Re: [PATCH] coresight: etm4x: return 0 instead of using local ret variable

[Shile Zhang]

On 2022/4/22 00:42, Mathieu Poirier wrote:
> Hi Shile,
>
> On Wed, Apr 20, 2022 at 01:28:31PM +0800, Shile Zhang wrote:
> > The etm4_remove function (now it's rename to etm4_remove_dev) always
> > return 0, and it has been changed to void in commit 4fd269e74f2f
> > ("amba: Make the remove callback return void"). But its weird that the
> > changes is gone in mainline. which is remained in 5.10.y branch.
>
> Commit 4fd269e74f2f is not valid upstream.  Changes that don't have a critical
> impact on user experience or fix a bug aren't backported to longterm kernels.
Sorry, I used wrong commit id, but this patch is merged in mainline with 
3fd269e74f2fe ("amba: Make the remove callback return void")

It can be checked from here: 
https://github.com/torvalds/linux/commit/3fd269e74f2fe#diff-937c525e02a9fea8fe99b3d91b9f87ca097b3392de5fea3d4ff56cec8b08c94aL1683

> > Just backport the changes of etm4_remove_dev and return 0 directly in it's
> > caller function etm4_remove_platform_dev.
>
> I'm not sure why the work "backport" is used here since this patch is destine
> for mainline.
Yes, maybe "backport" here is not OK, I'll re-work it in next version 
later. Thanks!

> > Signed-off-by: Shile Zhang <shile.zhang@xxxxxxxxxxxxxxxxx>
> > ---
> >   drivers/hwtracing/coresight/coresight-etm4x-core.c | 11 +++++------
> >   1 file changed, 5 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c
> > index 7f416a12000e..141f8209a152 100644
> > --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c
> > @@ -2104,7 +2104,7 @@ static void clear_etmdrvdata(void *info)
> >   	etmdrvdata[cpu] = NULL;
> >   }
> >   
> > -static int __exit etm4_remove_dev(struct etmv4_drvdata *drvdata)
> > +static void __exit etm4_remove_dev(struct etmv4_drvdata *drvdata)
> >   {
> >   	etm_perf_symlink(drvdata->csdev, false);
> >   	/*
> > @@ -2125,8 +2125,6 @@ static int __exit etm4_remove_dev(struct etmv4_drvdata *drvdata)
> >   
> >   	cscfg_unregister_csdev(drvdata->csdev);
> >   	coresight_unregister(drvdata->csdev);
> > -
> > -	return 0;
> >   }
> >   
> >   static void __exit etm4_remove_amba(struct amba_device *adev)
> > @@ -2139,13 +2137,14 @@ static void __exit etm4_remove_amba(struct amba_device *adev)
> >   
> >   static int __exit etm4_remove_platform_dev(struct platform_device *pdev)
> >   {
> > -	int ret = 0;
> >   	struct etmv4_drvdata *drvdata = dev_get_drvdata(&pdev->dev);
> >   
> >   	if (drvdata)
> > -		ret = etm4_remove_dev(drvdata);
> > +		etm4_remove_dev(drvdata);
> > +
> >   	pm_runtime_disable(&pdev->dev);
> > -	return ret;
> > +
> > +	return 0;
>
> I'm fine with the code but the changelog needs to be re-worked.  The only
> rational for this patch is that the return value for function etm4_remove_dev()
> is never used and as such being removed.
Thanks for your comments, I'll send V2 later.

> Thanks,
> Mathieu
>
> >   }
> >   
> >   static const struct amba_id etm4_ids[] = {
> > --
> > 2.33.0.rc2