Tweetovi
- Tweetovi, trenutna stranica.
- Tweetovi i odgovori
- Medijski sadržaj
Blokirali ste korisnika/cu @jkbs0
Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @jkbs0
-
Oh
@Chelsio1, be a little creative :-) ~/src/linux $ git grep -Phe 'int chtls(_listen|_start){2}' -- drivers/crypto/chelsio/\*.c int chtls_listen_start(struct chtls_dev *cdev, struct sock *sk) static int chtls_start_listen(struct chtls_dev *cdev, struct sock *sk)Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
If you think you don't need tracing or perf to read LKML, you are mistaken :)https://www.codeblueprint.co.uk/2016/12/19/a-kernel-devs-approach-to-improving.html …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Awesome talk on digging into micro-benchmarks with perf tool https://www.youtube.com/watch?v=nXaxk27zwlk … - what does perf report -g show by default and why is it not what you would expect? - what does asm volatile("": : :"memory") trick do? - why likely()/unlikely() annotations matter?
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
More than I ever wanted to know about strict aliasing, or why __attribute__((__may_alias__)) is useful -https://gist.github.com/shafik/848ae25ee209f698763cffee272a58f8 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
How function tracing in Linux (ftrace, perf, SystemTap) works under the hood. Level expert.https://twitter.com/KernelRecipes/status/1181465265348993024 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Anyone looking for some weekend reading?https://twitter.com/Cloudflare/status/1183029263773114368 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Best work-around I know of - from user-space: clock_getres(CLOCK_REALTIME_COARSE, &res); jiffy = http://res.tv _sec + (((double)http://res.tv _nsec)/1.0e9); Based on https://www.mail-archive.com/kernelnewbies@nl.linux.org/msg08850.html …
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Do we need a BPF helper to get HZ (jiffy duration)? SYN-RTO timeout is in jiffies and TCP-BPF progs can set it - https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/diff/samples/bpf/tcp_synrto_kern.c?id=61bc4d8daa7af018d7ea084ea38648828a5a90d5 …
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
jkbs proslijedio/la je Tweet
We decided to figure out what TCP_USER_TIMEOUT and SO_KEEPALIVE socket options actually do... Ten days later I'm confident nobody really understands it (and I didn't even look at the funny states like FIN-WAIT, LAST-ACK) Code: https://github.com/cloudflare/cloudflare-blog/tree/master/2019-09-tcp-keepalives … https://blog.cloudflare.com/when-tcp-sockets-refuse-to-die/ …https://twitter.com/Cloudflare/status/1175075764443721728 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
jkbs proslijedio/la je Tweet
The demo worked flawlessly! Here is a video of the same thing: https://www.youtube.com/watch?v=4sRECNvnFFA … inet-tool: https://github.com/majek/inet-tool#readme … kernel patches: https://github.com/jsitnicki/linux/commits/bpf-inet-lookup …https://twitter.com/davem_dokebi/status/1171035546057154564 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Here is a foretaste of what programming the socket lookup can do. Map a listening socket to any address or port without re-bind()'ing, any way you want it https://asciinema.org/a/U4WyzNEbFSr2FMJouv05SUmfZ … Demo put together by
@majek04https://twitter.com/majek04/status/1170312588980346883 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
A shot at breaking up the 1:1 mapping between (address, port) pairs and sockets. Program your socket lookup any way you want with BPF inet_lookup. RFCv2 now ready to try https://lore.kernel.org/bpf/20190828072250.29828-1-jakub@cloudflare.com/ … Come see our talk at LPC '19 https://linuxplumbersconf.org/event/4/contributions/487/ …
@majek04Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
IETF Technology Deep Dive: How Network Interface Cards (NICs) Work Today session is now livehttps://www.youtube.com/watch?v=wHM7RVk3-yk …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
(1) Contention on listener socket was removed by Eric Dumazet back in 2016 https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=3b24d854cb35383c30642116e5992fd619bdc9bc … Florian Westphal explains how the same could be done for TPROXY https://lore.kernel.org/netdev/20190621125155.2sdw7pugepj3ityx@breakpoint.cc/ … (2) Early demux for TCP done by David Miller dates back to 2012 https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=41063e9dd11956f2d285e12e4342e1d232ba0ea2 …
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Two TPROXY sins: 1. grabs a reference to listener socket - contention under SYN flood, 2. skips early demux - additional lookups in the routing tables. What is TPROXY? A socket dispatch https://www.kernel.org/doc/Documentation/networking/tproxy.txt … Why is it useful? Listening on all portshttps://blog.cloudflare.com/how-we-built-spectrum/ …
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
What if you could steer packets to listening sockets with BPF?https://twitter.com/majek04/status/1141774561429590023 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
This weekend let
@majek04 take you on a tour through@Cloudflare's edge. We will take a stop at every point where BPF is already in use or might be in the future.https://twitter.com/Cloudflare/status/1129765309848670208 …Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Bookmark for reference! Comprehensive overview of tunnels in Linux by
@liuhangbin https://developers.redhat.com/blog/2019/03/27/an-introduction-to-linux-virtual-interfaces-tunnels/ … Caveat lector, ATM most commands in the blog need fixing. Get workings ones from the ever so useful iproute2 cheat sheet https://baturin.org/docs/iproute2/#Tunnel%20management …Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
What?! GCC & Clang support ranges in switch statements! As language extension, of course: https://gcc.gnu.org/onlinedocs/gcc/Case-Ranges.html …pic.twitter.com/DjuypK7ncS
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
How to check if peer TCP closed its end of connection w/o reading from it? Polling for RDHUP event with zero timeout is one way: struct pollfd pfd = { .fd = peer_fd, .events = POLLRDHUP }; n = poll(&pfd, 1, 0 /*timeout*/); if (n > 0) { /* connection closed by peer */ }
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.