Re: [PATCH 08/12] drm: bridge/dw_hdmi: avoid enabling interface in mode_set

From: Yakir Yang <ykk@rock-chips.com>
To: Russell King - ARM Linux <linux@arm.linux.org.uk>
Cc: linux-rockchip@lists.infradead.org,
	dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	Fabio Estevam <fabio.estevam@freescale.com>,
	David Airlie <airlied@linux.ie>,
	Sascha Hauer <s.hauer@pengutronix.de>,
	Philipp Zabel <p.zabel@pengutronix.de>,
	Andy Yan <andy.yan@rock-chips.com>,
	Jon Nettleton <jon.nettleton@gmail.com>
Subject: Re: [PATCH 08/12] drm: bridge/dw_hdmi: avoid enabling interface in mode_set
Date: Wed, 07 Oct 2015 17:35:29 +0800	[thread overview]
Message-ID: <5614E761.8020706@rock-chips.com> (raw)
In-Reply-To: <20151007091830.GE21513@n2100.arm.linux.org.uk>

On 10/07/2015 05:18 PM, Russell King - ARM Linux wrote:
> On Wed, Oct 07, 2015 at 11:50:53AM +0800, Yakir Yang wrote:
>>
>> On 08/09/2015 12:04 AM, Russell King wrote:
>>> On a mode set, DRM makes the following sequence of calls:
>>> * for_each_encoder
>>> *   bridge	mode_fixup
>>> *   encoder	mode_fixup
>>> * crtc		mode_fixup
>>> * for_each_encoder
>>> *   bridge	disable
>>> *   encoder	prepare
>>> *   bridge	post_disable
>>> * disable unused encoders
>>> * crtc		prepare
>>> * crtc		mode_set
>>> * for_each_encoder
>>> *   encoder	mode_set
>>> *   bridge	mode_set
>>> * crtc		commit
>>> * for_each_encoder
>>> *   bridge	pre_enable
>>> *   encoder	commit
>>> *   bridge	enable
>>>
>>> dw_hdmi enables the HDMI output in both the bridge mode_set() and also
>>> the bridge enable() step.  This is duplicated work - we can avoid the
>>> setup in mode_set() and just do it in the enable() stage.  This
>>> simplifies the code a little.
>>>
>>> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
>> I have noticed that dw_hdmi driver have poweron/poweroff when
>> driver detect HPD event in irq thread, that's also duplicated works,
>> would you like to collect that changes into this one:
>>
>> static irqreturn_t dw_hdmi_irq(int irq, void *dev_id)
>> {
>>          ......
>>
>>          if (intr_stat & HDMI_IH_PHY_STAT0_HPD) {
>>                  if (phy_int_pol & HDMI_PHY_HPD) {
>>                          dev_dbg(hdmi->dev, "EVENT=plugin\n");
>>
>>                          hdmi_modb(hdmi, 0, HDMI_PHY_HPD, HDMI_PHY_POL0);
>>
>>                          dw_hdmi_poweron(hdmi);    // no need here
>>                  } else {
>>                          dev_dbg(hdmi->dev, "EVENT=plugout\n");
>>
>>                          hdmi_modb(hdmi, HDMI_PHY_HPD, HDMI_PHY_HPD,
>>                                    HDMI_PHY_POL0);
>>
>>                          dw_hdmi_poweroff(hdmi);    // no need here
>>                  }
>>                  drm_helper_hpd_irq_event(hdmi->connector.dev);
>>          }
>>          ......
>> }
> I'm very much of the opinion of making small logical changes.  This
> patch is one small logical change to the DRM-side logic to get rid
> of the identified duplication there without touching anything else.
> If removing the above calls to dw_hdmi_poweron()/dw_hdmi_poweroff()
> were found to cause a regression, then the whole change would end
> up being reverted, which would be annoying.

Hmm... Yeah, it do make some driver logical changes, but I
thought that's good, just make a clean on HPD thread, and I
do give lots of test on chrome tree about this changes, guess
a separate patch would be better.

If you don't feel good enough about this, okay, I would give
more test on that changes, and send upstream to request
comment later.

- Yakir