sweetjuly

10 ms·

91.
by sweetjuly 2y ago
There is no real safe way to use lock hold timeouts. While a waiter can timeout and possibly handle failing to acquire the lock, there's no generic safe way to steal the lock from the holder after a timeout since the holder may still
92.
by sweetjuly 2y ago
Not entirely. Cryptographic primitives are based on hardness assumptions . For example, we assume discrete log cannot be computed in probabilistic polynomial time when we leverage things like Diffie Hellman. No one has yet proven whether t
93.
by sweetjuly 2y ago
It's not completely to blame on the browsers. We had the concept of green threads [1] well before that, albeit for equally dubious reasons: OS threads were believed to be too slow, and so green threads were composed onto OS threads. It
94.
by sweetjuly 2y ago
Boeing has really gone down hill since the 2800s
95.
by sweetjuly 2y ago
Note though that both of these solutions are racy and so should not be done if "someone symlinking really fast and swapping the binaries" is in your threat model. Linux proc/self is safe though, just not the result from readl
96.
by sweetjuly 2y ago
Even so, you shouldn't be waking all 24 cores, the GPU, etc. just to record video. Let the cameras DMA into their buffers and wake up a single core when the buffers hit a high water line. The core only needs to be awake long enough to
97.
by sweetjuly 2y ago
This would require specifying a cache line size in the ABI, which is a somewhat odd uarch detail to bubble up. While 64-bytes is conventional for large application processors and has been for a long time, I wouldn't want to make it a r
98.
by sweetjuly 2y ago
The C extension authors did consider requiring alignment/padding to prevent the misaligned 32-bit instruction issues, but they specifically mention rejecting it since it ate up all the code size savings.
99.
by sweetjuly 2y ago
Does Google actually administer the insurance plan? Usually the way it works is that they hire some traditional insurance company to administer is and then they just payout. Administering the plan themselves is very risky and has tons of co
100.
by sweetjuly 2y ago
> It provides both opt-out protections as an attempt to mitigate XSS attacks nitpick, I guess, but CORS is about CSRF (Cross Site Request Forgery) and not XSS (Cross Site Scripting). If you have an XSS bug, CORS can't save you since
101.
by sweetjuly 2y ago
It's a little weirder. At least one translation granule is required but it is up to the implementation to choose which one(s) they want. Many older Arm cores only support 4KB and 64KB but newer ones support all three. The size of the t
102.
by sweetjuly 2y ago
Hmm, I'm not sure that's quite right. ARMv8 supports per TTBR translation granules [1] and so you can have 4K and 16K user processes coexisting under an arbitrary page size kernel by just context switching TCR.TG0 at the same time
103.
by sweetjuly 2y ago
Intel's relationship to power efficiency is really odd. They only started doing heterogenous/big.LITTLE in 2020 despite ARM based SoCs using it for more than a decade to great advantage. It's not as if portable, battery power
104.
by sweetjuly 2y ago
Is clockless useful in the face of modern ASIC tools? Not that you were suggesting in either direction, but it's been something in the back of my head for a while now. Sure, there's a slight power and latency advantage in avoiding
105.
by sweetjuly 2y ago
> As we have seen, enabling SMT on a CPU core requires sharing many of the buffers and execution resources between the two logical processors. Even if there is only one thread running on an SMT enabled core, these resources remain unavai
106.
by sweetjuly 2y ago
NSUserDefaults, the modern programmer's fork DB :)
107.
by sweetjuly 2y ago
There's also the theory that obvious scams and spam is helpful to adversaries because it filters out people who will realize it's a scam later and not follow through. Like, making the Nigerian Prince emails more convincing might l
108.
by sweetjuly 2y ago
It can really be anywhere. Intel/AMD probably have done due diligence in ensuring that the SoC itself sleeps correctly, but once system integrators start glueing on external hardware like radios, external graphics, floppy drives ( :P
109.
by sweetjuly 2y ago
Terrifying, I love it :) How was the performance in the end? Did you get a good speculation success rate? It'd be cool to predict which resources are speculation safe (ie the cache headers don't permit it, but the content in pract
110.
by sweetjuly 2y ago
Yep :) The real trick is to not be vulnerable to known issues, and then mitigate post-compromise like crazy on the off chance you get patch gapped or (very unlikely) zero dayed. Blocking IP addresses is extremely silly, especially in an IPv
111.
by sweetjuly 2y ago
A typical operating system should not have stability issues just because of badly prioritized interrupts. You may have severe performance issues, sure, but if you're crashing because someone didn't handle an interrupt in time, som
112.
by sweetjuly 2y ago
This is a different feature than a simple logout button though (something along the lines of "sign me out everywhere"), and many (smaller) websites don't actually support this. As far as I can tell, the very website we'r
113.
by sweetjuly 2y ago
I think you need to take this a step further and really define your threat model instead of being despaired :) If an attacker is able to steal a victim's cookie database, their system (or at the very least, their browser) is already de
114.
by sweetjuly 2y ago
Plus, a compromised browser could just block the logout request. It's a nonsense and indefensible threat model. If you need to worry about compromised browsers, you need a mitigation which doesn't rely on the browser or that'
115.
by sweetjuly 2y ago
Why would you need to revoke on logout? Forgetting seems to be enough in all cases except maybe SSO revocation because in all other cases you can (and indeed often must) trust the client to protect the credential.
116.
by sweetjuly 2y ago
This makes sense for nation states playing war games but what about smaller actors? Terror groups seem content blowing up shopping malls and driving cars into parades (which are all things which I imagine are planned on a much shorter scale
117.
by sweetjuly 2y ago
You can definitely go considerably faster. A lot of these FOSS cores are either outright unoptimized or target ASICs and so end up performing very badly on FPGAs. A well designed core on a modern FPGA (not one of these bottom of the barrel
118.
by sweetjuly 2y ago
Why don't we see these attacks though? I know they're worryingly practical and the West certainly has enough enemies (especially from extremist groups who don't have the same peace keeping concerns as a nation state), and y
119.
by sweetjuly 2y ago
it's good for the economy!
120.
by sweetjuly 2y ago
Yes, the textual IR format is not an API. It's an internal implementation detail which changes somewhat rapidly. It's well specified but there's no guarantees that if you write some IR which works in LLVM 16 that it'll w

More ›