From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?UTF-8?Q?Christian_K=c3=b6nig?= Subject: Re: [PATCH] ttm: wait mem space if user allow while gpu busy Date: Tue, 23 Apr 2019 16:42:13 +0200 Message-ID: References: <20190422103836.4300-1-david1.zhou@amd.com> <1c98b839-362d-c279-1abb-c022aed3abf1@gmail.com> Reply-To: christian.koenig@amd.com Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1485116869==" Return-path: Received: from mail-wm1-x343.google.com (mail-wm1-x343.google.com [IPv6:2a00:1450:4864:20::343]) by gabe.freedesktop.org (Postfix) with ESMTPS id B10C189104 for ; Tue, 23 Apr 2019 14:42:16 +0000 (UTC) Received: by mail-wm1-x343.google.com with SMTP id z6so2418510wmi.0 for ; Tue, 23 Apr 2019 07:42:16 -0700 (PDT) In-Reply-To: Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" To: "Zhou, David(ChunMing)" , "Koenig, Christian" , "Liang, Prike" , "dri-devel@lists.freedesktop.org" List-Id: dri-devel@lists.freedesktop.org This is a multi-part message in MIME format. --===============1485116869== Content-Type: multipart/alternative; boundary="------------094E65C91CD530A1623807BE" Content-Language: en-US This is a multi-part message in MIME format. --------------094E65C91CD530A1623807BE Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 8bit Well that's not so easy of hand. The basic problem here is that when you busy wait at this place you can easily run into situations where application A busy waits for B while B busy waits for A -> deadlock. So what we need here is the deadlock detection logic of the ww_mutex. To use this we at least need to do the following steps: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of its reservation object) has a ticket assigned. 3. If we have a ticket we grab a reference to the first BO on the LRU, drop the LRU lock and try to grab the reservation lock with the ticket. 4. If getting the reservation lock with the ticket succeeded we check if the BO is still the first one on the LRU in question (the BO could have moved). 5. If the BO is still the first one on the LRU in question we try to evict it as we would evict any other BO. 6. If any of the "If's" above fail we just back off and return -EBUSY. Steps 2-5 are certainly not trivial, but doable as far as I can see. Have fun :) Christian. Am 23.04.19 um 15:19 schrieb Zhou, David(ChunMing): > How about adding more condition ctx->resv inline to address your > concern? As well as don't wait from same user, shouldn't lead to deadlock. > > Otherwise, any other idea? > > -------- Original Message -------- > Subject: Re: [PATCH] ttm: wait mem space if user allow while gpu busy > From: Christian König > To: "Liang, Prike" ,"Zhou, David(ChunMing)" > ,dri-devel@lists.freedesktop.org > CC: > > Well that is certainly a NAK because it can lead to deadlock in the > memory management. > > You can't just busy wait with all those locks held. > > Regards, > Christian. > > Am 23.04.19 um 03:45 schrieb Liang, Prike: > > Acked-by: Prike Liang > > > > Thanks, > > Prike > > -----Original Message----- > > From: Chunming Zhou > > Sent: Monday, April 22, 2019 6:39 PM > > To: dri-devel@lists.freedesktop.org > > Cc: Liang, Prike ; Zhou, David(ChunMing) > > > Subject: [PATCH] ttm: wait mem space if user allow while gpu busy > > > > heavy gpu job could occupy memory long time, which could lead to > other user fail to get memory. > > > > Change-Id: I0b322d98cd76e5ac32b00462bbae8008d76c5e11 > > Signed-off-by: Chunming Zhou > > --- > >   drivers/gpu/drm/ttm/ttm_bo.c | 6 ++++-- > >   1 file changed, 4 insertions(+), 2 deletions(-) > > > > diff --git a/drivers/gpu/drm/ttm/ttm_bo.c > b/drivers/gpu/drm/ttm/ttm_bo.c index 7c484729f9b2..6c596cc24bec 100644 > > --- a/drivers/gpu/drm/ttm/ttm_bo.c > > +++ b/drivers/gpu/drm/ttm/ttm_bo.c > > @@ -830,8 +830,10 @@ static int ttm_bo_mem_force_space(struct > ttm_buffer_object *bo, > >                if (mem->mm_node) > >                        break; > >                ret = ttm_mem_evict_first(bdev, mem_type, place, ctx); > > -             if (unlikely(ret != 0)) > > -                     return ret; > > +             if (unlikely(ret != 0)) { > > +                     if (!ctx || ctx->no_wait_gpu || ret != -EBUSY) > > +                             return ret; > > +             } > >        } while (1); > >        mem->mem_type = mem_type; > >        return ttm_bo_add_move_fence(bo, man, mem); > > -- > > 2.17.1 > > > > _______________________________________________ > > dri-devel mailing list > > dri-devel@lists.freedesktop.org > > https://lists.freedesktop.org/mailman/listinfo/dri-devel > > > _______________________________________________ > dri-devel mailing list > dri-devel@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/dri-devel --------------094E65C91CD530A1623807BE Content-Type: text/html; charset=windows-1252 Content-Transfer-Encoding: 8bit
Well that's not so easy of hand.

The basic problem here is that when you busy wait at this place you can easily run into situations where application A busy waits for B while B busy waits for A -> deadlock.

So what we need here is the deadlock detection logic of the ww_mutex. To use this we at least need to do the following steps:

1. Reserve the BO in DC using a ww_mutex ticket (trivial).

2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of its reservation object) has a ticket assigned.

3. If we have a ticket we grab a reference to the first BO on the LRU, drop the LRU lock and try to grab the reservation lock with the ticket.

4. If getting the reservation lock with the ticket succeeded we check if the BO is still the first one on the LRU in question (the BO could have moved).

5. If the BO is still the first one on the LRU in question we try to evict it as we would evict any other BO.

6. If any of the "If's" above fail we just back off and return -EBUSY.

Steps 2-5 are certainly not trivial, but doable as far as I can see.

Have fun :)
Christian.

Am 23.04.19 um 15:19 schrieb Zhou, David(ChunMing):
How about adding more condition ctx->resv inline to address your concern? As well as don't wait from same user, shouldn't lead to deadlock.

Otherwise, any other idea?

-------- Original Message --------
Subject: Re: [PATCH] ttm: wait mem space if user allow while gpu busy
From: Christian König
To: "Liang, Prike" ,"Zhou, David(ChunMing)" ,dri-devel@lists.freedesktop.org
CC:

Well that is certainly a NAK because it can lead to deadlock in the
memory management.

You can't just busy wait with all those locks held.

Regards,
Christian.

Am 23.04.19 um 03:45 schrieb Liang, Prike:
> Acked-by: Prike Liang <Prike.Liang@amd.com>
>
> Thanks,
> Prike
> -----Original Message-----
> From: Chunming Zhou <david1.zhou@amd.com>
> Sent: Monday, April 22, 2019 6:39 PM
> To: dri-devel@lists.freedesktop.org
> Cc: Liang, Prike <Prike.Liang@amd.com>; Zhou, David(ChunMing) <David1.Zhou@amd.com>
> Subject: [PATCH] ttm: wait mem space if user allow while gpu busy
>
> heavy gpu job could occupy memory long time, which could lead to other user fail to get memory.
>
> Change-Id: I0b322d98cd76e5ac32b00462bbae8008d76c5e11
> Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
> ---
>   drivers/gpu/drm/ttm/ttm_bo.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 7c484729f9b2..6c596cc24bec 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
> @@ -830,8 +830,10 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
>                if (mem->mm_node)
>                        break;
>                ret = ttm_mem_evict_first(bdev, mem_type, place, ctx);
> -             if (unlikely(ret != 0))
> -                     return ret;
> +             if (unlikely(ret != 0)) {
> +                     if (!ctx || ctx->no_wait_gpu || ret != -EBUSY)
> +                             return ret;
> +             }
>        } while (1);
>        mem->mem_type = mem_type;
>        return ttm_bo_add_move_fence(bo, man, mem);
> --
> 2.17.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

 
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

--------------094E65C91CD530A1623807BE-- --===============1485116869== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: base64 Content-Disposition: inline X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KZHJpLWRldmVs IG1haWxpbmcgbGlzdApkcmktZGV2ZWxAbGlzdHMuZnJlZWRlc2t0b3Aub3JnCmh0dHBzOi8vbGlz dHMuZnJlZWRlc2t0b3Aub3JnL21haWxtYW4vbGlzdGluZm8vZHJpLWRldmVs --===============1485116869==--