Re: [PATCH 1/3] PCI: rcar: Add the initialization of PCIe link in resume_noirq

[Marek Vasut <marek.vasut@gmail.com>]

On 11/10/2017 09:59 AM, Simon Horman wrote:
> Hi Marek,

Hi Simon,

> On Wed, Nov 08, 2017 at 10:28:04AM +0100, Marek Vasut wrote:
>> From: Kazufumi Ikeda <kaz-ikeda@xc.jp.nec.com>
>>
>> Reestablish the PCIe link very early in the resume process in case it
>> went down to prevent PCI accesses from hanging the bus. Such accesses
>> can happen early in the PCI resume process, in the resume_noirq, thus
>> the link must be reestablished in the resume_noirq callback of the
>> driver.
>>
>> Signed-off-by: Kazufumi Ikeda <kaz-ikeda@xc.jp.nec.com>
>> Signed-off-by: Gaku Inami <gaku.inami.xw@bp.renesas.com>
>> Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
>> Cc: Geert Uytterhoeven <geert+renesas@glider.be>
>> Cc: Simon Horman <horms+renesas@verge.net.au>
>> Cc: Wolfram Sang <wsa@the-dreams.de>
>> Cc: linux-renesas-soc@vger.kernel.org
> 
> For patch-sets (with more than one patch) please provide a cover-letter.
> The --cover-letter option to git format-patch can help.
> 
>> ---
>>  drivers/pci/host/pcie-rcar.c | 31 ++++++++++++++++++++++++++++---
>>  1 file changed, 28 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/pci/host/pcie-rcar.c b/drivers/pci/host/pcie-rcar.c
>> index 889603783f01..aa588a7d4811 100644
>> --- a/drivers/pci/host/pcie-rcar.c
>> +++ b/drivers/pci/host/pcie-rcar.c
>> @@ -43,6 +43,7 @@
>>  
>>  /* Transfer control */
>>  #define PCIETCTLR		0x02000
>> +#define  DL_DOWN		(1 << 3)
> 
> Can you use the BIT() macro here?

Yes, I'll also add a patch to the series to convert all the others to
the BIT() macro, to keep things consistent.

>>  #define  CFINIT			1
>>  #define PCIETSTR		0x02004
>>  #define  DATA_LINK_ACTIVE	1
>> @@ -529,7 +530,7 @@ static void phy_write_reg(struct rcar_pcie *pcie,
>>  	phy_wait_for_ack(pcie);
>>  }
>>  
>> -static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
>> +static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie, int atomic)
> 
> As Sergei mentioned bool seems like a more appropriate type for atomic.

I can actually get rid of this altogether.

>>  {
>>  	unsigned int timeout = 10;
>>  
>> @@ -537,7 +538,10 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
>>  		if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
>>  			return 0;
>>  
>> -		msleep(5);
>> +		if (atomic)
>> +			mdelay(5);
>> +		else
>> +			msleep(5);
> 
> If we must delay, then I suppose this is reasonable.

See above

>>  	}
>>  
>>  	return -ETIMEDOUT;
>> @@ -595,7 +599,7 @@ static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
>>  	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
>>  
>>  	/* This will timeout if we don't have a link. */
>> -	err = rcar_pcie_wait_for_dl(pcie);
>> +	err = rcar_pcie_wait_for_dl(pcie, 0);
>>  	if (err)
>>  		return err;
>>  
>> @@ -1110,6 +1114,7 @@ static int rcar_pcie_probe(struct platform_device *pdev)
>>  	pcie = pci_host_bridge_priv(bridge);
>>  
>>  	pcie->dev = dev;
>> +	platform_set_drvdata(pdev, pcie);
>>  
>>  	INIT_LIST_HEAD(&pcie->resources);
>>  
>> @@ -1173,10 +1178,30 @@ static int rcar_pcie_probe(struct platform_device *pdev)
>>  	return err;
>>  }
>>  
>> +static int rcar_pcie_resume_noirq(struct device *dev)
>> +{
>> +	struct rcar_pcie *pcie = dev_get_drvdata(dev);
>> +	u32 val = rcar_pci_read_reg(pcie, PMSR);
>> +	int ret = 0;
>> +
>> +	if ((val == 0) || (rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN)) {
> 
> Please remove the unnecessary parentheses from the line above.
> 
> Also, I would prefer if the function returned early.
> Something like (completely untested!):

This should work.

> 	if (rcar_pci_read_reg(pcie, PMSR) &&
> 	    !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
> 		return 0;
> 
> 	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> 	return rcar_pcie_wait_for_dl(pcie, 1);
> 
>> +		/* Re-establish the PCIe link */
>> +		rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
>> +		ret = rcar_pcie_wait_for_dl(pcie, 1);
>> +	}
>> +
>> +	return ret;
>> +}
>> +
>> +static const struct dev_pm_ops rcar_pcie_pm_ops = {
>> +	.resume_noirq = rcar_pcie_resume_noirq,
>> +};
>> +
>>  static struct platform_driver rcar_pcie_driver = {
>>  	.driver = {
>>  		.name = "rcar-pcie",
>>  		.of_match_table = rcar_pcie_of_match,
>> +		.pm = &rcar_pcie_pm_ops,
>>  		.suppress_bind_attrs = true,
>>  	},
>>  	.probe = rcar_pcie_probe,
>> -- 
>> 2.11.0
>>


-- 
Best regards,
Marek Vasut