ezdiy

4 ms·

1.
by ezdiy 8y ago
TLB is only an indirect cause. This is because kernel scheduler preempts processes fairly infrequently (100 or 1000hz, or dynamic, but still capped to a small number). Scheduling quantums are so large precisely to keep TLB flush overhead of
2.
by ezdiy 8y ago
HTTP2 is all about "tcp-inside-tcp" states. https://github.com/golang/net/blob/master/http2/http2.go#L81 Wherever possible, they do the sensible thing - just goroutines for each flow, pipe
3.
by ezdiy 9y ago
What do you mean "accidentaly"? You either have the file pinned, or not. If its not pinned, you're not seeding anything. If it's pinned, you're liable same as with BT.
4.
by ezdiy 9y ago
> IPFS feels closest to it as of now. I'm rather skeptic. IMO, IPFS is just overengineered bittorrent. At the lowest layer, the semantics are the same, except with really bad ideas thrown into the mix - wantlist instead of bitmaps,
5.
by ezdiy 9y ago
GIT would come closest. Note that while it's not common now (aside from some first canaries, like scuttlebutt), GIT works perfectly fine as a P2P opennet. You are much less trapped on github, because you can just take your repo and t
6.
by ezdiy 9y ago
>Then when we've all sunk into all these convenient cloud services and "easy to use" disposable devices, we'll have lost all of our privacy and power. Convenience can be engineered into P2P, too. With Bitcoin, people
7.
by ezdiy 9y ago
All pubs on the wiki are indeed overloaded. Interestingly if one sets up their own, the other pubs eventually sync with it, only the desktop client seems to be unhappy with laggy pubs. Is that by design? FWIW, you can use pub.lua.cz:8008:@x
8.
by ezdiy 9y ago
If you want a "modern web" example, nntpchan - despite the name, its not related to usenet directly, only uses the same method of federated pub/sub replication.
9.
by ezdiy 10y ago
> using different types for things that are web observable, especially Can you be more specific? > has UI lagginess with large numbers of tabs due to IPC The IPC lock contention is part of the issue (as well as tab per process model),
10.
by ezdiy 10y ago
The answer is - it's sorta comparing apples and oranges - libraries, or general C++ inlining cruft can inflate the binary size a lot indeed. A much fairer comparison is to compare source side by side. Chromium source is about 4x larger
11.
by ezdiy 10y ago
I'm posting this from chropera as we speak, and I used presto for a decade before that. While your points are spot on, there's one thing about Presto - performance, or more specifically resource usage. Modern layouting engines are
12.
Bypassing driver signing enforcement on almost all modern Windows versions(kat.lua.cz)
1 points by ezdiy 10y ago|0 comments
13.
by ezdiy 11y ago
Love how the algorithm is simple compared to iDCT based ones, very good job! > No text at all Or indeed anything which poorly maps to gradients. For this I'm thinking about instead of storing pixel values per sample vertex, store on
14.
by ezdiy 11y ago
> A fully concurrent GC is impossible That's a common misconception, of course mutability gc barriers can be made atomic. But it comes at significant synchronization cost, plus using full shared world like in C does not seem like a
15.
by ezdiy 11y ago
Current channels implementation uses fibers, ie everything runs on single thread and only stack is switched. There is no GIL. If you want true multithreading, simply use pthreads like you would in C. Just make sure to mark gc roots of new t
16.
by ezdiy 11y ago
Indeed it's becoming quite promising (using crystal already for few pet scripts which were too slow even for rubinius). Where it's most lacking at the moment is gc - it uses stop-world off the shelf boehmgc which is ok but not exa
17.
by ezdiy 12y ago
While the efforts of Matasano to make wargaming cool and hip again certainly is commendable, I think HR drones of the world also appreciate a game of "pretend to be sociable teamplayer/brogrammer". Too often hackers don'
18.
by ezdiy 13y ago
Spender did excellent write-up about how it exactly works: http://www.reddit.com/r/netsec/comments/1eb9iw/sdfucksheepor...
19.
by ezdiy 13y ago
Try dmesg.
20.
by ezdiy 13y ago
You're correct. Bug is in 2.6.37-3.8.8, fixed in 3.8.9. http://lxr.linux.no/linux+v3.8.9/kernel/events/core.c#L5331
21.
by ezdiy 13y ago
Not that I know of, hence 0day.
22.
by ezdiy 13y ago
Either your kernel is patched, or the exploit does not work on your system. It certainly fails without -O2, glance over the code indicates horrible reliance on various UB on how gcc compiles it. $ grep -A 10 'int perf_swevent_init' linux-2.
23.
by ezdiy 13y ago
Than you so much for this. You actually implemented Haskell/Caml for us, mere mortals, and as close to bare metal as possible. Bravo!
24.
by ezdiy 13y ago
Silent patched a while ago: Kernel must be compiled with PERF_EVENTS (default on most modern distros). Bug fixed in 3.8.10. This is the kill log: http://lkml.indiana.edu/hypermail/linux/kernel/1304.1/04302.... This bug got apparently back
25.
Linux local privilege escalation 0day, 2.6.37 - 3.8.10(fucksheep.org)
140 points by ezdiy 13y ago|92 comments