Re: [PATCH v2 2/3] vu_handle_tx: Pass actual remaining out_sg capacity to vu_queue_pop()

On Fri, 13 Mar 2026 19:26:17 +0100 Laurent Vivier wrote:

In vu_handle_tx(), pass the actual remaining iovec capacity (ARRAY_SIZE(out_sg) - out_sg_count) to vu_queue_pop() rather than a fixed VU_MAX_TX_BUFFER_NB.

This enables dynamic allocation of iovec entries to each element rather than reserving a fixed number of slots per descriptor.

Signed-off-by: Laurent Vivier --- vu_common.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/vu_common.c b/vu_common.c index 4d809ac38a4b..ed0033d6bb11 100644 --- a/vu_common.c +++ b/vu_common.c @@ -20,8 +20,6 @@ #include "migrate.h" #include "epoll_ctl.h"

-#define VU_MAX_TX_BUFFER_NB 2 - /** * vu_packet_check_range() - Check if a given memory zone is contained in * a mapped guest memory region @@ -177,13 +175,14 @@ static void vu_handle_tx(struct vu_dev *vdev, int index,

count = 0; out_sg_count = 0; - while (count < VIRTQUEUE_MAX_SIZE && - out_sg_count + VU_MAX_TX_BUFFER_NB <= VIRTQUEUE_MAX_SIZE) { - int ret; + while (count < ARRAY_SIZE(elem) && + out_sg_count < ARRAY_SIZE(out_sg)) {

Nit: you could do this now: while (count < ARRAY_SIZE(elem) && out_sg_count < ARRAY_SIZE(out_sg)) { I still need a bit of time to review 3/3, if you don't need to re-spin for other reasons I can also fix this up on merge.

struct iov_tail data; + int ret;

ret = vu_queue_pop(vdev, vq, &elem[count], NULL, 0, - &out_sg[out_sg_count], VU_MAX_TX_BUFFER_NB); + &out_sg[out_sg_count], + ARRAY_SIZE(out_sg) - out_sg_count); if (ret < 0) break; out_sg_count += elem[count].out_num;

-- Stefano