- gerrit-log - lists.osmocom.org

Change in osmo-ttcn3-hacks[master]: bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this...
by neels 12 Apr '22

12 Apr '22

neels has submitted this change. ( https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27741 ) Change subject: bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this_bsc() ...................................................................... bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this_bsc() Deactivate as_Media() once the handover is completed, so that it does not fail on the expected MGCP DLCX from release. Fix fallout seen in these tests: SCCPlite BSC_Tests.TC_ho_into_this_bsc BSC_Tests.TC_ho_into_this_bsc_a5_0 BSC_Tests.TC_ho_into_this_bsc_a5_1 BSC_Tests.TC_ho_into_this_bsc_a5_3 BSC_Tests.TC_ho_into_this_bsc_a5_4 BSC_Tests.TC_ho_into_this_bsc_a5_1_3_no_chosen_enc_alg BSC_Tests.TC_ho_into_this_bsc_a5_1_3 BSC_Tests.TC_srvcc_eutran_to_geran BSC_Tests.TC_srvcc_eutran_to_geran_a5_3 BSC_Tests.TC_srvcc_eutran_to_geran_src_sai BSC_Tests.TC_srvcc_eutran_to_geran_forbid_fast_return BSC_Tests.TC_ho_into_this_bsc_sccp_cr_without_bssap All of these tests use f_ho_into_this_bsc(). (It is not clear to me why only the SCCPlite tests show the fallout, the AoIP should be similarly affected, but isn't.) The failures were introduced by recent merge of I0633f60f09d58802f6be0238ef41a632d93a4327, which made as_Media() stricter by failing on early DLCX. Related: SYS#5916 Change-Id: Ic5650a48eb3d90f2b42f16685178c97b54473429 --- M bsc/BSC_Tests.ttcn 1 file changed, 2 insertions(+), 1 deletion(-) Approvals: Jenkins Builder: Verified neels: Looks good to me, approved diff --git a/bsc/BSC_Tests.ttcn b/bsc/BSC_Tests.ttcn index a2add34..3a40d8d 100644 --- a/bsc/BSC_Tests.ttcn +++ b/bsc/BSC_Tests.ttcn @@ -5997,7 +5997,7 @@ f_create_mgcp_expect(ExpectCriteria:{omit,omit,omit}); f_MscConnHdlr_init(g_pars.media_nr, "127.0.0.2", "127.0.0.3", FR_AMR); - activate(as_Media()); + var default as_media := activate(as_Media()); var PDU_BSSAP ho_req := f_gen_handover_req(aoip_tla := g_pars.host_aoip_tla, cell_id_source := g_pars.cell_id_source, @@ -6130,6 +6130,7 @@ enc_PDU_ML3_MS_NW(l3_tx))); BSSAP.receive(tr_BSSMAP_HandoverComplete); + deactivate(as_media); setverdict(pass); } -- To view, visit https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27741 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-ttcn3-hacks Gerrit-Branch: master Gerrit-Change-Id: Ic5650a48eb3d90f2b42f16685178c97b54473429 Gerrit-Change-Number: 27741 Gerrit-PatchSet: 1 Gerrit-Owner: neels <nhofmeyr(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de> Gerrit-MessageType: merged

1 0

Change in osmo-ttcn3-hacks[master]: bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this...
by neels 12 Apr '22

12 Apr '22

neels has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27741 ) Change subject: bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this_bsc() ...................................................................... Patch Set 1: Code-Review+2 -- To view, visit https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27741 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-ttcn3-hacks Gerrit-Branch: master Gerrit-Change-Id: Ic5650a48eb3d90f2b42f16685178c97b54473429 Gerrit-Change-Number: 27741 Gerrit-PatchSet: 1 Gerrit-Owner: neels <nhofmeyr(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de> Gerrit-Comment-Date: Mon, 11 Apr 2022 23:54:55 +0000 Gerrit-HasComments: No Gerrit-Has-Labels: Yes Gerrit-MessageType: comment

1 0

Change in osmo-ttcn3-hacks[master]: bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this...
by neels 12 Apr '22

12 Apr '22

neels has uploaded this change for review. ( https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27741 ) Change subject: bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this_bsc() ...................................................................... bsc: fix sccplite test fallout: deactivate as_Media in f_ho_into_this_bsc() Deactivate as_Media() once the handover is completed, so that it does not fail on the expected MGCP DLCX from release. Fix fallout seen in these tests: SCCPlite BSC_Tests.TC_ho_into_this_bsc BSC_Tests.TC_ho_into_this_bsc_a5_0 BSC_Tests.TC_ho_into_this_bsc_a5_1 BSC_Tests.TC_ho_into_this_bsc_a5_3 BSC_Tests.TC_ho_into_this_bsc_a5_4 BSC_Tests.TC_ho_into_this_bsc_a5_1_3_no_chosen_enc_alg BSC_Tests.TC_ho_into_this_bsc_a5_1_3 BSC_Tests.TC_srvcc_eutran_to_geran BSC_Tests.TC_srvcc_eutran_to_geran_a5_3 BSC_Tests.TC_srvcc_eutran_to_geran_src_sai BSC_Tests.TC_srvcc_eutran_to_geran_forbid_fast_return BSC_Tests.TC_ho_into_this_bsc_sccp_cr_without_bssap All of these tests use f_ho_into_this_bsc(). (It is not clear to me why only the SCCPlite tests show the fallout, the AoIP should be similarly affected, but isn't.) The failures were introduced by recent merge of I0633f60f09d58802f6be0238ef41a632d93a4327, which made as_Media() stricter by failing on early DLCX. Related: SYS#5916 Change-Id: Ic5650a48eb3d90f2b42f16685178c97b54473429 --- M bsc/BSC_Tests.ttcn 1 file changed, 2 insertions(+), 1 deletion(-) git pull ssh://gerrit.osmocom.org:29418/osmo-ttcn3-hacks refs/changes/41/27741/1 diff --git a/bsc/BSC_Tests.ttcn b/bsc/BSC_Tests.ttcn index a2add34..3a40d8d 100644 --- a/bsc/BSC_Tests.ttcn +++ b/bsc/BSC_Tests.ttcn @@ -5997,7 +5997,7 @@ f_create_mgcp_expect(ExpectCriteria:{omit,omit,omit}); f_MscConnHdlr_init(g_pars.media_nr, "127.0.0.2", "127.0.0.3", FR_AMR); - activate(as_Media()); + var default as_media := activate(as_Media()); var PDU_BSSAP ho_req := f_gen_handover_req(aoip_tla := g_pars.host_aoip_tla, cell_id_source := g_pars.cell_id_source, @@ -6130,6 +6130,7 @@ enc_PDU_ML3_MS_NW(l3_tx))); BSSAP.receive(tr_BSSMAP_HandoverComplete); + deactivate(as_media); setverdict(pass); } -- To view, visit https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27741 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-ttcn3-hacks Gerrit-Branch: master Gerrit-Change-Id: Ic5650a48eb3d90f2b42f16685178c97b54473429 Gerrit-Change-Number: 27741 Gerrit-PatchSet: 1 Gerrit-Owner: neels <nhofmeyr(a)sysmocom.de> Gerrit-MessageType: newchange

1 0

Change in osmo-uecups[master]: Fix use-after-free by tun thread after tun obj destroyed
by fixeria 12 Apr '22

12 Apr '22

Attention is currently required from: osmith, pespin, daniel, lynxis lazus. fixeria has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-uecups/+/27739 ) Change subject: Fix use-after-free by tun thread after tun obj destroyed ...................................................................... Patch Set 2: Code-Review+1 -- To view, visit https://gerrit.osmocom.org/c/osmo-uecups/+/27739 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-uecups Gerrit-Branch: master Gerrit-Change-Id: Idf005359afb41d3413b09281a9ff937d5eafcc7c Gerrit-Change-Number: 27739 Gerrit-PatchSet: 2 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: daniel <dwillmann(a)sysmocom.de> Gerrit-Reviewer: fixeria <vyanitskiy(a)sysmocom.de> Gerrit-Reviewer: lynxis lazus <lynxis(a)fe80.eu> Gerrit-Reviewer: osmith <osmith(a)sysmocom.de> Gerrit-Attention: osmith <osmith(a)sysmocom.de> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Attention: daniel <dwillmann(a)sysmocom.de> Gerrit-Attention: lynxis lazus <lynxis(a)fe80.eu> Gerrit-Comment-Date: Mon, 11 Apr 2022 18:29:41 +0000 Gerrit-HasComments: No Gerrit-Has-Labels: Yes Gerrit-MessageType: comment

1 0

Change in osmo-uecups[master]: main.c: Improve logging
by fixeria 12 Apr '22

12 Apr '22

Attention is currently required from: pespin. fixeria has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-uecups/+/27738 ) Change subject: main.c: Improve logging ...................................................................... Patch Set 1: Code-Review+2 -- To view, visit https://gerrit.osmocom.org/c/osmo-uecups/+/27738 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-uecups Gerrit-Branch: master Gerrit-Change-Id: I6d9a355fef78906cee1ddf65ebb00a738b43c596 Gerrit-Change-Number: 27738 Gerrit-PatchSet: 1 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: fixeria <vyanitskiy(a)sysmocom.de> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Comment-Date: Mon, 11 Apr 2022 18:27:17 +0000 Gerrit-HasComments: No Gerrit-Has-Labels: Yes Gerrit-MessageType: comment

1 0

Change in osmo-uecups[master]: Fix use-after-free by tun thread after tun obj destroyed
by pespin 12 Apr '22

12 Apr '22

Attention is currently required from: osmith, fixeria, daniel, lynxis lazus. pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-uecups/+/27739 ) Change subject: Fix use-after-free by tun thread after tun obj destroyed ...................................................................... Patch Set 2: This change is ready for review. -- To view, visit https://gerrit.osmocom.org/c/osmo-uecups/+/27739 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-uecups Gerrit-Branch: master Gerrit-Change-Id: Idf005359afb41d3413b09281a9ff937d5eafcc7c Gerrit-Change-Number: 27739 Gerrit-PatchSet: 2 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: daniel <dwillmann(a)sysmocom.de> Gerrit-Reviewer: fixeria <vyanitskiy(a)sysmocom.de> Gerrit-Reviewer: lynxis lazus <lynxis(a)fe80.eu> Gerrit-Reviewer: osmith <osmith(a)sysmocom.de> Gerrit-Attention: osmith <osmith(a)sysmocom.de> Gerrit-Attention: fixeria <vyanitskiy(a)sysmocom.de> Gerrit-Attention: daniel <dwillmann(a)sysmocom.de> Gerrit-Attention: lynxis lazus <lynxis(a)fe80.eu> Gerrit-Comment-Date: Mon, 11 Apr 2022 18:19:44 +0000 Gerrit-HasComments: No Gerrit-Has-Labels: No Gerrit-MessageType: comment

1 0

Change in osmo-ttcn3-hacks[master]: GTPv2_Emulation: Increase reset_all_stats timeout
by pespin 12 Apr '22

12 Apr '22

pespin has uploaded this change for review. ( https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27740 ) Change subject: GTPv2_Emulation: Increase reset_all_stats timeout ...................................................................... GTPv2_Emulation: Increase reset_all_stats timeout Release of 256 tuns + threads takes quite a reasonable amount of time. Let's increase the timeout since reset_all_state can take around 10 seconds sometimes. Related: OS#5523 Change-Id: Ie01e07bd698cb5c386f757f4ec315f4892ad61cb --- M library/GTPv2_Emulation.ttcn 1 file changed, 3 insertions(+), 3 deletions(-) git pull ssh://gerrit.osmocom.org:29418/osmo-ttcn3-hacks refs/changes/40/27740/1 diff --git a/library/GTPv2_Emulation.ttcn b/library/GTPv2_Emulation.ttcn index f07afcb..5bb3b1e 100644 --- a/library/GTPv2_Emulation.ttcn +++ b/library/GTPv2_Emulation.ttcn @@ -353,9 +353,9 @@ } private function f_uecups_xceive(template (value) PDU_UECUPS tx, - template PDU_UECUPS rx_t := ?) + template PDU_UECUPS rx_t := ?, float time_out := 10.0) runs on GTPv2_Emulation_CT return PDU_UECUPS { - timer T := 10.0; + timer T := time_out; var UECUPS_RecvFrom mrf; UECUPS.send(t_UECUPS_Send(g_uecups_conn_id, tx)); @@ -399,7 +399,7 @@ g_uecups_conn_id := res.connId; /* clear all tunnel state in the daemon at start */ - f_uecups_xceive({reset_all_state := {}}, {reset_all_state_res:=?}); + f_uecups_xceive({reset_all_state := {}}, {reset_all_state_res:=?}, 30.0); } /* make sure we always pass incoming UECUPS indications whenever receiving fom the UECUPS port */ -- To view, visit https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/27740 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-ttcn3-hacks Gerrit-Branch: master Gerrit-Change-Id: Ie01e07bd698cb5c386f757f4ec315f4892ad61cb Gerrit-Change-Number: 27740 Gerrit-PatchSet: 1 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-MessageType: newchange

1 0

Change in osmo-uecups[master]: main.c: Improve logging
by pespin 12 Apr '22

12 Apr '22

pespin has uploaded this change for review. ( https://gerrit.osmocom.org/c/osmo-uecups/+/27738 ) Change subject: main.c: Improve logging ...................................................................... main.c: Improve logging Change-Id: I6d9a355fef78906cee1ddf65ebb00a738b43c596 --- M daemon/main.c 1 file changed, 4 insertions(+), 0 deletions(-) git pull ssh://gerrit.osmocom.org:29418/osmo-uecups refs/changes/38/27738/1 diff --git a/daemon/main.c b/daemon/main.c index 328d074..76aeab5 100644 --- a/daemon/main.c +++ b/daemon/main.c @@ -72,6 +72,8 @@ /* kill the specified subprocess and forget about it */ static void subprocess_destroy(struct subprocess *p, int signal) { + LOGCC(p->cups_client, LOGL_DEBUG, "Kill subprocess pid %llu with signal %u\n", + (unsigned long long)p->pid, signal); kill(p->pid, signal); llist_del(&p->list); talloc_free(p); @@ -478,6 +480,7 @@ struct subprocess *p, *p2; json_t *jres; + LOGCC(cc, LOGL_DEBUG, "Destroying all tunnels\n"); pthread_rwlock_wrlock(&d->rwlock); llist_for_each_entry_safe(t, t2, &d->gtp_tunnels, list) { _gtp_tunnel_destroy(t); @@ -485,6 +488,7 @@ pthread_rwlock_unlock(&d->rwlock); /* no locking needed as this list is only used by main thread */ + LOGCC(cc, LOGL_DEBUG, "Destroying all subprocesses\n"); llist_for_each_entry_safe(p, p2, &d->subprocesses, list) { subprocess_destroy(p, SIGKILL); } -- To view, visit https://gerrit.osmocom.org/c/osmo-uecups/+/27738 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-uecups Gerrit-Branch: master Gerrit-Change-Id: I6d9a355fef78906cee1ddf65ebb00a738b43c596 Gerrit-Change-Number: 27738 Gerrit-PatchSet: 1 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-MessageType: newchange

1 0

Change in osmo-uecups[master]: Fix use-after-free by tun thread after tun obj destroyed
by pespin 12 Apr '22

12 Apr '22

pespin has uploaded this change for review. ( https://gerrit.osmocom.org/c/osmo-uecups/+/27739 ) Change subject: Fix use-after-free by tun thread after tun obj destroyed ...................................................................... Fix use-after-free by tun thread after tun obj destroyed The main thread calls pthread_cancel before freeing the tun object. However, pthread_cancel doesn't kill the thread synchronously (man pthread_cancel). Hence, the tun thread may still be running for a while after the tun object is/has been(ing) freed. Let's avoid this by making sure the thread is stopped before freeing the object. To accomplish it, we must wait for the thread to be cancelled. A cleanup routie is added which will signal the "tun_released" message to the main thread through an osmo_itq, which will then free the object (since talloc context is managed by the main thread). Related: SYS#5523 Change-Id: Idf005359afb41d3413b09281a9ff937d5eafcc7c --- M daemon/daemon_vty.c M daemon/internal.h M daemon/main.c M daemon/tun_device.c 4 files changed, 85 insertions(+), 15 deletions(-) git pull ssh://gerrit.osmocom.org:29418/osmo-uecups refs/changes/39/27739/1 diff --git a/daemon/daemon_vty.c b/daemon/daemon_vty.c index 3c44ed1..cfbe421 100644 --- a/daemon/daemon_vty.c +++ b/daemon/daemon_vty.c @@ -99,7 +99,7 @@ vty_out(vty, "Cannot destrory non-existant TUN%s", VTY_NEWLINE); return CMD_WARNING; } - _tun_device_deref_destroy(tun); + _tun_device_deref_release(tun); pthread_rwlock_unlock(&g_daemon->rwlock); return CMD_SUCCESS; diff --git a/daemon/internal.h b/daemon/internal.h index 09ba52e..fc5708d 100644 --- a/daemon/internal.h +++ b/daemon/internal.h @@ -7,6 +7,7 @@ #include <sys/socket.h> #include <osmocom/core/linuxlist.h> #include <osmocom/core/write_queue.h> +#include <osmocom/core/it_q.h> #include <osmocom/core/utils.h> struct nl_sock; @@ -84,6 +85,13 @@ /*********************************************************************** * TUN Device ***********************************************************************/ +/* Message sent tun thread -> main thread through osmo_itq */ +struct gtp_daemon_itq_msg { + struct llist_head list; + struct { + struct tun_device *tun; + } tun_released; /* tun became stopped and can be freed */ +}; struct tun_device { /* entry in global list */ @@ -110,6 +118,9 @@ /* the thread handling Rx from the tun fd */ pthread_t thread; + + /* Used to store messages to be sent to main thread, since tun thread doesn't allocate through talloc */ + struct gtp_daemon_itq_msg itq_msg; }; struct tun_device * @@ -121,9 +132,10 @@ struct tun_device * _tun_device_find(struct gtp_daemon *d, const char *devname); -void _tun_device_deref_destroy(struct tun_device *tun); +void _tun_device_destroy(struct tun_device *tun); bool _tun_device_release(struct tun_device *tun); +void _tun_device_deref_release(struct tun_device *tun); bool tun_device_release(struct tun_device *tun); @@ -222,6 +234,12 @@ struct osmo_stream_srv_link *cups_link; struct osmo_signalfd *signalfd; + /* inter-thread queue between main thread and workers, pass struct gtp_daemon_itq_msg: */ + struct osmo_it_q *itq; + + /* Number of tunnels in progrress of being released: */ + unsigned int reset_all_state_tun_remaining; + struct { char *cups_local_ip; uint16_t cups_local_port; diff --git a/daemon/main.c b/daemon/main.c index 76aeab5..c8f63f6 100644 --- a/daemon/main.c +++ b/daemon/main.c @@ -58,6 +58,7 @@ /* client socket */ struct osmo_stream_srv *srv; char sockname[OSMO_SOCK_NAME_MAXLEN]; + bool reset_all_state_res_pending; }; struct subprocess { @@ -493,8 +494,12 @@ subprocess_destroy(p, SIGKILL); } - jres = gen_uecups_result("reset_all_state_res", "OK"); - cups_client_tx_json(cc, jres); + if (d->reset_all_state_tun_remaining == 0) { + jres = gen_uecups_result("reset_all_state_res", "OK"); + cups_client_tx_json(cc, jres); + } else { + cc->reset_all_state_res_pending = true; + } return 0; } @@ -669,6 +674,31 @@ } } +static void gtp_daemon_itq_read_cb(struct osmo_it_q *q, struct llist_head *item) +{ + struct gtp_daemon *d = (struct gtp_daemon *)q->data; + struct gtp_daemon_itq_msg *itq_msg = container_of(item, struct gtp_daemon_itq_msg, list); + + LOGP(DTUN, LOGL_DEBUG, "Rx new itq message from %s\n", + itq_msg->tun_released.tun->devname); + + _tun_device_destroy(itq_msg->tun_released.tun); + if (d->reset_all_state_tun_remaining > 0) { + d->reset_all_state_tun_remaining--; + if (d->reset_all_state_tun_remaining == 0) { + struct cups_client *cc; + llist_for_each_entry(cc, &d->cups_clients, list) { + json_t * jres; + if (!cc->reset_all_state_res_pending) + continue; + cc->reset_all_state_res_pending = false; + jres = gen_uecups_result("reset_all_state_res", "OK"); + cups_client_tx_json(cc, jres); + } + } + } +} + static struct gtp_daemon *gtp_daemon_alloc(void *ctx) { struct gtp_daemon *d = talloc_zero(ctx, struct gtp_daemon); @@ -682,6 +712,9 @@ pthread_rwlock_init(&d->rwlock, NULL); d->main_thread = pthread_self(); + d->itq = osmo_it_q_alloc(d, "itq", 4096, gtp_daemon_itq_read_cb, d); + osmo_fd_register(&d->itq->event_ofd); + INIT_LLIST_HEAD(&d->cups_clients); d->cfg.cups_local_ip = talloc_strdup(d, "localhost"); diff --git a/daemon/tun_device.c b/daemon/tun_device.c index 7d1948f..86be67f 100644 --- a/daemon/tun_device.c +++ b/daemon/tun_device.c @@ -120,6 +120,14 @@ return 0; } +static void tun_device_pthread_cleanup_routine(void *data) +{ + struct tun_device *tun = data; + LOGTUN(tun, LOGL_DEBUG, "pthread_cleanup\n"); + int rc = osmo_it_q_enqueue(tun->d->itq, &tun->itq_msg, list); + OSMO_ASSERT(rc == 0); +} + /* one thread for reading from each TUN device (TUN -> GTP encapsulation) */ static void *tun_device_thread(void *arg) { @@ -136,6 +144,8 @@ gtph->flags = 0x30; gtph->type = GTP_TPDU; + pthread_cleanup_push(tun_device_pthread_cleanup_routine, tun); + while (1) { struct gtp_tunnel *t; struct pkt_info pinfo; @@ -187,6 +197,8 @@ exit(1); } } + + pthread_cleanup_pop(tun_device_pthread_cleanup_routine); } static int tun_open(int flags, const char *name) @@ -376,24 +388,24 @@ return tun; } -/* UNLOCKED hard/forced destroy; caller must make sure references are cleaned up */ -static void _tun_device_destroy(struct tun_device *tun) +/* UNLOCKED hard/forced destroy; caller must make sure references are cleaned + * up, and tun thread is stopped beforehand by calling + * _tun_device_{deref_}release */ +void _tun_device_destroy(struct tun_device *tun) { /* talloc is not thread safe, all alloc/free must come from main thread */ ASSERT_MAIN_THREAD(tun->d); + LOGTUN(tun, LOGL_INFO, "Destroying\n"); - pthread_cancel(tun->thread); - llist_del(&tun->list); if (tun->netns_name) close(tun->netns_fd); close(tun->fd); nl_socket_free(tun->nl); - LOGTUN(tun, LOGL_INFO, "Destroying\n"); talloc_free(tun); } -/* UNLOCKED remove all objects referencing this tun and then destroy */ -void _tun_device_deref_destroy(struct tun_device *tun) +/* UNLOCKED remove all objects referencing this tun and then start async tun release procedure */ +void _tun_device_deref_release(struct tun_device *tun) { struct gtp_daemon *d = tun->d; char *devname = talloc_strdup(d, tun->devname); @@ -412,12 +424,12 @@ * check if the tun can still be found in the list */ tun2 = _tun_device_find(d, devname); if (tun2 && tun2 == tun) - _tun_device_destroy(tun2); + _tun_device_release(tun2); talloc_free(devname); } -/* UNLOCKED release a reference; destroy if refcount drops to 0 */ +/* UNLOCKED release a reference; start async tun release procedure if refcount drops to 0 */ bool _tun_device_release(struct tun_device *tun) { bool released = false; @@ -427,10 +439,17 @@ tun->use_count--; if (tun->use_count == 0) { - _tun_device_destroy(tun); + LOGTUN(tun, LOGL_INFO, "Releasing\n"); + llist_del(&tun->list); + tun->itq_msg.tun_released.tun = tun; + tun->d->reset_all_state_tun_remaining++; + /* We cancel the thread: the pthread_cleanup routing will send a message + * back to us (main thread) when finally cancelled. */ + pthread_cancel(tun->thread); released = true; - } else + } else { LOGTUN(tun, LOGL_DEBUG, "Release; new use_count=%lu\n", tun->use_count); + } return released; } -- To view, visit https://gerrit.osmocom.org/c/osmo-uecups/+/27739 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-uecups Gerrit-Branch: master Gerrit-Change-Id: Idf005359afb41d3413b09281a9ff937d5eafcc7c Gerrit-Change-Number: 27739 Gerrit-PatchSet: 1 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-MessageType: newchange

1 0

Change in osmo-uecups[master]: main.c: Fix typo in comment
by pespin 12 Apr '22

12 Apr '22

pespin has submitted this change. ( https://gerrit.osmocom.org/c/osmo-uecups/+/27736 ) Change subject: main.c: Fix typo in comment ...................................................................... main.c: Fix typo in comment Change-Id: I6fb65b0468ebab8c35bb918f1d04d62b641edf69 --- M daemon/main.c 1 file changed, 1 insertion(+), 1 deletion(-) Approvals: Jenkins Builder: Verified fixeria: Looks good to me, approved diff --git a/daemon/main.c b/daemon/main.c index 788ed70..39d60e6 100644 --- a/daemon/main.c +++ b/daemon/main.c @@ -370,7 +370,7 @@ child_terminated(d, fdsi->ssi_pid, fdsi->ssi_status); /* it is known that classic signals coalesce: If you get multiple signals of the - * same type before a process is scheduled, the subsequent signaals are dropped. This + * same type before a process is scheduled, the subsequent signals are dropped. This * makes sense for SIGINT or something like this, but for SIGCHLD carrying the PID of * the terminated process, it doesn't really. Linux had the chance to fix this when * introducing signalfd() - but the developers decided not to fix it. So the signalfd_siginfo -- To view, visit https://gerrit.osmocom.org/c/osmo-uecups/+/27736 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-uecups Gerrit-Branch: master Gerrit-Change-Id: I6fb65b0468ebab8c35bb918f1d04d62b641edf69 Gerrit-Change-Number: 27736 Gerrit-PatchSet: 1 Gerrit-Owner: pespin <pespin(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: fixeria <vyanitskiy(a)sysmocom.de> Gerrit-Reviewer: pespin <pespin(a)sysmocom.de> Gerrit-MessageType: merged

1 0

2026

2025

2024

2023

2022

gerrit-log