Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Dave Hansen <dave.hansen@linux.intel.com>
To:     linux-kernel@vger.kernel.org
Cc:     Dave Hansen <dave.hansen@linux.intel.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
        Kees Cook <keescook@chromium.org>,
        Peter Zijlstra <peterz@infradead.org>
Subject: [PATCH] x86/mm: Ease W^X enforcement back to just a warning
Date:   Mon,  3 Oct 2022 14:06:59 -0700
Message-Id: <20221003210659.3837868-1-dave.hansen@linux.intel.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

I'd like to get to a point where we can turn on W^X enforcement
and keep it on.  But, we're not quite there yet.  I'm planning
on applying this to x86/mm before it goes to Linus during this
merge window.  Any objections?

--

Currently, the "change_page_attr" (CPA) code refuses to create
W+X mappings on 64-bit kernels.  There have been reports both
from 32-bit[1] and from BPF[2] users where this change kept the
system from booting.

These reports are showing up even after about a month of soak
time in -next.

To avoid breaking anything, never enforce W^X.  Always warn
and return the requested permissions even if a problem is
detected.

1. https://lore.kernel.org/all/CAMj1kXHcF_iK_g0OZSkSv56Wmr=eQGQwNstcNjLEfS=mm7a06w@mail.gmail.com/
2. https://lore.kernel.org/bpf/c84cc27c1a5031a003039748c3c099732a718aec.camel@kernel.org/T/#u

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: x86@kernel.org
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 arch/x86/mm/pat/set_memory.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index efe882c753ca..97342c42dda8 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -580,7 +580,7 @@ static inline pgprot_t static_protections(pgprot_t prot, unsigned long start,
 }
 
 /*
- * Validate and enforce strict W^X semantics.
+ * Validate strict W^X semantics.
  */
 static inline pgprot_t verify_rwx(pgprot_t old, pgprot_t new, unsigned long start,
 				  unsigned long pfn, unsigned long npg)
@@ -595,7 +595,7 @@ static inline pgprot_t verify_rwx(pgprot_t old, pgprot_t new, unsigned long star
 	if (IS_ENABLED(CONFIG_X86_32))
 		return new;
 
-	/* Only enforce when NX is supported: */
+	/* Only verify when NX is supported: */
 	if (!(__supported_pte_mask & _PAGE_NX))
 		return new;
 
@@ -606,13 +606,17 @@ static inline pgprot_t verify_rwx(pgprot_t old, pgprot_t new, unsigned long star
 		return new;
 
 	end = start + npg * PAGE_SIZE - 1;
-	WARN_ONCE(1, "CPA refuse W^X violation: %016llx -> %016llx range: 0x%016lx - 0x%016lx PFN %lx\n",
+	WARN_ONCE(1, "CPA detected W^X violation: %016llx -> %016llx range: 0x%016lx - 0x%016lx PFN %lx\n",
 		  (unsigned long long)pgprot_val(old),
 		  (unsigned long long)pgprot_val(new),
 		  start, end, pfn);
 
-	/* refuse the transition into WX */
-	return old;
+	/*
+	 * For now, allow all permission change attempts by returning the
+	 * attempted permissions.  This can 'return old' to actively
+	 * refuse the permission change at a later time.
+	 */
+	return new;
 }
 
 /*
-- 
2.34.1