I need to look up this grill to see what its embedded controller does. Aside from…
reporting its current settings and temperature, and
some limited ability to control it remotely (very limited, if at all — the computers in IoT devices are cheap and insecure, and attackers can cause all sorts of mischief with a networked propane tank)…
…what else does it do that needs an update, never mind an update big enough to interfere with cooking?
When it came time to set the passcode to unlock the phone, I found out that the longest device unlock passcode that even the most recent version of Android will accept is 16 characters.That was the case five years ago, and it’s still the case today.
Android’s “Choose Lock Password” screen is part of AOSP (Android Open Source Project), which means that its source code is easy to find online. It’s ChooseLockPassword.java, and the limitation is a constant defined in a class named ChooseLockPasswordFragment, which defines the portion of the screen where you enter a new passcode.
Here are the lines from that class that define passcode requirements and limitations:
Note the values assigned to these variables. It turns out that there are only two constraints on Android passcodes that are currently in effect:
The minimum length, stored in mPasswordMinLength, which is set to the value stored in the constant LockPatternUtils.MIN_LOCK_PASSWORD_SIZE. This is currently set to 6.
The maximum length, stored in mPasswordMaxLength, which is set to 16.
As you might have inferred from the other variable names, there may eventually be other constraints on passcodes — namely, minimums for the number of letters, uppercase letters, lowercase letters, symbol characters, numeric characters, and non-letter characters — but they’re currently not in effect.
Why 16 characters?
16 is a power of 2, and to borrow a line from Snow Crash, powers of 2 are numbers that a programmer would recognize “more readily than his own mother’s date of birth”. This might lead you to believe that 16 characters would be some kind of technical limit or requirement, but…
…Android (and in fact, every current non-homemade operating system) doesn’t store things like passcodes and passwords as-is. Instead, it stores the hashes of those passcodes and passwords. The magic of hash functions is that no matter how short or long the text you feed into them, their output is always the same fixed size (and a relatively compact size, too).
For example, consider SHA-256, from the SHA-2 family of hash functions:
No matter the length of the input text, the output of the SHA-256 function is always the same length: 64 characters, each one a hexadecimal digit.
Under the 16-character limit, the password will always be shorter than the hash that actually gets stored! There’s also the fact that in a time when storage is measured in gigabytes, we could store a hash that was thousands of characters long and not even notice.
My guess is that the Android passcode size limit of 16 characters is purely arbitrary. Perhaps they thought that 16-character passwords like the ones below were the longest that anyone would want to memorize:
In my opinion, when it comes to getting the best bang and build quality for the buck on an Android phone, check out Motorola’s phones. Lenovo — the same company who took Right now, they’ve got discounts on many of their mobiles, including $100 off any of the Motorola One family — the Action, the Zoom, and the one I got: the Hyper.
With the discount, the unlocked Hyper goes for US$299 when purchased directly from Motorola. That’s a pretty good price for an Android phone with mid-level specs.
Here’s a quick video review of this chipset from Android Authority’s Gary Sims:
As a point of reference, this chipset is also used in Samsung’s Galaxy A70, A60, and M40, and LG’s Q70.
This chipset puts the Moto One Hyper firmly in the middle of the road of current Android offerings, making it a reasonably representative device for an indie Android developer/article author like Yours Truly.
The phone’s “Hyper” name is a reference to its “hyper charging” — high-speed charging thanks to its ability to take a higher level of power during the charging process. It comes with an 18 watt charger (the same level of power provided by the current iPad Pro and iPhone 11 chargers), but if you have a 45 watt charger handy, the phone’s 4,000 mAh battery will charge in just over 10 minutes.
The phone also comes with the usual literature and SIM extraction pin:
There is one additional goodie that I didn’t expect: a clear, flexible, rubber-like plastic case. It’s nothing fancy, but it was still a nice surprise.
I’ll post more details about the phone as I use it and start doing development work (native stuff in Kotlin, as well as some cross-platform work in Flutter, and maybe even Kivy).
Day 4 of the Hardware 101 component of the UC Baseline cybersecurity program was all about security for the enterprise, which naturally included topics such as servers. Not everyone in the class has had the opportunity to tour a server room or data center, and this was their chance to see these machines up close.
Unlike the previous days, we did not attempt to dismantle and then reassemble the servers — this was a “look, but don’t touch” sort of lesson.
We also had a guest lecturer who gave us a pretty thorough walkthrough of the sorts of things involved in an enterprise server/data center setup, some of which went way over my head. I don’t see a sysadmin/system architect role in my future, but it might not hurt for me to do some supplementary reading on this topic.
Day 5 was the final day of Hardware 101 and started with something that I’ve always been terrible at: Making networking cables.
We also spent some time looking over all sorts of intrusion devices, such as the incredibly cute “Pwnagotchi”, a Raspberry Pi Zero-based device that “listens” to wifi chatter to feed its machine learning program in order to figure out wifi passwords.
It uses an e-paper screen, which is quite legible and consumes little power.
It’s incredibly small:
Here’s a Pwnagotchi beside a U.S. quarter for size reference:
A great way to steal information to gain access to people’s accounts and systems is to set up a fake wifi hotspot at a place that offers free wifi, such as Starbucks. That’s what the Wifi Pineapple is for — people connect to it, thinking they’re connecting to Starbucks wifi. You route their signals through to the real Starbucks wifi, but you’re the go-between, and can “see” everything that your marks are sending on the internet: the data they’re passing back and forth, including stuff like user IDs and passwords:
It sends out a signal that causes devices currently connected to wifi to disconnect. You could use it in tandem with a Wifi Pineapple to force people to disconnect from the real wifi and then connect to the Pineapple instead, enabling you to read their internet communications.
If you really want to “sniff” all the wifi traffic in the room, you’ll want one of these — a high-gain antenna system hooked to a network interface controller (NIC) that reads signals in “promiscuous mode”, a capability that’s disabled in most NICs. In promiscuous mode, you can capture all wifi traffic instead of the bits of data that you’re authorized to receive. It’s a good network diagnostics tool — and it’s also useful for getting up to no good:
And finally, the Shark Jack. Plug it into someone’s network, either via the ethernet jack or USB, and it will execute scripts to get a map of the network or even deliver a payload somewhere onto the system:
It’s basically a real-world version of the device that Tony Stark slipped onto the command console of the SHIELD helicarrier in the first Avengers movie (it’s at the 0:44 mark):
I may have to invest in one of those bad boys. For research purposes, you understand.
We also had a guest lecturer who delivered a very thorough and informative presentation on getting started in cybersecurity. I’ll have to post notes on it later:
For the benefit of my classmates in the UC Baseline program (see this earlier post to find out what it’s about), I’m posting a regular series of notes here on Global Nerdy to supplement the class material. As our instructor Tremere said, what’s covered in the class merely scratches the surface, and that we should use it as a launching point for our own independent study.
There was a lot of introductory material to cover on day one of the Hardware 101 portion of the program, and there’s one bit of basic but important material that I think deserves a closer look, especially for my fellow classmates who’ve never had to deal with it before: How binary and hexadecimal numbers are related.
Here’s the same number, expressed in base 2, a.k.a. the binary system: 1010001111011100101101001.
That’s the problem with binary numbers: Because they use only two digits, 0 and 1, they grow in length extremely quickly, which makes them hard for humans to read. Can you tell the difference between 100000000000000000000000 and 1000000000000000000000000? Be careful, because those two numbers are significantly different — one is twice the size of the other!
(Think about it: In the decimal system, you make a number ten times as large by tacking a 0 onto the end. For the exact same reason, tacking a 0 onto the end of binary number doubles that number.)
Hexadecimal is an easier way to write binary numbers
Once again, the problem is that:
Binary numbers, because they use only two digits — 0 and 1 — get really long really quickly, and
Decimal numbers don’t convert easily to binary.
What we need is a numerical system that:
Can represent really big numbers with relatively few characters, and
Converts easily to binary.
Luckily for us, there’s a numerical system that fits this description: Hexadecimal. The root words for hexadecimal are hexa (Greek for “six”) and decimal (from Latin for “ten”), and it means base 16.
Using 4 binary digits, you can represent the numbers 0 through 15:
Hexadecimal is the answer to the question “What if we had a set of digits that represented the 16 numbers of 0 through 15?”
Let’s repeat the above table, this time with hexadecimal digits:
Hexadecimal gives us easier-to-read numbers where each digit represents a group of 4 binary digits. Because of this, it’s easy to convert back and forth between binary and hexadecimal.
Since we’re creatures of base 10, we have the single characters to represent the digits 0 through 9, but no single character to represent 10, 11, 12, 13, 14, and 15, which are digits in hexadecimal. To work around this problem, hexadecimal uses the first 6 letters from the Roman alphabet: A, B, C, D, E, and F.
How to indicate if you’re writing a number in decimal, binary, or hexadecimal form
Because we’re base 10 creatures, we simply write decimal numbers as-is:
To indicate that a number is in binary, we prefix it with the number zero followed by a lowercase b:
#Thiswill print"49833"inthe console
Or if you prefer, Python:
# This will print "49833" in the console
To indicate that a number is in hexadecimal, we prefix it with the number zero followed by a lowercase x:
#Thiswill print"49833"inthe console
# Both of these will print "49833" in the console
Common grouping of binary numbers and hexadecimal
4 bits: A half-byte, tetrade, or nybble
A single hexadecimal digit represents 4 bits, and my favorite term for a group of 4 bits is nybble. The 4 bits that make up a nybble can represent the numbers 0 through 15.
“Nybble” is one of those computer science-y jokes that’s based on the fact that a group of 8 bits is called a byte. I’ve seen the terms half-byte and tetrade also used.
8 bits: A byte
Two hexadecimal digits represent 8 bits, and a group of 8 bits is called a byte. The 8 bits that make up a byte can represent the numbers 0 through 255, or the numbers -128 through 127.
In the era of the first general-purpose microprocessors, the data bus was 8 bits wide, and so byte was the standard unit of data. Every character in the ASCII character set can be expressed in a single byte. Each of the 4 numbers in an IPv4 address is a byte.
16 bits: A word
Four hexadecimal digits represent 16 bits, and a group of 16 bits is most often called a word. The 16 bits that make up a word can represent the numbers 0 through 65,535 (a number sometimes referred to as “64K”), or the numbers -32,768 through 32,767.
If you were computing in the late ’80s or early ’90s — the era covered by Windows 1 through 3 or Macs in the classic chassis — you were using a 16-bit machine. That meant that it stored data a word at a time.
32 bits: A double word or DWORD
Eight hexadecimal digits represent 32 bits, and a group of 32 bits is often called a double word or DWORD; I’ve also heard the unimaginative term “32-bit word”. The 32 bits that make up a word can represent the numbers 0 through 4,294,967,295 (a number sometimes referred to as “4 gigs”), or the numbers −2,147,483,648 through 2,147,483,647.
32-bit operating systems and computers came about in the mid-1990s. Some are still in use today, although they’d now be considered older or “legacy” systems.
The IPv4 address system uses 32 bits, which means that it can represent a maximum of 4,294,967,29 internet addresses. That’s fewer addresses than there are people on earth, and as you might expect, we’re running out of these addresses. There are all manner of workarounds, but the real solution is for everyone to switch to IPv6, which uses 128 bits, which allows for over 3 × 1038 addresses — enough to assign 100 addresses to every atom on the surface of the earth.
64 bits: A quadruple word or QWORD
16 hexadecimal digits represent 64 bits, and a group of 64 bits is often called a quadruple word, quad word, or QWORD; I’ve also heard the unimaginative term “64-bit word”. The 64 bits that make up a word can represent the numbers 0 through 18,446,744,073,709,551,615 (about 18.4 quintillion), or the numbers -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807 (minus 9.2 quintillion through 9.2 quintillion).
If you have a Mac and it dates from 2007 or later, it’s probably a 64-bit machine. macOS has supported 32- and 64-bit applications, but from macOS Catalina (which came out in 2019) onward, it’s 64-bit only. As for Windows-based machines, if your processor is an Intel Core 2/i3/i5/i7/i9 or AMD Athlon 64/Opteron/Sempron/Turion 64/Phenom/Athlon II/Phenom II/FX/Ryzen/Epyc, you have a 64-bit processor.
Need more explanation?
The Khan Academy has a pretty good explainer of the decimal, binary, and hexadecimal number systems:
Wednesday: Day 3 continued the heavy hands-on portion of Hardware 101, the first segment of my five weeks at UC Baseline, the cybersecurity training program offered by Tampa Bay’s security guild, The Undercroft.
After taking apart and reassembling a desktop, it was time to up the ante and do the same with at least one laptop. I started with a Dell Latitude E5500, a bulky beast by today’s laptop standards, but one that’s more user-serviceable — and more easily taken apart — than most.
First step: Removing the battery.
The bottom panel was easy to pop open. It was held in place by nothing fancier than standard Phillips screws, which provided easy access to the RAM.
Next on the removal list: The optical drive. Once again, pretty straightforward — remove some anchoring screws, and then use a flathead screwdriver tip to push the the drive casing out.
The fan was quite easy to remove, as was the CPU heat sink.
Unlike the previous day’s desktop machines’ CPUs, which were in ZIF (zero insertion force) slots, laptop CPUs aren’t typically swappable, as they’re generally soldered onto the motherboard. This machine had a notebook-grade Core 2 Duo, which was typical for a mid-level laptop in the Windows 7 era.
It was also pretty easy to remove the keyboard…
…and once that was done, detaching the screen was a simple process.
With the disassembly complete, I laid out and labeled the parts that I’d extracted:
“All right, next challenge,” said Tremere, our instructor for the Hardware 101 portion of the course. “Disassemble, then reassemble the small one…”
I flipped it over, pleasantly surprised to see standard Phillips screws that were easy to access:
At this size, a laptop’s battery-to-actual-computer ratio jumps significantly:
This machine was still intended to be somewhat user-serviceable, so the battery and RAM were still easy to remove:
The drive didn’t take much effort to liberate, either:
The fan/heat sink combo didn’t put up much of a fight:
This is a machine made specifically for writing TPS reports and not much else, judging from its CPU. Still, I’m sure it could still do a serviceable job running a modern lightweight Linux — assuming it survives my disassembly and subsequent attempt to put it back together again.
Here are both patients, spread out across the operating table…
Re-assembly took a little longer, and I didn’t bother with photos of that process. I did manage to get it back together again, and with no extra parts!
I even the screen reattached! Later, I found a power adapter, and the machine managed start and get up to the BIOS screen, although the screen looked a little dim. Since I’m not trying out for a CompTIA hardware certificate, I’ll simply declare the procedure a success and not get too bogged down with fussy minutae such as “functioning” and “usable”.
Tuesday was Day 2 of the UC Baseline cybersecurity training program offered by Tampa Bay’s security guild, The Undercroft. I lucked out and got into the inaugural cohort, which means that I’ll spend 8 hours each business day in the classroom (masked and distanced, of course) for the next four weeks.
UC Baseline is made up of a number of separate units, which The Undercroft also provides individually. Week 1 is taken up by the Hardware 101 course, which is all about hardware and providing the class — some of whom have a deep technical background, while others don’t — a baseline knowledge of how the machines that make up the systems that we’re trying to secure.
I suspect that there’s an additional goal of removing any fear of tinkering.
Day 1 of Hardware 101 was mostly lectures about hardware, starting with logic gates and working all the way up to CPUs and SOCs, and Days 2 and 3 were the “tear down/rebuild” days. Day 2 focused on taking apart and then rebuilding desktops, and Day 3 took it up a notch by doing the same thing with laptops.
One of the goodies that we got (and get to keep) is the toolkit pictured below:
The first exercise was a teardown-only one. We could choose from a selection of old computers at the back of the room to tear apart, and I thought it might be fun to try and take apart this old Power Mac G5 from the mid-2000s. These machines are notoriously opaque, and I thought it might be fun to try to dig through its guts:
The Power Mac G5 was aimed at Apple’s “power use” customer — typically creatives who need serious computing horsepower. This particular machine was used by an advertising agency to do 3D rendering. As such, it’s one of the few Macs that’s easy to open, at least superficially. Take a look at this beautiful Jony Ive-designed latch:
Opening the latch reveals the machine’s aesthetically-pleasing innards, which were covered by a plastic shield. I popped off the shield and got to work.
By the way, that yellow clip in the photo above is connected to my anti-static wrist harness (another goodie we got as part of the course fee). Nobody expected these machines to survive the teardown process, but it never hurts to consistently follow standard safe electronics practices!
I then removed the cards from the two expansion slots. One was a high-speed network card; the other was pretty nice 2005-era graphics card:
Next up: The RAM!
After that came the Airport Extreme wireless NIC, freeing it from both the PCIe slot and its antenna wire:
That took care of the easy part. Time for a photo op:
Here’s what I yanked out so far. Note my screw management technique!
And now the hard part: getting to the processors. They’re encased in a pretty anodized aluminum box, and it turned out that the only way into it was to break the “warranty pin” — a plastic pin that acts as proof that a non-Apple-authorized person took a peek inside:
Behind the G5 door were the twin processors and their twin heat sinks:
I finished the teardown by identifying the components I’d extracted.
It was then time to move onto the next patient, a “TPS Reports”-writing desktop computer that we would have to disassemble and reassemble:
These are machines whose innards would need to be accessed by a mid-size office IT department, so it opens easily:
Modern computers largely fit together like Lego pieces. Even so, I kept notes on which cables went where.
Here, I’ve relieved the machine of its power supply and optical drive. It was missing a hard drive, so I retrieved one of the spare from the back of the room:
The final part of the assignment: Identify and retrieve the processor. It’s fairly obvious:
Here’s the processor, without the heat sink obscuring it. It’s an AMD Athlon II, which dates from around 2009 / 2010, when Windows 7 was a new thing:
The processor sat in a ZIF (zero insertion force) socket, which makes it easy to remove and then re-seat:
Look at all those pins. We’re a long way from my first processor, the 6502, which had only 40 pins.
Rebuild time! The machine had no RAM, so I grabbed two sticks from the back of the room and inserted into the primary slots, then put the rest of the machine back together again:
The final test — does it power up?
Success! A quick attachment to a monitor and keyboard showed an old Windows screen. Not bad for my first teardown/reassembly.