Dev/Chromium

From Kicksecure
< Dev


DISCLAIMER: This is only a collection of mostly user contributed notes. It will be reviewed, commented at a later time.

This page is a collection of notes, issues, criticism, advantages of Chromium. Development Considerations regarding default installation of Chromium in Kicksecure.

Firefox and Chromium Security[edit]

https://madaidans-insecurities.github.io/firefox-chromium.html

Daniel Micay, security researcher, KSPP member, developer of GrapheneOS, linux-hardened, hardened_malloc and more:

https://grapheneos.org/usage#web-browsing

https://twitter.com/DanielMicay/status/1154510616222142467

https://twitter.com/DanielMicay/status/1263895065224712192

https://twitter.com/DanielMicay/status/1264204763748589568

https://twitter.com/DanielMicay/status/1111034447824969728

https://old.reddit.com/r/firefox/comments/gokcis/firefox_is_insecure_refuted/frh286y/

https://old.reddit.com/r/GrapheneOS/comments/ciizae/vanadium_and_bromium_privacy/ev6m2ot/?context=3

https://old.reddit.com/r/GrapheneOS/comments/bx6h6s/apps_and_phone_set_up_grapheneos/eqcqayp/

https://old.reddit.com/r/GrapheneOS/comments/c4y8j2/web_browsing_section_added_to_the_official/erzhszu/

https://old.reddit.com/r/GrapheneOS/comments/c4y8j2/web_browsing_section_added_to_the_official/erzux4s/

https://old.reddit.com/r/GrapheneOS/comments/bg03np/browsers/elhivgg/

Thaddeus Grugq (thegrugq), information security researcher:

https://medium.com/@thegrugq/tor-and-its-discontents-ef5164845908

Dan Guido, CEO of Trail of Bits:

https://news.ycombinator.com/item?id=13623735

Matthew Green, cryptography expert:

https://twitter.com/matthew_d_green/status/830488564672626690

Thomas Ptacek, founder of Matasano Security, security researcher at Latacora:

https://twitter.com/tqbf/status/930807512609296384

https://twitter.com/tqbf/status/930860544927649792

https://twitter.com/tqbf/status/830511154950766595

https://twitter.com/tqbf/status/1107763320428417030

Theo de Raadt, founder and leader of OpenBSD:

https://marc.info/?l=openbsd-misc&m=152872551609819&w=2

PaXTeam/Pipacs of GRsecurity:

https://archive.fo/9aBLk

Chromium Debian Package Security[edit]

Contributed b @madaidan.

The Chromium package on Debian has massively crippled security and disables numerous, important security mechanisms, a few examples of which are documented below:

  • The Debian Chromium package is not a production build so basic security features like sandboxing, ASLR and CFI are crippled or nonexistent. [1]
  • Clang's type-based, forward-edge Control-Flow Integrity is not enabled. [2]
  • Automatic variable initialization is disabled with a source code patch. [3]
  • Debian disables Chromium's own hardened memory allocator (PartitionAlloc) and defaults to the likely non-hardened system malloc implementation (usually glibc). [4]
  • Debian applies many of its own patches to Chromium. Many of these are unnecessary and can potentially introduce new vulnerabilities. [5]
  • Furthermore, Debian's current Chromium package is extremely outdated, making it miss countless security fixes and new security features. It has even been susceptible to publicly known exploits being used in the wild. [6] In addition, Debian's Clang package is also severely outdated, making it impossible to enable any modern compiler mitigations even if the Debian maintainers cared enough to.

Thus, the Debian Chromium has substantially worsened security than an official version. However, despite this, it may still be more secure than Firefox (Firefox never had many of the disabled mitigations in the first place).

Chromium Doesnt give your Freedom of Modifications[edit]

Chromium doesnt has the easiness of about:config in Firefox for e.g if you want to disable certain TLS ciphers or WebRTC its not possible because its not there, This will force the users to only stick to whatever comes by default (unless recompile).

Note: Although Chromium has chrome://flags but its way less powerful/modifiable than Firefox about:config.

Chromium and Google API keys[edit]

Chromium uses API resides in google, only keys shipped within Debian/Chromium [7].

Someone suggested this [8]:

Why not modify chromium to read the api keys from a file, rather than

building them into the binary? The file could then be put in a separate package. If necessary in non-free.

This would have the additional benefit that those of us who want chromium to under no circumstances send every word we type and every website we visit to Google would no longer need to dig around in multiple prefereces dialogs to diable the multiple antifeatures enabled

by the keys.


Chromium Unknown Licenses with automated tool[edit]

Many of them comes with free software, but there is no indication all of them are [9]. (+10 years ticket)

Distribution of Adobe "Pepper" Flash Player proprietary plugin[edit]

Chromium comes with proprietary abilities within itself one of them is Adobe Flash Player [10].

This was resolved a long time ago. pepperflashplugin-nonfree is now its own separate package. https://packages.debian.org/buster/pepperflashplugin-nonfree

Chromium reduced capabilities to plugin with adblocker[edit]

Ad blocking poses an existential threat to publishers and big sellers of digital ads like Google — which is reported to have lost as much as $US6.6 billion in revenue to ad blockers last year.

Now one former Googler is fighting back against the blockers.

[11]

The move has angered Chrome users beyond belief, with many vowing to switch browsers, and many setting their eyes on Firefox, whose developers have been working to transform and rebrand the former fan-favorite into a privacy-first product.

But Google's planned Manifest V3 changes are being added to the Chromium base, meaning they'll also likely impact other Chromium-based browsers as well.

[12]

Reply by madaidan: [13]

The goal of manifest v3 is not to neuter ad blockers. It allows content filtering in a more secure way while also not allowing extensions to spy on users extensively. It doesn’t kill ad blocking - it only provides a safer way of doing it. Manifest v3 removes the legacy webRequest and replaces it with declarativeNetRequest.

https://developer.chrome.com/extensions/declarativeNetRequest

https://blog.chromium.org/2018/10/trustworthy-chrome-extensions-by-default.html

Besides, this hasn’t even been implemented yet and likely won’t be for a long time. Other browsers will probably follow suit too. For example, this is how it already works in Safari.

Reply by Patrick:

Technically too difficult to compare the old API with the new API. It cannot be deduced from media reports. Developers of existing adblocking extensions might also be less than happy if they have to rewrite their extensions, i.e. trash a lot of previous work.

In conclusion, it remains to be seen if the new API will result in a significantly worse adblocking experience for Chromium versus lets say Firefox users. Should that be the case, it would a a disadvantage. Until that happens, it doesn't speak against Chromium.

Another reply by Patrick:

"The goal of manifest v3 is not to neuter ad blockers." - Goals, intentions are not the most important point here. These will be speculative and hard to proof beyond reasonable denial. What however does matter here is the outcome.

Quote Raymond Hill, the developer behind uBlock Origin and uMatrix, explained in the Chromium bug tracker that one of the changes in Manifest v3 would break complex content filtering:

https://bugs.chromium.org/p/chromium/issues/detail?id=896897#c23

"Key portions of uBlock Origin and all of uMatrix use a different matching algorithm than that of the declarativeNetRequest API. Block/allow rules are enforced according to their *specificity*, whereas block/allow rules can override each others with no limit. This cannot be translated into a declarativeNetRequest API (assuming a 30,000 entries limit would not be a crippling limitation in itself)."

Many extension developers complaining about new API, saying it would break their extension:

https://groups.google.com/a/chromium.org/forum/#!topic/chromium-extensions/veJy9uAwS00%5B1-25%5D

Chromium: secretly stores referrer and URL for downloaded files[edit]

Can have impact on privacy and security [14], Bug was fixed later [15].

Reply by Patrick:

Bug happen. If now fixed, not sure it's even worth mentioning it.

Chromium: unconditionally downloads binary blob[edit]

Nasty bug,Got fixed [16].


Bug happen. If now fixed, not sure it's even worth mentioning it.

Questionable Chromium Privacy[edit]

Chromium privacy infrastructure is basically Google [17]:

"Additional Information on Chromium, Google Chrome, and Privacy Features that communicate with Google made available through the compilation of code in Chromium are subject to the Google Privacy Policy."

Reply by madaidan:

Chromium has some telemetry by default but it can all be disabled in the settings. By default, it's actually *less* invasive than others like Firefox.

https://www.scss.tcd.ie/Doug.Leith/pubs/browser_privacy.pdf

Google Chrome and (weird) DNS requests[edit]

This is part of chromium/chrome design [18] [19]:

When Chrome is started it will lookup domain names for previously opened web pages early in the startup process so if the user clicks on one of those links Chrome can connect to the target site immediately.

madaidan reply:

This isn’t weird. It improves page load time and is a standard thing that’s done in other browsers like Firefox. It can also be disabled.

https://support.mozilla.org/en-US/kb/how-stop-firefox-making-automatic-connections

Patrick reply:

This is called DNS Prefetching.

chromium does it: https://sites.google.com/a/chromium.org/dev/developers/design-documents/dns-prefetching

firefox also does it: https://support.mozilla.org/en-US/kb/how-stop-firefox-making-automatic-connections

What Chromium features are removed for privacy/security reasons? (Done by Brave Browser)[edit]

- Disable google services in privacy settings #244

- Disable Prediction Service - DNS prefetching #340

- Disable Chrome Google URL Tracker #248

- On startup disable connections to google domains #514

- Disable chrome.webstore.install for inline extensions #614

- Disable anchor ping attribute #764

- Disabling getBattery | bluetooth | credential.get/store | navigator.usb apis #114

- Disable network time tracker #792

- Block ChromeMetadataSource network access #769

- Remove dl.google.com repository from Linux packages #1078

- Disabling metrics reporting; Updating ad-block deps to fix audit_deps error; Updating crypto deps to fix audit_deps error; #2029

And Others you can check them here here.

Patrick reply:

Please sort this list according to different lists:

  • privacy impact only
  • security impact only
  • security and privacy impact (optional category)

Remotely Exploitable Chromium Security Vulnerability CVE-2021-21193 exploited in the wild[edit]

Summary[edit]

Viewing the situation as of 14 March 2021:

Viewing the situation as of 03 April 2021:

Sources[edit]

Quote Official Chrome Blog:

  • Google is aware of reports that an exploit for CVE-2021-21193 exists in the wild.

  • Fixed in version

    89.0.4389.90

Quote vulmon (Bold added.):

Google Chrome could allow a remote malicious user to execute arbitrary code on the system, caused by a use-after-free in Blink. By persuading a victim to visit a specially crafted Web site, a remote attacker could exploit this vulnerability to execute arbitrary code or cause a denial of service condition on the system.

version of Chromium on Flathub:

89.0.4389.82

version of Chromium on Snap Store

89.0.4389.82

Debian:

Chrome (non-freedom!) version.

dpkg -l | grep chrome

89.0.4389.90-1

Conclusion[edit]

TODO: write conclusion

notes:

Google Chrome Repository Insecurity[edit]

See Google Chrome Repository Insecurity.

Hardened Malloc[edit]

https://forums.whonix.org/t/hardened-malloc-hardened-memory-allocator/7474/147

  • Can Chromium from Flatpak be used with Hardened Malloc?
  • Which memory allocator is more secure, Chromium’s built-in or Hardened Malloc?

Chromium Forks[edit]

Development Discussion[edit]

Chromium Browser for Kicksecure ™ Discussions (not Whonix)

Related[edit]

Footnotes[edit]

  1. is_official_build=true is not set and its default value is false; it must be explicitly enabled which Chromium does not do:
  2. is_cfi=true is also not set: https://salsa.debian.org/search?utf8=%E2%9C%93&search=is_cfi&group_id=3323&project_id=20972 Compare this to another distribution which does such as Arch Linux: https://github.com/archlinux/svntogit-packages/blob/2cbe439471932d30ff2c8ded6b3dfd51b312bbc9/trunk/PKGBUILD#L145
  3. https://salsa.debian.org/chromium-team/chromium/-/blob/e55c310bb078d3c2b10fd27935b6a5a1a207f480/debian/patches/buster/clang7.patch
  4. https://salsa.debian.org/chromium-team/chromium/-/blob/7810576a1215c28d5daff0e0fbd0e3687fc43d72/debian/rules#L64
  5. https://salsa.debian.org/chromium-team/chromium/-/tree/7810576a1215c28d5daff0e0fbd0e3687fc43d72/debian/patches One of these patches have likely introduced a memory corruption bug that is currently causing hardened_malloc to kill Chromium:
  6. https://forums.whonix.org/t/chromium-browser-for-kicksecure-discussions-not-whonix/10388/49
  7. https://lists.debian.org/debian-legal/2013/10/msg00021.html
  8. https://lists.debian.org/debian-legal/2013/10/msg00023.html
  9. https://bugs.chromium.org/p/chromium/issues/detail?id=28291
  10. https://lists.debian.org/debian-legal/2013/02/msg00010.html
  11. https://www.businessinsider.com.au/former-google-exec-launches-sourcepoint-with-10-million-series-a-funding-2015-6
  12. https://www.zdnet.com/article/opera-brave-vivaldi-to-ignore-chromes-anti-ad-blocker-changes-despite-shared-codebase/
  13. https://forums.whonix.org/t/chromium-browser-for-kicksecure-discussions-not-whonix/10388
  14. https://green-possum-today.blogspot.com/2018/09/chromechromium-is-storing-url-and.html
  15. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=883746
  16. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786909
  17. https://www.chromium.org/Home/chromium-privacy
  18. https://isc.sans.edu/diary/Google+Chrome+and+%28weird%29+DNS+requests/10312
  19. https://sites.google.com/a/chromium.org/dev/developers/design-documents/dns-prefetching


Unfinished: This wiki is a work in progress. Please do not report broken links until this notice is removed, use Search Engines First and contribute improving this wiki.