Hi, Auger
> On 10/30/20 4:47 PM, Zhang Qilong wrote:
> > pm_runtime_get_sync() will increment pm usage counter even it failed.
> > Forgetting to call pm_runtime_put_noidle will result in reference leak
> > in vfio_platform_open, so we should fix it.
> >
> > Signed-off-by: Zhang Qilong <zhangqilong3@xxxxxxxxxx>
> > ---
> > drivers/vfio/platform/vfio_platform_common.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/vfio/platform/vfio_platform_common.c
> > b/drivers/vfio/platform/vfio_platform_common.c
> > index c0771a9567fb..aa97f1678981 100644
> > --- a/drivers/vfio/platform/vfio_platform_common.c
> > +++ b/drivers/vfio/platform/vfio_platform_common.c
> > @@ -266,8 +266,10 @@ static int vfio_platform_open(void *device_data)
> > goto err_irq;
> >
> > ret = pm_runtime_get_sync(vdev->device);
> > - if (ret < 0)
> > + if (ret < 0) {
> > + pm_runtime_put_noidle(vdev->device);
> can't we jump to err_rst then?
We could jump to err_rst here. The difference is that pm_runtime_put will leave this
device idle state if we decrease usage count failed that is meanless. Because this
module will be put. Generally speaking, the action of pm_runtime_put and
pm_runtime_put_noidle is the same here that only decreases usage count. Jumping to
err_rst could be more better for read view. I'll improve it in patch V2.
Thanks
Zhang Qilong
>
> Thanks
>
> Eric
> > goto err_pm;
> > + }
> >
> > ret = vfio_platform_call_reset(vdev, &extra_dbg);
> > if (ret && vdev->reset_required) {
> >
