Hi Jean, On Tue, 2024-02-13 at 12:05 +0100, Jean Delvare wrote: > Alexander, back when you contributed the code, you said the new command > was long-time tested, so it's hard to believe it includes the bug > reported by Piotr. Do you remember which Intel chipset you were using? > Is the code you submitted exactly what you were using on the hardware, > or is there a chance that you forgot one change when preparing the > upstream submission? from what I can tell, back than it was "Intel® Xeon® D-1500 product family Integrated PCH Logic". I'm not sure if it was WILDCATPOINT_SMBUS or LYNXPOINT_SMBUS, it's nevertheless FEATURES_ICH5, means FEATURE_BLOCK_BUFFER. Unfortunately, I cannot tell you the details of the tests, but the reason is that I had to rely on a colleague for this, so I cannot explain how this went unnoticed back than. Maybe the (incorrect) returned data was irrelevant for his tests, I can only speculate. I don't think there was a non-published downstream patch in play. -- Alexander Sverdlin.
