dri-devel.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: "Thomas Hellström (Intel)" <thomas_os@shipmail.org>
To: "Christian König" <ckoenig.leichtzumerken@gmail.com>,
	matthew.auld@intel.com, dri-devel@lists.freedesktop.org
Subject: Re: [PATCH 01/10] drm/ttm: allocate resource object instead of embedding it v2
Date: Tue, 8 Jun 2021 07:29:41 +0200	[thread overview]
Message-ID: <a08db95b-7c75-d998-7443-73d809121e47@shipmail.org> (raw)
In-Reply-To: <6d316dfe-0378-8284-1f8e-29caf5619e34@gmail.com>

Hi,

On 6/7/21 7:59 PM, Christian König wrote:
>
>
> Am 07.06.21 um 19:58 schrieb Thomas Hellström (Intel):
>>
>> On 6/7/21 7:54 PM, Christian König wrote:
>>>
>>>
>>> Am 07.06.21 um 19:06 schrieb Thomas Hellström (Intel):
>>>>
>>>> On 6/7/21 6:40 PM, Thomas Hellström (Intel) wrote:
>>>>>
>>>>> On 6/2/21 12:09 PM, Christian König wrote:
>>>>> ...
>>>>>> @@ -728,14 +728,15 @@ static int ttm_bo_add_move_fence(struct 
>>>>>> ttm_buffer_object *bo,
>>>>>>    */
>>>>>>   static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
>>>>>>                     const struct ttm_place *place,
>>>>>> -                  struct ttm_resource *mem,
>>>>>> +                  struct ttm_resource **mem,
>>>>>>                     struct ttm_operation_ctx *ctx)
>>>>>>   {
>>>>>>       struct ttm_device *bdev = bo->bdev;
>>>>>> -    struct ttm_resource_manager *man = ttm_manager_type(bdev, 
>>>>>> mem->mem_type);
>>>>>> +    struct ttm_resource_manager *man;
>>>>>>       struct ww_acquire_ctx *ticket;
>>>>>>       int ret;
>>>>>>   +    man = ttm_manager_type(bdev, (*mem)->mem_type);
>>>>>
>>>>> Isn't (*mem) uninitialized here? Should be place->mem_type? 
>>>>> Eviction is immediately sent to the bushes.
>>>>>
>>>>> Got at least one additional NULL pointer deref to track down in 
>>>>> the eviction code, but could be a merge error of mine as well.
>>>>
>>>> Actually this last one was probably due to a bad temporary fix of 
>>>> the above one.
>>>
>>> I've found one more warning during my testing now. But that is just 
>>> a false positive.
>>>
>>> Apart from that I haven't seen any other fallout, but fingers crossed.
>>
>> vmwgfx doesn't seem to happy. It works AFAICT., but warns in 
>> vmw_move() about ttm_bo_assign_mem() replacing an existing resource.
>
> Yeah, that's the one I've just fixed. Patch is underway.

If that's the move_to_ghost patch, I don't think that would fix the 
vmwgfx issue, as IIRC vmwgfx ever uses ghost objects.

/Thomas



  reply	other threads:[~2021-06-08  5:29 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-02 10:09 [PATCH 01/10] drm/ttm: allocate resource object instead of embedding it v2 Christian König
2021-06-02 10:09 ` [PATCH 02/10] drm/ttm: flip over the range manager to self allocated nodes Christian König
2021-06-02 11:44   ` Thomas Hellström (Intel)
2021-06-02 12:11     ` Christian König
2021-06-02 12:33       ` Thomas Hellström (Intel)
2021-06-02 13:07         ` Christian König
2021-06-02 14:13           ` Thomas Hellström (Intel)
2021-06-02 14:17             ` Christian König
2021-06-02 15:28               ` Thomas Hellström (Intel)
2021-06-02 18:41                 ` Christian König
2021-06-02 18:52                   ` Thomas Hellström (Intel)
2021-06-02 18:53                     ` Christian König
2021-06-02 10:09 ` [PATCH 03/10] drm/ttm: flip over the sys " Christian König
2021-06-03  7:51   ` Matthew Auld
2021-06-02 10:09 ` [PATCH 04/10] drm/amdgpu: revert "drm/amdgpu: stop allocating dummy GTT nodes" Christian König
2021-06-02 10:09 ` [PATCH 05/10] drm/amdkfd: use resource cursor in svm_migrate_copy_to_vram v2 Christian König
2021-06-03  9:44   ` Matthew Auld
2021-06-02 10:09 ` [PATCH 06/10] drm/amdgpu: switch the GTT backend to self alloc Christian König
2021-06-02 10:09 ` [PATCH 07/10] drm/amdgpu: switch the VRAM " Christian König
2021-06-02 10:09 ` [PATCH 08/10] drm/nouveau: switch the TTM backends " Christian König
2021-06-02 10:09 ` [PATCH 09/10] drm/vmwgfx: " Christian König
2021-06-02 10:09 ` [PATCH 10/10] drm/ttm: flip the switch for driver allocated resources v2 Christian König
2021-06-07 10:15   ` Thomas Hellström (Intel)
2021-06-07 10:37     ` Christian König
2021-06-07 10:44       ` Thomas Hellström (Intel)
2021-06-03  8:45 ` [PATCH 01/10] drm/ttm: allocate resource object instead of embedding it v2 Matthew Auld
2021-06-04 11:54   ` Christian König
2021-06-04  9:33 ` Thomas Hellström (Intel)
2021-06-07 16:40 ` Thomas Hellström (Intel)
2021-06-07 17:06   ` Thomas Hellström (Intel)
2021-06-07 17:54     ` Christian König
2021-06-07 17:58       ` Thomas Hellström (Intel)
2021-06-07 17:59         ` Christian König
2021-06-08  5:29           ` Thomas Hellström (Intel) [this message]
2021-06-08  7:14             ` Christian König
2021-06-08  7:17               ` Thomas Hellström (Intel)
2021-06-08  7:21                 ` Christian König
2021-06-08  9:38                   ` Das, Nirmoy
2021-06-08  9:40                     ` Das, Nirmoy
2021-06-08  9:42                       ` Christian König
2021-06-08  9:48                         ` Das, Nirmoy
2021-06-07 17:10   ` Christian König
2021-06-08  6:55 ` Thomas Hellström

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=a08db95b-7c75-d998-7443-73d809121e47@shipmail.org \
    --to=thomas_os@shipmail.org \
    --cc=ckoenig.leichtzumerken@gmail.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=matthew.auld@intel.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).