Fix: sync event enablers before choosing header type
authorMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Wed, 17 Oct 2018 19:41:37 +0000 (15:41 -0400)
committerMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Wed, 17 Oct 2018 21:02:22 +0000 (17:02 -0400)
commit4ef81e4ac900975a4627ec36e3693ccbce727a62
treea980293948a5912292b1ea542268aa192b00ff68
parent7502f47afaf537bbb67050b17ea095bb3976bf91
Fix: sync event enablers before choosing header type

On session start, we should allocate the event IDs before figuring
out the number of events per channel and select the proper header
type.

Without this, the number of events is always perceived to be 0,
which selects the "compact" header type. For a channel containing
many events (e.g. enable-event -k -a), this selects an inefficient
header type. With this fix, it selects the "large" header type,
which is more appropriate for a larger number of event IDs.

This will lead to a reduced trace throughput for tracing workloads
that have many events.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
lttng-events.c
This page took 0.025169 seconds and 4 git commands to generate.