On 05.06.20 13:52, Ard Biesheuvel wrote:
> EFI on ARM only supports short descriptors, and given that it mandates
> that the MMU and caches are on, it is implied that booting in HYP mode
> is not supported.
>
> However, implementations of EFI exist (i.e., U-Boot) that ignore this
> requirement, which is not entirely unreasonable, given that it does
> not make a lot of sense to begin with.
Hello Ard,
thanks for investigating the differences between EDK2 and U-Boot.
What I still do not understand is if there is a bug in U-Boot where
U-Boot does not conform to the UEFI specification? In this case I would
expect a fix in U-Boot.
Or is it simply a deficiency of Linux that it does not properly support
HYP/EL2 mode on 32-bit ARM?
Up to now I never experience a problem booting a 32bit board via U-Boot
-> GRUB-EFI -> Linux. Where did you have a problem when booting via
U-Boot's UEFI implementation and the current Linux kernel?
Does this patch relate to the retirement of KVM on 32 ARM in Linux 5.7
541ad0150ca4 ("arm: Remove 32bit KVM host support")? Without that patch
I would have expected Linux to work fine in HYP mode.
Best regards
Heinrich
>
> So let's make sure that we can deal with this condition gracefully.
> We already tolerate booting the EFI stub with the caches off (even
> though this violates the EFI spec as well), and so we should deal
> with HYP mode boot with MMU and caches either on or off.
>
> - When the MMU and caches are on, we can ignore the HYP stub altogether,
> since we can just use the existing mappings, and just branch into
> the decompressed kernel as usual after disabling MMU and caches.
>
> - When the MMU and caches are off, we have to drop to SVC mode so that
> we can set up the page tables using short descriptors. In this case,
> we need to install the HYP stub so that we can return to HYP mode
> when handing over to the kernel proper.
>
> Signed-off-by: Ard Biesheuvel <ardb@xxxxxxxxxx>
> ---
> arch/arm/boot/compressed/head.S | 29 +++++++++++++++++++++++++++++
> 1 file changed, 29 insertions(+)
>
> diff --git a/arch/arm/boot/compressed/head.S b/arch/arm/boot/compressed/head.S
> index c79db44ba128..986db86ba334 100644
> --- a/arch/arm/boot/compressed/head.S
> +++ b/arch/arm/boot/compressed/head.S
> @@ -1436,6 +1436,35 @@ ENTRY(efi_enter_kernel)
> mrc p15, 0, r0, c1, c0, 0 @ read SCTLR
> tst r0, #0x1 @ MMU enabled?
> orreq r4, r4, #1 @ set LSB if not
> +#ifdef CONFIG_ARM_VIRT_EXT
> + @
> + @ The EFI spec does not support booting on ARM in HYP mode,
> + @ since it mandates that the MMU and caches are on, with all
> + @ 32-bit addressable DRAM mapped 1:1 using short descriptors.
> + @ While the EDK2 reference implementation adheres to this,
> + @ U-Boot might decide to enter the EFI stub in HYP mode anyway,
> + @ with the MMU and caches either on or off.
> + @ In the former case, we're better off just carrying on using
> + @ the cached 1:1 mapping that the firmware provided, and pretend
> + @ that we are in SVC mode as far as the decompressor is
> + @ concerned. However, if the caches are off, we need to drop
> + @ into SVC mode now, and let the decompressor set up its cached
> + @ 1:1 mapping as usual.
> + @
> + mov r0, #SVC_MODE
> + msr spsr_cxsf, r0 @ record that we are in SVC mode
> + bne 1f @ skip HYP stub install if MMU is on
> +
> + mov r9, r4 @ preserve image base
> + bl __hyp_stub_install @ returns boot mode in r4
> + cmp r4, #HYP_MODE @ did we boot in HYP?
> + bne 1f @ skip drop to SVC if we did not
> +
> + safe_svcmode_maskall r0
> + msr spsr_cxsf, r4 @ record boot mode
> + mov r4, r9 @ restore image base
> +1:
> +#endif
>
> mov r0, r8 @ DT start
> add r1, r8, r2 @ DT end
>
