[Intel-gfx] [PATCH v2 2/2] drm: parse hf-vsdb
Sharma, Shashank
shashank.sharma at intel.com
Wed Dec 21 03:54:33 UTC 2016
Regards
Shashank
On 12/20/2016 8:05 PM, Jose Abreu wrote:
> Hi Shashank,
>
>
> On 20-12-2016 13:47, Shashank Sharma wrote:
>> HDMI 2.0 / CEA-861-F specs define a new CEA extension data block,
>> called hdmi-forum vendor specific data block (HF-VSDB). This block
>> contains information about sink's support for HDMI 2.0 compliant
>> features. These features are:
>> - Deep color YUV 420 support and BPC
>> - 3D flags for
>> - OSD Displarity
>> - Dual view signaling
>> - independent view signaling
>> - SCDC support
>> - Max TMDS char rate
>> - Scrambling support
>>
>> This patch adds a parser function for this block, and add flags to
>> indicate support for new features, in drm_display_info structure
>>
>> V2:
>> - Addressed review comments from Thierry
>> - remove len > 31 check
>> - remove version check
>> - fix duplicate values for macros of 36 and 30-bit depths
>> - Added a sub-class for HDMI related information within drm_display_info
>> (Thierry, Daniel) and populated it with HF-VSDB specific info.
>>
>> Cc: Thierry Reding <treding at nvidia.com>
>> Cc: Daniel Vetter <daniel.vetter at intel.com>
>> Cc: Jose Abreu <joabreu at synopsys.com>
>> Signed-off-by: Shashank Sharma <shashank.sharma at intel.com>
>> ---
>> drivers/gpu/drm/drm_edid.c | 70 ++++++++++++++++++++++++++++++++++++++++++++++
>> include/drm/drm_edid.h | 5 ++++
>> include/linux/hdmi.h | 1 +
>> 3 files changed, 76 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>> index b552197..59e04fb 100644
>> --- a/drivers/gpu/drm/drm_edid.c
>> +++ b/drivers/gpu/drm/drm_edid.c
>> @@ -3224,6 +3224,23 @@ static int add_3d_struct_modes(struct drm_connector *connector, u16 structure,
>> return 0;
>> }
>>
>> +static bool cea_db_is_hf_vsdb(const u8 *db)
>> +{
>> + u8 len;
>> + int hfvsdb_id;
>> +
>> + if (cea_db_tag(db) != VENDOR_BLOCK)
>> + return false;
>> +
>> + len = cea_db_payload_len(db);
>> + if (len < 7)
>> + return false;
>> +
>> + hfvsdb_id = db[1] | (db[2] << 8) | (db[3] << 16);
>> +
>> + return hfvsdb_id == HDMI_IEEE_OUI_HFVSDB;
>> +}
>> +
>> static bool cea_db_is_hdmi_vsdb(const u8 *db)
>> {
>> int hdmi_id;
>> @@ -3768,6 +3785,57 @@ bool drm_rgb_quant_range_selectable(struct edid *edid)
>> }
>> EXPORT_SYMBOL(drm_rgb_quant_range_selectable);
>>
>> +static void drm_parse_yuv420_deep_color_info(struct drm_connector *connector,
>> + const u8 *db)
>> +{
>> + struct drm_hdmi_info *info = &connector->display_info.hdmi_info;
>> +
>> + if (db[7] & DRM_EDID_YUV420_DC_48)
>> + info->edid_yuv420_dc_modes |= DRM_EDID_YUV420_DC_48;
>> + if (db[7] & DRM_EDID_YUV420_DC_36)
>> + info->edid_yuv420_dc_modes |= DRM_EDID_YUV420_DC_36;
>> + if (db[7] & DRM_EDID_YUV420_DC_30)
>> + info->edid_yuv420_dc_modes |= DRM_EDID_YUV420_DC_30;
>> +
>> + if (!info->edid_yuv420_dc_modes) {
>> + DRM_DEBUG("%s: No YUV 420 deep color support in sink.\n",
>> + connector->name);
>> + return;
>> + }
>> +}
>> +
>> +static void
>> +drm_parse_hf_vsdb(struct drm_connector *connector, const u8 *db)
>> +{
>> + struct drm_display_info *info = &connector->display_info;
>> + struct drm_hdmi_info *hdmi_info = &info->hdmi_info;
>> +
>> + if (db[5]) {
>> + /*
>> + * If the sink supplies max tmds char rate in db,
>> + * the actual max tmds rate = db[5] * 5Mhz.
>> + */
>> + info->max_tmds_clock = db[5] * 5000;
>> + DRM_DEBUG_KMS("HF-VSDB: max TMDS clock %d kHz\n",
>> + info->max_tmds_clock);
>> + }
>> +
>> + if (db[6] & DRM_HFVSDB_SCDC_SUPPORT)
>> + hdmi_info->scdc_supported = true;
>> + if (db[6] & DRM_HFVSDB_SCDC_RR_CAP)
>> + hdmi_info->scdc_rr_cap = true;
>> + if (db[6] & DRM_HFVSDB_SCRAMBLING)
>> + hdmi_info->scrambling = true;
> According to spec you should also check if scdc is supported in
> order to support scrambling.
Agree, rare case but fairly possible.
>> + if (db[6] & DRM_HFVSDB_INDEPENDENT_VIEW)
>> + hdmi_info->independent_view_3d = true;
>> + if (db[6] & DRM_HFVSDB_DUAL_VIEW)
>> + hdmi_info->dual_view_3d = true;
>> + if (db[6] & DRM_HFVSDB_3D_OSD)
>> + hdmi_info->osd_disparity_3d = true;
>> +
>> + drm_parse_yuv420_deep_color_info(connector, db);
>> +}
>> +
>> static void drm_parse_hdmi_deep_color_info(struct drm_connector *connector,
>> const u8 *hdmi)
>> {
>> @@ -3882,6 +3950,8 @@ static void drm_parse_cea_ext(struct drm_connector *connector,
>>
>> if (cea_db_is_hdmi_vsdb(db))
>> drm_parse_hdmi_vsdb_video(connector, db);
>> + if (cea_db_is_hf_vsdb(db))
>> + drm_parse_hf_vsdb(connector, db);
>> }
>> }
>>
>> diff --git a/include/drm/drm_edid.h b/include/drm/drm_edid.h
>> index 38eabf6..df606e3 100644
>> --- a/include/drm/drm_edid.h
>> +++ b/include/drm/drm_edid.h
>> @@ -212,6 +212,11 @@ struct detailed_timing {
>> #define DRM_EDID_HDMI_DC_30 (1 << 4)
>> #define DRM_EDID_HDMI_DC_Y444 (1 << 3)
>>
>> +/* YUV 420 deep color modes */
>> +#define DRM_EDID_YUV420_DC_48 (1 << 6)
>> +#define DRM_EDID_YUV420_DC_36 (1 << 5)
>> +#define DRM_EDID_YUV420_DC_30 (1 << 4)
> Hmm, I think this is wrong. According to spec DC 48 is at
> position 2, DC 36 at position 1 and DC 30 at position 0.
Oh, thanks for pointing this out, I dont know which version of spec did
I refer to while I was writing the code (or may be It was too late in
night :-))
>> +
>> /* ELD Header Block */
>> #define DRM_ELD_HEADER_BLOCK_SIZE 4
>>
>> diff --git a/include/linux/hdmi.h b/include/linux/hdmi.h
>> index edbb4fc..2d4e9ef 100644
>> --- a/include/linux/hdmi.h
>> +++ b/include/linux/hdmi.h
>> @@ -35,6 +35,7 @@ enum hdmi_infoframe_type {
>> };
>>
>> #define HDMI_IEEE_OUI 0x000c03
>> +#define HDMI_IEEE_OUI_HFVSDB 0xC45DD8
> Lower case, because HDMI_IEEE_OUI is also lower case.
Sure.
>
>> #define HDMI_INFOFRAME_HEADER_SIZE 4
>> #define HDMI_AVI_INFOFRAME_SIZE 13
>> #define HDMI_SPD_INFOFRAME_SIZE 25
> Best regards,
> Jose Miguel Abreu
More information about the Intel-gfx
mailing list