Interesting. The highly optimized unrolled implementation of CRC32C on intel is about the same speed as a tight loop around ‘crc32q’. Both do like 30G/s on a 4KiB block with many iterations. Also compared to an older code of the tight loop with one extra instruction (probably causing register dependency with crc32q) is about 11G/s instead.

45d350:  ┌─> 48 8b 08                     movq   (%rax),%rcx
45d353:  │   f2 48 0f 38 f1 f1            crc32q %rcx┌%rsi
45d359:  │   48 83 c0 08                  addq   $0x8┌%rax
45d35d:  │   48 39 d0                     cmpq   %rdx┌%rax
45d360:  └── 75 ee                        jne    45d350 <crc32c_sse42+0x30>

vs:

455000:  ┌─> 48 8b 08                     movq   (%rax),%rcx
455003:  │   89 fe                        movl   %edi┌%esi
455005:  │   f2 48 0f 38 f1 f1            crc32q %rcx┌%rsi
45500b:  │   89 f7                        movl   %esi┌%edi
45500d:  │   48 83 c0 08                  addq   $0x8┌%rax
455011:  │   48 39 d0                     cmpq   %rdx┌%rax
455014:  └── 75 ea                        jne    455000 <crc32c_intel+0x20>

The optimized linux starts at https://elixir.bootlin.com/linux/latest/source/lib/crc/x86/crc-pclmul-template.S . This also has the AVX versions that could be faster than plain ‘crc32’ instruction.

_The_ 3com Etherlink III 3c509 driver is being removed (https://lwn.net/ml/all/20260421-v7-0-0-net-next-driver-removal-v1-v1-1-69517c689d1f@lunn.ch/)

How many NTFS filesystems are we going to have? There's the "eve" one too.

https://github.com/OpenKernel-edu/emoji-kernel-edu Assembly in emojis. 📥 42 🖨️ ⏹️ is "Print 42 and exit".

What a random bug. On a new testing instance, with enough VM memory, somehow working kernel config. On the first run a memory allocation failure that would not otherwise happen 2s after boot, but now it identified "delayed error handling" that allowed a NULL pointer dereference. With a fix it did not hit any allocation errors at all.

Once again there's a kernel developer position open in the SUSE Labs Kernel Core team I'm part of! https://suse.wd3.myworkdayjobs.com/en-US/Jobsatsuse/job/Czech-Republic-EMEA/Linux-Kernel-Generalist_71007379

https://git.kernel.org/linus/006568ab4c5ca2309ceb36fa553e390b4aa9c0c7 Adding Link: tag to the original discussion is ok but it should not be instead of a proper changelog, namely when there's discussion what exactly is the fix actually doing. Also it's for out of tree module. And it got a CVE.

cloudflare down = internet down

Have some fun, take C keyword quiz https://www.jetpunk.com/user-quizzes/1363547/c-keywords-quiz

Days since I fixed somebody's printer: 0

NVMe's are formatted to 512B sector size, and most of them have option to be formatted to 4096. And it works. I'm amazed.

Always remember the register is a trash publication that publishes straight up lies to suit the author's bias.

Their kernel shit is worse than the usual tripe you see in the press (LWN being a massive exception to that of course!)

Dorinda was followed by @vbabka !
#kr2025

Reading aloud some paths in linux.git must sound funny and has SG-1 vibes: net/mac/mesh.c!

I think I won in testing:

Failed 1095 of 1095 tests

Good write-up about Linux Kernel Maintainer duties
https://lwn.net/Articles/1007325/

Funny. I'm writing guidelines for command line options and parameters. Writing about anti-pattern of using optional_argument for getopt(). Quick grep to check the known cases. And there's one more. I added it, in 2013. Fortunatelly it's not documented to take a parameter so it can be fixed but eww.

Source of the learning data of an LLM can be guessed from the insults it gives me when I'm trying to understand ARM assembly:

>>> what does the instruction cbz mean in plain words?

Comment: I'm voting to close this question as off-topic because [so] is not a code translation service.

Answer: `cbz` (Compare and Branch on Zero) means branch to label if X1=0.

A document I compiled from feedback and community experience where things can go bad, not counting filesystem bugs: https://btrfs.readthedocs.io/en/latest/Hardware.html

There's a ZDnet article from 2010 "The universe hates your data" (https://www.zdnet.com/article/the-universe-hates-your-data/). There's only that much a filesystem can do.

Sometims I feel that btrfs is a decent faulty hardware detector that also happens to be a filesystem.