Re: [v2] PCI: imx: make msi work without pcieportbus

[Baruch Siach]

Hi Richard,

Richard Zhu writes:
>> -----Original Message-----
>> From: Baruch Siach [mailto:baruch@xxxxxxxxxx]
>> Sent: 2018å12æ13æ 16:13
>> To: Richard Zhu <hongxing.zhu@xxxxxxx>
>> Cc: bhelgaas@xxxxxxxxxx; lorenzo.pieralisi@xxxxxxx;
>> l.stach@xxxxxxxxxxxxxx; andrew.smirnov@xxxxxxxxx;
>> linux-pci@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
>> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
>> Subject: Re: [v2] PCI: imx: make msi work without pcieportbus
>>
>> Richard Zhu writes:
>> > MSI_EN of iMX PCIe RC would be asserted when PCIEPORTBUS driver is
>> > selected.
>> > Thus, the MSI works fine on iMX PCIe before.
>> > Assert it unconditionally when MSI is supported.
>> > Otherwise, the MSI wouldn't be triggered although the EP is present
>> > and the MSIs are assigned.
>> >
>> > Signed-off-by: Richard Zhu <hongxing.zhu@xxxxxxx>
>> > ---
>> >  drivers/pci/controller/dwc/pci-imx6.c | 7 +++++++
>> >  1 file changed, 7 insertions(+)
>> >
>> > diff --git a/drivers/pci/controller/dwc/pci-imx6.c
>> > b/drivers/pci/controller/dwc/pci-imx6.c
>> > index 26087b3..d3e4296 100644
>> > --- a/drivers/pci/controller/dwc/pci-imx6.c
>> > +++ b/drivers/pci/controller/dwc/pci-imx6.c
>> > @@ -74,6 +74,7 @@ struct imx6_pcie {
>> >  #define PHY_PLL_LOCK_WAIT_USLEEP_MAX	200
>> >
>> >  /* PCIe Root Complex registers (memory-mapped) */
>> > +#define PCI_MSI_CAP				0x50
>> >  #define PCIE_RC_LCR				0x7c
>> >  #define PCIE_RC_LCR_MAX_LINK_SPEEDS_GEN1	0x1
>> >  #define PCIE_RC_LCR_MAX_LINK_SPEEDS_GEN2	0x2
>> > @@ -926,6 +927,7 @@ static int imx6_pcie_probe(struct platform_device
>> *pdev)
>> >  	struct resource *dbi_base;
>> >  	struct device_node *node = dev->of_node;
>> >  	int ret;
>> > +	u16 val;
>> >
>> >  	imx6_pcie = devm_kzalloc(dev, sizeof(*imx6_pcie), GFP_KERNEL);
>> >  	if (!imx6_pcie)
>> > @@ -1070,6 +1072,11 @@ static int imx6_pcie_probe(struct
>> platform_device *pdev)
>> >  	ret = imx6_add_pcie_port(imx6_pcie, pdev);
>> >  	if (ret < 0)
>> >  		return ret;
>> > +	if (IS_ENABLED(CONFIG_PCI_MSI)) {
>>
>> CONFIG_PCI_IMX6 depends on CONFIG_PCI_MSI_IRQ_DOMAIN that in turn
>> depends on CONFIG_PCI_MSI. So this condition should always be true. If so,
>> you can drop this one as well.
> [Richard Zhu] Thanks.
> Can we have the double check here?
> Thus, it's aligned to the MSI enable codes in the other place of the driver.
> For example:
> ...
> 	        if (IS_ENABLED(CONFIG_PCI_MSI))
>                 dw_pcie_msi_init(pp);
> ...
> Secondly, the driver has a better portability with the double check here.
> How do you think about it?

Kernel driver are not meant to be "portable" as far as I understand. But
I'll let the PCI maintainers comment on that. For reference, mainline
kernel makes CONFIG_PCI_MSI a hard dependency of pci-imx6 since commit
3ee803641e76 ("PCI/MSI: irqchip: Fix PCI_MSI dependencies") in v4.8.

In my opinion, if the driver actually compiles and is usable without
CONFIG_PCI_MSI then Kconfig dependencies should reflect that.

>> > +		val = dw_pcie_readw_dbi(pci, PCI_MSI_CAP + PCI_MSI_FLAGS);
>> > +		val |= PCI_MSI_FLAGS_ENABLE;
>> > +		dw_pcie_writew_dbi(pci, PCI_MSI_CAP + PCI_MSI_FLAGS, val);
>> > +	}
>> >
>> >  	return 0;
>> >  }

baruch

--
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch@xxxxxxxxxx - tel: +972.52.368.4656, http://www.tkos.co.il -