On 2012-08-16 18:21, Alex Williamson wrote:
> On Thu, 2012-08-16 at 15:54 +0200, Jan Kiszka wrote:
>> The refactored version cleanly hides the KVM IOCTL structure from the
>> users and also zeros out the padding field.
>>
>> Signed-off-by: Jan Kiszka <jan.kiszka@xxxxxxxxxxx>
>> ---
>> hw/device-assignment.c | 7 ++-----
>> qemu-kvm.c | 8 --------
>> qemu-kvm.h | 4 ----
>> target-i386/kvm.c | 13 +++++++++++++
>> target-i386/kvm_i386.h | 2 ++
>> 5 files changed, 17 insertions(+), 17 deletions(-)
>>
>> diff --git a/hw/device-assignment.c b/hw/device-assignment.c
>> index 0e2f8e6..af8a5aa 100644
>> --- a/hw/device-assignment.c
>> +++ b/hw/device-assignment.c
>> @@ -1045,7 +1045,6 @@ static int assigned_dev_update_msix_mmio(PCIDevice *pci_dev)
>> AssignedDevice *adev = DO_UPCAST(AssignedDevice, dev, pci_dev);
>> uint16_t entries_nr = 0;
>> int i, r = 0;
>> - struct kvm_assigned_msix_entry msix_entry;
>> MSIXTableEntry *entry = adev->msix_table;
>>
>> /* Get the usable entry number for allocating */
>> @@ -1075,7 +1074,6 @@ static int assigned_dev_update_msix_mmio(PCIDevice *pci_dev)
>> adev->irq_entries_nr = adev->msix_max;
>> adev->entry = g_malloc0(adev->msix_max * sizeof(*(adev->entry)));
>>
>> - msix_entry.assigned_dev_id = adev->dev_id;
>> entry = adev->msix_table;
>> for (i = 0; i < adev->msix_max; i++, entry++) {
>> if (msix_masked(entry)) {
>> @@ -1098,9 +1096,8 @@ static int assigned_dev_update_msix_mmio(PCIDevice *pci_dev)
>>
>> kvm_add_routing_entry(kvm_state, &adev->entry[i]);
>>
>> - msix_entry.gsi = adev->entry[i].gsi;
>> - msix_entry.entry = i;
>> - r = kvm_assign_set_msix_entry(kvm_state, &msix_entry);
>> + r = kvm_device_msix_set_vector(kvm_state, adev->dev_id, i,
>> + adev->entry[i].gsi);
>> if (r) {
>> fprintf(stderr, "fail to set MSI-X entry! %s\n", strerror(-r));
>> break;
>> diff --git a/qemu-kvm.c b/qemu-kvm.c
>> index 1a2a4fd..ec1911f 100644
>> --- a/qemu-kvm.c
>> +++ b/qemu-kvm.c
>> @@ -185,14 +185,6 @@ int kvm_get_irq_route_gsi(void)
>> #endif
>> }
>>
>> -#ifdef KVM_CAP_DEVICE_MSIX
>> -int kvm_assign_set_msix_entry(KVMState *s,
>> - struct kvm_assigned_msix_entry *entry)
>> -{
>> - return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_ENTRY, entry);
>> -}
>> -#endif
>> -
>> #if !defined(TARGET_I386)
>> void kvm_arch_init_irq_routing(KVMState *s)
>> {
>> diff --git a/qemu-kvm.h b/qemu-kvm.h
>> index 3fd6046..ad628d5 100644
>> --- a/qemu-kvm.h
>> +++ b/qemu-kvm.h
>> @@ -65,10 +65,6 @@ int kvm_del_routing_entry(struct kvm_irq_routing_entry *entry);
>> int kvm_update_routing_entry(struct kvm_irq_routing_entry *entry,
>> struct kvm_irq_routing_entry *newentry);
>>
>> -
>> -int kvm_assign_set_msix_entry(KVMState *s,
>> - struct kvm_assigned_msix_entry *entry);
>> -
>> #endif /* CONFIG_KVM */
>>
>> #endif
>> diff --git a/target-i386/kvm.c b/target-i386/kvm.c
>> index 676f45b..e9353ed 100644
>> --- a/target-i386/kvm.c
>> +++ b/target-i386/kvm.c
>> @@ -2173,6 +2173,19 @@ int kvm_device_msix_init_vectors(KVMState *s, uint32_t dev_id,
>> return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_NR, &msix_nr);
>> }
>>
>> +int kvm_device_msix_set_vector(KVMState *s, uint32_t dev_id, uint32_t vector,
>> + int virq)
>> +{
>> + struct kvm_assigned_msix_entry msix_entry = {
>> + .assigned_dev_id = dev_id,
>> + .gsi = virq,
>> + .entry = vector,
>> + };
>> +
>> + memset(msix_entry.padding, 0, sizeof(msix_entry.padding));
>
> nit, I think this can be done w/o a memset. .padding = { 0 }? Thanks,
I think to remember it has to be .padding = { 0, 0, 0 } (due to three
padding elements) to be standard conforming, but that would still be
nicer than the memset, indeed.
Jan
--
Siemens AG, Corporate Technology, CT RTC ITP SDP-DE
Corporate Competence Center Embedded Linux
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
