[CRIU] [PATCH 20/23] ptrace.c: don't seize the task that doesn't run in the userspace

[Pavel Emelyanov]

On 01/14/2013 02:28 PM, Alexander Kartashov wrote:
> On 01/14/2013 02:24 PM, Pavel Emelyanov wrote:
>> And is task stopped/trapped in this case?
> 
> Do you mean that the task may be ptrace()'d while being
> in this state? 

It can be seized without being in this state.

> Apparently it may, as demostrated by the test static/file_read
> that sometimes fail without this patch because of this.

This sounds like a BS. Task trapped/stopped inside the kernel
code? Can we "fix" it's registers while it's in the kernel and
make it execute some "bad" sequence of instructions leading to
some crash? I believe you must investigate the problem deeper.
Maybe you're right, but the explanation you provide doesn't
suit me.