Skip to contentSkip to conversation

intel_txt: Intel(R) TXT Sx shutdown support

RFC PATCH Superseded Joseph Cihula 2 participants 2 messages Jun 23, 2009 – Jun 27, 2009

Subsystems (4)

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

RFC PATCH v5 3/4 intel_txt: Intel(R) TXT Sx shutdown support

Support for graceful handling of sleep states (S3/S4/S5) after an Intel(R) TXT launch.

Without this patch, attempting to place the system in one of the ACPI sleep states (S3/S4/S5) will cause the TXT hardware to treat this as an attack and will cause a system reset, with memory locked. Not only may the subsequent memory scrub take some time, but the platform will be unable to enter the requested power state.

This patch calls back into the tboot so that it may properly and securely clean up system state and clear the secrets-in-memory flag, after which it will place the system into the requested sleep state using ACPI information passed by the kernel.

 arch/x86/kernel/smpboot.c     |    6 +++++-
 drivers/acpi/acpica/hwsleep.c |   13 +++++++++++++
 kernel/cpu.c                  |    6 +++++-
 3 files changed, 23 insertions(+), 2 deletions(-)

---

arch/x86/kernel/smpboot.c +5 -1
diff -uprN -X linus-2.6.git/Documentation/dontdiff linus-2.6.git/arch/x86/kernel/smpboot.c linus-2.6.git-txt/arch/x86/kernel/smpboot.c
@@ -62,6 +62,7 @@
 #include <asm/vmi.h>
 #include <asm/apic.h>
 #include <asm/setup.h>
+#include <asm/tboot.h>
 #include <asm/uv/uv.h>
 #include <linux/mc146818rtc.h>
 
@@ -1317,7 +1318,10 @@ void play_dead_common(void)
 void native_play_dead(void)
 {
 	play_dead_common();
-	wbinvd_halt();
+	if (tboot_in_measured_env())
+		tboot_shutdown(TB_SHUTDOWN_WFS);
+	else
+		wbinvd_halt();
 }
 
 #else /* ... !CONFIG_HOTPLUG_CPU */
drivers/acpi/acpica/hwsleep.c +13 -0
diff -uprN -X linus-2.6.git/Documentation/dontdiff linus-2.6.git/drivers/acpi/acpica/hwsleep.c linus-2.6.git-txt/drivers/acpi/acpica/hwsleep.c
@@ -45,6 +45,7 @@
 #include <acpi/acpi.h>
 #include "accommon.h"
 #include "actables.h"
+#include <asm/tboot.h>
 
 #define _COMPONENT          ACPI_HARDWARE
 ACPI_MODULE_NAME("hwsleep")
@@ -342,6 +343,18 @@ acpi_status asmlinkage acpi_enter_sleep_
 
 	ACPI_FLUSH_CPU_CACHE();
 
+	if (tboot_in_measured_env()) {
+		tboot_copy_fadt(&acpi_gbl_FADT);
+		tboot_shared->acpi_sinfo.pm1a_cnt_val = pm1a_control;
+		tboot_shared->acpi_sinfo.pm1b_cnt_val = pm1b_control;
+		/* we always use the 32b wakeup vector */
+		tboot_shared->acpi_sinfo.vector_width = 32;
+		tboot_shared->acpi_sinfo.kernel_s3_resume_vector =
+							acpi_wakeup_address;
+
+		tboot_sleep(sleep_state);
+	}
+
 	/* Write #2: Write both SLP_TYP + SLP_EN */
 
 	status = acpi_hw_write_pm1_control(pm1a_control, pm1b_control);
kernel/cpu.c +5 -1
diff -uprN -X linus-2.6.git/Documentation/dontdiff linus-2.6.git/kernel/cpu.c linus-2.6.git-txt/kernel/cpu.c
@@ -14,6 +14,7 @@
 #include <linux/kthread.h>
 #include <linux/stop_machine.h>
 #include <linux/mutex.h>
+#include <asm/tboot.h>
 
 #ifdef CONFIG_SMP
 /* Serializes the updates to cpu_online_mask, cpu_present_mask */
@@ -379,7 +380,7 @@ static cpumask_var_t frozen_cpus;
 
 int disable_nonboot_cpus(void)
 {
-	int cpu, first_cpu, error;
+	int cpu, first_cpu, error, num_cpus = 0;
 
 	error = stop_machine_create();
 	if (error)
@@ -394,6 +395,7 @@ int disable_nonboot_cpus(void)
 	for_each_online_cpu(cpu) {
 		if (cpu == first_cpu)
 			continue;
+		num_cpus++;
 		error = _cpu_down(cpu, 1);
 		if (!error) {
 			cpumask_set_cpu(cpu, frozen_cpus);
@@ -404,6 +406,8 @@ int disable_nonboot_cpus(void)
 			break;
 		}
 	}
+	/* ensure all CPUs have gone into wait-for-SIPI */
+	tboot_wait_for_aps(num_cpus);
 	if (!error) {
 		BUG_ON(num_online_cpus() > 1);
 		/* Make sure the CPUs won't be enabled by someone else */
#1 lore raw
Quoted 11 lines · Joseph Cihula

Support for graceful handling of sleep states (S3/S4/S5) after an Intel(R) TXT launch.

Without this patch, attempting to place the system in one of the ACPI sleep states (S3/S4/S5) will cause the TXT hardware to treat this as an attack and will cause a system reset, with memory locked. Not only may the subsequent memory scrub take some time, but the platform will be unable to enter the requested power state.

This patch calls back into the tboot so that it may properly and securely clean up system state and clear the secrets-in-memory flag, after which it will place the system into the requested sleep state using ACPI information passed by the kernel.

I don't get this. In case of hibernation... how do you 'protect' the data in memory?

This really needs big Documentation/ patch explaining the design before it can be reviewed properly.

What does it protect? All kernel memory? How does it handle S3? Who does encryption for hibernation?

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html