All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Christian König" <ckoenig.leichtzumerken@gmail.com>
To: Shashank Sharma <shashank.sharma@amd.com>,
	Pierre-Eric Pelloux-Prayer <pierre-eric@damsy.net>,
	amd-gfx@lists.freedesktop.org
Cc: Alex Deucher <alexander.deucher@amd.com>,
	Felix Kuehling <felix.kuehling@amd.com>,
	Christian Koenig <christian.koenig@amd.com>
Subject: Re: [PATCH v3 1/9] drm/amdgpu: UAPI for user queue management
Date: Thu, 30 Mar 2023 12:40:34 +0200	[thread overview]
Message-ID: <3c76552d-e6cf-cc89-1e57-e5ee789e79a6@gmail.com> (raw)
In-Reply-To: <48c60002-7f75-b0d0-fb7d-44b87d9cdedd@amd.com>

Am 30.03.23 um 10:15 schrieb Shashank Sharma:
> Hello Pierre-Eric,
> [SNIP]
>> I'm also wondering why the doorbell needs a handle+offset but
>> other buffers are passed in as virtual addresses?
>>
> As you know, doorbell offset here will be an relative offset in this 
> doorbell page, but the MQD needs the absolute offset on the doorbell 
> PCI BAR.
>
> So kernel needs both the object as well as relative offset to 
> calculate absolute offset.
>
> something like: absolute offset = base offset of this doorbell page + 
> relative offset of this doorbell.

Another much more obvious reason is that the doorbell doesn't have a 
virtual address.

At least for GFX the doorbell is used to signal to the hw that new 
commands are available. So as long as we don't want a shader to kick of 
other work we don't need to map the doorbell into the GPUVM address space.

Christian.

  reply	other threads:[~2023-03-30 10:40 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-29 16:04 [PATCH v3 0/9] AMDGPU Usermode queues Shashank Sharma
2023-03-29 16:04 ` [PATCH v3 1/9] drm/amdgpu: UAPI for user queue management Shashank Sharma
2023-03-29 17:25   ` Christian König
2023-03-29 17:57   ` Alex Deucher
2023-03-29 19:21     ` Shashank Sharma
2023-03-29 19:46       ` Alex Deucher
2023-03-30  6:13         ` Shashank Sharma
     [not found]   ` <71fc098c-c0cb-3097-4e11-c2d9bd9b4783@damsy.net>
2023-03-30  8:15     ` Shashank Sharma
2023-03-30 10:40       ` Christian König [this message]
2023-03-30 15:08         ` Alex Deucher
2023-03-29 16:04 ` [PATCH v3 2/9] drm/amdgpu: add usermode queue base code Shashank Sharma
2023-03-30 21:15   ` Alex Deucher
2023-03-31  8:52     ` Shashank Sharma
2023-04-04 16:05   ` Luben Tuikov
2023-03-29 16:04 ` [PATCH v3 3/9] drm/amdgpu: add new IOCTL for usermode queue Shashank Sharma
2023-04-10  0:02   ` Bas Nieuwenhuizen
2023-04-10 14:28     ` Shashank Sharma
2023-03-29 16:04 ` [PATCH v3 4/9] drm/amdgpu: create GFX-gen11 MQD for userqueue Shashank Sharma
2023-03-30 21:18   ` Alex Deucher
2023-03-31  8:49     ` Shashank Sharma
2023-04-04 16:21   ` Luben Tuikov
2023-03-29 16:04 ` [PATCH v3 5/9] drm/amdgpu: create context space for usermode queue Shashank Sharma
2023-03-30 21:23   ` Alex Deucher
2023-03-31  8:42     ` Shashank Sharma
2023-04-04 16:24   ` Luben Tuikov
2023-04-04 16:37     ` Shashank Sharma
2023-03-29 16:04 ` [PATCH v3 6/9] drm/amdgpu: add new parameters in v11_struct Shashank Sharma
2023-03-30 21:25   ` Alex Deucher
2023-03-31  6:39     ` Yadav, Arvind
2023-03-31  8:30     ` Shashank Sharma
2023-03-29 16:04 ` [PATCH v3 7/9] drm/amdgpu: map usermode queue into MES Shashank Sharma
2023-04-04 16:30   ` Luben Tuikov
2023-04-04 16:36     ` Shashank Sharma
2023-04-04 20:58       ` Luben Tuikov
2023-04-05 10:06         ` Shashank Sharma
2023-04-05 21:18           ` Luben Tuikov
2023-04-06  7:45             ` Shashank Sharma
2023-04-06 15:16               ` Felix Kuehling
2023-04-07  6:41                 ` Shashank Sharma
2023-03-29 16:04 ` [PATCH v3 8/9] drm/amdgpu: map wptr BO into GART Shashank Sharma
2023-04-10  0:00   ` Bas Nieuwenhuizen
2023-04-11  9:29     ` Christian König
2023-04-11 16:02       ` Shashank Sharma
2023-03-29 16:04 ` [PATCH v3 9/9] drm/amdgpu: generate doorbell index for userqueue Shashank Sharma
2023-04-10  0:36 ` [PATCH v3 0/9] AMDGPU Usermode queues Bas Nieuwenhuizen
2023-04-10  7:32   ` Sharma, Shashank
2023-04-10  9:25     ` Bas Nieuwenhuizen
2023-04-10 13:40       ` Sharma, Shashank
2023-04-10 13:46         ` Bas Nieuwenhuizen
2023-04-10 14:01           ` Shashank Sharma
2023-04-10 14:04             ` Bas Nieuwenhuizen
2023-04-10 14:26               ` Shashank Sharma
2023-04-11  9:37                 ` Christian König
2023-04-11  9:48                   ` Shashank Sharma
2023-04-11 10:00                     ` Bas Nieuwenhuizen
2023-04-11 10:55                       ` Shashank Sharma

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=3c76552d-e6cf-cc89-1e57-e5ee789e79a6@gmail.com \
    --to=ckoenig.leichtzumerken@gmail.com \
    --cc=alexander.deucher@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=christian.koenig@amd.com \
    --cc=felix.kuehling@amd.com \
    --cc=pierre-eric@damsy.net \
    --cc=shashank.sharma@amd.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 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.