1. Patchew
  2. linux
  3. [PATCH v2 0/2] SEV-SNP Unaccepted Memory Hotplug
  4. View patch

[PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add

Pratik R. Sampat posted 2 patches 3 weeks, 5 days ago
There is a newer version of this series
[PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by Pratik R. Sampat 3 weeks, 5 days ago 
Confidential computing guests require memory to be accepted before use.
The unaccepted memory bitmap maintained by firmware does not track
hotplugged memory ranges.

Call arch_accept_memory() during the hot-add path to explicitly validate
and transition the newly added memory to a private state, making it
usable by the guest.

Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
---
 mm/memory_hotplug.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index a63ec679d861..8cfbf0541430 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -38,6 +38,7 @@
 #include <linux/node.h>
 
 #include <asm/tlbflush.h>
+#include <asm/unaccepted_memory.h>
 
 #include "internal.h"
 #include "shuffle.h"
@@ -1567,6 +1568,9 @@ int add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
 	if (!strcmp(res->name, "System RAM"))
 		firmware_map_add_hotplug(start, start + size, "System RAM");
 
+	if (IS_ENABLED(CONFIG_UNACCEPTED_MEMORY))
+		arch_accept_memory(start, start + size);
+
 	/* device_online() will take the lock when calling online_pages() */
 	mem_hotplug_done();
 
-- 
2.52.0
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by David Hildenbrand (Red Hat) 3 weeks, 4 days ago 
On 1/12/26 21:22, Pratik R. Sampat wrote:
> Confidential computing guests require memory to be accepted before use.
> The unaccepted memory bitmap maintained by firmware does not track
> hotplugged memory ranges.
> 
> Call arch_accept_memory() during the hot-add path to explicitly validate
> and transition the newly added memory to a private state, making it
> usable by the guest.
> 
> Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
> ---
>   mm/memory_hotplug.c | 4 ++++
>   1 file changed, 4 insertions(+)
> 
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index a63ec679d861..8cfbf0541430 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -38,6 +38,7 @@
>   #include <linux/node.h>
>   
>   #include <asm/tlbflush.h>
> +#include <asm/unaccepted_memory.h>
>   
>   #include "internal.h"
>   #include "shuffle.h"
> @@ -1567,6 +1568,9 @@ int add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
>   	if (!strcmp(res->name, "System RAM"))
>   		firmware_map_add_hotplug(start, start + size, "System RAM");
>   
> +	if (IS_ENABLED(CONFIG_UNACCEPTED_MEMORY))
> +		arch_accept_memory(start, start + size);
> +
>   	/* device_online() will take the lock when calling online_pages() */
>   	mem_hotplug_done();
>   

As discussed, for things like virtio-mem or the HV-balloon this might be 
the wrong thing to do, but I don't expect these mechanisms to be used in 
CoCo environments just yet (and doing so would require enabling work for 
them).

So I'm fine with this for now.

-- 
Cheers

David
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by Pratik R. Sampat 3 weeks, 3 days ago 

On 1/14/26 4:30 AM, David Hildenbrand (Red Hat) wrote:
> On 1/12/26 21:22, Pratik R. Sampat wrote:
>> Confidential computing guests require memory to be accepted before use.
>> The unaccepted memory bitmap maintained by firmware does not track
>> hotplugged memory ranges.
>>
>> Call arch_accept_memory() during the hot-add path to explicitly validate
>> and transition the newly added memory to a private state, making it
>> usable by the guest.
>>
>> Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
>> ---
>>   mm/memory_hotplug.c | 4 ++++
>>   1 file changed, 4 insertions(+)
>>
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index a63ec679d861..8cfbf0541430 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -38,6 +38,7 @@
>>   #include <linux/node.h>
>>     #include <asm/tlbflush.h>
>> +#include <asm/unaccepted_memory.h>
>>     #include "internal.h"
>>   #include "shuffle.h"
>> @@ -1567,6 +1568,9 @@ int add_memory_resource(int nid, struct resource *res, mhp_t mhp_flags)
>>       if (!strcmp(res->name, "System RAM"))
>>           firmware_map_add_hotplug(start, start + size, "System RAM");
>>   +    if (IS_ENABLED(CONFIG_UNACCEPTED_MEMORY))
>> +        arch_accept_memory(start, start + size);
>> +
>>       /* device_online() will take the lock when calling online_pages() */
>>       mem_hotplug_done();
>>   
> 
> As discussed, for things like virtio-mem or the HV-balloon this might be the wrong thing to do, but I don't expect these mechanisms to be used in CoCo environments just yet (and doing so would require enabling work for them).
> 

Ack.

> So I'm fine with this for now.
> 

Thanks!
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by kernel test robot 3 weeks, 5 days ago 
Hi Pratik,

kernel test robot noticed the following build errors:

[auto build test ERROR on akpm-mm/mm-everything]
[also build test ERROR on tip/x86/core linus/master v6.19-rc5 next-20260109]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Pratik-R-Sampat/mm-memory_hotplug-Add-support-to-accept-memory-during-hot-add/20260113-042631
base:   https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link:    https://lore.kernel.org/r/20260112202300.43546-2-prsampat%40amd.com
patch subject: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
config: s390-defconfig (https://download.01.org/0day-ci/archive/20260113/202601131632.NrQzg2Wm-lkp@intel.com/config)
compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project 9b8addffa70cee5b2acc5454712d9cf78ce45710)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20260113/202601131632.NrQzg2Wm-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202601131632.NrQzg2Wm-lkp@intel.com/

All errors (new ones prefixed by >>):

>> mm/memory_hotplug.c:41:10: fatal error: 'asm/unaccepted_memory.h' file not found
      41 | #include <asm/unaccepted_memory.h>
         |          ^~~~~~~~~~~~~~~~~~~~~~~~~
   1 error generated.


vim +41 mm/memory_hotplug.c

    39	
    40	#include <asm/tlbflush.h>
  > 41	#include <asm/unaccepted_memory.h>
    42	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by kernel test robot 3 weeks, 5 days ago 
Hi Pratik,

kernel test robot noticed the following build errors:

[auto build test ERROR on akpm-mm/mm-everything]
[also build test ERROR on tip/x86/core linus/master v6.19-rc5 next-20260109]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Pratik-R-Sampat/mm-memory_hotplug-Add-support-to-accept-memory-during-hot-add/20260113-042631
base:   https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link:    https://lore.kernel.org/r/20260112202300.43546-2-prsampat%40amd.com
patch subject: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
config: s390-randconfig-001-20260113 (https://download.01.org/0day-ci/archive/20260113/202601131156.Kfi0QLIm-lkp@intel.com/config)
compiler: s390-linux-gcc (GCC) 8.5.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20260113/202601131156.Kfi0QLIm-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202601131156.Kfi0QLIm-lkp@intel.com/

All errors (new ones prefixed by >>):

>> mm/memory_hotplug.c:41:10: fatal error: asm/unaccepted_memory.h: No such file or directory
    #include <asm/unaccepted_memory.h>
             ^~~~~~~~~~~~~~~~~~~~~~~~~
   compilation terminated.


vim +41 mm/memory_hotplug.c

    39	
    40	#include <asm/tlbflush.h>
  > 41	#include <asm/unaccepted_memory.h>
    42	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by Andrew Morton 3 weeks, 5 days ago 
On Mon, 12 Jan 2026 14:22:59 -0600 "Pratik R. Sampat" <prsampat@amd.com> wrote:

> Confidential computing guests require memory to be accepted before use.
> The unaccepted memory bitmap maintained by firmware does not track
> hotplugged memory ranges.
> 
> Call arch_accept_memory() during the hot-add path to explicitly validate
> and transition the newly added memory to a private state, making it
> usable by the guest.
> 
> Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
> ---
>  mm/memory_hotplug.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index a63ec679d861..8cfbf0541430 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -38,6 +38,7 @@
>  #include <linux/node.h>
>  
>  #include <asm/tlbflush.h>
> +#include <asm/unaccepted_memory.h>

This only exists for x86!

Otherwise, the mm/ changes are minimal so I volunteer this patchset
for the x86 tree ;)
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by Pratik R. Sampat 3 weeks, 5 days ago 

On 1/12/26 3:04 PM, Andrew Morton wrote:
> On Mon, 12 Jan 2026 14:22:59 -0600 "Pratik R. Sampat" <prsampat@amd.com> wrote:
> 
>> Confidential computing guests require memory to be accepted before use.
>> The unaccepted memory bitmap maintained by firmware does not track
>> hotplugged memory ranges.
>>
>> Call arch_accept_memory() during the hot-add path to explicitly validate
>> and transition the newly added memory to a private state, making it
>> usable by the guest.
>>
>> Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
>> ---
>>  mm/memory_hotplug.c | 4 ++++
>>  1 file changed, 4 insertions(+)
>>
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index a63ec679d861..8cfbf0541430 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -38,6 +38,7 @@
>>  #include <linux/node.h>
>>  
>>  #include <asm/tlbflush.h>
>> +#include <asm/unaccepted_memory.h>
> 
> This only exists for x86!

Ah, I missed that entirely. Thanks for catching that.

Probably not the best option to have a generic unaccepted_memory.h as well.
Maybe, I should have arch_[un]accept_memory() definitions within mm.h wrapped
within CONFIG_UNACCEPTED_MEMORY instead so that its cleaner.

> 
> Otherwise, the mm/ changes are minimal so I volunteer this patchset
> for the x86 tree ;)

Ack!

--Pratik
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by Andrew Morton 3 weeks, 5 days ago 
On Mon, 12 Jan 2026 16:23:37 -0600 "Pratik R. Sampat" <prsampat@amd.com> wrote:

> 
> 
> On 1/12/26 3:04 PM, Andrew Morton wrote:
> > On Mon, 12 Jan 2026 14:22:59 -0600 "Pratik R. Sampat" <prsampat@amd.com> wrote:
> > 
> >> Confidential computing guests require memory to be accepted before use.
> >> The unaccepted memory bitmap maintained by firmware does not track
> >> hotplugged memory ranges.
> >>
> >> Call arch_accept_memory() during the hot-add path to explicitly validate
> >> and transition the newly added memory to a private state, making it
> >> usable by the guest.
> >>
> >> Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
> >> ---
> >>  mm/memory_hotplug.c | 4 ++++
> >>  1 file changed, 4 insertions(+)
> >>
> >> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> >> index a63ec679d861..8cfbf0541430 100644
> >> --- a/mm/memory_hotplug.c
> >> +++ b/mm/memory_hotplug.c
> >> @@ -38,6 +38,7 @@
> >>  #include <linux/node.h>
> >>  
> >>  #include <asm/tlbflush.h>
> >> +#include <asm/unaccepted_memory.h>
> > 
> > This only exists for x86!
> 
> Ah, I missed that entirely. Thanks for catching that.
> 
> Probably not the best option to have a generic unaccepted_memory.h as well.
> Maybe, I should have arch_[un]accept_memory() definitions within mm.h wrapped
> within CONFIG_UNACCEPTED_MEMORY instead so that its cleaner.

Something like that.

The idiomatic Linus way is to use

#ifndef arch_accept_memory
#define arch_accept_memory ...
#endif

Lots of prior art here:

	grep -r include/linux "ifndef arch_"


Oh, arch_get_idle_state_flags() got it all wrong.

	#ifdef CONFIG_ACPI_PROCESSOR_IDLE
	#ifndef arch_get_idle_state_flags
	static inline unsigned int arch_get_idle_state_flags(u32 arch_flags)
	{
		return 0;
	}
	#endif
	#endif /* CONFIG_ACPI_PROCESSOR_IDLE */

- shouldn't have needed "ifdef CONFIG_ACPI_PROCESSOR_IDLE"

- should have appended

	#define arch_get_idle_state_flags arch_get_idle_state_flags

  in case cpp hit the same lines a second time.
Re: [PATCH v2 1/2] mm/memory_hotplug: Add support to accept memory during hot-add
Posted by Pratik R. Sampat 3 weeks, 5 days ago 

On 1/12/26 4:43 PM, Andrew Morton wrote:
> On Mon, 12 Jan 2026 16:23:37 -0600 "Pratik R. Sampat" <prsampat@amd.com> wrote:
> 
>>
>>
>> On 1/12/26 3:04 PM, Andrew Morton wrote:
>>> On Mon, 12 Jan 2026 14:22:59 -0600 "Pratik R. Sampat" <prsampat@amd.com> wrote:
>>>
>>>> Confidential computing guests require memory to be accepted before use.
>>>> The unaccepted memory bitmap maintained by firmware does not track
>>>> hotplugged memory ranges.
>>>>
>>>> Call arch_accept_memory() during the hot-add path to explicitly validate
>>>> and transition the newly added memory to a private state, making it
>>>> usable by the guest.
>>>>
>>>> Signed-off-by: Pratik R. Sampat <prsampat@amd.com>
>>>> ---
>>>>  mm/memory_hotplug.c | 4 ++++
>>>>  1 file changed, 4 insertions(+)
>>>>
>>>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>>>> index a63ec679d861..8cfbf0541430 100644
>>>> --- a/mm/memory_hotplug.c
>>>> +++ b/mm/memory_hotplug.c
>>>> @@ -38,6 +38,7 @@
>>>>  #include <linux/node.h>
>>>>  
>>>>  #include <asm/tlbflush.h>
>>>> +#include <asm/unaccepted_memory.h>
>>>
>>> This only exists for x86!
>>
>> Ah, I missed that entirely. Thanks for catching that.
>>
>> Probably not the best option to have a generic unaccepted_memory.h as well.
>> Maybe, I should have arch_[un]accept_memory() definitions within mm.h wrapped
>> within CONFIG_UNACCEPTED_MEMORY instead so that its cleaner.
> 
> Something like that.
> 
> The idiomatic Linus way is to use
> 
> #ifndef arch_accept_memory
> #define arch_accept_memory ...
> #endif
> 
> Lots of prior art here:
> 
> 	grep -r include/linux "ifndef arch_"
> 
> 
> Oh, arch_get_idle_state_flags() got it all wrong.
> 
> 	#ifdef CONFIG_ACPI_PROCESSOR_IDLE
> 	#ifndef arch_get_idle_state_flags
> 	static inline unsigned int arch_get_idle_state_flags(u32 arch_flags)
> 	{
> 		return 0;
> 	}
> 	#endif
> 	#endif /* CONFIG_ACPI_PROCESSOR_IDLE */
> 
> - shouldn't have needed "ifdef CONFIG_ACPI_PROCESSOR_IDLE"
> 
> - should have appended
> 
> 	#define arch_get_idle_state_flags arch_get_idle_state_flags
> 
>   in case cpp hit the same lines a second time.
> 

Got it. Thanks for clearing that up. I'll make sure to do it this way
in the next iteration.

--Pratik

Patchew 2.1-devel-b67e4c2

© 2016 - 2026 Red Hat, Inc.