HNHacker News
TopNewBestAskShowJobs

kroeckx

170 karma · joined January 28, 2016

submissionscomments
kroeckx··on General Resolution: Statement Regarding Richard Stallman's FSF Readmission
Not as far as I know. If an option X beats an option Y, and Y beats option Z, and Z beats option X, you have a cycle. This is the case in the GR for option 1, 3, and 4.

If there are no other options beating X, Y or Z, they all 3 end up in the Schwartz set. We then have to determine which is the weakest defeat in the Schwartz set and remove it, and then determine the Schwartz set again.

Edit: Clearly I was confused, and changing it creates additional cycles.

kroeckx··on Debian votes to make no statement regarding RMS readmission to the FSF board
You can look at the Majority results if you want to say how many people think an option is acceptable. That ratio is also in the graph. For option 1, 203 people indicated that it's acceptable, 186 said it was unacceptable, leaving 31 that didn't say anything about it. The 203/186=1.09 is in the graph. You could also see that as 52% find that option acceptable.
kroeckx··on General Resolution: Statement Regarding Richard Stallman's FSF Readmission
That is not enough to create a cycle. Option 7 would still beat option 4 via option 3 with 8 and via option 2 with 4 votes. Note that there is already a cycle between option 1, 3, and 4. You're right that it doesn't take that many votes to create more cycles and complicate the process.
kroeckx··on Debian votes to make no statement regarding RMS readmission to the FSF board
The code does not draw lines between the options that didn't pass the quorum or majority test (drawn in pink), because they are dropped before the Schulze method is used. But it then adds the options again, to show all the options.

When multiple options are in the Schwartz set, the weakest defeats are removed until only 1 option is left in the Schwartz set. I think it only draws the graph after there is only 1 option left. But that didn't happen in this vote.

kroeckx··on OpenSSL will release a high-severity issue fix on 25th
We've had code in OpenSSL written in C that was constant time until a newer version of both gcc and clang recognized (mask & a | ~mask & b) as a conditional assign and could do things like turn it into a branch.

Edit: It can turn "b = foo(); c = (mask & a | ~mask & b)" into "if (mask) c = a; else c = foo();"

kroeckx··on OpenSSL will release a high-severity issue fix on 25th
As one of the OpenSSL maintainers, I would like to move away from C and probably to rust. But the problem is that C is the most portable language, and rust's platform support just isn't close enough.
kroeckx··on Ask HN: What is the most complex concept you understand?
I find it easiest to explain it in terms of electricity, which I assume most people here are familiar with. It acts just the same. Pressure, like voltage, is a difference. The flow rate is like the current, how much goes though. A pipe has a resistance that depends on it's diameter and length. You can calculate the resistance of the pipe as the pressure difference over the pipe divided by the flow rate.

One thing people might get confused by is why making the opening at the end of the pipe smaller makes the water go faster. What you do in electricity terms is add a series resistor, the total flow rate in the pipe will be lower, but the speed can be higher because it needs to go through smaller opening and so is more concentrated.

kroeckx··on Preserving data integrity: A ZFS-inspired storage system
It seems that a HGST/WD Ultastar, which is their data center drive line, has ERC disabled by default. I've now set it to the suggested 7 seconds.
kroeckx··on Preserving data integrity: A ZFS-inspired storage system
This not provide the same integrity as ZFS. Dm-integrity only protects against corruption on the disk itself, while ZFS attempts to provide protection against all sources of corruption, including software/firmware bugs, RAM and I/O path.

ZFS is using a Merkle tree and so knows which checksum to expect before reading the data. With raidz/mirror, when it detect corruption, it will attempt to read from the other drive(s) and combines those disks that give the correct checksum. If it can't find a combination that works, just what depends on that block is not available. If it can recover the block it will attempt to repair the disk(s) with corruption by writing the correct data to it again.

Standard mdraid can not repair an array in case of corruption since it doesn't know which of the data it has is correct or not, you need to figure out which of the disks has corrupt data on it, and manually remove the drive from the array and hope none of the other disks has corruption. In theory in case of a read error mdraid could attempt to fix it, but it doesn't and just removes the drive from the array. Dm-integrity provides more guarantees that the data you get is correct, but since mdraid and dm-integrity are different layers, they don't actually work together. There is no attempt to repair in case of corruption, dm-integrity will just return an read error and mdraid will remove the drive from the array.

kroeckx··on Netflix is now doing per-shot encoding for UHD content
As far as I know, with netflix the quality depends on things like the platform you're watching. Even with the highest subscription plan, if you watch it in Firefox or Chrome you're limited to 720p and you now get about 500 kbit/s, which looks awful. The older content seems to be at 1 or 2 mbit/s, and the 2 mbit/s looks good. But as you can see in the figures, they no longer seem to offer good quality 720p.
kroeckx··on Mozilla Lifeboat
That's not my experience at all. It got better when I left some large rooms. But I still regularly get things like messages that the server is offline, taking a minute before the UI shows that my message was sent, after my laptop was suspended, it takes minutes before element becomes usable again. I run my own IRC bridge, and if they send more than 5 messages a minute I start losing messages and they get out of order.
kroeckx··on Global air pollution maps by the European Space Agency
The map at https://maps.s5p-pal.com/ allows you to change the date. Comparing the 30-12-2019 - 13-01-2020 data to the latest makes a huge difference in certain areas.
kroeckx··on The radix 2^51 trick (2017)
Because in crypto, the math usually isn't mod 2^256. For instance in curve25519 you do do math mod 2^255-19, so they actually all fit in 51 bit.
kroeckx··on Dangerous SHA-1 crypto function is about to die in SSH
SHA1 is 160 bits and should have a complexity of 2^80 for collision resistance. The last paper puts a chosen-prefix collision at 2^63.4.
kroeckx··on OpenSSL 3.0
The plan at least is that at little as possible should break. If you do find a problem, please file an issue.
kroeckx··on OpenSSL 3.0
It's mostly about deprecating old APIs where a newer more general API is available, sometimes for more than 10 years.
kroeckx··on Covid-19 Open Research Dataset
I think https://www.coronawiki.org/ uses that data to visualize it.
kroeckx··on A Rust-based TLS library outperformed OpenSSL in almost every category
A lot of time is spend on the assembler versions doing the crypto, which is why they get used by various other projects like ring, boringssl, and the Linux kernel.

Less or no time is spend on improving the performance of the SSL library. From the article, it seems that the SSL library might have easy fixes.

kroeckx··on OpenSSH taking minutes to become available, booting takes half an hour (2018)
systemd didn't change much related to this. The seed file never got credited, at least not in Debian. So either you have the same problem on Gentoo, or Gentoo has other problems.
kroeckx··on OpenSSH taking minutes to become available, booting takes half an hour (2018)
The problem with /dev/urandom is that it's always available, even before any entropy was added. Fixing /dev/urandom to block until enough entropy is available once causes breakage like in the article. That is one of the reasons why getrandom() was added, so that existing applications can keep using the old and broken behavior but applications that care about the RNG being properly initialized can wait for it. So software that does crypto switched to it, and then things broke anyway.
kroeckx··on OpenSSH taking minutes to become available, booting takes half an hour (2018)
Why do you trust Intel less then any other hardware RNG you might have? Do you trust the one in the TPM? Or do you just not trust hardware to collect entropy?
kroeckx··on The X hole
At least Theo has a reputation for breaking embargoes. If he wants to get notified, he should work on changing that reputation.

[edit: It's probably more accurate to say that Theo doesn't want to agree to an embargo.]

kroeckx··on OpenSSL 1.1.1 Is Released
0-RTT / early data is supported. It is mentioned in the blog.

The wiki is really a document about issues you might run into when upgrading, and 0-RTT is not such a problem. To enable 0-RTT you need to use new functions.

kroeckx··on Why I Ripped the Same CD 300 Times
A CDDA has 2352 bytes of main data and 96 byte of sub-channel data per sector, so a total of 2448 bytes per sector. A sector is made up of 98 sub-frames. Each such sub-frame has 24 bytes of main data, 4 bytes of C1 error correction and 4 bytes of C2 error correction, and 1 "byte" of sub-channel data. The first 2 sub-frames have a special pattern in the sub channel data to recognize the start of the sector, leaving 96 bytes of sub-channel data available for other things. The error correction only covers the main data, the sub channel data just has a CRC.

The standard API to talk to the CD-ROM does not have a way to get the C1 and C2 error correction data. There is only a way to get 1 bit for each of the 2352 main data bytes after the C2 error correction to indicate that there was an error with it or not. But getting that data doesn't actually seem to be useful on the drives I tested it with, they all seem to have weird firmwares.

kroeckx··on STARTTLS Everywhere
There are multiple ways to set it up: - Reuse the key, set up the TLSA record using the public key. Certbot really doesn't seem to support this in an automated way. It requires that you generate a CSR, but there doesn't seem to be a way to renew this automatically. - Generate a new key for each certificate, which is what certbot. You would first need to add both the old and the new certificate to DNS, wait for the TTL of the old to expire, change to the new certificate, remove the old from DNS. There doesn't seem to be a tool to do this. - Set up the TLSA record to say you trust the CA.
kroeckx··on STARTTLS Everywhere
I'm still using a self-signed certificate because I'm using DANE / DNSSEC. I don't know of any tool that can properly help me automate the rollover of a let's encrypt certificate.
kroeckx··on Two-factor authentication is a mess
There are other companies like Google that first require you to set it up over SMS before they allow you to add a security token.
kroeckx··on TLS verification vulnerability in LibreSSL 2.5.1-2.5.3
It's not about client certificates, but it can be a reason why some software might want to do this. The return value of 1 allows the program to deal with verification errors instead of aborting the connection. You can for instance use this to display a proper error message that authentication failed, instead of just closing the connection and getting a generic TLS error message.
kroeckx··on TLS verification vulnerability in LibreSSL 2.5.1-2.5.3
I think there is a misunderstand of what the existing safety net is about. There are 2 error states: did the verification fail and should the connection be aborted. The safety net makes sure that if a function (the callback) says the connection must be aborted but didn't set the verification error, that it sets an unknown verification error. Note that the callback is external code.

The new "safety net" that libressl added said that if the connection doesn't need to be aborted there was no verification error.

kroeckx··on OpenSSL Key Recovery Attack on DH small subgroups (CVE-2016-0701)
I tried to be clear in when you're vulnerable and when not, so that you can decide if you're vulnerable or not. The answer in this case is "it depends, very likely not".
← PreviousPage 2 of 2