All of lore.kernel.org
 help / color / mirror / Atom feed
From: John Hubbard <jhubbard@nvidia.com>
To: Jerome Glisse <jglisse@redhat.com>
Cc: <linux-mm@kvack.org>, Andrew Morton <akpm@linux-foundation.org>,
	<linux-kernel@vger.kernel.org>,
	Ralph Campbell <rcampbell@nvidia.com>, <stable@vger.kernel.org>,
	Evgeny Baskakov <ebaskakov@nvidia.com>,
	Mark Hairgrove <mhairgrove@nvidia.com>
Subject: Re: [PATCH 03/15] mm/hmm: HMM should have a callback before MM is destroyed v2
Date: Wed, 21 Mar 2018 16:10:32 -0700	[thread overview]
Message-ID: <3f4e78a1-5a88-399d-e134-497229c42707@nvidia.com> (raw)
In-Reply-To: <20180321224620.GH3214@redhat.com>

On 03/21/2018 03:46 PM, Jerome Glisse wrote:
> On Wed, Mar 21, 2018 at 03:16:04PM -0700, John Hubbard wrote:
>> On 03/21/2018 11:03 AM, Jerome Glisse wrote:
>>> On Tue, Mar 20, 2018 at 09:14:34PM -0700, John Hubbard wrote:
>>>> On 03/19/2018 07:00 PM, jglisse@redhat.com wrote:
>>>>> From: Ralph Campbell <rcampbell@nvidia.com>
>>
>> <snip>
>>
>>>> Hi Jerome,
>>>>
>>>> This presents a deadlock problem (details below). As for solution ideas, 
>>>> Mark Hairgrove points out that the MMU notifiers had to solve the
>>>> same sort of problem, and part of the solution involves "avoid
>>>> holding locks when issuing these callbacks". That's not an entire 
>>>> solution description, of course, but it seems like a good start.
>>>>
>>>> Anyway, for the deadlock problem:
>>>>
>>>> Each of these ->release callbacks potentially has to wait for the 
>>>> hmm_invalidate_range() callbacks to finish. That is not shown in any
>>>> code directly, but it's because: when a device driver is processing 
>>>> the above ->release callback, it has to allow any in-progress operations 
>>>> to finish up (as specified clearly in your comment documentation above). 
>>>>
>>>> Some of those operations will invariably need to do things that result 
>>>> in page invalidations, thus triggering the hmm_invalidate_range() callback.
>>>> Then, the hmm_invalidate_range() callback tries to acquire the same 
>>>> hmm->mirrors_sem lock, thus leading to deadlock:
>>>>
>>>> hmm_invalidate_range():
>>>> // ...
>>>> 	down_read(&hmm->mirrors_sem);
>>>> 	list_for_each_entry(mirror, &hmm->mirrors, list)
>>>> 		mirror->ops->sync_cpu_device_pagetables(mirror, action,
>>>> 							start, end);
>>>> 	up_read(&hmm->mirrors_sem);
>>>
>>> That is just illegal, the release callback is not allowed to trigger
>>> invalidation all it does is kill all device's threads and stop device
>>> page fault from happening. So there is no deadlock issues. I can re-
>>> inforce the comment some more (see [1] for example on what it should
>>> be).
>>
>> That rule is fine, and it is true that the .release callback will not 
>> directly trigger any invalidations. However, the problem is in letting 
>> any *existing* outstanding operations finish up. We have to let 
>> existing operations "drain", in order to meet the requirement that 
>> everything is done when .release returns.
>>
>> For example, if a device driver thread is in the middle of working through
>> its fault buffer, it will call migrate_vma(), which will in turn unmap
>> pages. That will cause an hmm_invalidate_range() callback, which tries
>> to take hmm->mirrors_sems, and we deadlock.
>>
>> There's no way to "kill" such a thread while it's in the middle of
>> migrate_vma(), you have to let it finish up.
>>
>>> Also it is illegal for the sync callback to trigger any mmu_notifier
>>> callback. I thought this was obvious. The sync callback should only
>>> update device page table and do _nothing else_. No way to make this
>>> re-entrant.
>>
>> That is obvious, yes. I am not trying to say there is any problem with
>> that rule. It's the "drain outstanding operations during .release", 
>> above, that is the real problem.
> 
> Maybe just relax the release callback wording, it should stop any
> more processing of fault buffer but not wait for it to finish. In
> nouveau code i kill thing but i do not wait hence i don't deadlock.

But you may crash, because that approach allows .release to finish
up, thus removing the mm entirely, out from under (for example)
a migrate_vma call--or any other call that refers to the mm.

It doesn't seem too hard to avoid the problem, though: maybe we
can just drop the lock while doing the mirror->ops->release callback.
There are a few ways to do this, but one example is: 

    -- take the lock,
        -- copy the list to a local list, deleting entries as you go,
    -- drop the lock, 
    -- iterate through the local list copy and 
        -- issue the mirror->ops->release callbacks.

At this point, more items could have been added to the list, so repeat
the above until the original list is empty. 

This is subject to a limited starvation case if mirror keep getting 
registered, but I think we can ignore that, because it only lasts as long as 
mirrors keep getting added, and then it finishes up.

> 
> What matter is to stop any further processing. Yes some fault might
> be in flight but they will serialize on various lock. 

Those faults in flight could already be at a point where they have taken
whatever locks they need, so we don't dare let the mm get destroyed while
such fault handling is in progress.


So just do not
> wait in the release callback, kill thing. I might have a bug where i
> still fill in GPU page table in nouveau, i will check nouveau code
> for that.

Again, we can't "kill" a thread of execution (this would often be an
interrupt bottom half context, btw) while it is, for example,
in the middle of migrate_vma.

I really don't believe there is a safe way to do this without draining
the existing operations before .release returns, and for that, we'll need to 
issue the .release callbacks while not holding locks.

thanks,
-- 
John Hubbard
NVIDIA

> 
> Kill thing should also kill the channel (i don't do that in nouveau
> because i am waiting on some channel patchset) but i am not sure if
> hardware like it if we kill channel before stoping fault notification.
> 
> Cheers,
> Jérôme
> 

WARNING: multiple messages have this Message-ID (diff)
From: John Hubbard <jhubbard@nvidia.com>
To: Jerome Glisse <jglisse@redhat.com>
Cc: linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
	linux-kernel@vger.kernel.org,
	Ralph Campbell <rcampbell@nvidia.com>,
	stable@vger.kernel.org, Evgeny Baskakov <ebaskakov@nvidia.com>,
	Mark Hairgrove <mhairgrove@nvidia.com>
Subject: Re: [PATCH 03/15] mm/hmm: HMM should have a callback before MM is destroyed v2
Date: Wed, 21 Mar 2018 16:10:32 -0700	[thread overview]
Message-ID: <3f4e78a1-5a88-399d-e134-497229c42707@nvidia.com> (raw)
In-Reply-To: <20180321224620.GH3214@redhat.com>

On 03/21/2018 03:46 PM, Jerome Glisse wrote:
> On Wed, Mar 21, 2018 at 03:16:04PM -0700, John Hubbard wrote:
>> On 03/21/2018 11:03 AM, Jerome Glisse wrote:
>>> On Tue, Mar 20, 2018 at 09:14:34PM -0700, John Hubbard wrote:
>>>> On 03/19/2018 07:00 PM, jglisse@redhat.com wrote:
>>>>> From: Ralph Campbell <rcampbell@nvidia.com>
>>
>> <snip>
>>
>>>> Hi Jerome,
>>>>
>>>> This presents a deadlock problem (details below). As for solution ideas, 
>>>> Mark Hairgrove points out that the MMU notifiers had to solve the
>>>> same sort of problem, and part of the solution involves "avoid
>>>> holding locks when issuing these callbacks". That's not an entire 
>>>> solution description, of course, but it seems like a good start.
>>>>
>>>> Anyway, for the deadlock problem:
>>>>
>>>> Each of these ->release callbacks potentially has to wait for the 
>>>> hmm_invalidate_range() callbacks to finish. That is not shown in any
>>>> code directly, but it's because: when a device driver is processing 
>>>> the above ->release callback, it has to allow any in-progress operations 
>>>> to finish up (as specified clearly in your comment documentation above). 
>>>>
>>>> Some of those operations will invariably need to do things that result 
>>>> in page invalidations, thus triggering the hmm_invalidate_range() callback.
>>>> Then, the hmm_invalidate_range() callback tries to acquire the same 
>>>> hmm->mirrors_sem lock, thus leading to deadlock:
>>>>
>>>> hmm_invalidate_range():
>>>> // ...
>>>> 	down_read(&hmm->mirrors_sem);
>>>> 	list_for_each_entry(mirror, &hmm->mirrors, list)
>>>> 		mirror->ops->sync_cpu_device_pagetables(mirror, action,
>>>> 							start, end);
>>>> 	up_read(&hmm->mirrors_sem);
>>>
>>> That is just illegal, the release callback is not allowed to trigger
>>> invalidation all it does is kill all device's threads and stop device
>>> page fault from happening. So there is no deadlock issues. I can re-
>>> inforce the comment some more (see [1] for example on what it should
>>> be).
>>
>> That rule is fine, and it is true that the .release callback will not 
>> directly trigger any invalidations. However, the problem is in letting 
>> any *existing* outstanding operations finish up. We have to let 
>> existing operations "drain", in order to meet the requirement that 
>> everything is done when .release returns.
>>
>> For example, if a device driver thread is in the middle of working through
>> its fault buffer, it will call migrate_vma(), which will in turn unmap
>> pages. That will cause an hmm_invalidate_range() callback, which tries
>> to take hmm->mirrors_sems, and we deadlock.
>>
>> There's no way to "kill" such a thread while it's in the middle of
>> migrate_vma(), you have to let it finish up.
>>
>>> Also it is illegal for the sync callback to trigger any mmu_notifier
>>> callback. I thought this was obvious. The sync callback should only
>>> update device page table and do _nothing else_. No way to make this
>>> re-entrant.
>>
>> That is obvious, yes. I am not trying to say there is any problem with
>> that rule. It's the "drain outstanding operations during .release", 
>> above, that is the real problem.
> 
> Maybe just relax the release callback wording, it should stop any
> more processing of fault buffer but not wait for it to finish. In
> nouveau code i kill thing but i do not wait hence i don't deadlock.

But you may crash, because that approach allows .release to finish
up, thus removing the mm entirely, out from under (for example)
a migrate_vma call--or any other call that refers to the mm.

It doesn't seem too hard to avoid the problem, though: maybe we
can just drop the lock while doing the mirror->ops->release callback.
There are a few ways to do this, but one example is: 

    -- take the lock,
        -- copy the list to a local list, deleting entries as you go,
    -- drop the lock, 
    -- iterate through the local list copy and 
        -- issue the mirror->ops->release callbacks.

At this point, more items could have been added to the list, so repeat
the above until the original list is empty. 

This is subject to a limited starvation case if mirror keep getting 
registered, but I think we can ignore that, because it only lasts as long as 
mirrors keep getting added, and then it finishes up.

> 
> What matter is to stop any further processing. Yes some fault might
> be in flight but they will serialize on various lock. 

Those faults in flight could already be at a point where they have taken
whatever locks they need, so we don't dare let the mm get destroyed while
such fault handling is in progress.


So just do not
> wait in the release callback, kill thing. I might have a bug where i
> still fill in GPU page table in nouveau, i will check nouveau code
> for that.

Again, we can't "kill" a thread of execution (this would often be an
interrupt bottom half context, btw) while it is, for example,
in the middle of migrate_vma.

I really don't believe there is a safe way to do this without draining
the existing operations before .release returns, and for that, we'll need to 
issue the .release callbacks while not holding locks.

thanks,
-- 
John Hubbard
NVIDIA

> 
> Kill thing should also kill the channel (i don't do that in nouveau
> because i am waiting on some channel patchset) but i am not sure if
> hardware like it if we kill channel before stoping fault notification.
> 
> Cheers,
> Jérôme
> 

  reply	other threads:[~2018-03-21 23:10 UTC|newest]

Thread overview: 97+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-20  2:00 [PATCH 00/15] hmm: fixes and documentations v3 jglisse
2018-03-20  2:00 ` jglisse
2018-03-20  2:00 ` [PATCH 01/15] mm/hmm: documentation editorial update to HMM documentation jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 02/15] mm/hmm: fix header file if/else/endif maze v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 03/15] mm/hmm: HMM should have a callback before MM is destroyed v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-21  4:14   ` John Hubbard
2018-03-21  4:14     ` John Hubbard
2018-03-21 18:03     ` Jerome Glisse
2018-03-21 18:03       ` Jerome Glisse
2018-03-21 18:03       ` Jerome Glisse
2018-03-21 22:16       ` John Hubbard
2018-03-21 22:16         ` John Hubbard
2018-03-21 22:46         ` Jerome Glisse
2018-03-21 22:46           ` Jerome Glisse
2018-03-21 22:46           ` Jerome Glisse
2018-03-21 23:10           ` John Hubbard [this message]
2018-03-21 23:10             ` John Hubbard
2018-03-21 23:37             ` Jerome Glisse
2018-03-21 23:37               ` Jerome Glisse
2018-03-21 23:37               ` Jerome Glisse
2018-03-22  0:11               ` John Hubbard
2018-03-22  0:11                 ` John Hubbard
2018-03-22  1:32                 ` Jerome Glisse
2018-03-22  1:32                   ` Jerome Glisse
2018-03-22  1:32                   ` Jerome Glisse
2018-03-22  1:28   ` [PATCH 03/15] mm/hmm: HMM should have a callback before MM is destroyed v3 jglisse
2018-03-22  1:28     ` jglisse
2018-03-22  6:58     ` John Hubbard
2018-03-22  6:58       ` John Hubbard
2018-03-20  2:00 ` [PATCH 04/15] mm/hmm: unregister mmu_notifier when last HMM client quit jglisse
2018-03-20  2:00   ` jglisse
2018-03-21  4:24   ` John Hubbard
2018-03-21  4:24     ` John Hubbard
2018-03-21 18:12     ` Jerome Glisse
2018-03-21 18:12       ` Jerome Glisse
2018-03-21 18:16   ` [PATCH 04/15] mm/hmm: unregister mmu_notifier when last HMM client quit v2 jglisse
2018-03-21 18:16     ` jglisse
2018-03-21 23:22     ` John Hubbard
2018-03-21 23:22       ` John Hubbard
2018-03-21 23:41       ` Jerome Glisse
2018-03-21 23:41         ` Jerome Glisse
2018-03-22 22:47         ` John Hubbard
2018-03-22 22:47           ` John Hubbard
2018-03-22 23:37           ` Jerome Glisse
2018-03-22 23:37             ` Jerome Glisse
2018-03-23  0:13             ` John Hubbard
2018-03-23  0:13               ` John Hubbard
2018-03-23  0:50               ` Jerome Glisse
2018-03-23  0:50                 ` Jerome Glisse
2018-03-23  0:56                 ` John Hubbard
2018-03-23  0:56                   ` John Hubbard
2018-03-22  1:30     ` [PATCH 04/15] mm/hmm: unregister mmu_notifier when last HMM client quit v3 jglisse
2018-03-22  1:30       ` jglisse
2018-03-22 22:36       ` Andrew Morton
2018-03-20  2:00 ` [PATCH 05/15] mm/hmm: hmm_pfns_bad() was accessing wrong struct jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 06/15] mm/hmm: use struct for hmm_vma_fault(), hmm_vma_get_pfns() parameters v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 07/15] mm/hmm: remove HMM_PFN_READ flag and ignore peculiar architecture v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 08/15] mm/hmm: use uint64_t for HMM pfn instead of defining hmm_pfn_t to ulong v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 09/15] mm/hmm: cleanup special vma handling (VM_SPECIAL) jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 10/15] mm/hmm: do not differentiate between empty entry or missing directory v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-21  5:24   ` John Hubbard
2018-03-21  5:24     ` John Hubbard
2018-03-21 14:48     ` Jerome Glisse
2018-03-21 14:48       ` Jerome Glisse
2018-03-21 23:16       ` John Hubbard
2018-03-21 23:16         ` John Hubbard
2018-03-20  2:00 ` [PATCH 11/15] mm/hmm: rename HMM_PFN_DEVICE_UNADDRESSABLE to HMM_PFN_DEVICE_PRIVATE jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 12/15] mm/hmm: move hmm_pfns_clear() closer to where it is use jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 13/15] mm/hmm: factor out pte and pmd handling to simplify hmm_vma_walk_pmd() jglisse
2018-03-20  2:00   ` jglisse
2018-03-21  5:07   ` John Hubbard
2018-03-21  5:07     ` John Hubbard
2018-03-21 15:08     ` Jerome Glisse
2018-03-21 15:08       ` Jerome Glisse
2018-03-21 22:36       ` John Hubbard
2018-03-21 22:36         ` John Hubbard
2018-03-20  2:00 ` [PATCH 14/15] mm/hmm: change hmm_vma_fault() to allow write fault on page basis jglisse
2018-03-20  2:00   ` jglisse
2018-03-20  2:00 ` [PATCH 15/15] mm/hmm: use device driver encoding for HMM pfn v2 jglisse
2018-03-20  2:00   ` jglisse
2018-03-21  4:39   ` John Hubbard
2018-03-21  4:39     ` John Hubbard
2018-03-21 15:52     ` Jerome Glisse
2018-03-21 15:52       ` Jerome Glisse
2018-03-21 23:19       ` John Hubbard
2018-03-21 23:19         ` John Hubbard

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=3f4e78a1-5a88-399d-e134-497229c42707@nvidia.com \
    --to=jhubbard@nvidia.com \
    --cc=akpm@linux-foundation.org \
    --cc=ebaskakov@nvidia.com \
    --cc=jglisse@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhairgrove@nvidia.com \
    --cc=rcampbell@nvidia.com \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.