Cleanup: remove "disabled" flag for new release cycle
A message that is not in the archive (sent off-list or not yet received): [email protected]
PATCH 2/2 Cleanup: remove "disabled" flag for new release cycle
--- src/bin/lttng-sessiond/trace-ust.c | 5 ----- src/bin/lttng-sessiond/ust-app.c | 21 --------------------- 2 files changed, 26 deletions(-)
src/bin/lttng-sessiond/trace-ust.c +0 -5
@@ -419,11 +419,6 @@ struct ltt_ust_event *trace_ust_create_event(struct lttng_event *ev,ERR ("Unknown ust loglevel type (%d)" , ev->loglevel_type); goto error_free_event; } -/* -* Fix for enabler race. Enable is now done explicitly by -* sessiond after setting filter. -*/ - lue->attr.disabled =1 ;/* Same layout. */ lue->filter_expression = filter_expression;
src/bin/lttng-sessiond/ust-app.c +0 -21
@@ -1471,27 +1471,6 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, } goto error; } - }else { - ret =disable_ust_event (app, ua_sess, ua_event); -if (ret <0 ) { -/* -* If we hit an EPERM, something is wrong with our disable call. If -* we get an EEXIST, there is a problem on the tracer side since we -* just created it. -*/ -switch (ret) { -case -LTTNG_UST_ERR_PERM : -/* Code flow problem */ -assert (0 ); -case -LTTNG_UST_ERR_EXIST : -/* It's OK for our use case. */ - ret =0 ; -break ; -default : -break ; - } - goto error; - } } error:
Signature · 1 line
2.1.1
A message that is not in the archive (sent off-list or not yet received): [email protected]
Jérémie Galarneau
AppliedMerged, thanks!
Jérémie
Quoted 57 lines · Mathieu Desnoyers
Signed-off-by Mathieu Desnoyers <[email protected]>--- src/bin/lttng-sessiond/trace-ust.c | 5 ----- src/bin/lttng-sessiond/ust-app.c | 21 --------------------- 2 files changed, 26 deletions(-)src/bin/lttng-sessiond/trace-ust.c +0
-5@@ -419,11 +419,6 @@ struct ltt_ust_event *trace_ust_create_event(struct lttng_event *ev,ERR ("Unknown ust loglevel type (%d)" , ev->loglevel_type); goto error_free_event; } -/* -* Fix for enabler race. Enable is now done explicitly by -* sessiond after setting filter. -*/ - lue->attr.disabled =1 ;/* Same layout. */ lue->filter_expression = filter_expression;src/bin/lttng-sessiond/ust-app.c +0
-22@@ -1471,27 +1471,6 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, } goto error; } - }else { - ret =disable_ust_event (app, ua_sess, ua_event); -if (ret <0 ) { -/* -* If we hit an EPERM, something is wrong with our disable call. If -* we get an EEXIST, there is a problem on the tracer side since we -* just created it. -*/ -switch (ret) { -case -LTTNG_UST_ERR_PERM : -/* Code flow problem */ -assert (0 ); -case -LTTNG_UST_ERR_EXIST : -/* It's OK for our use case. */ - ret =0 ; -break ; -default : -break ; - } - goto error; - } } error: --2.1.1
Signature · 8 lines
Jérémie Galarneau EfficiOS Inc. http://www.efficios.com _______________________________________________ lttng-dev mailing list [email protected] http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev
Re: [PATCH lttng-tools 1/2] Fix: filter attach vs event enable race
In reply to a message not in the archiveMerged all the way back to 2.4, thanks!
Jérémie
Quoted 83 lines · Mathieu Desnoyers
In order to correctly handle the use-case where events are enabled _after_ trace is started, and _after_ applications are already being traced, the event should be created in a "disabled" state, so that it does not trace events until its filter is attached.
This fix needs to be done both in lttng-tools and lttng-ust. In order to keep ABI compatibility between tools and ust within a stable release cycle, we introduce a new "disabled" within struct
lttng_ust_eventpadding (previously zeroed). Newer LTTng-UST checks this flag, and fallback on the old racy behavior (enabling the event on creation) if it is unset.Therefore, old session daemon works with newer lttng-ust of the same stable release, and vice-versa. However, building lttng-tools requires an upgraded lttng-ust, which contains the communication protocol with the new "disabled" field.
This patch should be backported to stable-2.4, stable-2.5, stable-2.6 branches.
Signed-off-by Mathieu Desnoyers <[email protected]>--- src/bin/lttng-sessiond/trace-ust.c | 5 +++++ src/bin/lttng-sessiond/ust-app.c | 27 ++++++++++++++++++++++++++- 2 files changed, 31 insertions(+), 1 deletion(-)src/bin/lttng-sessiond/trace-ust.c +5
-0@@ -419,6 +419,11 @@ struct ltt_ust_event *trace_ust_create_event(struct lttng_event *ev,ERR ("Unknown ust loglevel type (%d)" , ev->loglevel_type); goto error_free_event; } +/* +* Fix for enabler race. Enable is now done explicitly by +* sessiond after setting filter. +*/ + lue->attr.disabled =1 ;/* Same layout. */ lue->filter_expression = filter_expression;src/bin/lttng-sessiond/ust-app.c +26
-2@@ -1446,7 +1446,32 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, }/* If event not enabled, disable it on the tracer */ -if (ua_event->enabled ==0 ) { +if (ua_event->enabled) { +/* +* We now need to explicitly enable the event, since it +* is now disabled at creation. +*/ + ret =enable_ust_event (app, ua_sess, ua_event); +if (ret <0 ) { +/* +* If we hit an EPERM, something is wrong with our enable call. If +* we get an EEXIST, there is a problem on the tracer side since we +* just created it. +*/ +switch (ret) { +case -LTTNG_UST_ERR_PERM : +/* Code flow problem */ +assert (0 ); +case -LTTNG_UST_ERR_EXIST : +/* It's OK for our use case. */ + ret =0 ; +break ; +default : +break ; + } + goto error; + } + }else { ret =disable_ust_event (app, ua_sess, ua_event);if (ret <0 ) {/* -- 2.1.1
Signature · 8 lines
Jérémie Galarneau EfficiOS Inc. http://www.efficios.com _______________________________________________ lttng-dev mailing list [email protected] http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev