Add more debugging for RFH created with same SIG as BFCache
We're seeing cases where a speculative RFH is created on a non-bfcache
history navigation, and the RFH shares SIG with a BFCached pages. This
should not be possible because we evict BFCached pages with related
SiteInstance on such navigations, so add some debugging to investigate.
Also, remove the previous DwC added for this that gets hit on proxy
creation, since it's triggered a lot and is blocking Beta release.
Bug: 368720558, 369714361
Change-Id: I5a6d9bd4ad264ecb984156f51fad368229f99edc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5892154
Reviewed-by: Alex Moshchuk <[email protected]>
Commit-Queue: Rakina Zata Amni <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1360933}
diff --git a/content/browser/renderer_host/navigation_controller_impl.cc b/content/browser/renderer_host/navigation_controller_impl.cc
index d5cfee8..8ad91545 100644
--- a/content/browser/renderer_host/navigation_controller_impl.cc
+++ b/content/browser/renderer_host/navigation_controller_impl.cc
@@ -3332,6 +3332,15 @@
// BrowsingInstance/SiteInstance, which is unsupported by
// RenderFrameHostManager::CommitPending(). To avoid this conundrum, we evict
// A2 from the cache.
+ SCOPED_CRASH_KEY_NUMBER("rvh_double", "pending_entry_si",
+ pending_entry_->site_instance()
+ ? pending_entry_->site_instance()->GetId().value()
+ : -1);
+ SCOPED_CRASH_KEY_NUMBER(
+ "rvh_double", "pending_entry_bi",
+ pending_entry_->site_instance()
+ ? pending_entry_->site_instance()->GetBrowsingInstanceId().value()
+ : -1);
if (pending_entry_->site_instance()) {
back_forward_cache_.EvictFramesInRelatedSiteInstances(
pending_entry_->site_instance());
