majke

26 ms·

271.
When eBPF stops doing simple subtractions(blog.cloudflare.com)
1 points by majke 7y ago|0 comments
272.
by majke 7y ago
Ha! Thanks for asking. We have a number of tests in place to test this eBPF. We are proud of them since, well testing eBPF is hard. Sadly, the tests aren't really worth much here - the bug was in the _kernel_ not the eBPF code! Once ag
273.
by majke 7y ago
Apologies. Indeed, not all generated clang eBPF is loadable eBPF. So we can sum up: - Clang can compile C - eBPF backend can generate the eBPF bytecode for subset of IR (AFAIU not all IR, not all "C" is okay for eBPF) - verifier c
274.
by majke 7y ago
Compiling to eBPF is hard! The compiler must: - avoid loops (backwards jumps) - unroll everything - inline everything - no function calls Basically, it's impossible to use clang to generate eBPF without -O2. Sorry.
275.
Hull: An alternative to shell that I'll never have time to implement(250bpm.com)
1 points by majke 8y ago|0 comments
276.
LPC 2015 – Using seccomp to limit thekernel attack surface [pdf](man7.org)
3 points by majke 8y ago|0 comments
277.
by majke 8y ago
(full disclosure: I'm affiliated to Cloudflare, but opinions here are my own of course) Thanks for posting this. I knew Paul is against DoH, but never understood his specific arguments. He has great comment about DoT (dns over TLS) cou
278.
by majke 8y ago
lwn articles https://lwn.net/Articles/646950/ https://lwn.net/Articles/152436/
279.
by majke 8y ago
> Limiting ANY queries to just TCP would have been a reasonable middle This was proposed solution, the problem is that in case of attack, against a valid Authoritative service, launched via open resolvers, the open resolvers would just d
280.
by majke 8y ago
(A) This days Cloudflare can sustain all the DNS amplifications without even blinking. This work doesn't help us in any way on the DoS front. This work helps smaller players that don't have spare network capacity. If anything, mak
281.
by majke 8y ago
Qmail used to do ANY. Qmail, as opposed to other software, was sane enough to understand that lack of MX in ANY didn't mean it wasn't there. It meant it wasn't in a cache, so it would retry with just MX. Our initial proposal
282.
by majke 8y ago
> SipHash is not secure enough for security purposes and not fast enough for general usage. I beg your pardon? The author seems very confused.
283.
by majke 8y ago
There are couple of interesting question: - backpressure - what if both userspace and SOCKMAP are doing read() on a socket? - can one use SOCKMAP to splice only a selected amount of data, and pick up the rest with read() - What are the pars
284.
by majke 8y ago
We specifically want to splice userspace TCP sockets. The unix way. I'm also not a fan of custom TCP/IP stacks. https://blog.cloudflare.com/why-we-use-the-linux-kernels-tcp...
285.
by majke 8y ago
I really really wanted to prove SOCKMAP rocks. Oh, well. Not just yet. In one of the linked videos the SOCKMAP authors say they didn't optimize part of the code. So there definitely is plenty of room. I don't see any fundamental r
286.
by majke 8y ago
Apples and oranges. Most important problem with DPDK and Netmap is that they require a dedicated network card. We don't have a network card to spare for each of our applications. Read on more: https://blog.cloudflare.com&#x2
287.
by majke 8y ago
There was a number of problems, but mostly boiling down to the sheer scale of our deployment. As an engineer I would say that KT gave us surprisingly good mileage. We were able to grow the company from couple of servers to hundreds, but at
288.
by majke 8y ago
I’m trying to convince Antirez to use this technique in redis: https://twitter.com/antirez/status/1081197002573139968 > There was a problem with that… I don’t remember what exactly, but it was like, the structu
289.
Io_submit and Linux AIO – An epoll alternative(blog.cloudflare.com)
121 points by majke 8y ago|66 comments
290.
EBPF, Sockets, TTL Distance and Manually Writing EBPF Assembly(blog.cloudflare.com)
2 points by majke 8y ago|0 comments
291.
The curious case of CLTQ instruction(joshpeterson.github.io)
2 points by majke 8y ago|0 comments
292.
by majke 8y ago
There is some relationship between motivation and enthusiasm. I think enthusiasm is the single most important thing. Naively, you could state that showing enthusiasm is an expression of internal motivation. But enthusiasm can be faked, whil
293.
by majke 8y ago
For me the highlight was the ~15:14 mark, building up on the the diving buddy example. First he mentions that creative process needs a buddy (in "pioneering conditions"): - buddy must be passionately curious - buddy needs to be ki
294.
Using FreeRTOS in small embedded systems(freertos.org)
1 points by majke 8y ago|0 comments
295.
by majke 8y ago
> Why is it impossible to parallelize this? One of the mentioned articles touches on this: http://utaharch.blogspot.com/2013/11/a-dram-refresh-tutorial... > Upon receiving a refresh command, the DRAM chips e
296.
by majke 8y ago
Watch out, there is a bit of cargo cult there. Particularly around the operations preparing the data for the FFT. The C part is rather straightforward. The pinning is helpful, but not really needed. The MTRR was a dead end (but I left the c
297.
How physical addresses map to rows and banks in DRAM (2015)(lackingrhoticity.blogspot.com)
1 points by majke 8y ago|0 comments
298.
by majke 8y ago
This reminds me of ... pthread api implemented as a standalone library! https://manpages.debian.org/jessie/libpth-dev/pth.3.en.html > The thread scheduling itself is done in a cooperative way, i.e., the threads
299.
by majke 8y ago
The title brings nostalgia! For me it all started when Michael Carter worked on orbited, years ago (was it 2008?). Michael made a number of important "discoveries", like the famous document.domain trick: https://develop
300.
The rise of multivector DDoS attacks(blog.cloudflare.com)
112 points by majke 8y ago|18 comments

More ›