Re: [PATCH] uacce: add print information if not enable sva

[Greg KH <gregkh@linuxfoundation.org>]

On Tue, Jun 08, 2021 at 09:39:52AM +0800, yekai(A) wrote:
> 
> 
> On 2021/6/4 15:51, Greg KH wrote:
> > On Fri, Jun 04, 2021 at 03:46:09PM +0800, Kai Ye wrote:
> > > Add print information necessary if user not enable sva.
> > > 
> > > Signed-off-by: Kai Ye <yekai13@huawei.com>
> > > ---
> > >  drivers/misc/uacce/uacce.c | 11 +++++++++--
> > >  1 file changed, 9 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/misc/uacce/uacce.c b/drivers/misc/uacce/uacce.c
> > > index bae18ef0..fe38af8 100644
> > > --- a/drivers/misc/uacce/uacce.c
> > > +++ b/drivers/misc/uacce/uacce.c
> > > @@ -387,15 +387,22 @@ static void uacce_release(struct device *dev)
> > > 
> > >  static unsigned int uacce_enable_sva(struct device *parent, unsigned int flags)
> > >  {
> > > +	int ret;
> > > +
> > >  	if (!(flags & UACCE_DEV_SVA))
> > >  		return flags;
> > > 
> > >  	flags &= ~UACCE_DEV_SVA;
> > > 
> > > -	if (iommu_dev_enable_feature(parent, IOMMU_DEV_FEAT_IOPF))
> > > +	ret = iommu_dev_enable_feature(parent, IOMMU_DEV_FEAT_IOPF);
> > > +	if (ret) {
> > > +		dev_err(parent, "failed to enable IOPF feature! ret = %d\n", ret);
> > 
> > Why is this needed?  Has this ever happened in real life such that the
> > log message is now required?
> > 
> > 
> > 
> > >  		return flags;
> > > +	}
> > > 
> > > -	if (iommu_dev_enable_feature(parent, IOMMU_DEV_FEAT_SVA)) {
> > > +	ret = iommu_dev_enable_feature(parent, IOMMU_DEV_FEAT_SVA);
> > > +	if (ret) {
> > > +		dev_err(parent, "failed to enable SVA feature! ret = %d\n", ret);
> > 
> > Same here, does this happen in real systems?
> > 
> > thanks,
> > 
> > greg k-h
> > .
> > 
> In a other debug version, the SVA feature failed to be enabled, and no
> related information was printed. we don't know the cause of the problem.
> Finally, Although, it's not the problem here. but we find that the sva
> enable function doesn't have debug information. Therefore, it is
> inconvenient to locate the fault.
> so i think the log message is required.

Ok, can you resend with the changes that Joe suggested to make this even
easier to use?

thanks,

greg k-h