I filed  Issue 679755  about this not too long ago about that.
I never realized that it has been duped against Issue 661257.
I didn't manage to look at the bug content bug but the relationship between the two is not clear to me.
That bug points to https://codereview.chromium.org/2459733002/diff/240001/components/discardable_memory/service/discardable_shared_memory_manager.cc#newcode48

I need to read all that bug and both the CLs, I'll do as soon as I have some spare moment. But maybe +ssid has a clearer view of the overall situation.
In the meantime thanks a lot for the bisect!

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/efed310cc305f0d89f9d1aa49fb5e0d4ec9ff2cd

commit efed310cc305f0d89f9d1aa49fb5e0d4ec9ff2cd
Author: ssid <ssid@chromium.org>
Date: Tue Feb 07 05:29:58 2017

[memory-infra] Make client discardable segments non-weak

The discardable segments shared between the processes do not add the
same global dump ids since the process id of the client is not available
in the manager. Temporarily we do not create weak dumps since the dump
provider is broken.

BUG=661257,  687399 

Review-Url: https://codereview.chromium.org/2668193002
Cr-Commit-Position: refs/heads/master@{#448554}

[modify] https://crrev.com/efed310cc305f0d89f9d1aa49fb5e0d4ec9ff2cd/components/discardable_memory/common/discardable_shared_memory_heap.cc

So this must not be blocking the other bug now since we should have the discardable provider report details on renderer after the CL.
Though the issue is not fully fixed. The issue 661257 tracks the rest of the missing bits. So closing this issue.