Commit ac498fa3 authored by Darek Stojaczyk's avatar Darek Stojaczyk Committed by Changpeng Liu
Browse files

vhost: use DPDK APIs to split non-contiguous virtual memory buffers 



Currently, we translate each 2MB chunk to manually check
if it's contiguous with the previous one, but there are
rte_vhost APIs that do it way more efficiently.

rte_vhost_va_from_guest_pa() was introduced in DPDK 18.02,
but was backported to 17.11 as well, so we don't even need
any RTE_VERSION ifdefs to use it now. This function
calculates the remaining region size instead of trying to
translate subsequent 2MB chunks over and over.

The previous rte_vhost_gpa_to_vva() was deprecated a long
time ago and after this patch we no longer make any use of
it.

DPDK usages of this new function check if the translated
memory region has 0 length, which seems very silly, but
let's just do it in SPDK as well.

Change-Id: Ifae8daa5f810b5a2ba1524958ad2399af700b532
Signed-off-by: default avatarDarek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/454878


Tested-by: default avatarSPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: default avatarBen Walker <benjamin.walker@intel.com>
Reviewed-by: default avatarChangpeng Liu <changpeng.liu@intel.com>
parent 5931284f

lib/vhost/vhost.c

+5 −25
Original line number Diff line number Diff line
@@ -479,9 +479,8 @@ int 
spdk_vhost_vring_desc_to_iov(struct spdk_vhost_session *vsession, struct iovec *iov,

			     uint16_t *iov_index, const struct vring_desc *desc)

{

	uint32_t remaining = desc->len;

	uint32_t to_boundary;

	uint32_t len;

	uint64_t len;

	uint64_t remaining = desc->len;

	uintptr_t payload = desc->addr;

	uintptr_t vva;
@@ -490,31 +489,12 @@ spdk_vhost_vring_desc_to_iov(struct spdk_vhost_session *vsession, struct iovec * 
			SPDK_ERRLOG("SPDK_VHOST_IOVS_MAX(%d) reached\n", SPDK_VHOST_IOVS_MAX);

			return -1;

		}

		vva = (uintptr_t)rte_vhost_gpa_to_vva(vsession->mem, payload);

		if (vva == 0) {

		len = remaining;

		vva = (uintptr_t)rte_vhost_va_from_guest_pa(vsession->mem, payload, &len);

		if (vva == 0 || len == 0) {

			SPDK_ERRLOG("gpa_to_vva(%p) == NULL\n", (void *)payload);

			return -1;

		}

		to_boundary = VALUE_2MB - _2MB_OFFSET(payload);

		if (spdk_likely(remaining <= to_boundary)) {

			len = remaining;

		} else {

			/*

			 * Descriptor crosses a 2MB hugepage boundary.  vhost memory regions are allocated

			 *  from hugepage memory, so this means this descriptor may be described by

			 *  discontiguous vhost memory regions.  Do not blindly split on the 2MB boundary,

			 *  only split it if the two sides of the boundary do not map to the same vhost

			 *  memory region.  This helps ensure we do not exceed the max number of IOVs

			 *  defined by SPDK_VHOST_IOVS_MAX.

			 */

			len = to_boundary;

			while (len < remaining) {

				if (vva + len != (uintptr_t)rte_vhost_gpa_to_vva(vsession->mem, payload + len)) {

					break;

				}

				len += spdk_min(remaining - len, VALUE_2MB);

			}

		}

		iov[*iov_index].iov_base = (void *)vva;

		iov[*iov_index].iov_len = len;

		remaining -= len;