It was done by Apple and OSF, and for a while was the only way to run Linux on NuBus Macs.
787 karma · joined May 13, 2008
It was done by Apple and OSF, and for a while was the only way to run Linux on NuBus Macs.
So the solution is to setjmp before each call to an X function that might notice the connection is closed, and longjmp out of the error callback. Ugly, but it works.
These days, you should really just use XCB, or even Wayland.
It creates what looks like a virtual NIC (literally in qemu's case, indirectly by SLIP in the original Slirp), and reassembles the packets it gets coming in from the guest OS or SLIP user. For example, a SYN packet gets turned into a call to connect(), and a data packet gets turned into a write() on the appropriate TCP socket FD. An RST packet gets turned into a call to close(). The reverse happens in the other direction, based on the read() data, fake TCP packets are generated, a closed socket gets mapped to a RST packet, and so forth.
From the outside, it looks like the guest (or SLIP user) is NATed through the host's IP address. But really it is just reassembling the intention of the guest based on the packets it is sending and calling host kernel functions to cause the same effects.
It shouldn't be surprising that it is very oriented around VMs hosted and manged by public cloud providers. Bare metal is at best an afterthought, and at worst explicitly disadvantaged.
Of course, when ESXi shipped without a console OS, but with a vmklinux and GPL'd drivers, a new take was needed.
Sun had a similar app that would fax orders to Tony and Alba's pizza, though that closed a few years ago.
It was as simple as pumping water into the compressor stage, cooling the air after the compressor had heated it. This allowed more fuel to be burned without melting the turbine stage. I think the extra mass of the water (steam) may have helped as well.
This is where the terms "Dry Thrust" and "Wet Thrust" came from.
Also, keep in mind, in the timeframe that is relevant here, the only apps were Mail and Safari. Apple making Mail a native app from the start was another tell that "you should just build web apps" was disingenuous.
Apple knew that there would be huge pushback against a walled garden appstore with Apple in absolute control, and huge pushback on them taking a 30% cut. So they started with "just write web apps". But of course, they also inexplicably made the first iPhone 2G only! Flagship phones that shipped many months before the iPhone had 3G. Why cut that particular corner, and not any others in their otherwise $$$-is-no-object new wonderphone?
You can do a much better job hiding low bandwidth and high latency in a native app, especially since app assets get downloaded once (probably over wifi) at app install time. So by essentially making app developers beg for access to their native platform, they radically reduced the anger at their wildly locked-down appstore and their 30% cut.
And it seemed a little suspicious how quickly they were able to deliver a 3rd-party SDK and developer documentation.
You can do that today with Cumulus Linux, on Mellanox switches, no less. Switches from other manufacturers as well: https://cumulusnetworks.com/hcl
This has pictures: http://www.mindspring.com/~sportrocketry/ggsg/sequence.html
Last time I looked into this, legally speaking it is payload that matters, not guidance, at least until you try to export. Though there is a general bias in the hobby against doing things that attract negative attention, even if it is, strictly speaking, legal.
There have been other sun-seekers as well.
Wikipedia has an updated version as well: https://upload.wikimedia.org/wikipedia/commons/8/88/Itanium_...
It won't prevent correlation attacks, but it does make metadata attacks in general harder and less confident. An improvement is an improvement even if it doesn't completely solve a problem.
For all of initrd/initramfs's history, you've only needed one if your monolithic kernel couldn't mount the root filesystem. So if you needed a module, or had some special config that needed to happen that the kernel couldn't do internally.
to dual 100G attach their servers to a pair of ToR switches. As a result, they'd really love to have something faster than 100G for their ToR to spine uplinks, but that isn't quite available yet.
Insanely (and awesomely), Mellanox already has these: http://www.mellanox.com/page/products_dyn?product_family=266... dual 200G NIC.