linux-kernel - [tip:x86/mm] x86/fault: Make error

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:   Tue, 20 Nov 2018 00:17:24 -0800
From:   tip-bot for Andy Lutomirski <tipbot@...or.com>
To:     linux-tip-commits@...r.kernel.org
Cc:     luto@...nel.org, tglx@...utronix.de, yu-cheng.yu@...el.com,
        dvlasenk@...hat.com, peterz@...radead.org, riel@...riel.com,
        brgerst@...il.com, hpa@...or.com, mingo@...nel.org,
        luto@...capital.net, dave.hansen@...ux.intel.com,
        linux-kernel@...r.kernel.org, bp@...en8.de,
        torvalds@...ux-foundation.org
Subject: [tip:x86/mm] x86/fault: Make error_code sanitization more robust

Commit-ID:  e49d3cbef0176c182b86206185f137a87f16ab91
Gitweb:     https://git.kernel.org/tip/e49d3cbef0176c182b86206185f137a87f16ab91
Author:     Andy Lutomirski <luto@...nel.org>
AuthorDate: Mon, 19 Nov 2018 14:45:31 -0800
Committer:  Ingo Molnar <mingo@...nel.org>
CommitDate: Tue, 20 Nov 2018 08:44:29 +0100

x86/fault: Make error_code sanitization more robust

The error code in a page fault on a kernel address indicates
whether that address is mapped, which should not be revealed in a signal.

The normal code path for a page fault on a kernel address sanitizes the bit,
but the paths for vsyscall emulation and SIGBUS do not.  Both are
harmless, but for subtle reasons.  SIGBUS is never sent for a kernel
address, and vsyscall emulation will never fault on a kernel address
per se because it will fail an access_ok() check instead.

Make the code more robust by adding a helper that sets the relevant
fields and sanitizing the error code in the helper.  This also
cleans up the code -- we had three copies of roughly the same thing.

Signed-off-by: Andy Lutomirski <luto@...nel.org>
Cc: Andy Lutomirski <luto@...capital.net>
Cc: Borislav Petkov <bp@...en8.de>
Cc: Brian Gerst <brgerst@...il.com>
Cc: Dave Hansen <dave.hansen@...ux.intel.com>
Cc: Denys Vlasenko <dvlasenk@...hat.com>
Cc: H. Peter Anvin <hpa@...or.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Peter Zijlstra <peterz@...radead.org>
Cc: Rik van Riel <riel@...riel.com>
Cc: Thomas Gleixner <tglx@...utronix.de>
Cc: Yu-cheng Yu <yu-cheng.yu@...el.com>
Link: http://lkml.kernel.org/r/b31159bd55bd0c4fa061a20dfd6c429c094bebaa.1542667307.git.luto@kernel.org
Signed-off-by: Ingo Molnar <mingo@...nel.org>
---
 arch/x86/mm/fault.c | 30 +++++++++++++++++++++---------
 1 file changed, 21 insertions(+), 9 deletions(-)

diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 3c9aed03d18e..b5ec1ca2f4a0 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -631,6 +631,24 @@ pgtable_bad(struct pt_regs *regs, unsigned long error_code,
 	oops_end(flags, regs, sig);
 }
 
+static void set_signal_archinfo(unsigned long address,
+				unsigned long error_code)
+{
+	struct task_struct *tsk = current;
+
+	/*
+	 * To avoid leaking information about the kernel page
+	 * table layout, pretend that user-mode accesses to
+	 * kernel addresses are always protection faults.
+	 */
+	if (address >= TASK_SIZE_MAX)
+		error_code |= X86_PF_PROT;
+
+	tsk->thread.trap_nr = X86_TRAP_PF;
+	tsk->thread.error_code = error_code | X86_PF_USER;
+	tsk->thread.cr2 = address;
+}
+
 static noinline void
 no_context(struct pt_regs *regs, unsigned long error_code,
 	   unsigned long address, int signal, int si_code)
@@ -656,9 +674,7 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 		 * faulting through the emulate_vsyscall() logic.
 		 */
 		if (current->thread.sig_on_uaccess_err && signal) {
-			tsk->thread.trap_nr = X86_TRAP_PF;
-			tsk->thread.error_code = error_code | X86_PF_USER;
-			tsk->thread.cr2 = address;
+			set_signal_archinfo(address, error_code);
 
 			/* XXX: hwpoison faults will set the wrong code. */
 			force_sig_fault(signal, si_code, (void __user *)address,
@@ -821,9 +837,7 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 		if (likely(show_unhandled_signals))
 			show_signal_msg(regs, error_code, address, tsk);
 
-		tsk->thread.cr2		= address;
-		tsk->thread.error_code	= error_code;
-		tsk->thread.trap_nr	= X86_TRAP_PF;
+		set_signal_archinfo(address, error_code);
 
 		if (si_code == SEGV_PKUERR)
 			force_sig_pkuerr((void __user *)address, pkey);
@@ -937,9 +951,7 @@ do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address,
 	if (is_prefetch(regs, error_code, address))
 		return;
 
-	tsk->thread.cr2		= address;
-	tsk->thread.error_code	= error_code;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
+	set_signal_archinfo(address, error_code);
 
 #ifdef CONFIG_MEMORY_FAILURE
 	if (fault & (VM_FAULT_HWPOISON|VM_FAULT_HWPOISON_LARGE)) {