[PATCH 12/59] Staging: hv: vmbus: Cleanup vmbus_uevent() code

Greg KH greg at kroah.com
Thu Aug 25 20:59:30 UTC 2011


On Thu, Aug 25, 2011 at 09:48:38AM -0700, K. Y. Srinivasan wrote:
> Now generate appropriate uevent based on the modalias string. As part of this,
> cleanup the existing uevent code.
> 
> Signed-off-by: K. Y. Srinivasan <kys at microsoft.com>
> Signed-off-by: Haiyang Zhang <haiyangz at microsoft.com>
> ---
>  drivers/staging/hv/vmbus_drv.c |   60 ++++++++--------------------------------
>  1 files changed, 12 insertions(+), 48 deletions(-)
> 
> diff --git a/drivers/staging/hv/vmbus_drv.c b/drivers/staging/hv/vmbus_drv.c
> index b651968..a6e7dc5 100644
> --- a/drivers/staging/hv/vmbus_drv.c
> +++ b/drivers/staging/hv/vmbus_drv.c
> @@ -237,58 +237,22 @@ static struct device_attribute vmbus_device_attrs[] = {
>   * This routine is invoked when a device is added or removed on the vmbus to
>   * generate a uevent to udev in the userspace. The udev will then look at its
>   * rule and the uevent generated here to load the appropriate driver
> + *
> + * The alias string will be of the form vmbus:guid where guid is the string
> + * representation of the device guid (each byte of the guid will be
> + * represented with two hex characters.
>   */
>  static int vmbus_uevent(struct device *device, struct kobj_uevent_env *env)
>  {
>  	struct hv_device *dev = device_to_hv_device(device);
> -	int ret;
> -
> -	ret = add_uevent_var(env, "VMBUS_DEVICE_CLASS_GUID={"
> -			     "%02x%02x%02x%02x-%02x%02x-%02x%02x-"
> -			     "%02x%02x%02x%02x%02x%02x%02x%02x}",
> -			     dev->dev_type.b[3],
> -			     dev->dev_type.b[2],
> -			     dev->dev_type.b[1],
> -			     dev->dev_type.b[0],
> -			     dev->dev_type.b[5],
> -			     dev->dev_type.b[4],
> -			     dev->dev_type.b[7],
> -			     dev->dev_type.b[6],
> -			     dev->dev_type.b[8],
> -			     dev->dev_type.b[9],
> -			     dev->dev_type.b[10],
> -			     dev->dev_type.b[11],
> -			     dev->dev_type.b[12],
> -			     dev->dev_type.b[13],
> -			     dev->dev_type.b[14],
> -			     dev->dev_type.b[15]);
> -
> -	if (ret)
> -		return ret;
> +	int i, ret;
> +	char alias_name[((sizeof(struct hv_vmbus_device_id) + 1)) * 2];
>  
> -	ret = add_uevent_var(env, "VMBUS_DEVICE_DEVICE_GUID={"
> -			     "%02x%02x%02x%02x-%02x%02x-%02x%02x-"
> -			     "%02x%02x%02x%02x%02x%02x%02x%02x}",
> -			     dev->dev_instance.b[3],
> -			     dev->dev_instance.b[2],
> -			     dev->dev_instance.b[1],
> -			     dev->dev_instance.b[0],
> -			     dev->dev_instance.b[5],
> -			     dev->dev_instance.b[4],
> -			     dev->dev_instance.b[7],
> -			     dev->dev_instance.b[6],
> -			     dev->dev_instance.b[8],
> -			     dev->dev_instance.b[9],
> -			     dev->dev_instance.b[10],
> -			     dev->dev_instance.b[11],
> -			     dev->dev_instance.b[12],
> -			     dev->dev_instance.b[13],
> -			     dev->dev_instance.b[14],
> -			     dev->dev_instance.b[15]);
> -	if (ret)
> -		return ret;
> +	for (i = 0; i < (sizeof(struct hv_vmbus_device_id) * 2); i += 2)
> +		sprintf(&alias_name[i], "%02x", dev->dev_type.b[i/2]);

I have to edit this to get it to work properly with the fact that I
added the driver_data field to hv_vmbus_device_id.

Arguably, one could say that this patch was always broken as you were
assuming the size of an individual field was the same size as the whole
structure, which I don't think is always the case, or at least it's not
a safe thing to assume :)

thanks,

greg k-h



More information about the devel mailing list