Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 380851bc authored by Andi Kleen's avatar Andi Kleen Committed by H. Peter Anvin
Browse files

x86, mce: use force_sig_info to kill process in machine check 



Impact: bug fix (with tolerant == 3)

do_exit cannot be called directly from the exception handler because
it can sleep and the exception handler runs on the exception stack.
Use force_sig() instead.

Based on a earlier patch by Ying Huang who debugged the problem.

Signed-off-by: default avatarAndi Kleen <ak@linux.intel.com>
Acked-by: default avatarThomas Gleixner <tglx@linutronix.de>
Signed-off-by: default avatarH. Peter Anvin <hpa@zytor.com>
parent 6ec68bff

arch/x86/kernel/cpu/mcheck/mce_64.c

+2 −2
Original line number Original line Diff line number Diff line
@@ -295,11 +295,11 @@ void do_machine_check(struct pt_regs * regs, long error_code) 
		 * If we know that the error was in user space, send a

		 * If we know that the error was in user space, send a

		 * SIGBUS.  Otherwise, panic if tolerance is low.

		 * SIGBUS.  Otherwise, panic if tolerance is low.

		 *

		 *

		 * do_exit() takes an awful lot of locks and has a slight

		 * force_sig() takes an awful lot of locks and has a slight

		 * risk of deadlocking.

		 * risk of deadlocking.

		 */

		 */

		if (user_space) {

		if (user_space) {

			do_exit(SIGBUS);

			force_sig(SIGBUS, current);

		} else if (panic_on_oops || tolerant < 2) {

		} else if (panic_on_oops || tolerant < 2) {

			mce_panic("Uncorrected machine check",

			mce_panic("Uncorrected machine check",

				&panicm, mcestart);

				&panicm, mcestart);