All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Christian König" <ckoenig.leichtzumerken@gmail.com>
To: Qiang Yu <yuq825@gmail.com>,
	dri-devel@lists.freedesktop.org, devicetree@vger.kernel.org
Cc: Simon Shields <simon@lineageos.org>, Marek Vasut <marex@denx.de>,
	Connor Abbott <cwabbott0@gmail.com>,
	Neil Armstrong <narmstrong@baylibre.com>,
	Andrei Paulau <7134956@gmail.com>,
	Vasily Khoruzhick <anarsoul@gmail.com>,
	Erico Nunes <nunes.erico@gmail.com>
Subject: Re: [PATCH RFC 05/24] Revert "drm: Nerf the preclose callback for modern drivers"
Date: Wed, 23 May 2018 11:35:04 +0200	[thread overview]
Message-ID: <9b48f845-9854-1029-ed21-df7b3590935b@gmail.com> (raw)
In-Reply-To: <20180518092815.25280-6-yuq825@gmail.com>

Well NAK, that brings back a callback we worked quite hard on getting 
rid of.

It looks like the problem isn't that you need the preclose callback, but 
you rather seem to misunderstood how TTM works.

All you need to do is to cleanup your command submission path so that 
the caller of lima_sched_context_queue_task() adds the resulting 
scheduler fence to TTMs buffer objects.

Regards,
Christian.

Am 18.05.2018 um 11:27 schrieb Qiang Yu:
> This reverts commit 45c3d213a400c952ab7119f394c5293bb6877e6b.
>
> lima driver need preclose to wait all task in the context
> created within closing file to finish before free all the
> buffer object. Otherwise pending tesk may fail and get
> noisy MMU fault message.
>
> Move this wait to each buffer object free function can
> achieve the same result but some buffer object is shared
> with other file context, but we only want to wait the
> closing file context's tasks. So the implementation is
> not that straight forword compared to the preclose one.
>
> Signed-off-by: Qiang Yu <yuq825@gmail.com>
> ---
>   drivers/gpu/drm/drm_file.c |  8 ++++----
>   include/drm/drm_drv.h      | 23 +++++++++++++++++++++--
>   2 files changed, 25 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_file.c b/drivers/gpu/drm/drm_file.c
> index e394799979a6..0a43107396b9 100644
> --- a/drivers/gpu/drm/drm_file.c
> +++ b/drivers/gpu/drm/drm_file.c
> @@ -361,8 +361,9 @@ void drm_lastclose(struct drm_device * dev)
>    *
>    * This function must be used by drivers as their &file_operations.release
>    * method. It frees any resources associated with the open file, and calls the
> - * &drm_driver.postclose driver callback. If this is the last open file for the
> - * DRM device also proceeds to call the &drm_driver.lastclose driver callback.
> + * &drm_driver.preclose and &drm_driver.lastclose driver callbacks. If this is
> + * the last open file for the DRM device also proceeds to call the
> + * &drm_driver.lastclose driver callback.
>    *
>    * RETURNS:
>    *
> @@ -382,8 +383,7 @@ int drm_release(struct inode *inode, struct file *filp)
>   	list_del(&file_priv->lhead);
>   	mutex_unlock(&dev->filelist_mutex);
>   
> -	if (drm_core_check_feature(dev, DRIVER_LEGACY) &&
> -	    dev->driver->preclose)
> +	if (dev->driver->preclose)
>   		dev->driver->preclose(dev, file_priv);
>   
>   	/* ========================================================
> diff --git a/include/drm/drm_drv.h b/include/drm/drm_drv.h
> index d23dcdd1bd95..8d6080f97ed4 100644
> --- a/include/drm/drm_drv.h
> +++ b/include/drm/drm_drv.h
> @@ -107,6 +107,23 @@ struct drm_driver {
>   	 */
>   	int (*open) (struct drm_device *, struct drm_file *);
>   
> +	/**
> +	 * @preclose:
> +	 *
> +	 * One of the driver callbacks when a new &struct drm_file is closed.
> +	 * Useful for tearing down driver-private data structures allocated in
> +	 * @open like buffer allocators, execution contexts or similar things.
> +	 *
> +	 * Since the display/modeset side of DRM can only be owned by exactly
> +	 * one &struct drm_file (see &drm_file.is_master and &drm_device.master)
> +	 * there should never be a need to tear down any modeset related
> +	 * resources in this callback. Doing so would be a driver design bug.
> +	 *
> +	 * FIXME: It is not really clear why there's both @preclose and
> +	 * @postclose. Without a really good reason, use @postclose only.
> +	 */
> +	void (*preclose) (struct drm_device *, struct drm_file *file_priv);
> +
>   	/**
>   	 * @postclose:
>   	 *
> @@ -118,6 +135,9 @@ struct drm_driver {
>   	 * one &struct drm_file (see &drm_file.is_master and &drm_device.master)
>   	 * there should never be a need to tear down any modeset related
>   	 * resources in this callback. Doing so would be a driver design bug.
> +	 *
> +	 * FIXME: It is not really clear why there's both @preclose and
> +	 * @postclose. Without a really good reason, use @postclose only.
>   	 */
>   	void (*postclose) (struct drm_device *, struct drm_file *);
>   
> @@ -134,7 +154,7 @@ struct drm_driver {
>   	 * state changes, e.g. in conjunction with the :ref:`vga_switcheroo`
>   	 * infrastructure.
>   	 *
> -	 * This is called after @postclose hook has been called.
> +	 * This is called after @preclose and @postclose have been called.
>   	 *
>   	 * NOTE:
>   	 *
> @@ -601,7 +621,6 @@ struct drm_driver {
>   	/* List of devices hanging off this driver with stealth attach. */
>   	struct list_head legacy_dev_list;
>   	int (*firstopen) (struct drm_device *);
> -	void (*preclose) (struct drm_device *, struct drm_file *file_priv);
>   	int (*dma_ioctl) (struct drm_device *dev, void *data, struct drm_file *file_priv);
>   	int (*dma_quiescent) (struct drm_device *);
>   	int (*context_dtor) (struct drm_device *dev, int context);

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

  reply	other threads:[~2018-05-23  9:35 UTC|newest]

Thread overview: 89+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-18  9:27 [PATCH RFC 00/24] Lima DRM driver Qiang Yu
2018-05-18  9:27 ` [PATCH RFC 01/24] ARM: dts: add gpu node to exynos4 Qiang Yu
2018-05-23 17:06   ` Rob Herring
2018-05-18  9:27 ` [PATCH RFC 02/24] dt-bindings: add switch-delay property for mali-utgard Qiang Yu
2018-05-23 17:04   ` Rob Herring
2018-05-24  1:52     ` Qiang Yu
2018-05-18  9:27 ` [PATCH RFC 03/24] arm64/dts: add switch-delay for meson mali Qiang Yu
2018-05-21 14:16   ` Neil Armstrong
2018-05-21 14:16     ` Neil Armstrong
2018-05-22  0:48     ` Qiang Yu
2018-05-22  0:48       ` Qiang Yu
2018-05-18  9:27 ` [PATCH RFC 04/24] " Qiang Yu
2018-05-21 14:16   ` Neil Armstrong
2018-05-21 14:16     ` Neil Armstrong
2018-05-18  9:27 ` [PATCH RFC 05/24] Revert "drm: Nerf the preclose callback for modern drivers" Qiang Yu
2018-05-23  9:35   ` Christian König [this message]
2018-05-23 13:13     ` Qiang Yu
2018-05-23 13:41       ` Christian König
2018-05-24  1:38         ` Qiang Yu
2018-05-24  6:46           ` Christian König
2018-05-24  9:24             ` Qiang Yu
2018-05-24  9:41               ` Christian König
2018-05-24 12:54                 ` Qiang Yu
2018-05-18  9:27 ` [PATCH RFC 06/24] drm/lima: add lima uapi header Qiang Yu
2018-05-18  9:33   ` Marek Vasut
2018-05-20  7:22     ` Qiang Yu
2018-05-20  9:52       ` Marek Vasut
2018-05-20  7:25     ` Qiang Yu
2018-05-18  9:27 ` [PATCH RFC 07/24] drm/lima: add mali 4xx GPU hardware regs Qiang Yu
2018-05-23 17:24   ` Rob Herring
2018-05-23 17:31     ` Vasily Khoruzhick
2018-05-24  0:58     ` Qiang Yu
2018-05-24 14:31       ` Rob Herring
2018-05-18  9:27 ` [PATCH RFC 08/24] drm/lima: add lima core driver Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 09/24] drm/lima: add GPU device functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 10/24] drm/lima: add PMU related functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 11/24] drm/lima: add L2 cache functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 12/24] drm/lima: add GP related functions Qiang Yu
2018-05-23 17:12   ` Marek Vasut
2018-05-24  0:38     ` Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 13/24] drm/lima: add PP " Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 14/24] drm/lima: add MMU " Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 15/24] drm/lima: add BCAST related function Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 16/24] drm/lima: add DLBU related functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 17/24] drm/lima: add GPU virtual memory space handing Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 18/24] drm/lima: add TTM subsystem functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 19/24] drm/lima: add buffer object functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 20/24] drm/lima: add GEM related functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 21/24] drm/lima: add GEM Prime " Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 22/24] drm/lima: add GPU schedule using DRM_SCHED Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 23/24] drm/lima: add context related functions Qiang Yu
2018-05-18  9:28 ` [PATCH RFC 24/24] drm/lima: add makefile and kconfig Qiang Yu
2018-05-23 17:16   ` Marek Vasut
2018-05-23 17:26     ` Rob Herring
2018-05-24  0:49       ` Qiang Yu
2018-06-15 17:23     ` Andre Przywara
2018-07-14  1:14       ` Qiang Yu
2018-07-14 12:06         ` André Przywara
2018-07-14 14:18           ` Qiang Yu
2018-07-14 19:15             ` André Przywara
2018-07-15  2:23               ` Qiang Yu
2018-05-23  9:02 ` [PATCH RFC 00/24] Lima DRM driver Daniel Vetter
2018-05-23 13:24   ` Qiang Yu
2018-05-23  9:29 ` Christian König
2018-05-23 13:52   ` Qiang Yu
2018-05-23 13:59     ` Christian König
2018-05-23 14:13       ` Qiang Yu
2018-05-23 14:19         ` Christian König
2018-05-23 14:27           ` Qiang Yu
2018-05-23 15:44     ` Daniel Vetter
2018-05-24  0:31       ` Qiang Yu
2018-05-24  6:27         ` Christian König
2018-05-24  7:25           ` Daniel Vetter
2018-05-24  9:53             ` Christian König
2018-05-19  6:52 Qiang Yu
2018-05-19  6:52 ` [PATCH RFC 05/24] Revert "drm: Nerf the preclose callback for modern drivers" Qiang Yu
2018-05-21 19:37   ` Eric Anholt
2018-05-22  1:04     ` Qiang Yu
2018-05-23  9:04       ` Daniel Vetter
2018-05-23 12:59         ` Qiang Yu
2018-05-23 20:31           ` Daniel Vetter
2018-05-24  1:18             ` Qiang Yu
2018-05-24  7:51               ` Daniel Vetter
2018-05-24  8:55                 ` Qiang Yu
2018-05-30 18:13                   ` Eric Anholt
2018-05-31 14:04                     ` Qiang Yu
2018-05-31 17:51                       ` Eric Anholt
2018-05-31 18:04                         ` Keith Packard
2018-06-01  2:03                           ` Qiang Yu
2018-06-01  1:58                         ` Qiang Yu

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=9b48f845-9854-1029-ed21-df7b3590935b@gmail.com \
    --to=ckoenig.leichtzumerken@gmail.com \
    --cc=7134956@gmail.com \
    --cc=anarsoul@gmail.com \
    --cc=christian.koenig@amd.com \
    --cc=cwabbott0@gmail.com \
    --cc=devicetree@vger.kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=marex@denx.de \
    --cc=narmstrong@baylibre.com \
    --cc=nunes.erico@gmail.com \
    --cc=simon@lineageos.org \
    --cc=yuq825@gmail.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.