Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Thu, 21 Sep 2023 22:52:50 +0200
From: Solar Designer <>
Subject: CVE-2023-4863: libwebp: Heap buffer overflow in WebP Codec


Ideally this should have been brought in here ~10 days ago, but I guess
better late than never, and I'd like to encourage others to be bringing
relevant stuff to the list.

On September 11, Google announced an update to Chrome:

fixing this issue:

> Critical CVE-2023-4863: Heap buffer overflow in WebP. Reported by Apple
> Security Engineering and Architecture (SEAR) and The Citizen Lab at The
> University of Toronto's Munk School on 2023-09-06
> We would also like to thank all security researchers that worked with us
> during the development cycle to prevent security bugs from ever reaching
> the stable channel.
> Google is aware that an exploit for CVE-2023-4863 exists in the wild.

With the bug being in a library used by many projects, this made people
wonder why a CVE was assigned to Chrome rather than to libwebp:

Adam Caudill, Security Architect at 1Password:

> The latest vulnerability causing headaches across the world is
> CVE-2023-4863, issued by Google Chrome and described as "Heap buffer
> overflow in WebP in Google Chrome prior to 116.0.5845.187 allowed a
> remote attacker to perform an out of bounds memory write via a crafted
> HTML page". This same CVE is cited by a number of other vendors as they
> are impacted as well. But, is this really a Google Chrome vulnerability?
> The fix for this issue is not actually part of Chrome or even Chromium,
> but rather the libwebp library, which is used by Chromium, and a number
> of other projects. The library is maintained by the WebM Project, which
> is a joint effort between Google and a number of other companies.
> Firefox, for example, uses also uses libwebp, and is impacted by this
> vulnerability - though they don't mention Chrome in their advisory, but
> they do mention Chrome's CVE.

and more recently:

Will Dormann:

> Out-loud wonders:
> Citizen Lab reported an ITW 0day to Apple, who assigned CVE-2023-41064
> to ImageIo. Apple and Citizen Lab reported a libwebp issue to Google,
> who filed CVE-2023-4863 to describe it, saying it's for "Chrome" only.
> Are these both the same vulnerability?
> If they are, then:
> Apple got a libwebp vulnerability, and decided to assign a CVE to their
> product rather than libwebp.
> Google got a libwebp vulnerability, and also decided to assign a
> different CVE to their product rather than libwebp.
> Surely this isn't how CVE is to be used??

The official CVE record still says:

> Heap buffer overflow in WebP in Google Chrome prior to 116.0.5845.187
> allowed a remote attacker to perform an out of bounds memory write via a
> crafted HTML page. (Chromium security severity: Critical)

However, "GitHub Advisory Database / GitHub Reviewed / CVE-2023-4863"
correctly attributes this to libwebp, and also gives affected version
ranges for some other projects.  It otherwise looks CVE-entry based and
weirdly refers to "a crafted HTML page", even though that's specific to
just some uses of the library:

> Heap buffer overflow in libwebp allow a remote attacker to perform an
> out of bounds memory write via a crafted HTML page.

What a mess.  Finally, distros are releasing update packages that use
CVE-2023-4863 to refer to the libwebp vulnerability, e.g. Red Hat:

> The libwebp packages provide a library and tools for the WebP graphics
> format. WebP is an image format with a lossy compression of digital
> photographic images. WebP consists of a codec based on the VP8 format, and
> a container based on the Resource Interchange File Format (RIFF).
> Webmasters, web developers and browser developers can use WebP to compress,
> archive, and distribute digital images more efficiently.
> Security Fix(es):
> * libwebp: Heap buffer overflow in WebP Codec (CVE-2023-4863)

This is probably the best way to go.

Somehow none of the above refers to vulnerable/fixed upstream libwebp
version numbers.  As far as I can see, the main bug is fixed in 1.3.2:

NEWS file entry:

> - 9/13/2023: version 1.3.2
>   This is a binary compatible release.
>   * security fix for lossless decoder (chromium: #1479274, CVE-2023-4863)

Going through recent commits, the main one appears to be:

commit 2af26267cdfcb63a88e5c74a85927a12d6ca1d76
Author: Vincent Rabaud <>
Date:   Thu Sep 7 21:16:03 2023 +0200

    Fix OOB write in BuildHuffmanTable.
    First, BuildHuffmanTable is called to check if the data is valid.
    If it is and the table is not big enough, more memory is allocated.
    This will make sure that valid (but unoptimized because of unbalanced
    codes) streams are still decodable.
    Bug: chromium:1479274
    Change-Id: I31c36dbf3aa78d35ecf38706b50464fd3d375741
    (cherry picked from commit 902bc9190331343b2017211debcec8d2ab87e17a)

However, another maybe-important one also made it into 1.3.2:

commit 95ea5226c870449522240ccff26f0b006037c520
Author: Vincent Rabaud <>
Date:   Mon Sep 11 16:06:08 2023 +0200

    Fix invalid incremental decoding check.
    The first condition is only necessary if we have not read enough
    (enough being defined by src_last, not src_end which is the end
    of the image).
    The second condition now fits the comment below: "if not
    incremental, and we are past the end of buffer".
    Change-Id: I0700f67c62db8e1c02c2e429a069a71e606a5e4f

and then these are post-1.3.2 (not in that release), but possibly also

commit dce8397fec159c9edfeec7c6388cb81428c87ed8
Author: Masahiro Hanada <>
Date:   Thu Sep 14 19:37:24 2023 +0900

    Fix next is invalid pointer when WebPSafeMalloc fails
    When WebPSafeMalloc fails on VP8LHuffmanTablesAllocate,
    next is not initialized to NULL.
    VP8LHuffmanTablesDeallocate uses next to know the following nodes.
    A patch fixes this issue.
    Change-Id: I144ae84cd97e5bca227018ef1afa95361267902c

commit 433c7dca11bb5b001ce5ad36ac1afd2906a2f13e
Author: Vincent Rabaud <>
Date:   Thu Sep 14 09:31:19 2023 +0200

    Fix static analyzer warnings.
    Change-Id: I45f0db2310b1188809963af93240e3d438f807b8

It looks like libwebp is having more eyes and static analyzers and
fuzzers on it now and more (potential) issues are being fixed (great),
but those additional fixes might not yet be making it into distros and
into larger third-party projects using libwebp such as apparently
Firefox, Signal, and Electron (to focus on FOSS here):

> WebP seemed to be all connected to Chrome, Mozilla, Edge, Brave, Signal,
> 1Password and Apple Zero-Days. 
> "any software that uses the libwebp library" is affected by this
> vulnerability, including Electron-based applications such as 1Password
> and Signal.

I wish someone more involved made this posting, not me.  It is possible
that I got something wrong.  I'd appreciate any relevant follow-ups.


Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.