On Tue, Dec 18, 2018 at 09:39:16AM +0800, gchen chen wrote:
> Yes, I think so.
> and i think the point is that unuse_mm() can't directly set tsk->mm=NULL.
So why can't unuse_mm call mm_update_next_owner?
--
MST
This seems like an issue all the unuse_mm users (at least those
outside of swapfile.c) have, so it should be solved in the core.
Bonus points for moving the set_fs magic into use_mm()..
>> Under normal circumstances,When do_exit exits, mm->owner will
>> be updated on exit_mm(). but when the kernel process calls
>> unuse_mm() and then exits,mm->owner cannot be updated. And it
>> will point to a task that has been released.
>>
>> Below is my issue on vhost_net:
>> A,
On 2018/12/13 下午12:47, gchen.guo...@gmail.com wrote:
From: guomin chen
Under normal circumstances,When do_exit exits, mm->owner will
be updated on exit_mm(). but when the kernel process calls
unuse_mm() and then exits,mm->owner cannot be updated. And it
will point to a task that has b
From: guomin chen
Under normal circumstances,When do_exit exits, mm->owner will
be updated on exit_mm(). but when the kernel process calls
unuse_mm() and then exits,mm->owner cannot be updated. And it
will point to a task that has been released.
Below is my issue on vhost_net:
A, B are