RE: [PATCH V4 12/13] hv_netvsc: Add Isolation VM support for netvsc driver

[Michael Kelley <mikelley@microsoft.com>]

From: Tianyu Lan <ltykernel@gmail.com> Sent: Friday, August 27, 2021 10:21 AM
> 
> In Isolation VM, all shared memory with host needs to mark visible
> to host via hvcall. vmbus_establish_gpadl() has already done it for
> netvsc rx/tx ring buffer. The page buffer used by vmbus_sendpacket_
> pagebuffer() stills need to be handled. Use DMA API to map/umap
> these memory during sending/receiving packet and Hyper-V swiotlb
> bounce buffer dma adress will be returned. The swiotlb bounce buffer
> has been masked to be visible to host during boot up.
> 
> Signed-off-by: Tianyu Lan <Tianyu.Lan@microsoft.com>
> ---
> Change since v3:
> 	* Add comment to explain why not to use dma_map_sg()
> 	* Fix some error handle.
> ---
>  arch/x86/hyperv/ivm.c             |   1 +
>  drivers/net/hyperv/hyperv_net.h   |   5 ++
>  drivers/net/hyperv/netvsc.c       | 135 +++++++++++++++++++++++++++++-
>  drivers/net/hyperv/rndis_filter.c |   2 +
>  include/linux/hyperv.h            |   5 ++
>  5 files changed, 145 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/hyperv/ivm.c b/arch/x86/hyperv/ivm.c
> index 84563b3c9f3a..08d8e01de017 100644
> --- a/arch/x86/hyperv/ivm.c
> +++ b/arch/x86/hyperv/ivm.c
> @@ -317,6 +317,7 @@ void *hv_map_memory(void *addr, unsigned long size)
> 
>  	return vaddr;
>  }
> +EXPORT_SYMBOL_GPL(hv_map_memory);
> 
>  void hv_unmap_memory(void *addr)
>  {
> diff --git a/drivers/net/hyperv/hyperv_net.h b/drivers/net/hyperv/hyperv_net.h
> index aa7c9962dbd8..862419912bfb 100644
> --- a/drivers/net/hyperv/hyperv_net.h
> +++ b/drivers/net/hyperv/hyperv_net.h
> @@ -164,6 +164,7 @@ struct hv_netvsc_packet {
>  	u32 total_bytes;
>  	u32 send_buf_index;
>  	u32 total_data_buflen;
> +	struct hv_dma_range *dma_range;
>  };
> 
>  #define NETVSC_HASH_KEYLEN 40
> @@ -1074,6 +1075,7 @@ struct netvsc_device {
> 
>  	/* Receive buffer allocated by us but manages by NetVSP */
>  	void *recv_buf;
> +	void *recv_original_buf;
>  	u32 recv_buf_size; /* allocated bytes */
>  	u32 recv_buf_gpadl_handle;
>  	u32 recv_section_cnt;
> @@ -1082,6 +1084,7 @@ struct netvsc_device {
> 
>  	/* Send buffer allocated by us */
>  	void *send_buf;
> +	void *send_original_buf;
>  	u32 send_buf_size;
>  	u32 send_buf_gpadl_handle;
>  	u32 send_section_cnt;
> @@ -1731,4 +1734,6 @@ struct rndis_message {
>  #define RETRY_US_HI	10000
>  #define RETRY_MAX	2000	/* >10 sec */
> 
> +void netvsc_dma_unmap(struct hv_device *hv_dev,
> +		      struct hv_netvsc_packet *packet);
>  #endif /* _HYPERV_NET_H */
> diff --git a/drivers/net/hyperv/netvsc.c b/drivers/net/hyperv/netvsc.c
> index f19bffff6a63..edd336b08c2c 100644
> --- a/drivers/net/hyperv/netvsc.c
> +++ b/drivers/net/hyperv/netvsc.c
> @@ -153,8 +153,21 @@ static void free_netvsc_device(struct rcu_head *head)
>  	int i;
> 
>  	kfree(nvdev->extension);
> -	vfree(nvdev->recv_buf);
> -	vfree(nvdev->send_buf);
> +
> +	if (nvdev->recv_original_buf) {
> +		vunmap(nvdev->recv_buf);

In patch 11, you have added a hv_unmap_memory()
function as the inverse of hv_map_memory().  Since this
buffer was mapped with hv_map_memory() and you have
added that function, the cleanup should use
hv_unmap_memory() rather than calling vunmap() directly.

> +		vfree(nvdev->recv_original_buf);
> +	} else {
> +		vfree(nvdev->recv_buf);
> +	}
> +
> +	if (nvdev->send_original_buf) {
> +		vunmap(nvdev->send_buf);

Same here.

> +		vfree(nvdev->send_original_buf);
> +	} else {
> +		vfree(nvdev->send_buf);
> +	}
> +
>  	kfree(nvdev->send_section_map);
> 
>  	for (i = 0; i < VRSS_CHANNEL_MAX; i++) {
> @@ -347,6 +360,7 @@ static int netvsc_init_buf(struct hv_device *device,
>  	unsigned int buf_size;
>  	size_t map_words;
>  	int i, ret = 0;
> +	void *vaddr;
> 
>  	/* Get receive buffer area. */
>  	buf_size = device_info->recv_sections * device_info->recv_section_size;
> @@ -382,6 +396,17 @@ static int netvsc_init_buf(struct hv_device *device,
>  		goto cleanup;
>  	}
> 
> +	if (hv_isolation_type_snp()) {
> +		vaddr = hv_map_memory(net_device->recv_buf, buf_size);

Since the netvsc driver is architecture neutral, this code also needs
to compile for ARM64.  A stub will be needed for hv_map_memory()
on the ARM64 side.  Same for hv_unmap_memory() as suggested
above.  Or better, move hv_map_memory() and hv_unmap_memory()
to an architecture neutral module such as hv_common.c.

Or if Christop's approach of creating the vmap_phys_addr() helper
comes to fruition, that's an even better approach since it will already
handle multiple architectures.

> +		if (!vaddr) {
> +			ret = -ENOMEM;
> +			goto cleanup;
> +		}
> +
> +		net_device->recv_original_buf = net_device->recv_buf;
> +		net_device->recv_buf = vaddr;
> +	}
> +
>  	/* Notify the NetVsp of the gpadl handle */
>  	init_packet = &net_device->channel_init_pkt;
>  	memset(init_packet, 0, sizeof(struct nvsp_message));
> @@ -485,6 +510,17 @@ static int netvsc_init_buf(struct hv_device *device,
>  		goto cleanup;
>  	}
> 
> +	if (hv_isolation_type_snp()) {
> +		vaddr = hv_map_memory(net_device->send_buf, buf_size);
> +		if (!vaddr) {
> +			ret = -ENOMEM;
> +			goto cleanup;
> +		}
> +
> +		net_device->send_original_buf = net_device->send_buf;
> +		net_device->send_buf = vaddr;
> +	}
> +
>  	/* Notify the NetVsp of the gpadl handle */
>  	init_packet = &net_device->channel_init_pkt;
>  	memset(init_packet, 0, sizeof(struct nvsp_message));
> @@ -775,7 +811,7 @@ static void netvsc_send_tx_complete(struct net_device *ndev,
> 
>  	/* Notify the layer above us */
>  	if (likely(skb)) {
> -		const struct hv_netvsc_packet *packet
> +		struct hv_netvsc_packet *packet
>  			= (struct hv_netvsc_packet *)skb->cb;
>  		u32 send_index = packet->send_buf_index;
>  		struct netvsc_stats *tx_stats;
> @@ -791,6 +827,7 @@ static void netvsc_send_tx_complete(struct net_device *ndev,
>  		tx_stats->bytes += packet->total_bytes;
>  		u64_stats_update_end(&tx_stats->syncp);
> 
> +		netvsc_dma_unmap(ndev_ctx->device_ctx, packet);
>  		napi_consume_skb(skb, budget);
>  	}
> 
> @@ -955,6 +992,87 @@ static void netvsc_copy_to_send_buf(struct netvsc_device *net_device,
>  		memset(dest, 0, padding);
>  }
> 
> +void netvsc_dma_unmap(struct hv_device *hv_dev,
> +		      struct hv_netvsc_packet *packet)
> +{
> +	u32 page_count = packet->cp_partial ?
> +		packet->page_buf_cnt - packet->rmsg_pgcnt :
> +		packet->page_buf_cnt;
> +	int i;
> +
> +	if (!hv_is_isolation_supported())
> +		return;
> +
> +	if (!packet->dma_range)
> +		return;
> +
> +	for (i = 0; i < page_count; i++)
> +		dma_unmap_single(&hv_dev->device, packet->dma_range[i].dma,
> +				 packet->dma_range[i].mapping_size,
> +				 DMA_TO_DEVICE);
> +
> +	kfree(packet->dma_range);
> +}
> +
> +/* netvsc_dma_map - Map swiotlb bounce buffer with data page of
> + * packet sent by vmbus_sendpacket_pagebuffer() in the Isolation
> + * VM.
> + *
> + * In isolation VM, netvsc send buffer has been marked visible to
> + * host and so the data copied to send buffer doesn't need to use
> + * bounce buffer. The data pages handled by vmbus_sendpacket_pagebuffer()
> + * may not be copied to send buffer and so these pages need to be
> + * mapped with swiotlb bounce buffer. netvsc_dma_map() is to do
> + * that. The pfns in the struct hv_page_buffer need to be converted
> + * to bounce buffer's pfn. The loop here is necessary becuase the

s/becuase/because/

> + * entries in the page buffer array are not necessarily full
> + * pages of data.  Each entry in the array has a separate offset and
> + * len that may be non-zero, even for entries in the middle of the
> + * array.  And the entries are not physically contiguous.  So each
> + * entry must be individually mapped rather than as a contiguous unit.
> + * So not use dma_map_sg() here.
> + */
> +int netvsc_dma_map(struct hv_device *hv_dev,
> +		   struct hv_netvsc_packet *packet,
> +		   struct hv_page_buffer *pb)
> +{
> +	u32 page_count =  packet->cp_partial ?
> +		packet->page_buf_cnt - packet->rmsg_pgcnt :
> +		packet->page_buf_cnt;
> +	dma_addr_t dma;
> +	int i;
> +
> +	if (!hv_is_isolation_supported())
> +		return 0;
> +
> +	packet->dma_range = kcalloc(page_count,
> +				    sizeof(*packet->dma_range),
> +				    GFP_KERNEL);
> +	if (!packet->dma_range)
> +		return -ENOMEM;
> +
> +	for (i = 0; i < page_count; i++) {
> +		char *src = phys_to_virt((pb[i].pfn << HV_HYP_PAGE_SHIFT)
> +					 + pb[i].offset);
> +		u32 len = pb[i].len;
> +
> +		dma = dma_map_single(&hv_dev->device, src, len,
> +				     DMA_TO_DEVICE);
> +		if (dma_mapping_error(&hv_dev->device, dma)) {
> +			kfree(packet->dma_range);
> +			return -ENOMEM;
> +		}
> +
> +		packet->dma_range[i].dma = dma;
> +		packet->dma_range[i].mapping_size = len;
> +		pb[i].pfn = dma >> HV_HYP_PAGE_SHIFT;
> +		pb[i].offset = offset_in_hvpage(dma);
> +		pb[i].len = len;
> +	}

Just to confirm, this driver does *not* set the DMA min_align_mask
like storvsc does.  So after the call to dma_map_single(), the offset
in the page could be different.  That's why you are updating
the pb[i].offset value.  Alternatively, you could set the DMA
min_align_mask, which would ensure the offset is unchanged.
I'm OK with either approach, though perhaps a comment is
warranted to explain, as this is a subtle issue.

> +
> +	return 0;
> +}
> +
>  static inline int netvsc_send_pkt(
>  	struct hv_device *device,
>  	struct hv_netvsc_packet *packet,
> @@ -995,14 +1113,24 @@ static inline int netvsc_send_pkt(
> 
>  	trace_nvsp_send_pkt(ndev, out_channel, rpkt);
> 
> +	packet->dma_range = NULL;
>  	if (packet->page_buf_cnt) {
>  		if (packet->cp_partial)
>  			pb += packet->rmsg_pgcnt;
> 
> +		ret = netvsc_dma_map(ndev_ctx->device_ctx, packet, pb);
> +		if (ret) {
> +			ret = -EAGAIN;
> +			goto exit;
> +		}
> +
>  		ret = vmbus_sendpacket_pagebuffer(out_channel,
>  						  pb, packet->page_buf_cnt,
>  						  &nvmsg, sizeof(nvmsg),
>  						  req_id);
> +
> +		if (ret)
> +			netvsc_dma_unmap(ndev_ctx->device_ctx, packet);
>  	} else {
>  		ret = vmbus_sendpacket(out_channel,
>  				       &nvmsg, sizeof(nvmsg),
> @@ -1010,6 +1138,7 @@ static inline int netvsc_send_pkt(
>  				       VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED);
>  	}
> 
> +exit:
>  	if (ret == 0) {
>  		atomic_inc_return(&nvchan->queue_sends);
> 
> diff --git a/drivers/net/hyperv/rndis_filter.c b/drivers/net/hyperv/rndis_filter.c
> index f6c9c2a670f9..448fcc325ed7 100644
> --- a/drivers/net/hyperv/rndis_filter.c
> +++ b/drivers/net/hyperv/rndis_filter.c
> @@ -361,6 +361,8 @@ static void rndis_filter_receive_response(struct net_device *ndev,
>  			}
>  		}
> 
> +		netvsc_dma_unmap(((struct net_device_context *)
> +			netdev_priv(ndev))->device_ctx, &request->pkt);
>  		complete(&request->wait_event);
>  	} else {
>  		netdev_err(ndev,
> diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h
> index 724a735d722a..139a43ad65a1 100644
> --- a/include/linux/hyperv.h
> +++ b/include/linux/hyperv.h
> @@ -1596,6 +1596,11 @@ struct hyperv_service_callback {
>  	void (*callback)(void *context);
>  };
> 
> +struct hv_dma_range {
> +	dma_addr_t dma;
> +	u32 mapping_size;
> +};
> +
>  #define MAX_SRV_VER	0x7ffffff
>  extern bool vmbus_prep_negotiate_resp(struct icmsg_hdr *icmsghdrp, u8 *buf, u32 buflen,
>  				const int *fw_version, int fw_vercnt,
> --
> 2.25.1