[Bug gas/22871] Encode instructions of 64-bit operand without the REX_W bit

[address@hidden]

https://sourceware.org/bugzilla/show_bug.cgi?id=22871

--- Comment #18 from Linus Torvalds <address@hidden> ---
(In reply to H.J. Lu from comment #17)
> A testb microbenchmark

Very interesting. I can confirm that testb seems slower on Skylake too.

And it's not some odd effect of just the call/ret sequence - I did a ".rept
100; .endr" around the test instructions to see, and it shows the same pattern.

Finally, it's not even about %dil and the REX prefix from your choice of
register, it's true even for the one-byte version of "testb %al" (well, two
bytes with the immediate, of course).

-- 
You are receiving this mail because:
You are on the CC list for the bug.