tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
1What: /sys/devices/system/machinecheck/machinecheckX/tolerant
2Contact: Borislav Petkov <bp@suse.de>
3Date: Dec, 2021
4Description:
5 Unused and obsolete after the advent of recoverable machine
6 checks (see last sentence below) and those are present since
7 2010 (Nehalem).
8
9 Original description:
10
11 The entries appear for each CPU, but they are truly shared
12 between all CPUs.
13
14 Tolerance level. When a machine check exception occurs for a
15 non corrected machine check the kernel can take different
16 actions.
17
18 Since machine check exceptions can happen any time it is
19 sometimes risky for the kernel to kill a process because it
20 defies normal kernel locking rules. The tolerance level
21 configures how hard the kernel tries to recover even at some
22 risk of deadlock. Higher tolerant values trade potentially
23 better uptime with the risk of a crash or even corruption
24 (for tolerant >= 3).
25
26 == ===========================================================
27 0 always panic on uncorrected errors, log corrected errors
28 1 panic or SIGBUS on uncorrected errors, log corrected errors
29 2 SIGBUS or log uncorrected errors, log corrected errors
30 3 never panic or SIGBUS, log all errors (for testing only)
31 == ===========================================================
32
33 Default: 1
34
35 Note this only makes a difference if the CPU allows recovery
36 from a machine check exception. Current x86 CPUs generally
37 do not.