Re: [PATCH] spi: Fix double-free on spi_unregister_master

[David Brownell <david-b@pacbell.net>]

On Thursday 08 November 2007, Atsushi Nemoto wrote:
> After 49dce689ad4ef0fd1f970ef762168e4bd46f69a3, device_for_each_child
> iteration hits the master device itself.  Do not call
> spi_unregister_device() for the master device.
> 
> Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>

ACK, and thanks.  Regression against 2.6.23...

I should have searched for other device_for_each_child() calls
when I had to do the same thing to fix the mmc_spi breakage.

Looks like the pxa2xx_spi suspend() code probably broke too.
I CC'd the maintainer of that code, though I'm not sure how
much time he has for this lately.

I'll hope someone else checks for non-SPI driver stacks which
such "class_dev" elimination may have broke...


> ---
>  drivers/spi/spi.c |    8 +++++---
>  1 files changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
> index 89769ce..b31f443 100644
> --- a/drivers/spi/spi.c
> +++ b/drivers/spi/spi.c
> @@ -457,10 +457,11 @@ done:
>  EXPORT_SYMBOL_GPL(spi_register_master);
>  
>  
> -static int __unregister(struct device *dev, void *unused)
> +static int __unregister(struct device *dev, void *master_dev)
>  {
>  	/* note: before about 2.6.14-rc1 this would corrupt memory: */
> -	spi_unregister_device(to_spi_device(dev));
> +	if (dev != master_dev)
> +		spi_unregister_device(to_spi_device(dev));
>  	return 0;
>  }
>  
> @@ -478,7 +479,8 @@ void spi_unregister_master(struct spi_master *master)
>  {
>  	int dummy;
>  
> -	dummy = device_for_each_child(master->dev.parent, NULL, __unregister);
> +	dummy = device_for_each_child(master->dev.parent, &master->dev,
> +					__unregister);
>  	device_unregister(&master->dev);
>  }
>  EXPORT_SYMBOL_GPL(spi_unregister_master);
>