Nouveau - Sep 2019 - Is Nouveau really using the io_reserve_lru?

Hi guys,

while working through more old TTM functionality I stumbled over the 
io_reserve_lru.

Basic idea is that when this flag is set the driver->io_mem_reserve() 
callback can return -EAGAIN resulting in unmapping of other BOs.

But Nouveau doesn't seem to return -EAGAIN in the call path of 
io_mem_reserve anywhere.

So is this unused or am I missing something?

Regards,
Christian.

On Tue, 24 Sep 2019 at 22:19, Christian König
<ckoenig.leichtzumerken at gmail.com> wrote:
>
> Hi guys,
>
> while working through more old TTM functionality I stumbled over the
> io_reserve_lru.
>
> Basic idea is that when this flag is set the driver->io_mem_reserve()
> callback can return -EAGAIN resulting in unmapping of other BOs.
>
> But Nouveau doesn't seem to return -EAGAIN in the call path of
> io_mem_reserve anywhere.
I believe this is a bug in Nouveau.  We *should* be returning -EAGAIN
if we fail to find space in BAR1 to map the BO into.

Ben.
>
> So is this unused or am I missing something?
>
> Regards,
> Christian.
> _______________________________________________
> Nouveau mailing list
> Nouveau at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/nouveau

Am 26.09.19 um 23:44 schrieb Ben Skeggs:
> On Tue, 24 Sep 2019 at 22:19, Christian König
> <ckoenig.leichtzumerken at gmail.com> wrote:
>> Hi guys,
>>
>> while working through more old TTM functionality I stumbled over the
>> io_reserve_lru.
>>
>> Basic idea is that when this flag is set the
driver->io_mem_reserve()
>> callback can return -EAGAIN resulting in unmapping of other BOs.
>>
>> But Nouveau doesn't seem to return -EAGAIN in the call path of
>> io_mem_reserve anywhere.
> I believe this is a bug in Nouveau.  We *should* be returning -EAGAIN
> if we fail to find space in BAR1 to map the BO into.
Yeah, thought that this is rather odd. But I can't fix it since I don't 
really know the code.

On the other hand do you have any objections that I move the remaining 
io_reserve_lru functionality into Nouveau then?

It's only used by this driver and since Thomas cleaned up the VM 
handling that shouldn't be a problem any more.

Thanks,
Christian.
>
> Ben.
>> So is this unused or am I missing something?
>>
>> Regards,
>> Christian.
>> _______________________________________________
>> Nouveau mailing list
>> Nouveau at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/nouveau

On Thu, Sep 26, 2019 at 5:44 PM Ben Skeggs <skeggsb at gmail.com>
wrote:
>
> On Tue, 24 Sep 2019 at 22:19, Christian König
> <ckoenig.leichtzumerken at gmail.com> wrote:
> >
> > Hi guys,
> >
> > while working through more old TTM functionality I stumbled over the
> > io_reserve_lru.
> >
> > Basic idea is that when this flag is set the
driver->io_mem_reserve()
> > callback can return -EAGAIN resulting in unmapping of other BOs.
> >
> > But Nouveau doesn't seem to return -EAGAIN in the call path of
> > io_mem_reserve anywhere.
> I believe this is a bug in Nouveau.  We *should* be returning -EAGAIN
> if we fail to find space in BAR1 to map the BO into.
Could this lead to SIGBUS in userspace, esp related to resume and
similar situations? A user has been experiencing this in a
tricky-to-reproduce scenario with a ton of vram dedicated to
framebuffers and so on (3x 4K), and the nouveau ddx falls back to
memcpy in certain cases.

  -ilia