| Message ID | 20220809025010.200289-1-chen.zhang@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [V3] net/colo.c: Fix the pointer issue reported by Coverity. | expand |
On Tue, Aug 9, 2022 at 11:05 AM Zhang Chen <chen.zhang@intel.com> wrote: > > When enable the virtio-net-pci, guest network packet will > load the vnet_hdr. In COLO status, the primary VM's network > packet maybe redirect to another VM, it need filter-redirect > enable the vnet_hdr flag at the same time, COLO-proxy will > correctly parse the original network packet. If have any > misconfiguration here, the vnet_hdr_len is wrong for parse > the packet, the data+offset will point to wrong place. > > Signed-off-by: Zhang Chen <chen.zhang@intel.com> > --- > net/colo.c | 18 ++++++++++-------- > net/colo.h | 1 + > 2 files changed, 11 insertions(+), 8 deletions(-) > > diff --git a/net/colo.c b/net/colo.c > index 6b0ff562ad..4f1b4a61f6 100644 > --- a/net/colo.c > +++ b/net/colo.c > @@ -44,21 +44,23 @@ int parse_packet_early(Packet *pkt) > { > int network_length; > static const uint8_t vlan[] = {0x81, 0x00}; > - uint8_t *data = pkt->data + pkt->vnet_hdr_len; > + uint8_t *data = pkt->data; > uint16_t l3_proto; > ssize_t l2hdr_len; > > - if (data == NULL) { > - trace_colo_proxy_main_vnet_info("This packet is not parsed correctly, " > + assert(data); > + > + /* Check the received vnet_hdr_len then add the offset */ > + if ((pkt->vnet_hdr_len > sizeof(struct virtio_net_hdr)) || Virtio-net starts to support RSS so it means the vnet_hdr size could be greater than virtio_net_hdr now. Or did you actually mean "<" here? Thanks > + (pkt->size < sizeof(struct eth_header) + sizeof(struct vlan_header) > + + pkt->vnet_hdr_len)) { > + trace_colo_proxy_main_vnet_info("This packet may be load wrong " > "pkt->vnet_hdr_len", pkt->vnet_hdr_len); > return 1; > } > - l2hdr_len = eth_get_l2_hdr_length(data); > + data += pkt->vnet_hdr_len; > > - if (pkt->size < ETH_HLEN + pkt->vnet_hdr_len) { > - trace_colo_proxy_main("pkt->size < ETH_HLEN"); > - return 1; > - } > + l2hdr_len = eth_get_l2_hdr_length(data); > > /* > * TODO: support vlan. > diff --git a/net/colo.h b/net/colo.h > index 8b3e8d5a83..22fc3031f7 100644 > --- a/net/colo.h > +++ b/net/colo.h > @@ -18,6 +18,7 @@ > #include "qemu/jhash.h" > #include "qemu/timer.h" > #include "net/eth.h" > +#include "standard-headers/linux/virtio_net.h" > > #define HASHTABLE_MAX_SIZE 16384 > > -- > 2.25.1 >
> -----Original Message----- > From: Jason Wang <jasowang@redhat.com> > Sent: Tuesday, August 9, 2022 4:12 PM > To: Zhang, Chen <chen.zhang@intel.com> > Cc: Peter Maydell <peter.maydell@linaro.org>; Li Zhijian > <lizhijian@fujitsu.com>; qemu-dev <qemu-devel@nongnu.org> > Subject: Re: [PATCH V3] net/colo.c: Fix the pointer issue reported by Coverity. > > On Tue, Aug 9, 2022 at 11:05 AM Zhang Chen <chen.zhang@intel.com> wrote: > > > > When enable the virtio-net-pci, guest network packet will load the > > vnet_hdr. In COLO status, the primary VM's network packet maybe > > redirect to another VM, it need filter-redirect enable the vnet_hdr > > flag at the same time, COLO-proxy will correctly parse the original > > network packet. If have any misconfiguration here, the vnet_hdr_len is > > wrong for parse the packet, the data+offset will point to wrong place. > > > > Signed-off-by: Zhang Chen <chen.zhang@intel.com> > > --- > > net/colo.c | 18 ++++++++++-------- > > net/colo.h | 1 + > > 2 files changed, 11 insertions(+), 8 deletions(-) > > > > diff --git a/net/colo.c b/net/colo.c > > index 6b0ff562ad..4f1b4a61f6 100644 > > --- a/net/colo.c > > +++ b/net/colo.c > > @@ -44,21 +44,23 @@ int parse_packet_early(Packet *pkt) { > > int network_length; > > static const uint8_t vlan[] = {0x81, 0x00}; > > - uint8_t *data = pkt->data + pkt->vnet_hdr_len; > > + uint8_t *data = pkt->data; > > uint16_t l3_proto; > > ssize_t l2hdr_len; > > > > - if (data == NULL) { > > - trace_colo_proxy_main_vnet_info("This packet is not parsed correctly, " > > + assert(data); > > + > > + /* Check the received vnet_hdr_len then add the offset */ > > + if ((pkt->vnet_hdr_len > sizeof(struct virtio_net_hdr)) || > > Virtio-net starts to support RSS so it means the vnet_hdr size could be greater > than virtio_net_hdr now. > > Or did you actually mean "<" here? No, I just try to avoid overflow with a maliciously over-large vnet_hdr_len value as Peter's comments. If enabled the RSS, how to get the maximum of vnet_hdr size? Thanks Chen > > Thanks > > > + (pkt->size < sizeof(struct eth_header) + sizeof(struct vlan_header) > > + + pkt->vnet_hdr_len)) { > > + trace_colo_proxy_main_vnet_info("This packet may be load wrong " > > "pkt->vnet_hdr_len", pkt->vnet_hdr_len); > > return 1; > > } > > - l2hdr_len = eth_get_l2_hdr_length(data); > > + data += pkt->vnet_hdr_len; > > > > - if (pkt->size < ETH_HLEN + pkt->vnet_hdr_len) { > > - trace_colo_proxy_main("pkt->size < ETH_HLEN"); > > - return 1; > > - } > > + l2hdr_len = eth_get_l2_hdr_length(data); > > > > /* > > * TODO: support vlan. > > diff --git a/net/colo.h b/net/colo.h > > index 8b3e8d5a83..22fc3031f7 100644 > > --- a/net/colo.h > > +++ b/net/colo.h > > @@ -18,6 +18,7 @@ > > #include "qemu/jhash.h" > > #include "qemu/timer.h" > > #include "net/eth.h" > > +#include "standard-headers/linux/virtio_net.h" > > > > #define HASHTABLE_MAX_SIZE 16384 > > > > -- > > 2.25.1 > >
On Tue, Aug 9, 2022 at 4:29 PM Zhang, Chen <chen.zhang@intel.com> wrote: > > > > > -----Original Message----- > > From: Jason Wang <jasowang@redhat.com> > > Sent: Tuesday, August 9, 2022 4:12 PM > > To: Zhang, Chen <chen.zhang@intel.com> > > Cc: Peter Maydell <peter.maydell@linaro.org>; Li Zhijian > > <lizhijian@fujitsu.com>; qemu-dev <qemu-devel@nongnu.org> > > Subject: Re: [PATCH V3] net/colo.c: Fix the pointer issue reported by Coverity. > > > > On Tue, Aug 9, 2022 at 11:05 AM Zhang Chen <chen.zhang@intel.com> wrote: > > > > > > When enable the virtio-net-pci, guest network packet will load the > > > vnet_hdr. In COLO status, the primary VM's network packet maybe > > > redirect to another VM, it need filter-redirect enable the vnet_hdr > > > flag at the same time, COLO-proxy will correctly parse the original > > > network packet. If have any misconfiguration here, the vnet_hdr_len is > > > wrong for parse the packet, the data+offset will point to wrong place. > > > > > > Signed-off-by: Zhang Chen <chen.zhang@intel.com> > > > --- > > > net/colo.c | 18 ++++++++++-------- > > > net/colo.h | 1 + > > > 2 files changed, 11 insertions(+), 8 deletions(-) > > > > > > diff --git a/net/colo.c b/net/colo.c > > > index 6b0ff562ad..4f1b4a61f6 100644 > > > --- a/net/colo.c > > > +++ b/net/colo.c > > > @@ -44,21 +44,23 @@ int parse_packet_early(Packet *pkt) { > > > int network_length; > > > static const uint8_t vlan[] = {0x81, 0x00}; > > > - uint8_t *data = pkt->data + pkt->vnet_hdr_len; > > > + uint8_t *data = pkt->data; > > > uint16_t l3_proto; > > > ssize_t l2hdr_len; > > > > > > - if (data == NULL) { > > > - trace_colo_proxy_main_vnet_info("This packet is not parsed correctly, " > > > + assert(data); > > > + > > > + /* Check the received vnet_hdr_len then add the offset */ > > > + if ((pkt->vnet_hdr_len > sizeof(struct virtio_net_hdr)) || > > > > Virtio-net starts to support RSS so it means the vnet_hdr size could be greater > > than virtio_net_hdr now. > > > > Or did you actually mean "<" here? > > No, I just try to avoid overflow with a maliciously over-large > vnet_hdr_len value as Peter's comments. > If enabled the RSS, how to get the maximum of vnet_hdr size? With hash_report the maximum is sizeof(virtio_net_hdr_v1_hash). But it might be extended in the future. We can probably start from it. Thanks > > Thanks > Chen > > > > > Thanks > > > > > + (pkt->size < sizeof(struct eth_header) + sizeof(struct vlan_header) > > > + + pkt->vnet_hdr_len)) { > > > + trace_colo_proxy_main_vnet_info("This packet may be load wrong " > > > "pkt->vnet_hdr_len", pkt->vnet_hdr_len); > > > return 1; > > > } > > > - l2hdr_len = eth_get_l2_hdr_length(data); > > > + data += pkt->vnet_hdr_len; > > > > > > - if (pkt->size < ETH_HLEN + pkt->vnet_hdr_len) { > > > - trace_colo_proxy_main("pkt->size < ETH_HLEN"); > > > - return 1; > > > - } > > > + l2hdr_len = eth_get_l2_hdr_length(data); > > > > > > /* > > > * TODO: support vlan. > > > diff --git a/net/colo.h b/net/colo.h > > > index 8b3e8d5a83..22fc3031f7 100644 > > > --- a/net/colo.h > > > +++ b/net/colo.h > > > @@ -18,6 +18,7 @@ > > > #include "qemu/jhash.h" > > > #include "qemu/timer.h" > > > #include "net/eth.h" > > > +#include "standard-headers/linux/virtio_net.h" > > > > > > #define HASHTABLE_MAX_SIZE 16384 > > > > > > -- > > > 2.25.1 > > > >
> -----Original Message----- > From: Jason Wang <jasowang@redhat.com> > Sent: Tuesday, August 9, 2022 4:43 PM > To: Zhang, Chen <chen.zhang@intel.com> > Cc: Peter Maydell <peter.maydell@linaro.org>; Li Zhijian > <lizhijian@fujitsu.com>; qemu-dev <qemu-devel@nongnu.org> > Subject: Re: [PATCH V3] net/colo.c: Fix the pointer issue reported by Coverity. > > On Tue, Aug 9, 2022 at 4:29 PM Zhang, Chen <chen.zhang@intel.com> wrote: > > > > > > > > > -----Original Message----- > > > From: Jason Wang <jasowang@redhat.com> > > > Sent: Tuesday, August 9, 2022 4:12 PM > > > To: Zhang, Chen <chen.zhang@intel.com> > > > Cc: Peter Maydell <peter.maydell@linaro.org>; Li Zhijian > > > <lizhijian@fujitsu.com>; qemu-dev <qemu-devel@nongnu.org> > > > Subject: Re: [PATCH V3] net/colo.c: Fix the pointer issue reported by > Coverity. > > > > > > On Tue, Aug 9, 2022 at 11:05 AM Zhang Chen <chen.zhang@intel.com> > wrote: > > > > > > > > When enable the virtio-net-pci, guest network packet will load the > > > > vnet_hdr. In COLO status, the primary VM's network packet maybe > > > > redirect to another VM, it need filter-redirect enable the > > > > vnet_hdr flag at the same time, COLO-proxy will correctly parse > > > > the original network packet. If have any misconfiguration here, > > > > the vnet_hdr_len is wrong for parse the packet, the data+offset will point > to wrong place. > > > > > > > > Signed-off-by: Zhang Chen <chen.zhang@intel.com> > > > > --- > > > > net/colo.c | 18 ++++++++++-------- net/colo.h | 1 + > > > > 2 files changed, 11 insertions(+), 8 deletions(-) > > > > > > > > diff --git a/net/colo.c b/net/colo.c index 6b0ff562ad..4f1b4a61f6 > > > > 100644 > > > > --- a/net/colo.c > > > > +++ b/net/colo.c > > > > @@ -44,21 +44,23 @@ int parse_packet_early(Packet *pkt) { > > > > int network_length; > > > > static const uint8_t vlan[] = {0x81, 0x00}; > > > > - uint8_t *data = pkt->data + pkt->vnet_hdr_len; > > > > + uint8_t *data = pkt->data; > > > > uint16_t l3_proto; > > > > ssize_t l2hdr_len; > > > > > > > > - if (data == NULL) { > > > > - trace_colo_proxy_main_vnet_info("This packet is not parsed > correctly, " > > > > + assert(data); > > > > + > > > > + /* Check the received vnet_hdr_len then add the offset */ > > > > + if ((pkt->vnet_hdr_len > sizeof(struct virtio_net_hdr)) || > > > > > > Virtio-net starts to support RSS so it means the vnet_hdr size could > > > be greater than virtio_net_hdr now. > > > > > > Or did you actually mean "<" here? > > > > No, I just try to avoid overflow with a maliciously over-large > > vnet_hdr_len value as Peter's comments. > > If enabled the RSS, how to get the maximum of vnet_hdr size? > > With hash_report the maximum is sizeof(virtio_net_hdr_v1_hash). But it might > be extended in the future. > > We can probably start from it. Thanks, I will change here to " if ((pkt->vnet_hdr_len > sizeof(virtio_net_hdr_v1_hash)) ||" in next version. Thanks Chen > > Thanks > > > > > Thanks > > Chen > > > > > > > > Thanks > > > > > > > + (pkt->size < sizeof(struct eth_header) + sizeof(struct vlan_header) > > > > + + pkt->vnet_hdr_len)) { > > > > + trace_colo_proxy_main_vnet_info("This packet may be load wrong > " > > > > "pkt->vnet_hdr_len", pkt->vnet_hdr_len); > > > > return 1; > > > > } > > > > - l2hdr_len = eth_get_l2_hdr_length(data); > > > > + data += pkt->vnet_hdr_len; > > > > > > > > - if (pkt->size < ETH_HLEN + pkt->vnet_hdr_len) { > > > > - trace_colo_proxy_main("pkt->size < ETH_HLEN"); > > > > - return 1; > > > > - } > > > > + l2hdr_len = eth_get_l2_hdr_length(data); > > > > > > > > /* > > > > * TODO: support vlan. > > > > diff --git a/net/colo.h b/net/colo.h index 8b3e8d5a83..22fc3031f7 > > > > 100644 > > > > --- a/net/colo.h > > > > +++ b/net/colo.h > > > > @@ -18,6 +18,7 @@ > > > > #include "qemu/jhash.h" > > > > #include "qemu/timer.h" > > > > #include "net/eth.h" > > > > +#include "standard-headers/linux/virtio_net.h" > > > > > > > > #define HASHTABLE_MAX_SIZE 16384 > > > > > > > > -- > > > > 2.25.1 > > > > > >
diff --git a/net/colo.c b/net/colo.c index 6b0ff562ad..4f1b4a61f6 100644 --- a/net/colo.c +++ b/net/colo.c @@ -44,21 +44,23 @@ int parse_packet_early(Packet *pkt) { int network_length; static const uint8_t vlan[] = {0x81, 0x00}; - uint8_t *data = pkt->data + pkt->vnet_hdr_len; + uint8_t *data = pkt->data; uint16_t l3_proto; ssize_t l2hdr_len; - if (data == NULL) { - trace_colo_proxy_main_vnet_info("This packet is not parsed correctly, " + assert(data); + + /* Check the received vnet_hdr_len then add the offset */ + if ((pkt->vnet_hdr_len > sizeof(struct virtio_net_hdr)) || + (pkt->size < sizeof(struct eth_header) + sizeof(struct vlan_header) + + pkt->vnet_hdr_len)) { + trace_colo_proxy_main_vnet_info("This packet may be load wrong " "pkt->vnet_hdr_len", pkt->vnet_hdr_len); return 1; } - l2hdr_len = eth_get_l2_hdr_length(data); + data += pkt->vnet_hdr_len; - if (pkt->size < ETH_HLEN + pkt->vnet_hdr_len) { - trace_colo_proxy_main("pkt->size < ETH_HLEN"); - return 1; - } + l2hdr_len = eth_get_l2_hdr_length(data); /* * TODO: support vlan. diff --git a/net/colo.h b/net/colo.h index 8b3e8d5a83..22fc3031f7 100644 --- a/net/colo.h +++ b/net/colo.h @@ -18,6 +18,7 @@ #include "qemu/jhash.h" #include "qemu/timer.h" #include "net/eth.h" +#include "standard-headers/linux/virtio_net.h" #define HASHTABLE_MAX_SIZE 16384
When enable the virtio-net-pci, guest network packet will load the vnet_hdr. In COLO status, the primary VM's network packet maybe redirect to another VM, it need filter-redirect enable the vnet_hdr flag at the same time, COLO-proxy will correctly parse the original network packet. If have any misconfiguration here, the vnet_hdr_len is wrong for parse the packet, the data+offset will point to wrong place. Signed-off-by: Zhang Chen <chen.zhang@intel.com> --- net/colo.c | 18 ++++++++++-------- net/colo.h | 1 + 2 files changed, 11 insertions(+), 8 deletions(-)