The patchset adds bits and pieces to get kexec (and crashkernel) work on
TDX guest.
The last patch implements CPU offlining according to the approved ACPI
spec change poposal[1]. It unlocks kexec with all CPUs visible in the target
kernel. It requires BIOS-side enabling. If it missing we fallback to booting
2nd kernel with single CPU.
Please review. I would be glad for any feedback.
[1] https://lore.kernel.org/all/13356251.uLZWGnKmhe@kreacher
v10:
- Rebased to current tip/master;
- Preserve CR4.MCE instead of setting it unconditionally;
- Fix build error in Hyper-V code after rebase;
- Include Ashish's patch for real;
v9:
- Rebased;
- Keep page tables that maps E820_TYPE_ACPI (Ashish);
- Ack/Reviewed/Tested-bys from Sathya, Kai, Tao;
- Minor printk() message adjustments;
v8:
- Rework serialization of around conversion memory back to private;
- Print ACPI_MADT_TYPE_MULTIPROC_WAKEUP in acpi_table_print_madt_entry();
- Drop debugfs interface to dump info on shared memory;
- Adjust comments and commit messages;
- Reviewed-bys by Baoquan, Dave and Thomas;
v7:
- Call enc_kexec_stop_conversion() and enc_kexec_unshare_mem() after shutting
down IO-APIC, lapic and hpet. It meets AMD requirements.
- Minor style changes;
- Add Acked/Reviewed-bys;
v6:
- Rebased to v6.8-rc1;
- Provide default noop callbacks from .enc_kexec_stop_conversion and
.enc_kexec_unshare_mem;
- Split off patch that introduces .enc_kexec_* callbacks;
- asm_acpi_mp_play_dead(): program CR3 directly from RSI, no MOV to RAX
required;
- Restructure how smp_ops.stop_this_cpu() hooked up in crash_nmi_callback();
- kvmclock patch got merged via KVM tree;
v5:
- Rename smp_ops.crash_play_dead to smp_ops.stop_this_cpu and use it in
stop_this_cpu();
- Split off enc_kexec_stop_conversion() from enc_kexec_unshare_mem();
- Introduce kernel_ident_mapping_free();
- Add explicit include for alternatives and stringify.
- Add barrier() after setting conversion_allowed to false;
- Mark cpu_hotplug_offline_disabled __ro_after_init;
- Print error if failed to hand over CPU to BIOS;
- Update comments and commit messages;
v4:
- Fix build for !KEXEC_CORE;
- Cleaner ATLERNATIVE use;
- Update commit messages and comments;
- Add Reviewed-bys;
v3:
- Rework acpi_mp_crash_stop_other_cpus() to avoid invoking hotplug state
machine;
- Free page tables if reset vector setup failed;
- Change asm_acpi_mp_play_dead() to pass reset vector and PGD as arguments;
- Mark acpi_mp_* variables as static and __ro_after_init;
- Use u32 for apicid;
- Disable CPU offlining if reset vector setup failed;
- Rename madt.S -> madt_playdead.S;
- Mark tdx_kexec_unshare_mem() as static;
- Rebase onto up-to-date tip/master;
- Whitespace fixes;
- Reorder patches;
- Add Reviewed-bys;
- Update comments and commit messages;
v2:
- Rework how unsharing hook ups into kexec codepath;
- Rework kvmclock_disable() fix based on Sean's;
- s/cpu_hotplug_not_supported()/cpu_hotplug_disable_offlining()/;
- use play_dead_common() to implement acpi_mp_play_dead();
- cond_resched() in tdx_shared_memory_show();
- s/target kernel/second kernel/;
- Update commit messages and comments;
Ashish Kalra (1):
x86/mm: Do not zap page table entries mapping unaccepted memory table
during kdump.
Kirill A. Shutemov (17):
x86/acpi: Extract ACPI MADT wakeup code into a separate file
x86/apic: Mark acpi_mp_wake_* variables as __ro_after_init
cpu/hotplug: Add support for declaring CPU offlining not supported
cpu/hotplug, x86/acpi: Disable CPU offlining for ACPI MADT wakeup
x86/kexec: Keep CR4.MCE set during kexec for TDX guest
x86/mm: Make x86_platform.guest.enc_status_change_*() return errno
x86/mm: Return correct level from lookup_address() if pte is none
x86/tdx: Account shared memory
x86/mm: Adding callbacks to prepare encrypted memory for kexec
x86/tdx: Convert shared memory back to private on kexec
x86/mm: Make e820_end_ram_pfn() cover E820_TYPE_ACPI ranges
x86/acpi: Rename fields in acpi_madt_multiproc_wakeup structure
x86/acpi: Do not attempt to bring up secondary CPUs in kexec case
x86/smp: Add smp_ops.stop_this_cpu() callback
x86/mm: Introduce kernel_ident_mapping_free()
x86/acpi: Add support for CPU offlining for ACPI MADT wakeup method
ACPI: tables: Print MULTIPROC_WAKEUP when MADT is parsed
arch/x86/Kconfig | 7 +
arch/x86/coco/core.c | 1 -
arch/x86/coco/tdx/tdx.c | 99 ++++++++-
arch/x86/hyperv/ivm.c | 22 +-
arch/x86/include/asm/acpi.h | 7 +
arch/x86/include/asm/init.h | 3 +
arch/x86/include/asm/pgtable.h | 5 +
arch/x86/include/asm/pgtable_types.h | 1 +
arch/x86/include/asm/set_memory.h | 3 +
arch/x86/include/asm/smp.h | 1 +
arch/x86/include/asm/x86_init.h | 6 +-
arch/x86/kernel/acpi/Makefile | 11 +-
arch/x86/kernel/acpi/boot.c | 86 +-------
arch/x86/kernel/acpi/madt_playdead.S | 28 +++
arch/x86/kernel/acpi/madt_wakeup.c | 292 +++++++++++++++++++++++++++
arch/x86/kernel/crash.c | 6 +
arch/x86/kernel/e820.c | 9 +-
arch/x86/kernel/process.c | 7 +
arch/x86/kernel/reboot.c | 18 ++
arch/x86/kernel/relocate_kernel_64.S | 8 +
arch/x86/kernel/x86_init.c | 8 +-
arch/x86/mm/ident_map.c | 73 +++++++
arch/x86/mm/init_64.c | 16 +-
arch/x86/mm/mem_encrypt_amd.c | 8 +-
arch/x86/mm/pat/set_memory.c | 59 ++++--
drivers/acpi/tables.c | 14 ++
include/acpi/actbl2.h | 19 +-
include/linux/cc_platform.h | 10 -
include/linux/cpu.h | 2 +
kernel/cpu.c | 12 +-
30 files changed, 685 insertions(+), 156 deletions(-)
create mode 100644 arch/x86/kernel/acpi/madt_playdead.S
create mode 100644 arch/x86/kernel/acpi/madt_wakeup.c
--
2.43.0
The ACPI MADT mailbox wakeup method doesn't allow to offline CPU after
it got woke up.
Currently offlining hotplug is prevented based on the confidential
computing attribute which is set for Intel TDX. But TDX is not
the only possible user of the wake up method. The MADT wakeup can be
implemented outside of a confidential computing environment. Offline
support is a property of the wakeup method, not the CoCo implementation.
Introduce cpu_hotplug_disable_offlining() that can be called to indicate
that CPU offlining should be disabled.
This function is going to replace CC_ATTR_HOTPLUG_DISABLED for ACPI
MADT wakeup method.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
include/linux/cpu.h | 2 ++
kernel/cpu.c | 13 ++++++++++++-
2 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/include/linux/cpu.h b/include/linux/cpu.h
index 272e4e79e15c..cfe29e52ce84 100644
--- a/include/linux/cpu.h
+++ b/include/linux/cpu.h
@@ -141,6 +141,7 @@ extern void cpus_read_lock(void);
extern void cpus_read_unlock(void);
extern int cpus_read_trylock(void);
extern void lockdep_assert_cpus_held(void);
+extern void cpu_hotplug_disable_offlining(void);
extern void cpu_hotplug_disable(void);
extern void cpu_hotplug_enable(void);
void clear_tasks_mm_cpumask(int cpu);
@@ -156,6 +157,7 @@ static inline void cpus_read_lock(void) { }
static inline void cpus_read_unlock(void) { }
static inline int cpus_read_trylock(void) { return true; }
static inline void lockdep_assert_cpus_held(void) { }
+static inline void cpu_hotplug_disable_offlining(void) { }
static inline void cpu_hotplug_disable(void) { }
static inline void cpu_hotplug_enable(void) { }
static inline int remove_cpu(unsigned int cpu) { return -EPERM; }
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 8f6affd051f7..08860baa6ce0 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -483,6 +483,8 @@ static int cpu_hotplug_disabled;
DEFINE_STATIC_PERCPU_RWSEM(cpu_hotplug_lock);
+static bool cpu_hotplug_offline_disabled __ro_after_init;
+
void cpus_read_lock(void)
{
percpu_down_read(&cpu_hotplug_lock);
@@ -542,6 +544,14 @@ static void lockdep_release_cpus_lock(void)
rwsem_release(&cpu_hotplug_lock.dep_map, _THIS_IP_);
}
+/* Declare CPU offlining not supported */
+void cpu_hotplug_disable_offlining(void)
+{
+ cpu_maps_update_begin();
+ cpu_hotplug_offline_disabled = true;
+ cpu_maps_update_done();
+}
+
/*
* Wait for currently running CPU hotplug operations to complete (if any) and
* disable future CPU hotplug (from sysfs). The 'cpu_add_remove_lock' protects
@@ -1518,7 +1528,8 @@ static int cpu_down_maps_locked(unsigned int cpu, enum cpuhp_state target)
* If the platform does not support hotplug, report it explicitly to
* differentiate it from a transient offlining failure.
*/
- if (cc_platform_has(CC_ATTR_HOTPLUG_DISABLED))
+ if (cc_platform_has(CC_ATTR_HOTPLUG_DISABLED) ||
+ cpu_hotplug_offline_disabled)
return -EOPNOTSUPP;
if (cpu_hotplug_disabled)
return -EBUSY;
--
2.43.0
In order to prepare for the expansion of support for the ACPI MADT
wakeup method, move the relevant code into a separate file.
Introduce a new configuration option to clearly indicate dependencies
without the use of ifdefs.
There have been no functional changes.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
Acked-by: Kai Huang <[email protected]>
Reviewed-by: Baoquan He <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/Kconfig | 7 +++
arch/x86/include/asm/acpi.h | 5 ++
arch/x86/kernel/acpi/Makefile | 11 ++--
arch/x86/kernel/acpi/boot.c | 86 +-----------------------------
arch/x86/kernel/acpi/madt_wakeup.c | 82 ++++++++++++++++++++++++++++
5 files changed, 101 insertions(+), 90 deletions(-)
create mode 100644 arch/x86/kernel/acpi/madt_wakeup.c
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 9c637f5b2c76..50ea71adb098 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -1116,6 +1116,13 @@ config X86_LOCAL_APIC
depends on X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_APIC || PCI_MSI
select IRQ_DOMAIN_HIERARCHY
+config X86_ACPI_MADT_WAKEUP
+ def_bool y
+ depends on X86_64
+ depends on ACPI
+ depends on SMP
+ depends on X86_LOCAL_APIC
+
config X86_IO_APIC
def_bool y
depends on X86_LOCAL_APIC || X86_UP_IOAPIC
diff --git a/arch/x86/include/asm/acpi.h b/arch/x86/include/asm/acpi.h
index f896eed4516c..2625b915ae7f 100644
--- a/arch/x86/include/asm/acpi.h
+++ b/arch/x86/include/asm/acpi.h
@@ -76,6 +76,11 @@ static inline bool acpi_skip_set_wakeup_address(void)
#define acpi_skip_set_wakeup_address acpi_skip_set_wakeup_address
+union acpi_subtable_headers;
+
+int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
+ const unsigned long end);
+
/*
* Check if the CPU can handle C2 and deeper
*/
diff --git a/arch/x86/kernel/acpi/Makefile b/arch/x86/kernel/acpi/Makefile
index fc17b3f136fe..8c7329c88a75 100644
--- a/arch/x86/kernel/acpi/Makefile
+++ b/arch/x86/kernel/acpi/Makefile
@@ -1,11 +1,12 @@
# SPDX-License-Identifier: GPL-2.0
-obj-$(CONFIG_ACPI) += boot.o
-obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
-obj-$(CONFIG_ACPI_APEI) += apei.o
-obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
+obj-$(CONFIG_ACPI) += boot.o
+obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
+obj-$(CONFIG_ACPI_APEI) += apei.o
+obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
+obj-$(CONFIG_X86_ACPI_MADT_WAKEUP) += madt_wakeup.o
ifneq ($(CONFIG_ACPI_PROCESSOR),)
-obj-y += cstate.o
+obj-y += cstate.o
endif
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 4bf82dbd2a6b..53b8802e01e7 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -67,13 +67,6 @@ static bool has_lapic_cpus __initdata;
static bool acpi_support_online_capable;
#endif
-#ifdef CONFIG_X86_64
-/* Physical address of the Multiprocessor Wakeup Structure mailbox */
-static u64 acpi_mp_wake_mailbox_paddr;
-/* Virtual address of the Multiprocessor Wakeup Structure mailbox */
-static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox;
-#endif
-
#ifdef CONFIG_X86_IO_APIC
/*
* Locks related to IOAPIC hotplug
@@ -341,60 +334,6 @@ acpi_parse_lapic_nmi(union acpi_subtable_headers * header, const unsigned long e
return 0;
}
-
-#ifdef CONFIG_X86_64
-static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
-{
- /*
- * Remap mailbox memory only for the first call to acpi_wakeup_cpu().
- *
- * Wakeup of secondary CPUs is fully serialized in the core code.
- * No need to protect acpi_mp_wake_mailbox from concurrent accesses.
- */
- if (!acpi_mp_wake_mailbox) {
- acpi_mp_wake_mailbox = memremap(acpi_mp_wake_mailbox_paddr,
- sizeof(*acpi_mp_wake_mailbox),
- MEMREMAP_WB);
- }
-
- /*
- * Mailbox memory is shared between the firmware and OS. Firmware will
- * listen on mailbox command address, and once it receives the wakeup
- * command, the CPU associated with the given apicid will be booted.
- *
- * The value of 'apic_id' and 'wakeup_vector' must be visible to the
- * firmware before the wakeup command is visible. smp_store_release()
- * ensures ordering and visibility.
- */
- acpi_mp_wake_mailbox->apic_id = apicid;
- acpi_mp_wake_mailbox->wakeup_vector = start_ip;
- smp_store_release(&acpi_mp_wake_mailbox->command,
- ACPI_MP_WAKE_COMMAND_WAKEUP);
-
- /*
- * Wait for the CPU to wake up.
- *
- * The CPU being woken up is essentially in a spin loop waiting to be
- * woken up. It should not take long for it wake up and acknowledge by
- * zeroing out ->command.
- *
- * ACPI specification doesn't provide any guidance on how long kernel
- * has to wait for a wake up acknowledgement. It also doesn't provide
- * a way to cancel a wake up request if it takes too long.
- *
- * In TDX environment, the VMM has control over how long it takes to
- * wake up secondary. It can postpone scheduling secondary vCPU
- * indefinitely. Giving up on wake up request and reporting error opens
- * possible attack vector for VMM: it can wake up a secondary CPU when
- * kernel doesn't expect it. Wait until positive result of the wake up
- * request.
- */
- while (READ_ONCE(acpi_mp_wake_mailbox->command))
- cpu_relax();
-
- return 0;
-}
-#endif /* CONFIG_X86_64 */
#endif /* CONFIG_X86_LOCAL_APIC */
#ifdef CONFIG_X86_IO_APIC
@@ -1124,29 +1063,6 @@ static int __init acpi_parse_madt_lapic_entries(void)
}
return 0;
}
-
-#ifdef CONFIG_X86_64
-static int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
- const unsigned long end)
-{
- struct acpi_madt_multiproc_wakeup *mp_wake;
-
- if (!IS_ENABLED(CONFIG_SMP))
- return -ENODEV;
-
- mp_wake = (struct acpi_madt_multiproc_wakeup *)header;
- if (BAD_MADT_ENTRY(mp_wake, end))
- return -EINVAL;
-
- acpi_table_print_madt_entry(&header->common);
-
- acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
-
- apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
-
- return 0;
-}
-#endif /* CONFIG_X86_64 */
#endif /* CONFIG_X86_LOCAL_APIC */
#ifdef CONFIG_X86_IO_APIC
@@ -1343,7 +1259,7 @@ static void __init acpi_process_madt(void)
smp_found_config = 1;
}
-#ifdef CONFIG_X86_64
+#ifdef CONFIG_X86_ACPI_MADT_WAKEUP
/*
* Parse MADT MP Wake entry.
*/
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
new file mode 100644
index 000000000000..7f164d38bd0b
--- /dev/null
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -0,0 +1,82 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+#include <linux/acpi.h>
+#include <linux/io.h>
+#include <asm/apic.h>
+#include <asm/barrier.h>
+#include <asm/processor.h>
+
+/* Physical address of the Multiprocessor Wakeup Structure mailbox */
+static u64 acpi_mp_wake_mailbox_paddr;
+
+/* Virtual address of the Multiprocessor Wakeup Structure mailbox */
+static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox;
+
+static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
+{
+ /*
+ * Remap mailbox memory only for the first call to acpi_wakeup_cpu().
+ *
+ * Wakeup of secondary CPUs is fully serialized in the core code.
+ * No need to protect acpi_mp_wake_mailbox from concurrent accesses.
+ */
+ if (!acpi_mp_wake_mailbox) {
+ acpi_mp_wake_mailbox = memremap(acpi_mp_wake_mailbox_paddr,
+ sizeof(*acpi_mp_wake_mailbox),
+ MEMREMAP_WB);
+ }
+
+ /*
+ * Mailbox memory is shared between the firmware and OS. Firmware will
+ * listen on mailbox command address, and once it receives the wakeup
+ * command, the CPU associated with the given apicid will be booted.
+ *
+ * The value of 'apic_id' and 'wakeup_vector' must be visible to the
+ * firmware before the wakeup command is visible. smp_store_release()
+ * ensures ordering and visibility.
+ */
+ acpi_mp_wake_mailbox->apic_id = apicid;
+ acpi_mp_wake_mailbox->wakeup_vector = start_ip;
+ smp_store_release(&acpi_mp_wake_mailbox->command,
+ ACPI_MP_WAKE_COMMAND_WAKEUP);
+
+ /*
+ * Wait for the CPU to wake up.
+ *
+ * The CPU being woken up is essentially in a spin loop waiting to be
+ * woken up. It should not take long for it wake up and acknowledge by
+ * zeroing out ->command.
+ *
+ * ACPI specification doesn't provide any guidance on how long kernel
+ * has to wait for a wake up acknowledgment. It also doesn't provide
+ * a way to cancel a wake up request if it takes too long.
+ *
+ * In TDX environment, the VMM has control over how long it takes to
+ * wake up secondary. It can postpone scheduling secondary vCPU
+ * indefinitely. Giving up on wake up request and reporting error opens
+ * possible attack vector for VMM: it can wake up a secondary CPU when
+ * kernel doesn't expect it. Wait until positive result of the wake up
+ * request.
+ */
+ while (READ_ONCE(acpi_mp_wake_mailbox->command))
+ cpu_relax();
+
+ return 0;
+}
+
+int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
+ const unsigned long end)
+{
+ struct acpi_madt_multiproc_wakeup *mp_wake;
+
+ mp_wake = (struct acpi_madt_multiproc_wakeup *)header;
+ if (BAD_MADT_ENTRY(mp_wake, end))
+ return -EINVAL;
+
+ acpi_table_print_madt_entry(&header->common);
+
+ acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
+
+ apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
+
+ return 0;
+}
--
2.43.0
ACPI MADT doesn't allow to offline CPU after it got woke up.
Currently CPU hotplug is prevented based on the confidential computing
attribute which is set for Intel TDX. But TDX is not the only possible
user of the wake up method.
Disable CPU offlining on ACPI MADT wakeup enumeration.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/coco/core.c | 1 -
arch/x86/kernel/acpi/madt_wakeup.c | 3 +++
include/linux/cc_platform.h | 10 ----------
kernel/cpu.c | 3 +--
4 files changed, 4 insertions(+), 13 deletions(-)
diff --git a/arch/x86/coco/core.c b/arch/x86/coco/core.c
index b31ef2424d19..0f81f70aca82 100644
--- a/arch/x86/coco/core.c
+++ b/arch/x86/coco/core.c
@@ -29,7 +29,6 @@ static bool noinstr intel_cc_platform_has(enum cc_attr attr)
{
switch (attr) {
case CC_ATTR_GUEST_UNROLL_STRING_IO:
- case CC_ATTR_HOTPLUG_DISABLED:
case CC_ATTR_GUEST_MEM_ENCRYPT:
case CC_ATTR_MEM_ENCRYPT:
return true;
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
index cf79ea6f3007..d222be8d7a07 100644
--- a/arch/x86/kernel/acpi/madt_wakeup.c
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -1,5 +1,6 @@
// SPDX-License-Identifier: GPL-2.0-or-later
#include <linux/acpi.h>
+#include <linux/cpu.h>
#include <linux/io.h>
#include <asm/apic.h>
#include <asm/barrier.h>
@@ -76,6 +77,8 @@ int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
+ cpu_hotplug_disable_offlining();
+
apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
return 0;
diff --git a/include/linux/cc_platform.h b/include/linux/cc_platform.h
index 60693a145894..caa4b4430634 100644
--- a/include/linux/cc_platform.h
+++ b/include/linux/cc_platform.h
@@ -81,16 +81,6 @@ enum cc_attr {
*/
CC_ATTR_GUEST_SEV_SNP,
- /**
- * @CC_ATTR_HOTPLUG_DISABLED: Hotplug is not supported or disabled.
- *
- * The platform/OS is running as a guest/virtual machine does not
- * support CPU hotplug feature.
- *
- * Examples include TDX Guest.
- */
- CC_ATTR_HOTPLUG_DISABLED,
-
/**
* @CC_ATTR_HOST_SEV_SNP: AMD SNP enabled on the host.
*
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 08860baa6ce0..a70767aee9d0 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -1528,8 +1528,7 @@ static int cpu_down_maps_locked(unsigned int cpu, enum cpuhp_state target)
* If the platform does not support hotplug, report it explicitly to
* differentiate it from a transient offlining failure.
*/
- if (cc_platform_has(CC_ATTR_HOTPLUG_DISABLED) ||
- cpu_hotplug_offline_disabled)
+ if (cpu_hotplug_offline_disabled)
return -EOPNOTSUPP;
if (cpu_hotplug_disabled)
return -EBUSY;
--
2.43.0
Depending on setup, TDX guests might be allowed to clear CR4.MCE.
Attempt to clear it leads to #VE.
Use alternatives to keep the flag during kexec for TDX guests.
The change doesn't affect non-TDX-guest environments.
Signed-off-by: Kirill A. Shutemov <[email protected]>
---
arch/x86/kernel/relocate_kernel_64.S | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
index 56cab1bb25f5..8e2037d78a1f 100644
--- a/arch/x86/kernel/relocate_kernel_64.S
+++ b/arch/x86/kernel/relocate_kernel_64.S
@@ -5,6 +5,8 @@
*/
#include <linux/linkage.h>
+#include <linux/stringify.h>
+#include <asm/alternative.h>
#include <asm/page_types.h>
#include <asm/kexec.h>
#include <asm/processor-flags.h>
@@ -145,11 +147,17 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
* Set cr4 to a known state:
* - physical address extension enabled
* - 5-level paging, if it was enabled before
+ * - Machine check exception on TDX guest, if it was enabled before.
+ * Clearing MCE might not allowed in TDX guests, depending on setup.
*/
movl $X86_CR4_PAE, %eax
testq $X86_CR4_LA57, %r13
jz 1f
orl $X86_CR4_LA57, %eax
+1:
+ testq $X86_CR4_MCE, %r13
+ jz 1f
+ ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
1:
movq %rax, %cr4
--
2.43.0
Currently, lookup_address() returns two things:
1. A "pte_t" (which might be a p[g4um]d_t)
2. The 'level' of the page tables where the "pte_t" was found
(returned via a pointer)
If no pte_t is found, 'level' is essentially garbage.
Always fill out the level. For NULL "pte_t"s, fill in the level where
the p*d_none() entry was found mirroring the "found" behavior.
Always filling out the level allows using lookup_address() to precisely
skip over holes when walking kernel page tables.
Add one more entry into enum pg_level to indicate the size of the VA
covered by one PGD entry in 5-level paging mode.
Update comments for lookup_address() and lookup_address_in_pgd() to
reflect changes in the interface.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Rick Edgecombe <[email protected]>
Reviewed-by: Baoquan He <[email protected]>
Reviewed-by: Dave Hansen <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/include/asm/pgtable_types.h | 1 +
arch/x86/mm/pat/set_memory.c | 16 ++++++++--------
2 files changed, 9 insertions(+), 8 deletions(-)
diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/pgtable_types.h
index 0b748ee16b3d..3f648ffdfbe5 100644
--- a/arch/x86/include/asm/pgtable_types.h
+++ b/arch/x86/include/asm/pgtable_types.h
@@ -548,6 +548,7 @@ enum pg_level {
PG_LEVEL_2M,
PG_LEVEL_1G,
PG_LEVEL_512G,
+ PG_LEVEL_256T,
PG_LEVEL_NUM
};
diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index e5b454036bf3..6c49f69c0368 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -657,7 +657,8 @@ static inline pgprot_t verify_rwx(pgprot_t old, pgprot_t new, unsigned long star
/*
* Lookup the page table entry for a virtual address in a specific pgd.
- * Return a pointer to the entry and the level of the mapping.
+ * Return a pointer to the entry (or NULL if the entry does not exist) and
+ * the level of the entry.
*/
pte_t *lookup_address_in_pgd(pgd_t *pgd, unsigned long address,
unsigned int *level)
@@ -666,32 +667,32 @@ pte_t *lookup_address_in_pgd(pgd_t *pgd, unsigned long address,
pud_t *pud;
pmd_t *pmd;
- *level = PG_LEVEL_NONE;
+ *level = PG_LEVEL_256T;
if (pgd_none(*pgd))
return NULL;
+ *level = PG_LEVEL_512G;
p4d = p4d_offset(pgd, address);
if (p4d_none(*p4d))
return NULL;
- *level = PG_LEVEL_512G;
if (p4d_leaf(*p4d) || !p4d_present(*p4d))
return (pte_t *)p4d;
+ *level = PG_LEVEL_1G;
pud = pud_offset(p4d, address);
if (pud_none(*pud))
return NULL;
- *level = PG_LEVEL_1G;
if (pud_leaf(*pud) || !pud_present(*pud))
return (pte_t *)pud;
+ *level = PG_LEVEL_2M;
pmd = pmd_offset(pud, address);
if (pmd_none(*pmd))
return NULL;
- *level = PG_LEVEL_2M;
if (pmd_leaf(*pmd) || !pmd_present(*pmd))
return (pte_t *)pmd;
@@ -704,9 +705,8 @@ pte_t *lookup_address_in_pgd(pgd_t *pgd, unsigned long address,
* Lookup the page table entry for a virtual address. Return a pointer
* to the entry and the level of the mapping.
*
- * Note: We return pud and pmd either when the entry is marked large
- * or when the present bit is not set. Otherwise we would return a
- * pointer to a nonexisting mapping.
+ * Note: the function returns p4d, pud or pmd either when the entry is marked
+ * large or when the present bit is not set. Otherwise it returns NULL.
*/
pte_t *lookup_address(unsigned long address, unsigned int *level)
{
--
2.43.0
TDX is going to have more than one reason to fail
enc_status_change_prepare().
Change the callback to return errno instead of assuming -EIO;
enc_status_change_finish() changed too to keep the interface symmetric.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Dave Hansen <[email protected]>
Reviewed-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/coco/tdx/tdx.c | 20 +++++++++++---------
arch/x86/hyperv/ivm.c | 22 ++++++++++------------
arch/x86/include/asm/x86_init.h | 4 ++--
arch/x86/kernel/x86_init.c | 4 ++--
arch/x86/mm/mem_encrypt_amd.c | 8 ++++----
arch/x86/mm/pat/set_memory.c | 8 +++++---
6 files changed, 34 insertions(+), 32 deletions(-)
diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c
index c1cb90369915..26fa47db5782 100644
--- a/arch/x86/coco/tdx/tdx.c
+++ b/arch/x86/coco/tdx/tdx.c
@@ -798,28 +798,30 @@ static bool tdx_enc_status_changed(unsigned long vaddr, int numpages, bool enc)
return true;
}
-static bool tdx_enc_status_change_prepare(unsigned long vaddr, int numpages,
- bool enc)
+static int tdx_enc_status_change_prepare(unsigned long vaddr, int numpages,
+ bool enc)
{
/*
* Only handle shared->private conversion here.
* See the comment in tdx_early_init().
*/
- if (enc)
- return tdx_enc_status_changed(vaddr, numpages, enc);
- return true;
+ if (enc && !tdx_enc_status_changed(vaddr, numpages, enc))
+ return -EIO;
+
+ return 0;
}
-static bool tdx_enc_status_change_finish(unsigned long vaddr, int numpages,
+static int tdx_enc_status_change_finish(unsigned long vaddr, int numpages,
bool enc)
{
/*
* Only handle private->shared conversion here.
* See the comment in tdx_early_init().
*/
- if (!enc)
- return tdx_enc_status_changed(vaddr, numpages, enc);
- return true;
+ if (!enc && !tdx_enc_status_changed(vaddr, numpages, enc))
+ return -EIO;
+
+ return 0;
}
void __init tdx_early_init(void)
diff --git a/arch/x86/hyperv/ivm.c b/arch/x86/hyperv/ivm.c
index 768d73de0d09..b4a851d27c7c 100644
--- a/arch/x86/hyperv/ivm.c
+++ b/arch/x86/hyperv/ivm.c
@@ -523,9 +523,9 @@ static int hv_mark_gpa_visibility(u16 count, const u64 pfn[],
* transition is complete, hv_vtom_set_host_visibility() marks the pages
* as "present" again.
*/
-static bool hv_vtom_clear_present(unsigned long kbuffer, int pagecount, bool enc)
+static int hv_vtom_clear_present(unsigned long kbuffer, int pagecount, bool enc)
{
- return !set_memory_np(kbuffer, pagecount);
+ return set_memory_np(kbuffer, pagecount);
}
/*
@@ -536,20 +536,19 @@ static bool hv_vtom_clear_present(unsigned long kbuffer, int pagecount, bool enc
* with host. This function works as wrap of hv_mark_gpa_visibility()
* with memory base and size.
*/
-static bool hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bool enc)
+static int hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bool enc)
{
enum hv_mem_host_visibility visibility = enc ?
VMBUS_PAGE_NOT_VISIBLE : VMBUS_PAGE_VISIBLE_READ_WRITE;
u64 *pfn_array;
phys_addr_t paddr;
+ int i, pfn, err;
void *vaddr;
int ret = 0;
- bool result = true;
- int i, pfn;
pfn_array = kmalloc(HV_HYP_PAGE_SIZE, GFP_KERNEL);
if (!pfn_array) {
- result = false;
+ ret = -ENOMEM;
goto err_set_memory_p;
}
@@ -568,10 +567,8 @@ static bool hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bo
if (pfn == HV_MAX_MODIFY_GPA_REP_COUNT || i == pagecount - 1) {
ret = hv_mark_gpa_visibility(pfn, pfn_array,
visibility);
- if (ret) {
- result = false;
+ if (ret)
goto err_free_pfn_array;
- }
pfn = 0;
}
}
@@ -586,10 +583,11 @@ static bool hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bo
* order to avoid leaving the memory range in a "broken" state. Setting
* the PRESENT bits shouldn't fail, but return an error if it does.
*/
- if (set_memory_p(kbuffer, pagecount))
- result = false;
+ err = set_memory_p(kbuffer, pagecount);
+ if (err && !ret)
+ ret = err;
- return result;
+ return ret;
}
static bool hv_vtom_tlb_flush_required(bool private)
diff --git a/arch/x86/include/asm/x86_init.h b/arch/x86/include/asm/x86_init.h
index 6149eabe200f..28ac3cb9b987 100644
--- a/arch/x86/include/asm/x86_init.h
+++ b/arch/x86/include/asm/x86_init.h
@@ -151,8 +151,8 @@ struct x86_init_acpi {
* @enc_cache_flush_required Returns true if a cache flush is needed before changing page encryption status
*/
struct x86_guest {
- bool (*enc_status_change_prepare)(unsigned long vaddr, int npages, bool enc);
- bool (*enc_status_change_finish)(unsigned long vaddr, int npages, bool enc);
+ int (*enc_status_change_prepare)(unsigned long vaddr, int npages, bool enc);
+ int (*enc_status_change_finish)(unsigned long vaddr, int npages, bool enc);
bool (*enc_tlb_flush_required)(bool enc);
bool (*enc_cache_flush_required)(void);
};
diff --git a/arch/x86/kernel/x86_init.c b/arch/x86/kernel/x86_init.c
index d5dc5a92635a..a7143bb7dd93 100644
--- a/arch/x86/kernel/x86_init.c
+++ b/arch/x86/kernel/x86_init.c
@@ -134,8 +134,8 @@ struct x86_cpuinit_ops x86_cpuinit = {
static void default_nmi_init(void) { };
-static bool enc_status_change_prepare_noop(unsigned long vaddr, int npages, bool enc) { return true; }
-static bool enc_status_change_finish_noop(unsigned long vaddr, int npages, bool enc) { return true; }
+static int enc_status_change_prepare_noop(unsigned long vaddr, int npages, bool enc) { return 0; }
+static int enc_status_change_finish_noop(unsigned long vaddr, int npages, bool enc) { return 0; }
static bool enc_tlb_flush_required_noop(bool enc) { return false; }
static bool enc_cache_flush_required_noop(void) { return false; }
static bool is_private_mmio_noop(u64 addr) {return false; }
diff --git a/arch/x86/mm/mem_encrypt_amd.c b/arch/x86/mm/mem_encrypt_amd.c
index 422602f6039b..e7b67519ddb5 100644
--- a/arch/x86/mm/mem_encrypt_amd.c
+++ b/arch/x86/mm/mem_encrypt_amd.c
@@ -283,7 +283,7 @@ static void enc_dec_hypercall(unsigned long vaddr, unsigned long size, bool enc)
#endif
}
-static bool amd_enc_status_change_prepare(unsigned long vaddr, int npages, bool enc)
+static int amd_enc_status_change_prepare(unsigned long vaddr, int npages, bool enc)
{
/*
* To maintain the security guarantees of SEV-SNP guests, make sure
@@ -292,11 +292,11 @@ static bool amd_enc_status_change_prepare(unsigned long vaddr, int npages, bool
if (cc_platform_has(CC_ATTR_GUEST_SEV_SNP) && !enc)
snp_set_memory_shared(vaddr, npages);
- return true;
+ return 0;
}
/* Return true unconditionally: return value doesn't matter for the SEV side */
-static bool amd_enc_status_change_finish(unsigned long vaddr, int npages, bool enc)
+static int amd_enc_status_change_finish(unsigned long vaddr, int npages, bool enc)
{
/*
* After memory is mapped encrypted in the page table, validate it
@@ -308,7 +308,7 @@ static bool amd_enc_status_change_finish(unsigned long vaddr, int npages, bool e
if (!cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT))
enc_dec_hypercall(vaddr, npages << PAGE_SHIFT, enc);
- return true;
+ return 0;
}
static void __init __set_clr_pte_enc(pte_t *kpte, int level, bool enc)
diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index 80c9037ffadf..e5b454036bf3 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -2156,7 +2156,8 @@ static int __set_memory_enc_pgtable(unsigned long addr, int numpages, bool enc)
cpa_flush(&cpa, x86_platform.guest.enc_cache_flush_required());
/* Notify hypervisor that we are about to set/clr encryption attribute. */
- if (!x86_platform.guest.enc_status_change_prepare(addr, numpages, enc))
+ ret = x86_platform.guest.enc_status_change_prepare(addr, numpages, enc);
+ if (ret)
goto vmm_fail;
ret = __change_page_attr_set_clr(&cpa, 1);
@@ -2174,7 +2175,8 @@ static int __set_memory_enc_pgtable(unsigned long addr, int numpages, bool enc)
return ret;
/* Notify hypervisor that we have successfully set/clr encryption attribute. */
- if (!x86_platform.guest.enc_status_change_finish(addr, numpages, enc))
+ ret = x86_platform.guest.enc_status_change_finish(addr, numpages, enc);
+ if (ret)
goto vmm_fail;
return 0;
@@ -2183,7 +2185,7 @@ static int __set_memory_enc_pgtable(unsigned long addr, int numpages, bool enc)
WARN_ONCE(1, "CPA VMM failure to convert memory (addr=%p, numpages=%d) to %s.\n",
(void *)addr, numpages, enc ? "private" : "shared");
- return -EIO;
+ return ret;
}
static int __set_memory_enc_dec(unsigned long addr, int numpages, bool enc)
--
2.43.0
From: Ashish Kalra <[email protected]>
During crashkernel boot only pre-allocated crash memory is presented as
E820_TYPE_RAM. This can cause page table entries mapping unaccepted memory
table to be zapped during phys_pte_init(), phys_pmd_init(), phys_pud_init()
and phys_p4d_init() as SNP/TDX guest use E820_TYPE_ACPI to store the
unaccepted memory table and pass it between the kernels on
kexec/kdump.
E820_TYPE_ACPI covers not only ACPI data, but also EFI tables and might
be required by kernel to function properly.
The problem was discovered during debugging kdump for SNP guest. The
unaccepted memory table stored with E820_TYPE_ACPI and passed between
the kernels on kdump was getting zapped as the PMD entry mapping this
is above the E820_TYPE_RAM range for the reserved crashkernel memory.
Signed-off-by: Ashish Kalra <[email protected]>
Signed-off-by: Kirill A. Shutemov <[email protected]>
---
arch/x86/mm/init_64.c | 16 ++++++++++++----
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 7e177856ee4f..28002cc7a37d 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -469,7 +469,9 @@ phys_pte_init(pte_t *pte_page, unsigned long paddr, unsigned long paddr_end,
!e820__mapped_any(paddr & PAGE_MASK, paddr_next,
E820_TYPE_RAM) &&
!e820__mapped_any(paddr & PAGE_MASK, paddr_next,
- E820_TYPE_RESERVED_KERN))
+ E820_TYPE_RESERVED_KERN) &&
+ !e820__mapped_any(paddr & PAGE_MASK, paddr_next,
+ E820_TYPE_ACPI))
set_pte_init(pte, __pte(0), init);
continue;
}
@@ -524,7 +526,9 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long paddr, unsigned long paddr_end,
!e820__mapped_any(paddr & PMD_MASK, paddr_next,
E820_TYPE_RAM) &&
!e820__mapped_any(paddr & PMD_MASK, paddr_next,
- E820_TYPE_RESERVED_KERN))
+ E820_TYPE_RESERVED_KERN) &&
+ !e820__mapped_any(paddr & PMD_MASK, paddr_next,
+ E820_TYPE_ACPI))
set_pmd_init(pmd, __pmd(0), init);
continue;
}
@@ -611,7 +615,9 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
!e820__mapped_any(paddr & PUD_MASK, paddr_next,
E820_TYPE_RAM) &&
!e820__mapped_any(paddr & PUD_MASK, paddr_next,
- E820_TYPE_RESERVED_KERN))
+ E820_TYPE_RESERVED_KERN) &&
+ !e820__mapped_any(paddr & PUD_MASK, paddr_next,
+ E820_TYPE_ACPI))
set_pud_init(pud, __pud(0), init);
continue;
}
@@ -698,7 +704,9 @@ phys_p4d_init(p4d_t *p4d_page, unsigned long paddr, unsigned long paddr_end,
!e820__mapped_any(paddr & P4D_MASK, paddr_next,
E820_TYPE_RAM) &&
!e820__mapped_any(paddr & P4D_MASK, paddr_next,
- E820_TYPE_RESERVED_KERN))
+ E820_TYPE_RESERVED_KERN) &&
+ !e820__mapped_any(paddr & P4D_MASK, paddr_next,
+ E820_TYPE_ACPI))
set_p4d_init(p4d, __p4d(0), init);
continue;
}
--
2.43.0
The kernel will convert all shared memory back to private during kexec.
The direct mapping page tables will provide information on which memory
is shared.
It is extremely important to convert all shared memory. If a page is
missed, it will cause the second kernel to crash when it accesses it.
Keep track of the number of shared pages. This will allow for
cross-checking against the shared information in the direct mapping and
reporting if the shared bit is lost.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/coco/tdx/tdx.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c
index 26fa47db5782..979891e97d83 100644
--- a/arch/x86/coco/tdx/tdx.c
+++ b/arch/x86/coco/tdx/tdx.c
@@ -38,6 +38,8 @@
#define TDREPORT_SUBTYPE_0 0
+static atomic_long_t nr_shared;
+
/* Called from __tdx_hypercall() for unrecoverable failure */
noinstr void __noreturn __tdx_hypercall_failed(void)
{
@@ -821,6 +823,11 @@ static int tdx_enc_status_change_finish(unsigned long vaddr, int numpages,
if (!enc && !tdx_enc_status_changed(vaddr, numpages, enc))
return -EIO;
+ if (enc)
+ atomic_long_sub(numpages, &nr_shared);
+ else
+ atomic_long_add(numpages, &nr_shared);
+
return 0;
}
--
2.43.0
ACPI MADT doesn't allow to offline a CPU after it was onlined. This
limits kexec: the second kernel won't be able to use more than one CPU.
To prevent a kexec kernel from onlining secondary CPUs invalidate the
mailbox address in the ACPI MADT wakeup structure which prevents a
kexec kernel to use it.
This is safe as the booting kernel has the mailbox address cached
already and acpi_wakeup_cpu() uses the cached value to bring up the
secondary CPUs.
Note: This is a Linux specific convention and not covered by the
ACPI specification.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Kai Huang <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/kernel/acpi/madt_wakeup.c | 29 ++++++++++++++++++++++++++++-
1 file changed, 28 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
index 004801b9b151..30820f9de5af 100644
--- a/arch/x86/kernel/acpi/madt_wakeup.c
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -14,6 +14,11 @@ static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox __ro_afte
static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
{
+ if (!acpi_mp_wake_mailbox_paddr) {
+ pr_warn_once("No MADT mailbox: cannot bringup secondary CPUs. Booting with kexec?\n");
+ return -EOPNOTSUPP;
+ }
+
/*
* Remap mailbox memory only for the first call to acpi_wakeup_cpu().
*
@@ -64,6 +69,28 @@ static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
return 0;
}
+static void acpi_mp_disable_offlining(struct acpi_madt_multiproc_wakeup *mp_wake)
+{
+ cpu_hotplug_disable_offlining();
+
+ /*
+ * ACPI MADT doesn't allow to offline a CPU after it was onlined. This
+ * limits kexec: the second kernel won't be able to use more than one CPU.
+ *
+ * To prevent a kexec kernel from onlining secondary CPUs invalidate the
+ * mailbox address in the ACPI MADT wakeup structure which prevents a
+ * kexec kernel to use it.
+ *
+ * This is safe as the booting kernel has the mailbox address cached
+ * already and acpi_wakeup_cpu() uses the cached value to bring up the
+ * secondary CPUs.
+ *
+ * Note: This is a Linux specific convention and not covered by the
+ * ACPI specification.
+ */
+ mp_wake->mailbox_address = 0;
+}
+
int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
const unsigned long end)
{
@@ -77,7 +104,7 @@ int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
acpi_mp_wake_mailbox_paddr = mp_wake->mailbox_address;
- cpu_hotplug_disable_offlining();
+ acpi_mp_disable_offlining(mp_wake);
apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
--
2.43.0
TDX guests allocate shared buffers to perform I/O. It is done by
allocating pages normally from the buddy allocator and converting them
to shared with set_memory_decrypted().
The second kernel has no idea what memory is converted this way. It only
sees E820_TYPE_RAM.
Accessing shared memory via private mapping is fatal. It leads to
unrecoverable TD exit.
On kexec walk direct mapping and convert all shared memory back to
private. It makes all RAM private again and second kernel may use it
normally.
The conversion occurs in two steps: stopping new conversions and
unsharing all memory. In the case of normal kexec, the stopping of
conversions takes place while scheduling is still functioning. This
allows for waiting until any ongoing conversions are finished. The
second step is carried out when all CPUs except one are inactive and
interrupts are disabled. This prevents any conflicts with code that may
access shared memory.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Rick Edgecombe <[email protected]>
Reviewed-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/coco/tdx/tdx.c | 72 +++++++++++++++++++++++++++++++
arch/x86/include/asm/pgtable.h | 5 +++
arch/x86/include/asm/set_memory.h | 3 ++
arch/x86/mm/pat/set_memory.c | 35 +++++++++++++--
4 files changed, 112 insertions(+), 3 deletions(-)
diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c
index 979891e97d83..59776ce1c1d7 100644
--- a/arch/x86/coco/tdx/tdx.c
+++ b/arch/x86/coco/tdx/tdx.c
@@ -7,6 +7,7 @@
#include <linux/cpufeature.h>
#include <linux/export.h>
#include <linux/io.h>
+#include <linux/kexec.h>
#include <asm/coco.h>
#include <asm/tdx.h>
#include <asm/vmx.h>
@@ -14,6 +15,7 @@
#include <asm/insn.h>
#include <asm/insn-eval.h>
#include <asm/pgtable.h>
+#include <asm/set_memory.h>
/* MMIO direction */
#define EPT_READ 0
@@ -831,6 +833,73 @@ static int tdx_enc_status_change_finish(unsigned long vaddr, int numpages,
return 0;
}
+/* Stop new private<->shared conversions */
+static void tdx_kexec_stop_conversion(bool crash)
+{
+ /*
+ * Crash kernel reaches here with interrupts disabled: can't wait for
+ * conversions to finish.
+ *
+ * If race happened, just report and proceed.
+ */
+ bool wait_for_lock = !crash;
+
+ if (!stop_memory_enc_conversion(wait_for_lock))
+ pr_warn("Failed to stop shared<->private conversions\n");
+}
+
+static void tdx_kexec_unshare_mem(void)
+{
+ unsigned long addr, end;
+ long found = 0, shared;
+
+ /*
+ * Walk direct mapping and convert all shared memory back to private,
+ */
+
+ addr = PAGE_OFFSET;
+ end = PAGE_OFFSET + get_max_mapped();
+
+ while (addr < end) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ pte = lookup_address(addr, &level);
+ size = page_level_size(level);
+
+ if (pte && pte_decrypted(*pte)) {
+ int pages = size / PAGE_SIZE;
+
+ /*
+ * Touching memory with shared bit set triggers implicit
+ * conversion to shared.
+ *
+ * Make sure nobody touches the shared range from
+ * now on.
+ */
+ set_pte(pte, __pte(0));
+
+ if (!tdx_enc_status_changed(addr, pages, true)) {
+ pr_err("Failed to unshare range %#lx-%#lx\n",
+ addr, addr + size);
+ }
+
+ found += pages;
+ }
+
+ addr += size;
+ }
+
+ __flush_tlb_all();
+
+ shared = atomic_long_read(&nr_shared);
+ if (shared != found) {
+ pr_err("shared page accounting is off\n");
+ pr_err("nr_shared = %ld, nr_found = %ld\n", shared, found);
+ }
+}
+
void __init tdx_early_init(void)
{
struct tdx_module_args args = {
@@ -890,6 +959,9 @@ void __init tdx_early_init(void)
x86_platform.guest.enc_cache_flush_required = tdx_cache_flush_required;
x86_platform.guest.enc_tlb_flush_required = tdx_tlb_flush_required;
+ x86_platform.guest.enc_kexec_stop_conversion = tdx_kexec_stop_conversion;
+ x86_platform.guest.enc_kexec_unshare_mem = tdx_kexec_unshare_mem;
+
/*
* TDX intercepts the RDMSR to read the X2APIC ID in the parallel
* bringup low level code. That raises #VE which cannot be handled
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 315535ffb258..17f4d97fae06 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -140,6 +140,11 @@ static inline int pte_young(pte_t pte)
return pte_flags(pte) & _PAGE_ACCESSED;
}
+static inline bool pte_decrypted(pte_t pte)
+{
+ return cc_mkdec(pte_val(pte)) == pte_val(pte);
+}
+
#define pmd_dirty pmd_dirty
static inline bool pmd_dirty(pmd_t pmd)
{
diff --git a/arch/x86/include/asm/set_memory.h b/arch/x86/include/asm/set_memory.h
index 9aee31862b4a..44b6d711296c 100644
--- a/arch/x86/include/asm/set_memory.h
+++ b/arch/x86/include/asm/set_memory.h
@@ -49,8 +49,11 @@ int set_memory_wb(unsigned long addr, int numpages);
int set_memory_np(unsigned long addr, int numpages);
int set_memory_p(unsigned long addr, int numpages);
int set_memory_4k(unsigned long addr, int numpages);
+
+bool stop_memory_enc_conversion(bool wait);
int set_memory_encrypted(unsigned long addr, int numpages);
int set_memory_decrypted(unsigned long addr, int numpages);
+
int set_memory_np_noalias(unsigned long addr, int numpages);
int set_memory_nonglobal(unsigned long addr, int numpages);
int set_memory_global(unsigned long addr, int numpages);
diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index 6c49f69c0368..21835339c0e6 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -2188,12 +2188,41 @@ static int __set_memory_enc_pgtable(unsigned long addr, int numpages, bool enc)
return ret;
}
+static DECLARE_RWSEM(mem_enc_lock);
+
+/*
+ * Stop new private<->shared conversions.
+ *
+ * Taking the exclusive mem_enc_lock waits for in-flight conversions to complete.
+ * The lock is not released to prevent new conversions from being started.
+ *
+ * If sleep is not allowed, as in a crash scenario, try to take the lock.
+ * Failure indicates that there is a race with the conversion.
+ */
+bool stop_memory_enc_conversion(bool wait)
+{
+ if (!wait)
+ return down_write_trylock(&mem_enc_lock);
+
+ down_write(&mem_enc_lock);
+
+ return true;
+}
+
static int __set_memory_enc_dec(unsigned long addr, int numpages, bool enc)
{
- if (cc_platform_has(CC_ATTR_MEM_ENCRYPT))
- return __set_memory_enc_pgtable(addr, numpages, enc);
+ int ret = 0;
- return 0;
+ if (cc_platform_has(CC_ATTR_MEM_ENCRYPT)) {
+ if (!down_read_trylock(&mem_enc_lock))
+ return -EBUSY;
+
+ ret =__set_memory_enc_pgtable(addr, numpages, enc);
+
+ up_read(&mem_enc_lock);
+ }
+
+ return ret;
}
int set_memory_encrypted(unsigned long addr, int numpages)
--
2.43.0
To prepare for the addition of support for MADT wakeup structure version
1, it is necessary to provide more appropriate names for the fields in
the structure.
The field 'mailbox_version' renamed as 'version'. This field signifies
the version of the structure and the related protocols, rather than the
version of the mailbox. This field has not been utilized in the code
thus far.
The field 'base_address' renamed as 'mailbox_address' to clarify the
kind of address it represents. In version 1, the structure includes the
reset vector address. Clear and distinct naming helps to prevent any
confusion.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Kai Huang <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/kernel/acpi/madt_wakeup.c | 2 +-
include/acpi/actbl2.h | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
index d222be8d7a07..004801b9b151 100644
--- a/arch/x86/kernel/acpi/madt_wakeup.c
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -75,7 +75,7 @@ int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
acpi_table_print_madt_entry(&header->common);
- acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
+ acpi_mp_wake_mailbox_paddr = mp_wake->mailbox_address;
cpu_hotplug_disable_offlining();
diff --git a/include/acpi/actbl2.h b/include/acpi/actbl2.h
index 9775384d61c6..e1a395af7591 100644
--- a/include/acpi/actbl2.h
+++ b/include/acpi/actbl2.h
@@ -1117,9 +1117,9 @@ struct acpi_madt_generic_translator {
struct acpi_madt_multiproc_wakeup {
struct acpi_subtable_header header;
- u16 mailbox_version;
+ u16 version;
u32 reserved; /* reserved - must be zero */
- u64 base_address;
+ u64 mailbox_address;
};
#define ACPI_MULTIPROC_WAKEUP_MB_OS_SIZE 2032
--
2.43.0
e820__end_of_ram_pfn() is used to calculate max_pfn which, among other
things, guides where direct mapping ends. Any memory above max_pfn is
not going to be present in the direct mapping.
e820__end_of_ram_pfn() finds the end of the ram based on the highest
E820_TYPE_RAM range. But it doesn't includes E820_TYPE_ACPI ranges into
calculation.
Despite the name, E820_TYPE_ACPI covers not only ACPI data, but also EFI
tables and might be required by kernel to function properly.
Usually the problem is hidden because there is some E820_TYPE_RAM memory
above E820_TYPE_ACPI. But crashkernel only presents pre-allocated crash
memory as E820_TYPE_RAM on boot. If the preallocated range is small, it
can fit under the last E820_TYPE_ACPI range.
Modify e820__end_of_ram_pfn() and e820__end_of_low_ram_pfn() to cover
E820_TYPE_ACPI memory.
The problem was discovered during debugging kexec for TDX guest. TDX
guest uses E820_TYPE_ACPI to store the unaccepted memory bitmap and pass
it between the kernels on kexec.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Dave Hansen <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/kernel/e820.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 6f1b379e3b38..f29969428443 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -827,7 +827,7 @@ u64 __init e820__memblock_alloc_reserved(u64 size, u64 align)
/*
* Find the highest page frame number we have available
*/
-static unsigned long __init e820_end_pfn(unsigned long limit_pfn, enum e820_type type)
+static unsigned long __init e820_end_ram_pfn(unsigned long limit_pfn)
{
int i;
unsigned long last_pfn = 0;
@@ -838,7 +838,8 @@ static unsigned long __init e820_end_pfn(unsigned long limit_pfn, enum e820_type
unsigned long start_pfn;
unsigned long end_pfn;
- if (entry->type != type)
+ if (entry->type != E820_TYPE_RAM &&
+ entry->type != E820_TYPE_ACPI)
continue;
start_pfn = entry->addr >> PAGE_SHIFT;
@@ -864,12 +865,12 @@ static unsigned long __init e820_end_pfn(unsigned long limit_pfn, enum e820_type
unsigned long __init e820__end_of_ram_pfn(void)
{
- return e820_end_pfn(MAX_ARCH_PFN, E820_TYPE_RAM);
+ return e820_end_ram_pfn(MAX_ARCH_PFN);
}
unsigned long __init e820__end_of_low_ram_pfn(void)
{
- return e820_end_pfn(1UL << (32 - PAGE_SHIFT), E820_TYPE_RAM);
+ return e820_end_ram_pfn(1UL << (32 - PAGE_SHIFT));
}
static void __init early_panic(char *msg)
--
2.43.0
AMD SEV and Intel TDX guests allocate shared buffers for performing I/O.
This is done by allocating pages normally from the buddy allocator and
then converting them to shared using set_memory_decrypted().
On kexec, the second kernel is unaware of which memory has been
converted in this manner. It only sees E820_TYPE_RAM. Accessing shared
memory as private is fatal.
Therefore, the memory state must be reset to its original state before
starting the new kernel with kexec.
The process of converting shared memory back to private occurs in two
steps:
- enc_kexec_stop_conversion() stops new conversions.
- enc_kexec_unshare_mem() unshares all existing shared memory, reverting
it back to private.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>x
Reviewed-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/include/asm/x86_init.h | 2 ++
arch/x86/kernel/crash.c | 6 ++++++
arch/x86/kernel/reboot.c | 12 ++++++++++++
arch/x86/kernel/x86_init.c | 4 ++++
4 files changed, 24 insertions(+)
diff --git a/arch/x86/include/asm/x86_init.h b/arch/x86/include/asm/x86_init.h
index 28ac3cb9b987..c731e6bc4343 100644
--- a/arch/x86/include/asm/x86_init.h
+++ b/arch/x86/include/asm/x86_init.h
@@ -155,6 +155,8 @@ struct x86_guest {
int (*enc_status_change_finish)(unsigned long vaddr, int npages, bool enc);
bool (*enc_tlb_flush_required)(bool enc);
bool (*enc_cache_flush_required)(void);
+ void (*enc_kexec_stop_conversion)(bool crash);
+ void (*enc_kexec_unshare_mem)(void);
};
/**
diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
index e74d0c4286c1..7a1560d7e62d 100644
--- a/arch/x86/kernel/crash.c
+++ b/arch/x86/kernel/crash.c
@@ -128,6 +128,12 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
#ifdef CONFIG_HPET_TIMER
hpet_disable();
#endif
+
+ if (cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT)) {
+ x86_platform.guest.enc_kexec_stop_conversion(true);
+ x86_platform.guest.enc_kexec_unshare_mem();
+ }
+
crash_save_cpu(regs, safe_smp_processor_id());
}
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index f3130f762784..1ec478f40963 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -12,6 +12,7 @@
#include <linux/delay.h>
#include <linux/objtool.h>
#include <linux/pgtable.h>
+#include <linux/kexec.h>
#include <acpi/reboot.h>
#include <asm/io.h>
#include <asm/apic.h>
@@ -716,6 +717,14 @@ static void native_machine_emergency_restart(void)
void native_machine_shutdown(void)
{
+ /*
+ * Call enc_kexec_stop_conversion() while all CPUs are still active and
+ * interrupts are enabled. This will allow all in-flight memory
+ * conversions to finish cleanly.
+ */
+ if (cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT) && kexec_in_progress)
+ x86_platform.guest.enc_kexec_stop_conversion(false);
+
/* Stop the cpus and apics */
#ifdef CONFIG_X86_IO_APIC
/*
@@ -752,6 +761,9 @@ void native_machine_shutdown(void)
#ifdef CONFIG_X86_64
x86_platform.iommu_shutdown();
#endif
+
+ if (cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT) && kexec_in_progress)
+ x86_platform.guest.enc_kexec_unshare_mem();
}
static void __machine_emergency_restart(int emergency)
diff --git a/arch/x86/kernel/x86_init.c b/arch/x86/kernel/x86_init.c
index a7143bb7dd93..045ce1c70070 100644
--- a/arch/x86/kernel/x86_init.c
+++ b/arch/x86/kernel/x86_init.c
@@ -138,6 +138,8 @@ static int enc_status_change_prepare_noop(unsigned long vaddr, int npages, bool
static int enc_status_change_finish_noop(unsigned long vaddr, int npages, bool enc) { return 0; }
static bool enc_tlb_flush_required_noop(bool enc) { return false; }
static bool enc_cache_flush_required_noop(void) { return false; }
+static void enc_kexec_stop_conversion_noop(bool crash) {}
+static void enc_kexec_unshare_mem_noop(void) {}
static bool is_private_mmio_noop(u64 addr) {return false; }
struct x86_platform_ops x86_platform __ro_after_init = {
@@ -161,6 +163,8 @@ struct x86_platform_ops x86_platform __ro_after_init = {
.enc_status_change_finish = enc_status_change_finish_noop,
.enc_tlb_flush_required = enc_tlb_flush_required_noop,
.enc_cache_flush_required = enc_cache_flush_required_noop,
+ .enc_kexec_stop_conversion = enc_kexec_stop_conversion_noop,
+ .enc_kexec_unshare_mem = enc_kexec_unshare_mem_noop,
},
};
--
2.43.0
The helper complements kernel_ident_mapping_init(): it frees the
identity mapping that was previously allocated. It will be used in the
error path to free a partially allocated mapping or if the mapping is no
longer needed.
The caller provides a struct x86_mapping_info with the free_pgd_page()
callback hooked up and the pgd_t to free.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Acked-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/include/asm/init.h | 3 ++
arch/x86/mm/ident_map.c | 73 +++++++++++++++++++++++++++++++++++++
2 files changed, 76 insertions(+)
diff --git a/arch/x86/include/asm/init.h b/arch/x86/include/asm/init.h
index cc9ccf61b6bd..14d72727d7ee 100644
--- a/arch/x86/include/asm/init.h
+++ b/arch/x86/include/asm/init.h
@@ -6,6 +6,7 @@
struct x86_mapping_info {
void *(*alloc_pgt_page)(void *); /* allocate buf for page table */
+ void (*free_pgt_page)(void *, void *); /* free buf for page table */
void *context; /* context for alloc_pgt_page */
unsigned long page_flag; /* page flag for PMD or PUD entry */
unsigned long offset; /* ident mapping offset */
@@ -16,4 +17,6 @@ struct x86_mapping_info {
int kernel_ident_mapping_init(struct x86_mapping_info *info, pgd_t *pgd_page,
unsigned long pstart, unsigned long pend);
+void kernel_ident_mapping_free(struct x86_mapping_info *info, pgd_t *pgd);
+
#endif /* _ASM_X86_INIT_H */
diff --git a/arch/x86/mm/ident_map.c b/arch/x86/mm/ident_map.c
index 968d7005f4a7..3996af7b4abf 100644
--- a/arch/x86/mm/ident_map.c
+++ b/arch/x86/mm/ident_map.c
@@ -4,6 +4,79 @@
* included by both the compressed kernel and the regular kernel.
*/
+static void free_pte(struct x86_mapping_info *info, pmd_t *pmd)
+{
+ pte_t *pte = pte_offset_kernel(pmd, 0);
+
+ info->free_pgt_page(pte, info->context);
+}
+
+static void free_pmd(struct x86_mapping_info *info, pud_t *pud)
+{
+ pmd_t *pmd = pmd_offset(pud, 0);
+ int i;
+
+ for (i = 0; i < PTRS_PER_PMD; i++) {
+ if (!pmd_present(pmd[i]))
+ continue;
+
+ if (pmd_leaf(pmd[i]))
+ continue;
+
+ free_pte(info, &pmd[i]);
+ }
+
+ info->free_pgt_page(pmd, info->context);
+}
+
+static void free_pud(struct x86_mapping_info *info, p4d_t *p4d)
+{
+ pud_t *pud = pud_offset(p4d, 0);
+ int i;
+
+ for (i = 0; i < PTRS_PER_PUD; i++) {
+ if (!pud_present(pud[i]))
+ continue;
+
+ if (pud_leaf(pud[i]))
+ continue;
+
+ free_pmd(info, &pud[i]);
+ }
+
+ info->free_pgt_page(pud, info->context);
+}
+
+static void free_p4d(struct x86_mapping_info *info, pgd_t *pgd)
+{
+ p4d_t *p4d = p4d_offset(pgd, 0);
+ int i;
+
+ for (i = 0; i < PTRS_PER_P4D; i++) {
+ if (!p4d_present(p4d[i]))
+ continue;
+
+ free_pud(info, &p4d[i]);
+ }
+
+ if (pgtable_l5_enabled())
+ info->free_pgt_page(pgd, info->context);
+}
+
+void kernel_ident_mapping_free(struct x86_mapping_info *info, pgd_t *pgd)
+{
+ int i;
+
+ for (i = 0; i < PTRS_PER_PGD; i++) {
+ if (!pgd_present(pgd[i]))
+ continue;
+
+ free_p4d(info, &pgd[i]);
+ }
+
+ info->free_pgt_page(pgd, info->context);
+}
+
static void ident_pmd_init(struct x86_mapping_info *info, pmd_t *pmd_page,
unsigned long addr, unsigned long end)
{
--
2.43.0
When MADT is parsed, print MULTIPROC_WAKEUP information:
ACPI: MP Wakeup (version[1], mailbox[0x7fffd000], reset[0x7fffe068])
This debug information will be very helpful during bring up.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Baoquan He <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
Acked-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
drivers/acpi/tables.c | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/drivers/acpi/tables.c b/drivers/acpi/tables.c
index b976e5fc3fbc..9e1b01c35070 100644
--- a/drivers/acpi/tables.c
+++ b/drivers/acpi/tables.c
@@ -198,6 +198,20 @@ void acpi_table_print_madt_entry(struct acpi_subtable_header *header)
}
break;
+ case ACPI_MADT_TYPE_MULTIPROC_WAKEUP:
+ {
+ struct acpi_madt_multiproc_wakeup *p =
+ (struct acpi_madt_multiproc_wakeup *)header;
+ u64 reset_vector = 0;
+
+ if (p->version >= ACPI_MADT_MP_WAKEUP_VERSION_V1)
+ reset_vector = p->reset_vector;
+
+ pr_debug("MP Wakeup (version[%d], mailbox[%#llx], reset[%#llx])\n",
+ p->version, p->mailbox_address, reset_vector);
+ }
+ break;
+
case ACPI_MADT_TYPE_CORE_PIC:
{
struct acpi_madt_core_pic *p = (struct acpi_madt_core_pic *)header;
--
2.43.0
MADT Multiprocessor Wakeup structure version 1 brings support of CPU
offlining: BIOS provides a reset vector where the CPU has to jump to
for offlining itself. The new TEST mailbox command can be used to test
whether the CPU offlined itself which means the BIOS has control over
the CPU and can online it again via the ACPI MADT wakeup method.
Add CPU offling support for the ACPI MADT wakeup method by implementing
custom cpu_die(), play_dead() and stop_this_cpu() SMP operations.
CPU offlining makes is possible to hand over secondary CPUs over kexec,
not limiting the second kernel to a single CPU.
The change conforms to the approved ACPI spec change proposal. See the
Link.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Link: https://lore.kernel.org/all/13356251.uLZWGnKmhe@kreacher
Acked-by: Kai Huang <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/include/asm/acpi.h | 2 +
arch/x86/kernel/acpi/Makefile | 2 +-
arch/x86/kernel/acpi/madt_playdead.S | 28 ++++
arch/x86/kernel/acpi/madt_wakeup.c | 184 ++++++++++++++++++++++++++-
include/acpi/actbl2.h | 15 ++-
5 files changed, 227 insertions(+), 4 deletions(-)
create mode 100644 arch/x86/kernel/acpi/madt_playdead.S
diff --git a/arch/x86/include/asm/acpi.h b/arch/x86/include/asm/acpi.h
index 2625b915ae7f..021cafa214c2 100644
--- a/arch/x86/include/asm/acpi.h
+++ b/arch/x86/include/asm/acpi.h
@@ -81,6 +81,8 @@ union acpi_subtable_headers;
int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
const unsigned long end);
+void asm_acpi_mp_play_dead(u64 reset_vector, u64 pgd_pa);
+
/*
* Check if the CPU can handle C2 and deeper
*/
diff --git a/arch/x86/kernel/acpi/Makefile b/arch/x86/kernel/acpi/Makefile
index 8c7329c88a75..37b1f28846de 100644
--- a/arch/x86/kernel/acpi/Makefile
+++ b/arch/x86/kernel/acpi/Makefile
@@ -4,7 +4,7 @@ obj-$(CONFIG_ACPI) += boot.o
obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
obj-$(CONFIG_ACPI_APEI) += apei.o
obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
-obj-$(CONFIG_X86_ACPI_MADT_WAKEUP) += madt_wakeup.o
+obj-$(CONFIG_X86_ACPI_MADT_WAKEUP) += madt_wakeup.o madt_playdead.o
ifneq ($(CONFIG_ACPI_PROCESSOR),)
obj-y += cstate.o
diff --git a/arch/x86/kernel/acpi/madt_playdead.S b/arch/x86/kernel/acpi/madt_playdead.S
new file mode 100644
index 000000000000..4e498d28cdc8
--- /dev/null
+++ b/arch/x86/kernel/acpi/madt_playdead.S
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#include <linux/linkage.h>
+#include <asm/nospec-branch.h>
+#include <asm/page_types.h>
+#include <asm/processor-flags.h>
+
+ .text
+ .align PAGE_SIZE
+
+/*
+ * asm_acpi_mp_play_dead() - Hand over control of the CPU to the BIOS
+ *
+ * rdi: Address of the ACPI MADT MPWK ResetVector
+ * rsi: PGD of the identity mapping
+ */
+SYM_FUNC_START(asm_acpi_mp_play_dead)
+ /* Turn off global entries. Following CR3 write will flush them. */
+ movq %cr4, %rdx
+ andq $~(X86_CR4_PGE), %rdx
+ movq %rdx, %cr4
+
+ /* Switch to identity mapping */
+ movq %rsi, %cr3
+
+ /* Jump to reset vector */
+ ANNOTATE_RETPOLINE_SAFE
+ jmp *%rdi
+SYM_FUNC_END(asm_acpi_mp_play_dead)
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
index 30820f9de5af..6cfe762be28b 100644
--- a/arch/x86/kernel/acpi/madt_wakeup.c
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -1,10 +1,19 @@
// SPDX-License-Identifier: GPL-2.0-or-later
#include <linux/acpi.h>
#include <linux/cpu.h>
+#include <linux/delay.h>
#include <linux/io.h>
+#include <linux/kexec.h>
+#include <linux/memblock.h>
+#include <linux/pgtable.h>
+#include <linux/sched/hotplug.h>
#include <asm/apic.h>
#include <asm/barrier.h>
+#include <asm/init.h>
+#include <asm/intel_pt.h>
+#include <asm/nmi.h>
#include <asm/processor.h>
+#include <asm/reboot.h>
/* Physical address of the Multiprocessor Wakeup Structure mailbox */
static u64 acpi_mp_wake_mailbox_paddr __ro_after_init;
@@ -12,6 +21,154 @@ static u64 acpi_mp_wake_mailbox_paddr __ro_after_init;
/* Virtual address of the Multiprocessor Wakeup Structure mailbox */
static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox __ro_after_init;
+static u64 acpi_mp_pgd __ro_after_init;
+static u64 acpi_mp_reset_vector_paddr __ro_after_init;
+
+static void acpi_mp_stop_this_cpu(void)
+{
+ asm_acpi_mp_play_dead(acpi_mp_reset_vector_paddr, acpi_mp_pgd);
+}
+
+static void acpi_mp_play_dead(void)
+{
+ play_dead_common();
+ asm_acpi_mp_play_dead(acpi_mp_reset_vector_paddr, acpi_mp_pgd);
+}
+
+static void acpi_mp_cpu_die(unsigned int cpu)
+{
+ u32 apicid = per_cpu(x86_cpu_to_apicid, cpu);
+ unsigned long timeout;
+
+ /*
+ * Use TEST mailbox command to prove that BIOS got control over
+ * the CPU before declaring it dead.
+ *
+ * BIOS has to clear 'command' field of the mailbox.
+ */
+ acpi_mp_wake_mailbox->apic_id = apicid;
+ smp_store_release(&acpi_mp_wake_mailbox->command,
+ ACPI_MP_WAKE_COMMAND_TEST);
+
+ /* Don't wait longer than a second. */
+ timeout = USEC_PER_SEC;
+ while (READ_ONCE(acpi_mp_wake_mailbox->command) && --timeout)
+ udelay(1);
+
+ if (!timeout)
+ pr_err("Failed to hand over CPU %d to BIOS\n", cpu);
+}
+
+/* The argument is required to match type of x86_mapping_info::alloc_pgt_page */
+static void __init *alloc_pgt_page(void *dummy)
+{
+ return memblock_alloc(PAGE_SIZE, PAGE_SIZE);
+}
+
+static void __init free_pgt_page(void *pgt, void *dummy)
+{
+ return memblock_free(pgt, PAGE_SIZE);
+}
+
+/*
+ * Make sure asm_acpi_mp_play_dead() is present in the identity mapping at
+ * the same place as in the kernel page tables. asm_acpi_mp_play_dead() switches
+ * to the identity mapping and the function has be present at the same spot in
+ * the virtual address space before and after switching page tables.
+ */
+static int __init init_transition_pgtable(pgd_t *pgd)
+{
+ pgprot_t prot = PAGE_KERNEL_EXEC_NOENC;
+ unsigned long vaddr, paddr;
+ p4d_t *p4d;
+ pud_t *pud;
+ pmd_t *pmd;
+ pte_t *pte;
+
+ vaddr = (unsigned long)asm_acpi_mp_play_dead;
+ pgd += pgd_index(vaddr);
+ if (!pgd_present(*pgd)) {
+ p4d = (p4d_t *)alloc_pgt_page(NULL);
+ if (!p4d)
+ return -ENOMEM;
+ set_pgd(pgd, __pgd(__pa(p4d) | _KERNPG_TABLE));
+ }
+ p4d = p4d_offset(pgd, vaddr);
+ if (!p4d_present(*p4d)) {
+ pud = (pud_t *)alloc_pgt_page(NULL);
+ if (!pud)
+ return -ENOMEM;
+ set_p4d(p4d, __p4d(__pa(pud) | _KERNPG_TABLE));
+ }
+ pud = pud_offset(p4d, vaddr);
+ if (!pud_present(*pud)) {
+ pmd = (pmd_t *)alloc_pgt_page(NULL);
+ if (!pmd)
+ return -ENOMEM;
+ set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE));
+ }
+ pmd = pmd_offset(pud, vaddr);
+ if (!pmd_present(*pmd)) {
+ pte = (pte_t *)alloc_pgt_page(NULL);
+ if (!pte)
+ return -ENOMEM;
+ set_pmd(pmd, __pmd(__pa(pte) | _KERNPG_TABLE));
+ }
+ pte = pte_offset_kernel(pmd, vaddr);
+
+ paddr = __pa(vaddr);
+ set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, prot));
+
+ return 0;
+}
+
+static int __init acpi_mp_setup_reset(u64 reset_vector)
+{
+ struct x86_mapping_info info = {
+ .alloc_pgt_page = alloc_pgt_page,
+ .free_pgt_page = free_pgt_page,
+ .page_flag = __PAGE_KERNEL_LARGE_EXEC,
+ .kernpg_flag = _KERNPG_TABLE_NOENC,
+ };
+ pgd_t *pgd;
+
+ pgd = alloc_pgt_page(NULL);
+ if (!pgd)
+ return -ENOMEM;
+
+ for (int i = 0; i < nr_pfn_mapped; i++) {
+ unsigned long mstart, mend;
+
+ mstart = pfn_mapped[i].start << PAGE_SHIFT;
+ mend = pfn_mapped[i].end << PAGE_SHIFT;
+ if (kernel_ident_mapping_init(&info, pgd, mstart, mend)) {
+ kernel_ident_mapping_free(&info, pgd);
+ return -ENOMEM;
+ }
+ }
+
+ if (kernel_ident_mapping_init(&info, pgd,
+ PAGE_ALIGN_DOWN(reset_vector),
+ PAGE_ALIGN(reset_vector + 1))) {
+ kernel_ident_mapping_free(&info, pgd);
+ return -ENOMEM;
+ }
+
+ if (init_transition_pgtable(pgd)) {
+ kernel_ident_mapping_free(&info, pgd);
+ return -ENOMEM;
+ }
+
+ smp_ops.play_dead = acpi_mp_play_dead;
+ smp_ops.stop_this_cpu = acpi_mp_stop_this_cpu;
+ smp_ops.cpu_die = acpi_mp_cpu_die;
+
+ acpi_mp_reset_vector_paddr = reset_vector;
+ acpi_mp_pgd = __pa(pgd);
+
+ return 0;
+}
+
static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
{
if (!acpi_mp_wake_mailbox_paddr) {
@@ -97,14 +254,37 @@ int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
struct acpi_madt_multiproc_wakeup *mp_wake;
mp_wake = (struct acpi_madt_multiproc_wakeup *)header;
- if (BAD_MADT_ENTRY(mp_wake, end))
+
+ /*
+ * Cannot use the standard BAD_MADT_ENTRY() to sanity check the @mp_wake
+ * entry. 'sizeof (struct acpi_madt_multiproc_wakeup)' can be larger
+ * than the actual size of the MP wakeup entry in ACPI table because the
+ * 'reset_vector' is only available in the V1 MP wakeup structure.
+ */
+ if (!mp_wake)
+ return -EINVAL;
+ if (end - (unsigned long)mp_wake < ACPI_MADT_MP_WAKEUP_SIZE_V0)
+ return -EINVAL;
+ if (mp_wake->header.length < ACPI_MADT_MP_WAKEUP_SIZE_V0)
return -EINVAL;
acpi_table_print_madt_entry(&header->common);
acpi_mp_wake_mailbox_paddr = mp_wake->mailbox_address;
- acpi_mp_disable_offlining(mp_wake);
+ if (mp_wake->version >= ACPI_MADT_MP_WAKEUP_VERSION_V1 &&
+ mp_wake->header.length >= ACPI_MADT_MP_WAKEUP_SIZE_V1) {
+ if (acpi_mp_setup_reset(mp_wake->reset_vector)) {
+ pr_warn("Failed to setup MADT reset vector\n");
+ acpi_mp_disable_offlining(mp_wake);
+ }
+ } else {
+ /*
+ * CPU offlining requires version 1 of the ACPI MADT wakeup
+ * structure.
+ */
+ acpi_mp_disable_offlining(mp_wake);
+ }
apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
diff --git a/include/acpi/actbl2.h b/include/acpi/actbl2.h
index e1a395af7591..2aedda70ef88 100644
--- a/include/acpi/actbl2.h
+++ b/include/acpi/actbl2.h
@@ -1120,8 +1120,20 @@ struct acpi_madt_multiproc_wakeup {
u16 version;
u32 reserved; /* reserved - must be zero */
u64 mailbox_address;
+ u64 reset_vector;
};
+/* Values for Version field above */
+
+enum acpi_madt_multiproc_wakeup_version {
+ ACPI_MADT_MP_WAKEUP_VERSION_NONE = 0,
+ ACPI_MADT_MP_WAKEUP_VERSION_V1 = 1,
+ ACPI_MADT_MP_WAKEUP_VERSION_RESERVED = 2, /* 2 and greater are reserved */
+};
+
+#define ACPI_MADT_MP_WAKEUP_SIZE_V0 16
+#define ACPI_MADT_MP_WAKEUP_SIZE_V1 24
+
#define ACPI_MULTIPROC_WAKEUP_MB_OS_SIZE 2032
#define ACPI_MULTIPROC_WAKEUP_MB_FIRMWARE_SIZE 2048
@@ -1134,7 +1146,8 @@ struct acpi_madt_multiproc_wakeup_mailbox {
u8 reserved_firmware[ACPI_MULTIPROC_WAKEUP_MB_FIRMWARE_SIZE]; /* reserved for firmware use */
};
-#define ACPI_MP_WAKE_COMMAND_WAKEUP 1
+#define ACPI_MP_WAKE_COMMAND_WAKEUP 1
+#define ACPI_MP_WAKE_COMMAND_TEST 2
/* 17: CPU Core Interrupt Controller (ACPI 6.5) */
--
2.43.0
On Tue, 2024-04-09 at 14:29 +0300, Kirill A. Shutemov wrote:
> Depending on setup, TDX guests might be allowed to clear CR4.MCE.
> Attempt to clear it leads to #VE.
>
> Use alternatives to keep the flag during kexec for TDX guests.
>
> The change doesn't affect non-TDX-guest environments.
>
> Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Kai Huang <[email protected]>
On Tue, Apr 09, 2024, Kirill A. Shutemov wrote:
> Depending on setup, TDX guests might be allowed to clear CR4.MCE.
> Attempt to clear it leads to #VE.
>
> Use alternatives to keep the flag during kexec for TDX guests.
>
> The change doesn't affect non-TDX-guest environments.
>
> Signed-off-by: Kirill A. Shutemov <[email protected]>
> ---
> arch/x86/kernel/relocate_kernel_64.S | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
> index 56cab1bb25f5..8e2037d78a1f 100644
> --- a/arch/x86/kernel/relocate_kernel_64.S
> +++ b/arch/x86/kernel/relocate_kernel_64.S
> @@ -5,6 +5,8 @@
> */
>
> #include <linux/linkage.h>
> +#include <linux/stringify.h>
> +#include <asm/alternative.h>
> #include <asm/page_types.h>
> #include <asm/kexec.h>
> #include <asm/processor-flags.h>
> @@ -145,11 +147,17 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
> * Set cr4 to a known state:
> * - physical address extension enabled
> * - 5-level paging, if it was enabled before
> + * - Machine check exception on TDX guest, if it was enabled before.
> + * Clearing MCE might not allowed in TDX guests, depending on setup.
> */
> movl $X86_CR4_PAE, %eax
> testq $X86_CR4_LA57, %r13
> jz 1f
> orl $X86_CR4_LA57, %eax
> +1:
> + testq $X86_CR4_MCE, %r13
> + jz 1f
> + ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
The TEST+Jcc+OR sequences are rather odd, and require way more instructions and
thus way more copy+paste than is necessary.
movl $X86_CR4_LA57, %eax
ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
andl %r13d, %eax
orl $X86_CR4_PAE, %eax
movq %rax, %cr4
Then preserving new bits unconditionally only requires adding the flag to the
initial move, and feature-dependent bits only need a single ALTERNATIVE line.
And there's no branches, blazing fast kexec! ;-)
acpi_mp_wake_mailbox_paddr and acpi_mp_wake_mailbox initialized once
during ACPI MADT init and never changed.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Acked-by: Kai Huang <[email protected]>
Reviewed-by: Baoquan He <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/kernel/acpi/madt_wakeup.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
index 7f164d38bd0b..cf79ea6f3007 100644
--- a/arch/x86/kernel/acpi/madt_wakeup.c
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -6,10 +6,10 @@
#include <asm/processor.h>
/* Physical address of the Multiprocessor Wakeup Structure mailbox */
-static u64 acpi_mp_wake_mailbox_paddr;
+static u64 acpi_mp_wake_mailbox_paddr __ro_after_init;
/* Virtual address of the Multiprocessor Wakeup Structure mailbox */
-static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox;
+static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox __ro_after_init;
static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
{
--
2.43.0
If the helper is defined, it is called instead of halt() to stop the CPU
at the end of stop_this_cpu() and on crash CPU shutdown.
ACPI MADT will use it to hand over the CPU to BIOS in order to be able
to wake it up again after kexec.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Acked-by: Kai Huang <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/include/asm/smp.h | 1 +
arch/x86/kernel/process.c | 7 +++++++
arch/x86/kernel/reboot.c | 6 ++++++
3 files changed, 14 insertions(+)
diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h
index a35936b512fe..ca073f40698f 100644
--- a/arch/x86/include/asm/smp.h
+++ b/arch/x86/include/asm/smp.h
@@ -35,6 +35,7 @@ struct smp_ops {
int (*cpu_disable)(void);
void (*cpu_die)(unsigned int cpu);
void (*play_dead)(void);
+ void (*stop_this_cpu)(void);
void (*send_call_func_ipi)(const struct cpumask *mask);
void (*send_call_func_single_ipi)(int cpu);
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index b8441147eb5e..f63f8fd00a91 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -835,6 +835,13 @@ void __noreturn stop_this_cpu(void *dummy)
*/
cpumask_clear_cpu(cpu, &cpus_stop_mask);
+#ifdef CONFIG_SMP
+ if (smp_ops.stop_this_cpu) {
+ smp_ops.stop_this_cpu();
+ unreachable();
+ }
+#endif
+
for (;;) {
/*
* Use native_halt() so that memory contents don't change
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 1ec478f40963..293ded05a4b0 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -880,6 +880,12 @@ static int crash_nmi_callback(unsigned int val, struct pt_regs *regs)
cpu_emergency_disable_virtualization();
atomic_dec(&waiting_for_crash_ipi);
+
+ if (smp_ops.stop_this_cpu) {
+ smp_ops.stop_this_cpu();
+ unreachable();
+ }
+
/* Assume hlt works */
halt();
for (;;)
--
2.43.0
On Tue, Apr 09, 2024 at 07:22:24AM -0700, Sean Christopherson wrote:
> On Tue, Apr 09, 2024, Kirill A. Shutemov wrote:
> > Depending on setup, TDX guests might be allowed to clear CR4.MCE.
> > Attempt to clear it leads to #VE.
> >
> > Use alternatives to keep the flag during kexec for TDX guests.
> >
> > The change doesn't affect non-TDX-guest environments.
> >
> > Signed-off-by: Kirill A. Shutemov <[email protected]>
> > ---
> > arch/x86/kernel/relocate_kernel_64.S | 8 ++++++++
> > 1 file changed, 8 insertions(+)
> >
> > diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
> > index 56cab1bb25f5..8e2037d78a1f 100644
> > --- a/arch/x86/kernel/relocate_kernel_64.S
> > +++ b/arch/x86/kernel/relocate_kernel_64.S
> > @@ -5,6 +5,8 @@
> > */
> >
> > #include <linux/linkage.h>
> > +#include <linux/stringify.h>
> > +#include <asm/alternative.h>
> > #include <asm/page_types.h>
> > #include <asm/kexec.h>
> > #include <asm/processor-flags.h>
> > @@ -145,11 +147,17 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
> > * Set cr4 to a known state:
> > * - physical address extension enabled
> > * - 5-level paging, if it was enabled before
> > + * - Machine check exception on TDX guest, if it was enabled before.
> > + * Clearing MCE might not allowed in TDX guests, depending on setup.
> > */
> > movl $X86_CR4_PAE, %eax
> > testq $X86_CR4_LA57, %r13
> > jz 1f
> > orl $X86_CR4_LA57, %eax
> > +1:
> > + testq $X86_CR4_MCE, %r13
> > + jz 1f
> > + ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
>
> The TEST+Jcc+OR sequences are rather odd, and require way more instructions and
> thus way more copy+paste than is necessary.
>
> movl $X86_CR4_LA57, %eax
> ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
> andl %r13d, %eax
> orl $X86_CR4_PAE, %eax
> movq %rax, %cr4
>
> Then preserving new bits unconditionally only requires adding the flag to the
> initial move, and feature-dependent bits only need a single ALTERNATIVE line.
Thanks! It is much better.
> And there's no branches, blazing fast kexec! ;-)
kexec/sec STONKS! :D
Updated patch is below.
From 6be428e3b1c6fb494b2c48ba6a7c133514a0b2b4 Mon Sep 17 00:00:00 2001
From: "Kirill A. Shutemov" <[email protected]>
Date: Fri, 10 Feb 2023 12:53:11 +0300
Subject: [PATCHv10.1 05/18] x86/kexec: Keep CR4.MCE set during kexec for TDX guest
Depending on setup, TDX guests might be allowed to clear CR4.MCE.
Attempt to clear it leads to #VE.
Use alternatives to keep the flag during kexec for TDX guests.
The change doesn't affect non-TDX-guest environments.
Signed-off-by: Kirill A. Shutemov <[email protected]>
---
arch/x86/kernel/relocate_kernel_64.S | 15 +++++++++------
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
index 56cab1bb25f5..90246d544eb1 100644
--- a/arch/x86/kernel/relocate_kernel_64.S
+++ b/arch/x86/kernel/relocate_kernel_64.S
@@ -5,6 +5,8 @@
*/
#include <linux/linkage.h>
+#include <linux/stringify.h>
+#include <asm/alternative.h>
#include <asm/page_types.h>
#include <asm/kexec.h>
#include <asm/processor-flags.h>
@@ -143,14 +145,15 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
/*
* Set cr4 to a known state:
- * - physical address extension enabled
* - 5-level paging, if it was enabled before
+ * - Machine check exception on TDX guest, if it was enabled before.
+ * Clearing MCE might not allowed in TDX guests, depending on setup.
+ * - physical address extension enabled
*/
- movl $X86_CR4_PAE, %eax
- testq $X86_CR4_LA57, %r13
- jz 1f
- orl $X86_CR4_LA57, %eax
-1:
+ movl $X86_CR4_LA57, %eax
+ ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
+ andl %r13d, %eax
+ orl $X86_CR4_PAE, %eax
movq %rax, %cr4
jmp 1f
--
Kiryl Shutsemau / Kirill A. Shutemov
From: Ashish Kalra <[email protected]>
The patchset adds bits and pieces to get kexec (and crashkernel) work on
SNP guest.
v4:
- Rebased to current tip/master.
- Reviewed-bys from Sathya.
- Remove snp_kexec_unprep_rom_memory() as it is not needed any more as
SEV-SNP code is not validating the ROM range in probe_roms() anymore.
- Fix kernel test robot build error/warnings.
v3:
- Rebased;
- moved Keep page tables that maps E820_TYPE_ACPI patch to Kirill's tdx
guest kexec patch series.
- checking the md attribute instead of checking the efi_setup for
detecting if running under kexec kernel.
- added new sev_es_enabled() function.
- skip video memory access in decompressor for SEV-ES/SNP systems to
prevent guest termination as boot stage2 #VC handler does not handle
MMIO.
v2:
- address zeroing of unaccepted memory table mappings at all page table levels
adding phys_pte_init(), phys_pud_init() and phys_p4d_init().
- include skip efi_arch_mem_reserve() in case of kexec as part of this
patch set.
- rename last_address_shd_kexec to a more appropriate
kexec_last_address_to_make_private.
- remove duplicate code shared with TDX and use common interfaces
defined for SNP and TDX for kexec/kdump.
- remove set_pte_enc() dependency on pg_level_to_pfn() and make the
function simpler.
- rename unshare_pte() to make_pte_private().
- clarify and make the comment for using kexec_last_address_to_make_private
more understandable.
- general cleanup.
Ashish Kalra (4):
efi/x86: skip efi_arch_mem_reserve() in case of kexec.
x86/sev: add sev_es_enabled() function.
x86/boot/compressed: Skip Video Memory access in Decompressor for
SEV-ES/SNP.
x86/snp: Convert shared memory back to private on kexec
arch/x86/boot/compressed/misc.c | 6 +-
arch/x86/boot/compressed/misc.h | 1 +
arch/x86/boot/compressed/sev.c | 5 +
arch/x86/boot/compressed/sev.h | 2 +
arch/x86/include/asm/sev.h | 4 +
arch/x86/kernel/sev.c | 161 ++++++++++++++++++++++++++++++++
arch/x86/mm/mem_encrypt_amd.c | 3 +
arch/x86/platform/efi/quirks.c | 20 +++-
8 files changed, 198 insertions(+), 4 deletions(-)
--
2.34.1
From: Ashish Kalra <[email protected]>
For kexec use case, need to use and stick to the EFI memmap passed
from the first kernel via boot-params/setup data, hence,
skip efi_arch_mem_reserve() during kexec.
Additionally during SNP guest kexec testing discovered that EFI memmap
is corrupted during chained kexec. kexec_enter_virtual_mode() during
late init will remap the efi_memmap physical pages allocated in
efi_arch_mem_reserve() via memblock & then subsequently cause random
EFI memmap corruption once memblock is freed/teared-down.
Suggested-by: Dave Young <[email protected]>
[Dave Young: checking the md attribute instead of checking the efi_setup]
Signed-off-by: Ashish Kalra <[email protected]>
---
arch/x86/platform/efi/quirks.c | 20 ++++++++++++++++++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
index f0cc00032751..982f5e50a4b3 100644
--- a/arch/x86/platform/efi/quirks.c
+++ b/arch/x86/platform/efi/quirks.c
@@ -258,12 +258,28 @@ void __init efi_arch_mem_reserve(phys_addr_t addr, u64 size)
int num_entries;
void *new;
- if (efi_mem_desc_lookup(addr, &md) ||
- md.type != EFI_BOOT_SERVICES_DATA) {
+ /*
+ * For kexec use case, we need to use the EFI memmap passed from the first
+ * kernel via setup data, so we need to skip this.
+ * Additionally kexec_enter_virtual_mode() during late init will remap
+ * the efi_memmap physical pages allocated here via memboot & then
+ * subsequently cause random EFI memmap corruption once memblock is freed.
+ */
+
+ if (efi_mem_desc_lookup(addr, &md)) {
pr_err("Failed to lookup EFI memory descriptor for %pa\n", &addr);
return;
}
+ if (md.type != EFI_BOOT_SERVICES_DATA) {
+ pr_err("Skip reserving non EFI Boot Service Data memory for %pa\n", &addr);
+ return;
+ }
+
+ /* Kexec copied the efi memmap from the first kernel, thus skip the case */
+ if (md.attribute & EFI_MEMORY_RUNTIME)
+ return;
+
if (addr + size > md.phys_addr + (md.num_pages << EFI_PAGE_SHIFT)) {
pr_err("Region spans EFI memory descriptors, %pa\n", &addr);
return;
--
2.34.1
From: Ashish Kalra <[email protected]>
Add sev_es_enabled() function to detect if SEV-ES
support is enabled.
Signed-off-by: Ashish Kalra <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
---
arch/x86/boot/compressed/sev.c | 5 +++++
arch/x86/boot/compressed/sev.h | 2 ++
2 files changed, 7 insertions(+)
diff --git a/arch/x86/boot/compressed/sev.c b/arch/x86/boot/compressed/sev.c
index ec71846d28c9..4ae4cc51e6b8 100644
--- a/arch/x86/boot/compressed/sev.c
+++ b/arch/x86/boot/compressed/sev.c
@@ -134,6 +134,11 @@ bool sev_snp_enabled(void)
return sev_status & MSR_AMD64_SEV_SNP_ENABLED;
}
+bool sev_es_enabled(void)
+{
+ return sev_status & MSR_AMD64_SEV_ES_ENABLED;
+}
+
static void __page_state_change(unsigned long paddr, enum psc_op op)
{
u64 val;
diff --git a/arch/x86/boot/compressed/sev.h b/arch/x86/boot/compressed/sev.h
index fc725a981b09..5008c80e66e6 100644
--- a/arch/x86/boot/compressed/sev.h
+++ b/arch/x86/boot/compressed/sev.h
@@ -11,11 +11,13 @@
#ifdef CONFIG_AMD_MEM_ENCRYPT
bool sev_snp_enabled(void);
+bool sev_es_enabled(void);
void snp_accept_memory(phys_addr_t start, phys_addr_t end);
#else
static inline bool sev_snp_enabled(void) { return false; }
+static inline bool sev_es_enabled(void) { return false; }
static inline void snp_accept_memory(phys_addr_t start, phys_addr_t end) { }
#endif
--
2.34.1
From: Ashish Kalra <[email protected]>
Accessing guest video memory/RAM during kernel decompressor
causes guest termination as boot stage2 #VC handler for
SEV-ES/SNP systems does not support MMIO handling.
This issue is observed with SEV-ES/SNP guest kexec as
kexec -c adds screen_info to the boot parameters
passed to the kexec kernel, which causes console output to
be dumped to both video and serial.
As the decompressor output gets cleared really fast, it is
preferable to get the console output only on serial, hence,
skip accessing video RAM during decompressor stage to
prevent guest termination.
Serial console output during decompressor stage works as
boot stage2 #VC handler already supports handling port I/O.
Suggested-by: Thomas Lendacy <[email protected]>
Signed-off-by: Ashish Kalra <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
---
arch/x86/boot/compressed/misc.c | 6 ++++--
arch/x86/boot/compressed/misc.h | 1 +
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/arch/x86/boot/compressed/misc.c b/arch/x86/boot/compressed/misc.c
index b70e4a21c15f..47b4db200e1f 100644
--- a/arch/x86/boot/compressed/misc.c
+++ b/arch/x86/boot/compressed/misc.c
@@ -427,8 +427,10 @@ asmlinkage __visible void *extract_kernel(void *rmode, unsigned char *output)
vidport = 0x3d4;
}
- lines = boot_params_ptr->screen_info.orig_video_lines;
- cols = boot_params_ptr->screen_info.orig_video_cols;
+ if (!sev_es_enabled()) {
+ lines = boot_params_ptr->screen_info.orig_video_lines;
+ cols = boot_params_ptr->screen_info.orig_video_cols;
+ }
init_default_io_ops();
diff --git a/arch/x86/boot/compressed/misc.h b/arch/x86/boot/compressed/misc.h
index b353a7be380c..3c12ca987554 100644
--- a/arch/x86/boot/compressed/misc.h
+++ b/arch/x86/boot/compressed/misc.h
@@ -37,6 +37,7 @@
#include <asm/desc_defs.h>
#include "tdx.h"
+#include "sev.h"
#define BOOT_CTYPE_H
#include <linux/acpi.h>
--
2.34.1
From: Ashish Kalra <[email protected]>
SNP guests allocate shared buffers to perform I/O. It is done by
allocating pages normally from the buddy allocator and converting them
to shared with set_memory_decrypted().
The second kernel has no idea what memory is converted this way. It only
sees E820_TYPE_RAM.
Accessing shared memory via private mapping will cause unrecoverable RMP
page-faults.
On kexec walk direct mapping and convert all shared memory back to
private. It makes all RAM private again and second kernel may use it
normally. Additionally for SNP guests convert all bss decrypted section
pages back to private and switch back ROM regions to shared so that
their revalidation does not fail during kexec kernel boot.
The conversion occurs in two steps: stopping new conversions and
unsharing all memory. In the case of normal kexec, the stopping of
conversions takes place while scheduling is still functioning. This
allows for waiting until any ongoing conversions are finished. The
second step is carried out when all CPUs except one are inactive and
interrupts are disabled. This prevents any conflicts with code that may
access shared memory.
Signed-off-by: Ashish Kalra <[email protected]>
---
arch/x86/include/asm/sev.h | 4 +
arch/x86/kernel/sev.c | 161 ++++++++++++++++++++++++++++++++++
arch/x86/mm/mem_encrypt_amd.c | 3 +
3 files changed, 168 insertions(+)
diff --git a/arch/x86/include/asm/sev.h b/arch/x86/include/asm/sev.h
index 7f57382afee4..78d40d08d201 100644
--- a/arch/x86/include/asm/sev.h
+++ b/arch/x86/include/asm/sev.h
@@ -229,6 +229,8 @@ void snp_accept_memory(phys_addr_t start, phys_addr_t end);
u64 snp_get_unsupported_features(u64 status);
u64 sev_get_status(void);
void sev_show_status(void);
+void snp_kexec_unshare_mem(void);
+void snp_kexec_stop_conversion(bool crash);
#else
static inline void sev_es_ist_enter(struct pt_regs *regs) { }
static inline void sev_es_ist_exit(void) { }
@@ -258,6 +260,8 @@ static inline void snp_accept_memory(phys_addr_t start, phys_addr_t end) { }
static inline u64 snp_get_unsupported_features(u64 status) { return 0; }
static inline u64 sev_get_status(void) { return 0; }
static inline void sev_show_status(void) { }
+static inline void snp_kexec_unshare_mem(void) { }
+static inline void snp_kexec_stop_conversion(bool crash) { }
#endif
#ifdef CONFIG_KVM_AMD_SEV
diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
index 38ad066179d8..17f616963beb 100644
--- a/arch/x86/kernel/sev.c
+++ b/arch/x86/kernel/sev.c
@@ -42,6 +42,8 @@
#include <asm/apic.h>
#include <asm/cpuid.h>
#include <asm/cmdline.h>
+#include <asm/pgtable.h>
+#include <asm/set_memory.h>
#define DR7_RESET_VALUE 0x400
@@ -92,6 +94,9 @@ static struct ghcb *boot_ghcb __section(".data");
/* Bitmap of SEV features supported by the hypervisor */
static u64 sev_hv_features __ro_after_init;
+/* Last address to be switched to private during kexec */
+static unsigned long kexec_last_addr_to_make_private;
+
/* #VC handler runtime per-CPU data */
struct sev_es_runtime_data {
struct ghcb ghcb_page;
@@ -913,6 +918,162 @@ void snp_accept_memory(phys_addr_t start, phys_addr_t end)
set_pages_state(vaddr, npages, SNP_PAGE_STATE_PRIVATE);
}
+static bool set_pte_enc(pte_t *kpte, int level, void *va)
+{
+ pte_t new_pte;
+
+ if (pte_none(*kpte))
+ return false;
+
+ /*
+ * Change the physical page attribute from C=0 to C=1. Flush the
+ * caches to ensure that data gets accessed with the correct C-bit.
+ */
+ if (pte_present(*kpte))
+ clflush_cache_range(va, page_level_size(level));
+
+ new_pte = __pte(cc_mkenc(pte_val(*kpte)));
+ set_pte_atomic(kpte, new_pte);
+
+ return true;
+}
+
+static bool make_pte_private(pte_t *pte, unsigned long addr, int pages, int level)
+{
+ struct sev_es_runtime_data *data;
+ struct ghcb *ghcb;
+
+ data = this_cpu_read(runtime_data);
+ ghcb = &data->ghcb_page;
+
+ /* Check for GHCB for being part of a PMD range. */
+ if ((unsigned long)ghcb >= addr &&
+ (unsigned long)ghcb <= (addr + (pages * PAGE_SIZE))) {
+ /*
+ * Ensure that the current cpu's GHCB is made private
+ * at the end of unshared loop so that we continue to use the
+ * optimized GHCB protocol and not force the switch to
+ * MSR protocol till the very end.
+ */
+ pr_debug("setting boot_ghcb to NULL for this cpu ghcb\n");
+ kexec_last_addr_to_make_private = addr;
+ return true;
+ }
+
+ if (!set_pte_enc(pte, level, (void *)addr))
+ return false;
+
+ snp_set_memory_private(addr, pages);
+
+ return true;
+}
+
+static void unshare_all_memory(void)
+{
+ unsigned long addr, end;
+
+ /*
+ * Walk direct mapping and convert all shared memory back to private,
+ */
+
+ addr = PAGE_OFFSET;
+ end = PAGE_OFFSET + get_max_mapped();
+
+ while (addr < end) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ pte = lookup_address(addr, &level);
+ size = page_level_size(level);
+
+ /*
+ * pte_none() check is required to skip physical memory holes in direct mapped.
+ */
+ if (pte && pte_decrypted(*pte) && !pte_none(*pte)) {
+ int pages = size / PAGE_SIZE;
+
+ if (!make_pte_private(pte, addr, pages, level)) {
+ pr_err("Failed to unshare range %#lx-%#lx\n",
+ addr, addr + size);
+ }
+
+ }
+
+ addr += size;
+ }
+ __flush_tlb_all();
+
+}
+
+static void unshare_all_bss_decrypted_memory(void)
+{
+ unsigned long vaddr, vaddr_end;
+ unsigned int level;
+ unsigned int npages;
+ pte_t *pte;
+
+ vaddr = (unsigned long)__start_bss_decrypted;
+ vaddr_end = (unsigned long)__start_bss_decrypted_unused;
+ npages = (vaddr_end - vaddr) >> PAGE_SHIFT;
+ for (; vaddr < vaddr_end; vaddr += PAGE_SIZE) {
+ pte = lookup_address(vaddr, &level);
+ if (!pte || !pte_decrypted(*pte) || pte_none(*pte))
+ continue;
+
+ set_pte_enc(pte, level, (void *)vaddr);
+ }
+ vaddr = (unsigned long)__start_bss_decrypted;
+ snp_set_memory_private(vaddr, npages);
+}
+
+/* Stop new private<->shared conversions */
+void snp_kexec_stop_conversion(bool crash)
+{
+ /*
+ * Crash kernel reaches here with interrupts disabled: can't wait for
+ * conversions to finish.
+ *
+ * If race happened, just report and proceed.
+ */
+ bool wait_for_lock = !crash;
+
+ if (!stop_memory_enc_conversion(wait_for_lock))
+ pr_warn("Failed to finish shared<->private conversions\n");
+}
+
+void snp_kexec_unshare_mem(void)
+{
+ if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
+ return;
+
+ unshare_all_memory();
+
+ unshare_all_bss_decrypted_memory();
+
+ if (kexec_last_addr_to_make_private) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ /*
+ * Switch to using the MSR protocol to change this cpu's
+ * GHCB to private.
+ * All the per-cpu GHCBs have been switched back to private,
+ * so can't do any more GHCB calls to the hypervisor beyond
+ * this point till the kexec kernel starts running.
+ */
+ boot_ghcb = NULL;
+ sev_cfg.ghcbs_initialized = false;
+
+ pr_debug("boot ghcb 0x%lx\n", kexec_last_addr_to_make_private);
+ pte = lookup_address(kexec_last_addr_to_make_private, &level);
+ size = page_level_size(level);
+ set_pte_enc(pte, level, (void *)kexec_last_addr_to_make_private);
+ snp_set_memory_private(kexec_last_addr_to_make_private, (size / PAGE_SIZE));
+ }
+}
+
static int snp_set_vmsa(void *va, bool vmsa)
{
u64 attrs;
diff --git a/arch/x86/mm/mem_encrypt_amd.c b/arch/x86/mm/mem_encrypt_amd.c
index e7b67519ddb5..49c40c2ed809 100644
--- a/arch/x86/mm/mem_encrypt_amd.c
+++ b/arch/x86/mm/mem_encrypt_amd.c
@@ -468,6 +468,9 @@ void __init sme_early_init(void)
x86_platform.guest.enc_tlb_flush_required = amd_enc_tlb_flush_required;
x86_platform.guest.enc_cache_flush_required = amd_enc_cache_flush_required;
+ x86_platform.guest.enc_kexec_stop_conversion = snp_kexec_stop_conversion;
+ x86_platform.guest.enc_kexec_unshare_mem = snp_kexec_unshare_mem;
+
/*
* AMD-SEV-ES intercepts the RDMSR to read the X2APIC ID in the
* parallel bringup low level code. That raises #VC which cannot be
--
2.34.1
On Tue, Apr 09, 2024 at 08:42:38PM +0000, Ashish Kalra wrote:
> From: Ashish Kalra <[email protected]>
>
> Add sev_es_enabled() function to detect if SEV-ES
> support is enabled.
And use it exactly once?
Nah, use sev_status directly.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
Hi Ashish,
kernel test robot noticed the following build errors:
[auto build test ERROR on tip/master]
[also build test ERROR on linus/master next-20240410]
[cannot apply to tip/x86/core tip/x86/mm tip/auto-latest]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Ashish-Kalra/efi-x86-skip-efi_arch_mem_reserve-in-case-of-kexec/20240410-044512
base: tip/master
patch link: https://lore.kernel.org/r/b24885f5495f6b8ba2f9e825fda9188fcbf28231.1712694667.git.ashish.kalra%40amd.com
patch subject: [PATCH v4 4/4] x86/snp: Convert shared memory back to private on kexec
config: x86_64-rhel-8.3-rust (https://download.01.org/0day-ci/archive/20240410/[email protected]/config)
compiler: clang version 17.0.6 (https://github.com/llvm/llvm-project 6009708b4367171ccdbf4b5905cb6a803753fe18)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240410/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All errors (new ones prefixed by >>):
>> arch/x86/kernel/sev.c:993:14: error: call to undeclared function 'pte_decrypted'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
993 | if (pte && pte_decrypted(*pte) && !pte_none(*pte)) {
| ^
arch/x86/kernel/sev.c:1021:16: error: call to undeclared function 'pte_decrypted'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
1021 | if (!pte || !pte_decrypted(*pte) || pte_none(*pte))
| ^
>> arch/x86/kernel/sev.c:1041:7: error: call to undeclared function 'stop_memory_enc_conversion'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
1041 | if (!stop_memory_enc_conversion(wait_for_lock))
| ^
3 errors generated.
--
>> arch/x86/mm/mem_encrypt_amd.c:471:21: error: no member named 'enc_kexec_stop_conversion' in 'struct x86_guest'
471 | x86_platform.guest.enc_kexec_stop_conversion = snp_kexec_stop_conversion;
| ~~~~~~~~~~~~~~~~~~ ^
>> arch/x86/mm/mem_encrypt_amd.c:472:21: error: no member named 'enc_kexec_unshare_mem' in 'struct x86_guest'
472 | x86_platform.guest.enc_kexec_unshare_mem = snp_kexec_unshare_mem;
| ~~~~~~~~~~~~~~~~~~ ^
2 errors generated.
vim +/pte_decrypted +993 arch/x86/kernel/sev.c
970
971 static void unshare_all_memory(void)
972 {
973 unsigned long addr, end;
974
975 /*
976 * Walk direct mapping and convert all shared memory back to private,
977 */
978
979 addr = PAGE_OFFSET;
980 end = PAGE_OFFSET + get_max_mapped();
981
982 while (addr < end) {
983 unsigned long size;
984 unsigned int level;
985 pte_t *pte;
986
987 pte = lookup_address(addr, &level);
988 size = page_level_size(level);
989
990 /*
991 * pte_none() check is required to skip physical memory holes in direct mapped.
992 */
> 993 if (pte && pte_decrypted(*pte) && !pte_none(*pte)) {
994 int pages = size / PAGE_SIZE;
995
996 if (!make_pte_private(pte, addr, pages, level)) {
997 pr_err("Failed to unshare range %#lx-%#lx\n",
998 addr, addr + size);
999 }
1000
1001 }
1002
1003 addr += size;
1004 }
1005 __flush_tlb_all();
1006
1007 }
1008
1009 static void unshare_all_bss_decrypted_memory(void)
1010 {
1011 unsigned long vaddr, vaddr_end;
1012 unsigned int level;
1013 unsigned int npages;
1014 pte_t *pte;
1015
1016 vaddr = (unsigned long)__start_bss_decrypted;
1017 vaddr_end = (unsigned long)__start_bss_decrypted_unused;
1018 npages = (vaddr_end - vaddr) >> PAGE_SHIFT;
1019 for (; vaddr < vaddr_end; vaddr += PAGE_SIZE) {
1020 pte = lookup_address(vaddr, &level);
1021 if (!pte || !pte_decrypted(*pte) || pte_none(*pte))
1022 continue;
1023
1024 set_pte_enc(pte, level, (void *)vaddr);
1025 }
1026 vaddr = (unsigned long)__start_bss_decrypted;
1027 snp_set_memory_private(vaddr, npages);
1028 }
1029
1030 /* Stop new private<->shared conversions */
1031 void snp_kexec_stop_conversion(bool crash)
1032 {
1033 /*
1034 * Crash kernel reaches here with interrupts disabled: can't wait for
1035 * conversions to finish.
1036 *
1037 * If race happened, just report and proceed.
1038 */
1039 bool wait_for_lock = !crash;
1040
> 1041 if (!stop_memory_enc_conversion(wait_for_lock))
1042 pr_warn("Failed to finish shared<->private conversions\n");
1043 }
1044
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
From: Ashish Kalra <[email protected]>
The patchset adds bits and pieces to get kexec (and crashkernel) work on
SNP guest.
The series is based off of and tested against Kirill Shutemov's tree:
https://github.com/intel/tdx.git guest-kexec
----
v5:
- Removed sev_es_enabled() function and using sev_status directly to
check for SEV-ES/SEV-SNP guest.
- used --base option to generate patches to specify Kirill's TDX guest
kexec patches as prerequisite patches to fix kernel test robot
build errors.
v4:
- Rebased to current tip/master.
- Reviewed-bys from Sathya.
- Remove snp_kexec_unprep_rom_memory() as it is not needed any more as
SEV-SNP code is not validating the ROM range in probe_roms() anymore.
- Fix kernel test robot build error/warnings.
v3:
- Rebased;
- moved Keep page tables that maps E820_TYPE_ACPI patch to Kirill's tdx
guest kexec patch series.
- checking the md attribute instead of checking the efi_setup for
detecting if running under kexec kernel.
- added new sev_es_enabled() function.
- skip video memory access in decompressor for SEV-ES/SNP systems to
prevent guest termination as boot stage2 #VC handler does not handle
MMIO.
v2:
- address zeroing of unaccepted memory table mappings at all page table levels
adding phys_pte_init(), phys_pud_init() and phys_p4d_init().
- include skip efi_arch_mem_reserve() in case of kexec as part of this
patch set.
- rename last_address_shd_kexec to a more appropriate
kexec_last_address_to_make_private.
- remove duplicate code shared with TDX and use common interfaces
defined for SNP and TDX for kexec/kdump.
- remove set_pte_enc() dependency on pg_level_to_pfn() and make the
function simpler.
- rename unshare_pte() to make_pte_private().
- clarify and make the comment for using kexec_last_address_to_make_private
more understandable.
- general cleanup.
Ashish Kalra (3):
efi/x86: skip efi_arch_mem_reserve() in case of kexec.
x86/boot/compressed: Skip Video Memory access in Decompressor for
SEV-ES/SNP.
x86/snp: Convert shared memory back to private on kexec
arch/x86/boot/compressed/misc.c | 6 +-
arch/x86/include/asm/sev.h | 4 +
arch/x86/kernel/sev.c | 161 ++++++++++++++++++++++++++++++++
arch/x86/mm/mem_encrypt_amd.c | 3 +
arch/x86/platform/efi/quirks.c | 20 +++-
5 files changed, 190 insertions(+), 4 deletions(-)
base-commit: a18b42d8997abfd77aa1637c0de6850b0c30b1fe
prerequisite-patch-id: bd8e77f0f12223d21cb2f35b77bfcbdd9ad80b0f
prerequisite-patch-id: bfe2fa046349978ac1825275eb205acecfbc22f3
prerequisite-patch-id: 5e60d292457c7cd98fd3e45c23127e9463b56a69
prerequisite-patch-id: 1f97d0a2edb7509dd58276f628d1a4bda62c154c
prerequisite-patch-id: 8db559385c44e8b6670d74196e8d83d2dfad2f40
prerequisite-patch-id: cbdfea1e50ecb3b4cee3a25a27df4d35bd95d532
prerequisite-patch-id: 1cea0996e0dc3bb9f0059c927c405ca31003791e
prerequisite-patch-id: 469a0a3c78b0eca82527cd85e2205fb8fb89d645
prerequisite-patch-id: 2974ef211db5253d9782018e352d2a6ff0b0ef54
prerequisite-patch-id: 2cfffd80947941892421dae99b7fa0f9f9715884
prerequisite-patch-id: 466c2cb9f0a107bbd1dbd8526f4eff2bdb55f1ce
prerequisite-patch-id: d4966ae63e86d24b0bf578da4dae871cd9002b12
prerequisite-patch-id: fccde6f1fa385b5af0195f81fcb95acd71822428
prerequisite-patch-id: 16048ee15e392b0b9217b8923939b0059311abd2
prerequisite-patch-id: 5c9ae9aa294f72f63ae2c3551507dfbd92525803
prerequisite-patch-id: 758bdb686290c018cbd5b7d005354019f9d15248
prerequisite-patch-id: 4125b799fc9577b1a46427e45618fa0174f7a4b3
prerequisite-patch-id: 60760e0c98ab7ccd2ca22ae3e9f20ff5a94c6e91
--
2.34.1
From: Ashish Kalra <[email protected]>
For kexec use case, need to use and stick to the EFI memmap passed
from the first kernel via boot-params/setup data, hence,
skip efi_arch_mem_reserve() during kexec.
Additionally during SNP guest kexec testing discovered that EFI memmap
is corrupted during chained kexec. kexec_enter_virtual_mode() during
late init will remap the efi_memmap physical pages allocated in
efi_arch_mem_reserve() via memblock & then subsequently cause random
EFI memmap corruption once memblock is freed/teared-down.
Suggested-by: Dave Young <[email protected]>
[Dave Young: checking the md attribute instead of checking the efi_setup]
Signed-off-by: Ashish Kalra <[email protected]>
---
arch/x86/platform/efi/quirks.c | 20 ++++++++++++++++++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
index f0cc00032751..982f5e50a4b3 100644
--- a/arch/x86/platform/efi/quirks.c
+++ b/arch/x86/platform/efi/quirks.c
@@ -258,12 +258,28 @@ void __init efi_arch_mem_reserve(phys_addr_t addr, u64 size)
int num_entries;
void *new;
- if (efi_mem_desc_lookup(addr, &md) ||
- md.type != EFI_BOOT_SERVICES_DATA) {
+ /*
+ * For kexec use case, we need to use the EFI memmap passed from the first
+ * kernel via setup data, so we need to skip this.
+ * Additionally kexec_enter_virtual_mode() during late init will remap
+ * the efi_memmap physical pages allocated here via memboot & then
+ * subsequently cause random EFI memmap corruption once memblock is freed.
+ */
+
+ if (efi_mem_desc_lookup(addr, &md)) {
pr_err("Failed to lookup EFI memory descriptor for %pa\n", &addr);
return;
}
+ if (md.type != EFI_BOOT_SERVICES_DATA) {
+ pr_err("Skip reserving non EFI Boot Service Data memory for %pa\n", &addr);
+ return;
+ }
+
+ /* Kexec copied the efi memmap from the first kernel, thus skip the case */
+ if (md.attribute & EFI_MEMORY_RUNTIME)
+ return;
+
if (addr + size > md.phys_addr + (md.num_pages << EFI_PAGE_SHIFT)) {
pr_err("Region spans EFI memory descriptors, %pa\n", &addr);
return;
--
2.34.1
From: Ashish Kalra <[email protected]>
Accessing guest video memory/RAM during kernel decompressor
causes guest termination as boot stage2 #VC handler for
SEV-ES/SNP systems does not support MMIO handling.
This issue is observed with SEV-ES/SNP guest kexec as
kexec -c adds screen_info to the boot parameters
passed to the kexec kernel, which causes console output to
be dumped to both video and serial.
As the decompressor output gets cleared really fast, it is
preferable to get the console output only on serial, hence,
skip accessing video RAM during decompressor stage to
prevent guest termination.
Serial console output during decompressor stage works as
boot stage2 #VC handler already supports handling port I/O.
Suggested-by: Thomas Lendacy <[email protected]>
Signed-off-by: Ashish Kalra <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
---
arch/x86/boot/compressed/misc.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/arch/x86/boot/compressed/misc.c b/arch/x86/boot/compressed/misc.c
index b70e4a21c15f..3b9f96b3dbcc 100644
--- a/arch/x86/boot/compressed/misc.c
+++ b/arch/x86/boot/compressed/misc.c
@@ -427,8 +427,10 @@ asmlinkage __visible void *extract_kernel(void *rmode, unsigned char *output)
vidport = 0x3d4;
}
- lines = boot_params_ptr->screen_info.orig_video_lines;
- cols = boot_params_ptr->screen_info.orig_video_cols;
+ if (!(sev_status & MSR_AMD64_SEV_ES_ENABLED)) {
+ lines = boot_params_ptr->screen_info.orig_video_lines;
+ cols = boot_params_ptr->screen_info.orig_video_cols;
+ }
init_default_io_ops();
--
2.34.1
From: Ashish Kalra <[email protected]>
SNP guests allocate shared buffers to perform I/O. It is done by
allocating pages normally from the buddy allocator and converting them
to shared with set_memory_decrypted().
The second kernel has no idea what memory is converted this way. It only
sees E820_TYPE_RAM.
Accessing shared memory via private mapping will cause unrecoverable RMP
page-faults.
On kexec walk direct mapping and convert all shared memory back to
private. It makes all RAM private again and second kernel may use it
normally. Additionally for SNP guests convert all bss decrypted section
pages back to private and switch back ROM regions to shared so that
their revalidation does not fail during kexec kernel boot.
The conversion occurs in two steps: stopping new conversions and
unsharing all memory. In the case of normal kexec, the stopping of
conversions takes place while scheduling is still functioning. This
allows for waiting until any ongoing conversions are finished. The
second step is carried out when all CPUs except one are inactive and
interrupts are disabled. This prevents any conflicts with code that may
access shared memory.
Signed-off-by: Ashish Kalra <[email protected]>
---
arch/x86/include/asm/sev.h | 4 +
arch/x86/kernel/sev.c | 161 ++++++++++++++++++++++++++++++++++
arch/x86/mm/mem_encrypt_amd.c | 3 +
3 files changed, 168 insertions(+)
diff --git a/arch/x86/include/asm/sev.h b/arch/x86/include/asm/sev.h
index 7f57382afee4..78d40d08d201 100644
--- a/arch/x86/include/asm/sev.h
+++ b/arch/x86/include/asm/sev.h
@@ -229,6 +229,8 @@ void snp_accept_memory(phys_addr_t start, phys_addr_t end);
u64 snp_get_unsupported_features(u64 status);
u64 sev_get_status(void);
void sev_show_status(void);
+void snp_kexec_unshare_mem(void);
+void snp_kexec_stop_conversion(bool crash);
#else
static inline void sev_es_ist_enter(struct pt_regs *regs) { }
static inline void sev_es_ist_exit(void) { }
@@ -258,6 +260,8 @@ static inline void snp_accept_memory(phys_addr_t start, phys_addr_t end) { }
static inline u64 snp_get_unsupported_features(u64 status) { return 0; }
static inline u64 sev_get_status(void) { return 0; }
static inline void sev_show_status(void) { }
+static inline void snp_kexec_unshare_mem(void) { }
+static inline void snp_kexec_stop_conversion(bool crash) { }
#endif
#ifdef CONFIG_KVM_AMD_SEV
diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
index 38ad066179d8..17f616963beb 100644
--- a/arch/x86/kernel/sev.c
+++ b/arch/x86/kernel/sev.c
@@ -42,6 +42,8 @@
#include <asm/apic.h>
#include <asm/cpuid.h>
#include <asm/cmdline.h>
+#include <asm/pgtable.h>
+#include <asm/set_memory.h>
#define DR7_RESET_VALUE 0x400
@@ -92,6 +94,9 @@ static struct ghcb *boot_ghcb __section(".data");
/* Bitmap of SEV features supported by the hypervisor */
static u64 sev_hv_features __ro_after_init;
+/* Last address to be switched to private during kexec */
+static unsigned long kexec_last_addr_to_make_private;
+
/* #VC handler runtime per-CPU data */
struct sev_es_runtime_data {
struct ghcb ghcb_page;
@@ -913,6 +918,162 @@ void snp_accept_memory(phys_addr_t start, phys_addr_t end)
set_pages_state(vaddr, npages, SNP_PAGE_STATE_PRIVATE);
}
+static bool set_pte_enc(pte_t *kpte, int level, void *va)
+{
+ pte_t new_pte;
+
+ if (pte_none(*kpte))
+ return false;
+
+ /*
+ * Change the physical page attribute from C=0 to C=1. Flush the
+ * caches to ensure that data gets accessed with the correct C-bit.
+ */
+ if (pte_present(*kpte))
+ clflush_cache_range(va, page_level_size(level));
+
+ new_pte = __pte(cc_mkenc(pte_val(*kpte)));
+ set_pte_atomic(kpte, new_pte);
+
+ return true;
+}
+
+static bool make_pte_private(pte_t *pte, unsigned long addr, int pages, int level)
+{
+ struct sev_es_runtime_data *data;
+ struct ghcb *ghcb;
+
+ data = this_cpu_read(runtime_data);
+ ghcb = &data->ghcb_page;
+
+ /* Check for GHCB for being part of a PMD range. */
+ if ((unsigned long)ghcb >= addr &&
+ (unsigned long)ghcb <= (addr + (pages * PAGE_SIZE))) {
+ /*
+ * Ensure that the current cpu's GHCB is made private
+ * at the end of unshared loop so that we continue to use the
+ * optimized GHCB protocol and not force the switch to
+ * MSR protocol till the very end.
+ */
+ pr_debug("setting boot_ghcb to NULL for this cpu ghcb\n");
+ kexec_last_addr_to_make_private = addr;
+ return true;
+ }
+
+ if (!set_pte_enc(pte, level, (void *)addr))
+ return false;
+
+ snp_set_memory_private(addr, pages);
+
+ return true;
+}
+
+static void unshare_all_memory(void)
+{
+ unsigned long addr, end;
+
+ /*
+ * Walk direct mapping and convert all shared memory back to private,
+ */
+
+ addr = PAGE_OFFSET;
+ end = PAGE_OFFSET + get_max_mapped();
+
+ while (addr < end) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ pte = lookup_address(addr, &level);
+ size = page_level_size(level);
+
+ /*
+ * pte_none() check is required to skip physical memory holes in direct mapped.
+ */
+ if (pte && pte_decrypted(*pte) && !pte_none(*pte)) {
+ int pages = size / PAGE_SIZE;
+
+ if (!make_pte_private(pte, addr, pages, level)) {
+ pr_err("Failed to unshare range %#lx-%#lx\n",
+ addr, addr + size);
+ }
+
+ }
+
+ addr += size;
+ }
+ __flush_tlb_all();
+
+}
+
+static void unshare_all_bss_decrypted_memory(void)
+{
+ unsigned long vaddr, vaddr_end;
+ unsigned int level;
+ unsigned int npages;
+ pte_t *pte;
+
+ vaddr = (unsigned long)__start_bss_decrypted;
+ vaddr_end = (unsigned long)__start_bss_decrypted_unused;
+ npages = (vaddr_end - vaddr) >> PAGE_SHIFT;
+ for (; vaddr < vaddr_end; vaddr += PAGE_SIZE) {
+ pte = lookup_address(vaddr, &level);
+ if (!pte || !pte_decrypted(*pte) || pte_none(*pte))
+ continue;
+
+ set_pte_enc(pte, level, (void *)vaddr);
+ }
+ vaddr = (unsigned long)__start_bss_decrypted;
+ snp_set_memory_private(vaddr, npages);
+}
+
+/* Stop new private<->shared conversions */
+void snp_kexec_stop_conversion(bool crash)
+{
+ /*
+ * Crash kernel reaches here with interrupts disabled: can't wait for
+ * conversions to finish.
+ *
+ * If race happened, just report and proceed.
+ */
+ bool wait_for_lock = !crash;
+
+ if (!stop_memory_enc_conversion(wait_for_lock))
+ pr_warn("Failed to finish shared<->private conversions\n");
+}
+
+void snp_kexec_unshare_mem(void)
+{
+ if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
+ return;
+
+ unshare_all_memory();
+
+ unshare_all_bss_decrypted_memory();
+
+ if (kexec_last_addr_to_make_private) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ /*
+ * Switch to using the MSR protocol to change this cpu's
+ * GHCB to private.
+ * All the per-cpu GHCBs have been switched back to private,
+ * so can't do any more GHCB calls to the hypervisor beyond
+ * this point till the kexec kernel starts running.
+ */
+ boot_ghcb = NULL;
+ sev_cfg.ghcbs_initialized = false;
+
+ pr_debug("boot ghcb 0x%lx\n", kexec_last_addr_to_make_private);
+ pte = lookup_address(kexec_last_addr_to_make_private, &level);
+ size = page_level_size(level);
+ set_pte_enc(pte, level, (void *)kexec_last_addr_to_make_private);
+ snp_set_memory_private(kexec_last_addr_to_make_private, (size / PAGE_SIZE));
+ }
+}
+
static int snp_set_vmsa(void *va, bool vmsa)
{
u64 attrs;
diff --git a/arch/x86/mm/mem_encrypt_amd.c b/arch/x86/mm/mem_encrypt_amd.c
index e7b67519ddb5..49c40c2ed809 100644
--- a/arch/x86/mm/mem_encrypt_amd.c
+++ b/arch/x86/mm/mem_encrypt_amd.c
@@ -468,6 +468,9 @@ void __init sme_early_init(void)
x86_platform.guest.enc_tlb_flush_required = amd_enc_tlb_flush_required;
x86_platform.guest.enc_cache_flush_required = amd_enc_cache_flush_required;
+ x86_platform.guest.enc_kexec_stop_conversion = snp_kexec_stop_conversion;
+ x86_platform.guest.enc_kexec_unshare_mem = snp_kexec_unshare_mem;
+
/*
* AMD-SEV-ES intercepts the RDMSR to read the X2APIC ID in the
* parallel bringup low level code. That raises #VC which cannot be
--
2.34.1
On Tue, Apr 09, 2024 at 02:29:55PM +0300, Kirill A. Shutemov wrote:
> +/* Declare CPU offlining not supported */
> +void cpu_hotplug_disable_offlining(void)
> +{
> + cpu_maps_update_begin();
"/*
* The following two APIs (cpu_maps_update_begin/done) must be used when
* attempting to serialize the updates to cpu_online_mask & cpu_present_mask.
*/
void cpu_maps_update_begin(void)
.."
> + cpu_hotplug_offline_disabled = true;
but this doesn't do that here.
Are we doing a one-off here for that variable or what?
> + cpu_maps_update_done();
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Tue, Apr 09, 2024 at 02:29:53PM +0300, Kirill A. Shutemov wrote:
> diff --git a/arch/x86/kernel/acpi/Makefile b/arch/x86/kernel/acpi/Makefile
> index fc17b3f136fe..8c7329c88a75 100644
> --- a/arch/x86/kernel/acpi/Makefile
> +++ b/arch/x86/kernel/acpi/Makefile
> @@ -1,11 +1,12 @@
> # SPDX-License-Identifier: GPL-2.0
>
> -obj-$(CONFIG_ACPI) += boot.o
> -obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
> -obj-$(CONFIG_ACPI_APEI) += apei.o
> -obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
> +obj-$(CONFIG_ACPI) += boot.o
> +obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
> +obj-$(CONFIG_ACPI_APEI) += apei.o
> +obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
> +obj-$(CONFIG_X86_ACPI_MADT_WAKEUP) += madt_wakeup.o
If you drop the "_X86" from the config symbol, you won't have to
re-align them. And the other config symbols don't have to have "_X86" in
them either because this is all in arch/x86/.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Thu, Apr 18, 2024 at 06:03:24PM +0200, Borislav Petkov wrote:
> On Tue, Apr 09, 2024 at 02:29:53PM +0300, Kirill A. Shutemov wrote:
> > diff --git a/arch/x86/kernel/acpi/Makefile b/arch/x86/kernel/acpi/Makefile
> > index fc17b3f136fe..8c7329c88a75 100644
> > --- a/arch/x86/kernel/acpi/Makefile
> > +++ b/arch/x86/kernel/acpi/Makefile
> > @@ -1,11 +1,12 @@
> > # SPDX-License-Identifier: GPL-2.0
> >
> > -obj-$(CONFIG_ACPI) += boot.o
> > -obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
> > -obj-$(CONFIG_ACPI_APEI) += apei.o
> > -obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
> > +obj-$(CONFIG_ACPI) += boot.o
> > +obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
> > +obj-$(CONFIG_ACPI_APEI) += apei.o
> > +obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
> > +obj-$(CONFIG_X86_ACPI_MADT_WAKEUP) += madt_wakeup.o
>
> If you drop the "_X86" from the config symbol, you won't have to
> re-align them. And the other config symbols don't have to have "_X86" in
> them either because this is all in arch/x86/.
Okay, fair enough. Updated patch is below.
From b020800f89ea4fce8f3698bd4ef290bba8f40b37 Mon Sep 17 00:00:00 2001
From: "Kirill A. Shutemov" <[email protected]>
Date: Fri, 1 Sep 2023 15:42:55 +0300
Subject: [PATCHv10.1 01/18] x86/acpi: Extract ACPI MADT wakeup code into a separate file
In order to prepare for the expansion of support for the ACPI MADT
wakeup method, move the relevant code into a separate file.
Introduce a new configuration option to clearly indicate dependencies
without the use of ifdefs.
There have been no functional changes.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>
Acked-by: Kai Huang <[email protected]>
Reviewed-by: Baoquan He <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/Kconfig | 7 +++
arch/x86/include/asm/acpi.h | 5 ++
arch/x86/kernel/acpi/Makefile | 1 +
arch/x86/kernel/acpi/boot.c | 86 +-----------------------------
arch/x86/kernel/acpi/madt_wakeup.c | 82 ++++++++++++++++++++++++++++
5 files changed, 96 insertions(+), 85 deletions(-)
create mode 100644 arch/x86/kernel/acpi/madt_wakeup.c
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 2dac256b6e8d..723cd5285781 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -1116,6 +1116,13 @@ config X86_LOCAL_APIC
depends on X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_APIC || PCI_MSI
select IRQ_DOMAIN_HIERARCHY
+config ACPI_MADT_WAKEUP
+ def_bool y
+ depends on X86_64
+ depends on ACPI
+ depends on SMP
+ depends on X86_LOCAL_APIC
+
config X86_IO_APIC
def_bool y
depends on X86_LOCAL_APIC || X86_UP_IOAPIC
diff --git a/arch/x86/include/asm/acpi.h b/arch/x86/include/asm/acpi.h
index f896eed4516c..2625b915ae7f 100644
--- a/arch/x86/include/asm/acpi.h
+++ b/arch/x86/include/asm/acpi.h
@@ -76,6 +76,11 @@ static inline bool acpi_skip_set_wakeup_address(void)
#define acpi_skip_set_wakeup_address acpi_skip_set_wakeup_address
+union acpi_subtable_headers;
+
+int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
+ const unsigned long end);
+
/*
* Check if the CPU can handle C2 and deeper
*/
diff --git a/arch/x86/kernel/acpi/Makefile b/arch/x86/kernel/acpi/Makefile
index fc17b3f136fe..2feba7257665 100644
--- a/arch/x86/kernel/acpi/Makefile
+++ b/arch/x86/kernel/acpi/Makefile
@@ -4,6 +4,7 @@ obj-$(CONFIG_ACPI) += boot.o
obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup_$(BITS).o
obj-$(CONFIG_ACPI_APEI) += apei.o
obj-$(CONFIG_ACPI_CPPC_LIB) += cppc.o
+obj-$(CONFIG_ACPI_MADT_WAKEUP) += madt_wakeup.o
ifneq ($(CONFIG_ACPI_PROCESSOR),)
obj-y += cstate.o
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 4bf82dbd2a6b..9f4618dcd704 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -67,13 +67,6 @@ static bool has_lapic_cpus __initdata;
static bool acpi_support_online_capable;
#endif
-#ifdef CONFIG_X86_64
-/* Physical address of the Multiprocessor Wakeup Structure mailbox */
-static u64 acpi_mp_wake_mailbox_paddr;
-/* Virtual address of the Multiprocessor Wakeup Structure mailbox */
-static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox;
-#endif
-
#ifdef CONFIG_X86_IO_APIC
/*
* Locks related to IOAPIC hotplug
@@ -341,60 +334,6 @@ acpi_parse_lapic_nmi(union acpi_subtable_headers * header, const unsigned long e
return 0;
}
-
-#ifdef CONFIG_X86_64
-static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
-{
- /*
- * Remap mailbox memory only for the first call to acpi_wakeup_cpu().
- *
- * Wakeup of secondary CPUs is fully serialized in the core code.
- * No need to protect acpi_mp_wake_mailbox from concurrent accesses.
- */
- if (!acpi_mp_wake_mailbox) {
- acpi_mp_wake_mailbox = memremap(acpi_mp_wake_mailbox_paddr,
- sizeof(*acpi_mp_wake_mailbox),
- MEMREMAP_WB);
- }
-
- /*
- * Mailbox memory is shared between the firmware and OS. Firmware will
- * listen on mailbox command address, and once it receives the wakeup
- * command, the CPU associated with the given apicid will be booted.
- *
- * The value of 'apic_id' and 'wakeup_vector' must be visible to the
- * firmware before the wakeup command is visible. smp_store_release()
- * ensures ordering and visibility.
- */
- acpi_mp_wake_mailbox->apic_id = apicid;
- acpi_mp_wake_mailbox->wakeup_vector = start_ip;
- smp_store_release(&acpi_mp_wake_mailbox->command,
- ACPI_MP_WAKE_COMMAND_WAKEUP);
-
- /*
- * Wait for the CPU to wake up.
- *
- * The CPU being woken up is essentially in a spin loop waiting to be
- * woken up. It should not take long for it wake up and acknowledge by
- * zeroing out ->command.
- *
- * ACPI specification doesn't provide any guidance on how long kernel
- * has to wait for a wake up acknowledgement. It also doesn't provide
- * a way to cancel a wake up request if it takes too long.
- *
- * In TDX environment, the VMM has control over how long it takes to
- * wake up secondary. It can postpone scheduling secondary vCPU
- * indefinitely. Giving up on wake up request and reporting error opens
- * possible attack vector for VMM: it can wake up a secondary CPU when
- * kernel doesn't expect it. Wait until positive result of the wake up
- * request.
- */
- while (READ_ONCE(acpi_mp_wake_mailbox->command))
- cpu_relax();
-
- return 0;
-}
-#endif /* CONFIG_X86_64 */
#endif /* CONFIG_X86_LOCAL_APIC */
#ifdef CONFIG_X86_IO_APIC
@@ -1124,29 +1063,6 @@ static int __init acpi_parse_madt_lapic_entries(void)
}
return 0;
}
-
-#ifdef CONFIG_X86_64
-static int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
- const unsigned long end)
-{
- struct acpi_madt_multiproc_wakeup *mp_wake;
-
- if (!IS_ENABLED(CONFIG_SMP))
- return -ENODEV;
-
- mp_wake = (struct acpi_madt_multiproc_wakeup *)header;
- if (BAD_MADT_ENTRY(mp_wake, end))
- return -EINVAL;
-
- acpi_table_print_madt_entry(&header->common);
-
- acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
-
- apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
-
- return 0;
-}
-#endif /* CONFIG_X86_64 */
#endif /* CONFIG_X86_LOCAL_APIC */
#ifdef CONFIG_X86_IO_APIC
@@ -1343,7 +1259,7 @@ static void __init acpi_process_madt(void)
smp_found_config = 1;
}
-#ifdef CONFIG_X86_64
+#ifdef CONFIG_ACPI_MADT_WAKEUP
/*
* Parse MADT MP Wake entry.
*/
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
new file mode 100644
index 000000000000..7f164d38bd0b
--- /dev/null
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -0,0 +1,82 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+#include <linux/acpi.h>
+#include <linux/io.h>
+#include <asm/apic.h>
+#include <asm/barrier.h>
+#include <asm/processor.h>
+
+/* Physical address of the Multiprocessor Wakeup Structure mailbox */
+static u64 acpi_mp_wake_mailbox_paddr;
+
+/* Virtual address of the Multiprocessor Wakeup Structure mailbox */
+static struct acpi_madt_multiproc_wakeup_mailbox *acpi_mp_wake_mailbox;
+
+static int acpi_wakeup_cpu(u32 apicid, unsigned long start_ip)
+{
+ /*
+ * Remap mailbox memory only for the first call to acpi_wakeup_cpu().
+ *
+ * Wakeup of secondary CPUs is fully serialized in the core code.
+ * No need to protect acpi_mp_wake_mailbox from concurrent accesses.
+ */
+ if (!acpi_mp_wake_mailbox) {
+ acpi_mp_wake_mailbox = memremap(acpi_mp_wake_mailbox_paddr,
+ sizeof(*acpi_mp_wake_mailbox),
+ MEMREMAP_WB);
+ }
+
+ /*
+ * Mailbox memory is shared between the firmware and OS. Firmware will
+ * listen on mailbox command address, and once it receives the wakeup
+ * command, the CPU associated with the given apicid will be booted.
+ *
+ * The value of 'apic_id' and 'wakeup_vector' must be visible to the
+ * firmware before the wakeup command is visible. smp_store_release()
+ * ensures ordering and visibility.
+ */
+ acpi_mp_wake_mailbox->apic_id = apicid;
+ acpi_mp_wake_mailbox->wakeup_vector = start_ip;
+ smp_store_release(&acpi_mp_wake_mailbox->command,
+ ACPI_MP_WAKE_COMMAND_WAKEUP);
+
+ /*
+ * Wait for the CPU to wake up.
+ *
+ * The CPU being woken up is essentially in a spin loop waiting to be
+ * woken up. It should not take long for it wake up and acknowledge by
+ * zeroing out ->command.
+ *
+ * ACPI specification doesn't provide any guidance on how long kernel
+ * has to wait for a wake up acknowledgment. It also doesn't provide
+ * a way to cancel a wake up request if it takes too long.
+ *
+ * In TDX environment, the VMM has control over how long it takes to
+ * wake up secondary. It can postpone scheduling secondary vCPU
+ * indefinitely. Giving up on wake up request and reporting error opens
+ * possible attack vector for VMM: it can wake up a secondary CPU when
+ * kernel doesn't expect it. Wait until positive result of the wake up
+ * request.
+ */
+ while (READ_ONCE(acpi_mp_wake_mailbox->command))
+ cpu_relax();
+
+ return 0;
+}
+
+int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
+ const unsigned long end)
+{
+ struct acpi_madt_multiproc_wakeup *mp_wake;
+
+ mp_wake = (struct acpi_madt_multiproc_wakeup *)header;
+ if (BAD_MADT_ENTRY(mp_wake, end))
+ return -EINVAL;
+
+ acpi_table_print_madt_entry(&header->common);
+
+ acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
+
+ apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
+
+ return 0;
+}
--
Kiryl Shutsemau / Kirill A. Shutemov
On Thu, Apr 18, 2024 at 04:37:09PM +0200, Borislav Petkov wrote:
> On Tue, Apr 09, 2024 at 02:29:55PM +0300, Kirill A. Shutemov wrote:
> > +/* Declare CPU offlining not supported */
> > +void cpu_hotplug_disable_offlining(void)
> > +{
> > + cpu_maps_update_begin();
>
> "/*
> * The following two APIs (cpu_maps_update_begin/done) must be used when
> * attempting to serialize the updates to cpu_online_mask & cpu_present_mask.
> */
> void cpu_maps_update_begin(void)
> ..."
>
> > + cpu_hotplug_offline_disabled = true;
>
> but this doesn't do that here.
>
> Are we doing a one-off here for that variable or what?
Yes, it is one-off. I guess we could use READ_ONCE()/WRITE_ONCE() to
access the variable with the same result. I am not sure why it would be
better.
--
Kiryl Shutsemau / Kirill A. Shutemov
On Fri, Apr 19, 2024 at 04:31:39PM +0300, Kirill A. Shutemov wrote:
> Yes, it is one-off. I guess we could use READ_ONCE()/WRITE_ONCE() to
> access the variable with the same result. I am not sure why it would be
> better.
Nah, and it is not even the first one-off:
cpu_hotplug_disable/_enable() uses the same locking to update
cpu_hotplug_disabled.
I guess we need to update the comment over cpu_maps_update_begin().
I guess this is fine wrt big picture of the CPU hotplug universe. Lemme
point tglx to it just in case.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Tue, Apr 09, 2024 at 02:29:56PM +0300, Kirill A. Shutemov wrote:
> ACPI MADT doesn't allow to offline CPU after it got woke up.
In all your text: s/woke/woken/g
>
> Currently CPU hotplug is prevented based on the confidential computing
> attribute which is set for Intel TDX. But TDX is not the only possible
> user of the wake up method.
>
> Disable CPU offlining on ACPI MADT wakeup enumeration.
Something's missing in that "justification". It should explain why
CC_ATTR_HOTPLUG_DISABLED is not needed anymore.
And looking at patch 3, I'm still unclear as to why this change is done.
Is it that the "ACPI MADT mailbox wakeup method" is going to be used by
TDX guests now too so that you don't need CC_ATTR_HOTPLUG_DISABLED
anymore?
It seems that if acpi_parse_mp_wake() finds an ok wakeup entry, then
offlining is disabled...
Or is it something else?
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Tue, Apr 23, 2024 at 06:02:58PM +0200, Borislav Petkov wrote:
> >
> > Currently CPU hotplug is prevented based on the confidential computing
> > attribute which is set for Intel TDX. But TDX is not the only possible
> > user of the wake up method.
> >
> > Disable CPU offlining on ACPI MADT wakeup enumeration.
>
> Something's missing in that "justification". It should explain why
> CC_ATTR_HOTPLUG_DISABLED is not needed anymore.
It was wrong from beginning. If ACPI MADT wake up method is used on the
platform, we cannot handle offline, regardless if it is TDX or not.
> And looking at patch 3, I'm still unclear as to why this change is done.
> Is it that the "ACPI MADT mailbox wakeup method" is going to be used by
> TDX guests now too so that you don't need CC_ATTR_HOTPLUG_DISABLED
> anymore?
ACPI MADT is the only wakeup method supported in TDX guests. But offline
is broken is because of ACPI MADT, not because of TDX.
--
Kiryl Shutsemau / Kirill A. Shutemov
On Wed, Apr 24, 2024 at 11:38:42AM +0300, Kirill A. Shutemov wrote:
> It was wrong from beginning. If ACPI MADT wake up method is used on the
> platform, we cannot handle offline, regardless if it is TDX or not.
Sounds to me like this fact should be a prominent part of the commit
message and these 1-4 patches should be carved out as a separate set
fixing that ACPI MADT thing and I should take them separately now...?
Also, does this need to go to stable although it is kinda big for
stable. If stable, do we need a smaller fix first which is backportable?
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Wed, Apr 24, 2024 at 03:50:52PM +0200, Borislav Petkov wrote:
> On Wed, Apr 24, 2024 at 11:38:42AM +0300, Kirill A. Shutemov wrote:
> > It was wrong from beginning. If ACPI MADT wake up method is used on the
> > platform, we cannot handle offline, regardless if it is TDX or not.
>
> Sounds to me like this fact should be a prominent part of the commit
> message and these 1-4 patches should be carved out as a separate set
> fixing that ACPI MADT thing and I should take them separately now...?
>
> Also, does this need to go to stable although it is kinda big for
> stable. If stable, do we need a smaller fix first which is backportable?
Correct me, if I am wrong, but I believe TDX guest is the only user of
ACPI MADT wake up method. At least it was added into kernel for TDX guest.
So it wouldn't fix anything user-visible. It might affect a future
platform that uses this wake up method, but it is a guessing game.
--
Kiryl Shutsemau / Kirill A. Shutemov
On 4/24/24 07:35, Kirill A. Shutemov wrote:
>> Also, does this need to go to stable although it is kinda big for
>> stable. If stable, do we need a smaller fix first which is backportable?
> Correct me, if I am wrong, but I believe TDX guest is the only user of
> ACPI MADT wake up method. At least it was added into kernel for TDX guest.
> So it wouldn't fix anything user-visible. It might affect a future
> platform that uses this wake up method, but it is a guessing game.
Yeah, the MADT wakeup is pretty funky, highly TDX specific, and not in
use or planned to be _put_ into use anywhere else. The X86S "fix" for
running in real mode does something else entirely:
https://cdrdv2.intel.com/v1/dl/getContent/776648
On Mon, Apr 15, 2024 at 11:22:58PM +0000, Ashish Kalra wrote:
> From: Ashish Kalra <[email protected]>
>
> For kexec use case, need to use and stick to the EFI memmap passed
> from the first kernel via boot-params/setup data, hence,
> skip efi_arch_mem_reserve() during kexec.
Please use this or similar scheme when formulating your commit messages.
This above is too laconic.
1. Prepare the context for the explanation briefly.
2. Explain the problem at hand.
3. "It happens because of <...>"
4. "Fix it by doing X"
5. "(Potentially do Y)."
And some of those above are optional depending on the issue being
explained.
For more detailed info, see
Documentation/process/submitting-patches.rst,
Section "2) Describe your changes".
> Additionally during SNP guest kexec testing discovered that EFI memmap
> is corrupted during chained kexec.
That sentence needs sanitization.
> kexec_enter_virtual_mode() during late init will remap the efi_memmap
> physical pages allocated in efi_arch_mem_reserve() via memblock & then
s/&/and/
This is not code. Please take a greater care when writing commit
messages - they're not write-only.
> subsequently cause random EFI memmap corruption once memblock is
> freed/teared-down.
"torn down"
> Suggested-by: Dave Young <[email protected]>
> [Dave Young: checking the md attribute instead of checking the efi_setup]
> Signed-off-by: Ashish Kalra <[email protected]>
> ---
> arch/x86/platform/efi/quirks.c | 20 ++++++++++++++++++--
> 1 file changed, 18 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
> index f0cc00032751..982f5e50a4b3 100644
> --- a/arch/x86/platform/efi/quirks.c
> +++ b/arch/x86/platform/efi/quirks.c
> @@ -258,12 +258,28 @@ void __init efi_arch_mem_reserve(phys_addr_t addr, u64 size)
> int num_entries;
> void *new;
>
> - if (efi_mem_desc_lookup(addr, &md) ||
> - md.type != EFI_BOOT_SERVICES_DATA) {
> + /*
> + * For kexec use case, we need to use the EFI memmap passed from the first
Make all your text impersonal - no "we", "I", etc.
> + * kernel via setup data, so we need to skip this.
What exactly do we need to skip?
If the EFI memory descriptor lookup fails?
> + * Additionally kexec_enter_virtual_mode() during late init will remap
> + * the efi_memmap physical pages allocated here via memboot & then
> + * subsequently cause random EFI memmap corruption once memblock is freed.
> + */
Why is that comment here and what is its relevance to the line it is
above of?
> + if (efi_mem_desc_lookup(addr, &md)) {
> pr_err("Failed to lookup EFI memory descriptor for %pa\n", &addr);
> return;
> }
>
> + if (md.type != EFI_BOOT_SERVICES_DATA) {
> + pr_err("Skip reserving non EFI Boot Service Data memory for %pa\n", &addr);
What is this pr_err() useful for?
> + return;
> + }
> +
> + /* Kexec copied the efi memmap from the first kernel, thus skip the case */
kexec? This is a generic function - what does it have to do with kexec?
The subject of this patch is:
Subject: [PATCH v5 1/3] efi/x86: skip efi_arch_mem_reserve() in case of kexec
and yet, nothing skips this function - it adds a bunch of checks,
printks and early returns with the intent that those early returns
happen on kexec and thus the actual memremap doesn't happen there.
So it is some sort of: let's check things which will be true in
a kexec-ed kernel and thus avoid the function by returning early.
But I have no clue.
It sounds to me like you need to go back up, to the 10000ft view and
explain how exactly this efi_mem_reserve() causes trouble for the
kexec-ed kernel so that we can think of a proper solution, not some
random hackery.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Wed, Apr 24, 2024 at 07:40:26AM -0700, Dave Hansen wrote:
> On 4/24/24 07:35, Kirill A. Shutemov wrote:
> >> Also, does this need to go to stable although it is kinda big for
> >> stable. If stable, do we need a smaller fix first which is backportable?
> > Correct me, if I am wrong, but I believe TDX guest is the only user of
> > ACPI MADT wake up method. At least it was added into kernel for TDX guest.
> > So it wouldn't fix anything user-visible. It might affect a future
> > platform that uses this wake up method, but it is a guessing game.
>
> Yeah, the MADT wakeup is pretty funky, highly TDX specific, and not in
> use or planned to be _put_ into use anywhere else.
Then please make sure all that info is in the commit messages in the
next revision. Because as they are now, they're not even beginning to
explain what exactly is this fixing. /me thinking that this is some
generic fix needed in stable is case-in-point.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
Hello Boris,
On 4/24/2024 9:48 AM, Borislav Petkov wrote:
> On Mon, Apr 15, 2024 at 11:22:58PM +0000, Ashish Kalra wrote:
>> From: Ashish Kalra <[email protected]>
>>
>> For kexec use case, need to use and stick to the EFI memmap passed
>> from the first kernel via boot-params/setup data, hence,
>> skip efi_arch_mem_reserve() during kexec.
> Please use this or similar scheme when formulating your commit messages.
> This above is too laconic.
>
> 1. Prepare the context for the explanation briefly.
>
> 2. Explain the problem at hand.
>
> 3. "It happens because of <...>"
>
> 4. "Fix it by doing X"
>
> 5. "(Potentially do Y)."
>
> And some of those above are optional depending on the issue being
> explained.
>
> For more detailed info, see
> Documentation/process/submitting-patches.rst,
> Section "2) Describe your changes".
Here is the more detailed description of the issue:
With SNP guest kexec and during nested guest kexec, observe the
following efi memmap corruption :
[Â Â Â 0.000000] efi: EFI v2.7 by EDK II^M
[Â Â Â 0.000000] efi: SMBIOS=0x7e33f000 SMBIOS 3.0=0x7e33d000
ACPI=0x7e57e000 ACPI 2.0=0x7e57e014 MEMATTR=0x7cc3c018
Unaccepted=0x7c09e018 ^M
[Â Â Â 0.000000] efi: [Firmware Bug]: Invalid EFI memory map entries:^M
[Â Â Â 0.000000] efi: mem03: [type=269370880|attr=0x0e42100e42180e41]
range=[0x0486200e41038c18-0x200e898a0eee713ac17] (invalid)^M
[Â Â Â 0.000000] efi: mem04: [type=12336|attr=0x0e410686300e4105]
range=[0x100e420000000176-0x8c290f26248d200e175] (invalid)^M
[Â Â Â 0.000000] efi: mem06: [type=1124304408|attr=0x000030b400000028]
range=[0x0e51300e45280e77-0xb44ed2142f460c1e76] (invalid)^M
[Â Â Â 0.000000] efi: mem08: [type=68|attr=0x300e540583280e41]
range=[0x0000011affff3cd8-0x486200e54b38c0bcd7] (invalid)^M
[Â Â Â 0.000000] efi: mem10: [type=1107529240|attr=0x0e42280e41300e41]
range=[0x300e41058c280e42-0x38010ae54c5c328ee41] (invalid)^M
[Â Â Â 0.000000] efi: mem11: [type=189335566|attr=0x048d200e42038e18]
range=[0x0000318c00000048-0xe42029228ce4200047] (invalid)^M
[Â Â Â 0.000000] efi: mem12: [type=239142534|attr=0x0000002400000b4b]
range=[0x0e41380e0a7d700e-0x80f26238f22bfe500d] (invalid)^M
[Â Â Â 0.000000] efi: mem14: [type=239207055|attr=0x0e41300e43380e0a]
range=[0x8c280e42048d200e-0xc70b028f2f27cc0a00d] (invalid)^M
[Â Â Â 0.000000] efi: mem15: [type=239210510|attr=0x00080e660b47080e]
range=[0x0000324c0000001c-0xa78028634ce490001b] (invalid)^M
[Â Â Â 0.000000] efi: mem16: [type=4294848528|attr=0x0000329400000014]
range=[0x0e410286100e4100-0x80f252036a218f20ff] (invalid)^M
[Â Â Â 0.000000] efi: mem19: [type=2250772033|attr=0x42180e42200e4328]
range=[0x41280e0ab9020683-0xe0e538c28b39e62682] (invalid)^M
[Â Â Â 0.000000] efi: mem20: [type=16|Â Â |Â |Â |Â |Â |Â |Â |Â |Â | |WB|Â
|WC|Â ] range=[0x00000008ffff4438-0xffff44340090333c437] (invalid)^M
[   0.000000] efi: mem22: [Reserved   |attr=0x000000c1ffff4420]
range=[0xffff442400003398-0x1033a04240003f397] (invalid)^M
[Â Â Â 0.000000] efi: mem23: [type=1141080856|attr=0x080e41100e43180e]
range=[0x280e66300e4b280e-0x440dc5ee7141f4c080d] (invalid)^M
[   0.000000] efi: mem25: [Reserved   |attr=0x0000000affff44a0]
range=[0xffff44a400003428-0x1034304a400013427] (invalid)^M
[Â Â Â 0.000000] efi: mem28: [type=16|Â Â |Â |Â |Â |Â |Â |Â |Â |Â | |WB|Â
|WC|Â ] range=[0x0000000affff4488-0xffff448400b034bc487] (invalid)^M
[   0.000000] efi: mem30: [Reserved   |attr=0x0000000affff4470]
range=[0xffff447400003518-0x10352047400013517] (invalid)^M
[Â Â Â 0.000000] efi: mem33: [type=16|Â Â |Â |Â |Â |Â |Â |Â |Â |Â | |WB|Â
|WC|Â ] range=[0x0000000affff4458-0xffff445400b035ac457] (invalid)^M
[Â Â Â 0.000000] efi: mem35: [type=269372416|attr=0x0e42100e42180e41]
range=[0x0486200e44038c18-0x200e8b8a0eee823ac17] (invalid)^M
[Â Â Â 0.000000] efi: mem37: [type=2351435330|attr=0x0e42100e42180e42]
range=[0x470783380e410686-0x2002b2a041c2141e685] (invalid)^M
[Â Â Â 0.000000] efi: mem38: [type=1093668417|attr=0x100e420000000270]
range=[0x42100e42180e4220-0xfff366a4e421b78c21f] (invalid)^M
[Â Â Â 0.000000] efi: mem39: [type=76357646|attr=0x180e42200e42280e]
range=[0x0e410686300e4105-0x4130f251a0710ae5104] (invalid)^M
[Â Â Â 0.000000] efi: mem40: [type=940444268|attr=0x0e42200e42280e41]
range=[0x180e42200e42280e-0x300fc71c300b4f2480d] (invalid)^M
[Â Â Â 0.000000] efi: mem41: [MMIOÂ Â Â Â Â Â Â |attr=0x8c280e42048d200e]
range=[0xffff479400003728-0x42138e0c87820292727] (invalid)^M
[Â Â Â 0.000000] efi: mem42: [type=1191674680|attr=0x0000004c0000000b]
range=[0x300e41380e0a0246-0x470b0f26238f22b8245] (invalid)^M
[Â Â Â 0.000000] efi: mem43: [type=2010|attr=0x0301f00e4d078338]
range=[0x45038e180e42028f-0xe4556bf118f282528e] (invalid)^M
[Â Â Â 0.000000] efi: mem44: [type=1109921345|attr=0x300e44000000006c]
range=[0x44080e42100e4218-0xfff39254e42138ac217] (invalid)^M
[Â Â Â 0.000000] efi: mem45: [type=40|attr=0x0e41100e41180e0a]
range=[0x0000008affff5228-0x4702400e53b3830d227] (invalid)^M
[Â Â Â 0.000000] efi: mem47: [type=1107529240|attr=0x42280e41300e4138]
range=[0x300e44058c280e42-0xe0d049a435c728ee41] (invalid)^M
..
This EFI memap corruption is happening during efi_arch_mem_reserve()
invocation with the previous kexec-ed kernel boot.
( efi_arch_mem_reserve() is invoked with the following call-stack: )
[Â Â Â 0.310010]Â efi_arch_mem_reserve+0xb1/0x220^M
[Â Â Â 0.310686]Â ? memblock_add_range+0x2a0/0x2e0^M
[Â Â Â 0.311382]Â efi_mem_reserve+0x36/0x60^M
[Â Â Â 0.311973]Â efi_bgrt_init+0x17d/0x1a0^M
[Â Â Â 0.312565]Â ? __pfx_acpi_parse_bgrt+0x10/0x10^M
[Â Â Â 0.313265]Â acpi_parse_bgrt+0x12/0x20^M
[Â Â Â 0.313858]Â acpi_table_parse+0x77/0xd0^M
[Â Â Â 0.314463]Â acpi_boot_init+0x362/0x630^M
[Â Â Â 0.315069]Â setup_arch+0xa88/0xf80^M
[Â Â Â 0.315629]Â start_kernel+0x68/0xa90^M
[Â Â Â 0.316194]Â x86_64_start_reservations+0x1c/0x30^M
[Â Â Â 0.316921]Â x86_64_start_kernel+0xbf/0x110^M
[Â Â Â 0.317582]Â common_startup_64+0x13e/0x141^M
[Â Â Â 0.318231]Â </TASK>^M
Now, efi_arch_mem_reserve() calls efi_memmap_alloc() to allocate memory
for EFI memory map and due to early allocation it uses memblock allocation.
Later in the boot flow, efi_enter_virtual_mode() calls
kexec_enter_virtual_mode() in case of a kexec-ed kernel boot.
This function kexec_enter_virtual_mode() installs the new EFI memory map
by calling efi_memmap_init_late() which remaps the efi_memmap physically
allocated above in efi_arch_mem_reserve(), but please note that this
remapping is still using memblock allocation.
Subsequently, when memblock is freed later in boot flow, the above
remapped efi_memmap will have random corruption (similar to a
use-after-free scenario).
This corrupted EFI memory map is then passed to the next kexec-ed kernel
which causes a panic when trying to use the corrupted EFI memory map.
>> Additionally during SNP guest kexec testing discovered that EFI memmap
>> is corrupted during chained kexec.
> That sentence needs sanitization.
>
>> kexec_enter_virtual_mode() during late init will remap the efi_memmap
>> physical pages allocated in efi_arch_mem_reserve() via memblock & then
> s/&/and/
>
> This is not code. Please take a greater care when writing commit
> messages - they're not write-only.
>
>> subsequently cause random EFI memmap corruption once memblock is
>> freed/teared-down.
> "torn down"
>
>> Suggested-by: Dave Young <[email protected]>
>> [Dave Young: checking the md attribute instead of checking the efi_setup]
>> Signed-off-by: Ashish Kalra <[email protected]>
>> ---
>> arch/x86/platform/efi/quirks.c | 20 ++++++++++++++++++--
>> 1 file changed, 18 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
>> index f0cc00032751..982f5e50a4b3 100644
>> --- a/arch/x86/platform/efi/quirks.c
>> +++ b/arch/x86/platform/efi/quirks.c
>> @@ -258,12 +258,28 @@ void __init efi_arch_mem_reserve(phys_addr_t addr, u64 size)
>> int num_entries;
>> void *new;
>>
>> - if (efi_mem_desc_lookup(addr, &md) ||
>> - md.type != EFI_BOOT_SERVICES_DATA) {
>> + /*
>> + * For kexec use case, we need to use the EFI memmap passed from the first
> Make all your text impersonal - no "we", "I", etc.
>
>> + * kernel via setup data, so we need to skip this.
> What exactly do we need to skip?
>
> If the EFI memory descriptor lookup fails?
>
>> + * Additionally kexec_enter_virtual_mode() during late init will remap
>> + * the efi_memmap physical pages allocated here via memboot & then
>> + * subsequently cause random EFI memmap corruption once memblock is freed.
>> + */
> Why is that comment here and what is its relevance to the line it is
> above of?
>
>> + if (efi_mem_desc_lookup(addr, &md)) {
>> pr_err("Failed to lookup EFI memory descriptor for %pa\n", &addr);
>> return;
>> }
>>
>> + if (md.type != EFI_BOOT_SERVICES_DATA) {
>> + pr_err("Skip reserving non EFI Boot Service Data memory for %pa\n", &addr);
> What is this pr_err() useful for?
>
>> + return;
>> + }
>> +
>> + /* Kexec copied the efi memmap from the first kernel, thus skip the case */
> kexec? This is a generic function - what does it have to do with kexec?
>
> The subject of this patch is:
>
> Subject: [PATCH v5 1/3] efi/x86: skip efi_arch_mem_reserve() in case of kexec
>
> and yet, nothing skips this function - it adds a bunch of checks,
> printks and early returns with the intent that those early returns
> happen on kexec and thus the actual memremap doesn't happen there.
>
> So it is some sort of: let's check things which will be true in
> a kexec-ed kernel and thus avoid the function by returning early.
>
> But I have no clue.
>
> It sounds to me like you need to go back up, to the 10000ft view and
> explain how exactly this efi_mem_reserve() causes trouble for the
> kexec-ed kernel so that we can think of a proper solution, not some
> random hackery.
The above details explain why and how efi_arch_mem_reserve() causes
trouble for the (nested) kexec-ed kernel, additionally, there is a
another reason to skip efi_arch_mem_reserve() altogether for the kexec
case, as for kexec use case we need to use the EFI memmap passed from
the 1st kernel via setup_data and probably need to avoid any additional
EFI memory map additions/updates.
Therefore, the first revision of this patch had the following code to
skip efi_arch_mem_reserve():
void __init efi_arch_mem_reserve(..) {
+ if (efi_setup) + return;
But then based on upstream review/feedback, the second revision of this
patch, updated the patch to check the md attribute of the EFI memory
descriptor instead of checking for efi_setup for detecting if running
under kexec kernel and the checking of the md attribute of the EFI
memory descriptor introduces these additional checks and pr_err() which
you commented on above.
Hopefully, the above detailed explanation captures the reason to skip
efi_arch_mem_reserve() in case of (SNP) guest kexec, looking forward to
your comments/feedback on the same for me to rework this patch
(especially the commit message) and post it again.
Thanks, Ashish
>It sounds to me like you need to go back up, to the 10000ft view and
>> explain how exactly this efi_mem_reserve() causes trouble for the
>> kexec-ed kernel so that we can think of a proper solution, not some
>> random hackery.
>
> The above details explain why and how efi_arch_mem_reserve() causes
> trouble for the (nested) kexec-ed kernel, additionally, there is a
> another reason to skip efi_arch_mem_reserve() altogether for the kexec
> case, as for kexec use case we need to use the EFI memmap passed from
> the 1st kernel via setup_data and probably need to avoid any
> additional EFI memory map additions/updates.
>
> Therefore, the first revision of this patch had the following code to
> skip efi_arch_mem_reserve():
>
> void __init efi_arch_mem_reserve(..) {
>
> + if (efi_setup) + return;
>
> But then based on upstream review/feedback, the second revision of
> this patch, updated the patch to check the md attribute of the EFI
> memory descriptor instead of checking for efi_setup for detecting if
> running under kexec kernel and the checking of the md attribute of the
> EFI memory descriptor introduces these additional checks and pr_err()
> which you commented on above.
>
> Hopefully, the above detailed explanation captures the reason to skip
> efi_arch_mem_reserve() in case of (SNP) guest kexec, looking forward
> to your comments/feedback on the same for me to rework this patch
> (especially the commit message) and post it again.
<snip>
I am actually going to rename this patch to something more appropriate like:
Fix EFI memory map corruption during SNP guest kexec
And in the patch itself, go back to skipping efi_arch_mem_reserve() by
checking efi_setup to check for running under kexec kernel similar to
how it used by efi_enter_virtual_mode().
Thanks, Ashish
On Wed, Apr 24, 2024 at 04:17:09PM -0500, Kalra, Ashish wrote:
> With SNP guest kexec and during nested guest kexec, observe the following
> efi memmap corruption :
Before we delve any deeper here, lemme make sure I understand this
correctly:
* You're in a SNP guest and you're kexec-ing into a new kernel?
or
* You have a plain hypervisor which runs a non-CoCo guest and that guest
is a hypervisor too and it starts a level 2 guest and *in* *that* level
2 guest you kexec a kernel?
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On 4/26/2024 9:21 AM, Borislav Petkov wrote:
> On Wed, Apr 24, 2024 at 04:17:09PM -0500, Kalra, Ashish wrote:
>> With SNP guest kexec and during nested guest kexec, observe the following
>> efi memmap corruption :
> Before we delve any deeper here, lemme make sure I understand this
> correctly:
>
> * You're in a SNP guest and you're kexec-ing into a new kernel?
Yes this is the case, where i am in a SNP guest and kexec-ing into a new
kernel.
I should have mentioned *chained* guest kexec above instead of nested
guest kexec.
Thanks, Ashish
On 4/26/2024 10:22 AM, Borislav Petkov wrote:
> On Fri, Apr 26, 2024 at 09:47:02AM -0500, Kalra, Ashish wrote:
>> I should have mentioned *chained* guest kexec above instead of nested guest
>> kexec.
> What is a "chained guest kexec" now?
"Chained guest kexec" is when we are in a guest and kexec-ing into a new
kernel and then this kernel kexecs into another and so on ...
Thanks, Ashish
On Fri, Apr 26, 2024 at 10:28:41AM -0500, Kalra, Ashish wrote:
> "Chained guest kexec" is when we are in a guest and kexec-ing into a new
> kernel and then this kernel kexecs into another and so on ...
Make sure to explain your terminology:
$ git grep -rE "chained.*kexec"
$
and there's nothing "chained" about it - you're simply kexec-ing in
a loop.
Please don't make it sound more complicated than it is.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Fri, Apr 26, 2024 at 09:47:02AM -0500, Kalra, Ashish wrote:
> I should have mentioned *chained* guest kexec above instead of nested guest
> kexec.
What is a "chained guest kexec" now?
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On 4/26/2024 10:34 AM, Borislav Petkov wrote:
> On Fri, Apr 26, 2024 at 10:28:41AM -0500, Kalra, Ashish wrote:
>> "Chained guest kexec" is when we are in a guest and kexec-ing into a new
>> kernel and then this kernel kexecs into another and so on ...
> Make sure to explain your terminology:
>
> $ git grep -rE "chained.*kexec"
> $
>
> and there's nothing "chained" about it - you're simply kexec-ing in
> a loop.
>
> Please don't make it sound more complicated than it is.
I have posted another version of this patch series (v6) with
restructured and updated commit message for this patch, so please review
this v6 series.
Thanks, Ashish
From: Ashish Kalra <[email protected]>
The patchset adds bits and pieces to get kexec (and crashkernel) work on
SNP guest.
The series is based off of and tested against Kirill Shutemov's tree:
https://github.com/intel/tdx.git guest-kexec
----
v6:
- Updated and restructured the commit message for patch 1/3 to
explain the issue in detail.
- Updated inline comments in patch 1/3 to explain the issue in
detail.
- Moved back to checking efi_setup for detecting if running
under kexec kernel.
v5:
- Removed sev_es_enabled() function and using sev_status directly to
check for SEV-ES/SEV-SNP guest.
- used --base option to generate patches to specify Kirill's TDX guest
kexec patches as prerequisite patches to fix kernel test robot
build errors.
v4:
- Rebased to current tip/master.
- Reviewed-bys from Sathya.
- Remove snp_kexec_unprep_rom_memory() as it is not needed any more as
SEV-SNP code is not validating the ROM range in probe_roms() anymore.
- Fix kernel test robot build error/warnings.
v3:
- Rebased;
- moved Keep page tables that maps E820_TYPE_ACPI patch to Kirill's tdx
guest kexec patch series.
- checking the md attribute instead of checking the efi_setup for
detecting if running under kexec kernel.
- added new sev_es_enabled() function.
- skip video memory access in decompressor for SEV-ES/SNP systems to
prevent guest termination as boot stage2 #VC handler does not handle
MMIO.
v2:
- address zeroing of unaccepted memory table mappings at all page table levels
adding phys_pte_init(), phys_pud_init() and phys_p4d_init().
- include skip efi_arch_mem_reserve() in case of kexec as part of this
patch set.
- rename last_address_shd_kexec to a more appropriate
kexec_last_address_to_make_private.
- remove duplicate code shared with TDX and use common interfaces
defined for SNP and TDX for kexec/kdump.
- remove set_pte_enc() dependency on pg_level_to_pfn() and make the
function simpler.
- rename unshare_pte() to make_pte_private().
- clarify and make the comment for using kexec_last_address_to_make_private
more understandable.
- general cleanup.
Ashish Kalra (3):
efi/x86: Fix EFI memory map corruption with kexec
x86/boot/compressed: Skip Video Memory access in Decompressor for
SEV-ES/SNP.
x86/snp: Convert shared memory back to private on kexec
arch/x86/boot/compressed/misc.c | 6 +-
arch/x86/include/asm/sev.h | 4 +
arch/x86/kernel/sev.c | 161 ++++++++++++++++++++++++++++++++
arch/x86/mm/mem_encrypt_amd.c | 3 +
arch/x86/platform/efi/quirks.c | 20 ++++
5 files changed, 192 insertions(+), 2 deletions(-)
base-commit: 7fcd76de8a7bc12e930ef383a157ce99d711715d
prerequisite-patch-id: a911f230c2524bd791c47f62f17f0a93cbf726b6
prerequisite-patch-id: bfe2fa046349978ac1825275eb205acecfbc22f3
prerequisite-patch-id: 5e60d292457c7cd98fd3e45c23127e9463b56a69
prerequisite-patch-id: 1f97d0a2edb7509dd58276f628d1a4bda62c154c
prerequisite-patch-id: cbc2507b5c2810c3015aaf836d774d32f969c19a
prerequisite-patch-id: cbdfea1e50ecb3b4cee3a25a27df4d35bd95d532
prerequisite-patch-id: 99382c42348b9a076ba930eca0dfc9d000ec951d
prerequisite-patch-id: 469a0a3c78b0eca82527cd85e2205fb8fb89d645
prerequisite-patch-id: 2974ef211db5253d9782018e352d2a6ff0b0ef54
prerequisite-patch-id: 2cfffd80947941892421dae99b7fa0f9f9715884
prerequisite-patch-id: 466c2cb9f0a107bbd1dbd8526f4eff2bdb55f1ce
prerequisite-patch-id: d4966ae63e86d24b0bf578da4dae871cd9002b12
prerequisite-patch-id: fccde6f1fa385b5af0195f81fcb95acd71822428
prerequisite-patch-id: 16048ee15e392b0b9217b8923939b0059311abd2
prerequisite-patch-id: 5c9ae9aa294f72f63ae2c3551507dfbd92525803
prerequisite-patch-id: 758bdb686290c018cbd5b7d005354019f9d15248
prerequisite-patch-id: c85fd0bb6d183a40da73720eaa607481b1d51daf
prerequisite-patch-id: 60760e0c98ab7ccd2ca22ae3e9f20ff5a94c6e91
--
2.34.1
From: Ashish Kalra <[email protected]>
SNP guests allocate shared buffers to perform I/O. It is done by
allocating pages normally from the buddy allocator and converting them
to shared with set_memory_decrypted().
The second kernel has no idea what memory is converted this way. It only
sees E820_TYPE_RAM.
Accessing shared memory via private mapping will cause unrecoverable RMP
page-faults.
On kexec walk direct mapping and convert all shared memory back to
private. It makes all RAM private again and second kernel may use it
normally. Additionally for SNP guests convert all bss decrypted section
pages back to private and switch back ROM regions to shared so that
their revalidation does not fail during kexec kernel boot.
The conversion occurs in two steps: stopping new conversions and
unsharing all memory. In the case of normal kexec, the stopping of
conversions takes place while scheduling is still functioning. This
allows for waiting until any ongoing conversions are finished. The
second step is carried out when all CPUs except one are inactive and
interrupts are disabled. This prevents any conflicts with code that may
access shared memory.
Signed-off-by: Ashish Kalra <[email protected]>
---
arch/x86/include/asm/sev.h | 4 +
arch/x86/kernel/sev.c | 161 ++++++++++++++++++++++++++++++++++
arch/x86/mm/mem_encrypt_amd.c | 3 +
3 files changed, 168 insertions(+)
diff --git a/arch/x86/include/asm/sev.h b/arch/x86/include/asm/sev.h
index 7f57382afee4..78d40d08d201 100644
--- a/arch/x86/include/asm/sev.h
+++ b/arch/x86/include/asm/sev.h
@@ -229,6 +229,8 @@ void snp_accept_memory(phys_addr_t start, phys_addr_t end);
u64 snp_get_unsupported_features(u64 status);
u64 sev_get_status(void);
void sev_show_status(void);
+void snp_kexec_unshare_mem(void);
+void snp_kexec_stop_conversion(bool crash);
#else
static inline void sev_es_ist_enter(struct pt_regs *regs) { }
static inline void sev_es_ist_exit(void) { }
@@ -258,6 +260,8 @@ static inline void snp_accept_memory(phys_addr_t start, phys_addr_t end) { }
static inline u64 snp_get_unsupported_features(u64 status) { return 0; }
static inline u64 sev_get_status(void) { return 0; }
static inline void sev_show_status(void) { }
+static inline void snp_kexec_unshare_mem(void) { }
+static inline void snp_kexec_stop_conversion(bool crash) { }
#endif
#ifdef CONFIG_KVM_AMD_SEV
diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
index 995f94467101..891257fde810 100644
--- a/arch/x86/kernel/sev.c
+++ b/arch/x86/kernel/sev.c
@@ -42,6 +42,8 @@
#include <asm/apic.h>
#include <asm/cpuid.h>
#include <asm/cmdline.h>
+#include <asm/pgtable.h>
+#include <asm/set_memory.h>
#define DR7_RESET_VALUE 0x400
@@ -92,6 +94,9 @@ static struct ghcb *boot_ghcb __section(".data");
/* Bitmap of SEV features supported by the hypervisor */
static u64 sev_hv_features __ro_after_init;
+/* Last address to be switched to private during kexec */
+static unsigned long kexec_last_addr_to_make_private;
+
/* #VC handler runtime per-CPU data */
struct sev_es_runtime_data {
struct ghcb ghcb_page;
@@ -913,6 +918,162 @@ void snp_accept_memory(phys_addr_t start, phys_addr_t end)
set_pages_state(vaddr, npages, SNP_PAGE_STATE_PRIVATE);
}
+static bool set_pte_enc(pte_t *kpte, int level, void *va)
+{
+ pte_t new_pte;
+
+ if (pte_none(*kpte))
+ return false;
+
+ /*
+ * Change the physical page attribute from C=0 to C=1. Flush the
+ * caches to ensure that data gets accessed with the correct C-bit.
+ */
+ if (pte_present(*kpte))
+ clflush_cache_range(va, page_level_size(level));
+
+ new_pte = __pte(cc_mkenc(pte_val(*kpte)));
+ set_pte_atomic(kpte, new_pte);
+
+ return true;
+}
+
+static bool make_pte_private(pte_t *pte, unsigned long addr, int pages, int level)
+{
+ struct sev_es_runtime_data *data;
+ struct ghcb *ghcb;
+
+ data = this_cpu_read(runtime_data);
+ ghcb = &data->ghcb_page;
+
+ /* Check for GHCB for being part of a PMD range. */
+ if ((unsigned long)ghcb >= addr &&
+ (unsigned long)ghcb <= (addr + (pages * PAGE_SIZE))) {
+ /*
+ * Ensure that the current cpu's GHCB is made private
+ * at the end of unshared loop so that we continue to use the
+ * optimized GHCB protocol and not force the switch to
+ * MSR protocol till the very end.
+ */
+ pr_debug("setting boot_ghcb to NULL for this cpu ghcb\n");
+ kexec_last_addr_to_make_private = addr;
+ return true;
+ }
+
+ if (!set_pte_enc(pte, level, (void *)addr))
+ return false;
+
+ snp_set_memory_private(addr, pages);
+
+ return true;
+}
+
+static void unshare_all_memory(void)
+{
+ unsigned long addr, end;
+
+ /*
+ * Walk direct mapping and convert all shared memory back to private,
+ */
+
+ addr = PAGE_OFFSET;
+ end = PAGE_OFFSET + get_max_mapped();
+
+ while (addr < end) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ pte = lookup_address(addr, &level);
+ size = page_level_size(level);
+
+ /*
+ * pte_none() check is required to skip physical memory holes in direct mapped.
+ */
+ if (pte && pte_decrypted(*pte) && !pte_none(*pte)) {
+ int pages = size / PAGE_SIZE;
+
+ if (!make_pte_private(pte, addr, pages, level)) {
+ pr_err("Failed to unshare range %#lx-%#lx\n",
+ addr, addr + size);
+ }
+
+ }
+
+ addr += size;
+ }
+ __flush_tlb_all();
+
+}
+
+static void unshare_all_bss_decrypted_memory(void)
+{
+ unsigned long vaddr, vaddr_end;
+ unsigned int level;
+ unsigned int npages;
+ pte_t *pte;
+
+ vaddr = (unsigned long)__start_bss_decrypted;
+ vaddr_end = (unsigned long)__start_bss_decrypted_unused;
+ npages = (vaddr_end - vaddr) >> PAGE_SHIFT;
+ for (; vaddr < vaddr_end; vaddr += PAGE_SIZE) {
+ pte = lookup_address(vaddr, &level);
+ if (!pte || !pte_decrypted(*pte) || pte_none(*pte))
+ continue;
+
+ set_pte_enc(pte, level, (void *)vaddr);
+ }
+ vaddr = (unsigned long)__start_bss_decrypted;
+ snp_set_memory_private(vaddr, npages);
+}
+
+/* Stop new private<->shared conversions */
+void snp_kexec_stop_conversion(bool crash)
+{
+ /*
+ * Crash kernel reaches here with interrupts disabled: can't wait for
+ * conversions to finish.
+ *
+ * If race happened, just report and proceed.
+ */
+ bool wait_for_lock = !crash;
+
+ if (!stop_memory_enc_conversion(wait_for_lock))
+ pr_warn("Failed to finish shared<->private conversions\n");
+}
+
+void snp_kexec_unshare_mem(void)
+{
+ if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
+ return;
+
+ unshare_all_memory();
+
+ unshare_all_bss_decrypted_memory();
+
+ if (kexec_last_addr_to_make_private) {
+ unsigned long size;
+ unsigned int level;
+ pte_t *pte;
+
+ /*
+ * Switch to using the MSR protocol to change this cpu's
+ * GHCB to private.
+ * All the per-cpu GHCBs have been switched back to private,
+ * so can't do any more GHCB calls to the hypervisor beyond
+ * this point till the kexec kernel starts running.
+ */
+ boot_ghcb = NULL;
+ sev_cfg.ghcbs_initialized = false;
+
+ pr_debug("boot ghcb 0x%lx\n", kexec_last_addr_to_make_private);
+ pte = lookup_address(kexec_last_addr_to_make_private, &level);
+ size = page_level_size(level);
+ set_pte_enc(pte, level, (void *)kexec_last_addr_to_make_private);
+ snp_set_memory_private(kexec_last_addr_to_make_private, (size / PAGE_SIZE));
+ }
+}
+
static int snp_set_vmsa(void *va, bool vmsa)
{
u64 attrs;
diff --git a/arch/x86/mm/mem_encrypt_amd.c b/arch/x86/mm/mem_encrypt_amd.c
index e7b67519ddb5..49c40c2ed809 100644
--- a/arch/x86/mm/mem_encrypt_amd.c
+++ b/arch/x86/mm/mem_encrypt_amd.c
@@ -468,6 +468,9 @@ void __init sme_early_init(void)
x86_platform.guest.enc_tlb_flush_required = amd_enc_tlb_flush_required;
x86_platform.guest.enc_cache_flush_required = amd_enc_cache_flush_required;
+ x86_platform.guest.enc_kexec_stop_conversion = snp_kexec_stop_conversion;
+ x86_platform.guest.enc_kexec_unshare_mem = snp_kexec_unshare_mem;
+
/*
* AMD-SEV-ES intercepts the RDMSR to read the X2APIC ID in the
* parallel bringup low level code. That raises #VC which cannot be
--
2.34.1
On Tue, Apr 09, 2024 at 02:30:01PM +0300, Kirill A. Shutemov wrote:
> diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
> index e74d0c4286c1..7a1560d7e62d 100644
> --- a/arch/x86/kernel/crash.c
> +++ b/arch/x86/kernel/crash.c
> @@ -128,6 +128,12 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
> #ifdef CONFIG_HPET_TIMER
> hpet_disable();
> #endif
> +
> + if (cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT)) {
> + x86_platform.guest.enc_kexec_stop_conversion(true);
> + x86_platform.guest.enc_kexec_unshare_mem();
> + }
This is not how this is done - the point of those function pointers is
to avoid random checks in the code but simply unconditionally call them.
The platform which needs something special to happen, assigns to them
its own function pointers and the rest assigns dummy stubs.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
ACPI MADT doesn't allow to offline CPU after it got woke up.
Currently CPU hotplug is prevented based on the confidential computing
attribute which is set for Intel TDX. But TDX is not the only possible
user of the wake up method. Any platform that uses ACPI MADT wakeup
method cannot offline CPU.
Disable CPU offlining on ACPI MADT wakeup enumeration.
The change has no visible effects for users: currently, TDX guest is the
only platform that uses the ACPI MADT wakeup method.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
v10.1:
- Update commit message.
---
arch/x86/coco/core.c | 1 -
arch/x86/kernel/acpi/madt_wakeup.c | 3 +++
include/linux/cc_platform.h | 10 ----------
kernel/cpu.c | 3 +--
4 files changed, 4 insertions(+), 13 deletions(-)
diff --git a/arch/x86/coco/core.c b/arch/x86/coco/core.c
index b31ef2424d19..0f81f70aca82 100644
--- a/arch/x86/coco/core.c
+++ b/arch/x86/coco/core.c
@@ -29,7 +29,6 @@ static bool noinstr intel_cc_platform_has(enum cc_attr attr)
{
switch (attr) {
case CC_ATTR_GUEST_UNROLL_STRING_IO:
- case CC_ATTR_HOTPLUG_DISABLED:
case CC_ATTR_GUEST_MEM_ENCRYPT:
case CC_ATTR_MEM_ENCRYPT:
return true;
diff --git a/arch/x86/kernel/acpi/madt_wakeup.c b/arch/x86/kernel/acpi/madt_wakeup.c
index cf79ea6f3007..d222be8d7a07 100644
--- a/arch/x86/kernel/acpi/madt_wakeup.c
+++ b/arch/x86/kernel/acpi/madt_wakeup.c
@@ -1,5 +1,6 @@
// SPDX-License-Identifier: GPL-2.0-or-later
#include <linux/acpi.h>
+#include <linux/cpu.h>
#include <linux/io.h>
#include <asm/apic.h>
#include <asm/barrier.h>
@@ -76,6 +77,8 @@ int __init acpi_parse_mp_wake(union acpi_subtable_headers *header,
acpi_mp_wake_mailbox_paddr = mp_wake->base_address;
+ cpu_hotplug_disable_offlining();
+
apic_update_callback(wakeup_secondary_cpu_64, acpi_wakeup_cpu);
return 0;
diff --git a/include/linux/cc_platform.h b/include/linux/cc_platform.h
index 60693a145894..caa4b4430634 100644
--- a/include/linux/cc_platform.h
+++ b/include/linux/cc_platform.h
@@ -81,16 +81,6 @@ enum cc_attr {
*/
CC_ATTR_GUEST_SEV_SNP,
- /**
- * @CC_ATTR_HOTPLUG_DISABLED: Hotplug is not supported or disabled.
- *
- * The platform/OS is running as a guest/virtual machine does not
- * support CPU hotplug feature.
- *
- * Examples include TDX Guest.
- */
- CC_ATTR_HOTPLUG_DISABLED,
-
/**
* @CC_ATTR_HOST_SEV_SNP: AMD SNP enabled on the host.
*
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 6e1a9157d09c..09765f628b8b 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -1528,8 +1528,7 @@ static int cpu_down_maps_locked(unsigned int cpu, enum cpuhp_state target)
* If the platform does not support hotplug, report it explicitly to
* differentiate it from a transient offlining failure.
*/
- if (cc_platform_has(CC_ATTR_HOTPLUG_DISABLED) ||
- cpu_hotplug_offline_disabled)
+ if (cpu_hotplug_offline_disabled)
return -EOPNOTSUPP;
if (cpu_hotplug_disabled)
return -EBUSY;
--
2.43.0
AMD SEV and Intel TDX guests allocate shared buffers for performing I/O.
This is done by allocating pages normally from the buddy allocator and
then converting them to shared using set_memory_decrypted().
On kexec, the second kernel is unaware of which memory has been
converted in this manner. It only sees E820_TYPE_RAM. Accessing shared
memory as private is fatal.
Therefore, the memory state must be reset to its original state before
starting the new kernel with kexec.
The process of converting shared memory back to private occurs in two
steps:
- enc_kexec_stop_conversion() stops new conversions.
- enc_kexec_unshare_mem() unshares all existing shared memory, reverting
it back to private.
Signed-off-by: Kirill A. Shutemov <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>x
Reviewed-by: Kai Huang <[email protected]>
Tested-by: Tao Liu <[email protected]>
---
arch/x86/include/asm/x86_init.h | 2 ++
arch/x86/kernel/crash.c | 4 ++++
arch/x86/kernel/reboot.c | 12 ++++++++++++
arch/x86/kernel/x86_init.c | 4 ++++
4 files changed, 22 insertions(+)
diff --git a/arch/x86/include/asm/x86_init.h b/arch/x86/include/asm/x86_init.h
index 28ac3cb9b987..c731e6bc4343 100644
--- a/arch/x86/include/asm/x86_init.h
+++ b/arch/x86/include/asm/x86_init.h
@@ -155,6 +155,8 @@ struct x86_guest {
int (*enc_status_change_finish)(unsigned long vaddr, int npages, bool enc);
bool (*enc_tlb_flush_required)(bool enc);
bool (*enc_cache_flush_required)(void);
+ void (*enc_kexec_stop_conversion)(bool crash);
+ void (*enc_kexec_unshare_mem)(void);
};
/**
diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
index e74d0c4286c1..f1b261be78b4 100644
--- a/arch/x86/kernel/crash.c
+++ b/arch/x86/kernel/crash.c
@@ -128,6 +128,10 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
#ifdef CONFIG_HPET_TIMER
hpet_disable();
#endif
+
+ x86_platform.guest.enc_kexec_stop_conversion(true);
+ x86_platform.guest.enc_kexec_unshare_mem();
+
crash_save_cpu(regs, safe_smp_processor_id());
}
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index f3130f762784..c1920ec34f0c 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -12,6 +12,7 @@
#include <linux/delay.h>
#include <linux/objtool.h>
#include <linux/pgtable.h>
+#include <linux/kexec.h>
#include <acpi/reboot.h>
#include <asm/io.h>
#include <asm/apic.h>
@@ -716,6 +717,14 @@ static void native_machine_emergency_restart(void)
void native_machine_shutdown(void)
{
+ /*
+ * Call enc_kexec_stop_conversion() while all CPUs are still active and
+ * interrupts are enabled. This will allow all in-flight memory
+ * conversions to finish cleanly.
+ */
+ if (kexec_in_progress)
+ x86_platform.guest.enc_kexec_stop_conversion(false);
+
/* Stop the cpus and apics */
#ifdef CONFIG_X86_IO_APIC
/*
@@ -752,6 +761,9 @@ void native_machine_shutdown(void)
#ifdef CONFIG_X86_64
x86_platform.iommu_shutdown();
#endif
+
+ if (kexec_in_progress)
+ x86_platform.guest.enc_kexec_unshare_mem();
}
static void __machine_emergency_restart(int emergency)
diff --git a/arch/x86/kernel/x86_init.c b/arch/x86/kernel/x86_init.c
index a7143bb7dd93..045ce1c70070 100644
--- a/arch/x86/kernel/x86_init.c
+++ b/arch/x86/kernel/x86_init.c
@@ -138,6 +138,8 @@ static int enc_status_change_prepare_noop(unsigned long vaddr, int npages, bool
static int enc_status_change_finish_noop(unsigned long vaddr, int npages, bool enc) { return 0; }
static bool enc_tlb_flush_required_noop(bool enc) { return false; }
static bool enc_cache_flush_required_noop(void) { return false; }
+static void enc_kexec_stop_conversion_noop(bool crash) {}
+static void enc_kexec_unshare_mem_noop(void) {}
static bool is_private_mmio_noop(u64 addr) {return false; }
struct x86_platform_ops x86_platform __ro_after_init = {
@@ -161,6 +163,8 @@ struct x86_platform_ops x86_platform __ro_after_init = {
.enc_status_change_finish = enc_status_change_finish_noop,
.enc_tlb_flush_required = enc_tlb_flush_required_noop,
.enc_cache_flush_required = enc_cache_flush_required_noop,
+ .enc_kexec_stop_conversion = enc_kexec_stop_conversion_noop,
+ .enc_kexec_unshare_mem = enc_kexec_unshare_mem_noop,
},
};
--
2.43.0
On Tue, Apr 09, 2024 at 06:26:05PM +0300, Kirill A. Shutemov wrote:
> From 6be428e3b1c6fb494b2c48ba6a7c133514a0b2b4 Mon Sep 17 00:00:00 2001
> From: "Kirill A. Shutemov" <[email protected]>
> Date: Fri, 10 Feb 2023 12:53:11 +0300
> Subject: [PATCHv10.1 05/18] x86/kexec: Keep CR4.MCE set during kexec for TDX guest
>
> Depending on setup, TDX guests might be allowed to clear CR4.MCE.
> Attempt to clear it leads to #VE.
>
> Use alternatives to keep the flag during kexec for TDX guests.
>
> The change doesn't affect non-TDX-guest environments.
This is all fine and dandy but nothing explains *why* TDX needs this
special dance.
Why can't TDX do the usual CR4.MCE diddling like the normal kernel
during init and needs to do that here immediately?
> Signed-off-by: Kirill A. Shutemov <[email protected]>
> ---
> arch/x86/kernel/relocate_kernel_64.S | 15 +++++++++------
> 1 file changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
> index 56cab1bb25f5..90246d544eb1 100644
> --- a/arch/x86/kernel/relocate_kernel_64.S
> +++ b/arch/x86/kernel/relocate_kernel_64.S
> @@ -5,6 +5,8 @@
> */
>
> #include <linux/linkage.h>
> +#include <linux/stringify.h>
> +#include <asm/alternative.h>
> #include <asm/page_types.h>
> #include <asm/kexec.h>
> #include <asm/processor-flags.h>
> @@ -143,14 +145,15 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
>
> /*
> * Set cr4 to a known state:
> - * - physical address extension enabled
> * - 5-level paging, if it was enabled before
> + * - Machine check exception on TDX guest, if it was enabled before.
> + * Clearing MCE might not allowed in TDX guests, depending on setup.
... might not be allowed ...
> + * - physical address extension enabled
> */
> - movl $X86_CR4_PAE, %eax
> - testq $X86_CR4_LA57, %r13
> - jz 1f
> - orl $X86_CR4_LA57, %eax
> -1:
> + movl $X86_CR4_LA57, %eax
> + ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
> + andl %r13d, %eax
%r13 needs a comment here that it contains %cr4 read above in
relocate_kernel()
> + orl $X86_CR4_PAE, %eax
> movq %rax, %cr4
>
> jmp 1f
> --
> Kiryl Shutsemau / Kirill A. Shutemov
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Tue, Apr 09, 2024 at 02:29:58PM +0300, Kirill A. Shutemov wrote:
> TDX is going to have more than one reason to fail
> enc_status_change_prepare().
>
> Change the callback to return errno instead of assuming -EIO;
> enc_status_change_finish() changed too to keep the interface symmetric.
"Change enc_status_change_finish() too... "
"Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
to do frotz", as if you are giving orders to the codebase to change
its behaviour."
You should know this by now...
> Signed-off-by: Kirill A. Shutemov <[email protected]>
> Reviewed-by: Dave Hansen <[email protected]>
> Reviewed-by: Kai Huang <[email protected]>
> Tested-by: Tao Liu <[email protected]>
> ---
> arch/x86/coco/tdx/tdx.c | 20 +++++++++++---------
> arch/x86/hyperv/ivm.c | 22 ++++++++++------------
> arch/x86/include/asm/x86_init.h | 4 ++--
> arch/x86/kernel/x86_init.c | 4 ++--
> arch/x86/mm/mem_encrypt_amd.c | 8 ++++----
> arch/x86/mm/pat/set_memory.c | 8 +++++---
> 6 files changed, 34 insertions(+), 32 deletions(-)
Another thing you should long know by now: get_maintainer.pl. You do
know that when you send a patch which touches multiple different
"places", you run it through get_maintainer.pl to get some hints as to
who to CC, right?
Because you're touching HyperV code and yet none of the HyperV folks are
CCed.
Do I need to give you the spiel I give to kernel newbies? :)
Lemme Cc them for you now.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Sun, Apr 28, 2024 at 07:25:57PM +0200, Borislav Petkov wrote:
> On Tue, Apr 09, 2024 at 02:29:58PM +0300, Kirill A. Shutemov wrote:
> > TDX is going to have more than one reason to fail
> > enc_status_change_prepare().
> >
> > Change the callback to return errno instead of assuming -EIO;
> > enc_status_change_finish() changed too to keep the interface symmetric.
>
> "Change enc_status_change_finish() too... "
>
> "Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
> instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
> to do frotz", as if you are giving orders to the codebase to change
> its behaviour."
Hm. I considered the sentence to be in imperative mood already. I guess I
don't fully understand what imperative mood is. Will fix.
> > Signed-off-by: Kirill A. Shutemov <[email protected]>
> > Reviewed-by: Dave Hansen <[email protected]>
> > Reviewed-by: Kai Huang <[email protected]>
> > Tested-by: Tao Liu <[email protected]>
> > ---
> > arch/x86/coco/tdx/tdx.c | 20 +++++++++++---------
> > arch/x86/hyperv/ivm.c | 22 ++++++++++------------
> > arch/x86/include/asm/x86_init.h | 4 ++--
> > arch/x86/kernel/x86_init.c | 4 ++--
> > arch/x86/mm/mem_encrypt_amd.c | 8 ++++----
> > arch/x86/mm/pat/set_memory.c | 8 +++++---
> > 6 files changed, 34 insertions(+), 32 deletions(-)
>
> Another thing you should long know by now: get_maintainer.pl. You do
> know that when you send a patch which touches multiple different
> "places", you run it through get_maintainer.pl to get some hints as to
> who to CC, right?
You are right, I didn't run get_maintainer.pl this time. I never got it
integrated properly into my workflow. How do you use it? Is it part of
'git send-email' hooks or do you do it manually somehow.
I don't feel I can trust the script to do The Right Thing™ all the time
to put into my hooks. I expect it to blow up on tree-wide patches for
instance.
As result I only run it occasionally, when I remember to which is
suboptimal.
Any tips?
--
Kiryl Shutsemau / Kirill A. Shutemov
On Mon, Apr 29, 2024 at 05:29:23PM +0300, Kirill A. Shutemov wrote:
> Hm. I considered the sentence to be in imperative mood already. I guess I
> don't fully understand what imperative mood is. Will fix.
This:
https://en.wikipedia.org/wiki/Imperative_mood
but basically the sentence is a command.
> You are right, I didn't run get_maintainer.pl this time. I never got it
> integrated properly into my workflow. How do you use it? Is it part of
> 'git send-email' hooks or do you do it manually somehow.
So what I do after the whole set is applied, is:
git diff HEAD~<NUM>.. | ./scripts/get_maintainer.pl
where <NUM> is the number of patches which belong to the series.
IOW, you get a full diff of the set and you run that diff through
get_maintainer.pl.
It'll give you a whole lot of people but you can go through that list
and prune it to the people who are really relevant for the set.
And then you do
git send-email --cc-cmd=cccmd.sh ...
and that script simply echoes a "Cc: <name>" one per line. That is, if
there are a lot of people to Cc. If there are only 1-3ish or so, you can
supply each with the "--cc" option to git-send-email.
Anyway, this is what I do. Someone has probably a lot better flow tho.
> I don't feel I can trust the script to do The Right Thing™ all the time
> to put into my hooks. I expect it to blow up on tree-wide patches for
> instance.
Yeah, not even going there. :-)
HTH.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Mon, Apr 29, 2024 at 04:17:38PM +0300, Kirill A. Shutemov wrote:
> As I mentioned above, clearing CR4.MCE triggers #VE. It is quirk of the
> platform.
You mean when identity_mapped() runs as part of a kexec-ed kernel, it
might clear CR4.MCE and that would trigger the #VE?
So, if that is correct, you basically want to *preserve* the CR4.MCE
setting across kexec?
But then __mcheck_cpu_init_generic() will go and set it
unconditionally.
So what exactly is the correct flow here?
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Mon, Apr 29, 2024 at 04:45:08PM +0200, Borislav Petkov wrote:
> On Mon, Apr 29, 2024 at 04:17:38PM +0300, Kirill A. Shutemov wrote:
> > As I mentioned above, clearing CR4.MCE triggers #VE. It is quirk of the
> > platform.
>
> You mean when identity_mapped() runs as part of a kexec-ed kernel, it
> might clear CR4.MCE and that would trigger the #VE?
Yes, that's what happens in current upstream.
> So, if that is correct, you basically want to *preserve* the CR4.MCE
> setting across kexec?
Yes.
> But then __mcheck_cpu_init_generic() will go and set it
> unconditionally.
__mcheck_cpu_init_generic() will not change anything in this case as the
bit is already set. Everything is hunky-dory.
> So what exactly is the correct flow here?
TDX guest has CR4.MCE set from time 0 and it has to stay this way all the
time including kexec flow.
We have already modified early boot code to preserve CR4.MCE. See
77a512e35db7 ("x86/boot: Avoid #VE during boot for TDX platforms").
The patch extends it to kexec flow.
--
Kiryl Shutsemau / Kirill A. Shutemov
On Sun, Apr 28, 2024 at 07:11:11PM +0200, Borislav Petkov wrote:
> On Tue, Apr 09, 2024 at 06:26:05PM +0300, Kirill A. Shutemov wrote:
> > From 6be428e3b1c6fb494b2c48ba6a7c133514a0b2b4 Mon Sep 17 00:00:00 2001
> > From: "Kirill A. Shutemov" <[email protected]>
> > Date: Fri, 10 Feb 2023 12:53:11 +0300
> > Subject: [PATCHv10.1 05/18] x86/kexec: Keep CR4.MCE set during kexec for TDX guest
> >
> > Depending on setup, TDX guests might be allowed to clear CR4.MCE.
> > Attempt to clear it leads to #VE.
> >
> > Use alternatives to keep the flag during kexec for TDX guests.
> >
> > The change doesn't affect non-TDX-guest environments.
>
> This is all fine and dandy but nothing explains *why* TDX needs this
> special dance.
>
> Why can't TDX do the usual CR4.MCE diddling like the normal kernel
> during init and needs to do that here immediately?
As I mentioned above, clearing CR4.MCE triggers #VE. It is quirk of the
platform.
There's plan to allow it in newer TDX modules, but kernel still has to
assume we cannot touch it in TDX guest case.
> > Signed-off-by: Kirill A. Shutemov <[email protected]>
> > ---
> > arch/x86/kernel/relocate_kernel_64.S | 15 +++++++++------
> > 1 file changed, 9 insertions(+), 6 deletions(-)
> >
> > diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
> > index 56cab1bb25f5..90246d544eb1 100644
> > --- a/arch/x86/kernel/relocate_kernel_64.S
> > +++ b/arch/x86/kernel/relocate_kernel_64.S
> > @@ -5,6 +5,8 @@
> > */
> >
> > #include <linux/linkage.h>
> > +#include <linux/stringify.h>
> > +#include <asm/alternative.h>
> > #include <asm/page_types.h>
> > #include <asm/kexec.h>
> > #include <asm/processor-flags.h>
> > @@ -143,14 +145,15 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
> >
> > /*
> > * Set cr4 to a known state:
> > - * - physical address extension enabled
> > * - 5-level paging, if it was enabled before
> > + * - Machine check exception on TDX guest, if it was enabled before.
> > + * Clearing MCE might not allowed in TDX guests, depending on setup.
>
> ... might not be allowed ...
>
Oopsie. Thanks.
> > + * - physical address extension enabled
> > */
> > - movl $X86_CR4_PAE, %eax
> > - testq $X86_CR4_LA57, %r13
> > - jz 1f
> > - orl $X86_CR4_LA57, %eax
> > -1:
> > + movl $X86_CR4_LA57, %eax
> > + ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
> > + andl %r13d, %eax
>
> %r13 needs a comment here that it contains %cr4 read above in
> relocate_kernel()
Okay.
--
Kiryl Shutsemau / Kirill A. Shutemov
On Mon, Apr 29, 2024 at 06:16:54PM +0300, Kirill A. Shutemov wrote:
> Yes, that's what happens in current upstream.
Let's rewrite that commit message then:
"TDX guests run with MCA enabled (CR4.MCE=1b) from the very start. If
that bit is cleared during CR4 register reprogramming during boot or
kexec flows, a #VE exception will be raised which the guest kernel
cannot handle that early.
Therefore, make sure the CR4.MCE setting is preserved over kexec too and
avoid raising any #VEs."
without that bit about TDX guests might be allowed to clear CR4.MCE
which is only confusing and unnecessary.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Tue, Apr 09, 2024 at 02:29:57PM +0300, Kirill A. Shutemov wrote:
> +1:
> + testq $X86_CR4_MCE, %r13
> + jz 1f
> + ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
> 1:
Please add the below patch to your set. Those same-number labels are
just abominable.
Thx.
---
From: "Borislav Petkov (AMD)" <[email protected]>
Date: Tue, 30 Apr 2024 15:00:16 +0200
Subject: [PATCH] x86/relocate_kernel: Use named labels for less confusion
That identity_mapped() function was loving that "1" label to the point
of completely confusing its readers.
Use named labels in each place for clarity.
No functional changes.
Signed-off-by: Borislav Petkov (AMD) <[email protected]>
---
arch/x86/kernel/relocate_kernel_64.S | 18 ++++++++++--------
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
index 8e2037d78a1f..0077c9e562a7 100644
--- a/arch/x86/kernel/relocate_kernel_64.S
+++ b/arch/x86/kernel/relocate_kernel_64.S
@@ -152,13 +152,15 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
*/
movl $X86_CR4_PAE, %eax
testq $X86_CR4_LA57, %r13
- jz 1f
+ jz no_la57
orl $X86_CR4_LA57, %eax
-1:
+no_la57:
+
testq $X86_CR4_MCE, %r13
- jz 1f
+ jz mca_off
ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
-1:
+mca_off:
+
movq %rax, %cr4
jmp 1f
@@ -173,9 +175,9 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
* used by kexec. Flush the caches before copying the kernel.
*/
testq %r12, %r12
- jz 1f
+ jz sme_off
wbinvd
-1:
+sme_off:
movq %rcx, %r11
call swap_pages
@@ -195,7 +197,7 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
*/
testq %r11, %r11
- jnz 1f
+ jnz relocate
xorl %eax, %eax
xorl %ebx, %ebx
xorl %ecx, %ecx
@@ -216,7 +218,7 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
ret
int3
-1:
+relocate:
popq %rdx
leaq PAGE_SIZE(%r10), %rsp
ANNOTATE_RETPOLINE_SAFE
--
2.43.0
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On Tue, Apr 30, 2024 at 03:03:23PM +0200, Borislav Petkov wrote:
> On Tue, Apr 09, 2024 at 02:29:57PM +0300, Kirill A. Shutemov wrote:
> > +1:
> > + testq $X86_CR4_MCE, %r13
> > + jz 1f
> > + ALTERNATIVE "", __stringify(orl $X86_CR4_MCE, %eax), X86_FEATURE_TDX_GUEST
> > 1:
>
> Please add the below patch to your set. Those same-number labels are
> just abominable.
>
> Thx.
>
> ---
> From: "Borislav Petkov (AMD)" <[email protected]>
> Date: Tue, 30 Apr 2024 15:00:16 +0200
> Subject: [PATCH] x86/relocate_kernel: Use named labels for less confusion
>
> That identity_mapped() function was loving that "1" label to the point
> of completely confusing its readers.
>
> Use named labels in each place for clarity.
>
> No functional changes.
>
> Signed-off-by: Borislav Petkov (AMD) <[email protected]>
> ---
> arch/x86/kernel/relocate_kernel_64.S | 18 ++++++++++--------
> 1 file changed, 10 insertions(+), 8 deletions(-)
>
> diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
> index 8e2037d78a1f..0077c9e562a7 100644
> --- a/arch/x86/kernel/relocate_kernel_64.S
> +++ b/arch/x86/kernel/relocate_kernel_64.S
> @@ -152,13 +152,15 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
> */
> movl $X86_CR4_PAE, %eax
> testq $X86_CR4_LA57, %r13
> - jz 1f
> + jz no_la57
> orl $X86_CR4_LA57, %eax
> -1:
> +no_la57:
I assume all of these new labels have to be prefixed with ".L", right?
--
Kiryl Shutsemau / Kirill A. Shutemov