arm64: efi: Execute runtime services from a dedicated stack
authorArd Biesheuvel <ardb@kernel.org>
Mon, 5 Dec 2022 10:31:25 +0000 (11:31 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 24 Jan 2023 06:24:41 +0000 (07:24 +0100)
commitf75a91c82dc805af8f718ff106ec9c090234b37b
treed2f09bcfae8e1fb39b954554256609cccbb34811
parentff3b1a624380c14b81f4e51c48e404a45f047aab
arm64: efi: Execute runtime services from a dedicated stack

commit ff7a167961d1b97e0e205f245f806e564d3505e7 upstream.

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>
Cc: Lee Jones <lee@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/arm64/include/asm/efi.h
arch/arm64/kernel/efi-rt-wrapper.S
arch/arm64/kernel/efi.c