If the HPD signal never asserts in panel_simple_prepare() and we return an error, we should unset the enable GPIO and disable the regulator to make it consistent for the caller.

At the moment I have some hardware where HPD sometimes doesn't assert. Obviously that needs to be debugged, but this patch makes it so that if I add a retry that I can make things work.

Fixes: 48834e6084f1 ("drm/panel-simple: Support hpd-gpios for delaying prepare()") Signed-off-by: Douglas Anderson dianders@chromium.org --- Stephen: I didn't take your Reviewed-by tag since I made a small change. If you're OK with this change please re-add.

Changes in v2: - Set the "unprepared_time" so if we retry we give the proper delay.

drivers/gpu/drm/panel/panel-simple.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index 71ae200ac48a..581ab6810b70 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -406,7 +406,7 @@ static int panel_simple_prepare(struct drm_panel *panel) if (IS_ERR(p->hpd_gpio)) { err = panel_simple_get_hpd_gpio(panel->dev, p, false); if (err) - return err; + goto error; }

err = readx_poll_timeout(gpiod_get_value_cansleep, p->hpd_gpio, @@ -418,13 +418,20 @@ static int panel_simple_prepare(struct drm_panel *panel) if (err) { dev_err(panel->dev, "error waiting for hpd GPIO: %d\n", err); - return err; + goto error; } }

p->prepared_time = ktime_get();

return 0; + +error: + gpiod_set_value_cansleep(p->enable_gpio, 0); + regulator_disable(p->supply); + p->unprepared_time = ktime_get(); + + return err; }

static int panel_simple_enable(struct drm_panel *panel)

If a panel has an hpd_absent_delay specified then we know exactly how long the maximum time is before HPD must be asserted. That means we can use it as a timeout for polling the HPD pin instead of using an arbitrary timeout. This is especially useful for dealing with panels that periodically fail to power on and need to be retried. We can detect the problem sooner.

Signed-off-by: Douglas Anderson dianders@chromium.org ---

Changes in v2: - ("drm/panel-simple: Don't wait longer for HPD...") new for v2.

drivers/gpu/drm/panel/panel-simple.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index 581ab6810b70..30842cf6d414 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -382,6 +382,7 @@ static int panel_simple_prepare(struct drm_panel *panel) unsigned int delay; int err; int hpd_asserted; + unsigned long hpd_wait_us;

if (p->prepared_time != 0) return 0; @@ -409,9 +410,14 @@ static int panel_simple_prepare(struct drm_panel *panel) goto error; }

+ if (p->desc->delay.hpd_absent_delay) + hpd_wait_us = p->desc->delay.hpd_absent_delay * 1000UL; + else + hpd_wait_us = 2000000; + err = readx_poll_timeout(gpiod_get_value_cansleep, p->hpd_gpio, hpd_asserted, hpd_asserted, - 1000, 2000000); + 1000, hpd_wait_us); if (hpd_asserted < 0) err = hpd_asserted;

On an Innolux N116BCA panel that I have in front of me, sometimes HPD simply doesn't assert no matter how long you wait for it. As per the very wise advice of The IT Crowd ("Have you tried turning it off and on again?") it appears that power cycling is enough to kick this panel back into a sane state.

From tests on this panel, it appears that leaving it powered off for a

while stimulates the problem. Adding a 6 second sleep at the start of panel_simple_prepare_once() makes it happen fairly reliably and, with this delay, I saw up to 3 retries needed sometimes. Without the 6 second sleep, however, the panel came up much more reliably the first time or after only 1 retry.

While it's unknown what the problems are with this panel (and probably the hardware should be debugged), adding a few retries to the power on routine doesn't seem insane. Even if this panel's problems are attributed to the fact that it's pre-production and/or can be fixed, retries clearly can help in some cases and really don't hurt.

Signed-off-by: Douglas Anderson dianders@chromium.org ---

Changes in v2: - ("drm/panel-simple: Retry if we timeout waiting for HPD") new for v2.

drivers/gpu/drm/panel/panel-simple.c | 32 +++++++++++++++++++++++++--- 1 file changed, 29 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index 30842cf6d414..823177d89d1b 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -376,7 +376,7 @@ static int panel_simple_get_hpd_gpio(struct device *dev, return 0; }

-static int panel_simple_prepare(struct drm_panel *panel) +static int panel_simple_prepare_once(struct drm_panel *panel) { struct panel_simple *p = to_panel_simple(panel); unsigned int delay; @@ -422,8 +422,9 @@ static int panel_simple_prepare(struct drm_panel *panel) err = hpd_asserted;

if (err) { - dev_err(panel->dev, - "error waiting for hpd GPIO: %d\n", err); + if (err != -ETIMEDOUT) + dev_err(panel->dev, + "error waiting for hpd GPIO: %d\n", err); goto error; } } @@ -440,6 +441,31 @@ static int panel_simple_prepare(struct drm_panel *panel) return err; }

+/* + * Some panels simply don't always come up and need to be power cycled to + * work properly. We'll allow for a handful of retries. + */ +#define MAX_PANEL_PREPARE_TRIES 5 + +static int panel_simple_prepare(struct drm_panel *panel) +{ + int ret; + int try; + + for (try = 0; try < MAX_PANEL_PREPARE_TRIES; try++) { + ret = panel_simple_prepare_once(panel); + if (ret != -ETIMEDOUT) + break; + } + + if (ret == -ETIMEDOUT) + dev_err(panel->dev, "Prepare timeout after %d tries\n", try); + else if (try) + dev_warn(panel->dev, "Prepare needed %d retries\n", try); + + return ret; +} + static int panel_simple_enable(struct drm_panel *panel) { struct panel_simple *p = to_panel_simple(panel);