[PATCH v12 4/5] vhost_user: Function defs for add/rem mem regions

[fengchengwen]

The 3/5 commit is helper function, I doubt it could compile, why not merge with this commit.

I suggest placing the refactored code on 3/5, but it should be self-contained and not include
functions that will not be used in 3/5 (e.g. remove_guest_pages).

On 5/6/2026 11:37 AM, pravin.bathija at dell.com wrote:
> From: Pravin M Bathija <pravin.bathija at dell.com>
> 
> These changes cover the function definition for add/remove memory
> region calls which are invoked on receiving vhost user message from
> vhost user front-end (e.g. Qemu). In our case, in addition to testing
> with qemu front-end, the testing has also been performed with libblkio
> front-end and spdk/dpdk back-end. We did I/O using libblkio based device
> driver, to spdk based drives.
> There are also changes for set_mem_table and new definition for get memory
> slots. Our changes optimize the set memory table call to use common support
> functions. A new vhost_user_initialize_memory() function is introduced to
> factor out the common memory initialization logic from the function
> vhost_user_set_mem_table(), which is now called from both the SET_MEM_TABLE
> message handler and the ADD_MEM_REG handler (for the first region).
> Message get memory slots is how the vhost-user front-end queries the
> vhost-user back-end about the number of memory slots available to be
> registered by the back-end. In addition support function to invalidate
> vring is also defined which is used in add/remove memory region functions.
> 
> Signed-off-by: Pravin M Bathija <pravin.bathija at dell.com>
> ---
>  lib/vhost/vhost_user.c | 311 ++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 278 insertions(+), 33 deletions(-)
> 
> diff --git a/lib/vhost/vhost_user.c b/lib/vhost/vhost_user.c
> index 9f1abb1842..c6b404da75 100644
> --- a/lib/vhost/vhost_user.c
> +++ b/lib/vhost/vhost_user.c
> @@ -71,6 +71,9 @@ VHOST_MESSAGE_HANDLER(VHOST_USER_SET_FEATURES, vhost_user_set_features, false, t
>  VHOST_MESSAGE_HANDLER(VHOST_USER_SET_OWNER, vhost_user_set_owner, false, true) \
>  VHOST_MESSAGE_HANDLER(VHOST_USER_RESET_OWNER, vhost_user_reset_owner, false, false) \
>  VHOST_MESSAGE_HANDLER(VHOST_USER_SET_MEM_TABLE, vhost_user_set_mem_table, true, true) \
> +VHOST_MESSAGE_HANDLER(VHOST_USER_GET_MAX_MEM_SLOTS, vhost_user_get_max_mem_slots, false, false) \
> +VHOST_MESSAGE_HANDLER(VHOST_USER_ADD_MEM_REG, vhost_user_add_mem_reg, true, true) \
> +VHOST_MESSAGE_HANDLER(VHOST_USER_REM_MEM_REG, vhost_user_rem_mem_reg, true, true) \
>  VHOST_MESSAGE_HANDLER(VHOST_USER_SET_LOG_BASE, vhost_user_set_log_base, true, true) \
>  VHOST_MESSAGE_HANDLER(VHOST_USER_SET_LOG_FD, vhost_user_set_log_fd, true, true) \
>  VHOST_MESSAGE_HANDLER(VHOST_USER_SET_VRING_NUM, vhost_user_set_vring_num, false, true) \
> @@ -1431,6 +1434,52 @@ vhost_user_mmap_region(struct virtio_net *dev,
>  	return 0;
>  }
>  
> +static int
> +vhost_user_initialize_memory(struct virtio_net **pdev)
> +{
> +	struct virtio_net *dev = *pdev;
> +	int numa_node = SOCKET_ID_ANY;
> +
> +	if (dev->mem != NULL) {
> +		VHOST_CONFIG_LOG(dev->ifname, ERR,
> +			"memory already initialized, free it first");
> +		return -1;
> +	}
> +
> +	/*
> +	 * If VQ 0 has already been allocated, try to allocate on the same
> +	 * NUMA node. It can be reallocated later in numa_realloc().
> +	 */
> +	if (dev->nr_vring > 0)
> +		numa_node = dev->virtqueue[0]->numa_node;
> +
> +	dev->nr_guest_pages = 0;
> +	if (dev->guest_pages == NULL) {
> +		dev->max_guest_pages = 8;
> +		dev->guest_pages = rte_zmalloc_socket(NULL,
> +					dev->max_guest_pages *
> +					sizeof(struct guest_page),
> +					RTE_CACHE_LINE_SIZE,
> +					numa_node);
> +		if (dev->guest_pages == NULL) {
> +			VHOST_CONFIG_LOG(dev->ifname, ERR,
> +				"failed to allocate memory for dev->guest_pages");
> +			return -1;
> +		}
> +	}
> +
> +	dev->mem = rte_zmalloc_socket("vhost-mem-table", sizeof(struct rte_vhost_memory) +
> +		sizeof(struct rte_vhost_mem_region) * VHOST_MEMORY_MAX_NREGIONS, 0, numa_node);
> +	if (dev->mem == NULL) {
> +		VHOST_CONFIG_LOG(dev->ifname, ERR, "failed to allocate memory for dev->mem");
> +		rte_free(dev->guest_pages);
> +		dev->guest_pages = NULL;
> +		return -1;
> +	}
> +
> +	return 0;
> +}
> +
>  static int
>  vhost_user_set_mem_table(struct virtio_net **pdev,
>  			struct vhu_msg_context *ctx,
> @@ -1439,7 +1488,6 @@ vhost_user_set_mem_table(struct virtio_net **pdev,
>  	struct virtio_net *dev = *pdev;
>  	struct VhostUserMemory *memory = &ctx->msg.payload.memory;
>  	struct rte_vhost_mem_region *reg;
> -	int numa_node = SOCKET_ID_ANY;
>  	uint64_t mmap_offset;
>  	uint32_t i;
>  	bool async_notify = false;
> @@ -1484,39 +1532,13 @@ vhost_user_set_mem_table(struct virtio_net **pdev,
>  		if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM))
>  			vhost_user_iotlb_flush_all(dev);
>  
> -		free_mem_region(dev);
> +		free_all_mem_regions(dev);
>  		rte_free(dev->mem);
>  		dev->mem = NULL;
>  	}
>  
> -	/*
> -	 * If VQ 0 has already been allocated, try to allocate on the same
> -	 * NUMA node. It can be reallocated later in numa_realloc().
> -	 */
> -	if (dev->nr_vring > 0)
> -		numa_node = dev->virtqueue[0]->numa_node;
> -
> -	dev->nr_guest_pages = 0;
> -	if (dev->guest_pages == NULL) {
> -		dev->max_guest_pages = 8;
> -		dev->guest_pages = rte_zmalloc_socket(NULL,
> -					dev->max_guest_pages *
> -					sizeof(struct guest_page),
> -					RTE_CACHE_LINE_SIZE,
> -					numa_node);
> -		if (dev->guest_pages == NULL) {
> -			VHOST_CONFIG_LOG(dev->ifname, ERR,
> -				"failed to allocate memory for dev->guest_pages");
> -			goto close_msg_fds;
> -		}
> -	}
> -
> -	dev->mem = rte_zmalloc_socket("vhost-mem-table", sizeof(struct rte_vhost_memory) +
> -		sizeof(struct rte_vhost_mem_region) * memory->nregions, 0, numa_node);
> -	if (dev->mem == NULL) {
> -		VHOST_CONFIG_LOG(dev->ifname, ERR, "failed to allocate memory for dev->mem");
> -		goto free_guest_pages;
> -	}
> +	if (vhost_user_initialize_memory(pdev) < 0)
> +		goto close_msg_fds;
>  
>  	for (i = 0; i < memory->nregions; i++) {
>  		reg = &dev->mem->regions[i];
> @@ -1580,11 +1602,9 @@ vhost_user_set_mem_table(struct virtio_net **pdev,
>  	return RTE_VHOST_MSG_RESULT_OK;
>  
>  free_mem_table:
> -	free_mem_region(dev);
> +	free_all_mem_regions(dev);
>  	rte_free(dev->mem);
>  	dev->mem = NULL;
> -
> -free_guest_pages:
>  	rte_free(dev->guest_pages);
>  	dev->guest_pages = NULL;
>  close_msg_fds:
> @@ -1592,6 +1612,231 @@ vhost_user_set_mem_table(struct virtio_net **pdev,
>  	return RTE_VHOST_MSG_RESULT_ERR;
>  }
>  
> +
> +static int
> +vhost_user_get_max_mem_slots(struct virtio_net **pdev __rte_unused,
> +			struct vhu_msg_context *ctx,
> +			int main_fd __rte_unused)
> +{
> +	uint32_t max_mem_slots = VHOST_MEMORY_MAX_NREGIONS;
> +
> +	ctx->msg.payload.u64 = (uint64_t)max_mem_slots;
> +	ctx->msg.size = sizeof(ctx->msg.payload.u64);
> +	ctx->fd_num = 0;
> +
> +	return RTE_VHOST_MSG_RESULT_REPLY;
> +}
> +
> +static void
> +_dev_invalidate_vrings(struct virtio_net **pdev)
> +{
> +	struct virtio_net *dev = *pdev;
> +	uint32_t i;
> +
> +	for (i = 0; i < dev->nr_vring; i++) {
> +		struct vhost_virtqueue *vq = dev->virtqueue[i];
> +
> +		if (!vq)
> +			continue;
> +
> +		if (vq->desc || vq->avail || vq->used) {
> +			vq_assert_lock(dev, vq);
> +
> +			/*
> +			 * If the memory table got updated, the ring addresses
> +			 * need to be translated again as virtual addresses have
> +			 * changed.
> +			 */
> +			vring_invalidate(dev, vq);
> +
> +			translate_ring_addresses(&dev, &vq);
> +		}
> +	}
> +
> +	*pdev = dev;
> +}
> +
> +/*
> + * Macro wrapper that performs the compile-time lock assertion with the
> + * correct message ID at the call site, then calls the implementation.
> + */
> +#define dev_invalidate_vrings(pdev, id) do { \
> +	static_assert(id ## _LOCK_ALL_QPS, \
> +		#id " handler is not declared as locking all queue pairs"); \
> +	_dev_invalidate_vrings(pdev); \
> +} while (0)
> +
> +static int
> +vhost_user_add_mem_reg(struct virtio_net **pdev,
> +			struct vhu_msg_context *ctx,
> +			int main_fd __rte_unused)
> +{
> +	uint32_t i;
> +	struct virtio_net *dev = *pdev;
> +	struct VhostUserMemoryRegion *region = &ctx->msg.payload.memory_single.region;
> +
> +	/* convert first region add to normal memory table set */
> +	if (dev->mem == NULL) {
> +		if (vhost_user_initialize_memory(pdev) < 0)
> +			goto close_msg_fds;
> +	}
> +
> +	/* make sure new region will fit */
> +	if (dev->mem->nregions >= VHOST_MEMORY_MAX_NREGIONS) {
> +		VHOST_CONFIG_LOG(dev->ifname, ERR, "too many memory regions already (%u)",
> +									dev->mem->nregions);
> +		goto close_msg_fds;
> +	}
> +
> +	/* make sure supplied memory fd present */
> +	if (ctx->fd_num != 1) {
> +		VHOST_CONFIG_LOG(dev->ifname, ERR, "fd count makes no sense (%u)", ctx->fd_num);
> +		goto close_msg_fds;
> +	}
> +
> +	/* Make sure no overlap in guest virtual address space */
> +	for (i = 0; i < dev->mem->nregions; i++) {
> +		struct rte_vhost_mem_region *current_region = &dev->mem->regions[i];
> +		uint64_t current_region_guest_start = current_region->guest_user_addr;
> +		uint64_t current_region_guest_end = current_region_guest_start
> +							+ current_region->size - 1;
> +		uint64_t proposed_region_guest_start = region->userspace_addr;
> +		uint64_t proposed_region_guest_end = proposed_region_guest_start
> +							+ region->memory_size - 1;



> +
> +		if (!((proposed_region_guest_end < current_region_guest_start) ||
> +			(proposed_region_guest_start > current_region_guest_end))) {
> +			VHOST_CONFIG_LOG(dev->ifname, ERR,
> +				"requested memory region overlaps with another region");
> +			VHOST_CONFIG_LOG(dev->ifname, ERR,
> +				"\tRequested region address:0x%" PRIx64,
> +				region->userspace_addr);
> +			VHOST_CONFIG_LOG(dev->ifname, ERR,
> +				"\tRequested region size:0x%" PRIx64,
> +				region->memory_size);
> +			VHOST_CONFIG_LOG(dev->ifname, ERR,
> +				"\tOverlapping region address:0x%" PRIx64,
> +				current_region->guest_user_addr);
> +			VHOST_CONFIG_LOG(dev->ifname, ERR,
> +				"\tOverlapping region size:0x%" PRIx64,
> +				current_region->size);
> +			goto close_msg_fds;
> +		}
> +	}
> +
> +	/* New region goes at the end of the contiguous array */
> +	struct rte_vhost_mem_region *reg = &dev->mem->regions[dev->mem->nregions];
> +
> +	reg->guest_phys_addr = region->guest_phys_addr;
> +	reg->guest_user_addr = region->userspace_addr;
> +	reg->size            = region->memory_size;
> +	reg->fd              = ctx->fds[0];
> +	ctx->fds[0]          = -1;
> +
> +	if (vhost_user_mmap_region(dev, reg, region->mmap_offset) < 0) {
> +		VHOST_CONFIG_LOG(dev->ifname, ERR, "failed to mmap region");
> +		if (reg->mmap_addr) {
> +			/* mmap succeeded but a later step (e.g. add_guest_pages)
> +			 * failed; undo the mapping and any guest-page entries.
> +			 */
> +			remove_guest_pages(dev, reg);
> +			free_mem_region(reg);
> +		} else {
> +			close(reg->fd);
> +			reg->fd = -1;
> +		}
> +		goto close_msg_fds;
> +	}
> +
> +	dev->mem->nregions++;
> +
> +	if (dev->async_copy && rte_vfio_is_enabled("vfio")) {
> +		if (async_dma_map_region(dev, reg, true) < 0)
> +			goto free_new_region;

It will invoke in goto label

> +	}
> +
> +	if (dev->postcopy_listening) {
> +		/*
> +		 * Cannot use vhost_user_postcopy_register() here because it
> +		 * reads ctx->msg.payload.memory (SET_MEM_TABLE layout), but
> +		 * ADD_MEM_REG uses the memory_single payload.  Register the
> +		 * single new region directly instead.
> +		 */
> +		if (vhost_user_postcopy_region_register(dev, reg) < 0)
> +			goto free_new_region;
> +	}
> +
> +	dev_invalidate_vrings(pdev, VHOST_USER_ADD_MEM_REG);
> +	dev = *pdev;

no need to do this, there are: struct virtio_net *dev = *pdev; in beginning

> +	dump_guest_pages(dev);
> +
> +	return RTE_VHOST_MSG_RESULT_OK;
> +
> +free_new_region:
> +	if (dev->async_copy && rte_vfio_is_enabled("vfio"))
> +		async_dma_map_region(dev, reg, false);
> +	remove_guest_pages(dev, reg);
> +	free_mem_region(reg);
> +	dev->mem->nregions--;
> +close_msg_fds:
> +	close_msg_fds(ctx);
> +	return RTE_VHOST_MSG_RESULT_ERR;
> +}
> +
> +static int
> +vhost_user_rem_mem_reg(struct virtio_net **pdev,
> +			struct vhu_msg_context *ctx,
> +			int main_fd __rte_unused)
> +{
> +	uint32_t i;
> +	struct virtio_net *dev = *pdev;
> +	struct VhostUserMemoryRegion *region = &ctx->msg.payload.memory_single.region;
> +
> +	if (dev->mem == NULL || dev->mem->nregions == 0) {
> +		VHOST_CONFIG_LOG(dev->ifname, ERR, "no memory regions to remove");
> +		close_msg_fds(ctx);

no fd when remove, so please don't invoke close_msg_fds

> +		return RTE_VHOST_MSG_RESULT_ERR;
> +	}
> +
> +	for (i = 0; i < dev->mem->nregions; i++) {
> +		struct rte_vhost_mem_region *current_region = &dev->mem->regions[i];
> +
> +		/*
> +		 * According to the vhost-user specification:
> +		 * The memory region to be removed is identified by its GPA,
> +		 * user address and size. The mmap offset is ignored.
> +		 */
> +		if (region->userspace_addr == current_region->guest_user_addr
> +			&& region->guest_phys_addr == current_region->guest_phys_addr
> +			&& region->memory_size == current_region->size) {
> +			if (dev->async_copy && rte_vfio_is_enabled("vfio"))
> +				async_dma_map_region(dev, current_region, false);

I'm afaid it not enough, we should drain DMA's pending request. But I think the original
impl also has this problem.

> +			remove_guest_pages(dev, current_region);
> +			free_mem_region(current_region);
> +
> +			/* Compact the regions array to keep it contiguous */
> +			if (i < dev->mem->nregions - 1) {
> +				memmove(&dev->mem->regions[i],
> +					&dev->mem->regions[i + 1],
> +					(dev->mem->nregions - 1 - i) *
> +					sizeof(struct rte_vhost_mem_region));
> +				memset(&dev->mem->regions[dev->mem->nregions - 1],
> +					0, sizeof(struct rte_vhost_mem_region));
> +			}
> +
> +			dev->mem->nregions--;
> +			dev_invalidate_vrings(pdev, VHOST_USER_REM_MEM_REG);
> +			dev = *pdev;

no need to do this, there are: struct virtio_net *dev = *pdev; in beginning

> +			close_msg_fds(ctx);
> +			return RTE_VHOST_MSG_RESULT_OK;
> +		}
> +	}
> +
> +	VHOST_CONFIG_LOG(dev->ifname, ERR, "failed to find region");
> +	close_msg_fds(ctx);
> +	return RTE_VHOST_MSG_RESULT_ERR;
> +}
> +
>  static bool
>  vq_is_ready(struct virtio_net *dev, struct vhost_virtqueue *vq)
>  {