On Fri, Dec 18, 2020 at 5:39 PM Chris= tophe Leroy <christophe.l= eroy@csgroup.eu> wrote:

It can cause, or it causes ? Did you encounter the issue ?

Yes, in configs that result in the section layout I des= cribed, the crush is consistent.

>
> The init sections are ordered like this:
>=C2=A0 =C2=A0 =C2=A0 =C2=A0.init.text
>=C2=A0 =C2=A0 =C2=A0 =C2=A0.exit.text
>=C2=A0 =C2=A0 =C2=A0 =C2=A0.init.data
>
> Currently, these sections aren't page aligned.
>
> Because the init code is mapped read-only at runtime and because the > .init.text section can potentially reside on the same physical page as=
> .init.data, the beginning of .init.data might be mapped read-only alon= g
> with .init.text.

init code is mapped PAGE_KERNEL_TEXT.

Whether PAGE_KERNEL_TEXT is read-only or not depends on the selected option= s.

You are right, of course. Should I change the commi= t message to 'might be mapped' or something?

>
> Then when the kernel tries to modify a variable in .init.data (like > kthreadd_done, used in kernel_init()) the kernel panics.
>
> To avoid this, I made these sections page aligned.

Should write this unpersonal, something like "To avoid this, make thes= e sections page aligned"

Noted, thanks.

=C2=A0

>
> Fixes: 060ef9d89d18 ("powerpc32: PAGE_EXEC required for inittext&= quot;)
> Signed-off-by: Ariel Marcovitch <ariel.marcovitch@gmail.com>
> ---
>=C2=A0 =C2=A0arch/powerpc/kernel/vmlinux.lds.S | 7 +++++++
>=C2=A0 =C2=A01 file changed, 7 insertions(+)
>
> diff --git a/arch/powerpc/kernel/vmlinux.lds.S b/arch/powerpc/kernel/v= mlinux.lds.S
> index 326e113d2e45..e3a7c90c03f4 100644
> --- a/arch/powerpc/kernel/vmlinux.lds.S
> +++ b/arch/powerpc/kernel/vmlinux.lds.S
> @@ -179,6 +179,11 @@ SECTIONS
>=C2=A0 =C2=A0#endif
>=C2=A0 =C2=A0 =C2=A0 =C2=A0} :text
>=C2=A0 =C2=A0
> +=C2=A0 =C2=A0 =C2=A0/* .init.text is made RO and .exit.text is not, s= o we must
> +=C2=A0 =C2=A0 =C2=A0 * ensure these sections reside in separate physi= cal pages.
> +=C2=A0 =C2=A0 =C2=A0 */
> +=C2=A0 =C2=A0 =C2=A0. =3D ALIGN(PAGE_SIZE);
> +

In principle, as it is text, it should be made RO as well. But what happens= at the begining doesn't
really matter, anyway .exit.text should never be executed and is discarded = together with init text.
So, I think it is OK the live with it as is for the time being.

Making it page aligned makes sense anyway.

Should we make _einittext page aligned instead, just like _etext ?
Yes, this will probably be better (because when=C2=A0_einittext is= not aligned,
the part of the page after _einittext is mapped RO = implicitly, and it's
hard to notice from the code). I suppose= you mean something like this:
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0= =C2=A0 =C2=A0 =C2=A0 _sinittext =3D .;
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 INIT_TEXT
+
+ =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 . =3D ALIGN(.);
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 _einittext =3D .;

>=C2=A0 =C2=A0 =C2=A0 =C2=A0/* .exit.text is discarded at runtime, not l= ink time,
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 * to deal with references from __bug_table<= br> >=C2=A0 =C2=A0 =C2=A0 =C2=A0 */
> @@ -186,6 +191,8 @@ SECTIONS
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0EXIT_TEXT
>=C2=A0 =C2=A0 =C2=A0 =C2=A0}
>=C2=A0 =C2=A0
> +=C2=A0 =C2=A0 =C2=A0. =3D ALIGN(PAGE_SIZE);
> +

Here for sure, as you explain in the coming log, this needs to be separated= from init text. So
making it aligned is a must.=C2=A0

>=C2=A0 =C2=A0 =C2=A0 =C2=A0.init.data : AT(ADDR(.init.data) - LOAD_OFFS= ET) {
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0INIT_DATA
>=C2=A0 =C2=A0 =C2=A0 =C2=A0}
>
> base-commit: 1398820fee515873379809a6415930ad0764b2f6
>

Christophe
Thanks for your time,
=C2=A0 =C2= =A0 =C2=A0 =C2=A0 Ariel Marcovitch