Re: [PATCH, net-next] r8169: Disable some bits on pcie

[Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>]

On 8/20/2015 3:57 PM, Sergei Shtylyov wrote:

>> Disable legacy interrupt on pci express interface use msi
>>   disable some bits from pci express interface wich is not   need on this nic.
>>
>>
>> Signed-off-by: Corcodel Marian <corcodel.marian@gmail.com>
>>
>> diff --git a/drivers/net/ethernet/realtek/r8169.c
>> b/drivers/net/ethernet/realtek/r8169.c
>> index 6d16de3..b1fb54f 100644
>> --- a/drivers/net/ethernet/realtek/r8169.c
>> +++ b/drivers/net/ethernet/realtek/r8169.c
>> @@ -8164,6 +8164,13 @@ static int rtl_init_one(struct pci_dev *pdev, const
>> struct pci_device_id *ent)
>>       if (!pci_is_pcie(pdev))
>>           netif_info(tp, probe, dev, "not PCI Express\n");
>>
>> +    if (pci_is_pcie(pdev))
>> +        pci_write_config_word(pdev, PCI_COMMAND, ~(PCI_COMMAND_FAST_BACK |
>> PCI_COMMAND_WAIT |
>> +                    PCI_COMMAND_VGA_PALETTE | PCI_COMMAND_INVALIDATE |
>> PCI_COMMAND_SPECIAL));
>
>     Please start the continuation line under 'pdev' on the first line.

    Can all those bits indeed be set for a networking device? 
PCI_COMMAND_VGA_PALETTE is the most incredible...

>> +
>> +    if (pci_is_pcie(pdev))
>> +        pci_intx(pdev, 0);
>> +
>
>     Why two identical *if* statements? Please make it one.

    Ah, I missed the *if* at the top. You need just an *else* clause here.

MBR, Sergei