introduce variable acpi_lapic into ia64

PATCH In review Baoquan He 2 participants 6 messages May 5, 2014 – May 24, 2014 ACPI

Share & tools

Baoquan He

author

2014-05-05 04:48 UTC #0 lore raw

PATCH v3 1/2 introduce variable acpi_lapic into ia64

This variable was defined and assigned in x86, is used to indicate whether LAPIC exists in MADT. Now introduce it into ia64 to help make correct judgment when get information for acpi processor later.

Signed-off-by Baoquan He <[email protected]>

---
 arch/ia64/include/asm/acpi.h | 1 +
 arch/ia64/kernel/acpi.c      | 3 +++
 2 files changed, 4 insertions(+)

arch/ia64/include/asm/acpi.h +1 -0

index d651102..b478219 100644
@@ -85,6 +85,7 @@ ia64_acpi_release_global_lock (unsigned int *lock)
 	((Acq) = ia64_acpi_release_global_lock(&facs->global_lock))
 
 #ifdef	CONFIG_ACPI
+extern int acpi_lapic;
 #define acpi_disabled 0	/* ACPI always enabled on IA64 */
 #define acpi_noirq 0	/* ACPI always enabled on IA64 */
 #define acpi_pci_disabled 0 /* ACPI PCI always enabled on IA64 */

arch/ia64/kernel/acpi.c +3 -0

index 0d407b3..615ef81 100644
@@ -56,6 +56,7 @@
 
 #define PREFIX			"ACPI: "
 
+int acpi_lapic;
 unsigned int acpi_cpei_override;
 unsigned int acpi_cpei_phys_cpuid;

@@ -676,6 +677,8 @@ int __init early_acpi_boot_init(void)
 	if (ret < 1)
 		printk(KERN_ERR PREFIX
 		       "Error parsing MADT - no LAPIC entries\n");
+	else
+		acpi_lapic = 1;
 
 #ifdef CONFIG_SMP
 	if (available_cpus == 0) {

Signature · 1 line

1.8.5.3

Baoquan He

author

2014-05-05 04:48 UTC #1 lore raw

PATCH v3 2/2 lapic need be checked if available when initialize acpi processor id

In acpi_processor_get_info(), acpi processor info is initialized including id, namely cpu index. Currently, if on UP system running SMP kerenl with no LAPIC in MADT, cpu0_initialized is checked if acpi processor id is initialized.

However this check maybe is not sufficient for kdump kernel. Most of time only 1 CPU is supported because of known problems in kdump kernel. So in 1st kernel multiple CPUs are present, then crash happened in one specific CPU, say 2nd CPU. Then it jumped into kdump kernel with "nr_cpus=1" specified in cmdline. In this situation, since kdump kernel is warm reset, it will reuse the ACPI resource passed from crashed kernel directly, namely 1st kernel. It means in MADT all LAPIC is enabled while only 1 CPU is present in running system. The kdump kernel usually is the same as the crashed 1st kernel. So now in kdump kernel, x86_cpu_to_apicid stored the apicid and its related cpu id. If only check cpu0_initialized, it will assign 0 to the acpi processor id of 1st CPU, it's not correct.

So in this patch, check acpi_lapic too. If acpi_lapic is 0, then LAPIC in MADT is not available, assigne 0 to the handling acpi processor id. If acpi_lapic is 1, then LAPIC in MADT is available, let's get apic processor id from x86_cpu_to_apicid.

Signed-off-by Baoquan He <[email protected]>

---
 drivers/acpi/acpi_processor.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/acpi/acpi_processor.c +1 -1

index b06f5f5..bf0cce9 100644
@@ -268,7 +268,7 @@ static int acpi_processor_get_info(struct acpi_device *device)
 	pr->apic_id = apic_id;
 
 	cpu_index = acpi_map_cpuid(pr->apic_id, pr->acpi_id);
-	if (!cpu0_initialized) {
+	if (!cpu0_initialized && !acpi_lapic) {
 		cpu0_initialized = 1;
 		/* Handle UP system running SMP kernel, with no LAPIC in MADT */
 		if ((cpu_index = -1) && (num_online_cpus() = 1))

Signature · 1 line

1.8.5.3

Baoquan He

author

2014-05-05 05:14 UTC #2 lore raw

In reply to Baoquan He

Hi Rafael,

Thanks for previous comments and suggestions.

I added the acpi_lapic in ia64. However I didn't find ia64 machine to test it. Could you or anyone please help test this 2 patches?

I don't know how to test UP system running SMP kernel with no LAPIC in MADT when it's ia64 arch.

Test steps for ia64 kdump:

1) get a multi-cpus ia64 machine, build a upstream kernel with SMP and ACPI
2)install kexec-tools, and edit /etc/sysconfig/kdump to make sure "nr_cpus=1" is in KDUMP_COMMANDLINE_APPEND. Then load the kdump kernel by below command:

"kdumpctl restart" or "systemctl restart kdump"

3) After kdump kernel loaded, execute below shell command. This can make crash happen in 2nd cpu.

taskset -c 1 sh -c "echo c >/proc/sysrq-trigger"

4) From console, below error message should not be printed any more. And the cpu related to 2nd lapid is present, this can be checked by console message and adding debugging code.

"acpi LNXCPU:0a: BIOS reported wrong ACPI id 0 for the processor."

--------
For x86_64, the UP test is taken by adding "disableapic nr_cpus=1" into cmdline of grub. The test for kdump is the same as above ia64.

Thanks
Baoquan

Quoted 46 lines · Baoquan He

Signed-off-by Baoquan He <[email protected]>

---
 arch/ia64/include/asm/acpi.h | 1 +
 arch/ia64/kernel/acpi.c      | 3 +++
 2 files changed, 4 insertions(+)

arch/ia64/include/asm/acpi.h +1 -0

index d651102..b478219 100644
@@ -85,6 +85,7 @@ ia64_acpi_release_global_lock (unsigned int *lock)
 	((Acq) = ia64_acpi_release_global_lock(&facs->global_lock))
 
 #ifdef	CONFIG_ACPI
+extern int acpi_lapic;
 #define acpi_disabled 0	/* ACPI always enabled on IA64 */
 #define acpi_noirq 0	/* ACPI always enabled on IA64 */
 #define acpi_pci_disabled 0 /* ACPI PCI always enabled on IA64 */

arch/ia64/kernel/acpi.c +3 -0

index 0d407b3..615ef81 100644
@@ -56,6 +56,7 @@
 
 #define PREFIX			"ACPI: "
 
+int acpi_lapic;
 unsigned int acpi_cpei_override;
 unsigned int acpi_cpei_phys_cpuid;

@@ -676,6 +677,8 @@ int __init early_acpi_boot_init(void)
 	if (ret < 1)
 		printk(KERN_ERR PREFIX
 		       "Error parsing MADT - no LAPIC entries\n");
+	else
+		acpi_lapic = 1;
 
 #ifdef CONFIG_SMP
 	if (available_cpus = 0) {

Signature · 1 line

1.8.5.3

Baoquan He

author

2014-05-05 06:57 UTC #3 lore raw

Quoted 13 lines · Baoquan He

Hi Rafael,
Thanks for previous comments and suggestions.
I added the acpi_lapic in ia64. However I didn't find ia64 machine to test it. Could you or anyone please help test this 2 patches?
I don't know how to test UP system running SMP kernel with no LAPIC in MADT when it's ia64 arch.
Test steps for ia64 kdump:

Quoted 7 lines

1) get a multi-cpus ia64 machine, build a upstream kernel with SMP and ACPI
2)install kexec-tools, and edit /etc/sysconfig/kdump to make sure "nr_cpus=1" is in KDUMP_COMMANDLINE_APPEND. Then load the kdump kernel by below command:
"kdumpctl restart" or "systemctl restart kdump"

Sorry, here I assume people use fedora distros. The kdump execution command is like below string which is taken from my fedora automation script.

/sbin/kexec -p '--command-line=BOOT_IMAGE=/vmlinuz-3.13.5-202.fc20.x86_64 root=UUID=f170152e-de83-46ee-9546-8ccd53f9753b ro rd.md=0 rd.lvm=0 rd.dm=0 vconsole.keymap=us rd.luks=0 vconsole.font=latarcyrheb-sun16 console=tty0 rd.shell LANG=en_US.UTF-8 console=ttyS0,115200n8 intel_iommu=on irqpoll nr_cpus=1 reset_devices cgroup_disable=memory mce=off numa=off udev.children-max=2 panic=10 rootflags=nofail disable_cpu_apicid=0' --initrd=/boot/initramfs-3.13.5-202.fc20.x86_64kdump.img /boot/vmlinuz-3.13.5-202.fc20.x86_64

Thanks
Baoquan

Quoted 68 lines

3) After kdump kernel loaded, execute below shell command. This can make crash happen in 2nd cpu.
taskset -c 1 sh -c "echo c >/proc/sysrq-trigger"
4) From console, below error message should not be printed any more. And the cpu related to 2nd lapid is present, this can be checked by console message and adding debugging code.
"acpi LNXCPU:0a: BIOS reported wrong ACPI id 0 for the processor."
--------
For x86_64, the UP test is taken by adding "disableapic nr_cpus=1" into cmdline of grub. The test for kdump is the same as above ia64.
Thanks
Baoquan
Quoted 46 lines · Baoquan He
This variable was defined and assigned in x86, is used to indicate whether LAPIC exists in MADT. Now introduce it into ia64 to help make correct judgment when get information for acpi processor later.
Signed-off-by Baoquan He <[email protected]>
---
 arch/ia64/include/asm/acpi.h | 1 +
 arch/ia64/kernel/acpi.c      | 3 +++
 2 files changed, 4 insertions(+)
arch/ia64/include/asm/acpi.h +1 -0
index d651102..b478219 100644
@@ -85,6 +85,7 @@ ia64_acpi_release_global_lock (unsigned int *lock)
 	((Acq) = ia64_acpi_release_global_lock(&facs->global_lock))
 
 #ifdef	CONFIG_ACPI
+extern int acpi_lapic;
 #define acpi_disabled 0	/* ACPI always enabled on IA64 */
 #define acpi_noirq 0	/* ACPI always enabled on IA64 */
 #define acpi_pci_disabled 0 /* ACPI PCI always enabled on IA64 */
arch/ia64/kernel/acpi.c +3 -0
index 0d407b3..615ef81 100644
@@ -56,6 +56,7 @@
 
 #define PREFIX			"ACPI: "
 
+int acpi_lapic;
 unsigned int acpi_cpei_override;
 unsigned int acpi_cpei_phys_cpuid;
 
@@ -676,6 +677,8 @@ int __init early_acpi_boot_init(void)
 	if (ret < 1)
 		printk(KERN_ERR PREFIX
 		       "Error parsing MADT - no LAPIC entries\n");
+	else
+		acpi_lapic = 1;
 
 #ifdef CONFIG_SMP
 	if (available_cpus == 0) {
Signature · 1 line
1.8.5.3

Rafael J. Wysocki

2014-05-19 22:40 UTC #4 lore raw

In reply to Baoquan He

Quoted 5 lines · Baoquan He

This variable was defined and assigned in x86, is used to indicate whether LAPIC exists in MADT. Now introduce it into ia64 to help make correct judgment when get information for acpi processor later.
Signed-off-by Baoquan He <[email protected]>

This and [2/2] queued up for 3.16, thanks!

Quoted 39 lines

---
 arch/ia64/include/asm/acpi.h | 1 +
 arch/ia64/kernel/acpi.c      | 3 +++
 2 files changed, 4 insertions(+)

arch/ia64/include/asm/acpi.h +1 -0

index d651102..b478219 100644
@@ -85,6 +85,7 @@ ia64_acpi_release_global_lock (unsigned int *lock)
 	((Acq) = ia64_acpi_release_global_lock(&facs->global_lock))
 
 #ifdef	CONFIG_ACPI
+extern int acpi_lapic;
 #define acpi_disabled 0	/* ACPI always enabled on IA64 */
 #define acpi_noirq 0	/* ACPI always enabled on IA64 */
 #define acpi_pci_disabled 0 /* ACPI PCI always enabled on IA64 */

arch/ia64/kernel/acpi.c +3 -0

index 0d407b3..615ef81 100644
@@ -56,6 +56,7 @@
 
 #define PREFIX			"ACPI: "
 
+int acpi_lapic;
 unsigned int acpi_cpei_override;
 unsigned int acpi_cpei_phys_cpuid;

@@ -676,6 +677,8 @@ int __init early_acpi_boot_init(void)
 	if (ret < 1)
 		printk(KERN_ERR PREFIX
 		       "Error parsing MADT - no LAPIC entries\n");
+	else
+		acpi_lapic = 1;
 
 #ifdef CONFIG_SMP
 	if (available_cpus == 0) {

Signature · 2 lines

I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Baoquan He

author

2014-05-24 09:42 UTC #5 lore raw

Quoted 8 lines · Rafael J. Wysocki

Quoted 5 lines · Baoquan He
This variable was defined and assigned in x86, is used to indicate whether LAPIC exists in MADT. Now introduce it into ia64 to help make correct judgment when get information for acpi processor later.
Signed-off-by Baoquan He <[email protected]>
This and [2/2] queued up for 3.16, thanks!

Thanks, Rafael!

Quoted 44 lines

Quoted 40 lines

---
 arch/ia64/include/asm/acpi.h | 1 +
 arch/ia64/kernel/acpi.c      | 3 +++
 2 files changed, 4 insertions(+)

arch/ia64/include/asm/acpi.h +1 -0

index d651102..b478219 100644
@@ -85,6 +85,7 @@ ia64_acpi_release_global_lock (unsigned int *lock)
 	((Acq) = ia64_acpi_release_global_lock(&facs->global_lock))
 
 #ifdef	CONFIG_ACPI
+extern int acpi_lapic;
 #define acpi_disabled 0	/* ACPI always enabled on IA64 */
 #define acpi_noirq 0	/* ACPI always enabled on IA64 */
 #define acpi_pci_disabled 0 /* ACPI PCI always enabled on IA64 */

arch/ia64/kernel/acpi.c +3 -0

index 0d407b3..615ef81 100644
@@ -56,6 +56,7 @@
 
 #define PREFIX			"ACPI: "
 
+int acpi_lapic;
 unsigned int acpi_cpei_override;
 unsigned int acpi_cpei_phys_cpuid;

@@ -676,6 +677,8 @@ int __init early_acpi_boot_init(void)
 	if (ret < 1)
 		printk(KERN_ERR PREFIX
 		       "Error parsing MADT - no LAPIC entries\n");
+	else
+		acpi_lapic = 1;
 
 #ifdef CONFIG_SMP
 	if (available_cpus = 0) {

Signature · 2 lines

I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.