From: Jason Wang <jasowang@redhat.com> To: mst@redhat.com Cc: virtualization@lists.linux-foundation.org, linux-kernel@vger.kernel.org, xieyongji@bytedance.com, stefanha@redhat.com, file@sect.tu-berlin.de, ashish.kalra@amd.com, konrad.wilk@oracle.com, kvm@vger.kernel.org, hch@infradead.org, ak@linux.intel.com, luto@kernel.org, Jason Wang <jasowang@redhat.com> Subject: [PATCH 0/7] Do not read from descriptor ring Date: Fri, 4 Jun 2021 13:53:43 +0800 [thread overview] Message-ID: <20210604055350.58753-1-jasowang@redhat.com> (raw) Hi: The virtio driver should not trust the device. This beame more urgent for the case of encrtpyed VM or VDUSE[1]. In both cases, technology like swiotlb/IOMMU is used to prevent the poking/mangling of memory from the device. But this is not sufficient since current virtio driver may trust what is stored in the descriptor table (coherent mapping) for performing the DMA operations like unmap and bounce so the device may choose to utilize the behaviour of swiotlb to perform attacks[2]. To protect from a malicous device, this series store and use the descriptor metadata in an auxiliay structure which can not be accessed via swiotlb/device instead of the ones in the descriptor table. This means the descriptor table is write-only from the view of the driver. Actually, we've almost achieved that through packed virtqueue and we just need to fix a corner case of handling mapping errors. For split virtqueue we just follow what's done in the packed. Note that we don't duplicate descriptor medata for indirect descriptors since it uses stream mapping which is read only so it's safe if the metadata of non-indirect descriptors are correct. For split virtqueue, the change increase the footprint due the the auxiliary metadata but it's almost neglectlable in simple test like pktgen and netperf TCP stream (slightly noticed in a 40GBE environment with more CPU usage). Slightly tested with packed on/off, iommu on/of, swiotlb force/off in the guest. Note that this series tries to fix the attack via descriptor ring. The other cases (used ring and config space) will be fixed by other series or patches. Please review. Changes from RFC V2: - no code change - twaeak the commit log a little bit Changes from RFC V1: - Always use auxiliary metadata for split virtqueue - Don't read from descripto when detaching indirect descriptor Jason Wang (7): virtio-ring: maintain next in extra state for packed virtqueue virtio_ring: rename vring_desc_extra_packed virtio-ring: factor out desc_extra allocation virtio_ring: secure handling of mapping errors virtio_ring: introduce virtqueue_desc_add_split() virtio: use err label in __vring_new_virtqueue() virtio-ring: store DMA metadata in desc_extra for split virtqueue drivers/virtio/virtio_ring.c | 201 +++++++++++++++++++++++++---------- 1 file changed, 144 insertions(+), 57 deletions(-) -- 2.25.1
WARNING: multiple messages have this Message-ID (diff)
From: Jason Wang <jasowang@redhat.com> To: mst@redhat.com Cc: ashish.kalra@amd.com, ak@linux.intel.com, file@sect.tu-berlin.de, luto@kernel.org, kvm@vger.kernel.org, konrad.wilk@oracle.com, linux-kernel@vger.kernel.org, virtualization@lists.linux-foundation.org, hch@infradead.org, xieyongji@bytedance.com, stefanha@redhat.com Subject: [PATCH 0/7] Do not read from descriptor ring Date: Fri, 4 Jun 2021 13:53:43 +0800 [thread overview] Message-ID: <20210604055350.58753-1-jasowang@redhat.com> (raw) Hi: The virtio driver should not trust the device. This beame more urgent for the case of encrtpyed VM or VDUSE[1]. In both cases, technology like swiotlb/IOMMU is used to prevent the poking/mangling of memory from the device. But this is not sufficient since current virtio driver may trust what is stored in the descriptor table (coherent mapping) for performing the DMA operations like unmap and bounce so the device may choose to utilize the behaviour of swiotlb to perform attacks[2]. To protect from a malicous device, this series store and use the descriptor metadata in an auxiliay structure which can not be accessed via swiotlb/device instead of the ones in the descriptor table. This means the descriptor table is write-only from the view of the driver. Actually, we've almost achieved that through packed virtqueue and we just need to fix a corner case of handling mapping errors. For split virtqueue we just follow what's done in the packed. Note that we don't duplicate descriptor medata for indirect descriptors since it uses stream mapping which is read only so it's safe if the metadata of non-indirect descriptors are correct. For split virtqueue, the change increase the footprint due the the auxiliary metadata but it's almost neglectlable in simple test like pktgen and netperf TCP stream (slightly noticed in a 40GBE environment with more CPU usage). Slightly tested with packed on/off, iommu on/of, swiotlb force/off in the guest. Note that this series tries to fix the attack via descriptor ring. The other cases (used ring and config space) will be fixed by other series or patches. Please review. Changes from RFC V2: - no code change - twaeak the commit log a little bit Changes from RFC V1: - Always use auxiliary metadata for split virtqueue - Don't read from descripto when detaching indirect descriptor Jason Wang (7): virtio-ring: maintain next in extra state for packed virtqueue virtio_ring: rename vring_desc_extra_packed virtio-ring: factor out desc_extra allocation virtio_ring: secure handling of mapping errors virtio_ring: introduce virtqueue_desc_add_split() virtio: use err label in __vring_new_virtqueue() virtio-ring: store DMA metadata in desc_extra for split virtqueue drivers/virtio/virtio_ring.c | 201 +++++++++++++++++++++++++---------- 1 file changed, 144 insertions(+), 57 deletions(-) -- 2.25.1 _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
next reply other threads:[~2021-06-04 5:54 UTC|newest] Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-06-04 5:53 Jason Wang [this message] 2021-06-04 5:53 ` [PATCH 0/7] Do not read from descriptor ring Jason Wang 2021-06-04 5:53 ` [PATCH 1/7] virtio-ring: maintain next in extra state for packed virtqueue Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-04 5:53 ` [PATCH 2/7] virtio_ring: rename vring_desc_extra_packed Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-04 5:53 ` [PATCH 3/7] virtio-ring: factor out desc_extra allocation Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-04 5:53 ` [PATCH 4/7] virtio_ring: secure handling of mapping errors Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-04 5:53 ` [PATCH 5/7] virtio_ring: introduce virtqueue_desc_add_split() Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-04 5:53 ` [PATCH 6/7] virtio: use err label in __vring_new_virtqueue() Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-04 5:53 ` [PATCH 7/7] virtio-ring: store DMA metadata in desc_extra for split virtqueue Jason Wang 2021-06-04 5:53 ` Jason Wang 2021-06-08 16:24 ` [PATCH 0/7] Do not read from descriptor ring Andy Lutomirski 2021-06-08 16:24 ` Andy Lutomirski 2021-06-10 3:12 ` Jason Wang 2021-06-10 3:12 ` Jason Wang 2021-07-11 16:08 ` Michael S. Tsirkin 2021-07-11 16:08 ` Michael S. Tsirkin
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20210604055350.58753-1-jasowang@redhat.com \ --to=jasowang@redhat.com \ --cc=ak@linux.intel.com \ --cc=ashish.kalra@amd.com \ --cc=file@sect.tu-berlin.de \ --cc=hch@infradead.org \ --cc=konrad.wilk@oracle.com \ --cc=kvm@vger.kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=luto@kernel.org \ --cc=mst@redhat.com \ --cc=stefanha@redhat.com \ --cc=virtualization@lists.linux-foundation.org \ --cc=xieyongji@bytedance.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.