[PATCH 1/3 v2] drm/i2c/adv7511: Add audio support
Archit Taneja
architt at codeaurora.org
Tue Mar 29 10:03:19 PDT 2016
On 3/29/2016 4:22 PM, Jose Abreu wrote:
> Hi Archit,
>
> On 29-03-2016 09:05, Archit Taneja wrote:
>> Hi,
>>
>> On 03/28/2016 08:06 PM, Jose Abreu wrote:
>>> This patch adds audio support for the ADV7511 HDMI transmitter
>>> using ALSA SoC.
>>>
>>> The code was ported from Analog Devices linux tree from
>>> commit 1770c4a1e32b ("Merge remote-tracking branch
>>> 'xilinx/master' into xcomm_zynq"), which is available at:
>>> - https://github.com/analogdevicesinc/linux/
>>>
>>> The main core file was renamed from adv7511.c to adv7511_core.c
>>> so that audio and video compile into a single adv7511.ko module
>>> and to keep up with Analog Devices kernel tree.
>>>
>>> The audio can be disabled using menu-config so it is possible
>>> to use only video mode.
>>>
>>> The HDMI mode is automatically started at boot and the audio
>>> (when enabled) registers as a codec into ALSA.
>>
>> Is there a reason why we set the mode to HDMI at probe itself?
>> Shouldn't it be okay to set the mode later once we read the
>> EDID off the panel?
>>
>> Some more comments below.
>>
>
> Well, when I was using this in kernel 3.18 (with an older version of the driver)
> I noticed that DVI mode was being used even when HDMI was connected so I forced
> the driver to start in HDMI mode. There were some changes in the driver so it is
> possible that this is no longer needed. Should I drop it?
Mode selection works fine with ADV7533 on a 4.5 kernel. I'm assuming it
should work out of the box for ADV7511 too. We should drop this.
>
>>>
>>> SPDIF DAI format was also added to ASoC as it is required
>>> by adv7511 audio.
>>>
>>> Signed-off-by: Jose Abreu <joabreu at synopsys.com>
>>> ---
>>>
>>> No changes v1 -> v2.
>>>
>>> drivers/gpu/drm/i2c/Kconfig | 11 +
>>> drivers/gpu/drm/i2c/Makefile | 2 +
>>> drivers/gpu/drm/i2c/adv7511.c | 1024 -----------------------------------
>>> drivers/gpu/drm/i2c/adv7511.h | 41 ++
>>> drivers/gpu/drm/i2c/adv7511_audio.c | 310 +++++++++++
>>> drivers/gpu/drm/i2c/adv7511_core.c | 1005 ++++++++++++++++++++++++++++++++++
>>> include/sound/soc-dai.h | 1 +
>>> 7 files changed, 1370 insertions(+), 1024 deletions(-)
>>> delete mode 100644 drivers/gpu/drm/i2c/adv7511.c
>>> create mode 100644 drivers/gpu/drm/i2c/adv7511_audio.c
>>> create mode 100644 drivers/gpu/drm/i2c/adv7511_core.c
>>
>> <snip>
>>
>>> +
>>> +static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id
>>> *id)
>>> +{
>>> + struct adv7511_link_config link_config;
>>> + struct adv7511 *adv7511;
>>> + struct device *dev = &i2c->dev;
>>> + unsigned int val;
>>> + int ret;
>>> +
>>> + if (!dev->of_node)
>>> + return -EINVAL;
>>> +
>>> + adv7511 = devm_kzalloc(dev, sizeof(*adv7511), GFP_KERNEL);
>>> + if (!adv7511)
>>> + return -ENOMEM;
>>> +
>>> + adv7511->powered = false;
>>> + adv7511->status = connector_status_disconnected;
>>> +
>>> + ret = adv7511_parse_dt(dev->of_node, &link_config);
>>> + if (ret)
>>> + return ret;
>>> +
>>> + /*
>>> + * The power down GPIO is optional. If present, toggle it from active to
>>> + * inactive to wake up the encoder.
>>> + */
>>> + adv7511->gpio_pd = devm_gpiod_get_optional(dev, "pd", GPIOD_OUT_HIGH);
>>> + if (IS_ERR(adv7511->gpio_pd))
>>> + return PTR_ERR(adv7511->gpio_pd);
>>> +
>>> + if (adv7511->gpio_pd) {
>>> + mdelay(5);
>>> + gpiod_set_value_cansleep(adv7511->gpio_pd, 0);
>>> + }
>>> +
>>> + adv7511->regmap = devm_regmap_init_i2c(i2c, &adv7511_regmap_config);
>>> + if (IS_ERR(adv7511->regmap))
>>> + return PTR_ERR(adv7511->regmap);
>>> +
>>> + ret = regmap_read(adv7511->regmap, ADV7511_REG_CHIP_REVISION, &val);
>>> + if (ret)
>>> + return ret;
>>> + dev_dbg(dev, "Rev. %d\n", val);
>>> +
>>> + ret = regmap_register_patch(adv7511->regmap, adv7511_fixed_registers,
>>> + ARRAY_SIZE(adv7511_fixed_registers));
>>> + if (ret)
>>> + return ret;
>>> +
>>> + regmap_write(adv7511->regmap, ADV7511_REG_EDID_I2C_ADDR, edid_i2c_addr);
>>> + regmap_write(adv7511->regmap, ADV7511_REG_PACKET_I2C_ADDR,
>>> + packet_i2c_addr);
>>> + regmap_write(adv7511->regmap, ADV7511_REG_CEC_I2C_ADDR, cec_i2c_addr);
>>> + adv7511_packet_disable(adv7511, 0xffff);
>>> +
>>> + adv7511->i2c_main = i2c;
>>> + adv7511->i2c_edid = i2c_new_dummy(i2c->adapter, edid_i2c_addr >> 1);
>>> + if (!adv7511->i2c_edid)
>>> + return -ENOMEM;
>>> +
>>> + if (i2c->irq) {
>>> + init_waitqueue_head(&adv7511->wq);
>>> +
>>> + ret = devm_request_threaded_irq(dev, i2c->irq, NULL,
>>> + adv7511_irq_handler,
>>> + IRQF_ONESHOT, dev_name(dev),
>>> + adv7511);
>>> + if (ret)
>>> + goto err_i2c_unregister_device;
>>> + }
>>> +
>>> + /* CEC is unused for now */
>>> + regmap_write(adv7511->regmap, ADV7511_REG_CEC_CTRL,
>>> + ADV7511_CEC_CTRL_POWER_DOWN);
>>> +
>>> + adv7511_power_off(adv7511);
>>> +
>>> + i2c_set_clientdata(i2c, adv7511);
>>> +
>>> +#ifdef CONFIG_DRM_I2C_ADV7511_AUDIO
>>> + adv7511_audio_init(&i2c->dev);
>>> +#endif
>>
>> If we intend to have more audio funcs being used by the core in the
>> future, it would be nice to have NOP audio funcs rather than having
>> multiple #ifdef checks in the driver when CONFIG_DRM_I2C_ADV7511_AUDIO
>> isn't set.
>>
>
> I will move this ifdef to adv751_audio and use NOP functions.
Thanks, I think it should help in the longer run.
>
>>> +
>>> + adv7511_set_link_config(adv7511, &link_config);
>>> +
>>> + /* Enable HDMI mode */
>>> + regmap_update_bits(adv7511->regmap, ADV7511_REG_HDCP_HDMI_CFG,
>>> + ADV7511_HDMI_CFG_MODE_MASK,
>>> + ADV7511_HDMI_CFG_MODE_HDMI);
>>> +
>>> + return 0;
>>> +
>>> +err_i2c_unregister_device:
>>> + i2c_unregister_device(adv7511->i2c_edid);
>>> +
>>> + return ret;
>>> +}
>>> +
>>> +static int adv7511_remove(struct i2c_client *i2c)
>>> +{
>>> + struct adv7511 *adv7511 = i2c_get_clientdata(i2c);
>>> +
>>
>> Are we missing a call to adv7511_audio_exit() here?
>
> I followed the code in Analog Devices tree where there is no call to
> audio_exit() but indeed you are correct. I will add this call.
>
Since we have 3 files for adv7511 now, could we also move the driver
to a separate folder? The long term plan is to convert all the i2c
slave encoder drivers to bridges. Keeping them together would be nicer
when we migrate this driver to the bridge folder.
Thanks,
Archit
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
More information about the linux-snps-arc
mailing list