Am 19.01.2018 um 11:40 schrieb Michal Hocko:
On Fri 19-01-18 09:39:03, Christian König wrote:
Am 19.01.2018 um 09:20 schrieb Michal Hocko:
[...]
OK, in that case I would propose a different approach. We already
have rss_stat. So why do not we simply add a new counter there
MM_KERNELPAGES and consider those in oom_badness? The rule would be
that such a memory is bound to the process life time. I guess we will
find more users for this later.
I already tried that and the problem with that approach is that some buffers
are not created by the application which actually uses them.
For example X/Wayland is creating and handing out render buffers to
application which want to use OpenGL.
So the result is when you always account the application who created the
buffer the OOM killer will certainly reap X/Wayland first. And that is
exactly what we want to avoid here.
Then you have to find the target allocation context at the time of the
allocation and account it.
And exactly that's the root of the problem: The target allocation
context isn't known at the time of the allocation.
We could add callbacks so that when the memory is passed from the
allocator to the actual user of the memory. In other words when the
memory is passed from the X server to the client the driver would need
to decrement the X servers accounting and increment the clients accounting.
But I think that would go deep into the file descriptor handling (we
would at least need to handle dup/dup2 and passing the fd using unix
domain sockets) and most likely would be rather error prone.
The per file descriptor badness is/was just the much easier approach to
solve the issue, because the drivers already knew which client is
currently using which buffer objects.
I of course agree that file descriptors can be shared between processes
and are by themselves not killable. But at least for our graphics driven
use case I don't see much of a problem killing all processes when a file
descriptor is used by more than one at the same time.
Regards,
Christian.
As follow up emails show, implementations
might differ and any robust oom solution have to rely on the common
counters.
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel
