Re: [PATCH v9 4/8] vpci: Hide extended capability when it fails to initialize

[Jan Beulich <jbeulich@xxxxxxxx>]

On 30.07.2025 12:19, Andrew Cooper wrote:
> On 30/07/2025 10:50 am, Jan Beulich wrote:
>> On 28.07.2025 07:03, Jiqian Chen wrote:
>>> +static int vpci_ext_capability_hide(
>>> +    const struct pci_dev *pdev, unsigned int cap)
>>> +{
>>> +    const unsigned int offset = pci_find_ext_capability(pdev->sbdf, cap);
>>> +    struct vpci_register *r, *prev_r;
>>> +    struct vpci *vpci = pdev->vpci;
>>> +    uint32_t header, pre_header;
>>> +
>>> +    if ( offset < PCI_CFG_SPACE_SIZE )
>>> +    {
>>> +        ASSERT_UNREACHABLE();
>>> +        return 0;
>>> +    }
>>> +
>>> +    spin_lock(&vpci->lock);
>>> +    r = vpci_get_register(vpci, offset, 4);
>>> +    if ( !r )
>>> +    {
>>> +        spin_unlock(&vpci->lock);
>>> +        return -ENODEV;
>>> +    }
>>> +
>>> +    header = (uint32_t)(uintptr_t)r->private;
>>> +    if ( offset == PCI_CFG_SPACE_SIZE )
>>> +    {
>>> +        if ( PCI_EXT_CAP_NEXT(header) <= PCI_CFG_SPACE_SIZE )
>>> +            r->private = (void *)(uintptr_t)0;
>> Eclair regards this a Misra rule 11.9 violation. Elsewhere we use (void *)0,
>> which I then would conclude is "fine". But I can't say why that is. Cc-ing
>> Bugseng for a possible explanation.
> 
> Eclair is complaining that this isn't written r->private = NULL.
> 
> Given that private is a pointer, I don't understand why NULL isn't used
> either.

As with the various uses in calls to vpci_add_register(), the goal is to
indicate we want a value of 0 (could in principle be non-0 values as well,
but happens to be 0 in a number of cases), disguised as a pointer. Which
NULL doesn't quite express. And NULL there would also be inconsistent with
some (void *)0x25 that may need using elsewhere.

Jan