critical vulnerability in wifi chipsets

[@LowLevelTV]

now way haha lowlevel.academy is really cool !?

[@paulmichaelfreedman8334]

The community is divided about your name change. Here's my proposal for a compromise: Low Level Lear

[@TremereTT]

Are range checks allways on in Rust plus they also exist for all kind of buffers?

[@qriist1850]

It's neat that you went to Twitchcon with the CEO of Jurassic Park. 🥰

[@TechnoMinded-qp5in]

I think in the future businesses will run Linux servers and not Windows servers.

[@plank_length_architect9539]

Just out of interest. You provide good curses, but sadly, the only payment options are im gonna call them the American payment systems (cash app, visa, American bank). So to the question: will there ever be any other payment options ? (Seap, direct transfer, PayPal, etc) Anyways, nice video, always great content, and all of you have a great day.

[@williamdrum9899]

He dropped "learning" because it's been nearly half a century and people still haven't learned how not to index an array out of bounds

[@LowLevelTV]

HA

[@amadzarak7746]

This comment wins

[@besterspieler2285]

@@amadzarak7746 you win

[@BastetFurry]

>still haven't learned how not to index an array out of bounds Some people should only program in BASIC, it won't let you write out of bounds, you have to fight and sabotage BASIC to do so.

[@FinaISpartan]

Yeah... I think we should just collectively stick to bound-checked languages at this point...

[@s_9280]

He has stopped learning. There is nothing left to learn. He has reached the lowest level.

[@LowLevelTV]

it's true i am immortal

[@AffectionateLocomotive]

🤨🤨🤨

[@RokasSondaras]

Calling your level lowest possible is a huge compliment who understands it. :D

[@false_positive]

they say you can't fall lower than binaries

[@szaszm_]

I wonder if he's gonna correct the "I was right." video, where he was wrong.

[@_JohnHammond]

no more learning? stooping to this level is pretty low, double L

[@opposite342]

ikr, double L *is* double L-ed indeed

[@bearwolffish]

bruh, should have been a battle rapper with them bars.

[@sasjadevries]

At least the name didn't go out of bounds. From 3 L-words to 2 L-words is not overflowing.

[@Mousedahbidoof]

2 L's make a W

[@RustIsWinning]

The only L I see here is the underscore before JohnHammond 😂

[@zekicay]

Please stop referring to forks of OpenWrt as OpenWrt, it would be the same as referring to Mint, Zorin, Pop! as Ubuntu. OpenWrt is not vulnerable to this. Additionally OpenWrt 19.x and 21.x are no longer supported, but regardless they are not vulnerable as they are not using Mediatek's proprietary SDK - the only thing with the bug. The mitigations are available as they are used by default in OpenWrt and Mediatek forked from it.

[@falsemcnuggethope]

This. I've been clenching for a while because I thought I would need to patch my router before visiting the office. Turns out this is a complete non-story. Openwrt has nothing to do with this vuln. I will go and release my babies now.

[@vprwave]

It's pretty surprising how low level the fact checking is on this channel.

[@squirlmy]

@@vprwave I notice this especially in security issues. Clickbait is not discouraged, because it usually "well, they know more about security, right? Better prepared, right?" We need to stop pushing theoretical security vulnerabilities because making people paranoid is good for business

[@Xi_Pooh_Shill]

Good. I just finished reinstalling 2 routers some days ago.

[@jaimebondoza3710]

90% of tech youtube is dumb misinformation for clicks

[@phloxpluto]

rip low level learning. you would've loved low level 😔

[@agastyasanyal4026]

Nah I was his classmate in school, everyone called him low level Even the teacher calls him low level

[@dudedavid522]

Low level media has a nice spin to it, bonus "fundraising round" initials

[@sakurako-omuroo]

​Low level classmate​@@agastyasanyal4026

[@soyitiel]

next up: Low, just low

[@collin4555]

shorty got low

[@Mordecrox]

LPL: This is Lockpicking, and today... FW: Thanks for turning in for Forgotten, I'm iMac... And of course SteveMRE1988 becoming just Steve

[@bluexo256]

she hit the flo

[@Kane0123]

Low level Shillong amiright?!

[@PeterZaitcev]

Low. I am the Low.

[@Test_749]

low level is no longer learning

[@valentinn3507]

Yea, just low level content 😂 Nahh it‘s a fine name, the old one was good already tho.

[@AffectionateLocomotive]

Brain is no longer braining.

[@pluto8404]

a new low

[@BurgerKingHarkinian]

It's only downhill to a lower level from here

[@wasshattereddreams]

bro's gonna make a surprise video showing a modern CPU through microscopes and find hardware vulnerabilities

[@scientiac]

Aww, it's like when "it's okay to be smart" rebranded to "Be Smart" . Low level learning was quite good, low level is fine though. Low level learning is stuck on my tongue though it's so fun to say: lowlevellearning lowlevellearning

[@Dumby]

Hi, his name is edthisislowlevellearning, he teaches us about computer security

Vanilla OpenWRT is not affected by this, as it uses open source mt76 driver with hostapd. wapp is part of Mediatek's proprietary driver

[@rossjennings4755]

This is really important information, and should have been called out in the video. I'm running OpenWrt at home on a Linksys E8450, which has a MediaTek MT7622BV SoC and an MT7915E Wi-Fi chipset, both of which are called out as "affected chipsets." Judging by the video (and the blog post it refers to), I might very well have been affected. It took a trip to the OpenWrt forums for me to figure out that wasn't the case.

[@MegaTomPL2]

M i affected? Should i do something? Hostname OpenWrt Model Xiaomi Redmi Router AX6S Architecture ARMv8 Processor rev 4 Target Platform mediatek/mt7622 Firmware Version OpenWrt 23.05.0 r23497-6637af95aa / LuCI openwrt-23.05 branch git-23.236.53405-fc638c8 Kernel Version 5.15.134

[@MegaTomPL2]

Okay, I checked it. You're right: even if the processor is affected, the PoC is using wapp, so we’re safe for now. However, to be sure, I upgraded to the latest version of my OpenWRT. The strange part is that it is exploited via UDP, so it’s affected globally. Even if someone was affected and got a reverse shell, they could be backdoored after an update. Good thing that vanilla OpenWRT doesn’t use wapp.

[@natek4391]

is it good practice to update device firmware when something like this triggers your worst fears? I have a mediatek openwrt router serving multiple virtual ssids so am a little worried I'll break the complex setup in an update

[@prophetzarquon]

Indeed, some updates can introduce their own vulnerabilities... I'd say a _related_ update is a must-get; otherwise, minimize exposure & apply patches only after confirming stability.

[@tranthien3932]

Ed: *Drop "Learning"* Literally Everyone: My Disappointment is Immeasurable and My Day is Ruined.

[@Turalcar]

I, for one, just like alliteration

[@pranjal3727]

At this point, all IDEs and compilers should issue a warning for all memcpy, memmov and DMA operations if they do not have a length limit check immediately before it.

[@angrydachshund]

Ten minutes later, the devs will disable those warnings at the project level, because obviously they are smart enough that their code does not have such bugs.

[@garrickstokes]

@@angrydachshund That's what I do 👍

[@pranjal3727]

@@angrydachshund -Wno-warning-i-know-better

[@moveonvillain1080]

IDE alerts the FBI when you make such mistakes and you get swatted.

[@nickwallette6201]

Sounds simple, until you recognize that offsets and lengths are often passed as variables that are the sum or product of other variables that may all have been individually verified as in-bounds, but when combined, are not. The reason buffer overflows still exist is not because people don't ever sanitize inputs or enforce limits. Okay, sometimes it is, and that's just sad. But sometimes it's just a series of unforeseen circumstances. I once considered what it would take, in C, to ensure that no arithmetic ever overflowed the target variable. It was a genuine exercise of "what would it take for a single function to be literally bug-free?" Some hardware actually has an overflow status flag in the ALU, but you can't assume that if you're going to write portable code. You could use a data type that is guaranteed to be big enough to hold any possible input, and then cast down to whatever the target type is -- but only if a larger type existed. Otherwise you would need to roll your own int128 or whatever. Or, you could measure all the inputs and determine if there's a possibility of overflow based on their values, but then you're really better off using a language that already does this (slow) or overloading the basic arithmetic operators to do it yourself (AFAIK, not possible in C), or using library functions instead of operators. None of those are really attractive options, and the vast majority of code out there doesn't do any of this. It all just assumes that the inputs are sane. Ergo, most code out there has inherent flaws that could be considered "bugs." It's a sobering exercise.

[@bparker06]

"you know it's being more openly monitored", why is there an assumption that most open source code is actually monitored?

[@avarise5607]

Its monitored, by like 2 people, dev and the gvmt agent on his way to make name with new fancy backdoor 😂

[@tablettablete186]

​@@avarise5607So 2 eyes now!

[@benargee]

Most people confuse "it is monitored" with "it can be monitored"

[@Minsk_Accords]

@@tablettablete186 better than 5 eyes

[@rogo7330]

For me, it's not about someone done audit on the code - it's about that you can torn thing out and replace it with something else, including your own binary.

[@ImSquiggs]

Not sure how I feel about the rebrand, but I'm generally terrified of change, so even a KZbinr deciding to remove eight letters from their name can be too much for me sometimes, haha.

[@LowLevelTV]

I am too, but thats life :)

[@crusaderanimation6967]

Same, i have force awakening poster in my room, i haven't watch the movie (maybe snipptets on TV) i don't plan to since (from what i've heard) it's rather shitty movie, but i have it since middle school (or rather polish equivalent of there of) and i'm 22 in December so now removing it triggers my sentimentality

[@joshxwho]

It's just a nice change it's a rebrand without the word A.I or Cloud afterwards. 😅

[@moveonvillain1080]

You have replaced about 330 billion cells in your body over the last 24 hours. You are not even the same person you were 30 minutes ago 😅

[@squirlmy]

@@moveonvillain1080 you are really only a story you tell to yourself, or rather your new self. Ponder. Time for beer!

[@Lutz64]

Who goes to new Jersey for vacation?

[@angrydachshund]

You do if your priest assigned it as penance.

[@NotMarkKnopfler]

Was thinking the same thing. Who would want to go on holiday anywhere in the USA (Montana being a notable exception!)... I'd rather spend a week in a septic tank 😅

[@shreyansdoshi]

​@@NotMarkKnopfler Upper Michigan, New Hampshire and Pacific Northwest are quite beautiful too.

[@griffinfaulkner3514]

@@Lutz64 I'll deal with the state if I get to see the ship.

[@balsalmalberto8086]

When I went with my dad we partied with guidos all night long and got spray tans shot up steroids.

[@dehrk9024]

thank you for being here, i have no close friends my family kicked me out barely affording rent but your videos give me hope and keep me going

[@LowLevelTV]

you got this dude

[@innovationsanonymous8841]

Time to download ghidra and qemu, no?

[@tayzonday]

“Low Level Learning” speaks to SEO and reconciling your content vertical with viewer behavioral telemetry. Both might be more important than reconciling your content with how humans parse ease and memorability.

[@9hoot789]

Low Level Learning is both easy to remember and has a better ring to it, I really can't see any pros to dropping it, like, at all.

[@Ironically-Sarcastic]

Bro have you been hacked? What are you talking about? You starting a new "Tay SEO Tips" channel or something, and you're testing the waters in the KZbin comments?

[@KMRamirez]

​@@9hoot789 he is referring to discovery by search engines and recommendation algorithms, as opposed to traditional branding/word of mouth

[@balsalmalberto8086]

Keep bringing us that chocolate rain the comments bro.

[@massgrave_]

Common Tay Zonday W as per usual

[@arashai]

What’s with the super clickbait titles these days? Love the channel but feels bad, man.

[@avarise5607]

Blame is on you for getting baited, not the author on using it. You got free will, use it

[@BurgerKingHarkinian]

​@@avarise5607trash take. You can like a video and still hate the thumbnail. Boycotting a video that you think you might like just because of the thumbnail is a bit extreme, don't you think?

[@marcogenovesi8570]

clickbait works

[@bytefu]

@@marcogenovesi8570 I've stumbled on this channel in the recommended section, and liked some videos, but these clickbait titles with no substance deter me from subscribing. I want the title to be on point, so I wouldn't waste time watching half a video on something that I am not interested in just to figure out what it's about. Basically, this makes me watch fewer videos, not more. But if his goal is to attract an audience of people with too much free time and no filter, I guess that works.

[@da5e]

Might be showing my age, but WAP will always mean 'like the internet, on a Nokia 3310, but somehow worse than that implies' to me.

[@NoBug404]

Good old wait as you pay.

[@SpiritmanProductions]

Despite its limitations, Wireless Application Protocol was great in its day. I used it mainly for viewing real-time train info, which was really useful when the service was disrupted and I was stuck at a small station with no staff and no screens.

[@KvapuJanjalia]

Blindly casting a BLOB of bytes to a struct. This is genius level C.

[@Scriabin_fan]

Next rebrand: Low

[@pluto8404]

thats too hard to say. Next Next rebrand: L

[@MacDKB]

After that, "L".

[@AstoundLikesWoW]

Dunkin donuts was too hard to say as well. They rebranded to Dunkin. But I heard that too is much too difficult for the average person. Hopefully the Dunk rerebrand will be enough.

[@Audacity_69]

Having a good day LLL releases a new vid Misery resumes

[@kyeeehehe]

Having a good day LL releases a new vid Misery resumes

[@GREG_WHEREISTHEMAYO]

Actually 🤓☝️ it’s LL now Edit: Nooo someone beat me to it

[@kyeeehehe]

@@GREG_WHEREISTHEMAYO hahaha no worries

[@すどにむ]

The reason why lots of those "security features" like ASLR and oft mentioned IOMMU aren't enabled for low level things that interacts with real world is, because they're inconsistent against the real world they interact with. Lots of those makes jittery mess in terms of response times. Webdevs think those are just one click enables because Web is jittery mess anyway but not everything in the world is.

[@ahmednishaal9432]

@@すどにむ 🤣🤣🤣🤣

[@Иіск]

I understand why you rebranded to Low Level. However, my 2 cents is that I prefer Low Level Learning😔

[@AstoundLikesWoW]

There are a surprising amount of companies who have shortened their names or simplified their logo. Every company has the same generic basic font and logo style now. Lame

[@boldizsarszabo]

Should've rebranded to Low Level Yapping

[@linuxguy1199]

A few years ago I started breaking up and isolating key pieces of my network infrastructure from each other, and im thankful I did as I've pretty much been able to live worry free even with all these problems coming out.

[@user-qr4jf4tv2x]

Low Level: i'm beneath you but nothing is beneath me

[@paulmichaelfreedman8334]

Behold the Underminerrrrrrrr

[@stevemaricar4350]

The detailed rundown you gave on the potential implications and mitigations is much appreciated. Makes me think twice about the trust we put in brands when it comes to tech security.

[@mskiptr]

> use software provided by your hw vendor > get bad quality code who would expect that…

[@mskiptr]

and that is why we should push for FOSS firmware

[@sujimayne]

Implying that FOSS equates to quality production code is just a fundamental misunderstanding of FOSS.

[@mskiptr]

@@sujimayne It's very common for the board support package code to be extremely low quality, quickly slapped together "just to make Android work". That's in contrast to upstream code within major FOSS projects like the Linux kernel or Mesa. FOSS firmware is a different aspect (I wrote the top-level comment before I knew the vulnerability is specifically in the wlan firmware and not the main OS of these routers). It would still help a lot because then it could be audited much more easily _and_ it could be patched for good in all open projects using that firmware. It's much harder to find all blobs sharing the vulnerability than to just patch it in one codebase.

[@mskiptr]

@@sujimayne But yes, you're right that "just make it FOSS" likely wouldn't solve this issue. (but sharing code between more of such chips, coreboot-style, probably would)

[@marcogenovesi8570]

@@sujimayne SDKs provided by embedded hardware vendors are often low quality code, and more often than not the FOSS version is better (if they released enough hardware information about it). This hardware is not as complex as GPUs so it's still within the abilities of a few community members to develop

[@saturten]

in my head this guy will forever be the tech bro version of nick from the yard

[@jackthatmonkey8994]

Except he is a positive impact on society unlike most other tech bros

[@jameso7826]

I love your channel because it's so full of information and you always present it very well and make the learning experience enjoyable, which is why I have notifications on. I hate your channel for every notification pushing me toward being a paranoid mess about all the damn RCE's that keep popping up.

[@casperes0912]

Ridiculous thing is that we’ve got tools that make it practically impossible to make buffer overflow vulnerabilities

[@tom23rd]

Being from NJ I can say "there's your problem" lol

[@LowLevelTV]

same

[@MattDog_222]

Didnt know LLA was one time, thats a very nice thing to point out. I usually dont buy subscriptions since i'm too busy working full time to use it effectively

[@johneric2720]

I actually liked the "Learning" as part of it, but it's totally fine. I always love to see someone else who is as passionate about the C programming language as I am 🙂and is not moved by modern technologies that make things too easy, bringing a high level of abstraction that people don't care to know about. Sometimes I even try writing my own libraries in C. I have used C in tons of projects which I am so proud of, the most recent is an HTTP Web Server, and I would love to make it "Open Source".

[@ozzelot3349]

A MediaTek product has a hole in it. The world yawns.

[@krazyolie]

the guy who used to set next to me at work regularly complained about working with mediatek as a customer - gave impossible to understand instructions and information.

[@cryonim]

@@krazyolie Even as a consumer of its products, its bad. Custom roms on android aren't made that easily on their firmware because they don't release their kernel (modified Linux kernel) for each phone to the public, meaning if you have a phone with Snapdragon, you can atleast expect a custom rom or build one yourself but its much harder on MTK.

[@Tim_Small]

There is no chip bug - this is just a bug in a proprietary daemon they ship to their customers by default. Despite what's said in the video OpenWrt isn't vulnerable (only the hacked-about fork of OpenWrt which MediaTek shipped to their customers). On the whole, the wifi chipset MediaTek open source support is very good, and they employ one of the OpenWrt contributors to develop their open source wireless drivers (the same guy that did the ath9k driver before Qualcomm took most functionality into their proprietary firmware with ath10k and later chips).

[@ozzelot3349]

@@Tim_Small Comment edited accordingly. :D

[@unduloid]

Yeah, you write programs in a programming language that encourages memory leaks, and then there will be memory leaks. Curious, that.

[@poglavicas]

This only applies to stock factory drivers, not opensource ones.

[@mathgeniuszach]

> ed stops going on vacation > vulnerabilities cease to exist > ed goes on vacation again thinking he's safe > the world burns

[@stan.rarick8556]

As a longtime assembler programmer I have always touted learning machine architecture via assembler language as a useful benefit in terms of stability and performance

[@mastersingleton]

The following three WiFi 6 Routers released in 2020/2021 features the combination of MediaTek's MT7621A network accelerator and the affected MT7915 Wi-Fi 6 connectivity platform are as follows: - D-Link AXO AX1800 (DIR-X1860) - TP-Link Archer AX20 (AX1800 Dual-Band Wi-Fi 6 Router) released exclusively in China - Buffalo AX1800 (WSR-1800AX4) released exclusively in Japan Check if you own any of the aforementioned routers and please update its firmware ASAP to the latest version that patches up this vulnerability; if its not available then replace them with a WiFi 6E Router that features either a Broadcom or Qualcomm WiFi 6E SoC.

[@q1joe]

The real shock here is SonicWall found this.

[@warasilawombat]

Honestly, SoC driver code is usually the absolute worst code in my embedded projects. It’s crazy that we rely on it for literally everything.

[@MelodicMethod]

i appreciate the detailed mitigation explanation; thanks!

[@bFix]

actually ARM64 is starting to get mainstream. Mediatek's new cheap architecture mt7981/mt7986 etc. is fully aarch64. mt7922 was like their first ARM64 chip though. mt7981 isn't expensive. you can get it for way less than 100 bucks mt7986 likely costs around 100 bucks (comparable to mt7922 when it came out)

[@AlcatelVelikiy]

MediaTek is hilariously bad for security, all MediaTek phones bootloader security tools can be bypassed with open source tools and dumped partitions too 😂

[@marcogenovesi8570]

The article is wrong, OpenWrt isn't using and has never used the vendor driver, so this affects only Mediatek SDK and vendor/custom firmware derivated from that.

[@gg-gn3re]

a lot of countries like japan, india and majority of china are a good 20+ years behind in software development so buffer overflows are going to be all over for another few decades. The easiest solution is get them away from C

[@m3talHalide-rt2fz]

"on vacation in new jersey" seems like `if true: return false`

[@nexxai]

"I was on vacation in New Jersey" That seems like an oxymoron

[@alexhiatt3374]

extremely funny that we're blindly trusting the network provided size field in current year

[@Afif87123]

atleast this forces mediatek to push updates to their old network cards, mine haven't gotten new one in the last like 2 years lol

[@falsemcnuggethope]

Imagine how many new holes they will introduce

[@garanceadrosehn9691]

I can see getting rid of "Learning" in the name, but "Low Level" by itself seems kinda weird. Maybe add some other word there. Low Level Insights Low Level Vacations.

[@CentreMetre]

Thank you for sacrificing you vacation days for the sake of stopping any more computer security vulnerabilities appearing .

[@evanmurphy8531]

"On vacation in New Jersey" should be classified as an oxymoron

[@MrSlm1982]

"On vacation in New Jersey".... I don't understand that sentence... They are English words but don't make sense in that order 😂

[@MrSlm1982]

@JimAllen-Persona if you've been to Jersey Shore... And then been to any other beach on the east coast... Jersey Shore is pretty much a last resort...

[@sk9la]

Ed and John, my favorite red hair hackers😄 Love to see both of you in a video

[@Kiyuja]

gotta love the "but what would Rust do?", this never gets old.

[@Chris-on5bt]

Great content Ed. Keep at em you are doing great!

[@blehbleh9283]

Openwrt in thumbnail but more like a mediatek vuln which is scarier

[@falsemcnuggethope]

OpenWRT is not even affected.

[@blarghblargh]

@@falsemcnuggethope OpenWRT 19.07 and 21.02 are affected

[@existenceisillusion6528]

Were you in Texas before New Jersey? I met a cybersecurity person who looked a bit like you, and I recommended your channel. 😂

[@herzogsbuick]

ya know i'm from pennsylvania, minutes away from the banks of the delaware river that serves as our state line, and i've never heard of someone being "on vacation in new jersey"

[@BinaryAdventure]

Low Level is a great new name!

[@ongunkanat]

I think your interpretation of "Would this happen in Rust" is still a little bit misleading. As opposed to every other popular language in the world, C lacks common abstraction mechanisms on top of lacking memory safety. So creating generic data structures like Vec, std::vector or ArrayList is stupidly difficult and doesn't scale well with the human effort. So there are no standard library types like Vec. Therefore C forces developers to create their own instance of the same functionality for every single use case over and over again. The C array types are also extremely limited operation-wise. You cannot copy the contents without something like memcpy which always takes the size externally (i.e. the array type does not encapsulate) so there are some tens of implementations of the same Vec in the same codebase and probably thousands of independent calls to memcpy. In the other languages you implement one type that encapsulates all the size information and write one function for copying the contents. That function is easily unit-testable and the encapsulation prevents the developer from fat fingering the size. Creating these kinds of vulnerabilities (even in DoS form) are really hard with idiomatic Rust, C++, Java, Python ... not just because they are memory safe (C++ isn't) but they are much better languages to create well-tested abstract libraries in.

[@ryogo_kurogane]

like petition to turn back to Low Level Learning (please dont ban me i love your videos)

[@Serversurferz]

It had a nice ring to it for sure

[@AleksyGrabovski]

You should learn Ada if you really care about security

[@marklonergan3898]

CVE-2024-20017... So it's the 20,017th vulnerability found so far this year... Feels like there's been more than that this year! 🤣

[@jsdcool3401]

I was also on vacation in New Jersey when crowd strike happened

[@dave7f611]

Another example of the mutual exclusivity of "online" and "secure".

[@themichaelw]

Rust boys will have a field day with this one

[@nordgaren2358]

Could not agree more about learning C. C is a great language to learn how computers work, and that's coming from someone who uses Rust for just about everything. My suggestion would be to learn C, then Rust, then revisit C a year later and see how your habits have changed. I guarantee you will write different C code after using Rust for a year.

[@michaelogala4443]

I think L3 would have been a good name too. People who were here before the rebrand would know that it means Low Level Learning but it would also be a play on the L3 CPU cache which is pretty low level :)

[@chuckcrizer]

Code from a chipset mfg is just sample or test code, it is never meant to be used in a production environment. Such code often doesn't have bounds checking or full error handling. Many are written in plain K&R style C for simplicity.

[@marcogenovesi8570]

the vendor drivers are not sample/test code, it's what you are supposed to use because they don't provide documentation about the hardware so you can't develop your own anyway

[@nytr]

Low level learning was good too man :D

[@ciaduck]

"Every time I go to defcon there's a new exploit." Almost like it's a security conference, where people present exploits...

[@addmoreice]

If you use fuzz testing on that code(if you are processing something like packets or protocols or encryption, you should!) then the fuzz would 100% find this bug before it could become a DOS. Rust has a lot of *amazing* fuzz testers. Highly recommend fuzz testing for this kind of thing, regardless of the language. Not perfect, it would be better if the language blocked it from constructions, but it's a definite move forward.

[@bizzehdee]

Its easier to write safe code in rust, if you are willing and free to learn rust. People writing firmwares are usually in well established industries that are notoriously slow moving... so theoretically, rust would have turned this into a different bug, but, someone building this in rust, would likely not know what they are doing as well as they knew what they were doing in C, and would have created lots of other different bugs

[@metalunits]

It's because these get announced while big conventions happen, so that scam time is maximized before people are able to get home and start patching. Also, I think they like to cause cause chaos and promote word of mouth since everyone is together at the conventions

[@xbsidesx]

“I won’t have vacation because work always happen when I leave” is not a healthy thing to pass along in yellow September.

[@sasukesarutobi3862]

I feel like it shouldn't, but I still find it fascinating how the difference between an RCE and a DOS can just be a copy check.

[@samsupertaco]

Please give less click-bait titles. I want to watch videos I’m interested in. Thumbs down on every vid due to title.

[@arslanrasit]

@@samsupertaco same

[@ohmygoditskornn]

Is your course (the C course) would be suitable for an absolute beginner with no programming or IT experience? 31 year old neet here.

[@edinatl2008]

Might use your courses for CPEs, looks dope.

[@VictorGamerLOL]

Next up we will have Low and then we'll have .

[@RiwenX]

Next brand update: Low.

[@rursus8354]

I think the problem is burritos. Whenever you eat a burrito, Internet will crash, AI will crash, some plane somewhere will crash. Burritos are dangerous.

[@guest7329]

NZ is rarely enabled because memory permissions require a memory management unit

[@jaysonrees738]

Today I learned that my WAP has an Atheros chipset in it.

[@morezco]

Love your work so much dude

[@viktoreidrien7110]

superb video, thank you again

[@jeremylindemann5117]

Is it difficult to write code that prevents stack overflows or underflows or do programmers generally not do it for some reason? I'm a relatively non-technical person but it seems like I've been hearing of multiple exploits recently which were enable by stack flow issue and I've been wondering if these could have been easily avoided by writing better code in the first place. Also, not sure if you already explained it but, are those vulnerabilities on the control chips fixable by update or are they hard coded in firmware?

[@tsclly2377]

passed this in to one that can have problems with this..

[@kodekorp2064]

Bro… You know that the attacks and new attacks coming out is always non stop and more numerous than there are defenses and securities.

[@SourceCodeDeleted]

Man, you make me want to return to reversing again...

[@thisisnotok2100]

Ayyyy llg Low level gang

[@TheMinersAgain]

Low level learning was iconic. Your channel though 🤷

[@mrpocock]

These issues would end almost over-night if there was real liability for bugs with real-world costs. Imagine if the vendor was liable for monetary sanctions that are a multiple of the economic impact of the exploits they ship. Screw up a buffer overflow that enables an exploit that's going to cost a billion dollars to mitigate? n x million dollars sanction. Boom! I guarantee that within a year or two, companies would be taking code quality seriously. Most stuff would get ported to rust just so they can claim in court cases that they've done due diligence.

[@retromanc3r]

Excellent research and analysis. Thanks and remember to Hack the Planet! 😊

[@foobarf8766]

Before assignment check the length, it should be easy. Should...

[@kuroenekodemon]

Why can't these companies make good products that are also secure? Now I gotta check if there's a wifi chipset update for my Wifi card...