rfsimulator: fix model name for channel

bridge->nb_cnx++ is done before calling snprintf(modelname,... in the function,
so we take the model name of the next channel. This commit corrects the issue by
using bridge->nb_cnx - 1, ensuring that the model name corresponds to the currently
connected client.
This bug was introduced by MR !3455.

Co-authored-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>
Co-authored-by: Laurent THOMAS <laurent.thomas@open-cells.com>
This commit is contained in:
Jaroslava Fiedlerova
2025-07-22 18:21:03 +02:00
co-authored by Bartosz Podrygajlo Laurent THOMAS
parent fc1c762cf2
commit c5bc949a96
+1 -1
View File
@@ -218,7 +218,7 @@ static buffer_t *allocCirBuf(rfsimulator_state_t *bridge, int sock)
sizeofArray(modelname), sizeofArray(modelname),
"rfsimu_channel_%s%d", "rfsimu_channel_%s%d",
(bridge->role == SIMU_ROLE_SERVER) ? "ue" : "enB", (bridge->role == SIMU_ROLE_SERVER) ? "ue" : "enB",
bridge->nb_cnx); bridge->nb_cnx - 1);
ptr->channel_model = find_channel_desc_fromname(modelname); // path_loss in dB ptr->channel_model = find_channel_desc_fromname(modelname); // path_loss in dB
if (!ptr->channel_model) { if (!ptr->channel_model) {
// Use legacy method to find channel model - this will use the same channel model for all clients // Use legacy method to find channel model - this will use the same channel model for all clients