Re: [PATCH V7] ns16550: Add support for UART present in Broadcom TruManage capable NetXtreme chips

[Aravind Gopalakrishnan <aravind.gopalakrishnan@amd.com>]

On 12/3/2013 10:46 AM, Jan Beulich wrote:
>>>> On 03.12.13 at 17:40, Aravind Gopalakrishnan <Aravind.Gopalakrishnan@amd.com> wrote:
>> @@ -434,7 +477,22 @@ static void __init ns16550_endboot(struct serial_port *port)
>>       struct ns16550 *uart = port->uart;
>>   
>>       if ( uart->remapped_io_base )
>> +    {
>> +        if ( uart->enable_ro ) {
>> +            if ( rangeset_add_range(mmio_ro_ranges,
>> +                                    uart->io_base,
>> +                                    uart->io_base + uart->io_size - 1) )
>> +                printk(XENLOG_INFO "Error while adding MMIO range of device to mmio_ro_ranges\n");
>> +
>> +            if ( pci_ro_device(0, uart->ps_bdf[0],
>> +                               PCI_DEVFN(uart->ps_bdf[1], uart->ps_bdf[2])) )
> You didn't really fix what you said you fixed. Quoting from my
> reply to v6:

Ok, moved pci_ro_device also to ns16550_init_postirq now so that I do 
the hiding in one place..

>>> But, more importantly, did you overlook the use of pci_hide_device()
>>> in ns16550_init_postirq(): The hiding should be done in one place.
>>> And with pci_ro_device() implicitly hiding the device, you should
>>> probably make sure you call just one of the two.
>> +                printk(XENLOG_INFO "Could not mark config space of 0:%02x:%02x.%u read-only.\n",
>> +                                    uart->ps_bdf[0], uart->ps_bdf[1],
>> +                                    uart->ps_bdf[2]);
> The leading 0: is pretty pointless - an absent segment/domain
> identifier implies it being zero.
>
>
Fixed this. Sending out changes in V8.

-Aravind.