[CRIU] [PATCH] vma: don't skip vmas during searching a parent vma

[Pavel Emelyanov]

On 05/13/2014 05:49 PM, Andrew Vagin wrote:
> On Tue, May 13, 2014 at 05:48:07PM +0400, Pavel Emelyanov wrote:
>> On 05/13/2014 05:18 PM, Andrew Vagin wrote:
>>> On Tue, May 13, 2014 at 05:09:14PM +0400, Pavel Emelyanov wrote:
>>>> On 05/13/2014 04:25 PM, Andrey Vagin wrote:
>>>>> We stop searching if vma->start is bigger than a required one.
>>>>> The coursor is set on the last examined vma. When we are searching a
>>>>> parent vma for the next vma, we start examine vma-s starting from
>>>>> coursor->next, so we don't examine the vma, which is pointed by cursor.
>>>>>
>>>>> This patch replaces list_for_each_entry_continue on list_for_each_entry_from.
>>>>>
>>>>> Reported-by: Filipe Brandenburger <filbranden at google.com>
>>>>> Signed-off-by: Andrey Vagin <avagin at openvz.org>
>>>>
>>>> Andrey, the recent 3 patches are great, but can you explain why keeping track
>>>> of mm-dumpable bit breaks cow and we start requiring these fixes?
>>>
>>> The test reports errors like this:
>>> 17:14:08.431: 21816: ERR: cow01.c:78: Unable to open file /proc/21817/pagemap (errno = 13 (Permission denied))
>>>
>>> These errors are ignored. cow01 worked fine when it was committed. Then
>>> we decided to execute tests from a non-root user and the test starts
>>> reports these error and stop to do something useful.
>>
>> OK, but as you say -- test worked fine when it was committed. This means we
>> broke COW at some point after moving tests into non superuser?
> 
> Yes, we did.

Can we find out when did this happen?

Thanks,
Pavel