Just like the EFI reservations, ACPI tables can be stamped over
by kexec_file_load(), and the secondary kernel will be unable to
recover from such corruption.
It looks like our x86 friends have been there before, and have
recently added some infratructure that does what we need since
1a1c130ab757 ("ACPI: tables: x86: Reserve memory occupied by ACPI
tables"), though it appears to exist for different reasons.
Let's call into acpi_reserve_initial_tables() early so that
our tables are protected from the big bad kexec.
Signed-off-by: Marc Zyngier <maz@xxxxxxxxxx>
---
arch/arm64/kernel/acpi.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/arm64/kernel/acpi.c b/arch/arm64/kernel/acpi.c
index cada0b816c8a..5b5406c92ee4 100644
--- a/arch/arm64/kernel/acpi.c
+++ b/arch/arm64/kernel/acpi.c
@@ -233,6 +233,7 @@ void __init acpi_boot_table_init(void)
if (earlycon_acpi_spcr_enable)
early_init_dt_scan_chosen_stdout();
} else {
+ acpi_reserve_initial_tables();
acpi_parse_spcr(earlycon_acpi_spcr_enable, true);
if (IS_ENABLED(CONFIG_ACPI_BGRT))
acpi_table_parse(ACPI_SIG_BGRT, acpi_parse_bgrt);
--
2.29.2
_______________________________________________
kexec mailing list
kexec@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/kexec
