Re: [PATCH] gpu: drm: etnaviv: Change return type to vm_fault_t

From: Souptick Joarder
Date: Mon May 28 2018 - 11:01:36 EST


On Mon, May 28, 2018 at 6:20 PM, Lucas Stach <l.stach@xxxxxxxxxxxxxx> wrote:
> Hi Souptick,
>
> Am Montag, den 21.05.2018, 22:42 +0530 schrieb Souptick Joarder:
>> Use new return type vm_fault_t for fault handler. For
>> now, this is just documenting that the function returns
>> a VM_FAULT value rather than an errno. Once all instances
>> are converted, vm_fault_t will become a distinct type.
>>
>> Ref- commit 1c8f422059ae ("mm: change return type to vm_fault_t")
>>
>> Previously vm_insert_page() returns err which driver
>> mapped into VM_FAULT_* type. The new function
>> vmf_insert_page() will replace this inefficiency by
>> returning VM_FAULT_* type.
>>
>> A non-fatal signal being delivered to a task which is
>> in the middle of a page fault may well end up in an
>> infinite loop, attempting to handle the page fault and
>> failing forever.
>>
>> On seeing NOPAGE, the fault handler believes the PTE
>> is in the page table, so does nothing before it returns
>> to arch code. It will end up returning to userspace if
>> the signal is non-fatal, at which point it'll go right
>> back into the page fault handler, and mutex_lock_interruptible()
>> will immediately fail. So we've converted a sleeping lock
>> into the most expensive spinlock.
>>
>> I don't think the graphics drivers really want to be
>> interrupted by any signal. I think they want to be
>> interruptible by fatal signals and should use the
>> mutex_lock_killable / fatal_signal_pending family of
>> functions. So mutex_lock_interruptible() is replaced
>> by mutex_lock_killable()
>
> I don't think the other thread discussing this issue has reached a
> proper conclusion and I would rather not pull in a patch based on thin
> reasoning. If we want to change this behavior I would really like to
> see the issue it is solving and it should be a separate patch, not
> intermixed with some return value change.
>
> Regards,
> Lucas
>

Ok, I will separate the vm_fault_t type return value patch
and send v2 for it. We would like to get this patch in queue
for 4.18.

The patch consists, mutex_lock_interruptible/killable() can be keep
on hold until we reached to a proper conclusion.