The code for testing regexes with embedded Apache variables
(rule->re_precomp == 1) during request processing was utilizing the global
engine pool for the storage of temporary values. This approach is not
threadsafe, retains the temporary variables longer than they are usable,
and causes corruption of the global pool's "cleanups" linked-lists when
Apache is configured with a threaded MPM.
Instead of reading the fuzzy db on every invocation, read and store
the db contents during initialization and store the contents in memory.
The only significant behavior change here is that a change in db contents
now (obviously) requires a daemon restart, as no API is provided to
flush the list of ssdeep chunks.
Resources load mechanism as the SecRemoteRuels were not showing information
about the loaded rules while Apache was reloaded. This patch add such
information to reload in the same way that it was showing on restart.
As reported by Walter Hop on our dev- mailing list, remote resource download
was not validating the HTTP code, parsing errors pages as resources. This
commit fix this issue, from now one HTTP error codes will be verified and
treated as errors. Operators are now dealing well with empty values that may
be produced in consequence of a download error.
As reported by Rainer Jung, Curl may not be mandatory to build
ModSecurity core. This patch make it optional by:
- Concentrate all downloads using curl on msc_remote_rules.c
- Split Curl build definitions checks into: WITH_CURL, WITH_REMOTE_RULES
and WITH_CRYPTO.
- WITH_CURL: Contains Culr headers and binaries during the build time.
- WITH_REMOTE_RULES: Currently enabled if Curl is present.
- WITH_CRYPTO: Set if apr tool was compiled with crypto support.
- Renames msc_remote_grab_content to msc_remote_download_content
On IIS CA validation was not working as libcurl on windows does not look for a
certificate store, unless it is specified. The resource downloads are now
respecting the SecRemoteRulesFailAction.
This commit makes ModSecurity to refuse to download or install rules
(SecRemoteRules) from sites that are not running HTTPS with a valid and
trusted certificate.
Initially those directives were only able to load content from a
local file. This commit extends this functionality allowing the user to
provide an HTTP URI that can be downloaded and loaded by ModSecurity.
Initially the download is associated with a server restart. For next
versions we expect to load such resources as it become outdated (Without
need to resetart the server).
FuzzyHash operator is optional and only installed if the headers for libfuzzy
was found in the system. Otherwise, the FuzzyHash operator is disable during
the compilation. After this commit, if some rules tries to use it, ModSecurity
will produce an runtime error not a config time error, allowing the web server
to procede normal with its operations.
The fuzzyHash operator can be used to match files. In conjuntcion
with FILES_TMP_CONTENT collection it can scan uploaded files and
try to match it with a pre caculated list of know malicious content,
more details on how it works can be found on ssdeep website:
http://ssdeep.sourceforge.net/
The ipv4 representation was only accepting slash notation with masks
represented in 2 digits. In the ipv6 implementation several fixies were made:
The maximum value to a bitmask was 64 which is not the reality, as ipv6 can
handle 128 bits. The second change was also to enable mask representation with
more and less than 2 digits. A more general fix was added to allow the unit
tests to work even if a invalid ip/range was informed during the creation of
the "tree", now it is checking if the tree is NULL while performing the
execution of the operator. Initial problem was reported at the issue: #706.
Used by the operator @ipMatch and variants, this structure storage all the IPs
addresses for later comparison. Last version was using RadixTree only if the
set of IPs was specified from files. IPs specified as parameters, was using
a chained list. Chained lists may affect the performance, since lookups in
worst case will be O(n). RadixTrees could provide better results depending
on the amount of elements and its contents.
The operators @ipMatch, @ipMatchF and @ipMatchFromFile were
added to the functions: SecReadStateLimit and SecReadStateLimit,
by using them it is possible to declare a suspicious list. When
a suspicious list is given, the {Read|Write}StateLimit will be
applied just to the IPs that belongs to that restricted list.
Note that the negative of those operators (e.g. !@ipMatch) can be
used to place a whitelist. The {Read|Write}StateLimit
restrictions will not be applied to those in the whitelist.
This current version the Sec{Read|Write}StateLimit can be used
varios times to add elements to both lists, however, the
last informed limit will be applied for the entire group. This
feature is experimental, and suggestions on how to improve it
are very welcome. For further discussion use the issue: #353.
