dri-devel.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Daniel Vetter <daniel@ffwll.ch>
To: "Noralf Trønnes" <noralf@tronnes.org>
Cc: intel-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Subject: Re: [Intel-gfx] [RFC v4 19/25] drm/client: Finish the in-kernel client API
Date: Mon, 16 Apr 2018 10:27:15 +0200	[thread overview]
Message-ID: <20180416082715.GK31310@phenom.ffwll.local> (raw)
In-Reply-To: <20180414115318.14500-20-noralf@tronnes.org>

On Sat, Apr 14, 2018 at 01:53:12PM +0200, Noralf Trønnes wrote:
> The modesetting code is already present, this adds the rest of the API.

Mentioning the TODO in the commit message would be good. Helps readers
like me who have an attention span measured in seconds :-)

Just commenting on the create_buffer leak here

> +static struct drm_client_buffer *
> +drm_client_buffer_create(struct drm_client_dev *client, u32 width, u32 height, u32 format)
> +{
> +	struct drm_mode_create_dumb dumb_args = { };
> +	struct drm_prime_handle prime_args = { };
> +	struct drm_client_buffer *buffer;
> +	struct dma_buf *dma_buf;
> +	void *vaddr;
> +	int ret;
> +
> +	buffer = kzalloc(sizeof(*buffer), GFP_KERNEL);
> +	if (!buffer)
> +		return ERR_PTR(-ENOMEM);
> +
> +	buffer->client = client;
> +	buffer->width = width;
> +	buffer->height = height;
> +	buffer->format = format;
> +
> +	dumb_args.width = buffer->width;
> +	dumb_args.height = buffer->height;
> +	dumb_args.bpp = drm_format_plane_cpp(format, 0) * 8;
> +	ret = drm_mode_create_dumb(client->dev, &dumb_args, client->file);
> +	if (ret)
> +		goto err_free;
> +
> +	buffer->handle = dumb_args.handle;
> +	buffer->pitch = dumb_args.pitch;
> +	buffer->size = dumb_args.size;
> +
> +	prime_args.handle = dumb_args.handle;
> +	ret = drm_prime_handle_to_fd(client->dev, &prime_args, client->file);
> +	if (ret)
> +		goto err_delete;
> +
> +	dma_buf = dma_buf_get(prime_args.fd);
> +	if (IS_ERR(dma_buf)) {
> +		ret = PTR_ERR(dma_buf);
> +		goto err_delete;
> +	}
> +
> +	/*
> +	 * If called from a worker the dmabuf fd isn't closed and the ref
> +	 * doesn't drop to zero on free.
> +	 * If I use __close_fd() it's all fine, but that function is not exported.
> +	 *
> +	 * How do I get rid of this fd when in a worker/kernel thread?
> +	 * The fd isn't used beyond this function.
> +	 */
> +//	WARN_ON(__close_fd(current->files, prime_args.fd));

Hm, this isn't 100% what I had in mind as the sequence for generic buffer
creation. Pseudo-code:

	ret = drm_mode_create_dumb(client->dev, &dumb_args, client->file);
	if (ret)
		goto err_free;
	
	gem_bo = drm_gem_object_lookup(client->file, dumb_args.handle);

gives you _really_ directly the underlying gem_bo. Of course this doesn't
work for non-gem based driver, but reality is that (almost) all of them
are. And we will not accept any new drivers which aren't gem based. So
ignoring vmwgfx for this drm_client work is imo perfectly fine. We should
ofc keep the option in the fb helpers to use non-gem buffers (so that
vmwgfx could switch over from their own in-driver fbdev helpers). All we
need for that is to keep the fb_probe callback.

Was there any other reason than vmwgfx for using prime buffers instead of
just directly using gem?

Cheers, Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

  reply	other threads:[~2018-04-16  8:27 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-04-14 11:52 [RFC v4 00/25] drm: Add generic fbdev emulation Noralf Trønnes
2018-04-14 11:52 ` [RFC v4 01/25] drm: provide management functions for drm_file Noralf Trønnes
2018-04-14 11:52 ` [RFC v4 02/25] drm/file: Don't set master on in-kernel clients Noralf Trønnes
2018-04-14 11:52 ` [RFC v4 03/25] drm/fb-helper: No need to cache rotation and sw_rotations Noralf Trønnes
2018-04-14 11:52 ` [RFC v4 04/25] drm/fb-helper: Remove drm_fb_helper_debug_enter/leave() Noralf Trønnes
2018-04-14 11:52 ` [RFC v4 05/25] drm/fb-helper: dpms_legacy(): Only set on connectors in use Noralf Trønnes
2018-04-14 11:52 ` [RFC v4 06/25] drm/atomic: Move __drm_atomic_helper_disable_plane/set_config() Noralf Trønnes
2018-04-16  8:30   ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 07/25] drm: Begin an API for in-kernel clients Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 08/25] drm/fb-helper: Use struct drm_client_display Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 09/25] drm/fb-helper: Move modeset commit code to drm_client Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 10/25] drm/connector: Add drm_connector_has_preferred_mode/pick_cmdline_mode() Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 11/25] drm/connector: Add connector array functions Noralf Trønnes
2018-04-16  8:33   ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 12/25] drm/i915: Add drm_driver->initial_client_display callback Noralf Trønnes
2018-04-16  8:38   ` Daniel Vetter
2019-03-01 11:46     ` [Intel-gfx] " Noralf Trønnes
2019-03-01 11:51       ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 13/25] drm/fb-helper: Remove struct drm_fb_helper_crtc Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 14/25] drm/fb-helper: Remove struct drm_fb_helper_connector Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 15/25] drm/fb-helper: Move modeset config code to drm_client Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 16/25] drm: Make ioctls available for in-kernel clients Noralf Trønnes
2018-04-16  9:04   ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 17/25] drm/client: Bail out if there's a DRM master Noralf Trønnes
2018-04-16  8:45   ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 18/25] drm/client: Make the display modes available to clients Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 19/25] drm/client: Finish the in-kernel client API Noralf Trønnes
2018-04-16  8:27   ` Daniel Vetter [this message]
2018-04-16 15:58     ` Noralf Trønnes
2018-04-17  8:08       ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 20/25] drm/prime: Don't pin module on export for in-kernel clients Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 21/25] drm/fb-helper: Add drm_fb_helper_fb_open/release() Noralf Trønnes
2018-04-16  8:46   ` Daniel Vetter
2018-04-16 16:10     ` Noralf Trønnes
2018-04-17  8:14       ` [Intel-gfx] " Daniel Vetter
2018-04-14 11:53 ` [RFC v4 22/25] drm/fb-helper: Add generic fbdev emulation Noralf Trønnes
2018-04-16  8:52   ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 23/25] drm: Add DRM device registered notifier Noralf Trønnes
2018-04-17 10:16   ` Daniel Vetter
2018-04-14 11:53 ` [RFC v4 24/25] drm/client: Hack: Add bootsplash Noralf Trønnes
2018-04-14 11:53 ` [RFC v4 25/25] drm/client: Hack: Add DRM VT console client Noralf Trønnes
2018-04-16  8:21 ` [RFC v4 00/25] drm: Add generic fbdev emulation Daniel Vetter
2018-04-16 18:49   ` Noralf Trønnes
2018-04-17  8:10     ` [Intel-gfx] " Daniel Vetter

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=20180416082715.GK31310@phenom.ffwll.local \
    --to=daniel@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=noralf@tronnes.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 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).