On Wed, Oct 12, 2022 at 04:02:43PM +0800, Patrick Yingxi Pan wrote:
> Thank you Paul and Akira!
>
> Hmm, it seems that I successfully witnessed a force push onto an open
> repository. ;-p
That you did. Like this:
git rebase -i --onto cc128b0969f7 cc128b0969f7 master
# for commit 520ad603b0f3 s/pick/reword/
# in the editor add your Signed-off by
git push perfbook +master
git push github +master
# The "perfbook" and "github" have no significance outside of
# my local repository.
It is good practice to leave a branch behind, though I skipped that
this time because this was a simple transformation. So you might do
"git branch old-master" before the "git rebase". That allows you to
start over in case something breaks during the rebase.
Thanx, Paul
> On Wed, Oct 12, 2022 at 3:30 PM Paul E. McKenney <paulmck@xxxxxxxxxx> wrote:
> >
> > On Wed, Oct 12, 2022 at 12:19:04PM +0900, Akira Yokosawa wrote:
> > > On Tue, 11 Oct 2022 02:02:08 -0700, Paul E. McKenney wrote:
> > > > On Tue, Oct 11, 2022 at 11:20:49AM +0800, Patrick Yingxi Pan wrote:
> > > >> Hello!
> > > >>
> > > >> I have added the missing READ_ONCE(). Please take a look. Thanks.
> > > >>
> > > >> Signed-off-by: Patrick Pan <pyxchina92929@xxxxxxxxx>
> > > >
> > > > Queued and pushed, thank you!
> > >
> > > Paul, somehow Patrick's S-o-b tag is lost in the amended changelog of
> > > commit 520ad603b0f3 ("memorder: Correct the transformed Listing 15.30
> > > code").
> > >
> > > I don't think it is mandatory in perfbook, but if you can, please fix
> > > up.
> >
> > Gah! Thank you for catching this, fixed.
> >
> > Thanx, Paul
> >
> > > >> ---
> > > >> memorder/memorder.tex | 2 +-
> > > >> 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >>
> > > >> diff --git a/memorder/memorder.tex b/memorder/memorder.tex
> > > >> index a0a87d0a..50ff2882 100644
> > > >> --- a/memorder/memorder.tex
> > > >> +++ b/memorder/memorder.tex
> > > >> @@ -2869,7 +2869,7 @@ An equals comparison on \clnref{equ} might lead
> > > >> the compiler to (incorrectly)
> > > >> conclude that both pointers are equivalent, when in fact they carry
> > > >> different dependencies.
> > > >> This means that the compiler might well transform \clnref{pc} to instead
> > > >> -be \co{r2 = q->c}, which might well cause the value 44 to be loaded
> > > >> +be \co{r2 = READ_ONCE(q->c)}, which might well cause the value 44 to be loaded
> > > >> instead of the expected value 144.
> > > >> \end{fcvref}
> > > >>
> > > >>
> > > >> base-commit: 7f12a9358e220f3d0c3a0880d01bc283113d7a5b
> > > >> --
