Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Lee Jones]

On Mon, 05 Dec 2022, Ard Biesheuvel wrote:

> With the introduction of PRMT in the ACPI subsystem, the EFI rts
> workqueue is no longer the only caller of efi_call_virt_pointer() in the
> kernel. This means the EFI runtime services lock is no longer sufficient
> to manage concurrent calls into firmware, but also that firmware calls
> may occur that are not marshalled via the workqueue mechanism, but
> originate directly from the caller context.
> 
> For added robustness, and to ensure that the runtime services have 8 KiB
> of stack space available as per the EFI spec, introduce a spinlock
> protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> serialization between the EFI rts workqueue (which itself serializes EFI
> runtime calls) and other callers of efi_call_virt_pointer().
> 
> While at it, use the stack pivot to avoid reloading the shadow call
> stack pointer from the ordinary stack, as doing so could produce a
> gadget to defeat it.
> 
> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> ---
>  arch/arm64/include/asm/efi.h       |  3 +++
>  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
>  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
>  3 files changed, 40 insertions(+), 1 deletion(-)

Could we have this in Stable please?

Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")

Ard, do we need Patch 2 as well, or can this be applied on its own?

> diff --git a/arch/arm64/include/asm/efi.h b/arch/arm64/include/asm/efi.h
> index 7c12e01c2b312e7b..1c408ec3c8b3a883 100644
> --- a/arch/arm64/include/asm/efi.h
> +++ b/arch/arm64/include/asm/efi.h
> @@ -25,6 +25,7 @@ int efi_set_mapping_permissions(struct mm_struct *mm, efi_memory_desc_t *md);
>  ({									\
>  	efi_virtmap_load();						\
>  	__efi_fpsimd_begin();						\
> +	spin_lock(&efi_rt_lock);					\
>  })
>  
>  #undef arch_efi_call_virt
> @@ -33,10 +34,12 @@ int efi_set_mapping_permissions(struct mm_struct *mm, efi_memory_desc_t *md);
>  
>  #define arch_efi_call_virt_teardown()					\
>  ({									\
> +	spin_unlock(&efi_rt_lock);					\
>  	__efi_fpsimd_end();						\
>  	efi_virtmap_unload();						\
>  })
>  
> +extern spinlock_t efi_rt_lock;
>  efi_status_t __efi_rt_asm_wrapper(void *, const char *, ...);
>  
>  #define ARCH_EFI_IRQ_FLAGS_MASK (PSR_D_BIT | PSR_A_BIT | PSR_I_BIT | PSR_F_BIT)
> diff --git a/arch/arm64/kernel/efi-rt-wrapper.S b/arch/arm64/kernel/efi-rt-wrapper.S
> index 75691a2641c1c0f8..b2786b968fee68dd 100644
> --- a/arch/arm64/kernel/efi-rt-wrapper.S
> +++ b/arch/arm64/kernel/efi-rt-wrapper.S
> @@ -16,6 +16,12 @@ SYM_FUNC_START(__efi_rt_asm_wrapper)
>  	 */
>  	stp	x1, x18, [sp, #16]
>  
> +	ldr_l	x16, efi_rt_stack_top
> +	mov	sp, x16
> +#ifdef CONFIG_SHADOW_CALL_STACK
> +	str	x18, [sp, #-16]!
> +#endif
> +
>  	/*
>  	 * We are lucky enough that no EFI runtime services take more than
>  	 * 5 arguments, so all are passed in registers rather than via the
> @@ -29,6 +35,7 @@ SYM_FUNC_START(__efi_rt_asm_wrapper)
>  	mov	x4, x6
>  	blr	x8
>  
> +	mov	sp, x29
>  	ldp	x1, x2, [sp, #16]
>  	cmp	x2, x18
>  	ldp	x29, x30, [sp], #32
> @@ -42,6 +49,10 @@ SYM_FUNC_START(__efi_rt_asm_wrapper)
>  	 * called with preemption disabled and a separate shadow stack is used
>  	 * for interrupts.
>  	 */
> -	mov	x18, x2
> +#ifdef CONFIG_SHADOW_CALL_STACK
> +	ldr_l	x18, efi_rt_stack_top
> +	ldr	x18, [x18, #-16]
> +#endif
> +
>  	b	efi_handle_corrupted_x18	// tail call
>  SYM_FUNC_END(__efi_rt_asm_wrapper)
> diff --git a/arch/arm64/kernel/efi.c b/arch/arm64/kernel/efi.c
> index a908a37f03678b6b..8cb2e005f8aca589 100644
> --- a/arch/arm64/kernel/efi.c
> +++ b/arch/arm64/kernel/efi.c
> @@ -144,3 +144,28 @@ asmlinkage efi_status_t efi_handle_corrupted_x18(efi_status_t s, const char *f)
>  	pr_err_ratelimited(FW_BUG "register x18 corrupted by EFI %s\n", f);
>  	return s;
>  }
> +
> +DEFINE_SPINLOCK(efi_rt_lock);
> +
> +asmlinkage u64 *efi_rt_stack_top __ro_after_init;
> +
> +/* required by the EFI spec */
> +static_assert(THREAD_SIZE >= SZ_8K);
> +
> +int __init arm64_efi_rt_init(void)
> +{
> +	void *p = __vmalloc_node_range(THREAD_SIZE, THREAD_ALIGN,
> +				       VMALLOC_START, VMALLOC_END, GFP_KERNEL,
> +				       PAGE_KERNEL, 0, NUMA_NO_NODE,
> +				       __builtin_return_address(0));
> +
> +	if (!p) {
> +		pr_warn("Failed to allocate EFI runtime stack\n");
> +		clear_bit(EFI_RUNTIME_SERVICES, &efi.flags);
> +		return -ENOMEM;
> +	}
> +
> +	efi_rt_stack_top = p + THREAD_SIZE;
> +	return 0;
> +}
> +core_initcall(arm64_efi_rt_init);
> -- 
> 2.35.1
> 
> 

-- 
Lee Jones [李琼斯]

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Ard Biesheuvel]

On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
>
> On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
>
> > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > kernel. This means the EFI runtime services lock is no longer sufficient
> > to manage concurrent calls into firmware, but also that firmware calls
> > may occur that are not marshalled via the workqueue mechanism, but
> > originate directly from the caller context.
> >
> > For added robustness, and to ensure that the runtime services have 8 KiB
> > of stack space available as per the EFI spec, introduce a spinlock
> > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > serialization between the EFI rts workqueue (which itself serializes EFI
> > runtime calls) and other callers of efi_call_virt_pointer().
> >
> > While at it, use the stack pivot to avoid reloading the shadow call
> > stack pointer from the ordinary stack, as doing so could produce a
> > gadget to defeat it.
> >
> > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > ---
> >  arch/arm64/include/asm/efi.h       |  3 +++
> >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> >  3 files changed, 40 insertions(+), 1 deletion(-)
>
> Could we have this in Stable please?
>
> Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
>
> Ard, do we need Patch 2 as well, or can this be applied on its own?
>

Thanks for the reminder.

Only patch #1 is needed. It should be applied to v5.10 and later.

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Lee Jones]

On Wed, 04 Jan 2023, Ard Biesheuvel wrote:

> On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> >
> > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> >
> > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > to manage concurrent calls into firmware, but also that firmware calls
> > > may occur that are not marshalled via the workqueue mechanism, but
> > > originate directly from the caller context.
> > >
> > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > of stack space available as per the EFI spec, introduce a spinlock
> > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > runtime calls) and other callers of efi_call_virt_pointer().
> > >
> > > While at it, use the stack pivot to avoid reloading the shadow call
> > > stack pointer from the ordinary stack, as doing so could produce a
> > > gadget to defeat it.
> > >
> > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > ---
> > >  arch/arm64/include/asm/efi.h       |  3 +++
> > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > >  3 files changed, 40 insertions(+), 1 deletion(-)
> >
> > Could we have this in Stable please?
> >
> > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> >
> > Ard, do we need Patch 2 as well, or can this be applied on its own?
> >
> 
> Thanks for the reminder.
> 
> Only patch #1 is needed. It should be applied to v5.10 and later.

Perfect, thanks Ard.

-- 
Lee Jones [李琼斯]

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Greg KH]

On Wed, Jan 04, 2023 at 02:42:30PM +0000, Lee Jones wrote:
> On Wed, 04 Jan 2023, Ard Biesheuvel wrote:
> 
> > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > >
> > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > >
> > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > originate directly from the caller context.
> > > >
> > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > >
> > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > gadget to defeat it.
> > > >
> > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > ---
> > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > >
> > > Could we have this in Stable please?
> > >
> > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > >
> > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > >
> > 
> > Thanks for the reminder.
> > 
> > Only patch #1 is needed. It should be applied to v5.10 and later.
> 
> Perfect, thanks Ard.

Now queued up, thanks.

greg k-h

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Mark Rutland]

On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> >
> > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> >
> > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > to manage concurrent calls into firmware, but also that firmware calls
> > > may occur that are not marshalled via the workqueue mechanism, but
> > > originate directly from the caller context.
> > >
> > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > of stack space available as per the EFI spec, introduce a spinlock
> > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > runtime calls) and other callers of efi_call_virt_pointer().
> > >
> > > While at it, use the stack pivot to avoid reloading the shadow call
> > > stack pointer from the ordinary stack, as doing so could produce a
> > > gadget to defeat it.
> > >
> > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > ---
> > >  arch/arm64/include/asm/efi.h       |  3 +++
> > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > >  3 files changed, 40 insertions(+), 1 deletion(-)
> >
> > Could we have this in Stable please?
> >
> > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> >
> > Ard, do we need Patch 2 as well, or can this be applied on its own?
> >
> 
> Thanks for the reminder.
> 
> Only patch #1 is needed. It should be applied to v5.10 and later.

Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
the stack unwinder?

From your last reply to me there I was expecting a respin with that fixed.

Mark.

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Ard Biesheuvel]

On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
>
> On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > >
> > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > >
> > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > originate directly from the caller context.
> > > >
> > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > >
> > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > gadget to defeat it.
> > > >
> > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > ---
> > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > >
> > > Could we have this in Stable please?
> > >
> > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > >
> > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > >
> >
> > Thanks for the reminder.
> >
> > Only patch #1 is needed. It should be applied to v5.10 and later.
>
> Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> the stack unwinder?
>
> From your last reply to me there I was expecting a respin with that fixed.
>

Apologies for the confusion.

I have a patch for this queued up, but AIUI, that cannot be merged all
the way back to v5.10, so these need to remain separate changes in any
case.

https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Mark Rutland]

On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> >
> > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > >
> > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > >
> > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > originate directly from the caller context.
> > > > >
> > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > >
> > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > gadget to defeat it.
> > > > >
> > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > ---
> > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > >
> > > > Could we have this in Stable please?
> > > >
> > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > >
> > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > >
> > >
> > > Thanks for the reminder.
> > >
> > > Only patch #1 is needed. It should be applied to v5.10 and later.
> >
> > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > the stack unwinder?
> >
> > From your last reply to me there I was expecting a respin with that fixed.
> >
> 
> Apologies for the confusion.
> 
> I have a patch for this queued up, but AIUI, that cannot be merged all
> the way back to v5.10, so these need to remain separate changes in any
> case.
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013

Ah, ok, thanks for the pointer!

I'm a little uneasy here, still.

By backporting this we're also backporting the new breakage of the stack
unwinder, and the minimal change for backports would be to add the lock and not
the new stack (which was added for additinoal robustness, not to fix the bug
the lock fixes).

I do appreciate that the additional stack is likely more useful than the
occasional diagnostic output from the kernel, but it does seem like this has
traded off one bug for another, and I'm just a little annoyed because I pointed
that out before the first pull request was made.

I do know that this isn't malicious, and I'm not trying to start a fight, but
now we have to consider whether we want/need to backport a stack unwinder fix
to account for this, and we hadn't had that discussion before.

Thanks,
Mark.

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Ard Biesheuvel]

On Wed, 4 Jan 2023 at 17:30, Mark Rutland <mark.rutland@arm.com> wrote:
>
> On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> > On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> > >
> > > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > > >
> > > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > > >
> > > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > > originate directly from the caller context.
> > > > > >
> > > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > > >
> > > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > > gadget to defeat it.
> > > > > >
> > > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > > ---
> > > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > > >
> > > > > Could we have this in Stable please?
> > > > >
> > > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > > >
> > > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > > >
> > > >
> > > > Thanks for the reminder.
> > > >
> > > > Only patch #1 is needed. It should be applied to v5.10 and later.
> > >
> > > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > > the stack unwinder?
> > >
> > > From your last reply to me there I was expecting a respin with that fixed.
> > >
> >
> > Apologies for the confusion.
> >
> > I have a patch for this queued up, but AIUI, that cannot be merged all
> > the way back to v5.10, so these need to remain separate changes in any
> > case.
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013
>
> Ah, ok, thanks for the pointer!
>
> I'm a little uneasy here, still.
>
> By backporting this we're also backporting the new breakage of the stack
> unwinder, and the minimal change for backports would be to add the lock and not
> the new stack (which was added for additinoal robustness, not to fix the bug
> the lock fixes).
>
> I do appreciate that the additional stack is likely more useful than the
> occasional diagnostic output from the kernel, but it does seem like this has
> traded off one bug for another, and I'm just a little annoyed because I pointed
> that out before the first pull request was made.
>
> I do know that this isn't malicious, and I'm not trying to start a fight, but
> now we have to consider whether we want/need to backport a stack unwinder fix
> to account for this, and we hadn't had that discussion before.
>

In that case, let's drop these backports for the time being, and
collaborate on a solution that works for all of us.

Greg, could you please drop these again? Thanks.

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Greg KH]

On Wed, Jan 04, 2023 at 05:32:18PM +0100, Ard Biesheuvel wrote:
> On Wed, 4 Jan 2023 at 17:30, Mark Rutland <mark.rutland@arm.com> wrote:
> >
> > On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> > > On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> > > >
> > > > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > > > >
> > > > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > > > >
> > > > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > > > originate directly from the caller context.
> > > > > > >
> > > > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > > > >
> > > > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > > > gadget to defeat it.
> > > > > > >
> > > > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > > > ---
> > > > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > > > >
> > > > > > Could we have this in Stable please?
> > > > > >
> > > > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > > > >
> > > > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > > > >
> > > > >
> > > > > Thanks for the reminder.
> > > > >
> > > > > Only patch #1 is needed. It should be applied to v5.10 and later.
> > > >
> > > > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > > > the stack unwinder?
> > > >
> > > > From your last reply to me there I was expecting a respin with that fixed.
> > > >
> > >
> > > Apologies for the confusion.
> > >
> > > I have a patch for this queued up, but AIUI, that cannot be merged all
> > > the way back to v5.10, so these need to remain separate changes in any
> > > case.
> > >
> > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013
> >
> > Ah, ok, thanks for the pointer!
> >
> > I'm a little uneasy here, still.
> >
> > By backporting this we're also backporting the new breakage of the stack
> > unwinder, and the minimal change for backports would be to add the lock and not
> > the new stack (which was added for additinoal robustness, not to fix the bug
> > the lock fixes).
> >
> > I do appreciate that the additional stack is likely more useful than the
> > occasional diagnostic output from the kernel, but it does seem like this has
> > traded off one bug for another, and I'm just a little annoyed because I pointed
> > that out before the first pull request was made.
> >
> > I do know that this isn't malicious, and I'm not trying to start a fight, but
> > now we have to consider whether we want/need to backport a stack unwinder fix
> > to account for this, and we hadn't had that discussion before.
> >
> 
> In that case, let's drop these backports for the time being, and
> collaborate on a solution that works for all of us.
> 
> Greg, could you please drop these again? Thanks.

Dropped now from all queues, thanks.

greg k-h

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Lee Jones]

On Thu, 05 Jan 2023, Greg KH wrote:

> On Wed, Jan 04, 2023 at 05:32:18PM +0100, Ard Biesheuvel wrote:
> > On Wed, 4 Jan 2023 at 17:30, Mark Rutland <mark.rutland@arm.com> wrote:
> > >
> > > On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> > > > On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> > > > >
> > > > > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > > > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > > > > >
> > > > > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > > > > >
> > > > > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > > > > originate directly from the caller context.
> > > > > > > >
> > > > > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > > > > >
> > > > > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > > > > gadget to defeat it.
> > > > > > > >
> > > > > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > > > > ---
> > > > > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > > > > >
> > > > > > > Could we have this in Stable please?
> > > > > > >
> > > > > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > > > > >
> > > > > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > > > > >
> > > > > >
> > > > > > Thanks for the reminder.
> > > > > >
> > > > > > Only patch #1 is needed. It should be applied to v5.10 and later.
> > > > >
> > > > > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > > > > the stack unwinder?
> > > > >
> > > > > From your last reply to me there I was expecting a respin with that fixed.
> > > > >
> > > >
> > > > Apologies for the confusion.
> > > >
> > > > I have a patch for this queued up, but AIUI, that cannot be merged all
> > > > the way back to v5.10, so these need to remain separate changes in any
> > > > case.
> > > >
> > > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013
> > >
> > > Ah, ok, thanks for the pointer!
> > >
> > > I'm a little uneasy here, still.
> > >
> > > By backporting this we're also backporting the new breakage of the stack
> > > unwinder, and the minimal change for backports would be to add the lock and not
> > > the new stack (which was added for additinoal robustness, not to fix the bug
> > > the lock fixes).
> > >
> > > I do appreciate that the additional stack is likely more useful than the
> > > occasional diagnostic output from the kernel, but it does seem like this has
> > > traded off one bug for another, and I'm just a little annoyed because I pointed
> > > that out before the first pull request was made.
> > >
> > > I do know that this isn't malicious, and I'm not trying to start a fight, but
> > > now we have to consider whether we want/need to backport a stack unwinder fix
> > > to account for this, and we hadn't had that discussion before.
> > >
> > 
> > In that case, let's drop these backports for the time being, and
> > collaborate on a solution that works for all of us.
> > 
> > Greg, could you please drop these again? Thanks.
> 
> Dropped now from all queues, thanks.

Now in Mainline as:

  18bba1843fc7f efi: rt-wrapper: Add missing include
  ff7a167961d1b arm64: efi: Execute runtime services from a dedicated stack

Would you be kind enough to re-collect them please?

Thank you.

-- 
Lee Jones [李琼斯]

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Greg KH]

On Tue, Jan 17, 2023 at 04:56:19PM +0000, Lee Jones wrote:
> On Thu, 05 Jan 2023, Greg KH wrote:
> 
> > On Wed, Jan 04, 2023 at 05:32:18PM +0100, Ard Biesheuvel wrote:
> > > On Wed, 4 Jan 2023 at 17:30, Mark Rutland <mark.rutland@arm.com> wrote:
> > > >
> > > > On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> > > > > On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> > > > > >
> > > > > > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > > > > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > > > > > >
> > > > > > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > > > > > >
> > > > > > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > > > > > originate directly from the caller context.
> > > > > > > > >
> > > > > > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > > > > > >
> > > > > > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > > > > > gadget to defeat it.
> > > > > > > > >
> > > > > > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > > > > > ---
> > > > > > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > > > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > > > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > > > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > > > > > >
> > > > > > > > Could we have this in Stable please?
> > > > > > > >
> > > > > > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > > > > > >
> > > > > > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > > > > > >
> > > > > > >
> > > > > > > Thanks for the reminder.
> > > > > > >
> > > > > > > Only patch #1 is needed. It should be applied to v5.10 and later.
> > > > > >
> > > > > > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > > > > > the stack unwinder?
> > > > > >
> > > > > > From your last reply to me there I was expecting a respin with that fixed.
> > > > > >
> > > > >
> > > > > Apologies for the confusion.
> > > > >
> > > > > I have a patch for this queued up, but AIUI, that cannot be merged all
> > > > > the way back to v5.10, so these need to remain separate changes in any
> > > > > case.
> > > > >
> > > > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013
> > > >
> > > > Ah, ok, thanks for the pointer!
> > > >
> > > > I'm a little uneasy here, still.
> > > >
> > > > By backporting this we're also backporting the new breakage of the stack
> > > > unwinder, and the minimal change for backports would be to add the lock and not
> > > > the new stack (which was added for additinoal robustness, not to fix the bug
> > > > the lock fixes).
> > > >
> > > > I do appreciate that the additional stack is likely more useful than the
> > > > occasional diagnostic output from the kernel, but it does seem like this has
> > > > traded off one bug for another, and I'm just a little annoyed because I pointed
> > > > that out before the first pull request was made.
> > > >
> > > > I do know that this isn't malicious, and I'm not trying to start a fight, but
> > > > now we have to consider whether we want/need to backport a stack unwinder fix
> > > > to account for this, and we hadn't had that discussion before.
> > > >
> > > 
> > > In that case, let's drop these backports for the time being, and
> > > collaborate on a solution that works for all of us.
> > > 
> > > Greg, could you please drop these again? Thanks.
> > 
> > Dropped now from all queues, thanks.
> 
> Now in Mainline as:
> 
>   18bba1843fc7f efi: rt-wrapper: Add missing include
>   ff7a167961d1b arm64: efi: Execute runtime services from a dedicated stack
> 
> Would you be kind enough to re-collect them please?

Now queued up for 5.10.y and newer.

greg k-h

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[SDL]

Dear Linux Kernel Developers,
I have reason to believe that the commit 
4012603cbd469223f225637d265a233f034c567a, which was backported to the 
linux-5.10 branch, is introducing a bug that leads to a kernel crash.

Environment Details:

Kernel Version: linux-5.10.234
Architecture: aarch64
Kernel Configuration: Available at
https://syzkaller.appspot.com/bug?extid=75dc11b3aa0369757b7c
QEMU Version: 8.2.2

QEMU Launch Command:

qemu-system-aarch64 \
     -m 4G \
     -smp 4,sockets=1,cores=4,threads=1 \
     -cpu cortex-a57 \
     -machine virt,accel=tcg \
     -kernel /home/user/lvc/linux-stable/arch/arm64/boot/Image \
     -append "console=ttyS0 root=/dev/vda2 earlyprintk=serial 
net.ifnames=0 rw debug loglevel=8" \
     -drive 
file=/home/user/lvc/image/disk.raw,format=raw,if=virtio,index=1 \
     -bios /usr/share/AAVMF/AAVMF_CODE.fd \
     -netdev user,id=net0,hostfwd=tcp::10022-:22 \
     -device virtio-net-device,netdev=net0 \
     -nographic

Steps to Reproduce:

mount -t sysfs sysfs /sys
mount -t proc proc /proc
mount -t efivarfs efivarfs /sys/firmware/efi/efivars
tools/testing/selftests/efivarfs/efivarfs.sh

Crash log:

[  101.403235][   T38] Unable to handle kernel paging request at virtual 
address ffffa00023b20000
[  101.403905][   T38] Mem abort info:
[  101.404159][   T38]   ESR = 0x96000047
[  101.404448][   T38]   EC = 0x25: DABT (current EL), IL = 32 bits
[  101.404748][   T38]   SET = 0, FnV = 0
[  101.404988][   T38]   EA = 0, S1PTW = 0
[  101.405233][   T38] Data abort info:
[  101.405469][   T38]   ISV = 0, ISS = 0x00000047
[  101.405721][   T38]   CM = 0, WnR = 1
[  101.406070][   T38] swapper pgtable: 4k pages, 48-bit VAs, 
pgdp=0000000103336000
[  101.406415][   T38] [ffffa00023b20000] pgd=000000013ffff003, 
p4d=000000013ffff003, pud=000000013fffe003, pmd=000000013fff9003, 
pte=0000000000000000
[  101.407997][   T38] Internal error: Oops: 0000000096000047 [#1] 
PREEMPT SMP
[  101.408446][   T38] Modules linked in:
[  101.409010][   T38] CPU: 1 PID: 38 Comm: kworker/u8:2 Not tainted 
5.10.234 #61
[  101.409376][   T38] Hardware name: QEMU QEMU Virtual Machine, BIOS 
2024.02-2ubuntu0.1 10/25/2024
[  101.410689][   T38] Workqueue: efi_rts_wq efi_call_rts
[  101.411219][   T38] pstate: 000003c5 (nzcv DAIF -PAN -UAO -TCO BTYPE=--)
[  101.411574][   T38] pc : el1_sync+0xc/0x140
[  101.411930][   T38] lr : 0x0
[  101.412132][   T38] sp : ffffa00023b1ffd0
[  101.412363][   T38] x29: 0000000000000000 x28: ffff0000cc741a80
[  101.412761][   T38] x27: 0000000000000001 x26: 0000000000e800e8
[  101.413117][   T38] x25: 0000000000800080 x24: ffff940004765298
[  101.413465][   T38] x23: 00000000100003c5 x22: ffffa00018c95600
[  101.413809][   T38] x21: ffffa0001923a980 x20: ffffa00018d30380
[  101.414170][   T38] x19: ffffa0001bab8008 x18: 1fffe000198e8491
[  101.414515][   T38] x17: 0000000000000000 x16: ffffa000102a26c0
[  101.414882][   T38] x15: 0000000000000001 x14: ffff0000ccf34400
[  101.415277][   T38] x13: 000000000000003c x12: ffff0000ccf34000
[  101.415656][   T38] x11: 0000000000000018 x10: 0000000000000054
[  101.416072][   T38] x9 : 4fc79849210be57c x8 : 11d293ca8be4df61
[  101.416446][   T38] x7 : 0000000041b58ab3 x6 : ffff94000476522a
[  101.416820][   T38] x5 : 0000002200000000 x4 : 0000000200000000
[  101.417190][   T38] x3 : 1fffe000198e8352 x2 : ffffa0001923a820
[  101.417560][   T38] x1 : ffff0000cc741a80 x0 : 0000000000000000
[  101.418109][   T38] Call trace:
[  101.418601][   T38] Code: d503201f a90007e0 a9010fe2 a90217e4 (a9031fe6)
[  101.419579][   T38] ---[ end trace 61316cddfdbbcb95 ]---
[  103.623221][   T38] Kernel panic - not syncing: Oops: Fatal exception
[  103.623723][   T38] SMP: stopping secondary CPUs
[  105.221708][   T38] SMP: failed to stop secondary CPUs 0-3
[  105.222306][   T38] Kernel Offset: disabled
[  105.222711][   T38] CPU features: 0x28240022,61002082
[  105.222969][   T38] Memory Limit: none

I identified this commit as the potential cause of the crash using git 
bisect. The issue is reproducible with the above setup and 
configuration. If you need more details or logs, I am more than willing 
to provide them. Thank you for your attention to this matter.

Best regards,

Alexey Nepomnyashih

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Mark Rutland]

On Wed, Jan 04, 2023 at 05:32:18PM +0100, Ard Biesheuvel wrote:
> On Wed, 4 Jan 2023 at 17:30, Mark Rutland <mark.rutland@arm.com> wrote:
> >
> > On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> > > On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> > > >
> > > > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > > > >
> > > > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > > > >
> > > > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > > > originate directly from the caller context.
> > > > > > >
> > > > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > > > >
> > > > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > > > gadget to defeat it.
> > > > > > >
> > > > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > > > ---
> > > > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > > > >
> > > > > > Could we have this in Stable please?
> > > > > >
> > > > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > > > >
> > > > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > > > >
> > > > >
> > > > > Thanks for the reminder.
> > > > >
> > > > > Only patch #1 is needed. It should be applied to v5.10 and later.
> > > >
> > > > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > > > the stack unwinder?
> > > >
> > > > From your last reply to me there I was expecting a respin with that fixed.
> > > >
> > >
> > > Apologies for the confusion.
> > >
> > > I have a patch for this queued up, but AIUI, that cannot be merged all
> > > the way back to v5.10, so these need to remain separate changes in any
> > > case.
> > >
> > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013
> >
> > Ah, ok, thanks for the pointer!
> >
> > I'm a little uneasy here, still.
> >
> > By backporting this we're also backporting the new breakage of the stack
> > unwinder, and the minimal change for backports would be to add the lock and not
> > the new stack (which was added for additinoal robustness, not to fix the bug
> > the lock fixes).
> >
> > I do appreciate that the additional stack is likely more useful than the
> > occasional diagnostic output from the kernel, but it does seem like this has
> > traded off one bug for another, and I'm just a little annoyed because I pointed
> > that out before the first pull request was made.
> >
> > I do know that this isn't malicious, and I'm not trying to start a fight, but
> > now we have to consider whether we want/need to backport a stack unwinder fix
> > to account for this, and we hadn't had that discussion before.
> 
> In that case, let's drop these backports for the time being, and
> collaborate on a solution that works for all of us.

Thanks!

IIUC our options here are:

1) Create a cut-down patch for stable that just adds the new lock but leaves
   out the new stack.

   I may be missing a reason why that's insufficient or painful.

2) Backport this *but* also backport the follow-up fixes from your other series:
   https://lore.kernel.org/r/20230104174433.1259428-1-ardb@kernel.org

   Above you mentioned something about v5.10, was that just to say that some
   manual backporting was required, or that there was a structural problem that
   would require more invasive changes / prerequisites?

3) Something else?

My preference would be (1), but if we are encountering issue with stack size on
stable kernels, then I'd be happy to help with manual backporting effort for
(2), as long as we backported all the relevant bits in one go.

Does that make sense, and does that sound reasonable to you?

Thanks,
Mark.

Re: [PATCH 1/2] arm64: efi: Execute runtime services from a dedicated stack

[Ard Biesheuvel]

On Thu, 5 Jan 2023 at 13:56, Mark Rutland <mark.rutland@arm.com> wrote:
>
> On Wed, Jan 04, 2023 at 05:32:18PM +0100, Ard Biesheuvel wrote:
> > On Wed, 4 Jan 2023 at 17:30, Mark Rutland <mark.rutland@arm.com> wrote:
> > >
> > > On Wed, Jan 04, 2023 at 05:15:34PM +0100, Ard Biesheuvel wrote:
> > > > On Wed, 4 Jan 2023 at 17:13, Mark Rutland <mark.rutland@arm.com> wrote:
> > > > >
> > > > > On Wed, Jan 04, 2023 at 02:56:19PM +0100, Ard Biesheuvel wrote:
> > > > > > On Wed, 4 Jan 2023 at 11:40, Lee Jones <lee@kernel.org> wrote:
> > > > > > >
> > > > > > > On Mon, 05 Dec 2022, Ard Biesheuvel wrote:
> > > > > > >
> > > > > > > > With the introduction of PRMT in the ACPI subsystem, the EFI rts
> > > > > > > > workqueue is no longer the only caller of efi_call_virt_pointer() in the
> > > > > > > > kernel. This means the EFI runtime services lock is no longer sufficient
> > > > > > > > to manage concurrent calls into firmware, but also that firmware calls
> > > > > > > > may occur that are not marshalled via the workqueue mechanism, but
> > > > > > > > originate directly from the caller context.
> > > > > > > >
> > > > > > > > For added robustness, and to ensure that the runtime services have 8 KiB
> > > > > > > > of stack space available as per the EFI spec, introduce a spinlock
> > > > > > > > protected EFI runtime stack of 8 KiB, where the spinlock also ensures
> > > > > > > > serialization between the EFI rts workqueue (which itself serializes EFI
> > > > > > > > runtime calls) and other callers of efi_call_virt_pointer().
> > > > > > > >
> > > > > > > > While at it, use the stack pivot to avoid reloading the shadow call
> > > > > > > > stack pointer from the ordinary stack, as doing so could produce a
> > > > > > > > gadget to defeat it.
> > > > > > > >
> > > > > > > > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > > > > > > > ---
> > > > > > > >  arch/arm64/include/asm/efi.h       |  3 +++
> > > > > > > >  arch/arm64/kernel/efi-rt-wrapper.S | 13 +++++++++-
> > > > > > > >  arch/arm64/kernel/efi.c            | 25 ++++++++++++++++++++
> > > > > > > >  3 files changed, 40 insertions(+), 1 deletion(-)
> > > > > > >
> > > > > > > Could we have this in Stable please?
> > > > > > >
> > > > > > > Upstream commit: ff7a167961d1b ("arm64: efi: Execute runtime services from a dedicated stack")
> > > > > > >
> > > > > > > Ard, do we need Patch 2 as well, or can this be applied on its own?
> > > > > > >
> > > > > >
> > > > > > Thanks for the reminder.
> > > > > >
> > > > > > Only patch #1 is needed. It should be applied to v5.10 and later.
> > > > >
> > > > > Hold on, why did this go into mainline when I had an outstanding comment w.r.t.
> > > > > the stack unwinder?
> > > > >
> > > > > From your last reply to me there I was expecting a respin with that fixed.
> > > > >
> > > >
> > > > Apologies for the confusion.
> > > >
> > > > I have a patch for this queued up, but AIUI, that cannot be merged all
> > > > the way back to v5.10, so these need to remain separate changes in any
> > > > case.
> > > >
> > > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c2530a04a73e6b75ed71ed14d09d7b42d6300013
> > >
> > > Ah, ok, thanks for the pointer!
> > >
> > > I'm a little uneasy here, still.
> > >
> > > By backporting this we're also backporting the new breakage of the stack
> > > unwinder, and the minimal change for backports would be to add the lock and not
> > > the new stack (which was added for additinoal robustness, not to fix the bug
> > > the lock fixes).
> > >
> > > I do appreciate that the additional stack is likely more useful than the
> > > occasional diagnostic output from the kernel, but it does seem like this has
> > > traded off one bug for another, and I'm just a little annoyed because I pointed
> > > that out before the first pull request was made.
> > >
> > > I do know that this isn't malicious, and I'm not trying to start a fight, but
> > > now we have to consider whether we want/need to backport a stack unwinder fix
> > > to account for this, and we hadn't had that discussion before.
> >
> > In that case, let's drop these backports for the time being, and
> > collaborate on a solution that works for all of us.
>
> Thanks!
>
> IIUC our options here are:
>
> 1) Create a cut-down patch for stable that just adds the new lock but leaves
>    out the new stack.
>

The lock by itself does nothing useful - it is only needed because
there is now a single stack that is shared by all callers of EFI
runtime code. The existing ones are serialized already, but ACPI may
invoke efi_call_virt_pointer() as well, which is why the additional
spinlock is required for completeness. However, that ACPI feature is
relatively recent (and I am not aware of any arm64 systems that
actually implement it in their firmware)

>    I may be missing a reason why that's insufficient or painful.
>

The reason for the backport is that it also allows us to stash the
shadow call stack pointer elsewhere, as storing it on the normal stack
as we do currently defeats the purpose.

> 2) Backport this *but* also backport the follow-up fixes from your other series:
>    https://lore.kernel.org/r/20230104174433.1259428-1-ardb@kernel.org
>
>    Above you mentioned something about v5.10, was that just to say that some
>    manual backporting was required, or that there was a structural problem that
>    would require more invasive changes / prerequisites?
>

This is when the shadow call stack was introduced. Apologies for not
making this clearer in the commit log.

> 3) Something else?
>
> My preference would be (1), but if we are encountering issue with stack size on
> stable kernels, then I'd be happy to help with manual backporting effort for
> (2), as long as we backported all the relevant bits in one go.
>
> Does that make sense, and does that sound reasonable to you?
>

What I had in mind (but did not communicate clearly) is to backport
the patch that introduces the new stack to v5.10 and later, and to
backport the patch that adds the stacktrace declaration as well, but
separately (and probably not all the way back to v5.10)