linux-kernel - Re: [PATCH 1/3] x86/execmem: don't use PAGE

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: <20250530074453.GG39944@noisy.programming.kicks-ass.net>
Date: Fri, 30 May 2025 09:44:53 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Mike Rapoport <rppt@...nel.org>
Cc: Juergen Gross <jgross@...e.com>, linux-kernel@...r.kernel.org,
	x86@...nel.org, xin@...or.com,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	Andy Lutomirski <luto@...nel.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
	"H. Peter Anvin" <hpa@...or.com>, stable@...r.kernel.org
Subject: Re: [PATCH 1/3] x86/execmem: don't use PAGE_KERNEL protection for
 code pages

On Wed, May 28, 2025 at 08:27:19PM +0300, Mike Rapoport wrote:
> On Wed, May 28, 2025 at 02:35:55PM +0200, Juergen Gross wrote:
> > In case X86_FEATURE_PSE isn't available (e.g. when running as a Xen
> > PV guest), execmem_arch_setup() will fall back to use PAGE_KERNEL
> > protection for the EXECMEM_MODULE_TEXT range.
> > 
> > This will result in attempts to execute code with the NX bit set in
> > case of ITS mitigation being applied.
> > 
> > Avoid this problem by using PAGE_KERNEL_EXEC protection instead,
> > which will not set the NX bit.
> > 
> > Cc: <stable@...r.kernel.org>
> > Reported-by: Xin Li <xin@...or.com>
> > Fixes: 5185e7f9f3bd ("x86/module: enable ROX caches for module text on 64 bit")
> > Signed-off-by: Juergen Gross <jgross@...e.com>
> > ---
> >  arch/x86/mm/init.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
> > index 7456df985d96..f5012ae31d8b 100644
> > --- a/arch/x86/mm/init.c
> > +++ b/arch/x86/mm/init.c
> > @@ -1089,7 +1089,7 @@ struct execmem_info __init *execmem_arch_setup(void)
> >  		pgprot = PAGE_KERNEL_ROX;
> >  		flags = EXECMEM_KASAN_SHADOW | EXECMEM_ROX_CACHE;
> >  	} else {
> > -		pgprot = PAGE_KERNEL;
> > +		pgprot = PAGE_KERNEL_EXEC;
> 
> Please don't. Everything except ITS can work with PAGE_KENREL so the fix
> should be on ITS side. 

Well, this is early vs post make_ro again.

Does something like so work for you?

---
diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index 7456df985d96..f5012ae31d8b 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -1089,7 +1089,7 @@ struct execmem_info __init *execmem_arch_setup(void)
 		pgprot = PAGE_KERNEL_ROX;
 		flags = EXECMEM_KASAN_SHADOW | EXECMEM_ROX_CACHE;
 	} else {
-		pgprot = PAGE_KERNEL;
+		pgprot = PAGE_KERNEL_EXEC;
 		flags = EXECMEM_KASAN_SHADOW;
 	}
 
diff --git a/mm/execmem.c b/mm/execmem.c
index 6f7a2653b280..dbe2eedea0e6 100644
--- a/mm/execmem.c
+++ b/mm/execmem.c
@@ -258,6 +258,7 @@ static bool execmem_cache_rox = false;
 
 void execmem_cache_make_ro(void)
 {
+	struct execmem_range *module_text = &execmem_info->ranges[EXECMEM_MODULE_TEXT];
 	struct maple_tree *free_areas = &execmem_cache.free_areas;
 	struct maple_tree *busy_areas = &execmem_cache.busy_areas;
 	MA_STATE(mas_free, free_areas, 0, ULONG_MAX);
@@ -269,6 +270,9 @@ void execmem_cache_make_ro(void)
 
 	mutex_lock(mutex);
 
+	if (!(module_text->flags & EXECMEM_ROX_CACHE))
+		module_text->pgprot = PAGE_KERNEL;
+
 	mas_for_each(&mas_free, area, ULONG_MAX) {
 		unsigned long pages = mas_range_len(&mas_free) >> PAGE_SHIFT;
 		set_memory_ro(mas_free.index, pages);