[illumos-gate] 17096 CPUs mistakenly showing up with fully-associative caches -- Robert Mustacchi <rm⊙fo>

[illumos-gate] 17097 netstat: ignoring unreachable code -- Toomas Soome <tsoome⊙mc>

hm, 502 from code.illumos.org

I would like to burn down all AI crawlers

yes please

[illumos-gate] 17098 netstat: replace divide condition -- Toomas Soome <tsoome⊙mc>

6 CVEs in rsync

OmniOS had an update to rsync last week. I'm hoping that was to address the CVEs.

yes, they were announced last week, but distros had advance knowledge to prepare quick updates

ok, it just came up in my feed and i know rsync is used here so i mentioned it just in case

I don't sub to actual CVEs myself though

I'm just glad to hear the patches I applied applied.

the political troll here the other day was enough distraction :) evil black hats messing up our infrastructure would be much worse

hope "black hat" is still an acceptable term

if not i apologize but would like a better term suggested

cracker used to be a synonym for black hat but it's a racial slur against white people...

the renaming in git isn't a big deal but removing the same term from discussions about data busses and such leads to sometimes awkward constructions

initiator works in some contexts

anyway I don't use rsync in any of my personal workflow, glad to know it's addressed though

dangergrrl: fwiw black hat would stll be the term of art in security circles

still*. need to figure out what's going on with that switch :(

I try to use preferred terms :) most of them are not annoying

I'm playing with usb today, working on webcam support for dragonfly bsd. I'll try more on the X DRI/DRM stuff i'm doing for oi later.

illumos.org/docs/about/logo <--- this says the 'powered by' versions of the logo are not acceptable for use on a t-shirt

already contemplating what to wear to events.linuxfoundation.org/open-source-summit-north-america :)

looking at github.com/richlowe/arm64-gate

What about github.com/alhazred/illumos-arm ? Is this still relevant?

that looks pretty dead, last commit 4 years ago

richlowe's had a commit last week

I hadn't come across richlowe's arm64-gate. Thanks for sharing that

haven't even cloned a copy yet, looks like it might boot on a pi 4 which i have, currently booting retropie, might have to fight over snes access :)

np

Yeah, I suppose so. I saw the RISC-V stuff in there (link I sent) and it caught my eye

And DEC Alpha but not sure who is going to be running that

oh, a risc-v board with an open gpu is high on my (hypothetical) hardware purchase list

Images for rpi4 at downloads.omnios.org/media/braich

i'm still sad Compaq killed Alpha

ptribble, thanks. i might have to port everything retropie needs to use it much though

my spouse will get grumpy if i take her emulator away

I found a couple of good candidates but have zero actual budget so it might be a while

(rv64 boards with fully open source graphics drivers)

i'm a member of risc-v international but that's free (like the internet society) and i don't have any actual hardware as of yet

oh, yay, i have clearance to tie up the pi4 for a while

anyway, i still have to work on bsd first, i'm poor and they are offering a small bribe :)

What would it take to get those RISC-V bits and pieces from that link I shared running again?

Not sure how much four years changes how relevant they are (asking from the perspective of someone who really doesn't know)

Well you would want the upstream repo github.com/n-hys/illumos-gate

My guess is that it would work fine on the hardware it supports, you would be running a version of illumos that's a few years old

What level of effort might be expected in moving the RISC-V support from that version of illumos to the current version of illumos?

look at how long it's taken for aarch64, and figure at least that long

though we think we have learned stuff on the way that would actually really help, we aren't sure we're done learning it yet

i really like rv64 but most of the boards have gpu's that require blobs for linux which makes little sense to me, kinda defeats the purpose of choosing open hardware for the cpu imo

rv64 boards should ideally be open enough for potential gnu guix support

if you can't use an open bootloader and the GPU is closed, what's the point?

Not paying a license fee to ARM.

i own a pi and even a Xilinx Zedboard so I'm obviously not a strict open hardware person

And ultimately economics is what drives it.

I have a preference for hardware that I'm permitted full documentation for without NDA so I'm not tied to specific versions of linux etc

but obviously I'll buy things that don't fit that

Sure, I'd prefer that to. Just if you're asking why they're doing it well.

for me royalties to ARM aren't the biggest issue since I'm not manufacturing chips myself, I mainly like being able to modify my things, have my choice of operating systems, not just a small range of linux kernels

available rv64 products aren't really cheaper so the lack of royalties to ARM doesn't seem to get passed to me

rmustacc, I get that the seller can have a higher margin without paying ARM but they should keep in mind why the customer wants RISC-V in the first place is my point

I know how to program hardware, I want hardware I'm allowed to program. That's my purpose in potentially choosing RV, personally.

especially given that ARM hardware even with the royalties typically beats RV in price/performance on things that are available to me

but i do realize that people who have the skills to program the hardware themselves are actually a small minority