Skip to main content


 
Do you still care about floppy drives? Then it's you chance to get involved in #Linux #kernel development, as the floppy driver just got orphaned: git.kernel.org/torvalds/c/be2…

Bu be warned: Right before it was orphaned four bugs that syzkaller had found were fixed.



 





 
A small, but in a way most important change of the #realtime (RT) patchset for #Linux was submitted for #kernel 5.3:

Tglx asked Linus to merge a patch introducing CONFIG_PREEMPT_RT that, if enabled, will make the resulting kernel fully preemptible: lore.kernel.org/lkml/201907151…

If that will finally make it possible for realtime and Nvidia driver coexist, that would be awesome. I, for one, use my machine for sound AND video (and games), so that's been a pain point for me.



 



 



 
#Linux mainline #kernel aka pre-5.3 now supports time travel ;-)

"[…] useful to run with "time travel" inside the #UML instance, for example […] some tests […] wait for 120 seconds […] obviously goes faster if it need not actually wait that long[…]"

git.kernel.org/torvalds/c/065…

Heh, somebody found my changes there ...

hehe, yeah, it was changed a Kconfig file and was big enough to even get noticed during my first brief scan; the title did the rest ;-)

Side note: seems the syzkaller devs will take a closer look at it: https://twitter.com/dvyukov/status/1150780506721705985

Ah, interesting. We'll see what comes out of this. FWIW, this stuff isn't exactly unknown in the literature, but naming for it is confusing etc. and "time travel mode" was suggested during upstream review ... Also, I believe it's the first upstream implementation, others talk about doing it on QEMU but it seemed very tricky to get right to me, so UML was the better solution.

Also, support for it in wpa_s:
https://w1.fi/cgit/hostap/commit/src?id=1cefd89b37be43cf98f9f6291671b32b01ef5c93
https://w1.fi/cgit/hostap/commit/src?id=8528ad0c604cb15239a0dc9f4140c2208448e337
https://w1.fi/cgit/hostap/commit/src?id=d55035c9351b730a7f44fe2c1c6525686cd57e05



 



 
Today #Linux #Kernel 5.2.1 got released.

IOW: Yet again prepare for headlines like the ones I outlined when 5.1.1 arrived (see below).

Last time it actually took about two days until I saw the first story with one of the predicted headlines… #sigh

Must be from the MS-DOS age, when people thought "4.01" meant "More bugfixes than 4.0", when in reality it meant that MS had done the localization work. Now 6.22 vs. 6.0 was an entirely different matter.



 
Wolfgang Denk writes: "[…] this is a somewhat desperate call for help... We have been trying for some time to help some of our customers to
get patches into #glibc to fix the remaining #Y2038 issues. But progress is way too slow... […] In short, we are stuck, and not even […]"

Arnd Bergmann reshared this.



 
"[…] System firmware provides ACPI SLIT tables […] that describe the relative differences in access latency between #NUMA nodes. It’s basically a matrix that #Linux reads on boot to build a map of NUMA memory latencies, and you can view it multiple ways […]"



 
AMD released BIOS updates to partners that in the next few days are supposed to fix the #RDRAND/#RDSEED issue in the new #Ryzen 3000 CPUs that prevents many Linux distributions from booting: heise.de/newsticker/mel… (sorry, German; but it quotes AMDs original English statement)





 



 

Alexander Nolting reshared this.




 
#eBPF

Alexander Nolting reshared this.