Conversation

Notices

1. Embed this notice
   Pissed Hippo (sun@shitposter.world)'s status on Thursday, 08-Aug-2024 02:23:36 JST Pissed Hippo

   • Terry Hendrix II 🏹
   @thendrix I am also old enough to remember when embedded meant microcontroller not full os like linux lol
   • Embed this notice
     Pissed Hippo (sun@shitposter.world)'s status on Thursday, 08-Aug-2024 02:32:46 JST Pissed Hippo

     in reply to

     • Terry Hendrix II 🏹
     • Account: Computers
     @pro @thendrix I know but I wanted to flex my age. at university QNX released a free version of Photon and I ran it on my PC.
   • Embed this notice
     Account: Computers (pro@mu.zaitcev.nu)'s status on Thursday, 08-Aug-2024 02:32:48 JST Account: Computers

     in reply to

     • Terry Hendrix II 🏹
     @sun @thendrix QNX was released in 1980s, or 40 years ago, and provides POSIX.
   • Embed this notice
     翠星石 (suiseiseki@freesoftwareextremist.com)'s status on Thursday, 08-Aug-2024 02:37:24 JST 翠星石

     in reply to
     @sun That version of Photon was proprietary and I suspect under ridiculous terms that it could only be used for education.
   • Embed this notice
     Pissed Hippo (sun@shitposter.world)'s status on Thursday, 08-Aug-2024 02:38:50 JST Pissed Hippo

     in reply to

     • 翠星石
     @Suiseiseki yes it was only for educational or hobbyist or evaluation purposes
   • Embed this notice
     Pissed Hippo (sun@shitposter.world)'s status on Thursday, 08-Aug-2024 02:39:54 JST Pissed Hippo

     in reply to

     • Terry Hendrix II 🏹
     • Account: Computers
     @pro @thendrix Around that time Ubuntu had a fork for multimedia production that utilized the soft realtime kernel patches. I experimented with the Linux realtime kernel patches (usually getting rid of deadlocks) and for whatever reason they frequently caused breaks in userland applications.
   • Embed this notice
     Account: Computers (pro@mu.zaitcev.nu)'s status on Thursday, 08-Aug-2024 02:39:56 JST Account: Computers

     in reply to

     • Terry Hendrix II 🏹
     @sun @thendrix Red Hat looked at RT kernel for RHEL some 20 years ago. Tim Burke was supposed to lead it. But it gradually petered out as all such initiatives. They are working on "automotive RHEL" now and it's also not going far, I'm pretty sure. It's like Windows CE.
   • Embed this notice
     Account: Computers (pro@mu.zaitcev.nu)'s status on Thursday, 08-Aug-2024 02:51:24 JST Account: Computers

     in reply to

     • Terry Hendrix II 🏹
     @sun @thendrix One of the odd things many apps expect is for the child to run first out of fork().
     
     It always annoyed me from the standpoint of a mini-micro-unix emulation on something like 8080. Because of that nonsense you must have to have 2 processes runnable, because the parent will need to reach wait() eventually. It's not a problem on a normal computer, because you just COW everything anyway. Your fork() is actually clone() with parameters. Just a few more page faults. On the original PDP-11, fork() was a swap and un-swap, so the disk's DMA did the job of memcpy(), so that's the historic reason why.
     
     How apps sense which one runs first is by racing it with writing into signalling pipes. Which is just stupid and bugs, bugs all around.
     Pissed Hippo likes this.
   • Embed this notice
     翠星石 (suiseiseki@freesoftwareextremist.com)'s status on Thursday, 08-Aug-2024 03:01:33 JST 翠星石

     in reply to
     @sun Such terms violates freedom zero.
     
     Such licenses almost always are designed to violate all 4 freedoms; https://www.gnu.org/philosophy/free-sw.en.html#four-freedoms
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Thursday, 08-Aug-2024 04:16:01 JST feld

     in reply to

     • Terry Hendrix II 🏹
     • feld
     • Account: Computers
     @sun @pro @thendrix
     
     > It was back in February though that Intel acquired Linutronix and stated getting the PREEMPT_RT support upstreamed would still be among the goals for the Intel-owned firm. Thanks to the resources afforded by Intel, it looks like indeed that will soon be a reality.
     
     
     haha okay well maybe not since Intel is losing their asses
     Pissed Hippo likes this.
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Thursday, 08-Aug-2024 04:16:02 JST feld

     in reply to

     • Terry Hendrix II 🏹
     • Account: Computers
     @sun @pro @thendrix They've literally spent the last 20 years working on the realtime Linux patches and they're getting very close to being upstreamed. I believe the last problems they're dealing with have to do with kernel logging which is a blocking operation
     
     edit: found the remaining things
     
     > There are various patches around printk, a number of Intel i915 DRM patches, lazy preemption, various ARM and PowerPC changes, and also the actual bits for being able to enable the real-time kernel builds (setting ARCH_SUPPORTS_RT).
     
     printk is the kernel logging