HNHacker News
TopNewBestAskShowJobs

kijiki

787 karma · joined May 13, 2008

submissionscomments
kijiki··on MIPS provides highly scalable RISC processor IP
> Branch delay slots aren't a big deal.

Fair enough, but if the claim to fame over the current dominant ISA is that there isn't weird baggage (big deal or not), there is an opportunity for another new ISA without that baggage (important or not)...

kijiki··on Ubiquiti is suing Brian Krebs for his reporting on their breach
A lot of ubiquiti wifi gear can run OpenWRT. Would at least insulate you from their cloud insecurity.
kijiki··on Universal Chiplet Interconnect Express UCIe 1.0 Launched
https://ark.intel.com/content/www/us/en/ark/products/130409/...

Intel has already shipped a MCM with an Intel CPU and a AMD GPU, back in 2018.

That probably just used PCIe, since cache coherence wasn't needed.

kijiki··on United flies passenger flight on sustainable aviation fuel supplying one engine
True for pure hybrid Priuses, but Plug-in Priuses use Lithium-Ion batteries just like Teslas.
kijiki··on Improving first impressions on Signal
The issue with these sorts of HashCash schemes is that legitimate users need to PoW their messages on a battery powered phone with a basic CPU and GPU. Spammers can use big, mains-powered GPUs, or better yet, botnets full of stolen CPU-time from mains-powered desktops/servers/etc.
kijiki··on Improving first impressions on Signal
Unless your bots are doing things that get their messages reported by recipients, this shouldn't affect you at all.
kijiki··on Improving first impressions on Signal
No one can give you a 100% answer, but I can relay some personal anecdotes which might help.

I have consulted for the Signal Foundation in the past. None of the employees or consultants I worked with indicated that they'd had any problems.

Around 2010 timeframe, I hosted the RedPhone server backend on some spare colo hardware I owned. RedPhone, along with TextSecure, were the predecessors to Signal Messenger.

I have never had any (signal related) issues with the government.

kijiki··on Raspberry Pi 3 Fastboot – Less Than 2 Seconds
Caches aren't a problem, you can just persist memory. The problem is the non-RAM internal state of everything. You can preserve CPU register state and various operating modes, but persisting the state of every hardware device you use is a big, undocumented challenge.

Virtualization systems like qemu-system or VMware support suspend/resume, which is essentially the same problem.

kijiki··on The Ties that un-Bind: Decoupling IP from web services and sockets [pdf]
Your switches run a routing protocol (almost always BPG) and your load balancers all announce the same IP to them. This results in a routing table with many next hops for the LB IP, and traffic is distributed across them using ECMP, typically with some resilient hashing scheme so that a given flow will always go to the same physical LB, even if there are topology changes due to LB or switch failures.

At the edges, your switches or border routers announce the IP to all your peers and transit ISPs.

kijiki··on Stop using Zoom, Hamburg’s data protection agency warns state government
I haven't tried it lately, but last time I tried it around a year ago, if you try to use the web client and the moderator hasn't shown up yet, you have to solve a Google ReCaptcha every couple of minutes while you're waiting, or it kicks you out.

I usually try to go with the charitable interpretation, but when the dark patterns start to stack up that high...

kijiki··on Dumping and extracting the SpaceX Starlink user terminal firmware
I haven't been to snow ranch in years, but early Spacex investor Steve Jurvetson was a regular at the high power amateur rocket launches there. Probably other Spacex folks there as well.

Launches have all been shutdown for COVID reasons, and even before that, not infrequently cancelled due to fire risk.

kijiki··on The slow rise of robots in the data center
Plexxi tried something along those lines. It was cool, but very expensive, so it didn't really go anywhere.
kijiki··on Arm Introduces Its Confidential Compute Architecture
A legit concern. Basically everything that connects directly to the cache controllers has to get CCA right. Coherent GPUs are the big, complicated one.

Intel had an SGX bug where the GPU could access the last 16 bytes of every 64 byte cache line. If you need the GPU enabled, you have no choice but to design your enclave to only use the first 48 bytes of each cache line. Fortunately if you don't need the GPU, whether or not the GPU is enabled (among other things) is measured and included in the attestation quote, so clients can simply choose to refuse to talk to servers with the GPU enabled...

kijiki··on Arm Introduces Its Confidential Compute Architecture
Intel and AMD do it in the memory controller. The caches, by virtue of being entirely on-die, are assumed to be inaccessible.

It is unclear what AMD intends to do here for their TSV stacked mega-cache thing. Perhaps they'll declare the TSVs as not practically snoopable, similar to how on-die metal layer wires are treated now...

kijiki··on Prossimo – Memory safety for the Internet's most critical infrastructure
Big Internet BGP servers also aren't super exposed due to the common use of TCP MD5 (RFC 2385) or TCP AO (RFC 5925). The userspace daemon won't even see packets from attackers who don't have the shared secret for that peer link.
kijiki··on The number of cities with municipal broadband has jumped over 4x in two years
Sonic sells gigabit for $40/month if you're lucky enough to live somewhere they're allowed to get pole access.
kijiki··on Arm Announces Neoverse V1, N2 Platforms and CPUs, CMN-700 Mesh
amd64 is more RISC now than ia32 was in 2001 as well.
kijiki··on How many x86 instructions are there? (2016)
Tried it out on Debian Bullseye x86_64:

  $ objdump -w -j .text --no-show-raw-insn -d /usr/bin/emacs-gtk | egrep '^ *[0-9]+:' | awk '{print $2}' | sort | uniq | wc -l
  130

  $ objdump -w -j .text --no-show-raw-insn -d /bin/ls | egrep '^ *[0-9]+:' | awk '{print $2}' | sort | uniq | wc -l
  97

  $ objdump -w -j .text --no-show-raw-insn -d firefox-bin | egrep '^ *[0-9]+:' | awk '{print $2}' | sort | uniq | wc -l
  136
kijiki··on Signal Server code on GitHub is up to date again
TLS is used, but there is another layer of encryption e2e from the client to inside the enclave. Your MITM server B can decrypt the TLS layer, but still can't see the actual traffic.
kijiki··on „We were chased out of SF after trying to give free Internet to the poorest“
I used a burner gmail account to try it out. It was slow enough the idea of creating new burner gmail accounts every time I wanted to use it was uninteresting.

IIRC, it was below 128kbit/s for me, though there was at least one mesh hop away from a node that had a directional antenna pointed towards 444 Castro. Perhaps it was faster if your local node happened to have a direct uplink to either 444 or the 'plex.

kijiki··on „We were chased out of SF after trying to give free Internet to the poorest“
I don't know anything about the SF program, but I lived in Mountain View at the time (still do). They put one of the mesh nodes on a street lamp about 200 feet away from my house. I played around with using it as a backup for when the DSL was down, but I gave up because it required you to log in with a google account.
kijiki··on BPF for storage: an exokernel-inspired approach
Everything old is new again.

https://en.wikipedia.org/wiki/Channel_I/O

kijiki··on Mars becomes the 2nd planet that has more computers running Linux than Windows
> Looks like Wind River discontinued RTLinux

Xenomai is still under somewhat active development.

kijiki··on Spy pixels in emails 'have become endemic'
Inline images. Faster loading, and will still work when I look at the email years later when all the img links have rotted.
kijiki··on Spy pixels in emails 'have become endemic'
gmail presumably doesn't need to use a tracking pixel to determine if someone opened an email in the gmail interface...
kijiki··on Help users in Iran reconnect to Signal
These blocks tend to be reactive, so if a blocked app starts using Tor, Iran will block Tor fairly quickly. So in return for a short-lived regain of the use of Signal, all of Tor gets blocked.

This proxy arrangement is better because folks who start them tell their friends in Iran, who tell their friends, but it isn't listed publicly, like most Tor entry nodes are. When the authorities find a proxy and block it, they only disconnect a subset of Signal users, who hopefully have other proxies they've learned from other friends or friends-of-friends. So now the blocking is trying to put out a thousand small fires that they have to find one-by-one.

<I have consulted for the Signal Foundation in the past, but not recently and haven't talked with anyone there about this>

kijiki··on IBM paperweight teardown: Reverse-engineering 1970s memory chips
https://assets.moma.org/documents/moma_catalogue_2101_300062...

Unsurprisingly it was funded by Intel.

kijiki··on Japanese companies fight for share of EUV chip technology sector
> silicon valley is thousands of miles away from all of them.

Intel does 10nm production at their fab complex in Chandler, AZ. About 600 miles away from Silicon Valley.

kijiki··on Intel outsources Core i3 to TSMC's 5nm process
The govt did it back when it was Japanese DRAM manufactures that were outcompeting the US: https://en.wikipedia.org/wiki/SEMATECH
kijiki··on AMD Patent Reveals Hybrid CPU-FPGA Design That Could Be Enabled by Xilinx Tech
Really? I used a HT FPGA back then, but it came on an HTX card that fit into the HTX slot on the motherboard.

Never heard of putting it into the socket, would be a real pain to attach JTAG to program/debug your design...

← PreviousPage 4 of 12Next →