mm/mremap_pages: Fix static key devmap_managed_key updates

PATCH In review Aneesh Kumar K.V 5 participants 10 messages Oct 22, 2020 – Oct 23, 2020

Subsystems (3)

Memory management libnvdimm: non-volatile memory device Memory hot(un)plug

lore

Share & tools

lore

Aneesh Kumar K.V

author

2020-10-22 06:08 UTC #0 lore raw

PATCH 1/1 mm/mremap_pages: Fix static key devmap_managed_key updates

commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().

devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.

jump label: negative count!

 WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
 Modules linked in:
 ....

NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0 LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0

 Call Trace:
 [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
 [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
 [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
 [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
 [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
 [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
 [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
 [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
 [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
 [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
 [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
 [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
 [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
 [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
 [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
 [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
 [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
 [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
 [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
 [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
 [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c

Cc Christoph Hellwig <[email protected]>

Cc Dan Williams <[email protected]>

Cc Sachin Sant <[email protected]>

Cc [email protected]

Cc Ira Weiny <[email protected]>

Cc Jason Gunthorpe <[email protected]>

Signed-off-by Aneesh Kumar K.V <[email protected]>

---
 mm/memremap.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

mm/memremap.c +15 -4

index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
 {
 	unsigned long pfn;
 	int i;
+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}
 
 	dev_pagemap_kill(pgmap);
 	for (i = 0; i < pgmap->nr_range; i++)

@@ -169,7 +179,8 @@ void memunmap_pages(struct dev_pagemap *pgmap)
 		pageunmap_range(pgmap, i);
 
 	WARN_ONCE(pgmap->altmap.alloc, "failed to free all reserved pages\n");
-	devmap_managed_enable_put();
+	if (need_devmap_managed)
+		devmap_managed_enable_put();
 }
 EXPORT_SYMBOL_GPL(memunmap_pages);

@@ -307,7 +318,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		.pgprot = PAGE_KERNEL,
 	};
 	const int nr_range = pgmap->nr_range;
-	bool need_devmap_managed = true;
+	bool need_devmap_managed = false;
 	int error, i;
 
 	if (WARN_ONCE(!nr_range, "nr_range must be specified\n"))

@@ -327,6 +338,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 			WARN(1, "Missing owner\n");
 			return ERR_PTR(-EINVAL);
 		}
+		need_devmap_managed = true;
 		break;
 	case MEMORY_DEVICE_FS_DAX:
 		if (!IS_ENABLED(CONFIG_ZONE_DEVICE) ||

@@ -334,13 +346,12 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 			WARN(1, "File system DAX not supported\n");
 			return ERR_PTR(-EINVAL);
 		}
+		need_devmap_managed = true;
 		break;
 	case MEMORY_DEVICE_GENERIC:
-		need_devmap_managed = false;
 		break;
 	case MEMORY_DEVICE_PCI_P2PDMA:
 		params.pgprot = pgprot_noncached(params.pgprot);
-		need_devmap_managed = false;
 		break;
 	default:
 		WARN(1, "Invalid pgmap type %d\n", pgmap->type);

Signature · 1 line

2.26.2

Sachin Sant

Tested-by on v1 1/1

2020-10-22 08:34 UTC #1 lore raw

Quoted 7 lines

jump label: negative count!

WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
Modules linked in:

….

Quoted 8 lines

Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
—

Tested-by Sachin Sant <[email protected]>

-Sachin

Christoph Hellwig

Reviewed-by on v1 1/1

2020-10-22 13:26 UTC #2 lore raw

In reply to Aneesh Kumar K.V

Looks good,

Reviewed-by Christoph Hellwig <[email protected]>

Ira Weiny

2020-10-22 15:41 UTC #3 lore raw

In reply to Aneesh Kumar K.V

Quoted 59 lines · Aneesh Kumar K.V

commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().
devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.
jump label: negative count!
 WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
 Modules linked in:
 ....
NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0 LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0
 Call Trace:
 [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
 [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
 [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
 [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
 [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
 [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
 [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
 [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
 [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
 [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
 [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
 [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
 [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
 [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
 [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
 [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
 [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
 [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
 [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
 [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
 [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
---
 mm/memremap.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)
mm/memremap.c +0 -0
index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
 {
 	unsigned long pfn;
 	int i;

mm/memremap.c void memunmap_pages(struct dev_pagemap *pgmap) · line 170

Quoted 10 lines

mm/memremap.c +10 -0

+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}

Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?

Quoted 14 lines

mm/memremap.c +2 -1

 	dev_pagemap_kill(pgmap);
 	for (i = 0; i < pgmap->nr_range; i++)

@@ -169,7 +179,8 @@ void memunmap_pages(struct dev_pagemap *pgmap)
 		pageunmap_range(pgmap, i);
 
 	WARN_ONCE(pgmap->altmap.alloc, "failed to free all reserved pages\n");
-	devmap_managed_enable_put();
+	if (need_devmap_managed)
+		devmap_managed_enable_put();
 }
 EXPORT_SYMBOL_GPL(memunmap_pages);

mm/memremap.c void *memremap_pages(struct dev_pagemap *pgmap, int nid) · line 321

Quoted 5 lines

mm/memremap.c +1 -1

@@ -307,7 +318,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		.pgprot = PAGE_KERNEL,
 	};
 	const int nr_range = pgmap->nr_range;
-	bool need_devmap_managed = true;
+	bool need_devmap_managed = false;

I'm CC'ing Ralph Campbell because I think some of his work has proposed this same change.

Ira

Quoted 29 lines

mm/memremap.c +2 -2

 	int error, i;
 
 	if (WARN_ONCE(!nr_range, "nr_range must be specified\n"))

@@ -327,6 +338,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 			WARN(1, "Missing owner\n");
 			return ERR_PTR(-EINVAL);
 		}
+		need_devmap_managed = true;
 		break;
 	case MEMORY_DEVICE_FS_DAX:
 		if (!IS_ENABLED(CONFIG_ZONE_DEVICE) ||

@@ -334,13 +346,12 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 			WARN(1, "File system DAX not supported\n");
 			return ERR_PTR(-EINVAL);
 		}
+		need_devmap_managed = true;
 		break;
 	case MEMORY_DEVICE_GENERIC:
-		need_devmap_managed = false;
 		break;
 	case MEMORY_DEVICE_PCI_P2PDMA:
 		params.pgprot = pgprot_noncached(params.pgprot);
-		need_devmap_managed = false;
 		break;
 	default:
 		WARN(1, "Invalid pgmap type %d\n", pgmap->type);

Signature · 1 line

2.26.2

Ralph Campbell

2020-10-22 18:20 UTC #4 lore raw

Quoted 65 lines · Ira Weiny

Quoted 64 lines · Aneesh Kumar K.V
commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().
devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.
jump label: negative count!
  WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
  Modules linked in:
  ....
NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0 LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0
  Call Trace:
  [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
  [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
  [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
  [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
  [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
  [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
  [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
  [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
  [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
  [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
  [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
  [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
  [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
  [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
  [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
  [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
  [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
  [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
  [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
  [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
  [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
---
  mm/memremap.c | 19 +++++++++++++++----
  1 file changed, 15 insertions(+), 4 deletions(-)
mm/memremap.c +5 -0
index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
  {
  	unsigned long pfn;
  	int i;
+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:

Quoted 8 lines

mm/memremap.c void memunmap_pages(struct dev_pagemap *pgmap) · line 170
Quoted 5 lines
mm/memremap.c +5 -0
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}
Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?

Perhaps. I can imagine a helper defined in include/linux/mm.h which page_is_devmap_managed() could also call but that would impact a lot of places that include mm.h. Since memremap.c already has to have intimate knowledge of the pgmap->type, I think limiting the change to just what is needed is better for now. So the patch looks OK to me.

Looking at this some more, I would suggest changing devmap_managed_enable_get() and devmap_managed_enable_put() to do the special case checking instead of doing it in memremap_pages() and memunmap_pages().
Then devmap_managed_enable_get() doesn't need to return an error if CONFIG_DEV_PAGEMAP_OPS isn't defined. I have only compile tested the following.

Signed-off-by Ralph Campbell <[email protected]>

---
 mm/memremap.c | 39 ++++++++++++++++-----------------------
 1 file changed, 16 insertions(+), 23 deletions(-)

mm/memremap.c +16 ~~-23~~

index 73a206d0f645..16b2fb482da1 100644
@@ -41,28 +41,24 @@ EXPORT_SYMBOL_GPL(memremap_compat_align);
 DEFINE_STATIC_KEY_FALSE(devmap_managed_key);
 EXPORT_SYMBOL(devmap_managed_key);
 
-static void devmap_managed_enable_put(void)
+static void devmap_managed_enable_put(struct dev_pagemap *pgmap)
 {
-	static_branch_dec(&devmap_managed_key);
+	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
+	    pgmap->type == MEMORY_DEVICE_FS_DAX)
+		static_branch_dec(&devmap_managed_key);
 }
 
-static int devmap_managed_enable_get(struct dev_pagemap *pgmap)
+static void devmap_managed_enable_get(struct dev_pagemap *pgmap)
 {
-	if (pgmap->type == MEMORY_DEVICE_PRIVATE &&
-	    (!pgmap->ops || !pgmap->ops->page_free)) {
-		WARN(1, "Missing page_free method\n");
-		return -EINVAL;
-	}
-
-	static_branch_inc(&devmap_managed_key);
-	return 0;
+	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
+	    pgmap->type == MEMORY_DEVICE_FS_DAX)
+		static_branch_inc(&devmap_managed_key);
 }
 #else
-static int devmap_managed_enable_get(struct dev_pagemap *pgmap)
+static void devmap_managed_enable_get(struct dev_pagemap *pgmap)
 {
-	return -EINVAL;
 }
-static void devmap_managed_enable_put(void)
+static void devmap_managed_enable_put(struct dev_pagemap *pgmap)
 {
 }
 #endif /* CONFIG_DEV_PAGEMAP_OPS */

@@ -169,7 +165,7 @@ void memunmap_pages(struct dev_pagemap *pgmap)
 		pageunmap_range(pgmap, i);
 
 	WARN_ONCE(pgmap->altmap.alloc, "failed to free all reserved pages\n");
-	devmap_managed_enable_put();
+	devmap_managed_enable_put(pgmap);
 }
 EXPORT_SYMBOL_GPL(memunmap_pages);

@@ -307,7 +303,6 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		.pgprot = PAGE_KERNEL,
 	};
 	const int nr_range = pgmap->nr_range;
-	bool need_devmap_managed = true;
 	int error, i;
 
 	if (WARN_ONCE(!nr_range, "nr_range must be specified\n"))

@@ -323,6 +318,10 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 			WARN(1, "Missing migrate_to_ram method\n");
 			return ERR_PTR(-EINVAL);
 		}
+		if (!pgmap->ops->page_free) {
+			WARN(1, "Missing page_free method\n");
+			return ERR_PTR(-EINVAL);
+		}
 		if (!pgmap->owner) {
 			WARN(1, "Missing owner\n");
 			return ERR_PTR(-EINVAL);

@@ -336,11 +335,9 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		}
 		break;
 	case MEMORY_DEVICE_GENERIC:
-		need_devmap_managed = false;
 		break;
 	case MEMORY_DEVICE_PCI_P2PDMA:
 		params.pgprot = pgprot_noncached(params.pgprot);
-		need_devmap_managed = false;
 		break;
 	default:
 		WARN(1, "Invalid pgmap type %d\n", pgmap->type);

@@ -364,11 +361,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		}
 	}
 
-	if (need_devmap_managed) {
-		error = devmap_managed_enable_get(pgmap);
-		if (error)
-			return ERR_PTR(error);
-	}
+	devmap_managed_enable_get(pgmap);
 
 	/*
 	 * Clear the pgmap nr_range as it will be incremented for each

Signature · 1 line

2.20.1

Quoted 23 lines

Quoted 14 lines

mm/memremap.c +2 -1

  	dev_pagemap_kill(pgmap);
  	for (i = 0; i < pgmap->nr_range; i++)

@@ -169,7 +179,8 @@ void memunmap_pages(struct dev_pagemap *pgmap)
  		pageunmap_range(pgmap, i);
 
  	WARN_ONCE(pgmap->altmap.alloc, "failed to free all reserved pages\n");
-	devmap_managed_enable_put();
+	if (need_devmap_managed)
+		devmap_managed_enable_put();
  }
  EXPORT_SYMBOL_GPL(memunmap_pages);

mm/memremap.c void *memremap_pages(struct dev_pagemap *pgmap, int nid) · line 321

Quoted 5 lines

mm/memremap.c +1 -1

@@ -307,7 +318,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
  		.pgprot = PAGE_KERNEL,
  	};
  	const int nr_range = pgmap->nr_range;
-	bool need_devmap_managed = true;
+	bool need_devmap_managed = false;

I'm CC'ing Ralph Campbell because I think some of his work has proposed this same change.

Quoted 1 line

Ira

This part of the patch isn't strictly needed, it just reverses the default value of need_devmap_managed.

Quoted 29 lines

mm/memremap.c +2 -2

  	int error, i;
 
  	if (WARN_ONCE(!nr_range, "nr_range must be specified\n"))

@@ -327,6 +338,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
  			WARN(1, "Missing owner\n");
  			return ERR_PTR(-EINVAL);
  		}
+		need_devmap_managed = true;
  		break;
  	case MEMORY_DEVICE_FS_DAX:
  		if (!IS_ENABLED(CONFIG_ZONE_DEVICE) ||

@@ -334,13 +346,12 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
  			WARN(1, "File system DAX not supported\n");
  			return ERR_PTR(-EINVAL);
  		}
+		need_devmap_managed = true;
  		break;
  	case MEMORY_DEVICE_GENERIC:
-		need_devmap_managed = false;
  		break;
  	case MEMORY_DEVICE_PCI_P2PDMA:
  		params.pgprot = pgprot_noncached(params.pgprot);
-		need_devmap_managed = false;
  		break;
  	default:
  		WARN(1, "Invalid pgmap type %d\n", pgmap->type);

Signature · 1 line

2.26.2

Ira Weiny

Reviewed-by

2020-10-22 19:10 UTC #5 lore raw

Quoted 82 lines · Ralph Campbell

Quoted 65 lines · Ira Weiny
Quoted 64 lines · Aneesh Kumar K.V
commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().
devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.
jump label: negative count!
  WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
  Modules linked in:
  ....
NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0 LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0
  Call Trace:
  [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
  [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
  [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
  [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
  [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
  [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
  [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
  [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
  [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
  [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
  [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
  [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
  [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
  [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
  [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
  [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
  [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
  [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
  [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
  [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
  [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
---
  mm/memremap.c | 19 +++++++++++++++----
  1 file changed, 15 insertions(+), 4 deletions(-)
mm/memremap.c +5 -0
index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
  {
  	unsigned long pfn;
  	int i;
+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:
Quoted 8 lines
mm/memremap.c void memunmap_pages(struct dev_pagemap *pgmap) · line 170
Quoted 5 lines
mm/memremap.c +5 -0
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}
Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?
Perhaps. I can imagine a helper defined in include/linux/mm.h which page_is_devmap_managed() could also call but that would impact a lot of places that include mm.h. Since memremap.c already has to have intimate knowledge of the pgmap->type, I think limiting the change to just what is needed is better for now. So the patch looks OK to me.

Quoted 6 lines

Looking at this some more, I would suggest changing devmap_managed_enable_get() and devmap_managed_enable_put() to do the special case checking instead of doing it in memremap_pages() and memunmap_pages().
Then devmap_managed_enable_get() doesn't need to return an error if CONFIG_DEV_PAGEMAP_OPS isn't defined. I have only compile tested the following.

This looks cleaner to me. Aneesh?

FWIW:

Reviewed-by Ira Weiny <[email protected]>

Ira

Quoted 151 lines

Signed-off-by Ralph Campbell <[email protected]>

---
 mm/memremap.c | 39 ++++++++++++++++-----------------------
 1 file changed, 16 insertions(+), 23 deletions(-)

mm/memremap.c +16 ~~-23~~

index 73a206d0f645..16b2fb482da1 100644
@@ -41,28 +41,24 @@ EXPORT_SYMBOL_GPL(memremap_compat_align);
 DEFINE_STATIC_KEY_FALSE(devmap_managed_key);
 EXPORT_SYMBOL(devmap_managed_key);
-static void devmap_managed_enable_put(void)
+static void devmap_managed_enable_put(struct dev_pagemap *pgmap)
 {
-	static_branch_dec(&devmap_managed_key);
+	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
+	    pgmap->type == MEMORY_DEVICE_FS_DAX)
+		static_branch_dec(&devmap_managed_key);
 }
-static int devmap_managed_enable_get(struct dev_pagemap *pgmap)
+static void devmap_managed_enable_get(struct dev_pagemap *pgmap)
 {
-	if (pgmap->type == MEMORY_DEVICE_PRIVATE &&
-	    (!pgmap->ops || !pgmap->ops->page_free)) {
-		WARN(1, "Missing page_free method\n");
-		return -EINVAL;
-	}
-
-	static_branch_inc(&devmap_managed_key);
-	return 0;
+	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
+	    pgmap->type == MEMORY_DEVICE_FS_DAX)
+		static_branch_inc(&devmap_managed_key);
 }
 #else
-static int devmap_managed_enable_get(struct dev_pagemap *pgmap)
+static void devmap_managed_enable_get(struct dev_pagemap *pgmap)
 {
-	return -EINVAL;
 }
-static void devmap_managed_enable_put(void)
+static void devmap_managed_enable_put(struct dev_pagemap *pgmap)
 {
 }
 #endif /* CONFIG_DEV_PAGEMAP_OPS */

@@ -169,7 +165,7 @@ void memunmap_pages(struct dev_pagemap *pgmap)
 		pageunmap_range(pgmap, i);
 	WARN_ONCE(pgmap->altmap.alloc, "failed to free all reserved pages\n");
-	devmap_managed_enable_put();
+	devmap_managed_enable_put(pgmap);
 }
 EXPORT_SYMBOL_GPL(memunmap_pages);

@@ -307,7 +303,6 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		.pgprot = PAGE_KERNEL,
 	};
 	const int nr_range = pgmap->nr_range;
-	bool need_devmap_managed = true;
 	int error, i;
 	if (WARN_ONCE(!nr_range, "nr_range must be specified\n"))

@@ -323,6 +318,10 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 			WARN(1, "Missing migrate_to_ram method\n");
 			return ERR_PTR(-EINVAL);
 		}
+		if (!pgmap->ops->page_free) {
+			WARN(1, "Missing page_free method\n");
+			return ERR_PTR(-EINVAL);
+		}
 		if (!pgmap->owner) {
 			WARN(1, "Missing owner\n");
 			return ERR_PTR(-EINVAL);

@@ -336,11 +335,9 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		}
 		break;
 	case MEMORY_DEVICE_GENERIC:
-		need_devmap_managed = false;
 		break;
 	case MEMORY_DEVICE_PCI_P2PDMA:
 		params.pgprot = pgprot_noncached(params.pgprot);
-		need_devmap_managed = false;
 		break;
 	default:
 		WARN(1, "Invalid pgmap type %d\n", pgmap->type);

@@ -364,11 +361,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
 		}
 	}
-	if (need_devmap_managed) {
-		error = devmap_managed_enable_get(pgmap);
-		if (error)
-			return ERR_PTR(error);
-	}
+	devmap_managed_enable_get(pgmap);
 	/*
 	 * Clear the pgmap nr_range as it will be incremented for each

Signature · 1 line

2.20.1

Quoted 20 lines

Quoted 11 lines

mm/memremap.c +2 -1

  	dev_pagemap_kill(pgmap);
  	for (i = 0; i < pgmap->nr_range; i++)

@@ -169,7 +179,8 @@ void memunmap_pages(struct dev_pagemap *pgmap)
  		pageunmap_range(pgmap, i);
  	WARN_ONCE(pgmap->altmap.alloc, "failed to free all reserved pages\n");
-	devmap_managed_enable_put();
+	if (need_devmap_managed)
+		devmap_managed_enable_put();
  }
  EXPORT_SYMBOL_GPL(memunmap_pages);

mm/memremap.c void *memremap_pages(struct dev_pagemap *pgmap, int nid) · line 321

Quoted 5 lines

mm/memremap.c +1 -1

@@ -307,7 +318,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
  		.pgprot = PAGE_KERNEL,
  	};
  	const int nr_range = pgmap->nr_range;
-	bool need_devmap_managed = true;
+	bool need_devmap_managed = false;

I'm CC'ing Ralph Campbell because I think some of his work has proposed this same change.

Quoted 1 line

Ira

This part of the patch isn't strictly needed, it just reverses the default value of need_devmap_managed.

Quoted 20 lines

mm/memremap.c +2 -1

  	int error, i;
  	if (WARN_ONCE(!nr_range, "nr_range must be specified\n"))

@@ -327,6 +338,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
  			WARN(1, "Missing owner\n");
  			return ERR_PTR(-EINVAL);
  		}
+		need_devmap_managed = true;
  		break;
  	case MEMORY_DEVICE_FS_DAX:
  		if (!IS_ENABLED(CONFIG_ZONE_DEVICE) ||

@@ -334,13 +346,12 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid)
  			WARN(1, "File system DAX not supported\n");
  			return ERR_PTR(-EINVAL);
  		}
+		need_devmap_managed = true;
  		break;
  	case MEMORY_DEVICE_GENERIC:
-		need_devmap_managed = false;
  		break;
  	case MEMORY_DEVICE_PCI_P2PDMA:

Quoted 8 lines

mm/memremap.c +0 -1

  		params.pgprot = pgprot_noncached(params.pgprot);
-		need_devmap_managed = false;
  		break;
  	default:
  		WARN(1, "Invalid pgmap type %d\n", pgmap->type);

Signature · 1 line

2.26.2

_______________________________________________
Linux-nvdimm mailing list -- [email protected] To unsubscribe send an email to [email protected]

Aneesh Kumar K.V

author Reviewed-by

2020-10-23 02:57 UTC #6 lore raw

Quoted 90 lines · Ira Weiny

Quoted 82 lines · Ralph Campbell
Quoted 65 lines · Ira Weiny
Quoted 64 lines · Aneesh Kumar K.V
commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().
devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.
   jump label: negative count!
   WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
   Modules linked in:
   ....

   NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0
   LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0
   Call Trace:
   [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
   [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
   [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
   [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
   [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
   [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
   [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
   [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
   [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
   [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
   [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
   [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
   [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
   [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
   [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
   [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
   [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
   [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
   [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
   [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
   [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
---
   mm/memremap.c | 19 +++++++++++++++----
   1 file changed, 15 insertions(+), 4 deletions(-)
mm/memremap.c +5 -0
index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
   {
   	unsigned long pfn;
   	int i;
+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:
Quoted 8 lines
mm/memremap.c void memunmap_pages(struct dev_pagemap *pgmap) · line 170
Quoted 5 lines
mm/memremap.c +5 -0
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}
Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?
Perhaps. I can imagine a helper defined in include/linux/mm.h which page_is_devmap_managed() could also call but that would impact a lot of places that include mm.h. Since memremap.c already has to have intimate knowledge of the pgmap->type, I think limiting the change to just what is needed is better for now. So the patch looks OK to me.
Quoted 6 lines
Looking at this some more, I would suggest changing devmap_managed_enable_get() and devmap_managed_enable_put() to do the special case checking instead of doing it in memremap_pages() and memunmap_pages().
Then devmap_managed_enable_get() doesn't need to return an error if CONFIG_DEV_PAGEMAP_OPS isn't defined. I have only compile tested the following.

Quoted 6 lines

This looks cleaner to me. Aneesh?
FWIW:
Reviewed-by Ira Weiny <[email protected]>

Yes. You can add

Reviewed-by Aneesh Kumar K.V <[email protected]>

-aneesh

Sachin Sant

Tested-by

2020-10-23 06:39 UTC #7 lore raw

In reply to Ralph Campbell

Quoted 17 lines

Quoted 3 lines
Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?
Perhaps. I can imagine a helper defined in include/linux/mm.h which page_is_devmap_managed() could also call but that would impact a lot of places that include mm.h. Since memremap.c already has to have intimate knowledge of the pgmap->type, I think limiting the change to just what is needed is better for now. So the patch looks OK to me.
Looking at this some more, I would suggest changing devmap_managed_enable_get() and devmap_managed_enable_put() to do the special case checking instead of doing it in memremap_pages() and memunmap_pages().
Then devmap_managed_enable_get() doesn't need to return an error if CONFIG_DEV_PAGEMAP_OPS isn't defined. I have only compile tested the following.

Quoted 2 lines

Signed-off-by Ralph Campbell <[email protected]>
—

This patch fixes the warning for me.

Tested-by Sachin Sant <[email protected]>

-Sachin

Christoph Hellwig

2020-10-23 06:46 UTC #8 lore raw

In reply to Ralph Campbell

Quoted 82 lines · Ralph Campbell

Quoted 65 lines · Ira Weiny
Quoted 64 lines · Aneesh Kumar K.V
commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().
devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.
jump label: negative count!
  WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
  Modules linked in:
  ....
NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0 LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0
  Call Trace:
  [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
  [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
  [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
  [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
  [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
  [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
  [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
  [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
  [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
  [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
  [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
  [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
  [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
  [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
  [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
  [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
  [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
  [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
  [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
  [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
  [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
---
  mm/memremap.c | 19 +++++++++++++++----
  1 file changed, 15 insertions(+), 4 deletions(-)
mm/memremap.c +5 -0
index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
  {
  	unsigned long pfn;
  	int i;
+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:
Quoted 8 lines
mm/memremap.c void memunmap_pages(struct dev_pagemap *pgmap) · line 170
Quoted 5 lines
mm/memremap.c +5 -0
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}
Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?
Perhaps. I can imagine a helper defined in include/linux/mm.h which page_is_devmap_managed() could also call but that would impact a lot of places that include mm.h. Since memremap.c already has to have intimate knowledge of the pgmap->type, I think limiting the change to just what is needed is better for now. So the patch looks OK to me.

Quoted 8 lines

Looking at this some more, I would suggest changing devmap_managed_enable_get() and devmap_managed_enable_put() to do the special case checking instead of doing it in memremap_pages() and memunmap_pages().
Then devmap_managed_enable_get() doesn't need to return an error if CONFIG_DEV_PAGEMAP_OPS isn't defined. I have only compile tested the following.
Signed-off-by Ralph Campbell <[email protected]>

This looks ok as well. Can you submit it as a proper standalone patch?

Ralph Campbell

2020-10-23 17:29 UTC #9 lore raw

Quoted 90 lines · Christoph Hellwig

Quoted 89 lines · Ralph Campbell
Quoted 65 lines · Ira Weiny
Quoted 64 lines · Aneesh Kumar K.V
commit 6f42193fd86e ("memremap: don't use a separate devm action for devmap_managed_enable_get") changed the static key updates such that we now call devmap_managed_enable_put() without doing the equivalent devmap_managed_enable_get().
devmap_managed_enable_get() is only called for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_FS_DAX, But memunmap_pages() get called for other pgmap types too. This results in the below warning when switching between system-ram and devdax mode for devdax namespace.
   jump label: negative count!
   WARNING: CPU: 52 PID: 1335 at kernel/jump_label.c:235 static_key_slow_try_dec+0x88/0xa0
   Modules linked in:
   ....

   NIP [c000000000433318] static_key_slow_try_dec+0x88/0xa0
   LR [c000000000433314] static_key_slow_try_dec+0x84/0xa0
   Call Trace:
   [c000000025c1f660] [c000000000433314] static_key_slow_try_dec+0x84/0xa0 (unreliable)
   [c000000025c1f6d0] [c000000000433664] __static_key_slow_dec_cpuslocked+0x34/0xd0
   [c000000025c1f700] [c0000000004337a4] static_key_slow_dec+0x54/0xf0
   [c000000025c1f770] [c00000000059c49c] memunmap_pages+0x36c/0x500
   [c000000025c1f820] [c000000000d91d10] devm_action_release+0x30/0x50
   [c000000025c1f840] [c000000000d92e34] release_nodes+0x2f4/0x3e0
   [c000000025c1f8f0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
   [c000000025c1f930] [c000000000d883a4] bus_remove_device+0x124/0x210
   [c000000025c1f9b0] [c000000000d80ef4] device_del+0x1d4/0x530
   [c000000025c1fa70] [c000000000e341e8] unregister_dev_dax+0x48/0xe0
   [c000000025c1fae0] [c000000000d91d10] devm_action_release+0x30/0x50
   [c000000025c1fb00] [c000000000d92e34] release_nodes+0x2f4/0x3e0
   [c000000025c1fbb0] [c000000000d8b15c] device_release_driver_internal+0x17c/0x280
   [c000000025c1fbf0] [c000000000d87000] unbind_store+0x130/0x170
   [c000000025c1fc30] [c000000000d862a0] drv_attr_store+0x40/0x60
   [c000000025c1fc50] [c0000000006d316c] sysfs_kf_write+0x6c/0xb0
   [c000000025c1fc90] [c0000000006d2328] kernfs_fop_write+0x118/0x280
   [c000000025c1fce0] [c0000000005a79f8] vfs_write+0xe8/0x2a0
   [c000000025c1fd30] [c0000000005a7d94] ksys_write+0x84/0x140
   [c000000025c1fd80] [c00000000003a430] system_call_exception+0x120/0x270
   [c000000025c1fe20] [c00000000000c540] system_call_common+0xf0/0x27c
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Sachin Sant <[email protected]>
Cc [email protected]
Cc Ira Weiny <[email protected]>
Cc Jason Gunthorpe <[email protected]>
Signed-off-by Aneesh Kumar K.V <[email protected]>
---
   mm/memremap.c | 19 +++++++++++++++----
   1 file changed, 15 insertions(+), 4 deletions(-)
mm/memremap.c +5 -0
index 73a206d0f645..d4402ff3e467 100644
@@ -158,6 +158,16 @@ void memunmap_pages(struct dev_pagemap *pgmap)
   {
   	unsigned long pfn;
   	int i;
+	bool need_devmap_managed = false;
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_FS_DAX:
Quoted 8 lines
mm/memremap.c void memunmap_pages(struct dev_pagemap *pgmap) · line 170
Quoted 5 lines
mm/memremap.c +5 -0
+		need_devmap_managed = true;
+		break;
+	default:
+		break;
+	}
Is it overkill to avoid duplicating this switch logic in page_is_devmap_managed() by creating another call which can be used here?
Perhaps. I can imagine a helper defined in include/linux/mm.h which page_is_devmap_managed() could also call but that would impact a lot of places that include mm.h. Since memremap.c already has to have intimate knowledge of the pgmap->type, I think limiting the change to just what is needed is better for now. So the patch looks OK to me.
Looking at this some more, I would suggest changing devmap_managed_enable_get() and devmap_managed_enable_put() to do the special case checking instead of doing it in memremap_pages() and memunmap_pages().
Then devmap_managed_enable_get() doesn't need to return an error if CONFIG_DEV_PAGEMAP_OPS isn't defined. I have only compile tested the following.

Quoted 4 lines

Quoted 1 line
Signed-off-by Ralph Campbell <[email protected]>
This looks ok as well. Can you submit it as a proper standalone patch?

Yes. I was just about to ask whether I should do that and you beat me to it.