Re: [PATCH v7 01/22] x86/decompressor: Don't rely on upper 32 bits of GPRs being preserved

[Ard Biesheuvel]

On Mon, 31 Jul 2023 at 12:07, Borislav Petkov <bp@xxxxxxxxx> wrote:
>
> On Fri, Jul 28, 2023 at 11:08:55AM +0200, Ard Biesheuvel wrote:
> > The 4-to-5 level mode switch trampoline disables long mode and paging in
> > order to be able to flick the LA57 bit. According to section 3.4.1.1 of
> > the x86 architecture manual [0], we should not rely on 64-bit GPRs
>
> Please use passive voice in your commit message: no "we" or "I", etc,
> and describe your changes in imperative mood.
>

Sure.

> > retaining the upper 32 bits of their contents across such a mode switch.
> >
> > Given that RBP, RBX and RSI are live at this point, let's preserve them
> > on the stack, along with the return address that might be above 4G as
> > well.
> >
> > [0] Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 1: Basic Architecture
> >
> >   "Because the upper 32 bits of 64-bit general-purpose registers are
> >    undefined in 32-bit modes, the upper 32 bits of any general-purpose
> >    register are not preserved when switching from 64-bit mode to a 32-bit
> >    mode (to protected mode or compatibility mode). Software must not
> >    depend on these bits to maintain a value after a 64-bit to 32-bit
> >    mode switch."
> >
> > Fixes: 194a9749c73d650c ("x86/boot/compressed/64: Handle 5-level paging boot if kernel is above 4G")
> > Signed-off-by: Ard Biesheuvel <ardb@xxxxxxxxxx>
> > ---
> >  arch/x86/boot/compressed/head_64.S | 23 +++++++++++++++-----
> >  1 file changed, 17 insertions(+), 6 deletions(-)
> >
> > diff --git a/arch/x86/boot/compressed/head_64.S b/arch/x86/boot/compressed/head_64.S
> > index 03c4328a88cbd5d0..f7c11a0018477de8 100644
> > --- a/arch/x86/boot/compressed/head_64.S
> > +++ b/arch/x86/boot/compressed/head_64.S
> > @@ -459,11 +459,22 @@ SYM_CODE_START(startup_64)
> >       /* Save the trampoline address in RCX */
> >       movq    %rax, %rcx
> >
> > +     /* Set up 32-bit addressable stack */
> > +     leaq    TRAMPOLINE_32BIT_STACK_END(%rcx), %rsp
> > +
> >       /*
> > -      * Load the address of trampoline_return() into RDI.
> > -      * It will be used by the trampoline to return to the main code.
> > +      * Load the address of trampoline_return() into RDI and push it onto
> > +      * the stack so it will survive 32-bit truncation due to the 32-bit
>
> I think you should explain here what that SDM excerpt above says so that
> it is clear what "32-bit truncation" is meant.
>

Ok

> > +      * protected mode switch. It will be used by the trampoline to return
> > +      * to the main code.
> >        */
> >       leaq    trampoline_return(%rip), %rdi
> > +     pushq   %rdi
> > +
> > +     /* Preserve other live 64-bit registers */
> > +     pushq   %rbp
> > +     pushq   %rbx
> > +     pushq   %rsi
> >
> >       /* Switch to compatibility mode (CS.L = 0 CS.D = 1) via far return */
> >       pushq   $__KERNEL32_CS
>
> What is more interesting is why is this trampoline crap unconditional?
>
> /me goes and does git archeology...
>
> 194a9749c73d ("x86/boot/compressed/64: Handle 5-level paging boot if kernel is above 4G")
>
>     We go through the trampoline even if we don't have to: if we're already
>     in 5-level paging mode or if we don't need to switch to it. This way the
>     trampoline gets tested on every boot.
>
> Well, f*ck no.
>
> All my machines don't have 5level pages. And I bet 5level pages is still
> not in the majority of the machines out there.
>
> This all needs to be abstracted away into a separate function which
> exits early if no 5 level support.
>
> Kirill, please fix this.
>

This is already fixed further down in the series.