Notices by LisPi (lispi314@udongein.xyz)

#Gentoo wiki is wrong about btrfs & VMs and it annoys me.

No, one does not disable CoW on btrfs to improve performance. That nukes not only the CoW properly but also the integrity. The durability of data is also entirely ruined by doing that, btrfs' model of atomicity is entirely built on copy-on-write.

(swap is effectively the only-use case that makes sense for that file attribute.)

What one does is adjust commit frequency, dirty page cache ratios and in-VM workloads that unnecessarily request synchronous writes (get rid of systemd-journald persistent/on-disk IO, or the entire program, among others).

(One could just entirely disable the VMs' ability to do synchronous writes, but then they may truly need them sometimes. So it's best to actually understand one's workload properly.)

If one wants to use a faster filesystem that doesn't do integrity checking, don't mangle btrfs, configure it properly or just use xfs instead.

@gantua @dalias @jschauma @mirabilos > If you have a stupid signal handler that sets a global variable its_time_to_exit = 1 which is just read elsewhere as a volatile, then once you add threads, suddenly you have an memory ordering issue

That's very much not the way I would do it or inter-thread communication, but if you want to do it that way then pthread_key_create, pthread_setspecific and pthread_getspecific are probably the answer to that problem.

@dalias @navi @jschauma @gantua @mirabilos After using a few languages with working green threads, I do hate how OS threads are expensive and one can't just use the naive option with them without all sorts of problems.

@gantua @dalias @jschauma @mirabilos I do think that portable libraries should be preferred for addressing that issue to instead using Linux/platform-specific options directly (building on standards, rather than disregarding them).

Given the Linux kernel project produces libraries to make their custom APIs less painful to use, people are probably already using a library anyway.

@dalias @Robotistry @datenwolf @jpoesen I've been told in other threads that some countries don't have that, but here it's an extremely common malicious practice. Effectively normalized.

@icedquinn @SuperDicq @enigmatico @mstar @mars Dragons are tasty according to Siegfried.

In any case, if direct connections to other peers are a problem, low-latency mixnets are indeed no longer an option. Delay-tolerant asynchronous networking, including considerable amounts of sneakernet-like methods, becomes necessary.

That requires a considerably harsher authoritarian stance and outside of North Korea no one really does it fully. (China does not investigate all connections nor all foreign connections.)