4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Guenter Roeck <linux@xxxxxxxxxxxx>
If CONFIG_SECCOMP=n, /proc/self/status includes an empty line. This causes
the iotop application to bail out with an error message.
File "/usr/local/lib64/python2.7/site-packages/iotop/data.py", line 196,
in parse_proc_pid_status
key, value = line.split(':\t', 1)
ValueError: need more than 1 value to unpack
The problem is seen in v4.9.y but not upstream because commit af884cd4a5ae6
("proc: report no_new_privs state") has not been backported to v4.9.y.
The backport of commit fae1fa0fc6cc ("proc: Provide details on speculation
flaw mitigations") tried to address the resulting differences but was
wrong, introducing the problem.
Fixes: 51ef9af2a35b ("proc: Provide details on speculation flaw mitigations")
Cc: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Gwendal Grignou <gwendal@xxxxxxxxxxxx>
Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx>
Acked-by: Kees Cook <keescook@xxxxxxxxxxxx>
---
This patch only applies to v4.9.y. v4.4.y also needs to be fixed (see
https://www.spinics.net/lists/stable/msg279131.html), but the fix
is slightly different. v4.14.y and later are not affected.
fs/proc/array.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -346,8 +346,9 @@ static inline void task_seccomp(struct s
{
#ifdef CONFIG_SECCOMP
seq_put_decimal_ull(m, "Seccomp:\t", p->seccomp.mode);
+ seq_putc(m, '\n');
#endif
- seq_printf(m, "\nSpeculation_Store_Bypass:\t");
+ seq_printf(m, "Speculation_Store_Bypass:\t");
switch (arch_prctl_spec_ctrl_get(p, PR_SPEC_STORE_BYPASS)) {
case -EINVAL:
seq_printf(m, "unknown");
