Re: [PATCH] staging: media: atomisp: Fix error path in lm3554_probe()

[Dan Carpenter <dan.carpenter@oracle.com>]

On Thu, Sep 03, 2020 at 04:48:41PM +0100, Alex Dewar wrote:
> Good point about the timer!
> 
> > >  
> > > -	err = lm3554_gpio_init(client);
> > > -	if (err) {
> > > +	ret = lm3554_gpio_init(client);
> > > +	if (ret) {
> > >  		dev_err(&client->dev, "gpio request/direction_output fail");
> > > -		goto fail2;
> > > +		goto err_cleanup_entity;
> > >  	}
> > >  	return atomisp_register_i2c_module(&flash->sd, NULL, LED_FLASH);
> > 
> > If atomisp_register_i2c_module() fails then we need to call
> > lm3554_gpio_uninit(client) and do other cleanup.
> 
> I'm probably showing my ignorance here, but I can't see what cleanup we
> need. Inside lm3554_gpio_init we have:
> 
> 	ret = gpiod_direction_output(pdata->gpio_reset, 0);
> 	if (ret < 0)
> 		return ret;
> 	dev_info(&client->dev, "flash led reset successfully\n");
> 
> 	if (!pdata->gpio_strobe)
> 		return -EINVAL;
> 
> 	ret = gpiod_direction_output(pdata->gpio_strobe, 0);
> 	if (ret < 0)
> 		return ret;
> 
> I'm not sure how you "undo" a call to gpiod_direction_output, but I'm
> thinking you won't need to do anything because it should be ok to leave
> a gpio to output 0... right?

You're right.  I wonder if there is really any need for the
lm3554_gpio_uninit() function at all?  It's basically the same as
lm3554_gpio_init() except for the order of function calls.  Probably
we could just rename lm3554_gpio_init() to something like
lm3554_gpio_set_default() and use it in both the probe() and remove
functions()...

But I don't know the code and can't test it so let's leave that for
another day.

We still do need to clean up if atomisp_register_i2c_module() fails
though, and the timer as well so could you resend a v2?

regards,
dan carpenter