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  PHC 
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sun, 26 Nov 2017 18:26:06 +0530
From:   Archit Taneja <>
To:     John Stultz <>, Arnd Bergmann <>
Cc:     Laurent Pinchart <>,
        David Airlie <>,
        Lars-Peter Clausen <>,
        Bhumika Goyal <>,
        Inki Dae <>,,
        lkml <>,
        Xinliang Liu <>,
        Dan Carpenter <>,
        Sean Paul <>,
        Hans Verkuil <>
Subject: Re: [RFC][PATCH] drm: adv7511/33: Fix adv7511_cec_init() failure


On 11/17/2017 04:29 AM, John Stultz wrote:
> From: Arnd Bergmann <>
> An otherwise correct cleanup patch from Dan Carpenter turned a broken
> failure handling from a feature patch by Hans Verkuil into a kernel
> Oops, so bisection points to commit 7af35b0addbc ("drm/kirin: Checking
> for IS_ERR() instead of NULL") rather than 3b1b975003e4 ("drm:
> adv7511/33: add HDMI CEC support").
> I've managed to piece together several partial problems, though
> I'm still struggling with the bigger picture:
> adv7511_probe() registers a drm_bridge structure that was allocated
> with devm_kzalloc(). It calls adv7511_cec_init(), which fails for an
> unknown reason, which in turn triggers the registered structure to be
> removed.
> Elsewhere, kirin_drm_platform_probe() gets called, which calls
> of_graph_get_remote_node(), and that returns NULL. Before Dan's
> patch we would go on with a NULL pointer here and register that,
> now kirin_drm_platform_probe() fails with -ENODEV.
> In a third driver, dsi_parse_dt() calls drm_of_find_panel_or_bridge(),
> which after not finding a panel goes on to call of_drm_find_bridge(),
> and that crashes due to the earlier list corruption.
> This addresses the first issue by making sure that adv7511_probe()
> does not completely fail when the adv7511_cec_init() function fails,
> and instead we just disable the CEC feature. This avoids having the
> driver entirely fail to load if just the CEC initialization fails.
> Reported-by: Naresh Kamboju <>
> Cc: Xinliang Liu <>
> Cc: Dan Carpenter <>
> Cc: Sean Paul <>
> Cc: Hans Verkuil <>
> Cc: Archit Taneja <>
> Link:
> Link:
> Fixes: 7af35b0addbc ("drm/kirin: Checking for IS_ERR() instead of NULL")
> Fixes: 3b1b975003e4 ("drm: adv7511/33: add HDMI CEC support")
> Signed-off-by: Arnd Bergmann <>
> [jstultz: Reworked so when adv7511_cec_init() fails, we disable the feature instead
> of disabling the entire driver, which causes graphics to not funciton]
> Signed-off-by: John Stultz <>
> ---
> Just wanted to send out my rework of Arnd's patch here.
> Feedback would be welcome.
> thanks
> -john
>   drivers/gpu/drm/bridge/adv7511/adv7511_drv.c | 8 ++++----
>   1 file changed, 4 insertions(+), 4 deletions(-)
> diff --git a/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c b/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c
> index 0e14f15..939c3b9 100644
> --- a/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c
> +++ b/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c
> @@ -1203,12 +1203,12 @@ static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id)
>   #ifdef CONFIG_DRM_I2C_ADV7511_CEC
>   	ret = adv7511_cec_init(dev, adv7511, offset);
> -	if (ret)
> -		goto err_unregister_cec;
>   #else
> -	regmap_write(adv7511->regmap, ADV7511_REG_CEC_CTRL + offset,
> -		     ADV7511_CEC_CTRL_POWER_DOWN);
> +	ret = 1;
>   #endif
> +	if (ret)
> +		regmap_write(adv7511->regmap, ADV7511_REG_CEC_CTRL + offset,
> +			     ADV7511_CEC_CTRL_POWER_DOWN);

This would force CEC to be powered off even if adv7511_cec_init() returned 0, right?
We wouldn't want that if we want to use CEC on a platform that supports it.

Do we know why the call to adv7511_cec_init() is failing on the Hikey board? If it's
because there isn't a "cec" clock specified in DT, it's not really a fatal error, it
just means that the platform hasn't been set up to support CEC. In that case, we
should just power down the CEC block. So, if adv7511_cec_init() would return a
-ENOENT, which we could use as a hint to power down CEC. So, maybe something like this?:

         ret = adv7511_cec_init(dev, adv7511, offset);
         if (ret && ret != -ENOENT)	
		goto err_unregister_cec;
	if (ret)
             regmap_write(adv7511->regmap, ADV7511_REG_CEC_CTRL + offset,

Apart from this, we should also move adv7511_cec_init() up in the probe so that
it's called before the drm_bridge is registered.


>   	return 0;

Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project

Powered by blists - more mailing lists