[roll] Roll fuchsia [loadbench] Reset trace buffer before running unittests
The loadbench unittests sometimes invoke zx_ktrace_read and verify
that the resulting trace data can be successfully parse.
This generally works, but can fail in the following scenario:
1. A prior test starts tracing, which causes all of the metadata to
be emitted.
2. The previous test consumes some of this trace data.
3. The previous test stops tracing, but does not clear out all of
the trace buffer contents.
4. This test then reads the contents of the trace buffer, which lacks
some of the metadata needed to parse the contents.
This only shows up when streaming data because that's the only way
a previous test could possibly consume the metadata. In saturating
mode, the ktrace implementation ensures that metadata is emitted
before reading.
Original-Bug: 404539312
Original-Reviewed-on: https://fuchsia-review.googlesource.com/c/fuchsia/+/1275677
Original-Revision: 57409c89c22ac68bf790f13ca90b58a81b5bd740
GitOrigin-RevId: 0cbf6598ef6deed32af61f51ff65242fdbfde1d0
Change-Id: I21886dc0eafbd83485bc2256df27212ab3ddc0b8
diff --git a/stem b/stem
index b874665..dde28c6 100644
--- a/stem
+++ b/stem
@@ -8,7 +8,7 @@
root=""
remote="https://fuchsia.googlesource.com/fuchsia"
gerrithost="https://fuchsia-review.googlesource.com"
- revision="cd270aa0a1332d3a1967eb02500c922392675f9c"
+ revision="57409c89c22ac68bf790f13ca90b58a81b5bd740"
/>
</imports>
<projects>
