On 06/30, Taehee Yoo wrote:
The devmem TCP requires the hds-thresh value to be 0, but it doesn't change it automatically. Therefore, configure_hds_thresh() is added to handle this.
The run_devmem_tests() now tests hds_thresh, but it skips test if the hds_thresh_max value is 0.
Signed-off-by: Taehee Yoo ap420073@gmail.com
.../selftests/drivers/net/hw/ncdevmem.c | 86 +++++++++++++++++++ 1 file changed, 86 insertions(+)
diff --git a/tools/testing/selftests/drivers/net/hw/ncdevmem.c b/tools/testing/selftests/drivers/net/hw/ncdevmem.c index cc9b40d9c5d5..d78b5e5697d7 100644 --- a/tools/testing/selftests/drivers/net/hw/ncdevmem.c +++ b/tools/testing/selftests/drivers/net/hw/ncdevmem.c @@ -349,6 +349,72 @@ static int configure_headersplit(bool on) return ret; } +static int configure_hds_thresh(int len) +{
- struct ethtool_rings_get_req *get_req;
- struct ethtool_rings_get_rsp *get_rsp;
- struct ethtool_rings_set_req *req;
- struct ynl_error yerr;
- struct ynl_sock *ys;
- int ret;
- ys = ynl_sock_create(&ynl_ethtool_family, &yerr);
- if (!ys) {
fprintf(stderr, "YNL: %s\n", yerr.msg);
return -1;
- }
- req = ethtool_rings_set_req_alloc();
- ethtool_rings_set_req_set_header_dev_index(req, ifindex);
[..]
- ethtool_rings_set_req_set_hds_thresh(req, len);
Any reason not to add this to the existing configure_headersplit? Is it only for tests?