A recent change moved the microcode loader hotplug callback into the early
startup phase, which is running with interrupts disabled. It missed that
the callbacks invoke sysfs functions which might sleep causing nice 'might
sleep' splats with proper debugging enabled.
Split the callbacks and only load the microcode in the early startup phase
and move the sysfs handling back into the later threaded and preemptible
bringup phase where it was before.
Fixes: 78f4e932f776 ("x86/microcode, cpuhotplug: Add a microcode loader CPU hotplug callback")
Signed-off-by: Thomas Gleixner <[email protected]>
Cc: [email protected]
---
arch/x86/kernel/cpu/microcode/core.c | 15 ++++++++++-----
1 file changed, 10 insertions(+), 5 deletions(-)
--- a/arch/x86/kernel/cpu/microcode/core.c
+++ b/arch/x86/kernel/cpu/microcode/core.c
@@ -789,13 +789,16 @@ static struct syscore_ops mc_syscore_ops
.resume = mc_bp_resume,
};
-static int mc_cpu_online(unsigned int cpu)
+static int mc_cpu_starting(unsigned int cpu)
{
- struct device *dev;
-
- dev = get_cpu_device(cpu);
microcode_update_cpu(cpu);
pr_debug("CPU%d added\n", cpu);
+ return 0;
+}
+
+static int mc_cpu_online(unsigned int cpu)
+{
+ struct device *dev = get_cpu_device(cpu);
if (sysfs_create_group(&dev->kobj, &mc_attr_group))
pr_err("Failed to create group for CPU%d\n", cpu);
@@ -872,7 +875,9 @@ int __init microcode_init(void)
goto out_ucode_group;
register_syscore_ops(&mc_syscore_ops);
- cpuhp_setup_state_nocalls(CPUHP_AP_MICROCODE_LOADER, "x86/microcode:online",
+ cpuhp_setup_state_nocalls(CPUHP_AP_MICROCODE_LOADER, "x86/microcode:starting",
+ mc_cpu_starting, NULL);
+ cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "x86/microcode:online",
mc_cpu_online, mc_cpu_down_prep);
pr_info("Microcode Update Driver: v%s.", DRIVER_VERSION);
On Tue, Jun 18, 2019 at 10:31:40PM +0200, Thomas Gleixner wrote:
> A recent change moved the microcode loader hotplug callback into the early
> startup phase, which is running with interrupts disabled. It missed that
> the callbacks invoke sysfs functions which might sleep causing nice 'might
> sleep' splats with proper debugging enabled.
>
> Split the callbacks and only load the microcode in the early startup phase
> and move the sysfs handling back into the later threaded and preemptible
> bringup phase where it was before.
>
> Fixes: 78f4e932f776 ("x86/microcode, cpuhotplug: Add a microcode loader CPU hotplug callback")
> Signed-off-by: Thomas Gleixner <[email protected]>
> Cc: [email protected]
> ---
> arch/x86/kernel/cpu/microcode/core.c | 15 ++++++++++-----
> 1 file changed, 10 insertions(+), 5 deletions(-)
Acked-by: Borislav Petkov <[email protected]>
Thanks for fixing that!
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
Commit-ID: 5423f5ce5ca410b3646f355279e4e937d452e622
Gitweb: https://git.kernel.org/tip/5423f5ce5ca410b3646f355279e4e937d452e622
Author: Thomas Gleixner <[email protected]>
AuthorDate: Tue, 18 Jun 2019 22:31:40 +0200
Committer: Borislav Petkov <[email protected]>
CommitDate: Wed, 19 Jun 2019 09:16:35 +0200
x86/microcode: Fix the microcode load on CPU hotplug for real
A recent change moved the microcode loader hotplug callback into the early
startup phase which is running with interrupts disabled. It missed that
the callbacks invoke sysfs functions which might sleep causing nice 'might
sleep' splats with proper debugging enabled.
Split the callbacks and only load the microcode in the early startup phase
and move the sysfs handling back into the later threaded and preemptible
bringup phase where it was before.
Fixes: 78f4e932f776 ("x86/microcode, cpuhotplug: Add a microcode loader CPU hotplug callback")
Signed-off-by: Thomas Gleixner <[email protected]>
Signed-off-by: Borislav Petkov <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: [email protected]
Cc: x86-ml <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
---
arch/x86/kernel/cpu/microcode/core.c | 15 ++++++++++-----
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/arch/x86/kernel/cpu/microcode/core.c b/arch/x86/kernel/cpu/microcode/core.c
index a813987b5552..cb0fdcaf1415 100644
--- a/arch/x86/kernel/cpu/microcode/core.c
+++ b/arch/x86/kernel/cpu/microcode/core.c
@@ -789,13 +789,16 @@ static struct syscore_ops mc_syscore_ops = {
.resume = mc_bp_resume,
};
-static int mc_cpu_online(unsigned int cpu)
+static int mc_cpu_starting(unsigned int cpu)
{
- struct device *dev;
-
- dev = get_cpu_device(cpu);
microcode_update_cpu(cpu);
pr_debug("CPU%d added\n", cpu);
+ return 0;
+}
+
+static int mc_cpu_online(unsigned int cpu)
+{
+ struct device *dev = get_cpu_device(cpu);
if (sysfs_create_group(&dev->kobj, &mc_attr_group))
pr_err("Failed to create group for CPU%d\n", cpu);
@@ -872,7 +875,9 @@ int __init microcode_init(void)
goto out_ucode_group;
register_syscore_ops(&mc_syscore_ops);
- cpuhp_setup_state_nocalls(CPUHP_AP_MICROCODE_LOADER, "x86/microcode:online",
+ cpuhp_setup_state_nocalls(CPUHP_AP_MICROCODE_LOADER, "x86/microcode:starting",
+ mc_cpu_starting, NULL);
+ cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "x86/microcode:online",
mc_cpu_online, mc_cpu_down_prep);
pr_info("Microcode Update Driver: v%s.", DRIVER_VERSION);
On Tue, Jun 18, 2019 at 1:31 PM Thomas Gleixner <[email protected]> wrote:
>
> A recent change moved the microcode loader hotplug callback into the early
> startup phase, which is running with interrupts disabled. It missed that
> the callbacks invoke sysfs functions which might sleep causing nice 'might
> sleep' splats with proper debugging enabled.
>
> Split the callbacks and only load the microcode in the early startup phase
> and move the sysfs handling back into the later threaded and preemptible
> bringup phase where it was before.
>
> Fixes: 78f4e932f776 ("x86/microcode, cpuhotplug: Add a microcode loader CPU hotplug callback")
> Signed-off-by: Thomas Gleixner <[email protected]>
I ran into the bug fixed by this. For 4.19:
Tested-by: Evan Green <[email protected]>