linux-kernel - Re: [PATCHv2 02/14] fork: allow arch-override of VMAP stack alignment

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: <20170815163036.GJ6090@leverpostej>
Date:   Tue, 15 Aug 2017 17:30:36 +0100
From:   Mark Rutland <mark.rutland@....com>
To:     Andy Lutomirski <luto@...capital.net>
Cc:     "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        Catalin Marinas <catalin.marinas@....com>,
        James Morse <james.morse@....com>,
        Laura Abbott <labbott@...hat.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Matt Fleming <matt@...eblueprint.co.uk>,
        Will Deacon <will.deacon@....com>,
        "kernel-hardening@...ts.openwall.com" 
        <kernel-hardening@...ts.openwall.com>,
        Kees Cook <keescook@...omium.org>
Subject: Re: [PATCHv2 02/14] fork: allow arch-override of VMAP stack alignment

On Tue, Aug 15, 2017 at 09:09:36AM -0700, Andy Lutomirski wrote:
> On Tue, Aug 15, 2017 at 5:50 AM, Mark Rutland <mark.rutland@....com> wrote:
> > In some cases, an architecture might wish its stacks to be aligned to a
> > boundary larger than THREAD_SIZE. For example, using an alignment of
> > double THREAD_SIZE can allow for stack overflows smaller than
> > THREAD_SIZE to be detected by checking a single bit of the stack
> > pointer.
> >
> > This patch allows architectures to override the alignment of VMAP'd
> > stacks, by defining THREAD_ALIGN. Where not defined, this defaults to
> > THREAD_SIZE, as is the case today.
> 
> This looks okay, but it might make sense to move that to a header file
> so THREAD_ALIGN is always available.

I was a little worried about breaking things, since arches don't define
THREAD_SIZE in a consistent location.

Looking again, it looks like those are all transitiviely included into
each arch's <asm/thread_info.h>, so I think I can move this into
<linux/thread_info.h>, which'll have to be added to kernel.fork.c's
includes.

Are you happy with the below fixup?

Thanks,
Mark.

---->8----
diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h
index 250a276..905d769 100644
--- a/include/linux/thread_info.h
+++ b/include/linux/thread_info.h
@@ -38,6 +38,10 @@ enum {
 
 #ifdef __KERNEL__
 
+#ifndef THREAD_ALIGN
+#define THREAD_ALIGN   THREAD_SIZE
+#endif
+
 #ifdef CONFIG_DEBUG_STACK_USAGE
 # define THREADINFO_GFP                (GFP_KERNEL_ACCOUNT | __GFP_NOTRACK | \
                                 __GFP_ZERO)
diff --git a/kernel/fork.c b/kernel/fork.c
index 696d692..f12882a 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -88,6 +88,7 @@
 #include <linux/sysctl.h>
 #include <linux/kcov.h>
 #include <linux/livepatch.h>
+#include <linux/thread_info.h>
 
 #include <asm/pgtable.h>
 #include <asm/pgalloc.h>
@@ -217,9 +218,6 @@ static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node)
                return s->addr;
        }
 
-#ifndef THREAD_ALIGN
-#define THREAD_ALIGN   THREAD_SIZE
-#endif
        stack = __vmalloc_node_range(THREAD_SIZE, THREAD_ALIGN,
                                     VMALLOC_START, VMALLOC_END,
                                     THREADINFO_GFP,