Conversation

Krzysztof Kozlowski

ZTE or zte.com.cn apparently decided to flood kernel with extremely poor quality, automated patches. Reminds me previous flood from Vivo.com. Some of the patches were never built and clearly do not compile. Some are looking correct, but are wrong if you open the context of the file (e.g. using sysfs_emit in procfs or some other code). None of the review feedbacks were responded to, they just keep sending the code, sometimes 5 same patches without noticeable changes or without changelog.

And I bet somewhere within this flood of automation-generated stuff, is actual fix or actual bug pretending to be a fix but introducing a vulnerability.
2
2
5
And at least two patches are so obviously wrong and so obviously not compiled, that it is actually reassuring. Why? Because it cannot be a try of sneaking faulty patches into the kernel (see UMN.edu and their hypocrite commits). It's just to plain junk code, too obvious for real attack:

https://lore.kernel.org/all/20250313160959132rMOoPpNjSNTlB-E6rMOuj@zte.com.cn/
https://lore.kernel.org/all/20250313144430115dzzemB3yO4jZvI29ZZebY@zte.com.cn/

Although, exhaustion of resources is also sort of an attack...
1
0
4

@krzk IRL DDoS :> .

More seriously, maintaining is already so much work, having to deal with this kind of non sense is ... terrible for a lack of better word.

0
0
1

@krzk I've seen bizarre patches in the past that were 'obviously' wrong with little explanation and when you reply saying so getting little to no push back.

Sometimes you seriously wonder if there is some ongoing attempt to find weaknesses in review process.

There are other patches that are more subtly wrong, but then obviously you wonder if author just made mistakes.

But these days, post-Jia Tan, one can't help wonder...

0
0
1

@krzk @krzk tag yourself im "kconfig: replace strncpy with strscpy"

(obviously wrong because it's a userspace tool and strscpy is kernel api)

https://lore.kernel.org/all/20250317152257605EN-VIC5_IyrowJu481BsZ@zte.com.cn/

1
0
0
@cas Yup, that's another patch coming from some tools or coccinelle without even understanding the code.
1
0
1

@krzk @cas

> without even understanding the code

And even compiling it with the patches applied, it seems...

0
0
0