majke

20 ms·

211.
BSD Socket Addressing(idea.popcount.org)
2 points by majke 6y ago|0 comments
212.
Managing 10x Engineers – The Wolf(randsinrepose.com)
12 points by majke 6y ago|1 comments
213.
by majke 6y ago
This is great. Thanks
214.
by majke 6y ago
I'm jealous. Dan is reading my twitter and re-posting stuff here :) https://twitter.com/majek04/status/1267388797278400512 You should definitely follow me on twitter :) Anyway, the question is about optimally
215.
by majke 6y ago
Funnily enough this is the code I was reading today https://github.com/constructor-igor/cudafy/blob/11cdd4def4e7... s = a ^ b; // sum bits r = a & 0x7f7f7f7f; // cle
216.
Build simple fuzzer – part 4(carstein.github.io)
1 points by majke 6y ago|0 comments
217.
by majke 6y ago
Isn't TLS 1.3 something OS should provide? Isn't QUIC something an OS should provide? Isn't the OS responsible for preventing antivirus software installing insecure browser plugins? Isn't the OS responsible for process s
218.
by majke 6y ago
Which one? Which STUN server are you using?
219.
by majke 6y ago
The article starts with GFW and then goes onto RST packet validation.It's great to see people re-learning the classics. Awesome! Extra points for using "scapy". The injected RST are often easy to drop once you know what you a
220.
by majke 7y ago
This is excellent. I never thought about it that way. Indeed, with syn cookies the receiving party has no idea what was the original sequence number, so it will treat _any_ delivered packet as first. Couple of caveats: - you can jam syn coo
221.
by majke 7y ago
I personally don't like golang's defer. For me it obscures the flow of the program. For example when I acquire a lock, I like to see where exactly it's released. For me "defer" only makes sense in the context of exc
222.
by majke 7y ago
The ack handling (loose) seem to have some undesired corner cases. Apart from that the problem with conntrack is not semantics but lack of understanding. This blog post aims to narrow the gap and explain in clear language what you can and w
223.
by majke 7y ago
Conntrack evolved recently and is thought to be fast. Florian Westphal once said that an iptable rule doing -j NOTRACK is slower than hitting conntrack. The rationale is that iptables need to do indirect jumps, which with spetre mitigations
224.
by majke 7y ago
Author here. There are more conntrack stories to tell... In this one I tried to do a quick "what happens when table is full", and ended up with a rather rich and complex blog post. In it I discuss: - EPERM - implicitly dropped pac
225.
by majke 7y ago
<update> Crap. That gist contained some draft version of aesnihash. SORRY. See updated version https://gist.github.com/majek/96dd615ed6c8aa64f60aac14e3f6ab... </update> Thanks for spending time on this
226.
by majke 7y ago
Oh boy. This is drastic. Thanks for great advice: marek:~$ time (cat logs-popcount-org.txt | sort -u | wc -l) 39057531 real 2m37.387s user 2m35.626s sys 0m2.937s marek:~$ time (cat logs-popcount-org.txt | LANG=C sort -u -S6G
227.
by majke 7y ago
2 minutes was for 40M items. I had 1B items to sort/uniq.
228.
by majke 7y ago
Ok, I'll bite again: marek:~$ cat logs-popcount-org.txt | perf stat -d awk '!a[$0] { a[$0]=1; print }' > /dev/null Performance counter stats for 'awk !a[$0] { a[$0]=1; print }': 4
229.
by majke 7y ago
Apologies, maybe I oversimplified the original problem. I'm dealing with IP's (both v4 and v6), subnets, ranges (which may or may not align to subnets). These map to one or more datacenter numbers. I could indeed define data model
230.
by majke 7y ago
`sort -u` is slower for me. Doing cat from left to right helps readability - it's important. Performance-wise cat gives you 64KiB blocks of data, while direct pipe can give more. My programs (mmuniq-*) use 512KiB input buffer, so indee
231.
by majke 7y ago
The secret is that I used a bloom filter anyway, as the next stage in the pipeline. I needed to load the IP whitelist/blacklist data quickly and rapidly compare packets against it. I had tooling to do bloom filters already. I also had
232.
by majke 7y ago
Here you go: marek:~$ time (cat logs-popcount-org.txt | awk '!a[$0] {a[$0]=1; print }'|wc -l) 39057531 real 0m41.236s user 0m38.179s sys 0m5.447s So: sort: 2m, awk 41 seconds. Also, awk used 6.1G of RAM at peak.
233.
by majke 7y ago
In the final program `mmuniq` I did a couple of, I think, interesting hacks. https://github.com/cloudflare/cloudflare-blog/blob/master/20... First, I used a hash function using aesni (aesenc) instruction
234.
by majke 7y ago
This may be a controversial opinion but I strongly disagree with using bug priority/severity/urgency/impact etc to do work prioritization. I think there are two separate things about tickets: one is the about technical merit
235.
Iommu resistant DMA attacks – all modern IOMMU implementations are broken [pdf](cs.technion.ac.il)
2 points by majke 7y ago|0 comments
236.
by majke 7y ago
I'm afraid it only got worse. marek@mrnew:~$ unshare -Ur root@mrnew:~# Not only it's easier to do than LD_PRELOAD these days, you actually _do_ get real elevated capabilities.
237.
by majke 7y ago
A similar situation happened to me, and the problem was not hints. I just wasn't happy with that particular problem because I knew how much time it took me the last day, when preparing, to solve it! I was really not into doing it again
238.
by majke 7y ago
> In most async systems … you end up in a world where you chain a bunch of async functions together with no regard of back pressure. yup. Back pressure doesn’t compose in the world of callbacks / async. It does compose if designed
239.
by majke 7y ago
You are totally right, as confirmed by Mr Kerrisk here: https://twitter.com/mkerrisk/status/1207018543193063424
240.
Linux 3.8 introduced unprivileged user namespaces [pdf](man7.org)
166 points by majke 7y ago|51 comments

More ›