From: Tomasz Figa <tomasz.figa@gmail.com>
To: Sean Paul <seanpaul@chromium.org>
Cc: dri-devel@lists.freedesktop.org, marcheu@chromium.org
Subject: Re: [PATCH v3 14/32] drm/exynos: Remove exynos_drm_hdmi shim
Date: Sun, 10 Nov 2013 22:24:05 +0100 [thread overview]
Message-ID: <1526769.j7lMfrmRht@flatron> (raw)
In-Reply-To: <1383063198-10526-15-git-send-email-seanpaul@chromium.org>
Hi Sean,
On Tuesday 29 of October 2013 12:13:00 Sean Paul wrote:
> This patch trims exynos_drm_hdmi out of the driver. The reason it
> existed in the first place was to make up for the mixture of
> display/overlay/manager ops being spread across hdmi and mixer. With
> that code now rationalized, mixer and hdmi map directly to
> exynos_drm_crtc and exynos_drm_encoder, respectively. Since there is a
> 1:1 mapping, we no longer need this layer.
>
> Signed-off-by: Sean Paul <seanpaul@chromium.org>
> ---
>
> Changes in v2:
> - hdmi/mixer ops now take display/manager instead of context
> Changes in v3:
> - Moved removal of exynos_drm_hdmi.c file here
>
> drivers/gpu/drm/exynos/Makefile | 3 +-
> drivers/gpu/drm/exynos/exynos_drm_drv.c | 13 -
> drivers/gpu/drm/exynos/exynos_drm_hdmi.c | 416 -------------------------------
> drivers/gpu/drm/exynos/exynos_drm_hdmi.h | 69 -----
> drivers/gpu/drm/exynos/exynos_hdmi.c | 162 +++++++-----
> drivers/gpu/drm/exynos/exynos_mixer.c | 204 ++++++++-------
> drivers/gpu/drm/exynos/exynos_mixer.h | 20 ++
> 7 files changed, 222 insertions(+), 665 deletions(-)
> delete mode 100644 drivers/gpu/drm/exynos/exynos_drm_hdmi.c
> delete mode 100644 drivers/gpu/drm/exynos/exynos_drm_hdmi.h
> create mode 100644 drivers/gpu/drm/exynos/exynos_mixer.h
>
> diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
> index e106309..9e12fb7 100644
> --- a/drivers/gpu/drm/exynos/exynos_hdmi.c
> +++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
[snip]
> +static struct exynos_drm_display hdmi_display = {
> + .type = EXYNOS_DISPLAY_TYPE_HDMI,
> + .ops = &hdmi_display_ops,
> +};
> +
> static irqreturn_t hdmi_irq_thread(int irq, void *arg)
> {
> - struct exynos_drm_hdmi_context *ctx = arg;
> - struct hdmi_context *hdata = ctx->ctx;
> + struct hdmi_context *hdata = arg;
>
> mutex_lock(&hdata->hdmi_mutex);
> hdata->hpd = gpio_get_value(hdata->hpd_gpio);
> @@ -1887,7 +1944,6 @@ static struct of_device_id hdmi_match_types[] = {
> static int hdmi_probe(struct platform_device *pdev)
> {
> struct device *dev = &pdev->dev;
> - struct exynos_drm_hdmi_context *drm_hdmi_ctx;
> struct hdmi_context *hdata;
> struct s5p_hdmi_platform_data *pdata;
> struct resource *res;
> @@ -1902,20 +1958,13 @@ static int hdmi_probe(struct platform_device *pdev)
> if (!pdata)
> return -EINVAL;
>
> - drm_hdmi_ctx = devm_kzalloc(dev, sizeof(*drm_hdmi_ctx), GFP_KERNEL);
> - if (!drm_hdmi_ctx)
> - return -ENOMEM;
> -
> hdata = devm_kzalloc(dev, sizeof(struct hdmi_context), GFP_KERNEL);
> if (!hdata)
> return -ENOMEM;
>
> mutex_init(&hdata->hdmi_mutex);
>
> - drm_hdmi_ctx->ctx = (void *)hdata;
> - hdata->parent_ctx = (void *)drm_hdmi_ctx;
> -
> - platform_set_drvdata(pdev, drm_hdmi_ctx);
> + platform_set_drvdata(pdev, &hdmi_display);
This looks like a step backwards. Originally the driver had kind of per
instance data, while now it has a single global struct, making it harder
to properly support multiple HDMI instances, what is not unlikely to show
up in future SoCs.
So instead, I would embed an exynos_drm_display struct inside struct
hdmi_context and initialize it at runtime and use it to get access to
hdata, using container_of.
> diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
> index 60935c4..39aed3e 100644
> --- a/drivers/gpu/drm/exynos/exynos_mixer.c
> +++ b/drivers/gpu/drm/exynos/exynos_mixer.c
[snip]
> @@ -1184,21 +1190,17 @@ static struct of_device_id mixer_match_types[] = {
> static int mixer_probe(struct platform_device *pdev)
> {
> struct device *dev = &pdev->dev;
> - struct exynos_drm_hdmi_context *drm_hdmi_ctx;
> struct mixer_context *ctx;
> struct mixer_drv_data *drv;
> int ret;
>
> dev_info(dev, "probe start\n");
>
> - drm_hdmi_ctx = devm_kzalloc(dev, sizeof(*drm_hdmi_ctx),
> - GFP_KERNEL);
> - if (!drm_hdmi_ctx)
> - return -ENOMEM;
> -
> - ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
> - if (!ctx)
> + ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_KERNEL);
> + if (!ctx) {
> + DRM_ERROR("failed to alloc mixer context.\n");
> return -ENOMEM;
> + }
>
> mutex_init(&ctx->mixer_mutex);
>
> @@ -1211,18 +1213,14 @@ static int mixer_probe(struct platform_device *pdev)
> platform_get_device_id(pdev)->driver_data;
> }
>
> - ctx->dev = dev;
> - ctx->parent_ctx = (void *)drm_hdmi_ctx;
> - drm_hdmi_ctx->ctx = (void *)ctx;
> + ctx->dev = &pdev->dev;
> ctx->vp_enabled = drv->is_vp_enabled;
> ctx->mxr_ver = drv->version;
> DRM_INIT_WAITQUEUE(&ctx->wait_vsync_queue);
> atomic_set(&ctx->wait_vsync_event, 0);
>
> - platform_set_drvdata(pdev, drm_hdmi_ctx);
> -
> /* acquire resources: regs, irqs, clocks */
> - ret = mixer_resources_init(drm_hdmi_ctx, pdev);
> + ret = mixer_resources_init(ctx, pdev);
> if (ret) {
> DRM_ERROR("mixer_resources_init failed\n");
> goto fail;
> @@ -1230,18 +1228,16 @@ static int mixer_probe(struct platform_device *pdev)
>
> if (ctx->vp_enabled) {
> /* acquire vp resources: regs, irqs, clocks */
> - ret = vp_resources_init(drm_hdmi_ctx, pdev);
> + ret = vp_resources_init(ctx, pdev);
> if (ret) {
> DRM_ERROR("vp_resources_init failed\n");
> goto fail;
> }
> }
>
> - /* attach mixer driver to common hdmi. */
> - exynos_mixer_drv_attach(drm_hdmi_ctx);
> -
> - /* register specific callback point to common hdmi. */
> - exynos_mixer_ops_register(&mixer_ops);
> + mixer_manager.ctx = ctx;
> + platform_set_drvdata(pdev, &mixer_manager);
> + exynos_drm_manager_register(&mixer_manager);
Ditto.
Best regards,
Tomasz
next prev parent reply other threads:[~2013-11-11 6:16 UTC|newest]
Thread overview: 124+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-10-29 16:12 [PATCH v3 00/32] drm/exynos: Refactor parts of the exynos driver Sean Paul
2013-10-29 16:12 ` [PATCH v3 01/32] drm/exynos: Remove useless slab.h include Sean Paul
2013-10-31 10:24 ` Inki Dae
2013-10-31 23:32 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 02/32] drm/exynos: Merge overlay_ops into manager_ops Sean Paul
2013-10-31 23:39 ` Tomasz Figa
2013-11-01 19:50 ` Sean Paul
2013-11-01 19:55 ` Tomasz Figa
2013-11-04 7:44 ` Inki Dae
2013-10-29 16:12 ` [PATCH v3 03/32] drm/exynos: Add an initialize function to manager and display Sean Paul
2013-10-31 23:42 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 04/32] drm/exynos: Use manager_op initialize in fimd Sean Paul
2013-10-31 23:49 ` Tomasz Figa
2013-11-01 19:51 ` Sean Paul
2013-11-01 19:57 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 05/32] drm/exynos: hdmi: Implement initialize op for hdmi Sean Paul
2013-10-31 23:53 ` Tomasz Figa
2013-11-01 19:54 ` Sean Paul
2013-11-01 19:56 ` Tomasz Figa
2013-11-01 20:08 ` Sean Paul
2013-10-29 16:12 ` [PATCH v3 06/32] drm/exynos: Pass exynos_drm_manager in manager ops instead of dev Sean Paul
2013-11-01 0:19 ` Tomasz Figa
2013-11-01 20:01 ` Sean Paul
2013-11-01 20:11 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 07/32] drm/exynos: Remove apply manager callback Sean Paul
2013-11-08 21:05 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 08/32] drm/exynos: Remove dpms link between encoder/connector Sean Paul
2013-11-08 21:45 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 09/32] drm/exynos: Rename display_op power_on to dpms Sean Paul
2013-11-08 22:09 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 10/32] drm/exynos: Don't keep dpms state in encoder Sean Paul
2013-11-10 20:47 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 11/32] drm/exynos: Use unsigned long for possible_crtcs Sean Paul
2013-11-10 20:47 ` Tomasz Figa
2013-10-29 16:12 ` [PATCH v3 12/32] drm/exynos: Split manager/display/subdrv Sean Paul
2013-10-31 10:30 ` Inki Dae
2013-10-31 16:08 ` Sean Paul
2013-11-01 4:20 ` Inki Dae
2013-11-10 21:09 ` Tomasz Figa
2013-11-12 17:51 ` Sean Paul
2013-11-12 18:35 ` Tomasz Figa
2013-11-26 18:00 ` Olof Johansson
2013-11-27 10:04 ` Thierry Reding
2013-11-28 23:04 ` Tomasz Figa
2013-11-29 7:52 ` Daniel Vetter
2013-11-29 9:10 ` Tomasz Figa
2013-11-29 10:25 ` Daniel Vetter
2013-11-29 14:13 ` Rob Clark
2013-11-29 17:05 ` Tomasz Figa
2013-11-29 18:35 ` Rob Clark
2013-11-30 5:25 ` Inki Dae
2013-12-03 21:38 ` Sean Paul
2013-11-29 10:16 ` Thierry Reding
2013-10-29 16:12 ` [PATCH v3 13/32] drm/exynos: hdmi: remove the i2c drivers and use devtree Sean Paul
2013-11-10 20:46 ` Tomasz Figa
2013-11-11 8:44 ` Thierry Reding
2013-11-28 13:30 ` Tomasz Figa
2013-11-29 10:24 ` Thierry Reding
2013-12-03 0:37 ` Olof Johansson
2013-10-29 16:13 ` [PATCH v3 14/32] drm/exynos: Remove exynos_drm_hdmi shim Sean Paul
2013-11-10 21:24 ` Tomasz Figa [this message]
2013-10-29 16:13 ` [PATCH v3 15/32] drm/exynos: Use drm_mode_copy to copy modes Sean Paul
2013-11-10 21:27 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 16/32] drm/exynos: Disable unused crtc planes from crtc Sean Paul
2013-11-10 21:29 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 17/32] drm/exynos: Add mode_set manager operation Sean Paul
2013-11-10 21:31 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 18/32] drm/exynos: Implement mode_fixup " Sean Paul
2013-11-10 21:33 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 19/32] drm/exynos: Use mode_set to configure fimd Sean Paul
2013-11-10 22:03 ` Tomasz Figa
2013-11-15 13:49 ` Daniel Kurtz
2013-11-15 13:53 ` Daniel Kurtz
2013-11-28 22:57 ` Tomasz Figa
2013-12-04 22:37 ` Sean Paul
2013-10-29 16:13 ` [PATCH v3 20/32] drm/exynos: Remove unused/useless fimd_context members Sean Paul
2013-11-11 1:19 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 21/32] drm/exynos: Move dp driver from video/ to drm/ Sean Paul
2013-10-31 10:46 ` Inki Dae
2013-10-31 16:05 ` Sean Paul
2013-10-31 23:06 ` Jingoo Han
2013-10-31 23:11 ` Tomasz Figa
2013-10-31 23:23 ` Jingoo Han
2013-10-31 23:27 ` Tomasz Figa
2013-10-31 23:55 ` Jingoo Han
2013-11-01 0:01 ` Tomasz Figa
[not found] ` <3513711.0qTZKxmOZX@flatron>
2013-12-04 23:07 ` Sean Paul
2013-10-29 16:13 ` [PATCH v3 22/32] drm/exynos: Move display implementation into dp Sean Paul
[not found] ` <1383063198-10526-23-git-send-email-seanpaul-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
2013-11-11 1:53 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 23/32] ARM: dts: Move display-timings node from fimd to dp Sean Paul
2013-10-29 16:13 ` [PATCH v3 24/32] drm/exynos: Implement dpms display callback in DP Sean Paul
2013-11-11 2:04 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 25/32] drm/exynos: Clean up FIMD power on/off routines Sean Paul
2013-10-31 10:54 ` Inki Dae
[not found] ` <1630995.NnKzZB9Rl5@flatron>
2013-11-11 4:08 ` Inki Dae
2013-11-11 2:09 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 26/32] drm/exynos: Consolidate suspend/resume in drm_drv Sean Paul
2013-11-29 14:58 ` Tomasz Figa
2013-12-19 16:48 ` Inki Dae
2013-10-29 16:13 ` [PATCH v3 27/32] drm/exynos: Add create_connector callback Sean Paul
2013-11-11 2:19 ` Tomasz Figa
2013-12-03 5:01 ` Inki Dae
2013-10-29 16:13 ` [PATCH v3 28/32] drm/exynos: Implement drm_connector in hdmi directly Sean Paul
2013-11-29 15:58 ` Tomasz Figa
2013-12-02 9:46 ` Thierry Reding
2013-12-02 9:54 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 29/32] drm/exynos: Implement drm_connector directly in dp driver Sean Paul
2013-11-29 16:04 ` Tomasz Figa
2013-12-03 4:45 ` Inki Dae
2013-12-04 6:46 ` Inki Dae
2013-10-29 16:13 ` [PATCH v3 30/32] drm/exynos: Implement drm_connector directly in vidi driver Sean Paul
2013-11-29 16:13 ` Tomasz Figa
2013-12-03 4:47 ` Inki Dae
2013-12-04 6:47 ` Inki Dae
2013-10-29 16:13 ` [PATCH v3 31/32] drm/exynos: Move lvds bridge discovery into DP driver Sean Paul
2013-11-29 16:55 ` Tomasz Figa
2013-11-30 5:18 ` Inki Dae
2013-11-30 12:27 ` Tomasz Figa
2013-10-29 16:13 ` [PATCH v3 32/32] drm/exynos: Remove the exynos_drm_connector shim Sean Paul
2013-11-29 16:14 ` Tomasz Figa
2013-11-07 5:48 ` [PATCH v3 00/32] drm/exynos: Refactor parts of the exynos driver Inki Dae
2013-11-07 18:20 ` Sean Paul
2013-11-08 21:01 ` Tomasz Figa
2013-11-08 21:22 ` Sean Paul
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=1526769.j7lMfrmRht@flatron \
--to=tomasz.figa@gmail.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=marcheu@chromium.org \
--cc=seanpaul@chromium.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).