[EXT] Re: [PATCH 3/6] perf/marvell : Odyssey LLC-TAD performance monitor support

[Gowthami Thiagarajan]

Hi Mark/Will,

> -----Original Message-----
> From: Mark Rutland <mark.rutland at arm.com>
> Sent: Tuesday, August 15, 2023 6:54 PM
> To: Gowthami Thiagarajan <gthiagarajan at marvell.com>
> Cc: will at kernel.org; linux-arm-kernel at lists.infradead.org; linux-kernel at vger.kernel.org; Sunil Kovvuri
> Goutham <sgoutham at marvell.com>; Bharat Bhushan <bbhushan2 at marvell.com>; George Cherian
> <gcherian at marvell.com>; Linu Cherian <lcherian at marvell.com>
> Subject: [EXT] Re: [PATCH 3/6] perf/marvell : Odyssey LLC-TAD performance monitor support
> 
> External Email
> 
> ----------------------------------------------------------------------
> On Sat, Aug 12, 2023 at 01:51:00PM +0000, Gowthami Thiagarajan wrote:
> > > -----Original Message-----
> > > From: Mark Rutland <mark.rutland at arm.com>
> > > Sent: Friday, July 28, 2023 9:08 PM
> > > To: Gowthami Thiagarajan <gthiagarajan at marvell.com>
> > > Cc: will at kernel.org; linux-arm-kernel at lists.infradead.org; linux-kernel at vger.kernel.org; Sunil
> Kovvuri
> > > Goutham <sgoutham at marvell.com>; Bharat Bhushan <bbhushan2 at marvell.com>; George Cherian
> > > <gcherian at marvell.com>; Linu Cherian <lcherian at marvell.com>
> > > Subject: [EXT] Re: [PATCH 3/6] perf/marvell : Odyssey LLC-TAD performance monitor support
> > >
> > > External Email
> > >
> > > ----------------------------------------------------------------------
> > > On Fri, Jun 30, 2023 at 05:33:48PM +0530, Gowthami Thiagarajan wrote:
> 
> > > > +static int tad_pmu_event_counter_add(struct perf_event *event, int flags)
> > > > +{
> 
> > > > +	if (!event->attr.disabled)
> > > > +		return -EINVAL;
> > >
> > > Why?
> > Just checks the default disabled attribute.
> 
> Why does it matter?
> 
> What's the problem with opening an event which is *not* disabled?
> 
> [...]
Checked this. Don't find the significance of the validation. Can be removed.
> 
> > > > +static int tad_pmu_probe(struct platform_device *pdev)
> > > > +{
> > > > +	struct device *dev = &pdev->dev;
> > > > +	struct tad_region *regions;
> > > > +	struct tad_pmu *tad_pmu;
> > > > +	struct resource *res;
> > > > +	u32 tad_pmu_page_size;
> > > > +	u32 tad_page_size;
> > > > +	u32 tad_cnt;
> > > > +	int i, ret;
> > > > +	char *name;
> > > > +
> > > > +	tad_pmu = devm_kzalloc(&pdev->dev, sizeof(*tad_pmu), GFP_KERNEL);
> > > > +	if (!tad_pmu)
> > > > +		return -ENOMEM;
> > > > +
> > > > +	platform_set_drvdata(pdev, tad_pmu);
> > > > +
> > > > +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > > > +	if (!res) {
> > > > +		dev_err(&pdev->dev, "Mem resource not found\n");
> > > > +		return -ENODEV;
> > > > +	}
> > > > +
> > > > +	ret = device_property_read_u32(dev, "marvell,tad-page-size", &tad_page_size);
> > > > +	if (ret) {
> > > > +		dev_err(&pdev->dev, "Can't find tad-page-size property\n");
> > > > +		return ret;
> > > > +	}
> > > > +
> > > > +	ret = device_property_read_u32(dev, "marvell,tad-pmu-page-size",
> > > > +				       &tad_pmu_page_size);
> > > > +	if (ret) {
> > > > +		dev_err(&pdev->dev, "Can't find tad-pmu-page-size property\n");
> > > > +		return ret;
> > > > +	}
> > >
> > > Why do you think these properties are necessary?
> > >
> > > These should almost certainly be provided by IO resources, and shouldn't need a
> > > custom property.
> >
> > IO resources don't provide all the information in this case. Need to vary these values
> > at different boot times. So, kept these custom properties.
> 
> I think the only rason the information is missing is that your DT binding isn't quite right.
> 
> Later on you do:
> 
> +       for (i = 0; i < tad_cnt && res->start < res->end; i++) {
> +               regions[i].base = devm_ioremap(&pdev->dev,
> +                                              res->start,
> +                                              tad_pmu_page_size);
> +               if (IS_ERR(regions[i].base)) {
> +                       dev_err(&pdev->dev, "TAD%d ioremap fail\n", i);
> +                       return -ENOMEM;
> +               }
> +               res->start += tad_page_size;
> +       }
> 
> ... which means you're splitting one reg entry into multiple mappings, whereas
> you could have multiple reg entries, one per TAD page.

These properties are updated by the firmware at runtime. In order to avoid the firmware updating multiple nodes, a single node for the entire system has been kept. So that the firmware can just update at a single place  .

Thanks,
Gowthami

> 
> Thanks,
> Mark.