Transceiver52M: UHD: Exit on receive more than 100 timeout errors
Receiving timeout is not a fatal error and should be recoverable on network devices.
But for some reasons these errors are not always recoverable, in this case osmo-trx should be restarted.
diff --git a/Transceiver52M/UHDDevice.cpp b/Transceiver52M/UHDDevice.cpp
index bdeb1b5..72cad4d 100644
--- a/Transceiver52M/UHDDevice.cpp
+++ b/Transceiver52M/UHDDevice.cpp
@@ -823,6 +823,7 @@
int uhd_device::check_rx_md_err(uhd::rx_metadata_t &md, ssize_t num_smpls)
{
uhd::time_spec_t ts;
+ static int err_count = 0;
if (!num_smpls) {
LOG(ERR) << str_code(md);
@@ -830,6 +831,11 @@
switch (md.error_code) {
case uhd::rx_metadata_t::ERROR_CODE_TIMEOUT:
LOG(ALERT) << "UHD: Receive timed out";
+ if (err_count > 100) {
+ err_count = 0;
+ return ERROR_UNRECOVERABLE;
+ }
+ err_count++;
case uhd::rx_metadata_t::ERROR_CODE_OVERFLOW:
case uhd::rx_metadata_t::ERROR_CODE_LATE_COMMAND:
case uhd::rx_metadata_t::ERROR_CODE_BROKEN_CHAIN: