The CPU field will be moved back into thread_info even when
THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
of struct thread_info.
Signed-off-by: Ard Biesheuvel <[email protected]>
---
arch/powerpc/include/asm/thread_info.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/arch/powerpc/include/asm/thread_info.h b/arch/powerpc/include/asm/thread_info.h
index b4ec6c7dd72e..5725029aaa29 100644
--- a/arch/powerpc/include/asm/thread_info.h
+++ b/arch/powerpc/include/asm/thread_info.h
@@ -47,6 +47,9 @@
struct thread_info {
int preempt_count; /* 0 => preemptable,
<0 => BUG */
+#ifdef CONFIG_SMP
+ unsigned int cpu;
+#endif
unsigned long local_flags; /* private flags for thread */
#ifdef CONFIG_LIVEPATCH
unsigned long *livepatch_sp;
--
2.30.2
On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
>
> The CPU field will be moved back into thread_info even when
> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
> of struct thread_info.
>
> Signed-off-by: Ard Biesheuvel <[email protected]>
Michael,
Do you have any objections or issues with this patch or the subsequent
ones cleaning up the task CPU kludge for ppc32? Christophe indicated
that he was happy with it.
Thanks,
Ard.
> ---
> arch/powerpc/include/asm/thread_info.h | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/arch/powerpc/include/asm/thread_info.h b/arch/powerpc/include/asm/thread_info.h
> index b4ec6c7dd72e..5725029aaa29 100644
> --- a/arch/powerpc/include/asm/thread_info.h
> +++ b/arch/powerpc/include/asm/thread_info.h
> @@ -47,6 +47,9 @@
> struct thread_info {
> int preempt_count; /* 0 => preemptable,
> <0 => BUG */
> +#ifdef CONFIG_SMP
> + unsigned int cpu;
> +#endif
> unsigned long local_flags; /* private flags for thread */
> #ifdef CONFIG_LIVEPATCH
> unsigned long *livepatch_sp;
> --
> 2.30.2
>
Ard Biesheuvel <[email protected]> writes:
> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
>>
>> The CPU field will be moved back into thread_info even when
>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
>> of struct thread_info.
>>
>> Signed-off-by: Ard Biesheuvel <[email protected]>
>
> Michael,
>
> Do you have any objections or issues with this patch or the subsequent
> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
> that he was happy with it.
No objections, it looks good to me, thanks for cleaning up that horror :)
It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
git tree with the dependencies I'd be happy to run some tests over it.
cheers
Michael Ellerman <[email protected]> writes:
> Ard Biesheuvel <[email protected]> writes:
>> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
>>>
>>> The CPU field will be moved back into thread_info even when
>>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
>>> of struct thread_info.
>>>
>>> Signed-off-by: Ard Biesheuvel <[email protected]>
>>
>> Michael,
>>
>> Do you have any objections or issues with this patch or the subsequent
>> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
>> that he was happy with it.
>
> No objections, it looks good to me, thanks for cleaning up that horror :)
>
> It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
> git tree with the dependencies I'd be happy to run some tests over it.
Actually I realised I can just drop the last patch.
So that looks fine, passes my standard quick build & boot on qemu tests,
and builds with/without stack protector enabled.
cheers
On Tue, 28 Sept 2021 at 02:16, Michael Ellerman <[email protected]> wrote:
>
> Michael Ellerman <[email protected]> writes:
> > Ard Biesheuvel <[email protected]> writes:
> >> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
> >>>
> >>> The CPU field will be moved back into thread_info even when
> >>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
> >>> of struct thread_info.
> >>>
> >>> Signed-off-by: Ard Biesheuvel <[email protected]>
> >>
> >> Michael,
> >>
> >> Do you have any objections or issues with this patch or the subsequent
> >> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
> >> that he was happy with it.
> >
> > No objections, it looks good to me, thanks for cleaning up that horror :)
> >
> > It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
> > git tree with the dependencies I'd be happy to run some tests over it.
>
> Actually I realised I can just drop the last patch.
>
> So that looks fine, passes my standard quick build & boot on qemu tests,
> and builds with/without stack protector enabled.
>
Thanks.
Do you have any opinion on how this series should be merged? Kees Cook
is willing to take them via his cross-arch tree, or you could carry
them if you prefer. Taking it via multiple trees at the same time is
going to be tricky, or take two cycles, with I'd prefer to avoid.
--
Ard.
On Thu, Sep 30, 2021 at 08:46:04AM +1000, Michael Ellerman wrote:
> Ard Biesheuvel <[email protected]> writes:
> > On Tue, 28 Sept 2021 at 02:16, Michael Ellerman <[email protected]> wrote:
> >>
> >> Michael Ellerman <[email protected]> writes:
> >> > Ard Biesheuvel <[email protected]> writes:
> >> >> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
> >> >>>
> >> >>> The CPU field will be moved back into thread_info even when
> >> >>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
> >> >>> of struct thread_info.
> >> >>>
> >> >>> Signed-off-by: Ard Biesheuvel <[email protected]>
> >> >>
> >> >> Michael,
> >> >>
> >> >> Do you have any objections or issues with this patch or the subsequent
> >> >> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
> >> >> that he was happy with it.
> >> >
> >> > No objections, it looks good to me, thanks for cleaning up that horror :)
> >> >
> >> > It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
> >> > git tree with the dependencies I'd be happy to run some tests over it.
> >>
> >> Actually I realised I can just drop the last patch.
> >>
> >> So that looks fine, passes my standard quick build & boot on qemu tests,
> >> and builds with/without stack protector enabled.
> >>
> >
> > Thanks.
> >
> > Do you have any opinion on how this series should be merged? Kees Cook
> > is willing to take them via his cross-arch tree, or you could carry
> > them if you prefer. Taking it via multiple trees at the same time is
> > going to be tricky, or take two cycles, with I'd prefer to avoid.
>
> I don't really mind. If Kees is happy to take it then that's OK by me.
>
> If Kees put the series in a topic branch based off rc2 then I could
> merge that, and avoid any conflicts.
If that helps, yeah, I can make a separate stable branch. Thanks!
-Kees
--
Kees Cook
Ard Biesheuvel <[email protected]> writes:
> On Tue, 28 Sept 2021 at 02:16, Michael Ellerman <[email protected]> wrote:
>>
>> Michael Ellerman <[email protected]> writes:
>> > Ard Biesheuvel <[email protected]> writes:
>> >> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
>> >>>
>> >>> The CPU field will be moved back into thread_info even when
>> >>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
>> >>> of struct thread_info.
>> >>>
>> >>> Signed-off-by: Ard Biesheuvel <[email protected]>
>> >>
>> >> Michael,
>> >>
>> >> Do you have any objections or issues with this patch or the subsequent
>> >> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
>> >> that he was happy with it.
>> >
>> > No objections, it looks good to me, thanks for cleaning up that horror :)
>> >
>> > It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
>> > git tree with the dependencies I'd be happy to run some tests over it.
>>
>> Actually I realised I can just drop the last patch.
>>
>> So that looks fine, passes my standard quick build & boot on qemu tests,
>> and builds with/without stack protector enabled.
>>
>
> Thanks.
>
> Do you have any opinion on how this series should be merged? Kees Cook
> is willing to take them via his cross-arch tree, or you could carry
> them if you prefer. Taking it via multiple trees at the same time is
> going to be tricky, or take two cycles, with I'd prefer to avoid.
I don't really mind. If Kees is happy to take it then that's OK by me.
If Kees put the series in a topic branch based off rc2 then I could
merge that, and avoid any conflicts.
cheers
On Thu, Sep 30, 2021 at 08:46:04AM +1000, Michael Ellerman wrote:
> Ard Biesheuvel <[email protected]> writes:
> > On Tue, 28 Sept 2021 at 02:16, Michael Ellerman <[email protected]> wrote:
> >>
> >> Michael Ellerman <[email protected]> writes:
> >> > Ard Biesheuvel <[email protected]> writes:
> >> >> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
> >> >>>
> >> >>> The CPU field will be moved back into thread_info even when
> >> >>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
> >> >>> of struct thread_info.
> >> >>>
> >> >>> Signed-off-by: Ard Biesheuvel <[email protected]>
> >> >>
> >> >> Michael,
> >> >>
> >> >> Do you have any objections or issues with this patch or the subsequent
> >> >> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
> >> >> that he was happy with it.
> >> >
> >> > No objections, it looks good to me, thanks for cleaning up that horror :)
> >> >
> >> > It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
> >> > git tree with the dependencies I'd be happy to run some tests over it.
> >>
> >> Actually I realised I can just drop the last patch.
> >>
> >> So that looks fine, passes my standard quick build & boot on qemu tests,
> >> and builds with/without stack protector enabled.
> >>
> >
> > Thanks.
> >
> > Do you have any opinion on how this series should be merged? Kees Cook
> > is willing to take them via his cross-arch tree, or you could carry
> > them if you prefer. Taking it via multiple trees at the same time is
> > going to be tricky, or take two cycles, with I'd prefer to avoid.
>
> I don't really mind. If Kees is happy to take it then that's OK by me.
>
> If Kees put the series in a topic branch based off rc2 then I could
> merge that, and avoid any conflicts.
I've created:
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git for-next/thread_info/cpu
it includes a --no-ff merge commit, which I'm not sure is desirable? Let
me know if I should adjust this, or if Linus will yell about this if I
send him a PR containing a merge commit? I'm not sure what's right here.
Thanks!
--
Kees Cook
Kees Cook <[email protected]> writes:
> On Thu, Sep 30, 2021 at 08:46:04AM +1000, Michael Ellerman wrote:
>> Ard Biesheuvel <[email protected]> writes:
>> > On Tue, 28 Sept 2021 at 02:16, Michael Ellerman <[email protected]> wrote:
>> >>
>> >> Michael Ellerman <[email protected]> writes:
>> >> > Ard Biesheuvel <[email protected]> writes:
>> >> >> On Tue, 14 Sept 2021 at 14:11, Ard Biesheuvel <[email protected]> wrote:
>> >> >>>
>> >> >>> The CPU field will be moved back into thread_info even when
>> >> >>> THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
>> >> >>> of struct thread_info.
>> >> >>>
>> >> >>> Signed-off-by: Ard Biesheuvel <[email protected]>
>> >> >>
>> >> >> Michael,
>> >> >>
>> >> >> Do you have any objections or issues with this patch or the subsequent
>> >> >> ones cleaning up the task CPU kludge for ppc32? Christophe indicated
>> >> >> that he was happy with it.
>> >> >
>> >> > No objections, it looks good to me, thanks for cleaning up that horror :)
>> >> >
>> >> > It didn't apply cleanly to master so I haven't tested it at all, if you can point me at a
>> >> > git tree with the dependencies I'd be happy to run some tests over it.
>> >>
>> >> Actually I realised I can just drop the last patch.
>> >>
>> >> So that looks fine, passes my standard quick build & boot on qemu tests,
>> >> and builds with/without stack protector enabled.
>> >>
>> >
>> > Thanks.
>> >
>> > Do you have any opinion on how this series should be merged? Kees Cook
>> > is willing to take them via his cross-arch tree, or you could carry
>> > them if you prefer. Taking it via multiple trees at the same time is
>> > going to be tricky, or take two cycles, with I'd prefer to avoid.
>>
>> I don't really mind. If Kees is happy to take it then that's OK by me.
>>
>> If Kees put the series in a topic branch based off rc2 then I could
>> merge that, and avoid any conflicts.
>
> I've created:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git for-next/thread_info/cpu
>
> it includes a --no-ff merge commit, which I'm not sure is desirable? Let
> me know if I should adjust this, or if Linus will yell about this if I
> send him a PR containing a merge commit? I'm not sure what's right here.
It looks good to me.
I don't think Linus will be bothered about that merge. It has useful
information, ie. explains why you're merging it and that arch
maintainers have acked it, and quotes Ard's cover letter.
cheers