Fun thing - SGI specifically used 256 color mode a lot, to reduce memory usage even if you used 24bit outputs. So long as you used defaults of their Motif fork, everything you didn't specifically request to use more colors would use 256 color visuals which then were composited in hardware.
This was the main driver of VGA memory size for a time - if you spent money on 2MB card instead of a 1MB, you could have higher resolution or bit depth.
if you had a big enough framebuffer in your display adapter, though, X11 could display more than your main ram could support - the design, when using "classic way", allowed X server to draw directly on framebuffer memory (just like GDI did)
In my experience[1], the fsck for given filesystem will happily replicate the errors, sometimes in random ways, because often it cannot figure which road to take in face of inconsistency. If anything, OpenZFS built upon that by now documenting the previously deeply hidden option to "rewind" ZFS uberblock if the breakage is recent enough.
[1] I've seen combination of ubuntu bug in packaging (of grub, of all things) and e2fsck nearly wipe a small company from existence, because e2fsck ended up trusting the data it got from superblock when it was not consistent.
> If anything, OpenZFS built upon that by now documenting the previously deeply hidden option to "rewind" ZFS uberblock if the breakage is recent enough.
One of the most "wizardry" moments in my career I've personally witnessed was a deep-level ZFS expert (core OpenZFS developer) we had on retainer come in during a sev0 emergency and rapidly diagnose/rollback a very broken ZFS filesystem to a previous version from a few hours before the incident happened.
This was entirely user error (an admin connected redundant ZFS "heads" to the same JBOD in an incorrect manner so both thought they were primary and both wrote to the disks) that we caught more or less immediately so the damage was somewhat limited. At the time we thought we were screwed and would have to restore from the previous days backup with a multi-day (at best) time to repair.
This was on illumos a few years after the Solaris fork, so I don't think this feature was documented at the time. It certainly was a surprise to me, even though I knew that "in theory" such capability existed. The CLI incantations though were pure wizardry level stuff, especially watching it in real time with terminal sharing with someone who very much knew what they were doing.
A very noticeable feature is that USB4 can tunnel USB3, which means it works like an USB hub, instead of an external PCIe USB controller (like in Thunderbolt). USB2 is still just physically separately transported over the D+/D- pins.
USB4 actually provides both USB 1/2 and 3 tunnelling, but it's incorrect to say it behaves like a hub because it involves needing an appropriate endpoint on the other end. Effectively a virtual cable, iirc, though there are at least two different mechanisms.
It's commonly trotted out, but the people who spearheaded the disastrous changes including mass outsourcing were Boeing for life - with McD people writing alarming memos about outsourcing goals set for Dreamliner
Main issue with opening it further is lack of DMU-level userland API, especially given how syscall heavy it could get (and iouring might be locked out due to politics)
The secret is that ZFS actually implements an object storage layer on top of block devices and only then implements ZVOL and ZPL (ZFS POSIX filesystem) on top of that.
A "zfs send" is essentially a serialized stream of objects sorted by dependency (objects later in stream will refer to objects earlier in stream, but not the other way around).
My understanding is that issues in scaling 68k line were already well known by then, same as with VAX (even if crucial people at Digital didn't want to believe).
The difference is that 68k was ubiquitous, reasonably cheap 32bit capable platform with MMU that had huge availability of parts and made porting software easy. Sun was working with 68k partially because they chose it in 1980, a year after it was made available, and by 1986 they published SPARC ISA and shipped first systems a year later
reply