On Thu, 2018-06-28 at 07:19 +0200, Johan Hovold wrote:
> On Thu, Jun 28, 2018 at 11:06:13AM +0800, Sean Wang wrote:
> > On Wed, 2018-06-27 at 20:04 +0300, Andy Shevchenko wrote:
> > > On Wed, Jun 27, 2018 at 7:59 PM, Andy Shevchenko
> > > <andy.shevche...@gmail.com> wrote:
> > > > On Wed, Jun 27, 2018 at 8:43 AM, <sean.w...@mediatek.com> wrote:
> > > >> From: Sean Wang <sean.w...@mediatek.com>
>
> > > >> +#include <net/bluetooth/bluetooth.h>
> > > >> +#include <net/bluetooth/hci_core.h>
> > >
> > > >> + /* Enable the power domain and clock the device requires. */
> > > >> + pm_runtime_enable(dev);
> > > >> + err = pm_runtime_get_sync(dev);
> > > >> + if (err < 0)
> > > >> + goto err_pm2;
>
> > > >> +err_pm1:
> > > >> + pm_runtime_put_sync(dev);
> > > >> +err_pm2:
> > > >> + pm_runtime_disable(dev);
>
> Please name error labels after what they do, not using numbers (see
> CodingStyle). Here you could use err_disable_rpm instead of err_pm2, for
> example.
>
> Also, if you really want to undo pm_runtime_get_sync() failing above,
> you still need a pm_runtime_put_noidle() to balance the usage count.
>
I'll give a reasonable naming for these labels and add a
pm_runtime_put_noidle() in the path undoing failing
pm_runtime_get_sync().
> > > >> +struct mtk_stp_hdr {
> > > >> + __u8 prefix;
> > > >> + __u8 dlen1:4;
> > > >> + __u8 type:4;
> > >
> > > >> + __u8 dlen2:8;
> > > >> + __u8 cs;
> > > >> +} __packed;
>
> Perhaps too much context have been lost here, but unless you're sharing
> this struct with user space, you should be using u8 (without __) above
> (and elsewhere).
>
These struct don't be shard with user space so I will turn __u8 into u8.
Thanks so much for all suggestions.
> Johan
