From: "Christian König" <christian.koenig@amd.com>
To: Dave Airlie <airlied@gmail.com>
Cc: Roland Scheidegger <sroland@vmware.com>,
Ben Skeggs <bskeggs@redhat.com>,
dri-devel <dri-devel@lists.freedesktop.org>,
Gerd Hoffmann <kraxel@redhat.com>
Subject: Re: [PATCH 39/49] drm/ttm: make ttm_bo_man_init/takedown take type + args
Date: Tue, 4 Aug 2020 12:26:27 +0200 [thread overview]
Message-ID: <b39fd32c-0c22-70c6-5abd-22eea744c769@amd.com> (raw)
In-Reply-To: <CAPM=9tydAjSwOWf_CRUYqM5HzuhWPH8HUYOmn0P1ufGFCB=Gvg@mail.gmail.com>
Am 04.08.20 um 03:42 schrieb Dave Airlie:
> On Fri, 31 Jul 2020 at 23:32, Christian König <christian.koenig@amd.com> wrote:
>> Am 31.07.20 um 06:05 schrieb Dave Airlie:
>>> From: Dave Airlie <airlied@redhat.com>
>>>
>>> This makes it easier to move these to a driver allocated system
>> No, sorry that looks like going into the wrong direction to me.
>>
>> I already wanted to suggest to get rid of the size argument instead.
> I'm not sure how it should look then, I don't want the driver poking
> around inside the range manager code, the range manager should be a
> generic object that driver inits and leaves alone,
>
> I definitely don't want the driver to be poking caching and size
> values into the man objects for it, since it isn't a driver object.
>
> Do you have some other view on how the generic range manager should work?
That was not what I was talking about. Take a look at what those fields
are used for :)
As far as I see the only usage of the size is in
ttm_resource_manager_debug(). But this size is actually totally opaque
to TTM, it could be pages, bytes, fried chicken wings or whatever. In
other words it would be much better to print it in the debug callback of
each resource manager.
The available_caching is completely superfluous as well. The original
idea what that a driver could specify multiple placements as flags in
one entry, but no driver ever used that as far as I know and it became
completely deprecated in 2014 when I moved the lpfn and fpfn into each
place.
The default_caching is used in ttm_bo_select_caching(), but this is
complete utterly nonsense. It just results in multiple possible cache
behaviors being selected, which are then fortunately ignored by
ttm_tt_set_placement_caching :)
To be honest I think just removing those parameters and the associated
caching flags all together is the next logical step.
Regards,
Christian.
>
> Dave.
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2020-08-04 10:26 UTC|newest]
Thread overview: 99+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-31 4:04 [PATCH 00/49] ttm mem manager refactoring Dave Airlie
2020-07-31 4:04 ` [PATCH 01/49] drm/qxl/ttm: call ttm manager debug Dave Airlie
2020-07-31 4:04 ` [PATCH 02/49] drm/vram-helper: call the ttm manager debug function Dave Airlie
2020-07-31 6:43 ` Thomas Zimmermann
2020-07-31 4:04 ` [PATCH 03/49] drm/ttm: split the mm manager init code Dave Airlie
2020-07-31 5:44 ` Sam Ravnborg
2020-07-31 5:51 ` Dave Airlie
2020-07-31 6:26 ` Dave Airlie
2020-07-31 6:43 ` Sam Ravnborg
2020-07-31 9:12 ` Christian König
2020-07-31 4:04 ` [PATCH 04/49] drm/ttm: provide a driver-led init path for generic mm manager Dave Airlie
2020-07-31 6:57 ` Thomas Zimmermann
2020-07-31 7:03 ` Thomas Zimmermann
2020-07-31 7:20 ` Dave Airlie
2020-07-31 13:00 ` Christian König
2020-07-31 4:04 ` [PATCH 05/49] drm/amdgpu/ttm: init managers from the driver side Dave Airlie
2020-07-31 4:04 ` [PATCH 06/49] drm/radeon: use new ttm man init path Dave Airlie
2020-07-31 4:04 ` [PATCH 07/49] drm/qxl/ttm: use new init path for manager Dave Airlie
2020-07-31 4:04 ` [PATCH 08/49] drm/vram_helper: use new ttm manager init function Dave Airlie
2020-07-31 4:04 ` [PATCH 09/49] drm/nouveau: use new memory manager init paths Dave Airlie
2020-07-31 4:04 ` [PATCH 10/49] drm/vmwgfx/ttm: convert vram mm init to new code paths Dave Airlie
2020-07-31 4:04 ` [PATCH 11/49] drm/vmwgfx/ttm: switch gmrid allocator to new init paths Dave Airlie
2020-07-31 4:04 ` [PATCH 12/49] drm/ttm: convert system manager init to new code Dave Airlie
2020-07-31 12:57 ` Christian König
2020-07-31 4:04 ` [PATCH 13/49] drm/ttm: purge old manager init path Dave Airlie
2020-07-31 12:58 ` Christian König
2020-07-31 4:04 ` [PATCH 14/49] drm/ttm: pass man around instead of mem_type in some places Dave Airlie
2020-07-31 13:04 ` Christian König
2020-07-31 4:04 ` [PATCH 15/49] drm/ttm: make some inline helper functions for cleanup paths Dave Airlie
2020-07-31 13:07 ` Christian König
2020-08-04 2:28 ` Dave Airlie
2020-07-31 4:04 ` [PATCH 16/49] drm/ttm: start allowing drivers to use new takedown path Dave Airlie
2020-07-31 13:09 ` Christian König
2020-07-31 4:04 ` [PATCH 17/49] drm/amdgpu/ttm: " Dave Airlie
2020-07-31 13:08 ` Christian König
2020-07-31 4:04 ` [PATCH 18/49] drm/vmwgfx: takedown vram manager Dave Airlie
2020-07-31 4:04 ` [PATCH 19/49] drm/vram_helper: call explicit mm takedown Dave Airlie
2020-07-31 4:04 ` [PATCH 20/49] drm/nouveau: use new cleanup paths Dave Airlie
2020-07-31 4:04 ` [PATCH 21/49] drm/radeon/ttm: use new takedown paths Dave Airlie
2020-07-31 13:10 ` Christian König
2020-07-31 4:04 ` [PATCH 22/49] drm/qxl/ttm: use new takedown path Dave Airlie
2020-07-31 4:04 ` [PATCH 23/49] drm/vmwgfx: fix gmrid takedown paths to new interface Dave Airlie
2020-07-31 4:04 ` [PATCH 24/49] drm/ttm: remove range manager legacy takedown path Dave Airlie
2020-07-31 13:11 ` Christian König
2020-07-31 4:04 ` [PATCH 25/49] drm/ttm: make TTM responsible for cleaning system only Dave Airlie
2020-07-31 13:19 ` Christian König
2020-07-31 4:04 ` [PATCH 26/49] drm/ttm: add wrapper to get manager from bdev Dave Airlie
2020-07-31 7:14 ` Thomas Zimmermann
2020-07-31 7:21 ` Dave Airlie
2020-07-31 13:18 ` daniel
2020-07-31 13:23 ` Christian König
2020-07-31 21:06 ` Dave Airlie
2020-07-31 4:04 ` [PATCH 27/49] drm/amdgfx/ttm: use wrapper to get ttm memory managers Dave Airlie
2020-07-31 4:04 ` [PATCH 28/49] drm/vram-helper: use wrapper to access " Dave Airlie
2020-07-31 4:05 ` [PATCH 29/49] drm/nouveau/ttm: " Dave Airlie
2020-07-31 4:05 ` [PATCH 30/49] drm/qxl/ttm: use wrapper to access memory manager Dave Airlie
2020-07-31 4:05 ` [PATCH 31/49] drm/radeon/ttm: " Dave Airlie
2020-07-31 4:05 ` [PATCH 32/49] drm/vmwgfx/ttm: " Dave Airlie
2020-07-31 4:05 ` [PATCH 33/49] drm/ttm: rename manager variable to make sure wrapper is used Dave Airlie
2020-07-31 4:05 ` [PATCH 34/49] drm/ttm: make manager debug function optional Dave Airlie
2020-07-31 9:46 ` daniel
2020-07-31 13:03 ` Christian König
2020-07-31 13:24 ` Christian König
2020-07-31 4:05 ` [PATCH 35/49] drm/nouveau/ttm: don't fill in blank ttm debug callback Dave Airlie
2020-07-31 13:25 ` Christian König
2020-07-31 4:05 ` [PATCH 36/49] drm/vmwgfx/gmrid: don't provide pointless " Dave Airlie
2020-07-31 13:26 ` Christian König
2020-07-31 4:05 ` [PATCH 37/49] drm/ttm: allow drivers to provide their own manager subclasses Dave Airlie
2020-07-31 4:05 ` [PATCH 38/49] drm/amdgpu/ttm: use bo manager subclassing for vram/gtt mgrs Dave Airlie
2020-07-31 13:29 ` Christian König
2020-07-31 4:05 ` [PATCH 39/49] drm/ttm: make ttm_bo_man_init/takedown take type + args Dave Airlie
2020-07-31 13:32 ` Christian König
2020-08-04 1:42 ` Dave Airlie
2020-08-04 10:26 ` Christian König [this message]
2020-08-06 3:46 ` Dave Airlie
2020-08-06 8:24 ` Christian König
2020-07-31 4:05 ` [PATCH 40/49] drm/ttm: move range manager to subclassed driver allocation Dave Airlie
2020-07-31 13:33 ` Christian König
2020-07-31 4:05 ` [PATCH 41/49] drm/vmwgfx/ttm: move thp to driver managed Dave Airlie
2020-07-31 4:05 ` [PATCH 42/49] drm/vmwgfx/gmrid: convert to driver controlled allocation Dave Airlie
2020-07-31 4:05 ` [PATCH 43/49] drm/nouveau/ttm: move to driver allocated manager Dave Airlie
2020-07-31 4:05 ` [PATCH 44/49] drm/ttm: drop priv pointer in memory manager Dave Airlie
2020-07-31 13:33 ` Christian König
2020-07-31 4:05 ` [PATCH 45/49] drm/amdgpu/ttm: remove man->bdev references Dave Airlie
2020-07-31 14:55 ` Christian König
2020-07-31 4:05 ` [PATCH 46/49] drm/ttm: drop man->bdev link Dave Airlie
2020-07-31 14:54 ` Christian König
2020-07-31 4:05 ` [PATCH 47/49] drm/ttm: drop list of memory managers from device Dave Airlie
2020-07-31 9:55 ` daniel
2020-07-31 14:57 ` Christian König
2020-07-31 4:05 ` [PATCH 48/49] drm/ttm: drop type manager has_type Dave Airlie
2020-07-31 4:05 ` [PATCH 49/49] drm/ttm: consolidate manager used apis into a set and get Dave Airlie
2020-07-31 9:51 ` daniel
2020-07-31 14:59 ` Christian König
2020-07-31 9:17 ` [PATCH 00/49] ttm mem manager refactoring Christian König
2020-07-31 9:29 ` daniel
2020-07-31 13:01 ` Christian König
2020-08-03 7:12 ` Dave Airlie
2020-08-03 11:13 ` Christian König
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=b39fd32c-0c22-70c6-5abd-22eea744c769@amd.com \
--to=christian.koenig@amd.com \
--cc=airlied@gmail.com \
--cc=bskeggs@redhat.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=kraxel@redhat.com \
--cc=sroland@vmware.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).