Closed
Bug 1420217
(efail)
Opened 8 years ago
Closed 7 years ago
[efail] Full Plaintext Leak in PGP
Categories
(Thunderbird :: Security, defect)
Tracking
(Not tracked)
RESOLVED
DUPLICATE
of bug 1419417
People
(Reporter: schinzel, Unassigned)
References
Details
(Keywords: sec-high)
Attachments
(4 files)
User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/62.0.3202.94 Safari/537.36
Steps to reproduce:
Send a modified PGP message to Thunderbird/Enigmail. For a detailed description, see the PDF attached. We will notify GnuPG as well.
Actual results:
The attacker gets the full plaintext of the PGP message.
Expected results:
The attacker should not get the full plaintext of the PGP message.
Updated•8 years ago
|
Component: Untriaged → Security
Comment 1•8 years ago
|
||
Jens Müller reported bug 1419417 which is a a subset of this report.
I suppose this and that are dupes of one another, but this has the more formal report.
See Also: → CVE-2018-12372
| Reporter | ||
Comment 2•8 years ago
|
||
This here is an extended version with another bug in PGP that should affect all implementations.
Comment 3•8 years ago
|
||
I also got this bug reported as Enigmail bug. Is this something that can/should be fixed in Enigmail, or is that on the level of the HTML engine? If I get it right, it also affects S/MIME.
| Reporter | ||
Comment 4•8 years ago
|
||
Please talk to Jens about the bugs he reported, we just included it in our report as a references. Sorry for the confusion.
Regarding the crypto gadget attack: the attacks depends on the attacker getting a backchannel. Right now, we use HTML for this. Disabling all backchannels would somehow mitigate the attack. From a crypto viewpoint, this would not be a very satisfying mitigation though. The core of the vulnerability is certainly the missing authenticated encryption.
Yes, this is the same attack as against S/MIME (ID 1411592).
Comment 5•8 years ago
|
||
1. It seems to me that the report was cut off at page 9. Section 7 is about "Email Clients" and mentions several tested clients. Section 7.1 is about Thunderbird, but there is no section 7.2, and nothing else -- there is no other mail client described. Is that on purpose?
2. The report mentions 3 emails attached, but I cannot find them. Could you attach the emails to the bug?
| Reporter | ||
Comment 6•8 years ago
|
||
| Reporter | ||
Comment 7•8 years ago
|
||
| Reporter | ||
Comment 8•8 years ago
|
||
| Reporter | ||
Comment 9•8 years ago
|
||
It is on purpose that we only describe findings for Thunderbird here.
I have uploaded the files.
Comment 10•8 years ago
|
||
Incidentally, bug 1412632 addresses section 7.1.1 of the document embedded for PGP/MIME parts: a warning is displayed that the user may be tricked into leaking encrypted information to the recipient. I might improve the warning to explicitly say that this may even be true if the user doesn't see any of the decrypted text.
Updated•8 years ago
|
Alias: efail
Assignee: nobody → ben.bucksch
Status: UNCONFIRMED → ASSIGNED
Depends on: CVE-2018-12372
Ever confirmed: true
Summary: Full Plaintext Leak in PGP → [efail] Full Plaintext Leak in PGP
Comment 11•7 years ago
|
||
Marking this as duplicate of bug 1419417. I think we're all done, there and in the dependent bugs.
Assignee: ben.bucksch → nobody
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
Resolution: --- → DUPLICATE
Updated•2 years ago
|
Group: mail-core-security
You need to log in
before you can comment on or make changes to this bug.
Description
•