| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <55FAA679.2030902@codeaurora.org>
Date: Thu, 17 Sep 2015 17:09:37 +0530
From: Archit Taneja <architt@...eaurora.org>
To: Xinwei Kong <kong.kongxinwei@...ilicon.com>, airlied@...ux.ie,
corbet@....net, catalin.marinas@....com, will.deacon@....com
Cc: dri-devel@...ts.freedesktop.org, linux-doc@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org, linuxarm@...wei.com,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
xinliang.liu@...aro.org, andy.green@...aro.org,
qijiwen@...ilicon.com, gongyu@...ilicon.com,
haojian.zhuang@...aro.org, liguozhu@...ilicon.com,
xuwei5@...ilicon.com, yinshengbao@...ilicon.com,
yanhaifeng@...ilicon.com, ml.yang@...ilicon.com, yimin@...wei.com,
w.f@...wei.com, puck.chen@...ilicon.com, bintian.wang@...wei.com,
benjamin.gaignard@...aro.org, xuyiping@...ilicon.com,
james.yanglong@...ilicon.com, fangdechun@...ilicon.com
Subject: Re: [PATCH RFC 5/8] drm: hisilicon: fill interface function of
encoder\connector part
On 9/15/2015 3:07 PM, Xinwei Kong wrote:
> This patch enables the adv7533 module which is connecting hisilicon SOC
> by dsi module. while using DSI module and adv7533 module to implement the
> encoder/connector interface of DRM\KMS.
>
> Signed-off-by: Xinliang Liu <xinliang.liu@...aro.org>
> Signed-off-by: Xinwei Kong <kong.kongxinwei@...ilicon.com>
> Signed-off-by: Andy Green <andy.green@...aro.org>
> Signed-off-by: Jiwen Qi <qijiwen@...ilicon.com>
> Signed-off-by: Yu Gong <gongyu@...ilicon.com>
> ---
> drivers/gpu/drm/hisilicon/Kconfig | 10 +
> drivers/gpu/drm/hisilicon/hisi_drm_connector.c | 34 ++
> drivers/gpu/drm/hisilicon/hisi_drm_connector.h | 8 +
> drivers/gpu/drm/hisilicon/hisi_drm_dsi.c | 670 +++++++++++++++++++++++++
> drivers/gpu/drm/hisilicon/hisi_drm_encoder.c | 52 ++
> drivers/gpu/drm/hisilicon/hisi_drm_encoder.h | 19 +
> drivers/gpu/drm/hisilicon/hisi_dsi_reg.h | 91 ++++
> 7 files changed, 884 insertions(+)
> create mode 100644 drivers/gpu/drm/hisilicon/hisi_dsi_reg.h
>
<snip>
> +struct hisi_connector_funcs hisi_dsi_connector_ops = {
> + .get_modes = dsi_connector_get_modes,
> +};
> +
> +struct hisi_encoder_funcs hisi_dsi_encoder_ops = {
> + .mode_set = dsi_encoder_mode_set,
> + .enable = dsi_encoder_enable,
> + .disable = dsi_encoder_disable,
> +};
> +
> static int hisi_dsi_bind(struct device *dev, struct device *master,
> void *data)
> {
> @@ -52,8 +682,23 @@ static int hisi_dsi_bind(struct device *dev, struct device *master,
>
> hisi_drm_encoder_init(ctx->dev, &ctx->dsi.hisi_encoder.base.base);
>
> +#ifdef CONFIG_DRM_HISI_HAS_SLAVE_ENCODER
> + ret = ctx->drm_i2c_driver->encoder_init(ctx->client, ctx->dev,
> + &ctx->dsi.hisi_encoder.base);
> + if (ret) {
> + DRM_ERROR("fail to init drm i2c encoder\n");
> + return ret;
> + }
> +
> + if (!ctx->dsi.hisi_encoder.base.slave_funcs) {
> + DRM_ERROR("failed check encoder function\n");
> + return -ENODEV;
> + }
> +#endif
A slave encoder isn't supposed to be used this way. A slave encoder is
supposed to register itself to drm as an encoder object. A slave
encoder connects directly to a crtc, and operate like a normal encoder.
In other words, a slave encoder isn't something that attaches to a
'master' encoder as the driver does here. 'Slave encoder' here means
that the encoder is a part of a different bus (like i2c etc).
Currently, the driver is directly calling slave encoder ops in
hisi_drm_encoder.c.
> +
> hisi_drm_connector_init(ctx->dev, &ctx->dsi.hisi_encoder.base.base,
> &ctx->dsi.hisi_connector.connector);
> +
> return ret;
> }
>
> @@ -102,6 +747,27 @@ static int hisi_dsi_probe(struct platform_device *pdev)
> return -EINVAL;
> }
>
> +#ifdef CONFIG_DRM_HISI_HAS_SLAVE_ENCODER
> + ctx->client = of_find_i2c_device_by_node(slave_node);
> + of_node_put(slave_node);
> + if (!ctx->client) {
> + DRM_ERROR("failed to find slave encoder i2c client\n");
> + return -EPROBE_DEFER;
> + }
> +
> + if (!ctx->client->dev.driver) {
> + DRM_ERROR("%s: NULL client driver\n", __func__);
> + return -EPROBE_DEFER;
> + }
> +
> + ctx->drm_i2c_driver = to_drm_i2c_encoder_driver(
> + to_i2c_driver(ctx->client->dev.driver));
> + if (IS_ERR(ctx->drm_i2c_driver)) {
> + DRM_ERROR("failed to initialize encoder driver");
> + return -EPROBE_DEFER;
> + }
> +#endif
> +
> dsi = &ctx->dsi;
> dsi->ctx = ctx;
> dsi->lanes = 3;
> @@ -110,6 +776,10 @@ static int hisi_dsi_probe(struct platform_device *pdev)
> dsi->format = MIPI_DSI_FMT_RGB888;
> dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_SYNC_PULSE;
>
> + dsi->hisi_encoder.ops = &hisi_dsi_encoder_ops;
> + dsi->hisi_connector.encoder = &dsi->hisi_encoder.base.base;
> + dsi->hisi_connector.ops = &hisi_dsi_connector_ops;
> +
> return component_add(&pdev->dev, &hisi_dsi_ops);
> }
>
> diff --git a/drivers/gpu/drm/hisilicon/hisi_drm_encoder.c b/drivers/gpu/drm/hisilicon/hisi_drm_encoder.c
> index 89fc73d..acd73d8 100644
> --- a/drivers/gpu/drm/hisilicon/hisi_drm_encoder.c
> +++ b/drivers/gpu/drm/hisilicon/hisi_drm_encoder.c
> @@ -15,18 +15,49 @@
>
> #include "hisi_drm_encoder.h"
>
> +#define to_hisi_encoder(encoder) \
> + container_of(encoder, struct hisi_encoder, base.base)
> +
> void hisi_drm_encoder_disable(struct drm_encoder *encoder)
> {
> + struct hisi_encoder *hencoder = to_hisi_encoder(encoder);
> + struct hisi_encoder_funcs *ops = hencoder->ops;
> + struct drm_encoder_slave_funcs *sfuncs = get_slave_funcs(encoder);
> +
> + if (ops->enable)
> + ops->disable(encoder);
> +
> + if (sfuncs && sfuncs->dpms)
> + sfuncs->dpms(encoder, DRM_MODE_DPMS_OFF);
> }
>
The encoder should represent only the DSI encoder hardware here. It
shouldn't be calling slave encoder ops.
What you want is to have the encoder connect to a bridge. Something
like:
crtc(ade)->encoder(dsi)->bridge(adv7533)->connector
Archit
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists