Re: [PATCH 4/5] scsi: libsas: Drop SAS_DPRINTK() and revise logs levels

[John Garry]

On 14/11/2018 14:53, Joe Perches wrote:
> On Wed, 2018-11-14 at 21:47 +0800, John Garry wrote:
> > Like sas_printk() did previously, SAS_DPRINTK() offers little value now
> > that libsas logs already have the "sas" prefix through pr_fmt(fmt). So it
> > can be dropped.
> >
> > However, after reviewing some logs in libsas, it is noticed that debug
> > level is too low in many instances.
> >
> > So this change drops SAS_DPRINTK() and revises some logs to a more
> > appropriate level. However many stay at debug level, although some
> > are significantly promoted.
> []
> > All the pre-existing checkpatch errors for spanning messages across
> > multiple lines are untouched.
>
> I think coalescing would be useful.

Sorry, I missed that. Do you mean that we stop spanning strings over 
multiple lines?

If yes, I tend to agree. It means we can grep for full strings vs just a 
different checkpatch issue (>80 lines or spanning multiple lines)

> Where there are already embedded "sas: "
> prefixes, those should be removed too.

Ah, I missed that guy. File sas_ata.c may also have some.

> You could verify this by using
>
> 	$ strings drivers/scsi/libsas/*.o | grep "^[0-7]"
>
> > @@ -186,10 +186,10 @@ int sas_notify_lldd_dev_found(struct domain_device *dev)
> >
> >  	res = i->dft->lldd_dev_found(dev);
> >  	if (res) {
> > -		printk("sas: driver on pcidev %s cannot handle "
> > -		       "device %llx, error:%d\n",
> > -		       dev_name(sas_ha->dev),
> > -		       SAS_ADDR(dev->sas_addr), res);
> > +		pr_warn("sas: driver on pcidev %s cannot handle "
> > +			"device %llx, error:%d\n",
> > +			dev_name(sas_ha->dev),
> > +			SAS_ADDR(dev->sas_addr), res);
>
> e.g.: this now emits "sas: sas: driver etc..."

Cheers,
John

> .