Messages in this thread | | | Date | Wed, 19 Feb 2014 11:24:29 -0500 | From | Boris Ostrovsky <> | Subject | Re: [PATCH] xen-pciback: Use pci_enable_msix_range() instead of pci_enable_msix() |
| |
On 02/19/2014 11:05 AM, Alexander Gordeev wrote: > On Wed, Feb 19, 2014 at 10:40:19AM -0500, Boris Ostrovsky wrote: >>>> diff --git a/drivers/xen/xen-pciback/pciback_ops.c >>>> b/drivers/xen/xen-pciback/pciback_ops.c >>>> index 64eb0cd..f5b4c3e 100644 >>>> --- a/drivers/xen/xen-pciback/pciback_ops.c >>>> +++ b/drivers/xen/xen-pciback/pciback_ops.c >>>> @@ -213,9 +213,15 @@ int xen_pcibk_enable_msix(struct >>>> xen_pcibk_device *pdev, >>>> entries[i].vector = op->msix_entries[i].vector; >>>> } >>>> - result = pci_enable_msix(dev, entries, op->value); >>>> + result = pci_enable_msix_range(dev, entries, op->value, op->value); >>>> + if (result < op->value) { >>> >>> I think it would be better to have 'if (result != op->value)', in >>> case op->value is negative (which presumably it should never be). >> >> Better yet, at the top of the routine we check 'if (op->value > >> SH_INFO_MAX_VEC)'. If you add '|| op->value < 0' we'd be all set. > xen_pci_op::value is uint32_t
Ah, OK --- then 'if (op->value > SH_INFO_MAX_VEC)' alone will catch this (hopefully its' not in billions).
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
| |