Re: [PATCH v8] PCI: exynos: Handle PCIe PHY initialization errors
From: Sreeraj S Kurup
Date: Thu Nov 19 2026 - 09:23:37 EST
Hi Krzysztof,
Thank you for your feedback. I would like to clarify a few things regarding the patches
and the concerns you raised.
The code is not entirely AI-written, as you claim. The reason I sent multiple patches on
the same day is that I was working on them over time and submitting them as they were ready.
They were not all written and sent at once. There was a time gap between the submissions, so
please don't assume that I was blindly sending random patches just to get past the review process.
Regarding the v2 to v9 revisions, I admit that I blindly trusted your Shashiko reviews. I submitted
the patches on my holiday, assuming that Shashiko was properly trained on the latest Linux kernel
source code. I also assumed that, since the PCI subsystem uses Shashiko for patch reviews, its
feedback would be reliable enough to trust.
If Shashiko repeatedly reports issues that are incorrect or misleading, why is it being used as a
review tool in the first place? Isn't relying on such feedback equally problematic? I trusted its
suggestions because I assumed that the review system was reliable and that its feedback had been
properly validated.
Manivannan Sadhashivam also advised me not to blindly trust automated review tools. However, he
explained his concerns constructively and helped me understand where I was going wrong.
He understood that I had trusted Shashiko's feedback in good faith. I appreciate that kind of
review because it helps contributors learn and improve rather than simply questioning their
understanding or dismissing their work as AI-generated slop.
I'm not claiming that my patches are perfect, and I'm willing to take responsibility for genuine
mistakes. But dismissing the work as "100% AI-written slop" simply because I submitted multiple
revisions is unfair. Humans can also submit multiple patches in a single day. That alone does not
prove that the code is AI-generated or that the author doesn't understand it.
I understand your concerns about code quality, and I will take greater care to verify every review
comment instead of blindly trusting automated feedback in the future.
I respect your experience and feedback, and I am willing to learn from genuine technical criticism.
My intention has always been to contribute quality code to the Linux kernel and improve through the
review process. I hope you can understand my perspective as well.
Best regards,
Sreeraj