Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1704690 > unrolled thread
| Started by | Aleksa Sarai <asarai@suse.com> |
|---|---|
| First post | 2017-08-05 19:00 +0200 |
| Last post | 2017-08-06 06:00 +0200 |
| Articles | 3 — 3 participants |
Back to article view | Back to linux.kernel
[PATCH] sched: debug: use task_pid_vnr in /proc/$pid/sched Aleksa Sarai <asarai@suse.com> - 2017-08-05 19:00 +0200
Re: [PATCH] sched: debug: use task_pid_vnr in /proc/$pid/sched Aleksa Sarai <asarai@suse.de> - 2017-08-05 20:00 +0200
Re: [PATCH] sched: debug: use task_pid_vnr in /proc/$pid/sched ebiederm@xmission.com (Eric W. Biederman) - 2017-08-06 06:00 +0200
| From | Aleksa Sarai <asarai@suse.com> |
|---|---|
| Date | 2017-08-05 19:00 +0200 |
| Subject | [PATCH] sched: debug: use task_pid_vnr in /proc/$pid/sched |
| Message-ID | <ubaAF-7wb-7@gated-at.bofh.it> |
It appears as though the addition of the PID namespace did not update
the output code for /proc/$pid/sched, which made it trivial to figure
out whether a process was inside &init_pid_ns from userspace (making
container detection trivial[1]). This lead to situations such as:
% unshare -pf head -n1 /proc/self/sched
head (10047, #threads: 1)
Fix this by just using task_pid_vnr for the output of /proc/$pid/sched.
All of the other uses of task_pid_nr in kernel/sched/debug.c are from a
sysctl context and thus don't need to be namespaced.
[1]: https://github.com/jessfraz/amicontained
Cc: <stable@vger.kernel.org>
Cc: Jess Frazelle <acidburn@google.com>
Signed-off-by: Aleksa Sarai <asarai@suse.com>
---
kernel/sched/debug.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 4fa66de52bd6..a06acbe33e16 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -876,7 +876,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
{
unsigned long nr_switches;
- SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p),
+ SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_vnr(p),
get_nr_threads(p));
SEQ_printf(m,
"---------------------------------------------------------"
--
2.13.3
[toc] | [next] | [standalone]
| From | Aleksa Sarai <asarai@suse.de> |
|---|---|
| Date | 2017-08-05 20:00 +0200 |
| Message-ID | <ubbwJ-863-1@gated-at.bofh.it> |
| In reply to | #1704690 |
On 2017-08-06 02:52, Aleksa Sarai wrote:
> It appears as though the addition of the PID namespace did not update
> the output code for /proc/$pid/sched, which made it trivial to figure
> out whether a process was inside &init_pid_ns from userspace (making
> container detection trivial[1]). This lead to situations such as:
>
> % unshare -pf head -n1 /proc/self/sched
> head (10047, #threads: 1)
>
> Fix this by just using task_pid_vnr for the output of /proc/$pid/sched.
> All of the other uses of task_pid_nr in kernel/sched/debug.c are from a
> sysctl context and thus don't need to be namespaced.
>
> [1]: https://github.com/jessfraz/amicontained
>
> Cc: <stable@vger.kernel.org>
> Cc: Jess Frazelle <acidburn@google.com>
> Signed-off-by: Aleksa Sarai <asarai@suse.com>
> ---
> kernel/sched/debug.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
> index 4fa66de52bd6..a06acbe33e16 100644
> --- a/kernel/sched/debug.c
> +++ b/kernel/sched/debug.c
> @@ -876,7 +876,7 @@ void proc_sched_show_task(struct task_struct *p,
> struct seq_file *m)
> {
> unsigned long nr_switches;
>
> - SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p),
> + SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_vnr(p),
> get_nr_threads(p));
> SEQ_printf(m,
> "---------------------------------------------------------"
Added Eric to Cc.
--
Aleksa Sarai
Software Engineer (Containers)
SUSE Linux GmbH
https://www.cyphar.com/
[toc] | [prev] | [next] | [standalone]
| From | ebiederm@xmission.com (Eric W. Biederman) |
|---|---|
| Date | 2017-08-06 06:00 +0200 |
| Message-ID | <ubkTn-5vg-3@gated-at.bofh.it> |
| In reply to | #1704693 |
Aleksa Sarai <asarai@suse.de> writes:
> On 2017-08-06 02:52, Aleksa Sarai wrote:
>> It appears as though the addition of the PID namespace did not update
>> the output code for /proc/$pid/sched, which made it trivial to figure
>> out whether a process was inside &init_pid_ns from userspace (making
>> container detection trivial[1]). This lead to situations such as:
>>
>> % unshare -pf head -n1 /proc/self/sched
>> head (10047, #threads: 1)
>>
>> Fix this by just using task_pid_vnr for the output of /proc/$pid/sched.
>> All of the other uses of task_pid_nr in kernel/sched/debug.c are from a
>> sysctl context and thus don't need to be namespaced.
>>
>> [1]: https://github.com/jessfraz/amicontained
>>
>> Cc: <stable@vger.kernel.org>
>> Cc: Jess Frazelle <acidburn@google.com>
>> Signed-off-by: Aleksa Sarai <asarai@suse.com>
>> ---
>> kernel/sched/debug.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
>> index 4fa66de52bd6..a06acbe33e16 100644
>> --- a/kernel/sched/debug.c
>> +++ b/kernel/sched/debug.c
>> @@ -876,7 +876,7 @@ void proc_sched_show_task(struct task_struct *p,
>> struct seq_file *m)
>> {
>> unsigned long nr_switches;
>>
>> - SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p),
>> + SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_vnr(p),
>> get_nr_threads(p));
>> SEQ_printf(m,
>> "---------------------------------------------------------"
>
> Added Eric to Cc.
Changing this to pid_nr_ns, to make this relative to the pid namespace
of proc would be very reasonable.
Making files content relative to current is in general a bug.
Hiding the fact you are contained is not an explicit goal. But making
those things that are reasonably in a namespace relative to that
namespace is.
Eric
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web