On Wed, Jun 11, 2025 at 03:03:35PM -0600, Keith Busch wrote: > On Tue, Jun 10, 2025 at 07:06:43AM +0200, Christoph Hellwig wrote: > > - iod->dma_len = bv->bv_len; > > + iod->cmd.common.dptr.prp1 = cpu_to_le64(dma_addr); > > + iod->cmd.common.dptr.prp2 = 0; > > + if (bv.bv_len > first_prp_len) > > + iod->cmd.common.dptr.prp2 = > > + cpu_to_le64(dma_addr + first_prp_len); > > Nit, set prp2 to 0 in an 'else' case instead of unconditionally before > overwriting it? unconditionally setting it actually generates slightly better code, assuming the compiler won't just optimize the other form to it anyway. But if you think it is more readable I can change it around.
