On Wed, Oct 04, 2023 at 12:14:10AM +1000, Greg Ungerer wrote:
> On 3/10/23 06:07, Matthew Wilcox wrote:
> > 00000918 <folio_unlock>:
> > 918: 206f 0004 moveal %sp@(4),%a0
> > 91c: 7001 moveq #1,%d0
> > 91e: b190 eorl %d0,%a0@
> > 920: 2010 movel %a0@,%d0
> > 922: 4a00 tstb %d0
> > 924: 6a0a bpls 930 <folio_unlock+0x18>
> > 926: 42a7 clrl %sp@-
> > 928: 2f08 movel %a0,%sp@-
> > 92a: 4eba fafa jsr %pc@(426 <folio_wake_bit>)
> > 92e: 508f addql #8,%sp
> > 930: 4e75 rts
fwiw, here's what folio_unlock looks like today without any of my
patches:
00000746 <folio_unlock>:
746: 206f 0004 moveal %sp@(4),%a0
74a: 43e8 0003 lea %a0@(3),%a1
74e: 0891 0000 bclr #0,%a1@
752: 2010 movel %a0@,%d0
754: 4a00 tstb %d0
756: 6a0a bpls 762 <folio_unlock+0x1c>
758: 42a7 clrl %sp@-
75a: 2f08 movel %a0,%sp@-
75c: 4eba fcc8 jsr %pc@(426 <folio_wake_bit>)
760: 508f addql #8,%sp
762: 4e75 rts
Same number of instructions, but today's code has slightly longer insns,
so I'm tempted to take the win?
> > We could use eori instead of eorl, at least according to table 3-9 on
> > page 3-8:
> >
> > EOR Dy,<ea>x L Source ^ Destination → Destination ISA_A
> > EORI #<data>,Dx L Immediate Data ^ Destination → Destination ISA_A
Oh. I misread. It only does EORI to a data register; it can't do EORI
to an address.
> 400413e6 <folio_unlock>:
> 400413e6: 206f 0004 moveal %sp@(4),%a0
> 400413ea: 2010 movel %a0@,%d0
> 400413ec: 0a80 0000 0001 eoril #1,%d0
> 400413f2: 2080 movel %d0,%a0@
> 400413f4: 2010 movel %a0@,%d0
> 400413f6: 4a00 tstb %d0
> 400413f8: 6c0a bges 40041404 <folio_unlock+0x1e>
> 400413fa: 42a7 clrl %sp@-
> 400413fc: 2f08 movel %a0,%sp@-
> 400413fe: 4eba ff30 jsr %pc@(40041330 <folio_wake_bit>)
> 40041402: 508f addql #8,%sp
> 40041404: 4e75 rts
>
> But that is still worse anyway.
Yup. Looks like the version I posted actually does the best! I'll
munge that into the patch series and repost. Thanks for your help!
