I think this is the first failing build:
https://ci.chromium.org/buildbot/chromium.clang/CrWinClang%28shared%29/948
The tests passed in this build:
https://ci.chromium.org/buildbot/chromium.clang/CrWinClang%28shared%29/947

I think the test failed when it landed: https://chromium.googlesource.com/chromium/src/+/0d4a27248ff0b9aace5be15c041cd21e9a0443fa

Verified locally that the test has been failing in this config since it landed.

The code looks like this:

scoped_refptr<SimpleFontData> CSSFontFaceSource::GetFontData(
    const FontDescription& font_description,
    const FontSelectionCapabilities& font_selection_capabilities) {
  // If the font hasn't loaded or an error occurred, then we've got nothing.
  if (!IsValid())
    return nullptr;

  if (IsLocal()) {
    // We're local. Just return a SimpleFontData from the normal cache.
    return CreateFontData(font_description, font_selection_capabilities);
  }

  FontCacheKey key = font_description.CacheKey(FontFaceCreationParams());

  scoped_refptr<SimpleFontData>& font_data =
      font_data_table_.insert(key, nullptr).stored_value->value; // << ---- 
  if (!font_data)
    font_data = CreateFontData(font_description, font_selection_capabilities);

  font_cache_key_age.PrependOrMoveToFirst(key);
  PruneOldestIfNeeded();

  DCHECK_LE(font_data_table_.size(), kMaxCachedFontData);
  // No release, because fontData is a reference to a RefPtr that is held in the
  // font_data_table_.
  return font_data;
}

I think what's happening is that MSVC extends the lifetime of the object returned by font_data_table_.insert() until the end of the function. When that object, a WTF::HashTableAddResult, destructs after the PruneOldestIfNeeded(), it trips the dcheck.

A potential fix: https://chromium-review.googlesource.com/#/c/chromium/src/+/921942

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ce5486ae573d41c2d29e1e58308b99355595b0be

commit ce5486ae573d41c2d29e1e58308b99355595b0be
Author: Hans Wennborg <hans@chromium.org>
Date: Thu Feb 15 18:43:23 2018

Fix CSSFontFaceSourceTest.UnboundedGrowth test in MSVC builds

The WTF::HashTableAddResult from calling font_data_table_.insert()
was being lifetime-extended in MSVC builds to the end of the function,
causing the HashTableAddResult's destructor to run after the hash table
had been changed, causing a dcheck.

Bug:  812600 
Change-Id: Ifa2b05ef92370c7ba043eea4258f4086274d1871
Reviewed-on: https://chromium-review.googlesource.com/921942
Reviewed-by: Dominik Röttsches <drott@chromium.org>
Reviewed-by: Emil A Eklund <eae@chromium.org>
Commit-Queue: Emil A Eklund <eae@chromium.org>
Cr-Commit-Position: refs/heads/master@{#537085}
[modify] https://crrev.com/ce5486ae573d41c2d29e1e58308b99355595b0be/third_party/WebKit/Source/core/css/CSSFontFaceSource.cpp

 Issue 799589  has been merged into this issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d93525cde8ce7151e584db7bd690a09c06b83715

commit d93525cde8ce7151e584db7bd690a09c06b83715
Author: Reid Kleckner <rnk@google.com>
Date: Thu Mar 01 01:26:11 2018

Fix dangling reference into font data hashtable

It's unlikely that pruning the cache will resize the hash table,
but this makes a copy of the scoped_refptr just in case.

R=drott@chromium.org
BUG= 812600 

Change-Id: I5a6256b9079e9c7f36249cadbb679189f0248afd
Reviewed-on: https://chromium-review.googlesource.com/852531
Reviewed-by: Hans Wennborg <hans@chromium.org>
Reviewed-by: Dominik Röttsches <drott@chromium.org>
Commit-Queue: Reid Kleckner <rnk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#539970}
[modify] https://crrev.com/d93525cde8ce7151e584db7bd690a09c06b83715/third_party/WebKit/Source/core/css/CSSFontFaceSource.cpp