800059000-0xffffff180005a000           4K     ro                     NX pte
> > 0xffffff180005a000-0xffffff1800069000          60K                               pte
> > 0xffffff1800069000-0xffffff180006a000           4K     ro                     NX pte
> > 0xffffff180006a000-0xffffff1800079000          60K                               pte
> > ... 131059 entries skipped ... 
> > ---[ High Kernel Mapping ]---
> > 0xffffffff80000000-0xffffffff81e00000          30M                               pmd
> > 0xffffffff81e00000-0xffffffff82000000           2M     ro         PSE     GLB x  pmd
> 
> Ha, this must be it. What I said yesterday about the guard hole
> hypervisor range was wrong because we're looking at VA slice [47:12] and
> this one matches:

That's the entry area, which is mapped into kernel _AND_ user space. Now
that's special because we switch CR3 while we are executing there.

And this one is:

0xffffffff81e00000-0xffffffff82000000           2M     ro         PSE     GLB x  pmd

and the one we switch to is:

0xffffffff81000000-0xffffffff82000000          16M     ro         PSE         x  pmd

Meelis, does the patch below fix it for you?

Thanks,

	tglx

8<-------------------

--- a/arch/x86/mm/pti.c
+++ b/arch/x86/mm/pti.c
@@ -367,7 +367,8 @@ static void __init pti_setup_espfix64(vo
 static void __init pti_clone_entry_text(void)
 {
 	pti_clone_pmds((unsigned long) __entry_text_start,
-			(unsigned long) __irqentry_text_end, _PAGE_RW);
+			(unsigned long) __irqentry_text_end,
+		       _PAGE_RW | _PAGE_GLOBAL);
 }
 
 /*