smp: Fix a potential usage of stale nr_cpus
PATCH v3 1/1 smp: Fix a potential usage of stale nr_cpus
When the cmdline of "nr_cpus" is not valid, the @nr_cpu_ids is assigned a stale value. The nr_cpus is only valid when get_option() return 1. So check the return value to prevent this.
---
changelog in v3:
1) Return -EINVAL when the parameter is bogus.
changelog in v2:
1) Rework the commit log.
2) Rework the return value check.
kernel/smp.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
kernel/smp.c +5 -1
@@ -772,9 +772,13 @@ static int __init nrcpus(char *str) {int nr_cpus; -get_option (&str, &nr_cpus); +if (get_option (&str, &nr_cpus) !=1 ) +return -EINVAL ; +if (nr_cpus >0 && nr_cpus < nr_cpu_ids) nr_cpu_ids = nr_cpus; +else +return -EINVAL ;return 0 ; }
Signature · 1 line
2.11.0
Quoted 22 lines · Muchun Song
When the cmdline of "
nr_cpus" is not valid, the @nr_cpu_ids is assigned a stale value. Thenr_cpusis only valid whenget_option()return 1. So check the return value to prevent this.Signed-off-by Muchun Song <[email protected]>---
changelog in v3:
1) Return-EINVALwhen the parameter is bogus.changelog in v2:
1) Rework the commit log.
2) Rework the return value check.kernel/smp.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)kernel/smp.c +0
-0@@ -772,9 +772,13 @@ static int __init nrcpus(char *str) {
kernel/smp.c static int __init nrcpus(char *str) · line 781
Quoted 10 lines
kernel/smp.c +5
-1int nr_cpus; -get_option (&str, &nr_cpus); +if (get_option (&str, &nr_cpus) !=1 ) +return -EINVAL ; +if (nr_cpus >0 && nr_cpus < nr_cpu_ids) nr_cpu_ids = nr_cpus; +else +return -EINVAL ;
Exactly what does 'not valid' mean, and why doesn't get_option() return -EINVAL in that case?
Thanks,
Ingo
Quoted 11 lines · Ingo Molnar
kernel/smp.c static int __init nrcpus(char *str) · line 781
Quoted 8 lines
kernel/smp.c +5
-1-get_option (&str, &nr_cpus); +if (get_option (&str, &nr_cpus) !=1 ) +return -EINVAL ; +if (nr_cpus >0 && nr_cpus < nr_cpu_ids) nr_cpu_ids = nr_cpus; +else +return -EINVAL ;Exactly what does 'not valid' mean, and why doesn't
get_option()return-EINVALin that case?
What's unclear about invalid? If you specify nr_cpus=-1 or nr_cpus=2000000 the its obviously invalid.
How should get_option() know that this is invalid? get_option() is a number parser and does not know about any restrictions on the parsed value obviously.
get_option() returns string parsing information:
0 -> not integer found
1 -> integer found, no trailing comma or hyphen
2 -> integer found and trailing comma
3 -> integer found and traling hyphen (range parsing)And that's what is checked in if (get_option() != 1), i.e. anything else than a plain integer is invalid for this command line option.
Thanks,
tglx
Ingo Molnar
Reviewed-byQuoted 13 lines · Thomas Gleixner
Quoted 11 lines · Ingo Molnar
kernel/smp.c static int __init nrcpus(char *str) · line 781
Quoted 8 lines
kernel/smp.c +5
-1-get_option (&str, &nr_cpus); +if (get_option (&str, &nr_cpus) !=1 ) +return -EINVAL ; +if (nr_cpus >0 && nr_cpus < nr_cpu_ids) nr_cpu_ids = nr_cpus; +else +return -EINVAL ;Exactly what does 'not valid' mean, and why doesn't
get_option()return-EINVALin that case?
Quoted 2 lines
What's unclear about invalid? If you specify nr_cpus=-1 or nr_cpus=2000000 the its obviously invalid.
So this was the old (buggy) code:
Quoted 6 lines
{
int nr_cpus;
get_option(&str, &nr_cpus);
if (nr_cpus > 0 && nr_cpus < nr_cpu_ids)
nr_cpu_ids = nr_cpus;And this was the explanation given in the changelog:
Quoted 3 lines
Quoted 3 lines
When the cmdline of "
nr_cpus" is not valid, the @nr_cpu_ids is assigned a stale value. Thenr_cpusis only valid whenget_option()return 1. So check the return value to prevent this.
The answer to my question is that the bug is that the return value of get_option() wasn't checked properly, and if get_option() returns an error then the nr_cpus local variable is not set - but we used it in the old code, which can result in essentially a random value for nr_cpu_ids.
Quoted 3 lines
How should
get_option()know that this is invalid?get_option()is a number parser and does not know about any restrictions on the parsed value obviously.
But that's apparently not the bug here, 'invalid' here was meant as per the parser's syntax. If nr_cpus is out of range (like the 2000000 example you gave), then nr_cpu_ids might not be set at all, and remains at the 0 initialized value. Which isn't good but not 'stale' either.
This is why I was puzzled where a 'stale' value might come from, at first sight I was assuming that some large value was written, like your 200000 example. The "stale value" happens if it's invalid syntax and get_option() returns an error, in which case 'nr_cpus' remains uninitialized.
And this is the explanation I didn't find at first reading, and which explanation future changelogs should perhaps include.
The new code does this:
int nr_cpus;
if (get_option(&str, &nr_cpus) != 1)
return -EINVAL;
if (nr_cpus > 0 && nr_cpus < nr_cpu_ids)
nr_cpu_ids = nr_cpus;
else
return -EINVAL;Which does all the proper error handling and fixes the uninitialized 'nr_cpus' local variable usage. So I agree with the fix:
Thanks,
Ingo