Re: [PATCH v10 01/11] drm/etnaviv: Add a dedicated function to register an irq handler

[Sui Jingfeng]

Hi,

On 2023/6/21 18:16, Lucas Stach wrote:
> Am Mittwoch, dem 21.06.2023 um 17:20 +0800 schrieb Sui Jingfeng:
> > Hi,
> >
> > On 2023/6/21 17:07, Lucas Stach wrote:
> > > Am Dienstag, dem 20.06.2023 um 17:47 +0800 schrieb Sui Jingfeng:
> > > > From: Sui Jingfeng <suijingfeng@xxxxxxxxxxx>
> > > >
> > > > Because getting IRQ from a device is platform-dependent, PCI devices have
> > > > different methods for getting an IRQ. This patch is a preparation to extend
> > > > this driver for supporting the PCI devices.
> > > >
> > > > Cc: Lucas Stach <l.stach@xxxxxxxxxxxxxx>
> > > > Cc: Christian Gmeiner <christian.gmeiner@xxxxxxxxx>
> > > > Cc: Philipp Zabel <p.zabel@xxxxxxxxxxxxxx>
> > > > Cc: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
> > > > Cc: Daniel Vetter <daniel@xxxxxxxx>
> > > > Signed-off-by: Sui Jingfeng <suijingfeng@xxxxxxxxxxx>
> > > > ---
> > > >    drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 32 +++++++++++++++++++--------
> > > >    1 file changed, 23 insertions(+), 9 deletions(-)
> > > >
> > > > diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
> > > > index de8c9894967c..a03e81337d8f 100644
> > > > --- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
> > > > +++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
> > > > @@ -1817,6 +1817,27 @@ static const struct of_device_id etnaviv_gpu_match[] = {
> > > >    };
> > > >    MODULE_DEVICE_TABLE(of, etnaviv_gpu_match);
> > > >    
> > > > +static int etnaviv_gpu_register_irq(struct etnaviv_gpu *gpu, int irq)
> > > > +{
> > > > +	struct device *dev = gpu->dev;
> > > > +	int err;
> > > > +
> > > > +	if (irq < 0)
> > > > +		return irq;
> > > > +
> > > > +	err = devm_request_irq(dev, irq, irq_handler, 0, dev_name(dev), gpu);
> > > > +	if (err) {
> > > > +		dev_err(dev, "failed to request irq %u: %d\n", irq, err);
> > > > +		return err;
> > > > +	}
> > > > +
> > > > +	gpu->irq = irq;
> > > > +
> > > > +	dev_info(dev, "irq(%d) handler registered\n", irq);
> > > There is no reason to put this into the kernel log.
> > I want to see the IRQ of the device when debugging,
> >
> > etnaviv actually print very less.
> >
> > This serve as a minimal signal  to us the etnaviv_gpu_register_irq()
> > function is successful at driver load time.
> And debugging is a very different use-case than normal operation. If
> it's needed at all, downgrade this to dev_dbg. This isn't interesting
> information for a ordinary user of a system.
>
> > >    It's no different
> > > than other resources to the driver and we don't log each one of those
> > > either.
> > >
> > > In fact I don't see any reason for this change in the first place.
> > > Effectively you are moving a single function call into a new function,
> > > which doesn't seem like an improvement.
> > This is to make the patch easy to review, each patch is only introduce a
> > small function,
> What I'm saying is that I don't see the need to introduce this function
> at all. All you need to do is move platform_get_irq out into the
> platform device code path. The devm_request_irq can stay where it is,
> as the only difference between platform and PCI device is how the irq
> number is retrieved from the platform.

Yes, you are right. I understand what are asking, but my point is:


This patch is paving the way for us to introduce the PCI device driver.

All of the patches before the patch 
v10-0006-drm-etnaviv-Add-driver-support-for-the-PCI-devic.patch

are actually doing the preparation.


Look at the patch 0006, I achieve the goal by 128 insertions and 7 
deletions.

while the only 7 deletions are actually for code shading(convert the 
static function to global function).

There is No large area diff and NO distortion.

The goal is adding a PCI device driver on the top of what we already have.


Before the cleanup, the etnaviv_gpu_platform_probe() function is just 
like is a *glue*.

Originally,  it integrate a lot of irrelevant part together.


1.  Mapping MMIO registers make it platform-dependent;

2.  Calling platform_get_irq(pdev, 0) make it platform-dependent;

3.  Getting Clocks by calling devm_clk_get() make it platform-dependent;

4.  Calling component_add() make it subsytem and framework-dependent;


All of above list item is deny us to introduce the PCI device driver 
wrapper.

It(sub-functional code) is not relevant to each other.

Hence the first five patch is actually do the clean,

for the clarify(and tidy and good looking) of the patch 6.


I will drop the printing, but keep the cleanup function there,

Is this acceptable?


> Regards,
> Lucas
>
> > which is paving the way for we introducing the PCI device driver.
> >
> > Otherwise when we introducing the PCI device driver, the patch is looks
> > ugly,
> >
> > It is difficult to review.
> >
> > > Regards,
> > > Lucas
> > >
> > > > +
> > > > +	return 0;
> > > > +}
> > > > +
> > > >    static int etnaviv_gpu_platform_probe(struct platform_device *pdev)
> > > >    {
> > > >    	struct device *dev = &pdev->dev;
> > > > @@ -1837,16 +1858,9 @@ static int etnaviv_gpu_platform_probe(struct platform_device *pdev)
> > > >    		return PTR_ERR(gpu->mmio);
> > > >    
> > > >    	/* Get Interrupt: */
> > > > -	gpu->irq = platform_get_irq(pdev, 0);
> > > > -	if (gpu->irq < 0)
> > > > -		return gpu->irq;
> > > > -
> > > > -	err = devm_request_irq(&pdev->dev, gpu->irq, irq_handler, 0,
> > > > -			       dev_name(gpu->dev), gpu);
> > > > -	if (err) {
> > > > -		dev_err(dev, "failed to request IRQ%u: %d\n", gpu->irq, err);
> > > > +	err = etnaviv_gpu_register_irq(gpu, platform_get_irq(pdev, 0));
> > > > +	if (err)
> > > >    		return err;
> > > > -	}
> > > >    
> > > >    	/* Get Clocks: */
> > > >    	gpu->clk_reg = devm_clk_get_optional(&pdev->dev, "reg");

--
Jingfeng