Started by upstream project "gerrit-osmo-bts" build number 1,1291129gerrit-osmo-btsjob/gerrit-osmo-bts/COMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImaster297269297269002316052316051.0303368303368053497360956095<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btshttps://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/1127/artifacthttps://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/changeshttps://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/1127/https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/1127/testReportfalse#1127231605224846gerrit-osmo-bts-build #11271127falsefalse112774530SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/1127/src/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/bts_shutdown_fsm.ceditsrc/common/tx_power.ceditinclude/osmo-bts/tx_power.hgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyfixeriaCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/000074055740551.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11277405576078gerrit-osmo-bts-build » BTS_MODEL=lc15,FIRMWARE_VERSION=litecell15,0,default,osmocom-gerrit #11271127falsefalse112774560SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=lc15,FIRMWARE_VERSION=litecell15,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.ceditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/00735457354561422614221.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276142274960gerrit-osmo-bts-build » BTS_MODEL=lc15,FIRMWARE_VERSION=master,0,default,osmocom-gerrit #11271127falsefalse112774565SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=lc15,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.ceditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/00689636896366048660481.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276604877186gerrit-osmo-bts-build » BTS_MODEL=lc15,FIRMWARE_VERSION=v2017.01,0,default,osmocom-gerrit #11271127falsefalse112774562SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=lc15,FIRMWARE_VERSION=v2017.01,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/bts_shutdown_fsm.ceditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/00689466894666064660641.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276606477639gerrit-osmo-bts-build » BTS_MODEL=oc2g,FIRMWARE_VERSION=master,0,default,osmocom-gerrit #11271127falsefalse112774561SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=oc2g,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/bts_shutdown_fsm.ceditsrc/common/tx_power.ceditinclude/osmo-bts/tx_power.hgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/00735607356060889608891.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276088978510gerrit-osmo-bts-build » BTS_MODEL=oc2g,FIRMWARE_VERSION=oc2g,0,default,osmocom-gerrit #11271127falsefalse112774566SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=oc2g,FIRMWARE_VERSION=oc2g,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.ceditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/001174061740611.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11277406172135gerrit-osmo-bts-build » BTS_MODEL=oc2g,FIRMWARE_VERSION=oc2g-next,0,default,osmocom-gerrit #11271127falsefalse112774553SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=oc2g,FIRMWARE_VERSION=oc2g-next,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/tx_power.ceditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/000073113731131.003333<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11277311366548gerrit-osmo-bts-build » BTS_MODEL=oct,FIRMWARE_VERSION=master,0,default,osmocom-gerrit #11271127falsefalse112774556SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=oct,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/bts_shutdown_fsm.ceditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/001175779757791.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11277577968530gerrit-osmo-bts-build » BTS_MODEL=oct+trx,FIRMWARE_VERSION=master,0,default,osmocom-gerrit #11271127falsefalse112774554SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=oct+trx,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.ceditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/001193782937821.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11279378291022gerrit-osmo-bts-build » BTS_MODEL=sysmo,FIRMWARE_VERSION=femtobts_v2.7,1,default,osmocom-gerrit #11271127falsefalse112774558SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=sysmo,FIRMWARE_VERSION=femtobts_v2.7,WITH_MANUALS=1,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.ceditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/000068631686311.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276863166861gerrit-osmo-bts-build » BTS_MODEL=sysmo,FIRMWARE_VERSION=master,0,default,osmocom-gerrit #11271127falsefalse112774551SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=sysmo,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/tx_power.ceditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/000073123731231.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11277312366870gerrit-osmo-bts-build » BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v2.4,0,default,osmocom-gerrit #11271127falsefalse112774559SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v2.4,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/bts_shutdown_fsm.ceditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/001174184741841.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11277418467593gerrit-osmo-bts-build » BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v3.0.1pre,0,default,osmocom-gerrit #11271127falsefalse112774552SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v3.0.1pre,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build5-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/tx_power.ceditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/000069328693281.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276932865669gerrit-osmo-bts-build » BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v3.1,0,default,osmocom-gerrit #11271127falsefalse112774555SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v3.1,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/bts_shutdown_fsm.ceditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/00696076960757506575061.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11275750668156gerrit-osmo-bts-build » BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v5.1,0,default,osmocom-gerrit #11271127falsefalse112774564SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=sysmo,FIRMWARE_VERSION=superfemto_v5.1,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/tx_power.ceditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/000068545685451.0000<_2023q1 _class='hudson.plugins.git.util.Build'>374bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1bfdc23b1c676d017da2c1dff26d937be9c9926e2bfdc23b1c676d017da2c1dff26d937be9c9926e22023q1794f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.1f79a23bab0e43959d25363d2a4903127e0a1cb3af79a23bab0e43959d25363d2a4903127e0a1cb3aosmith/1.7.187475a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.275a087d16ba2539b8bcf05102b41d820e818724175a087d16ba2539b8bcf05102b41d820e8187241osmith/1.7.211274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#11276854564908gerrit-osmo-bts-build » BTS_MODEL=trx,FIRMWARE_VERSION=master,0,default,osmocom-gerrit #11271127falsefalse112774557SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=trx,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=osmocom-gerrit/1127/build4-deb12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditsrc/common/tx_power.ceditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim YanitskiyCOMMENT_TYPEDISTROGERRIT_BRANCHmasterGERRIT_CHANGE_NUMBER42524GERRIT_HOSTgerrit.osmocom.orgGERRIT_PATCHSET_NUMBER1GERRIT_PATCHSET_REVISION4c6324213ac87f9979df7e6d0f596f175ddb27f6GERRIT_PATCHSET_UPLOADER_NAMEfixeriaGERRIT_PORT29418GERRIT_PROJECTosmo-btsGERRIT_REFSPECrefs/changes/24/42524/1GERRIT_REPO_URLssh://jenkins@gerrit.osmocom.org:29418/osmo-btsPIPELINE_BUILD_URLhttps://jenkins.osmocom.org/jenkins/job/gerrit-osmo-bts/1129/PROJECT_NAMEosmo-btsBRANCH_CImasterStarted by upstream project "gerrit-osmo-bts-build" build number 1,1271127gerrit-osmo-bts-buildjob/gerrit-osmo-bts-build/00112268672268671.000011274c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6master4c6324213ac87f9979df7e6d0f596f175ddb27f64c6324213ac87f9979df7e6d0f596f175ddb27f6masterssh://jenkins@gerrit.osmocom.org:29418/osmo-btsfalse#1127226867220304gerrit-osmo-bts-build » BTS_MODEL=trx,FIRMWARE_VERSION=master,0,default,raspbian #11271127falsefalse112774563SUCCESS1774792375960https://jenkins.osmocom.org/jenkins/view/all/job/gerrit-osmo-bts-build/BTS_MODEL=trx,FIRMWARE_VERSION=master,WITH_MANUALS=0,a4=default,label=raspbian/1127/lx2-raspbian12build-ansiblesrc/common/tx_power.cinclude/osmo-bts/tx_power.hsrc/common/bts_shutdown_fsm.c4c6324213ac87f9979df7e6d0f596f175ddb27f61774610689000https://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiyvyanitskiy@sysmocom.debts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSED With multiple TRXes ramping down in lockstep, both their final ramp-timer callbacks fire back-to-back in the same event loop pass, setting p_total_cur_mdBm to the target value for both before any async hardware acknowledgement arrives. When the first SETPOWER ack returns and fires ramp_down_compl_cb() for TRX0, the remaining-TRX check in st_wait_ramp_down_compl() inspects p_total_cur_mdBm for TRX1 and finds it already at the target - concluding that all TRXes are done. The FSM then transitions to WAIT_TRX_CLOSED, and the second ack (for TRX1) fires ramp_down_compl_cb() into the wrong state, producing: BTS_SHUTDOWN(...){WAIT_TRX_CLOSED}: Event BTS_SHUTDOWN_EV_TRX_RAMP_COMPL not permitted The root cause is that p_total_cur_mdBm is a *requested* value set in the timer callback, not a confirmed one. The hardware confirmation arrives asynchronously via power_trx_change_compl() -> power_ramp_do_step(), which is also where compl_cb() is invoked. Fix by adding a 'complete' flag to trx_power_params.ramp that is: * cleared when _power_ramp_start() begins a new ramp, and * set just before compl_cb() is called in power_ramp_do_step() The shutdown FSM remaining-TRX count then checks !ramp.complete instead of comparing p_total_cur_mdBm against the target, correctly reflecting which TRXes have actually received hardware confirmation of ramp completion. Change-Id: Ia71393e871187d6b44b7f520eb421bab354aafd1 2026-03-27 18:24:49 +07004c6324213ac87f9979df7e6d0f596f175ddb27f6bts_shutdown_fsm: fix spurious RAMP_COMPL events in WAIT_TRX_CLOSEDeditinclude/osmo-bts/tx_power.heditsrc/common/bts_shutdown_fsm.ceditsrc/common/tx_power.cgithttps://jenkins.osmocom.org/jenkins/user/fixeriaVadim Yanitskiy