Re: [PATCH 2/2] vpci: use named rangeset for BARs

[Jan Beulich <jbeulich@suse.com>]

On 22.11.2021 11:27, Roger Pau Monné wrote:
> On Mon, Nov 22, 2021 at 11:28:25AM +0200, Oleksandr Andrushchenko wrote:
>> --- a/xen/drivers/vpci/header.c
>> +++ b/xen/drivers/vpci/header.c
>> @@ -206,12 +206,16 @@ static void defer_map(struct domain *d, struct pci_dev *pdev,
>>  static int modify_bars(const struct pci_dev *pdev, uint16_t cmd, bool rom_only)
>>  {
>>      struct vpci_header *header = &pdev->vpci->header;
>> -    struct rangeset *mem = rangeset_new(NULL, NULL, 0);
>> +    struct rangeset *mem;
>> +    char str[32];
>>      struct pci_dev *tmp, *dev = NULL;
>>      const struct vpci_msix *msix = pdev->vpci->msix;
>>      unsigned int i;
>>      int rc;
>>  
>> +    snprintf(str, sizeof(str), "%pp", &pdev->sbdf);
>> +    mem = rangeset_new(NULL, str, RANGESETF_no_print);
> 
> You are still not adding the rangeset to the domain list, as the first
> parameter passed here in NULL instead of a domain struct.
> 
> Given the current short living of the rangesets I'm not sure it makes
> much sense to link them to the domain ATM, but I guess this is kind of
> a preparatory change as other patches you have will have the
> rangesets permanent as long as the device is assigned to a domain.
> 
> Likely the above reasoning (or the appropriate one) should be added to
> the commit message.

Or, as also suggested as an option, them getting accounted to the domain
could be folded into the patch making them long-lived.

Jan