| Age | Commit message (Collapse) | Author |
|---|
|
The integer overflow was detected by DonKult who added a check like this:
(std::numeric_limits<decltype(Itm.Size)>::max() - (2 * sizeof(Block)))
Which deals with the code as is, but also still is a fairly big limit,
and could become fragile if we change the code. Let's limit our file
sizes to 128 GiB, which should be sufficient for everyone.
Original comment by DonKult:
The code assumes that it can add sizeof(Block)-1 to the size of the item
later on, but if we are close to a 64bit overflow this is not possible.
Fixing this seems too complex compared to just ensuring there is enough
room left given that we will have a lot more problems the moment we will
be acting on files that large as if the item is that large, the (valid)
tar including it probably doesn't fit in 64bit either.
|
|
Like the code in arfile.cc, MemControlExtract also has buffer
overflows, in code allocating memory for parsing control files.
Specify an upper limit of 64 MiB for control files to both protect
against the Size overflowing (we allocate Size + 2 bytes), and
protect a bit against control files consisting only of zeroes.
|
|
Tarballs have long names and long link targets structured by a
special tar header with a GNU extension followed by the actual
content (padded to 512 bytes). Essentially, think of a name as
a special kind of file.
The limit of a file size in a header is 12 bytes, aka 10**12
or 1 TB. While this works OK-ish for file content that we stream
to extractors, we need to copy file names into memory, and this
opens us up to an OOM DoS attack.
Limit the file name size to 1 MiB, as libarchive does, to make
things safer.
|
|
GHSL-2020-169: This first hunk adds a check that we have more files
left to read in the file than the size of the member, ensuring that
(a) the number is not negative, which caused the crash here and (b)
ensures that we similarly avoid other issues with trying to read too
much data.
GHSL-2020-168: Long file names are encoded by a special marker in
the filename and then the real filename is part of what is normally
the data. We did not check that the length of the file name is within
the length of the member, which means that we got a overflow later
when subtracting the length from the member size to get the remaining
member size.
The file createdeb-lp1899193.cc was provided by GitHub Security Lab
and reformatted using apt coding style for inclusion in the test
case, both of these issues have an automated test case in
test/integration/test-ubuntu-bug-1899193-security-issues.
LP: #1899193
|
|
|
|
(CVE-2020-3810)
When normalizing ar member names by removing trailing whitespace
and slashes, an out-out-bound read can be caused if the ar member
name consists only of such characters, because the code did not
stop at 0, but would wrap around and continue reading from the
stack, without any limit.
Add a check to abort if we reached the first character in the
name, effectively rejecting the use of names consisting just
of slashes and spaces.
Furthermore, certain error cases in arfile.cc and extracttar.cc have
included member names in the output that were not checked at all and
might hence not be nul terminated, leading to further out of bound reads.
Fixes Debian/apt#111
LP: #1878177
|
|
Test from the fix for the regression in trusty for LP #1821308.
|
|
Verifying the content of Release.gpg made us fail on binary signatures
which were never officially supported (apt-secure manpage only documents
only the generation of ASCII armored), but silently accepted by gpgv as
we passed it on unchecked before.
The binary format is complex and is itself split into old and new
formats so adding support for this would not only add lots of code but
also a good opportunity for bugs and dubious benefit.
Reporting this issue explicitly should help repository creators figure
out the problem faster than the default NODATA message hinting at
captive portals.
Given that the binary format has no file magic or any other clear and
simple indication that this is a detached signature we guess based on
the first two bits only – and by that only supporting the "old" binary
format which seems to be the only one generated by gnupg in this case.
References: e2965b0b6bdd68ffcad0e06d11755412a7e16e50
Closes: #921685
|
|
|
|
A pin of -32768 overrides any other, disables repo
See merge request apt-team/apt!40
|
|
This new field allows a repository to declare that access to
packages requires authorization. The current implementation will
set the pin to -32768 if no authorization has been provided in
the auth.conf(.d) files.
This implementation is suboptimal in two aspects:
(1) A repository should behave more like NotSource repositories
(2) We only have the host name for the repository, we cannot use
paths yet.
- We can fix those after an ABI break.
The code also adds a check to acquire-item.cc to not use the
specified repository as a download source, mimicking NotSource.
|
|
This allows disabling a repository by pinning it to 'never',
which is internally translated to a value of -32768 (or whatever
the minimum of short is).
This overrides any other pin for that repository. It can be used
to make sure certain sources are never used; for example, in
unattended-upgrades.
To prevent semantic changes to existing files, we substitute
min + 1 for every pin-priority: <min>. This is a temporary
solution, as we are waiting for an ABI break.
To add pins with that value, the special Pin-Priority
"never" may be used for now. It's unclear if that will
persist, or if the interface will change eventually.
|
|
Fail if InRelease or Release.gpg contain unsigned lines
See merge request apt-team/apt!45
|
|
Implementing a parser with recursion isn't the best idea, but in
practice we should get away with it for the time being to avoid
needless codechurn.
Closes: #920317 #921037
|
|
We support dash-encoding even if we don't really work with files who
would need it as implementations are free to encode every line, but
otherwise a line starting with a dash must either be a header we parse
explicitly or the file is refused. This is against the RFC which says
clients should warn on such files, but given that we aren't expecting
any files with dash-started lines to begin with this looks a lot like a
we should not continue to touch the file as it smells like an attempt to
confuse different parsers by "hiding" headers in-between others.
The other slightly more reasonable explanation would be an armor header
key starting with a dash, but no existing key does that and it seems
unlikely that this could ever happen. Also, it is recommended that
clients warn about unknown keys, so new appearance is limited.
|
|
Having many rather similar implementations especially if one is exported
while others aren't (and the rest of it not factored out at all) seems
suboptimal.
|
|
The exploit for CVE-2019-3462 uses the fact that a Release.gpg file can
contain additional content beside the expected detached signature(s).
We were passing the file unchecked to gpgv which ignores these extras
without complains, so we reuse the same line-reading implementation we
use for InRelease splitting to detect if a Release.gpg file contains
unexpected data and fail in this case given that we in the previous
commit we established that we fail in the similar InRelease case now.
|
|
The warnings were introduced 2 years ago without any reports from the
wild about them actually appearing for anyone, so now seems to be an as
good time as any to switch them to errors.
This allows rewritting the code by failing earlier instead of trying to
keep going which makes the diff a bit hard to follow but should help
simplifying reasoning about it.
References: 6376dfb8dfb99b9d182c2fb13aa34b2ac89805e3
|
|
Report keys used to sign file from gpgv method to acquire system
See merge request apt-team/apt!44
|
|
This fixes a security issue that can be exploited to inject arbritrary debs
or other files into a signed repository as followed:
(1) Server sends a redirect to somewhere%0a<headers for the apt method> (where %0a is
\n encoded)
(2) apt method decodes the redirect (because the method encodes the URLs before
sending them out), writting something like
somewhere\n
<headers>
into its output
(3) apt then uses the headers injected for validation purposes.
Regression-Of: c34ea12ad509cb34c954ed574a301c3cbede55ec
LP: #1812353
|
|
Telling the acquire system which keys caused the gpgv method to
succeed allows us for now just a casual check if the gpgv method
really executed catching bugs like CVE-2018-0501, but we will make use
of the information for better features in the following commits.
|
|
Having a method take a bunch of string vectors is bad style, so we
change this to a wrapping struct and adapt the rest of the code brushing
it up slightly in the process, which results even in a slightly "better"
debug output, no practical change otherwise.
Gbp-Dch: Ignore
|
|
Set PATH=/usr/sbin:/usr/bin:/sbin:/bin when running dpkg
See merge request apt-team/apt!38
|
|
This avoids a lot of problems from local installations of
scripting languages and other stuff in /usr/local for which
maintainer scripts are not prepared.
[v3: Inherit PATH during tests, check overrides work]
[v2: Add testing]
|
|
This allows us to install matching auth files for sources.list.d
files, for example; very useful.
This converts aptmethod's authfd from one FileFd to a vector of
pointers to FileFd, as FileFd cannot be copied, and move operators
are hard.
|
|
Use quoted tagnames in config dumps
See merge request apt-team/apt!32
|
|
This adds a new "autopurge" command that will is a shortcut for
"autoremove --purge"
Thanks: Michael Vogt for the initial work
|
|
Try 10 times in a row
|
|
Tagnames in configuration can include spaces (and other nasties) e.g. in
repository-specific configuration options due to Origin/Label
potentially containing a space. The configuration file format supports
parsing quoted as well as encoded spaces, but the output generated by
apt-config and other places which might be feedback into apt via
parsing (e.g. before calling apt-key in our gpgv method) do not quote
and hence produce invalid configuration files.
Changing the default to be an encoded tagname ensures that the output of
dump can be used as a config file, but other users might not expect
this so that is technically a backward-breaking change.
|
|
No user visible change expect for some years old changelog entries,
so we don't really need to add a new one for this…
Reported-By: codespell
Gbp-Dch: Ignore
|
|
The filename can be overridden, but sometimes it is useful to do it only
for the directory-part of the filename – e.g. if you want to let a flat
archive directory (like /var/cache/apt/archives) serve a pool-based
request like /pool/a/apt_version.deb.
Gbp-Dch: Ignore
|
|
We sometimes autogenerate HTML pages e.g. for listing files in a
directory or for various error codes. If this would be a serious
webserver this would be a security problem (althrough a bit hard to
exploit), but as it is not shipped and intended to be used by our
testcases only the world hasn't ended & we can ignore it for
changelog and fix it for brownie points.
Gbp-Dch: Ignore
|
|
Browsing pages served via aptwebserver is working better if we tell the
browser the Content-Type which for this simple usecase we can just do by
guessing based on the file extension – and because hardcoding a list
would be boring we just reuse the mime.types data from mime-support if
available and allow it to be overridden by files and config.
Gbp-Dch: Ignore
|
|
Fixes Debian/apt#77
|
|
Support subkeys and multiple keyrings in Signed-By options
See merge request apt-team/apt!27
|
|
Closes: #910941
|
|
Some post-invoke scripts install packages, which fails because
the environment variable is not set. This sets the variable for
all three kinds of scripts {pre,post-}invoke and pre-install-pkgs,
but we will only allow post-invoke at a later time.
Gbp-Dch: full
|
|
The implementation of "apt-cache show" (not "apt show") incorrectly
resets the currently used parser if the record itself and the
description to show come from the same file (as it is the case if no
Translation-* files are available e.g. after debootstrap).
The code is more complex than you would hope to support some rather
unusual setups involving Descriptions and their translations as tested
for by ./test-bug-712435-missing-descriptions as otherwise this could
be a one-line change.
Regression-Of: bf53f39c9a0221b670ffff74053ed36fc502d5a0
Closes: #909155
|
|
Process all of --status-fd and don't expect duplicate status msg
See merge request apt-team/apt!26
|
|
The uniqueness in std::set containers is ensured by the ordering
operator we provide, but it was not considering that different versions
can have the same description like the different architectures for a
version of a package.
Closes: #908218
|
|
A user can specify multiple fingerprints for a while now, so its seems
counter-intuitive to support only one keyring, especially if this isn't
really checked or enforced and while unlikely mixtures of both should
work properly, too, instead of a kinda random behaviour.
|
|
If we limit a file to be signed by a certain key it should usually
accept also being signed by any of this keys subkeys instead of
requiring each subkey to be listed explicitly. If the later is really
wanted we support now also the same syntax as gpg does with appending an
exclamation mark at the end of the fingerprint to force no mapping.
|
|
We are seeing 'processing' messages from dpkg first, so it makes sense
to translate them to "Preparing" messages instead of using "Installing"
and co to override these shortly after with the "Preparing" messages.
The difference isn't all to visible as later messages tend to linger far
longer in the display than the ealier ones, but at least in a listing it
seems more logical.
|
|
The progress reporting relies on parsing the status reports of
dpkg which used to repeat being in the same state multiple times
in the same run, but by fixing #365921 it will stop doing so.
The problem is in theory just with 'config-files' in case we do purge as
this (can) do remove + purge in one step, but we remove this also for
the unpack + configure combination althrough we handle these currently
in two independent dpkg calls.
|
|
It is an uphill battle to "reset" the environment to a clean state
without making it needlessly hard to use 'good' environment variables,
so we just try a little harder here without really trying for
completeness.
Gbp-Dch: Ignore
|
|
Reported-By: Guillem Jover <guillem@debian.org>
Gbp-Dch: Ignore
|
|
No user-visible change as it effects mostly code comments and
not a single error message, manpage or similar.
Reported-By: codespell & spellintian
Gbp-Dch: Ignore
|
|
cppcheck reports: (error) Iterator 't' used after element has been erased.
The loop is actually fashioned to deal with this (not in the most
efficient way, but in simplest and speed isn't really a concern here)
IF this codepath had a "break" at the end… so I added one.
Note that the tests aren't failing before (and hopefully after) the
change as the undefined behavior we encounter is too stable.
Thanks: David Binderman for reporting
|
|
APT in 1.6 saw me rewriting the mirror:// transport method, which works
comparable to the decommissioned httpredir.d.o "just" that apt requests
a mirror list and performs all the redirections internally with all the
bells like parallel download and automatic fallback (more details in the
apt-transport-mirror manpage included in the 1.6 release).
The automatic fallback is the problem here: The intend is that if a file
fails to be downloaded (e.g. because the mirror is offline, broken,
out-of-sync, …) instead of erroring out the next mirror in the list is
contacted for a retry of the download.
Internally the acquire process of an InRelease file (works with the
Release/Release.gpg pair, too) happens in steps: 1) download file and 2)
verify file, both handled as URL requests passed around. Due to an
oversight the fallbacks for the first step are still active for the
second step, so that the successful download from another mirror stands
in for the failed verification… *facepalm*
Note that the attacker can not judge by the request arriving for the
InRelease file if the user is using the mirror method or not. If entire
traffic is observed Eve might be able to observe the request for
a mirror list, but that might or might not be telling if following
requests for InRelease files will be based on that list or for another
sources.list entry not using mirror (Users have also the option to have
the mirror list locally (via e.g. mirror+file://) instead of on a remote
host). If the user isn't using mirror:// for this InRelease file apt
will fail very visibly as intended.
(The mirror list needs to include at least two mirrors and to work
reliably the attacker needs to be able to MITM all mirrors in the list.
For remotely accessed mirror lists this is no limitation as the attacker
is in full control of the file in that case)
Fixed by clearing the alternatives after a step completes (and moving a pimpl
class further to the top to make that valid compilable code). mirror://
is at the moment the only method using this code infrastructure (for all
others this set is already empty) and the only method-independent user
so far is the download of deb files, but those are downloaded and
verified in a single step; so there shouldn't be much opportunity for
regression here even through a central code area is changed.
Upgrade instructions: Given all apt-based frontends are affected, even
additional restrictions like signed-by are bypassed and the attack in
progress is hardly visible in the progress reporting of an update
operation (the InRelease file is marked "Ign", but no fallback to
"Release/Release.gpg" is happening) and leaves no trace (expect files
downloaded from the attackers repository of course) the best course of
action might be to change the sources.list to not use the mirror family
of transports ({tor+,…}mirror{,+{http{,s},file,…}}) until a fixed
version of the src:apt packages are installed.
Regression-Of: 355e1aceac1dd05c4c7daf3420b09bd860fd169d,
57fa854e4cdb060e87ca265abd5a83364f9fa681
LP: #1787752
|
|
gpgs DETAILS documentation file declares that GOODSIG could report keyid
or fingerprint since gpg2, but for the time being it is still keyid
only. Who knows if that will ever change as that feels like an interface
break with dangerous security implications, but lets be better safe than
sorry especially as the code dealing with signed-by keyids is prepared
for this already. This code is rewritten still to have them all use the
same code for this type of problem.
|
