ISS: fix re-connection initiated by IcingSearchEngine calls
After max retries of re-conn triggered by vm binder conn death, the next
re-conn is triggered by any incoming IcingSearchEngine calls. In the
current code of this re-conn, we didn't replace all vm icing instances with new ones. So when next IcingSearchEngine call happens, we restart and reconnect to the VM but still call old vm icing instances, which results in DOE.
Bug: 424266583
Test: adb root; for i in {1..20}; do adb shell kill -9 "$(adb shell pidof crosvm_isolated_storage_service_vm)"; sleep 0.1; done; adb shell dumpsys app_search --query t
Test: atest AppSearchSessionCtsTest
Test: atest --rebuild-module-info CtsAppSearchTestCases AppSearchServicesTests
AppSearchMockingServicesTests AppSearchCoreTests
Flag: com.android.appsearch.flags.enable_isolated_storage
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:a15e92a014264ff22ae0b05dff96e3a9b15c170b)
Merged-In: I07d3c37e46bd55d8097d1a58c35d91f6f54741aa
Change-Id: I07d3c37e46bd55d8097d1a58c35d91f6f54741aa
1 file changed
