Re: [PATCH 2/5] kernel.h: Add non_block_start/end()

From: Daniel Vetter <daniel.vetter@ffwll.ch>
To: Jason Gunthorpe <jgg@ziepe.ca>
Cc: "Michal Hocko" <mhocko@kernel.org>,
	"Andrew Morton" <akpm@linux-foundation.org>,
	LKML <linux-kernel@vger.kernel.org>,
	"Linux MM" <linux-mm@kvack.org>,
	"DRI Development" <dri-devel@lists.freedesktop.org>,
	"Intel Graphics Development" <intel-gfx@lists.freedesktop.org>,
	"Peter Zijlstra" <peterz@infradead.org>,
	"Ingo Molnar" <mingo@redhat.com>,
	"David Rientjes" <rientjes@google.com>,
	"Christian König" <christian.koenig@amd.com>,
	"Jérôme Glisse" <jglisse@redhat.com>,
	"Masahiro Yamada" <yamada.masahiro@socionext.com>,
	"Wei Wang" <wvw@google.com>,
	"Andy Shevchenko" <andriy.shevchenko@linux.intel.com>,
	"Thomas Gleixner" <tglx@linutronix.de>,
	"Jann Horn" <jannh@google.com>, "Feng Tang" <feng.tang@intel.com>,
	"Kees Cook" <keescook@chromium.org>,
	"Randy Dunlap" <rdunlap@infradead.org>,
	"Daniel Vetter" <daniel.vetter@intel.com>
Subject: Re: [PATCH 2/5] kernel.h: Add non_block_start/end()
Date: Thu, 15 Aug 2019 16:43:38 +0200	[thread overview]
Message-ID: <CAKMK7uEJQ6mPQaOWbT_6M+55T-dCVbsOxFnMC6KzLAMQNa-RGg@mail.gmail.com> (raw)
In-Reply-To: <20190815143759.GG21596@ziepe.ca>

On Thu, Aug 15, 2019 at 4:38 PM Jason Gunthorpe <jgg@ziepe.ca> wrote:
>
> On Thu, Aug 15, 2019 at 03:12:11PM +0200, Daniel Vetter wrote:
> > On Thu, Aug 15, 2019 at 3:04 PM Jason Gunthorpe <jgg@ziepe.ca> wrote:
> > >
> > > On Thu, Aug 15, 2019 at 10:44:29AM +0200, Michal Hocko wrote:
> > >
> > > > As the oom reaper is the primary guarantee of the oom handling forward
> > > > progress it cannot be blocked on anything that might depend on blockable
> > > > memory allocations. These are not really easy to track because they
> > > > might be indirect - e.g. notifier blocks on a lock which other context
> > > > holds while allocating memory or waiting for a flusher that needs memory
> > > > to perform its work.
> > >
> > > But lockdep *does* track all this and fs_reclaim_acquire() was created
> > > to solve exactly this problem.
> > >
> > > fs_reclaim is a lock and it flows through all the usual lockdep
> > > schemes like any other lock. Any time the page allocator wants to do
> > > something the would deadlock with reclaim it takes the lock.
> > >
> > > Failure is expressed by a deadlock cycle in the lockdep map, and
> > > lockdep can handle arbitary complexity through layers of locks, work
> > > queues, threads, etc.
> > >
> > > What is missing?
> >
> > Lockdep doens't seen everything by far. E.g. a wait_event will be
> > caught by the annotations here, but not by lockdep.
>
> Sure, but the wait_event might be OK if its progress isn't contingent
> on fs_reclaim, ie triggered from interrupt, so why ban it?

For normal notifiers sure (but lockdep won't help you proof that at
all). For oom_reaper apparently not, but that's really Michal's thing,
I have not much idea about that.

> > And since we're talking about mmu notifiers here and gpus/dma engines.
> > We have dma_fence_wait, which can wait for any hw/driver in the system
> > that takes part in shared/zero-copy buffer processing. Which at least
> > on the graphics side is everything. This pulls in enormous amounts of
> > deadlock potential that lockdep simply is blind about and will never
> > see.
>
> It seems very risky to entagle a notifier widely like that.

That's why I've looked into all possible ways to annotate them with
debug infrastructure.

> It looks pretty sure that notifiers are fs_reclaim, so at a minimum
> that wait_event can't be contingent on anything that is doing
> GFP_KERNEL or it will deadlock - and blockable doesn't make that sleep
> safe.
>
> Avoiding an uncertain wait_event under notifiers would seem to be the
> only reasonable design here..

You have to wait for the gpu to finnish current processing in
invalidate_range_start. Otherwise there's no point to any of this
really. So the wait_event/dma_fence_wait are unavoidable really.

That's also why I'm throwing in the lockdep annotation on top, and why
it would be really nice if we somehow can get the cross-release work
landed. But it looks like all the people who could make it happen are
busy with smeltdown :-/
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch