I believe this is page-specific, so re-routing to offline pages. 

Dmitry, would you be able to take a look / route?

Have checked this with Nexus 9(7.1.1/NMF26F),Samsung Galaxy J7(SM-J700F)(6.0.1/MMB29K),Google Pixel(7.1.1/NMF26O). Able to reproduce the issue on latest M-57.0.2963.0 with few links of www.foxsports.com

But unable to reproduce the issue with other sites. 

Regression details:

1.Good build:57.0.2956.0
  Bad build:57.0.2957.0
2. This issue is not reproducible on latest M-56.0.2924.41

Bisect range:
https://chromium.googlesource.com/chromium/src/+log/57.0.2956.0..57.0.2957.0?pretty=fuller&n=10000

Bisect script is pointing to: https://chromium.googlesource.com/chromium/src/+/4f6b28049376a0e309fda4cbeb6e4d7aac98965c

Peter, could you please take a look? From the first video, it looks like RequestCoordinator picks up the request, then fails fast - and this repeats 3 times until "Download failed.". Might be a failure of prerenderer on this specific site?

It's hard to repro this since my new tab page isn't offering fox sports.  It could fail quickly if the network drops while the prerenderer is trying to fetch a resource early in the game (like the HTML file).

The second video seems to show only pausing already loading pages.
It can in fact take awhile for pages to load. 10 minutes might not be unreasonable for a big page over a slow connection (such as Gin2G or Gin2G-poor)

It looks like the bisect is pointing at the change which enabled the download link functionality, but the feature should not have been present before that, so the listed change is not particularly suspect.

Do other links to fox sports also fail the same way, and does it repro consistently?

I am able to reproduce this issue with Nexus 9(7.1.1/NMF26F),Samsung Galaxy J7(SM-J700F)(6.0.1/MMB29K),Google Pixel(7.1.1/NMF26O) on latest M-57.0.2987.8 with below mentioned steps,

1. Launch chrome> Go to www.foxsports.com.
2. Long tap on any link> Tap on 'Download link'.
3. Observe.

Repro-ed on Nexus 7 with ToT. Saw three download attempts each failing in prerenderer with FinalStatus code 47: FINAL_STATUS_NEW_NAVIGATION_ENTRY

Potential fix might be to revise the following check in DidNavigateMainFrame() to not apply to ORIGIN_OFFLINE:  https://cs.chromium.org/chromium/src/chrome/browser/prerender/prerender_contents.cc?rcl=1485260312&l=588
since ORIGIN_OFFLINE does not allow navigation to prerendered pages and so does not consume navigation history.

Tested that idea locally and was able to successfully download two fox sports articles that failed previously. Both did take 2 attempts to complete (one had FinalStatus 22 failure on first attempt and the other had prerenderer timeout (FinalStatus 1)). Will get CL out for review.

https://codereview.chromium.org/2656653002 out for review

URGENT - PLEASE REVIEW ASAP

Greetings from the release team!

This bug is marked as an M-57 beta blocker, which means it needs to be fixed on trunk by THIS FRIDAY, Jan 27 in order to be merged back to the M57 branch on time.  Please prioritize fixing this issue.

Unsure if this bug should block the beta release, or know it should block but you won't be able to fix it in time?  CC me to this bug and we can discuss.

If you're absolutely sure this should not block beta, the bug can be punted to stable (by changing ReleaseBlock-Beta to ReleaseBlock-Stable), or if the bug should not block the release at all simply remove the release block tag.

Thanks,
Alex

1. The fix for this just landed.
2. It only affects a small fraction of sites, we think it can be punted to Stable (but it is M57 still).

By the way, there may be a secondary issue with downloading some of these fox sports article pages. Some seem to be failing with prerenderer FINAL_STATUS_UNSUPPORTED_SCHEME code on good WIFI and I am investigating further. 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3ae6f734f3566484634a068cbe663bbccc24937c

commit 3ae6f734f3566484634a068cbe663bbccc24937c
Author: dougarnett <dougarnett@chromium.org>
Date: Thu Jan 26 18:22:29 2017

Suppress NEW_NAVIGATION_ENTRY prerenderer failures for ORIGIN_OFFLINE

We have noticed some FINAL_STATUS_NEW_NAVIGATION_ENTRY failures in UMA
for offline pages but didn't know how to trigger them.  Bug 675767  has
identified a site (foxsports.com) with articles/links that do have this
issue. The issue is repeatable for those pages and they are never
succeeding. I end-to-end tested this change with logging to confirm
we do repeatably see entry count going to 2 and that with this change
the pages can background load successfully.

BUG= 675767 

Review-Url: https://codereview.chromium.org/2656653002
Cr-Commit-Position: refs/heads/master@{#446379}

[modify] https://crrev.com/3ae6f734f3566484634a068cbe663bbccc24937c/chrome/browser/prerender/prerender_browsertest.cc
[modify] https://crrev.com/3ae6f734f3566484634a068cbe663bbccc24937c/chrome/browser/prerender/prerender_contents.cc

Fixing the NEW_NAVIGATION_ENTRY issue lets many foxsports.com article pages progress to another issue - UNSUPPORTED_SCHEME prerendering failure for a "blob:http://" url.

Some examples:

http://www.foxsports.com/mlb/story/trade-deadline-insane-andrew-mccutchen-brian-dozier-royals-tigers-pirates-twins-012617
fails on: blob:http://www.foxsports.com/80127a49-a408-4f8e-947f-97cc15d82558

http://www.foxsports.com/nba/story/cleveland-cavaliers-lebron-james-playmaker-kyrie-irving-trade-012517
fails on: blob:http://www.foxsports.com/5d41796e-bc70-4d88-93f3-51613defeeac

 http://www.foxsports.com/college-football/story/matthew-mcconaughey-says-hair-in-new-movie-is-inspired-by-wvu-head-coach-dana-holgorsen-012517
fails on: blob:http://www.foxsports.com/eab40866-94c2-421a-b0db-a60e05332758

Created separate bug for the "blob:" subresource issue ( Issue 686192 ).

I'm going to mark this bug as fixed to enable merging the landed NEW_NAVIGATION_ENTRY fix. This only solves the original complaint for a subset of pages and resolving the remaining subset (with "blob:" subresource) will be continued under  Issue 686192 .

[Auto-generated comment by a script] We noticed that this issue is targeted for M-57; it appears the fix may have landed after branch point, meaning a merge might be required. Please confirm if a merge is required here - if so add Merge-Request-57 label, otherwise remove Merge-TBD label. Thanks.

Your change meets the bar and is auto-approved for M57. Please go ahead and merge the CL to branch 2987 manually. Please contact milestone owner if you have questions.
Owners: amineer@(clank), cmasso@(bling), ketakid@(cros), govind@(desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/31f17d1b0f90bf0ed092d3b620b4d2434e20d14b

commit 31f17d1b0f90bf0ed092d3b620b4d2434e20d14b
Author: Pete Williamson <petewil@chromium.org>
Date: Fri Jan 27 22:57:26 2017

Suppress NEW_NAVIGATION_ENTRY prerenderer failures for ORIGIN_OFFLINE

We have noticed some FINAL_STATUS_NEW_NAVIGATION_ENTRY failures in UMA
for offline pages but didn't know how to trigger them.  Bug 675767  has
identified a site (foxsports.com) with articles/links that do have this
issue. The issue is repeatable for those pages and they are never
succeeding. I end-to-end tested this change with logging to confirm
we do repeatably see entry count going to 2 and that with this change
the pages can background load successfully.

BUG= 675767 

Review-Url: https://codereview.chromium.org/2656653002
Cr-Commit-Position: refs/heads/master@{#446379}
(cherry picked from commit 3ae6f734f3566484634a068cbe663bbccc24937c)

Review-Url: https://codereview.chromium.org/2652393005 .
Cr-Commit-Position: refs/branch-heads/2987@{#159}
Cr-Branched-From: ad51088c0e8776e8dcd963dbe752c4035ba6dab6-refs/heads/master@{#444943}

[modify] https://crrev.com/31f17d1b0f90bf0ed092d3b620b4d2434e20d14b/chrome/browser/prerender/prerender_browsertest.cc
[modify] https://crrev.com/31f17d1b0f90bf0ed092d3b620b4d2434e20d14b/chrome/browser/prerender/prerender_contents.cc

Verified with Chrome "57.0.2987.19" on nexus 6P.