[RESEND,v2,2/4] PCI: Do not Use IORESOURCE_STARTALIGN to identify bridge resources

Now we use the IORESOURCE_STARTALIGN to identify bridge resources
in __assign_resources_sorted(). That's quite fragile. We can't
make sure that the PCI devices' resources will not use
IORESOURCE_STARTALIGN any more.

In this patch, we try to use a more robust way to identify
bridge resources.

Signed-off-by: Yongji Xie <xyjxie@linux.vnet.ibm.com>
---
 drivers/pci/setup-bus.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

On Thu, Jun 02, 2016 at 01:46:49PM +0800, Yongji Xie wrote:
> Now we use the IORESOURCE_STARTALIGN to identify bridge resources
> in __assign_resources_sorted(). That's quite fragile. We can't
> make sure that the PCI devices' resources will not use
> IORESOURCE_STARTALIGN any more.

Can you explain this a little more?  I don't quite understand the
problem.  Maybe you can give an example of the problem?

> In this patch, we try to use a more robust way to identify
> bridge resources.
> 
> Signed-off-by: Yongji Xie <xyjxie@linux.vnet.ibm.com>
> ---
>  drivers/pci/setup-bus.c |    9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
> index 55641a3..216ddbc 100644
> --- a/drivers/pci/setup-bus.c
> +++ b/drivers/pci/setup-bus.c
> @@ -390,6 +390,7 @@ static void __assign_resources_sorted(struct list_head *head,
>  	struct pci_dev_resource *dev_res, *tmp_res, *dev_res2;
>  	unsigned long fail_type;
>  	resource_size_t add_align, align;
> +	int index;
>  
>  	/* Check if optional add_size is there */
>  	if (!realloc_head || list_empty(realloc_head))
> @@ -410,11 +411,13 @@ static void __assign_resources_sorted(struct list_head *head,
>  
>  		/*
>  		 * There are two kinds of additional resources in the list:
> -		 * 1. bridge resource  -- IORESOURCE_STARTALIGN
> -		 * 2. SR-IOV resource   -- IORESOURCE_SIZEALIGN
> +		 * 1. bridge resource
> +		 * 2. SR-IOV resource
>  		 * Here just fix the additional alignment for bridge
>  		 */
> -		if (!(dev_res->res->flags & IORESOURCE_STARTALIGN))
> +		index = dev_res->res - dev_res->dev->resource;
> +		if (index < PCI_BRIDGE_RESOURCES ||
> +			index > PCI_BRIDGE_RESOURCE_END)

I think the code looks OK; at least, it seems to match the comment.
I'd just like to understand the problem a little better.

>  			continue;
>  
>  		add_align = get_res_add_align(realloc_head, dev_res->res);
> -- 
> 1.7.9.5
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On 2016/6/21 9:50, Bjorn Helgaas wrote:

> On Thu, Jun 02, 2016 at 01:46:49PM +0800, Yongji Xie wrote:
>> Now we use the IORESOURCE_STARTALIGN to identify bridge resources
>> in __assign_resources_sorted(). That's quite fragile. We can't
>> make sure that the PCI devices' resources will not use
>> IORESOURCE_STARTALIGN any more.
> Can you explain this a little more?  I don't quite understand the
> problem.  Maybe you can give an example of the problem?

Now there are two kinds of additional resources in the list: bridge
resource and SR-IOV resource. Here we just want to fix the
additional alignment for bridge. But if SR-IOV resource get the
flag IORESOURCE_STARTALIGN which is not only for bridge, the
current check seems to be wrong. And kernel parameter
"resource_alignment" could set IORESOURCE_STARTALIGN for
SR-IOV resource with my next patch applied.

Thanks,
Yongji

>> In this patch, we try to use a more robust way to identify
>> bridge resources.
>>
>> Signed-off-by: Yongji Xie <xyjxie@linux.vnet.ibm.com>
>> ---
>>   drivers/pci/setup-bus.c |    9 ++++++---
>>   1 file changed, 6 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
>> index 55641a3..216ddbc 100644
>> --- a/drivers/pci/setup-bus.c
>> +++ b/drivers/pci/setup-bus.c
>> @@ -390,6 +390,7 @@ static void __assign_resources_sorted(struct list_head *head,
>>   	struct pci_dev_resource *dev_res, *tmp_res, *dev_res2;
>>   	unsigned long fail_type;
>>   	resource_size_t add_align, align;
>> +	int index;
>>   
>>   	/* Check if optional add_size is there */
>>   	if (!realloc_head || list_empty(realloc_head))
>> @@ -410,11 +411,13 @@ static void __assign_resources_sorted(struct list_head *head,
>>   
>>   		/*
>>   		 * There are two kinds of additional resources in the list:
>> -		 * 1. bridge resource  -- IORESOURCE_STARTALIGN
>> -		 * 2. SR-IOV resource   -- IORESOURCE_SIZEALIGN
>> +		 * 1. bridge resource
>> +		 * 2. SR-IOV resource
>>   		 * Here just fix the additional alignment for bridge
>>   		 */
>> -		if (!(dev_res->res->flags & IORESOURCE_STARTALIGN))
>> +		index = dev_res->res - dev_res->dev->resource;
>> +		if (index < PCI_BRIDGE_RESOURCES ||
>> +			index > PCI_BRIDGE_RESOURCE_END)
> I think the code looks OK; at least, it seems to match the comment.
> I'd just like to understand the problem a little better.
>
>>   			continue;
>>   
>>   		add_align = get_res_add_align(realloc_head, dev_res->res);
>> -- 
>> 1.7.9.5
>>
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html