From mboxrd@z Thu Jan  1 00:00:00 1970
From: rjui@broadcom.com (Ray Jui)
Date: Mon, 25 May 2015 17:32:48 -0700
Subject: [PATCH 1/2] PCI: iproc: directly add pci resources
In-Reply-To: <5563A6B3.2030600@hauke-m.de>
References: <1432499823-27369-1-git-send-email-hauke@hauke-m.de>
 <5563579E.80608@broadcom.com> <5563A6B3.2030600@hauke-m.de>
Message-ID: <5563BF30.2040106@broadcom.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Hi Hauke,

On 15-05-25 03:48 PM, Hauke Mehrtens wrote:
> On 05/25/2015 07:10 PM, Ray Jui wrote:
>> Hi Hauke,
>>
>> On 5/24/2015 1:37 PM, Hauke Mehrtens wrote:
>>> The resources member in the struct was pointing to a stack variable and
>>> is invalid after the the registration function returned. Remove this
>>> pointer and add it a a parameter to the function.
>>>
>>> Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
>>> ---
>>>   drivers/pci/host/pcie-iproc-bcma.c     | 4 +---
>>>   drivers/pci/host/pcie-iproc-platform.c | 4 +---
>>>   drivers/pci/host/pcie-iproc.c          | 4 ++--
>>>   drivers/pci/host/pcie-iproc.h          | 3 +--
>>>   4 files changed, 5 insertions(+), 10 deletions(-)
>>>
>>> diff --git a/drivers/pci/host/pcie-iproc-bcma.c b/drivers/pci/host/pcie-iproc-bcma.c
>>> index c318f19..f96b39e 100644
>>> --- a/drivers/pci/host/pcie-iproc-bcma.c
>>> +++ b/drivers/pci/host/pcie-iproc-bcma.c
>>> @@ -62,11 +62,9 @@ static int iproc_pcie_bcma_probe(struct bcma_device *bdev)
>>>   	res_mem.flags = IORESOURCE_MEM;
>>>   	pci_add_resource(&res, &res_mem);
>>>
>>> -	pcie->resources = &res;
>>> -
>>>   	pcie->map_irq = iproc_pcie_bcma_map_irq;
>>>
>>> -	ret = iproc_pcie_setup(pcie);
>>> +	ret = iproc_pcie_setup(pcie, &res);
>>>   	if (ret) {
>>>   		dev_err(pcie->dev, "PCIe controller setup failed\n");
>>>   		return ret;
>>> diff --git a/drivers/pci/host/pcie-iproc-platform.c b/drivers/pci/host/pcie-iproc-platform.c
>>> index c8aa06f..c5fe4c1 100644
>>> --- a/drivers/pci/host/pcie-iproc-platform.c
>>> +++ b/drivers/pci/host/pcie-iproc-platform.c
>>> @@ -69,11 +69,9 @@ static int iproc_pcie_pltfm_probe(struct platform_device *pdev)
>>>   		return ret;
>>>   	}
>>>
>>> -	pcie->resources = &res;
>>> -
>>>   	pcie->map_irq = of_irq_parse_and_map_pci;
>>>
>>> -	ret = iproc_pcie_setup(pcie);
>>> +	ret = iproc_pcie_setup(pcie, &res);
>>>   	if (ret) {
>>>   		dev_err(pcie->dev, "PCIe controller setup failed\n");
>>>   		return ret;
>>> diff --git a/drivers/pci/host/pcie-iproc.c b/drivers/pci/host/pcie-iproc.c
>>> index cef31f6..d77481e 100644
>>> --- a/drivers/pci/host/pcie-iproc.c
>>> +++ b/drivers/pci/host/pcie-iproc.c
>>> @@ -183,7 +183,7 @@ static void iproc_pcie_enable(struct iproc_pcie *pcie)
>>>   	writel(SYS_RC_INTX_MASK, pcie->base + SYS_RC_INTX_EN);
>>>   }
>>>
>>> -int iproc_pcie_setup(struct iproc_pcie *pcie)
>>> +int iproc_pcie_setup(struct iproc_pcie *pcie, struct list_head *res)
>>>   {
>>>   	int ret;
>>>   	struct pci_bus *bus;
>>> @@ -211,7 +211,7 @@ int iproc_pcie_setup(struct iproc_pcie *pcie)
>>>   	pcie->sysdata.private_data = pcie;
>>>
>>>   	bus = pci_create_root_bus(pcie->dev, 0, &iproc_pcie_ops,
>>> -				  &pcie->sysdata, pcie->resources);
>>> +				  &pcie->sysdata, res);
>>>   	if (!bus) {
>>>   		dev_err(pcie->dev, "unable to create PCI root bus\n");
>>>   		ret = -ENOMEM;
>>> diff --git a/drivers/pci/host/pcie-iproc.h b/drivers/pci/host/pcie-iproc.h
>>> index a333d4b..ba0a108 100644
>>> --- a/drivers/pci/host/pcie-iproc.h
>>> +++ b/drivers/pci/host/pcie-iproc.h
>>> @@ -29,7 +29,6 @@
>>>   struct iproc_pcie {
>>>   	struct device *dev;
>>>   	void __iomem *base;
>>> -	struct list_head *resources;
>>
>> This means we do not want to keep a copy of the resources. In the
>> future, if we need to add support to explicitly set up the
>> inbound/outbound mapping window, we need to do it in iproc_pcie_setup.
>> Is that the intention?
>
> I haven't really thought about where to configure the memory mapping,
> but I thought it would be somewhere in iproc_pcie_setup() or an method
> called from there.
> If you need it after iproc_pcie_setup() finished we should embed the
> list into the struct iproc_pcie, because currently this pointer points
> to the stack.
>
> Hauke
>

I think most likely I'll be adding the memory mapping logic in 
iproc_pcie_setup(), so your current patch should be fine.

Reviewed-by: Ray Jui <rjui@broadcom.com>
Tested-by: Ray Jui <rjui@broadcom.com>

Thanks.