Skip to contentSkip to conversation

x86/microcode: Remove microcode_mutex.

PATCH Superseded Sebastian Andrzej Siewior 3 participants 6 messages Aug 3, 2023 – Aug 8, 2023

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/microcode: Remove microcode_mutex.

microcode_mutex is only used by reload_store(). It has a comment saying "to synchronize with each other". This probably means the sysfs interface vs the legacy interface which was removed in commit 181b6f40e9ea8 ("x86/microcode: Rip out the OLD_INTERFACE").

The sysfs interface does not need additional synchronisation vs itself because it is provided as kernfs_ops::mutex which is acquired in kernfs_fop_write_iter().

Remove superfluous microcode_mutex.

---
This poped up as "defined but not used" on RT builds without CONFIG_MICROCODE_LATE_LOADING enabled.

 arch/x86/kernel/cpu/microcode/core.c | 6 ------
 1 file changed, 6 deletions(-)
arch/x86/kernel/cpu/microcode/core.c +0 -6
index 3afcf3de0dd49..2f9d35744bc41 100644
@@ -54,15 +54,12 @@ LIST_HEAD(microcode_cache);
  *
  * All non cpu-hotplug-callback call sites use:
  *
- * - microcode_mutex to synchronize with each other;
  * - cpus_read_lock/unlock() to synchronize with
  *   the cpu-hotplug-callback call sites.
  *
  * We guarantee that only a single cpu is being
  * updated at any particular moment of time.
  */
-static DEFINE_MUTEX(microcode_mutex);
-
 struct ucode_cpu_info		ucode_cpu_info[NR_CPUS];
 
 struct cpu_info_ctx {
@@ -488,10 +485,7 @@ static ssize_t reload_store(struct device *dev,
 	if (tmp_ret != UCODE_NEW)
 		goto put;
 
-	mutex_lock(&microcode_mutex);
 	ret = microcode_reload_late();
-	mutex_unlock(&microcode_mutex);
-
 put:
 	cpus_read_unlock();
 
Signature · 1 line
2.40.1
#1 lore raw

Nit: The full stop at the end is not needed.

Quoted 5 lines · Sebastian Andrzej Siewior

microcode_mutex is only used by reload_store(). It has a comment saying "to synchronize with each other". This probably means the sysfs interface vs the legacy interface which was removed in commit 181b6f40e9ea8 ("x86/microcode: Rip out the OLD_INTERFACE").

There is also commit b6f86689d5b7 ("x86/microcode: Rip out the subsys interface gunk") which last year removed another usage of microcode_mutex.

Quoted 5 lines

The sysfs interface does not need additional synchronisation vs itself because it is provided as kernfs_ops::mutex which is acquired in kernfs_fop_write_iter().

Remove superfluous microcode_mutex.

I agree, the current usage does look unnecessary.

Maybe reword the commit message to say that after these two Rip outs there are no of *other* usages of microcode_mutex to synchronize with?

Quoted 5 lines

---
This poped up as "defined but not used" on RT builds without CONFIG_MICROCODE_LATE_LOADING enabled.

This issue has been raised a couple of times recently but the justification has been deemed insufficient since it can't be reproduced with a .config file.

See:
https://lore.kernel.org/lkml/[email protected]/

and

https://lore.kernel.org/lkml/[email protected]/

However, your current justification of not needing the mutex itself seems reasonable to me.

Quoted 25 lines
 arch/x86/kernel/cpu/microcode/core.c | 6 ------
 1 file changed, 6 deletions(-)
arch/x86/kernel/cpu/microcode/core.c +0 -3
index 3afcf3de0dd49..2f9d35744bc41 100644
@@ -54,15 +54,12 @@ LIST_HEAD(microcode_cache);
  *
  * All non cpu-hotplug-callback call sites use:
  *
- * - microcode_mutex to synchronize with each other;
  * - cpus_read_lock/unlock() to synchronize with
  *   the cpu-hotplug-callback call sites.
  *
  * We guarantee that only a single cpu is being
  * updated at any particular moment of time.
  */
-static DEFINE_MUTEX(microcode_mutex);
-
 struct ucode_cpu_info		ucode_cpu_info[NR_CPUS];
 
 struct cpu_info_ctx {

arch/x86/kernel/cpu/microcode/core.c static ssize_t reload_store(struct device *dev, · line 491

Quoted 10 lines
arch/x86/kernel/cpu/microcode/core.c +0 -3
@@ -488,10 +485,7 @@ static ssize_t reload_store(struct device *dev,
 	if (tmp_ret != UCODE_NEW)
 		goto put;
 
-	mutex_lock(&microcode_mutex);
 	ret = microcode_reload_late();
-	mutex_unlock(&microcode_mutex);
-
 put:
 	cpus_read_unlock();
 

The code changes look fine to me.

You can also add below to the patch.

Quoted 14 lines
arch/x86/kernel/cpu/common.c +1 -2
index 52683fddafaf..777340724ec3 100644
@@ -2276,8 +2276,7 @@ void store_cpu_caps(struct cpuinfo_x86 *curr_info)
  * @prev_info: CPU capabilities stored before an update.
  *
  * The microcode loader calls this upon late microcode load to recheck features,
- * only when microcode has been updated. Caller holds microcode_mutex and CPU
- * hotplug lock.
+ * only when microcode has been updated. Caller holds CPU hotplug lock.
  *
  * Return: None
  */
#2 lore raw
Quoted 11 lines · Sohil Mehta

Nit: The full stop at the end is not needed.

Quoted 5 lines · Sebastian Andrzej Siewior

microcode_mutex is only used by reload_store(). It has a comment saying "to synchronize with each other". This probably means the sysfs interface vs the legacy interface which was removed in commit 181b6f40e9ea8 ("x86/microcode: Rip out the OLD_INTERFACE").

There is also commit b6f86689d5b7 ("x86/microcode: Rip out the subsys interface gunk") which last year removed another usage of microcode_mutex.

Okay.

Quoted 10 lines
Quoted 5 lines

The sysfs interface does not need additional synchronisation vs itself because it is provided as kernfs_ops::mutex which is acquired in kernfs_fop_write_iter().

Remove superfluous microcode_mutex.

I agree, the current usage does look unnecessary.

Maybe reword the commit message to say that after these two Rip outs there are no of *other* usages of microcode_mutex to synchronize with?

Okay.

Quoted 8 lines
Quoted 4 lines

---
This poped up as "defined but not used" on RT builds without CONFIG_MICROCODE_LATE_LOADING enabled.

This issue has been raised a couple of times recently but the justification has been deemed insufficient since it can't be reproduced with a .config file.

The PREEMPT_RT's implementation of struct mutex is different which makes it easier for the compiler to spot an used mutex. The !RT's implementation has list_head pointing to the mutex as part of MUTEX_INIT which marks the variable as used.

Quoted 9 lines

See:
https://lore.kernel.org/lkml/[email protected]/

and

https://lore.kernel.org/lkml/[email protected]/

However, your current justification of not needing the mutex itself seems reasonable to me.

So everyone tried to move it but me…

…

Quoted 3 lines

The code changes look fine to me.

You can also add below to the patch.

Will do.

Sebastian

#3 lore raw

PATCH v2 1/1 x86/microcode: Remove microcode_mutex.

microcode_mutex is only used by reload_store(). It has a comment saying "to synchronize with each other". Other user of this mutex have been removed in the commits

	181b6f40e9ea8 ("x86/microcode: Rip out the OLD_INTERFACE").
	b6f86689d5b74 ("x86/microcode: Rip out the subsys interface gunk")

The sysfs interface does not need additional synchronisation vs itself because it is provided as kernfs_ops::mutex which is acquired in kernfs_fop_write_iter().

Remove superfluous microcode_mutex.

---

v1…v2:
- Add another commit to the commit description which removed another user. Suggested by Sohil Mehta.
- Also update the comment in microcode_check(). Suggested by Sohil Mehta.

 arch/x86/kernel/cpu/common.c         | 3 +--
 arch/x86/kernel/cpu/microcode/core.c | 6 ------
 2 files changed, 1 insertion(+), 8 deletions(-)
arch/x86/kernel/cpu/common.c +1 -2
index 0ba1067f4e5f1..de49b9cd8077a 100644
@@ -2276,8 +2276,7 @@ void store_cpu_caps(struct cpuinfo_x86 *curr_info)
  * @prev_info:	CPU capabilities stored before an update.
  *
  * The microcode loader calls this upon late microcode load to recheck features,
- * only when microcode has been updated. Caller holds microcode_mutex and CPU
- * hotplug lock.
+ * only when microcode has been updated. Caller holds and CPU hotplug lock.
  *
  * Return: None
  */
arch/x86/kernel/cpu/microcode/core.c +0 -6
index 3afcf3de0dd49..2f9d35744bc41 100644
@@ -54,15 +54,12 @@ LIST_HEAD(microcode_cache);
  *
  * All non cpu-hotplug-callback call sites use:
  *
- * - microcode_mutex to synchronize with each other;
  * - cpus_read_lock/unlock() to synchronize with
  *   the cpu-hotplug-callback call sites.
  *
  * We guarantee that only a single cpu is being
  * updated at any particular moment of time.
  */
-static DEFINE_MUTEX(microcode_mutex);
-
 struct ucode_cpu_info		ucode_cpu_info[NR_CPUS];
 
 struct cpu_info_ctx {
@@ -488,10 +485,7 @@ static ssize_t reload_store(struct device *dev,
 	if (tmp_ret != UCODE_NEW)
 		goto put;
 
-	mutex_lock(&microcode_mutex);
 	ret = microcode_reload_late();
-	mutex_unlock(&microcode_mutex);
-
 put:
 	cpus_read_unlock();
 
Signature · 1 line
2.40.1
Reviewed-by on v2 1/1
#4 lore raw

arch/x86/kernel/cpu/microcode/core.c static ssize_t reload_store(struct device *dev, · line 494

Quoted 9 lines

struct cpu_info_ctx {

arch/x86/kernel/cpu/microcode/core.c +0 -3
@@ -488,10 +485,7 @@ static ssize_t reload_store(struct device *dev,
 	if (tmp_ret != UCODE_NEW)
 		goto put;
 
-	mutex_lock(&microcode_mutex);
 	ret = microcode_reload_late();
-	mutex_unlock(&microcode_mutex);
-

Maybe leave the new line in there before the put:? Makes it slightly easier to read the code.

arch/x86/kernel/cpu/microcode/core.c static ssize_t reload_store(struct device *dev, · line 491

Quoted 3 lines
arch/x86/kernel/cpu/microcode/core.c +0 -0
 put:
 	cpus_read_unlock();
 

Anyway,

#5 lore raw

[tip: x86/microcode] x86/microcode: Remove microcode_mutex

In reply to Sebastian Andrzej Siewior
The following commit has been merged into the x86/microcode branch of tip:

The following commit has been merged into the x86/microcode branch of tip:

Commit-ID:     80347cd515ca149f1da31786ec9a59b0dfd1e579
Gitweb:        https://git.kernel.org/tip/80347cd515ca149f1da31786ec9a59b0dfd1e579
Author:        Sebastian Andrzej Siewior <[email protected]>
AuthorDate:    Fri, 04 Aug 2023 09:58:53 +02:00
Committer:     Borislav Petkov (AMD) <[email protected]>
CommitterDate: Tue, 08 Aug 2023 19:06:29 +02:00

x86/microcode: Remove microcode_mutex

microcode_mutex is only used by reload_store(). It has a comment saying "to synchronize with each other". Other user of this mutex have been removed in the commits

181b6f40e9ea8 ("x86/microcode: Rip out the OLD_INTERFACE"). b6f86689d5b74 ("x86/microcode: Rip out the subsys interface gunk")

The sysfs interface does not need additional synchronisation vs itself because it is provided as kernfs_ops::mutex which is acquired in kernfs_fop_write_iter().

Remove the superfluous microcode_mutex.

---
 arch/x86/kernel/cpu/common.c         | 3 +--
 arch/x86/kernel/cpu/microcode/core.c | 6 ------
 2 files changed, 1 insertion(+), 8 deletions(-)
arch/x86/kernel/cpu/common.c +1 -2
index 52683fd..06015f3 100644
@@ -2276,8 +2276,7 @@ void store_cpu_caps(struct cpuinfo_x86 *curr_info)
  * @prev_info:	CPU capabilities stored before an update.
  *
  * The microcode loader calls this upon late microcode load to recheck features,
- * only when microcode has been updated. Caller holds microcode_mutex and CPU
- * hotplug lock.
+ * only when microcode has been updated. Caller holds and CPU hotplug lock.
  *
  * Return: None
  */
arch/x86/kernel/cpu/microcode/core.c +0 -6
index 192adf5..c9a53e3 100644
@@ -54,15 +54,12 @@ LIST_HEAD(microcode_cache);
  *
  * All non cpu-hotplug-callback call sites use:
  *
- * - microcode_mutex to synchronize with each other;
  * - cpus_read_lock/unlock() to synchronize with
  *   the cpu-hotplug-callback call sites.
  *
  * We guarantee that only a single cpu is being
  * updated at any particular moment of time.
  */
-static DEFINE_MUTEX(microcode_mutex);
-
 struct ucode_cpu_info		ucode_cpu_info[NR_CPUS];
 
 struct cpu_info_ctx {
@@ -488,10 +485,7 @@ static ssize_t reload_store(struct device *dev,
 	if (tmp_ret != UCODE_NEW)
 		goto put;
 
-	mutex_lock(&microcode_mutex);
 	ret = microcode_reload_late();
-	mutex_unlock(&microcode_mutex);
-
 put:
 	cpus_read_unlock();