[v7,2/3] mm/memory_hotplug: split memmap_on_memory requests across memblocks

Commit Message

Verma, Vishal L Oct. 25, 2023, 10:44 p.m. UTC

The MHP_MEMMAP_ON_MEMORY flag for hotplugged memory is restricted to
'memblock_size' chunks of memory being added. Adding a larger span of
memory precludes memmap_on_memory semantics.

For users of hotplug such as kmem, large amounts of memory might get
added from the CXL subsystem. In some cases, this amount may exceed the
available 'main memory' to store the memmap for the memory being added.
In this case, it is useful to have a way to place the memmap on the
memory being added, even if it means splitting the addition into
memblock-sized chunks.

Change add_memory_resource() to loop over memblock-sized chunks of
memory if caller requested memmap_on_memory, and if other conditions for
it are met. Teach try_remove_memory() to also expect that a memory
range being removed might have been split up into memblock sized chunks,
and to loop through those as needed.

This does preclude being able to use PUD mappings in the direct map; a
proposal to how this could be optimized in the future is laid out
here[1].

[1]: https://lore.kernel.org/linux-mm/[e-mail]/

Cc: Andrew Morton <[e-mail]>
Cc: David Hildenbrand <[e-mail]>
Cc: Michal Hocko <[e-mail]>
Cc: Oscar Salvador <[e-mail]>
Cc: Dan Williams <[e-mail]>
Cc: Dave Jiang <[e-mail]>
Cc: Dave Hansen <[e-mail]>
Cc: Huang Ying <[e-mail]>
Suggested-by: David Hildenbrand <[e-mail]>
Reviewed-by: Dan Williams <[e-mail]>
Signed-off-by: Vishal Verma <[e-mail]>
---
 mm/memory_hotplug.c | 209 ++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 144 insertions(+), 65 deletions(-)

Comments

Huang, Ying Oct. 30, 2023, 2:46 a.m. UTC | #1

Vishal Verma <[e-mail]> writes:

> The MHP_MEMMAP_ON_MEMORY flag for hotplugged memory is restricted to
> 'memblock_size' chunks of memory being added. Adding a larger span of
> memory precludes memmap_on_memory semantics.
>
> For users of hotplug such as kmem, large amounts of memory might get
> added from the CXL subsystem. In some cases, this amount may exceed the
> available 'main memory' to store the memmap for the memory being added.
> In this case, it is useful to have a way to place the memmap on the
> memory being added, even if it means splitting the addition into
> memblock-sized chunks.
>
> Change add_memory_resource() to loop over memblock-sized chunks of
> memory if caller requested memmap_on_memory, and if other conditions for
> it are met. Teach try_remove_memory() to also expect that a memory
> range being removed might have been split up into memblock sized chunks,
> and to loop through those as needed.
>
> This does preclude being able to use PUD mappings in the direct map; a
> proposal to how this could be optimized in the future is laid out
> here[1].
>
> [1]: https://lore.kernel.org/linux-mm/[e-mail]/
>
> Cc: Andrew Morton <[e-mail]>
> Cc: David Hildenbrand <[e-mail]>
> Cc: Michal Hocko <[e-mail]>
> Cc: Oscar Salvador <[e-mail]>
> Cc: Dan Williams <[e-mail]>
> Cc: Dave Jiang <[e-mail]>
> Cc: Dave Hansen <[e-mail]>
> Cc: Huang Ying <[e-mail]>
> Suggested-by: David Hildenbrand <[e-mail]>
> Reviewed-by: Dan Williams <[e-mail]>
> Signed-off-by: Vishal Verma <[e-mail]>
> ---
>  mm/memory_hotplug.c | 209 ++++++++++++++++++++++++++++++++++++----------------
>  1 file changed, 144 insertions(+), 65 deletions(-)
>
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 6be7de9efa55..b97035193090 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -1380,6 +1380,48 @@ static bool mhp_supports_memmap_on_memory(unsigned long size)
>  	return arch_supports_memmap_on_memory(vmemmap_size);
>  }
>  
> +static int create_altmaps_and_memory_blocks(int nid, struct memory_group *group,
> +					    u64 start, u64 size)
> +{
> +	unsigned long memblock_size = memory_block_size_bytes();
> +	u64 cur_start;
> +	int ret;
> +
> +	for (cur_start = start; cur_start < start + size;
> +	     cur_start += memblock_size) {
> +		struct mhp_params params = { .pgprot =
> +						     pgprot_mhp(PAGE_KERNEL) };
> +		struct vmem_altmap mhp_altmap = {
> +			.base_pfn = PHYS_PFN(cur_start),
> +			.end_pfn = PHYS_PFN(cur_start + memblock_size - 1),
> +		};
> +
> +		mhp_altmap.free = memory_block_memmap_on_memory_pages();
> +		params.altmap = kmemdup(&mhp_altmap, sizeof(struct vmem_altmap),
> +					GFP_KERNEL);
> +		if (!params.altmap)
> +			return -ENOMEM;
> +
> +		/* call arch's memory hotadd */
> +		ret = arch_add_memory(nid, cur_start, memblock_size, &params);
> +		if (ret < 0) {
> +			kfree(params.altmap);

Should we call

        remove_memory_blocks_and_altmaps(start, cur_start - start);

here to clean up resources?

--
Best Regards,
Huang, Ying

> +			return ret;
> +		}
> +
> +		/* create memory block devices after memory was added */
> +		ret = create_memory_block_devices(cur_start, memblock_size,
> +						  params.altmap, group);
> +		if (ret) {
> +			arch_remove_memory(cur_start, memblock_size, NULL);
> +			kfree(params.altmap);
> +			return ret;
> +		}
> +	}
> +
> +	return 0;
> +}
> +
>  /*
>   * NOTE: The caller must call lock_device_hotplug() to serialize hotplug
>   * and online/offline operations (triggered e.g. by sysfs).
> @@ -1390,10 +1432,6 @@ int __ref add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
>  {
>  	struct mhp_params params = { .pgprot = pgprot_mhp(PAGE_KERNEL) };
>  	enum memblock_flags memblock_flags = MEMBLOCK_NONE;
> -	struct vmem_altmap mhp_altmap = {
> -		.base_pfn =  PHYS_PFN(res->start),
> -		.end_pfn  =  PHYS_PFN(res->end),
> -	};
>  	struct memory_group *group = NULL;
>  	u64 start, size;
>  	bool new_node = false;
> @@ -1436,28 +1474,22 @@ int __ref add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
>  	/*
>  	 * Self hosted memmap array
>  	 */
> -	if (mhp_flags & MHP_MEMMAP_ON_MEMORY) {
> -		if (mhp_supports_memmap_on_memory(size)) {
> -			mhp_altmap.free = memory_block_memmap_on_memory_pages();
> -			params.altmap = kmemdup(&mhp_altmap,
> -						sizeof(struct vmem_altmap),
> -						GFP_KERNEL);
> -			if (!params.altmap)
> -				goto error;
> +	if ((mhp_flags & MHP_MEMMAP_ON_MEMORY) &&
> +	    mhp_supports_memmap_on_memory(memory_block_size_bytes())) {
> +		ret = create_altmaps_and_memory_blocks(nid, group, start, size);
> +		if (ret)
> +			goto error;
> +	} else {
> +		ret = arch_add_memory(nid, start, size, &params);
> +		if (ret < 0)
> +			goto error;
> +
> +		/* create memory block devices after memory was added */
> +		ret = create_memory_block_devices(start, size, NULL, group);
> +		if (ret) {
> +			arch_remove_memory(start, size, NULL);
> +			goto error;
>  		}
> -		/* fallback to not using altmap  */
> -	}
> -
> -	/* call arch's memory hotadd */
> -	ret = arch_add_memory(nid, start, size, &params);
> -	if (ret < 0)
> -		goto error_free;
> -
> -	/* create memory block devices after memory was added */
> -	ret = create_memory_block_devices(start, size, params.altmap, group);
> -	if (ret) {
> -		arch_remove_memory(start, size, NULL);
> -		goto error_free;
>  	}
>  
>  	if (new_node) {
> @@ -1494,8 +1526,6 @@ int __ref add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
>  		walk_memory_blocks(start, size, NULL, online_memory_block);
>  
>  	return ret;
> -error_free:
> -	kfree(params.altmap);
>  error:
>  	if (IS_ENABLED(CONFIG_ARCH_KEEP_MEMBLOCK))
>  		memblock_remove(start, size);
> @@ -2076,6 +2106,16 @@ static int test_has_altmap_cb(struct memory_block *mem, void *arg)
>  	return 0;
>  }
>  
> +static int count_memory_range_altmaps_cb(struct memory_block *mem, void *arg)
> +{
> +	u64 *num_altmaps = (u64 *)arg;
> +
> +	if (mem->altmap)
> +		*num_altmaps += 1;
> +
> +	return 0;
> +}
> +
>  static int check_cpu_on_node(int nid)
>  {
>  	int cpu;
> @@ -2146,11 +2186,69 @@ void try_offline_node(int nid)
>  }
>  EXPORT_SYMBOL(try_offline_node);
>  
> -static int __ref try_remove_memory(u64 start, u64 size)
> +static void __ref remove_memory_blocks_and_altmaps(u64 start, u64 size)
>  {
> -	struct memory_block *mem;
> -	int rc = 0, nid = NUMA_NO_NODE;
> +	unsigned long memblock_size = memory_block_size_bytes();
>  	struct vmem_altmap *altmap = NULL;
> +	struct memory_block *mem;
> +	u64 cur_start;
> +	int rc;
> +
> +	/*
> +	 * For memmap_on_memory, the altmaps could have been added on
> +	 * a per-memblock basis. Loop through the entire range if so,
> +	 * and remove each memblock and its altmap.
> +	 */
> +	for (cur_start = start; cur_start < start + size;
> +	     cur_start += memblock_size) {
> +		rc = walk_memory_blocks(cur_start, memblock_size, &mem,
> +					test_has_altmap_cb);
> +		if (rc) {
> +			altmap = mem->altmap;
> +			/*
> +			 * Mark altmap NULL so that we can add a debug
> +			 * check on memblock free.
> +			 */
> +			mem->altmap = NULL;
> +		}
> +
> +		remove_memory_block_devices(cur_start, memblock_size);
> +
> +		arch_remove_memory(cur_start, memblock_size, altmap);
> +
> +		/* Verify that all vmemmap pages have actually been freed. */
> +		if (altmap) {
> +			WARN(altmap->alloc, "Altmap not fully unmapped");
> +			kfree(altmap);
> +		}
> +	}
> +}
> +
> +static int memory_blocks_have_altmaps(u64 start, u64 size)
> +{
> +	u64 num_memblocks = size / memory_block_size_bytes();
> +	u64 num_altmaps = 0;
> +
> +	if (!mhp_memmap_on_memory())
> +		return 0;
> +
> +	walk_memory_blocks(start, size, &num_altmaps,
> +			   count_memory_range_altmaps_cb);
> +
> +	if (num_altmaps == 0)
> +		return 0;
> +
> +	if (num_memblocks != num_altmaps) {
> +		WARN_ONCE(1, "Not all memblocks in range have altmaps");
> +		return -EINVAL;
> +	}
> +
> +	return 1;
> +}
> +
> +static int __ref try_remove_memory(u64 start, u64 size)
> +{
> +	int rc, nid = NUMA_NO_NODE;
>  
>  	BUG_ON(check_hotplug_memory_range(start, size));
>  
> @@ -2167,45 +2265,25 @@ static int __ref try_remove_memory(u64 start, u64 size)
>  	if (rc)
>  		return rc;
>  
> -	/*
> -	 * We only support removing memory added with MHP_MEMMAP_ON_MEMORY in
> -	 * the same granularity it was added - a single memory block.
> -	 */
> -	if (mhp_memmap_on_memory()) {
> -		rc = walk_memory_blocks(start, size, &mem, test_has_altmap_cb);
> -		if (rc) {
> -			if (size != memory_block_size_bytes()) {
> -				pr_warn("Refuse to remove %#llx - %#llx,"
> -					"wrong granularity\n",
> -					start, start + size);
> -				return -EINVAL;
> -			}
> -			altmap = mem->altmap;
> -			/*
> -			 * Mark altmap NULL so that we can add a debug
> -			 * check on memblock free.
> -			 */
> -			mem->altmap = NULL;
> -		}
> -	}
> -
>  	/* remove memmap entry */
>  	firmware_map_remove(start, start + size, "System RAM");
>  
> -	/*
> -	 * Memory block device removal under the device_hotplug_lock is
> -	 * a barrier against racing online attempts.
> -	 */
> -	remove_memory_block_devices(start, size);
> -
>  	mem_hotplug_begin();
>  
> -	arch_remove_memory(start, size, altmap);
> -
> -	/* Verify that all vmemmap pages have actually been freed. */
> -	if (altmap) {
> -		WARN(altmap->alloc, "Altmap not fully unmapped");
> -		kfree(altmap);
> +	rc = memory_blocks_have_altmaps(start, size);
> +	if (rc < 0) {
> +		goto err;
> +	} else if (rc == 0) {
> +		/*
> +		 * Memory block device removal under the device_hotplug_lock is
> +		 * a barrier against racing online attempts.
> +		 * No altmaps present, do the removal directly
> +		 */
> +		remove_memory_block_devices(start, size);
> +		arch_remove_memory(start, size, NULL);
> +	} else {
> +		/* all memblocks in the range have altmaps */
> +		remove_memory_blocks_and_altmaps(start, size);
>  	}
>  
>  	if (IS_ENABLED(CONFIG_ARCH_KEEP_MEMBLOCK)) {
> @@ -2218,8 +2296,9 @@ static int __ref try_remove_memory(u64 start, u64 size)
>  	if (nid != NUMA_NO_NODE)
>  		try_offline_node(nid);
>  
> +err:
>  	mem_hotplug_done();
> -	return 0;
> +	return (rc < 0 ? rc : 0);
>  }
>  
>  /**

David Hildenbrand Oct. 30, 2023, 10:20 a.m. UTC | #2

On 26.10.23 00:44, Vishal Verma wrote:
> The MHP_MEMMAP_ON_MEMORY flag for hotplugged memory is restricted to
> 'memblock_size' chunks of memory being added. Adding a larger span of
> memory precludes memmap_on_memory semantics.
> 
> For users of hotplug such as kmem, large amounts of memory might get
> added from the CXL subsystem. In some cases, this amount may exceed the
> available 'main memory' to store the memmap for the memory being added.
> In this case, it is useful to have a way to place the memmap on the
> memory being added, even if it means splitting the addition into
> memblock-sized chunks.
> 
> Change add_memory_resource() to loop over memblock-sized chunks of
> memory if caller requested memmap_on_memory, and if other conditions for
> it are met. Teach try_remove_memory() to also expect that a memory
> range being removed might have been split up into memblock sized chunks,
> and to loop through those as needed.
> 
> This does preclude being able to use PUD mappings in the direct map; a
> proposal to how this could be optimized in the future is laid out
> here[1].


Almost there, I think :)

>   
> +static int create_altmaps_and_memory_blocks(int nid, struct memory_group *group,
> +					    u64 start, u64 size)
> +{
> +	unsigned long memblock_size = memory_block_size_bytes();
> +	u64 cur_start;
> +	int ret;
> +
> +	for (cur_start = start; cur_start < start + size;
> +	     cur_start += memblock_size) {
> +		struct mhp_params params = { .pgprot =
> +						     pgprot_mhp(PAGE_KERNEL) };
> +		struct vmem_altmap mhp_altmap = {
> +			.base_pfn = PHYS_PFN(cur_start),
> +			.end_pfn = PHYS_PFN(cur_start + memblock_size - 1),
> +		};
> +
> +		mhp_altmap.free = memory_block_memmap_on_memory_pages();
> +		params.altmap = kmemdup(&mhp_altmap, sizeof(struct vmem_altmap),
> +					GFP_KERNEL);
> +		if (!params.altmap)
> +			return -ENOMEM;

Best to cleanup here instead of handling it in the caller [as noted by 
Vishal, we might not be doing that yet]. Using 
remove_memory_blocks_and_altmaps() on the fully processed range sounds 
reasonable.

maybe something like

ret = arch_add_memory(nid, cur_start, memblock_size, &params);
if (ret) {
	kfree(params.altmap);
	goto out;
}

ret = create_memory_block_devices(cur_start, memblock_size,
				   params.altmap, group);
if (ret) {
	arch_remove_memory(cur_start, memblock_size, NULL);
	kfree(params.altmap);
	goto out;
}

if (ret && cur_start != start)
	remove_memory_blocks_and_altmaps(start, cur_start - start);
return ret;

> +
> +		/* call arch's memory hotadd */
> +		ret = arch_add_memory(nid, cur_start, memblock_size, &params);
> +		if (ret < 0) {
> +			kfree(params.altmap);
> +			return ret;
> +		}
> +
> +		/* create memory block devices after memory was added */
> +		ret = create_memory_block_devices(cur_start, memblock_size,
> +						  params.altmap, group);
> +		if (ret) {
> +			arch_remove_memory(cur_start, memblock_size, NULL);
> +			kfree(params.altmap);
> +			return ret;
> +		}
> +	}
> +
> +	return 0;
> +}
> +

[...]

>   static int check_cpu_on_node(int nid)
>   {
>   	int cpu;
> @@ -2146,11 +2186,69 @@ void try_offline_node(int nid)
>   }
>   EXPORT_SYMBOL(try_offline_node);
>   
> -static int __ref try_remove_memory(u64 start, u64 size)
> +static void __ref remove_memory_blocks_and_altmaps(u64 start, u64 size)
>   {
> -	struct memory_block *mem;
> -	int rc = 0, nid = NUMA_NO_NODE;
> +	unsigned long memblock_size = memory_block_size_bytes();
>   	struct vmem_altmap *altmap = NULL;
> +	struct memory_block *mem;
> +	u64 cur_start;
> +	int rc;
> +
> +	/*
> +	 * For memmap_on_memory, the altmaps could have been added on
> +	 * a per-memblock basis. Loop through the entire range if so,
> +	 * and remove each memblock and its altmap.
> +	 */

/*
  * altmaps where added on a per-memblock basis; we have to process
  * each individual memory block.
  */

> +	for (cur_start = start; cur_start < start + size;
> +	     cur_start += memblock_size) {
> +		rc = walk_memory_blocks(cur_start, memblock_size, &mem,
> +					test_has_altmap_cb);
> +		if (rc) {
> +			altmap = mem->altmap;
> +			/*
> +			 * Mark altmap NULL so that we can add a debug
> +			 * check on memblock free.
> +			 */
> +			mem->altmap = NULL;
> +		}

Simpler (especially, we know that there must be an altmap):

mem = find_memory_block(pfn_to_section_nr(cur_start));
altmap = mem->altmap;
mem->altmap = NULL;

I think we might be able to remove test_has_altmap_cb() then.

> +
> +		remove_memory_block_devices(cur_start, memblock_size);
> +
> +		arch_remove_memory(cur_start, memblock_size, altmap);
> +
> +		/* Verify that all vmemmap pages have actually been freed. */
> +		if (altmap) {

There must be an altmap, so this can be done unconditionally.

> +			WARN(altmap->alloc, "Altmap not fully unmapped");
> +			kfree(altmap);
> +		}
> +	}
> +}

Verma, Vishal L Oct. 31, 2023, 2:14 a.m. UTC | #3

On Mon, 2023-10-30 at 11:20 +0100, David Hildenbrand wrote:
> On 26.10.23 00:44, Vishal Verma wrote:
> > 
[..]

> > @@ -2146,11 +2186,69 @@ void try_offline_node(int nid)
> >   }
> >   EXPORT_SYMBOL(try_offline_node);
> >   
> > -static int __ref try_remove_memory(u64 start, u64 size)
> > +static void __ref remove_memory_blocks_and_altmaps(u64 start, u64 size)
> >   {
> > -       struct memory_block *mem;
> > -       int rc = 0, nid = NUMA_NO_NODE;
> > +       unsigned long memblock_size = memory_block_size_bytes();
> >         struct vmem_altmap *altmap = NULL;
> > +       struct memory_block *mem;
> > +       u64 cur_start;
> > +       int rc;
> > +
> > +       /*
> > +        * For memmap_on_memory, the altmaps could have been added on
> > +        * a per-memblock basis. Loop through the entire range if so,
> > +        * and remove each memblock and its altmap.
> > +        */
> 
> /*
>   * altmaps where added on a per-memblock basis; we have to process
>   * each individual memory block.
>   */
> 
> > +       for (cur_start = start; cur_start < start + size;
> > +            cur_start += memblock_size) {
> > +               rc = walk_memory_blocks(cur_start, memblock_size, &mem,
> > +                                       test_has_altmap_cb);
> > +               if (rc) {
> > +                       altmap = mem->altmap;
> > +                       /*
> > +                        * Mark altmap NULL so that we can add a debug
> > +                        * check on memblock free.
> > +                        */
> > +                       mem->altmap = NULL;
> > +               }
> 
> Simpler (especially, we know that there must be an altmap):
> 
> mem = find_memory_block(pfn_to_section_nr(cur_start));
> altmap = mem->altmap;
> mem->altmap = NULL;
> 
> I think we might be able to remove test_has_altmap_cb() then.
> 
> > +
> > +               remove_memory_block_devices(cur_start, memblock_size);
> > +
> > +               arch_remove_memory(cur_start, memblock_size, altmap);
> > +
> > +               /* Verify that all vmemmap pages have actually been freed. */
> > +               if (altmap) {
> 
> There must be an altmap, so this can be done unconditionally.

Hi David,

All other comments make sense, making those changes now.

However for this one, does the WARN() below go away then?

I was wondering if maybe arch_remove_memory() is responsible for
freeing the altmap here, and at this stage we're just checking if that
happened. If it didn't WARN and then free it.

I drilled down the path, and I don't see altmap actually getting freed
in vmem_altmap_free(), but I wasn't sure if <something else> was meant
to free it as altmap->alloc went down to 0.

> 
> > +                       WARN(altmap->alloc, "Altmap not fully unmapped");
> > +                       kfree(altmap);
> > +               }
> > +       }
> > +}
> 
>

David Hildenbrand Oct. 31, 2023, 10:13 a.m. UTC | #4

On 31.10.23 03:14, Verma, Vishal L wrote:
> On Mon, 2023-10-30 at 11:20 +0100, David Hildenbrand wrote:
>> On 26.10.23 00:44, Vishal Verma wrote:
>>>
> [..]
> 
>>> @@ -2146,11 +2186,69 @@ void try_offline_node(int nid)
>>>    }
>>>    EXPORT_SYMBOL(try_offline_node);
>>>    
>>> -static int __ref try_remove_memory(u64 start, u64 size)
>>> +static void __ref remove_memory_blocks_and_altmaps(u64 start, u64 size)
>>>    {
>>> -       struct memory_block *mem;
>>> -       int rc = 0, nid = NUMA_NO_NODE;
>>> +       unsigned long memblock_size = memory_block_size_bytes();
>>>          struct vmem_altmap *altmap = NULL;
>>> +       struct memory_block *mem;
>>> +       u64 cur_start;
>>> +       int rc;
>>> +
>>> +       /*
>>> +        * For memmap_on_memory, the altmaps could have been added on
>>> +        * a per-memblock basis. Loop through the entire range if so,
>>> +        * and remove each memblock and its altmap.
>>> +        */
>>
>> /*
>>    * altmaps where added on a per-memblock basis; we have to process
>>    * each individual memory block.
>>    */
>>
>>> +       for (cur_start = start; cur_start < start + size;
>>> +            cur_start += memblock_size) {
>>> +               rc = walk_memory_blocks(cur_start, memblock_size, &mem,
>>> +                                       test_has_altmap_cb);
>>> +               if (rc) {
>>> +                       altmap = mem->altmap;
>>> +                       /*
>>> +                        * Mark altmap NULL so that we can add a debug
>>> +                        * check on memblock free.
>>> +                        */
>>> +                       mem->altmap = NULL;
>>> +               }
>>
>> Simpler (especially, we know that there must be an altmap):
>>
>> mem = find_memory_block(pfn_to_section_nr(cur_start));
>> altmap = mem->altmap;
>> mem->altmap = NULL;
>>
>> I think we might be able to remove test_has_altmap_cb() then.
>>
>>> +
>>> +               remove_memory_block_devices(cur_start, memblock_size);
>>> +
>>> +               arch_remove_memory(cur_start, memblock_size, altmap);
>>> +
>>> +               /* Verify that all vmemmap pages have actually been freed. */
>>> +               if (altmap) {
>>
>> There must be an altmap, so this can be done unconditionally.
> 
> Hi David,

Hi!

> 
> All other comments make sense, making those changes now.
> 
> However for this one, does the WARN() below go away then?
> 
> I was wondering if maybe arch_remove_memory() is responsible for
> freeing the altmap here, and at this stage we're just checking if that
> happened. If it didn't WARN and then free it.

I think that has to stay, to make sure arch_remove_memory() did the 
right thing and we don't -- by BUG -- still have some altmap pages in 
use after they should have been completely freed.

> 
> I drilled down the path, and I don't see altmap actually getting freed
> in vmem_altmap_free(), but I wasn't sure if <something else> was meant
> to free it as altmap->alloc went down to 0.


vmem_altmap_free() does the "altmap->alloc -= nr_pfns", which is called 
when arch_remove_memory() frees the vmemmap pages and detects that they 
actually come from the altmap reserve and not from the buddy/earlyboot 
allocator.

Freeing an altmap is just unaccounting it in the altmap structure; and 
here we make sure that we are actually back down to 0 and don't have 
some weird altmap freeing BUG in arch_remove_memory().

Patch

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 6be7de9efa55..b97035193090 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1380,6 +1380,48 @@  static bool mhp_supports_memmap_on_memory(unsigned long size)
 	return arch_supports_memmap_on_memory(vmemmap_size);
 }
 
+static int create_altmaps_and_memory_blocks(int nid, struct memory_group *group,
+					    u64 start, u64 size)
+{
+	unsigned long memblock_size = memory_block_size_bytes();
+	u64 cur_start;
+	int ret;
+
+	for (cur_start = start; cur_start < start + size;
+	     cur_start += memblock_size) {
+		struct mhp_params params = { .pgprot =
+						     pgprot_mhp(PAGE_KERNEL) };
+		struct vmem_altmap mhp_altmap = {
+			.base_pfn = PHYS_PFN(cur_start),
+			.end_pfn = PHYS_PFN(cur_start + memblock_size - 1),
+		};
+
+		mhp_altmap.free = memory_block_memmap_on_memory_pages();
+		params.altmap = kmemdup(&mhp_altmap, sizeof(struct vmem_altmap),
+					GFP_KERNEL);
+		if (!params.altmap)
+			return -ENOMEM;
+
+		/* call arch's memory hotadd */
+		ret = arch_add_memory(nid, cur_start, memblock_size, &params);
+		if (ret < 0) {
+			kfree(params.altmap);
+			return ret;
+		}
+
+		/* create memory block devices after memory was added */
+		ret = create_memory_block_devices(cur_start, memblock_size,
+						  params.altmap, group);
+		if (ret) {
+			arch_remove_memory(cur_start, memblock_size, NULL);
+			kfree(params.altmap);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
 /*
  * NOTE: The caller must call lock_device_hotplug() to serialize hotplug
  * and online/offline operations (triggered e.g. by sysfs).
@@ -1390,10 +1432,6 @@  int __ref add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
 {
 	struct mhp_params params = { .pgprot = pgprot_mhp(PAGE_KERNEL) };
 	enum memblock_flags memblock_flags = MEMBLOCK_NONE;
-	struct vmem_altmap mhp_altmap = {
-		.base_pfn =  PHYS_PFN(res->start),
-		.end_pfn  =  PHYS_PFN(res->end),
-	};
 	struct memory_group *group = NULL;
 	u64 start, size;
 	bool new_node = false;
@@ -1436,28 +1474,22 @@  int __ref add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
 	/*
 	 * Self hosted memmap array
 	 */
-	if (mhp_flags & MHP_MEMMAP_ON_MEMORY) {
-		if (mhp_supports_memmap_on_memory(size)) {
-			mhp_altmap.free = memory_block_memmap_on_memory_pages();
-			params.altmap = kmemdup(&mhp_altmap,
-						sizeof(struct vmem_altmap),
-						GFP_KERNEL);
-			if (!params.altmap)
-				goto error;
+	if ((mhp_flags & MHP_MEMMAP_ON_MEMORY) &&
+	    mhp_supports_memmap_on_memory(memory_block_size_bytes())) {
+		ret = create_altmaps_and_memory_blocks(nid, group, start, size);
+		if (ret)
+			goto error;
+	} else {
+		ret = arch_add_memory(nid, start, size, &params);
+		if (ret < 0)
+			goto error;
+
+		/* create memory block devices after memory was added */
+		ret = create_memory_block_devices(start, size, NULL, group);
+		if (ret) {
+			arch_remove_memory(start, size, NULL);
+			goto error;
 		}
-		/* fallback to not using altmap  */
-	}
-
-	/* call arch's memory hotadd */
-	ret = arch_add_memory(nid, start, size, &params);
-	if (ret < 0)
-		goto error_free;
-
-	/* create memory block devices after memory was added */
-	ret = create_memory_block_devices(start, size, params.altmap, group);
-	if (ret) {
-		arch_remove_memory(start, size, NULL);
-		goto error_free;
 	}
 
 	if (new_node) {
@@ -1494,8 +1526,6 @@  int __ref add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
 		walk_memory_blocks(start, size, NULL, online_memory_block);
 
 	return ret;
-error_free:
-	kfree(params.altmap);
 error:
 	if (IS_ENABLED(CONFIG_ARCH_KEEP_MEMBLOCK))
 		memblock_remove(start, size);
@@ -2076,6 +2106,16 @@  static int test_has_altmap_cb(struct memory_block *mem, void *arg)
 	return 0;
 }
 
+static int count_memory_range_altmaps_cb(struct memory_block *mem, void *arg)
+{
+	u64 *num_altmaps = (u64 *)arg;
+
+	if (mem->altmap)
+		*num_altmaps += 1;
+
+	return 0;
+}
+
 static int check_cpu_on_node(int nid)
 {
 	int cpu;
@@ -2146,11 +2186,69 @@  void try_offline_node(int nid)
 }
 EXPORT_SYMBOL(try_offline_node);
 
-static int __ref try_remove_memory(u64 start, u64 size)
+static void __ref remove_memory_blocks_and_altmaps(u64 start, u64 size)
 {
-	struct memory_block *mem;
-	int rc = 0, nid = NUMA_NO_NODE;
+	unsigned long memblock_size = memory_block_size_bytes();
 	struct vmem_altmap *altmap = NULL;
+	struct memory_block *mem;
+	u64 cur_start;
+	int rc;
+
+	/*
+	 * For memmap_on_memory, the altmaps could have been added on
+	 * a per-memblock basis. Loop through the entire range if so,
+	 * and remove each memblock and its altmap.
+	 */
+	for (cur_start = start; cur_start < start + size;
+	     cur_start += memblock_size) {
+		rc = walk_memory_blocks(cur_start, memblock_size, &mem,
+					test_has_altmap_cb);
+		if (rc) {
+			altmap = mem->altmap;
+			/*
+			 * Mark altmap NULL so that we can add a debug
+			 * check on memblock free.
+			 */
+			mem->altmap = NULL;
+		}
+
+		remove_memory_block_devices(cur_start, memblock_size);
+
+		arch_remove_memory(cur_start, memblock_size, altmap);
+
+		/* Verify that all vmemmap pages have actually been freed. */
+		if (altmap) {
+			WARN(altmap->alloc, "Altmap not fully unmapped");
+			kfree(altmap);
+		}
+	}
+}
+
+static int memory_blocks_have_altmaps(u64 start, u64 size)
+{
+	u64 num_memblocks = size / memory_block_size_bytes();
+	u64 num_altmaps = 0;
+
+	if (!mhp_memmap_on_memory())
+		return 0;
+
+	walk_memory_blocks(start, size, &num_altmaps,
+			   count_memory_range_altmaps_cb);
+
+	if (num_altmaps == 0)
+		return 0;
+
+	if (num_memblocks != num_altmaps) {
+		WARN_ONCE(1, "Not all memblocks in range have altmaps");
+		return -EINVAL;
+	}
+
+	return 1;
+}
+
+static int __ref try_remove_memory(u64 start, u64 size)
+{
+	int rc, nid = NUMA_NO_NODE;
 
 	BUG_ON(check_hotplug_memory_range(start, size));
 
@@ -2167,45 +2265,25 @@  static int __ref try_remove_memory(u64 start, u64 size)
 	if (rc)
 		return rc;
 
-	/*
-	 * We only support removing memory added with MHP_MEMMAP_ON_MEMORY in
-	 * the same granularity it was added - a single memory block.
-	 */
-	if (mhp_memmap_on_memory()) {
-		rc = walk_memory_blocks(start, size, &mem, test_has_altmap_cb);
-		if (rc) {
-			if (size != memory_block_size_bytes()) {
-				pr_warn("Refuse to remove %#llx - %#llx,"
-					"wrong granularity\n",
-					start, start + size);
-				return -EINVAL;
-			}
-			altmap = mem->altmap;
-			/*
-			 * Mark altmap NULL so that we can add a debug
-			 * check on memblock free.
-			 */
-			mem->altmap = NULL;
-		}
-	}
-
 	/* remove memmap entry */
 	firmware_map_remove(start, start + size, "System RAM");
 
-	/*
-	 * Memory block device removal under the device_hotplug_lock is
-	 * a barrier against racing online attempts.
-	 */
-	remove_memory_block_devices(start, size);
-
 	mem_hotplug_begin();
 
-	arch_remove_memory(start, size, altmap);
-
-	/* Verify that all vmemmap pages have actually been freed. */
-	if (altmap) {
-		WARN(altmap->alloc, "Altmap not fully unmapped");
-		kfree(altmap);
+	rc = memory_blocks_have_altmaps(start, size);
+	if (rc < 0) {
+		goto err;
+	} else if (rc == 0) {
+		/*
+		 * Memory block device removal under the device_hotplug_lock is
+		 * a barrier against racing online attempts.
+		 * No altmaps present, do the removal directly
+		 */
+		remove_memory_block_devices(start, size);
+		arch_remove_memory(start, size, NULL);
+	} else {
+		/* all memblocks in the range have altmaps */
+		remove_memory_blocks_and_altmaps(start, size);
 	}
 
 	if (IS_ENABLED(CONFIG_ARCH_KEEP_MEMBLOCK)) {
@@ -2218,8 +2296,9 @@  static int __ref try_remove_memory(u64 start, u64 size)
 	if (nid != NUMA_NO_NODE)
 		try_offline_node(nid);
 
+err:
 	mem_hotplug_done();
-	return 0;
+	return (rc < 0 ? rc : 0);
 }
 
 /**