summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorMaxim Kuvyrkov <maxim.kuvyrkov@linaro.org>2019-03-15 06:19:22 +0000
committerMaxim Kuvyrkov <maxim.kuvyrkov@linaro.org>2019-03-15 06:19:22 +0000
commit82cbd12c5395b01d1a4ba3b7b62611d3065f7122 (patch)
treefaf5238da1e8fe79f0154aa97d689be7c5c4b903
parent7b44d0602565d7a6da313506637610318ce2bed4 (diff)
tcwg_kernel-bisect.sh: Always add $first_bad to interesting-commits.git
In case of last_good being bad we often have first_bad just on the boundary of bisection range, and last_good outside of bisection range, so we end up ignoring last_good commit in bisections of similar configurations. Adding first_bad commit will reduce bisection time in such cases. Change-Id: I33ab875ec751a161452634c8b989a39b4e8eb05a
-rwxr-xr-xtcwg_kernel-bisect.sh12
1 files changed, 5 insertions, 7 deletions
diff --git a/tcwg_kernel-bisect.sh b/tcwg_kernel-bisect.sh
index 12d48f68..5129d1f5 100755
--- a/tcwg_kernel-bisect.sh
+++ b/tcwg_kernel-bisect.sh
@@ -309,17 +309,15 @@ if [ x"$first_bad" != x"" ]; then
break
fi
done
+ # Add both $last_good and $first_bad to interesting commits.
+ push_interesting_commits $last_good $first_bad
if [ x"$res" = x"0" ]; then
- # Success!
- push_interesting_commits $first_bad $last_good
- # Touch $artifacts/first-bad as a marker of successful bisect.
+ # Success! Touch $artifacts/first-bad as a marker of successful bisect.
echo $first_bad > $artifacts/first-bad
else
# It seems $last_good was on a path that tested good, even though
- # it itself is bad. We re-trigger the bisection job with updated
- # parameters.
- push_interesting_commits $last_good
-
+ # it itself is bad.
+ #
# We need to be careful to avoid re-trigger loops, so verify that
# last_good is an ancestor of bad_rev.
assert git merge-base --is-ancestor $last_good $bad_rev