| Message ID | 20220420200720.434717-4-parri.andrea@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | hv_sock: Hardening changes | expand |
From: Andrea Parri (Microsoft) <parri.andrea@gmail.com> Sent: Wednesday, April 20, 2022 1:07 PM > > For additional robustness in the face of Hyper-V errors or malicious > behavior, validate all values that originate from packets that Hyper-V > has sent to the guest in the host-to-guest ring buffer. Ensure that > invalid values cannot cause data being copied out of the bounds of the > source buffer in hvs_stream_dequeue(). > > Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com> > --- > include/linux/hyperv.h | 5 +++++ > net/vmw_vsock/hyperv_transport.c | 11 +++++++++-- > 2 files changed, 14 insertions(+), 2 deletions(-) > > diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h > index fe2e0179ed51e..55478a6810b60 100644 > --- a/include/linux/hyperv.h > +++ b/include/linux/hyperv.h > @@ -1663,6 +1663,11 @@ static inline u32 hv_pkt_datalen(const struct > vmpacket_descriptor *desc) > return (desc->len8 << 3) - (desc->offset8 << 3); > } > > +/* Get packet length associated with descriptor */ > +static inline u32 hv_pkt_len(const struct vmpacket_descriptor *desc) > +{ > + return desc->len8 << 3; > +} > > struct vmpacket_descriptor * > hv_pkt_iter_first_raw(struct vmbus_channel *channel); > diff --git a/net/vmw_vsock/hyperv_transport.c b/net/vmw_vsock/hyperv_transport.c > index 8c37d07017fc4..092cadc2c866d 100644 > --- a/net/vmw_vsock/hyperv_transport.c > +++ b/net/vmw_vsock/hyperv_transport.c > @@ -577,12 +577,19 @@ static bool hvs_dgram_allow(u32 cid, u32 port) > static int hvs_update_recv_data(struct hvsock *hvs) > { > struct hvs_recv_buf *recv_buf; > - u32 payload_len; > + u32 pkt_len, payload_len; > + > + pkt_len = hv_pkt_len(hvs->recv_desc); > + > + /* Ensure the packet is big enough to read its header */ > + if (pkt_len < HVS_HEADER_LEN) > + return -EIO; > > recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); > payload_len = recv_buf->hdr.data_size; > > - if (payload_len > HVS_MTU_SIZE) > + /* Ensure the packet is big enough to read its payload */ > + if (payload_len > pkt_len - HVS_HEADER_LEN || payload_len > HVS_MTU_SIZE) > return -EIO; > > if (payload_len == 0) > -- > 2.25.1 Reviewed-by: Michael Kelley <mikelley@microsoft.com>
On Wed, Apr 20, 2022 at 10:07:18PM +0200, Andrea Parri (Microsoft) wrote: >For additional robustness in the face of Hyper-V errors or malicious >behavior, validate all values that originate from packets that Hyper-V >has sent to the guest in the host-to-guest ring buffer. Ensure that >invalid values cannot cause data being copied out of the bounds of the >source buffer in hvs_stream_dequeue(). > >Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com> >--- > include/linux/hyperv.h | 5 +++++ > net/vmw_vsock/hyperv_transport.c | 11 +++++++++-- > 2 files changed, 14 insertions(+), 2 deletions(-) > >diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h >index fe2e0179ed51e..55478a6810b60 100644 >--- a/include/linux/hyperv.h >+++ b/include/linux/hyperv.h >@@ -1663,6 +1663,11 @@ static inline u32 hv_pkt_datalen(const struct vmpacket_descriptor *desc) > return (desc->len8 << 3) - (desc->offset8 << 3); > } > >+/* Get packet length associated with descriptor */ >+static inline u32 hv_pkt_len(const struct vmpacket_descriptor *desc) >+{ >+ return desc->len8 << 3; >+} > > struct vmpacket_descriptor * > hv_pkt_iter_first_raw(struct vmbus_channel *channel); >diff --git a/net/vmw_vsock/hyperv_transport.c b/net/vmw_vsock/hyperv_transport.c >index 8c37d07017fc4..092cadc2c866d 100644 >--- a/net/vmw_vsock/hyperv_transport.c >+++ b/net/vmw_vsock/hyperv_transport.c >@@ -577,12 +577,19 @@ static bool hvs_dgram_allow(u32 cid, u32 port) > static int hvs_update_recv_data(struct hvsock *hvs) > { > struct hvs_recv_buf *recv_buf; >- u32 payload_len; >+ u32 pkt_len, payload_len; >+ >+ pkt_len = hv_pkt_len(hvs->recv_desc); >+ >+ /* Ensure the packet is big enough to read its header */ >+ if (pkt_len < HVS_HEADER_LEN) >+ return -EIO; > > recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); > payload_len = recv_buf->hdr.data_size; > >- if (payload_len > HVS_MTU_SIZE) >+ /* Ensure the packet is big enough to read its payload */ >+ if (payload_len > pkt_len - HVS_HEADER_LEN || payload_len > HVS_MTU_SIZE) checkpatch warns that we exceed 80 characters, I do not have a strong opinion on this, but if you have to resend better break the condition into 2 lines. Maybe even update or remove the comment? (it only describes the first condition, but the conditions are pretty clear, so I don't think it adds much). Thanks, Stefano
> > @@ -577,12 +577,19 @@ static bool hvs_dgram_allow(u32 cid, u32 port) > > static int hvs_update_recv_data(struct hvsock *hvs) > > { > > struct hvs_recv_buf *recv_buf; > > - u32 payload_len; > > + u32 pkt_len, payload_len; > > + > > + pkt_len = hv_pkt_len(hvs->recv_desc); > > + > > + /* Ensure the packet is big enough to read its header */ > > + if (pkt_len < HVS_HEADER_LEN) > > + return -EIO; > > > > recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); > > payload_len = recv_buf->hdr.data_size; > > > > - if (payload_len > HVS_MTU_SIZE) > > + /* Ensure the packet is big enough to read its payload */ > > + if (payload_len > pkt_len - HVS_HEADER_LEN || payload_len > HVS_MTU_SIZE) > > checkpatch warns that we exceed 80 characters, I do not have a strong > opinion on this, but if you have to resend better break the condition into 2 > lines. Will break if preferred. (but does it really warn?? I understand that the warning was deprecated and the "limit" increased to 100 chars...) > Maybe even update or remove the comment? (it only describes the first > condition, but the conditions are pretty clear, so I don't think it adds > much). Works for me. (taking it as this applies to the previous comment too.) Thanks, Andrea
On Thu, Apr 21, 2022 at 5:30 PM Andrea Parri <parri.andrea@gmail.com> wrote: > > > > @@ -577,12 +577,19 @@ static bool hvs_dgram_allow(u32 cid, u32 port) > > > static int hvs_update_recv_data(struct hvsock *hvs) > > > { > > > struct hvs_recv_buf *recv_buf; > > > - u32 payload_len; > > > + u32 pkt_len, payload_len; > > > + > > > + pkt_len = hv_pkt_len(hvs->recv_desc); > > > + > > > + /* Ensure the packet is big enough to read its header */ > > > + if (pkt_len < HVS_HEADER_LEN) > > > + return -EIO; > > > > > > recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); > > > payload_len = recv_buf->hdr.data_size; > > > > > > - if (payload_len > HVS_MTU_SIZE) > > > + /* Ensure the packet is big enough to read its payload */ > > > + if (payload_len > pkt_len - HVS_HEADER_LEN || payload_len > HVS_MTU_SIZE) > > > > checkpatch warns that we exceed 80 characters, I do not have a strong > > opinion on this, but if you have to resend better break the condition into 2 > > lines. > > Will break if preferred. (but does it really warn?? I understand that > the warning was deprecated and the "limit" increased to 100 chars...) I see the warn here: https://patchwork.kernel.org/project/netdevbpf/patch/20220420200720.434717-4-parri.andrea@gmail.com/ in the kernel doc [1] we still say we prefer 80 columns, so I try to follow, especially when it doesn't make things worse. [1] https://docs.kernel.org/process/coding-style.html#breaking-long-lines-and-strings > > > > Maybe even update or remove the comment? (it only describes the first > > condition, but the conditions are pretty clear, so I don't think it adds > > much). > > Works for me. (taking it as this applies to the previous comment too.) Yep. Thanks, Stefano
diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h index fe2e0179ed51e..55478a6810b60 100644 --- a/include/linux/hyperv.h +++ b/include/linux/hyperv.h @@ -1663,6 +1663,11 @@ static inline u32 hv_pkt_datalen(const struct vmpacket_descriptor *desc) return (desc->len8 << 3) - (desc->offset8 << 3); } +/* Get packet length associated with descriptor */ +static inline u32 hv_pkt_len(const struct vmpacket_descriptor *desc) +{ + return desc->len8 << 3; +} struct vmpacket_descriptor * hv_pkt_iter_first_raw(struct vmbus_channel *channel); diff --git a/net/vmw_vsock/hyperv_transport.c b/net/vmw_vsock/hyperv_transport.c index 8c37d07017fc4..092cadc2c866d 100644 --- a/net/vmw_vsock/hyperv_transport.c +++ b/net/vmw_vsock/hyperv_transport.c @@ -577,12 +577,19 @@ static bool hvs_dgram_allow(u32 cid, u32 port) static int hvs_update_recv_data(struct hvsock *hvs) { struct hvs_recv_buf *recv_buf; - u32 payload_len; + u32 pkt_len, payload_len; + + pkt_len = hv_pkt_len(hvs->recv_desc); + + /* Ensure the packet is big enough to read its header */ + if (pkt_len < HVS_HEADER_LEN) + return -EIO; recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); payload_len = recv_buf->hdr.data_size; - if (payload_len > HVS_MTU_SIZE) + /* Ensure the packet is big enough to read its payload */ + if (payload_len > pkt_len - HVS_HEADER_LEN || payload_len > HVS_MTU_SIZE) return -EIO; if (payload_len == 0)
For additional robustness in the face of Hyper-V errors or malicious behavior, validate all values that originate from packets that Hyper-V has sent to the guest in the host-to-guest ring buffer. Ensure that invalid values cannot cause data being copied out of the bounds of the source buffer in hvs_stream_dequeue(). Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com> --- include/linux/hyperv.h | 5 +++++ net/vmw_vsock/hyperv_transport.c | 11 +++++++++-- 2 files changed, 14 insertions(+), 2 deletions(-)