On Tue, Mar 11, 2025 at 10:26:24AM -0700, Eric Biggers wrote: > > In the !HIGHMEM case (i.e., the common case) this is just worse, though. It > expands into more instructions than before, only to get the same linear address > that it did before. You also seem to be ignoring the comment that explains that > we're working in the linear buffer of the whole sg entry. Thanks for letting me know. I had assumed that this would all get optimised away. Let me look into this atrocious code generation. Cheers, -- Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt