| Message ID | 20200527153152.9211-1-thuth@redhat.com |
|---|---|
| State | New |
| Headers | show |
| Series | hw/pci: Fix crash when running QEMU with "-nic model=rocker" | expand |
On 27/05/20 17:31, Thomas Huth wrote: > QEMU currently aborts when being started with "-nic model=rocker" or with > "-net nic,model=rocker". This happens because the "rocker" device is not > a normal NIC but a switch, which has different properties. Thus we should > only consider real NIC devices for "-nic" and "-net". These devices can > be identified by the "netdev" property, so check for this property before > adding the device to the list. > > Reported-by: Michael Tokarev <mjt@tls.msk.ru> > Fixes: 52310c3fa7dc854d ("net: allow using any PCI NICs in -net or -nic") > Signed-off-by: Thomas Huth <thuth@redhat.com> > --- > hw/pci/pci.c | 13 ++++++++++++- > 1 file changed, 12 insertions(+), 1 deletion(-) > > diff --git a/hw/pci/pci.c b/hw/pci/pci.c > index 70c66965f5..46214f8287 100644 > --- a/hw/pci/pci.c > +++ b/hw/pci/pci.c > @@ -1887,7 +1887,18 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, > if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) && > dc->user_creatable) { > const char *name = object_class_get_name(list->data); > - g_ptr_array_add(pci_nic_models, (gpointer)name); > + /* > + * A network device might also be something else than a NIC, see > + * e.g. the "rocker" device. Thus we have to look for the "netdev" > + * property, too. Unfortunately, some devices like virtio-net only > + * create this property during instance_init, so we have to create > + * a temporary instance here to be able to check it. > + */ > + Object *obj = object_new_with_class(OBJECT_CLASS(dc)); > + if (object_property_find(obj, "netdev", NULL)) { > + g_ptr_array_add(pci_nic_models, (gpointer)name); > + } > + object_unref(obj); > } > next = list->next; > g_slist_free_1(list); > Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Le 27/05/2020 à 17:31, Thomas Huth a écrit : > QEMU currently aborts when being started with "-nic model=rocker" or with > "-net nic,model=rocker". This happens because the "rocker" device is not > a normal NIC but a switch, which has different properties. Thus we should > only consider real NIC devices for "-nic" and "-net". These devices can > be identified by the "netdev" property, so check for this property before > adding the device to the list. > > Reported-by: Michael Tokarev <mjt@tls.msk.ru> > Fixes: 52310c3fa7dc854d ("net: allow using any PCI NICs in -net or -nic") > Signed-off-by: Thomas Huth <thuth@redhat.com> > --- > hw/pci/pci.c | 13 ++++++++++++- > 1 file changed, 12 insertions(+), 1 deletion(-) > > diff --git a/hw/pci/pci.c b/hw/pci/pci.c > index 70c66965f5..46214f8287 100644 > --- a/hw/pci/pci.c > +++ b/hw/pci/pci.c > @@ -1887,7 +1887,18 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, > if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) && > dc->user_creatable) { > const char *name = object_class_get_name(list->data); > - g_ptr_array_add(pci_nic_models, (gpointer)name); > + /* > + * A network device might also be something else than a NIC, see > + * e.g. the "rocker" device. Thus we have to look for the "netdev" > + * property, too. Unfortunately, some devices like virtio-net only > + * create this property during instance_init, so we have to create > + * a temporary instance here to be able to check it. > + */ > + Object *obj = object_new_with_class(OBJECT_CLASS(dc)); > + if (object_property_find(obj, "netdev", NULL)) { > + g_ptr_array_add(pci_nic_models, (gpointer)name); > + } > + object_unref(obj); > } > next = list->next; > g_slist_free_1(list); > Not really trivial, I will not pick up this patch via trivial-branch, unless PCI maintainers request it. Thanks, Laurent
On 09/06/20 19:21, Laurent Vivier wrote: > Le 27/05/2020 à 17:31, Thomas Huth a écrit : >> QEMU currently aborts when being started with "-nic model=rocker" or with >> "-net nic,model=rocker". This happens because the "rocker" device is not >> a normal NIC but a switch, which has different properties. Thus we should >> only consider real NIC devices for "-nic" and "-net". These devices can >> be identified by the "netdev" property, so check for this property before >> adding the device to the list. >> >> Reported-by: Michael Tokarev <mjt@tls.msk.ru> >> Fixes: 52310c3fa7dc854d ("net: allow using any PCI NICs in -net or -nic") >> Signed-off-by: Thomas Huth <thuth@redhat.com> >> --- >> hw/pci/pci.c | 13 ++++++++++++- >> 1 file changed, 12 insertions(+), 1 deletion(-) >> >> diff --git a/hw/pci/pci.c b/hw/pci/pci.c >> index 70c66965f5..46214f8287 100644 >> --- a/hw/pci/pci.c >> +++ b/hw/pci/pci.c >> @@ -1887,7 +1887,18 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, >> if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) && >> dc->user_creatable) { >> const char *name = object_class_get_name(list->data); >> - g_ptr_array_add(pci_nic_models, (gpointer)name); >> + /* >> + * A network device might also be something else than a NIC, see >> + * e.g. the "rocker" device. Thus we have to look for the "netdev" >> + * property, too. Unfortunately, some devices like virtio-net only >> + * create this property during instance_init, so we have to create >> + * a temporary instance here to be able to check it. >> + */ >> + Object *obj = object_new_with_class(OBJECT_CLASS(dc)); >> + if (object_property_find(obj, "netdev", NULL)) { >> + g_ptr_array_add(pci_nic_models, (gpointer)name); >> + } >> + object_unref(obj); >> } >> next = list->next; >> g_slist_free_1(list); >> > > Not really trivial, I will not pick up this patch via trivial-branch, > unless PCI maintainers request it. > > Thanks, > Laurent > I'll pick it up then. Paolo
On Tue, Jun 09, 2020 at 07:24:24PM +0200, Paolo Bonzini wrote: > On 09/06/20 19:21, Laurent Vivier wrote: > > Le 27/05/2020 à 17:31, Thomas Huth a écrit : > >> QEMU currently aborts when being started with "-nic model=rocker" or with > >> "-net nic,model=rocker". This happens because the "rocker" device is not > >> a normal NIC but a switch, which has different properties. Thus we should > >> only consider real NIC devices for "-nic" and "-net". These devices can > >> be identified by the "netdev" property, so check for this property before > >> adding the device to the list. > >> > >> Reported-by: Michael Tokarev <mjt@tls.msk.ru> > >> Fixes: 52310c3fa7dc854d ("net: allow using any PCI NICs in -net or -nic") > >> Signed-off-by: Thomas Huth <thuth@redhat.com> > >> --- > >> hw/pci/pci.c | 13 ++++++++++++- > >> 1 file changed, 12 insertions(+), 1 deletion(-) > >> > >> diff --git a/hw/pci/pci.c b/hw/pci/pci.c > >> index 70c66965f5..46214f8287 100644 > >> --- a/hw/pci/pci.c > >> +++ b/hw/pci/pci.c > >> @@ -1887,7 +1887,18 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, > >> if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) && > >> dc->user_creatable) { > >> const char *name = object_class_get_name(list->data); > >> - g_ptr_array_add(pci_nic_models, (gpointer)name); > >> + /* > >> + * A network device might also be something else than a NIC, see > >> + * e.g. the "rocker" device. Thus we have to look for the "netdev" > >> + * property, too. Unfortunately, some devices like virtio-net only > >> + * create this property during instance_init, so we have to create > >> + * a temporary instance here to be able to check it. > >> + */ > >> + Object *obj = object_new_with_class(OBJECT_CLASS(dc)); > >> + if (object_property_find(obj, "netdev", NULL)) { > >> + g_ptr_array_add(pci_nic_models, (gpointer)name); > >> + } > >> + object_unref(obj); > >> } > >> next = list->next; > >> g_slist_free_1(list); > >> > > > > Not really trivial, I will not pick up this patch via trivial-branch, > > unless PCI maintainers request it. > > > > Thanks, > > Laurent > > > > I'll pick it up then. > > Paolo I queued this already. Thanks!
diff --git a/hw/pci/pci.c b/hw/pci/pci.c index 70c66965f5..46214f8287 100644 --- a/hw/pci/pci.c +++ b/hw/pci/pci.c @@ -1887,7 +1887,18 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) && dc->user_creatable) { const char *name = object_class_get_name(list->data); - g_ptr_array_add(pci_nic_models, (gpointer)name); + /* + * A network device might also be something else than a NIC, see + * e.g. the "rocker" device. Thus we have to look for the "netdev" + * property, too. Unfortunately, some devices like virtio-net only + * create this property during instance_init, so we have to create + * a temporary instance here to be able to check it. + */ + Object *obj = object_new_with_class(OBJECT_CLASS(dc)); + if (object_property_find(obj, "netdev", NULL)) { + g_ptr_array_add(pci_nic_models, (gpointer)name); + } + object_unref(obj); } next = list->next; g_slist_free_1(list);
QEMU currently aborts when being started with "-nic model=rocker" or with "-net nic,model=rocker". This happens because the "rocker" device is not a normal NIC but a switch, which has different properties. Thus we should only consider real NIC devices for "-nic" and "-net". These devices can be identified by the "netdev" property, so check for this property before adding the device to the list. Reported-by: Michael Tokarev <mjt@tls.msk.ru> Fixes: 52310c3fa7dc854d ("net: allow using any PCI NICs in -net or -nic") Signed-off-by: Thomas Huth <thuth@redhat.com> --- hw/pci/pci.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)