On Wed, Dec 20, 2023 at 6:58 AM Daniel Borkmann <daniel@xxxxxxxxxxxxx> wrote: > > On 12/19/23 2:56 PM, Hou Tao wrote: > > From: Hou Tao <houtao1@xxxxxxxxxx> > > > > According to the implementation of atomic_xchg() under x86-64, the lock > > prefix is not necessary for BPF_XCHG atomic operation, so just remove > > it. > > It's probably a good idea for the commit message to explicitly quote the > Intel docs in here, so it's easier to find on why the lock prefix would > not be needed for the xchg op. It's a surprise to me as well. Definitely more info would be good. Also if xchg insn without lock somehow implies lock in the HW what is the harm of adding it explicitly? If it's a lock in HW than performance with and without lock prefix should be the same, right?
