diff mbox series

[v40,17/24] x86/fault: Add helper function to sanitize error code

Message ID 20201104145430.300542-18-jarkko.sakkinen@linux.intel.com (mailing list archive)
State New, archived
Headers show
Series Intel SGX foundations | expand

Commit Message

Jarkko Sakkinen Nov. 4, 2020, 2:54 p.m. UTC
From: Sean Christopherson <sean.j.christopherson@intel.com>

vDSO exception fixup is a replacement for signals in limited situations.
Signals and vDSO exception fixup need to provide similar information to
userspace, including the hardware error code.

That hardware error code needs to be sanitized.  For instance, if userspace
accesses a kernel address, the error code could indicate to userspace
whether the address had a Present=1 PTE.  That can leak information about
the kernel layout to userspace, which is bad.

The existing signal code does this sanitization, but fairly late in the
signal process.  The vDSO exception code runs before the sanitization

Move error code sanitization out of the signal code and into a helper.
Call the helper in the signal code.

Acked-by: Jethro Beekman <jethro@fortanix.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Changes from v39:
* Add the missing change to the set_signal_archinfo() that removes the
  snippet contained in sanitize_error_code().

 arch/x86/mm/fault.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)
diff mbox series


diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 9339fee83784..0161d4acf3ad 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -602,11 +602,9 @@  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)
+static void sanitize_error_code(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
@@ -617,7 +615,13 @@  static void set_signal_archinfo(unsigned long address,
 	 * information and does not appear to cause any problems.
 	if (address >= TASK_SIZE_MAX)
-		error_code |= X86_PF_PROT;
+		*error_code |= X86_PF_PROT;
+static void set_signal_archinfo(unsigned long address,
+				unsigned long error_code)
+	struct task_struct *tsk = current;
 	tsk->thread.trap_nr = X86_TRAP_PF;
 	tsk->thread.error_code = error_code | X86_PF_USER;
@@ -658,6 +662,8 @@  no_context(struct pt_regs *regs, unsigned long error_code,
 		 * faulting through the emulate_vsyscall() logic.
 		if (current->thread.sig_on_uaccess_err && signal) {
+			sanitize_error_code(address, &error_code);
 			set_signal_archinfo(address, error_code);
 			/* XXX: hwpoison faults will set the wrong code. */
@@ -806,13 +812,7 @@  __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 		if (is_errata100(regs, address))
-		/*
-		 * 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;
+		sanitize_error_code(address, &error_code);
 		if (likely(show_unhandled_signals))
 			show_signal_msg(regs, error_code, address, tsk);
@@ -931,6 +931,8 @@  do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address,
 	if (is_prefetch(regs, error_code, address))
+	sanitize_error_code(address, &error_code);
 	set_signal_archinfo(address, error_code);