pmooney

Brian Cantrill is my hero: youtu.be/-zRN7XLCRhc?t=2079 :D

hmm?

stupid question: how different is illumos to Solaris Nevada?

Do you mean specifically a given release of opensolaris from the past, or what became Solaris 11.0?

does crucible supports shared the disk between two VM? example Oracle RAC needs shared disks between two nodes

no

terabit, it has diverged

do you that? thank you

crucible is definitely not designed with multiple-attach in mind

is there any plan to support?

not at this point

just from a user experience and "feel" point of view

okay, so, there is no shared disk solution right now?

I'm sure the illumos devs have done a lot of work in the past decade+

tozhu: correct

Thank you

roll your own with iscsi, potentially

i just miss Solaris , trying to decide if I should install illumos or solaris 11

idc about all the nome+desktop stuff much (although it was real nice with opensolaris)

when using Clang from pkgsrc I get an error like this `ld: fatal: file /opt/local/lib/clang/16/lib/sunos/libclang_rt.builtins-x86_64.a: open failed: No such file or directory`. I'm trying to understand if it's a missing file or what. fyi I'm porting a compiler to illumos.

antranigv: I don't know much about clang but yes, that looks like a missing file.

sommerfeld I'll install clang from OmniOS extras now and try to compare them.

other question, I read the information from internet, your oxide computer booting does not need bios/uefi, is it really, is there any details for the progress, it’s interesting

my guess is that the libclang_rt.builtins file is for clang's equivalent of gcc's libgcc - some functions the compiler's code generator depends on that aren't in libc.

More followup on Stephan Althaus's zones problem, consistent with the previous analysis: openindiana.org/pipermail/oi-dev/2024-January/033535.html and openindiana.org/pipermail/oi-dev/2024-January/033536.html; the zone that failed to boot was at the commit right before the fexecve change while the global zone was four commits after.

[illumos-gate] 16213 i86pc improperly handles 64-bit config space -- Robert Mustacchi <rm⊙fo>

'NOTICE: igc0 link up, 1000 Mbps, full duplex' -- basic of a link working are there. Need to write up all the I/O. But it's a start.

igc?

i think those are the newer I2x0 parts that are 2.5Gbit

Yea that's what it looks like, seems they've been around at least since 2019. Hopefully they are not as complicated to program as other Intel parts.

i'm guessing since it's just 2.5gb, they probably didn't decide to throw in an entire network switch into the NIC *cough i40e*cough*ixl*cough*, so that probably helps keep the complexity down

though I'm going to assume their TSO support is equally as brain dead (or I guess maybe 'cheap' is likely more accurate) as their other parts

It's just an igb/e1000g that has a 2.5G MAC/phy.

Though that MAC/PHY working is a different question.

There's a lot of erratum there.

ooh exciting, if you have something that needs testing I can help

wiedi: Right now I just have links and interrupts going. So once I get the rest of the surround I'll make sure to make more noise.

cool

But the packet rx/tx is the same as igb, so that shouldn't be too bad.

[illumos-gate] 16114 ksh should not be isaexec'd -- Richard Lowe <richlowe⊙rn>