On Thu, Sep 08, 2011 at 07:24:29AM -0700, K. Y. Srinivasan wrote:
> Now, eliminate the usage of ext field in struct hv_device for storvsc driver.
> We do this by registering pointer to struct storvsc_device as the driver
> specific data and eliminating the current usage of driver specific data to
> save and retrieve the pointer to struct Scsi_Host.
> Additionally, all access to the driver specific data is through
> the vmbus wrapper functions.
>
> Signed-off-by: K. Y. Srinivasan <kys@xxxxxxxxxxxxx>
> Signed-off-by: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>
> ---
> drivers/staging/hv/storvsc_drv.c | 24 +++++++++++++++---------
> 1 files changed, 15 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/staging/hv/storvsc_drv.c b/drivers/staging/hv/storvsc_drv.c
> index 57c1035..98d47cd 100644
> --- a/drivers/staging/hv/storvsc_drv.c
> +++ b/drivers/staging/hv/storvsc_drv.c
> @@ -266,6 +266,7 @@ struct storvsc_device {
> bool destroy;
> bool drain_notify;
> atomic_t num_outstanding_req;
> + struct Scsi_Host *host;
You are properly reference counting this pointer, right?
> @@ -306,7 +307,8 @@ static inline struct storvsc_device *get_out_stor_device(
> {
> struct storvsc_device *stor_device;
>
> - stor_device = (struct storvsc_device *)device->ext;
> + stor_device =
> + (struct storvsc_device *)hv_get_drvdata(device);
Casting is not needed at all.
> @@ -1366,7 +1371,8 @@ static int storvsc_probe(struct hv_device *device,
> stor_device->destroy = false;
> init_waitqueue_head(&stor_device->waiting_to_drain);
> stor_device->device = device;
> - device->ext = stor_device;
> + stor_device->host = host;
> + hv_set_drvdata(device, stor_device);
Lookie there, no reference count incremented, right? Or did I miss it
somewhere else?
gotta love dynamic pointer lifetime rules...
greg k-h
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
