> On Fri, 12 Oct 2018, Christoph Hellwig wrote:
>
> >
> > > diff --git a/drivers/scsi/mac_esp.c b/drivers/scsi/mac_esp.c
> > > index eb551f3cc471..85d067889a9b 100644
> > > --- a/drivers/scsi/mac_esp.c
> > > +++ b/drivers/scsi/mac_esp.c
> > > @@ -58,8 +58,7 @@ static struct esp *esp_chips[2];
> > > static DEFINE_SPINLOCK(esp_chips_lock);
> > >
> > > #define MAC_ESP_GET_PRIV(esp) ((struct mac_esp_priv *) \
> > > - platform_get_drvdata((struct platform_device
> > > *) \
> > > - (esp->dev)))
> > > + dev_get_drvdata((esp)->dev))
> > >
> > > static inline void mac_esp_write8(struct esp *esp, u8 val, unsigned long
> > > reg)
> > > {
> > > @@ -508,7 +507,7 @@ static int esp_mac_probe(struct platform_device *dev)
> > > esp = shost_priv(host);
> > >
> > > esp->host = host;
> > > - esp->dev = dev;
> > > + esp->dev = &dev->dev;
> > >
> > > esp->command_block = kzalloc(16, GFP_KERNEL);
> > > if (!esp->command_block)
> >
> >
> > Isn't this missing the corresponding dev_set_drvdata() in esp_mac_probe()?
> > Also conversion to dev_get_drvdata() in esp_mac_remove()?
> >
>
> I don't think it matters at all, as nothing in the core esp_scsi.c code
> looks into the driver data to start with, but even if it did,
> platform_set_drvdata and platform_get_drvdata and just trivial wrappers
> around dev_set_drvdata and dev_get_drvdata and we rely on that fact all
> over the kernel.
>
Sorry, I got confused by the mix of dev_get_drvdata() and
platform_get_drvdata() but I see now that they all refer to the same
struct device.
> That being said if you want them converted I can throw in another patch
> at the end of the series.
>
No need. Thanks.
--
