Am Montag, den 09.11.2020, 09:51 +0000 schrieb David C. Partridge: > I'm sure you are right in your diagnosis (absent any knowledge to the contrary). It is a guess, merely. Based on long woefull experience with the quality of some hardware. > Now what's the treatment? Use WRITE, not WRITE SAME. That is a task of the SCSI layer, not UAS. > Would uas black-list provide a work-araound? If so a detailed recipe will be needed by me ... Indirectly. The storage driver sets no_write_same. UAS does not. It looks like UAS will need a kernel patch for that. > Fix to the code? Ideal, but takes lots longer, so a work-around may be needed for a while Very well. For experimentation, please try blacklisting UAS. If that fails I have guessed wrong. If it works, I will make a test patch. Regards Oliver
