Persistent Memory type vs Reserved Memory type
A message that is not in the archive (sent off-list or not yet received): [email protected]
Hi:
I am trying to debug an issue related to my private built kernel's (4.11 from upstream with ACPI/NFIT/NVDIMM enabled) interaction with a persistent memory device's type.
We have a way to emulate a software based NVDIMM device into a guest, which is running my private built kernel. When the host set this emulated NVDIMM device as Reserved type, I was able to see a pmem0 block device exposed and everything seems to function correctly.however, when it's exposed as a Persistent Memory type, the kernel does not boot and it has a black screen and no usefully logging could help root cause this. My questions are:
1. Is this a known issue for 4.11? 2. Is there any patches NIFT/NVDIMM beyond 4.11 that's needed to get Persistent Memory type supported? 3. In this case, is there anyway to debug this further?
ThanksCheng-mean
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm
Quoted 5 lines · Soccer Liu
Hi:
I am trying to debug an issue related to my private built kernel's (4.11 from upstream with ACPI/NFIT/NVDIMM enabled) interaction with a persistent memory device's type.
We have a way to emulate a software based NVDIMM device into a guest, which is running my private built kernel. When the host set this emulated NVDIMM device as Reserved type, I was able to see a pmem0 block device exposed and everything seems to function correctly.however, when it's exposed as a Persistent Memory type, the kernel does not boot and it has a black screen and no usefully logging could help root cause this.
My questions are:
1. Is this a known issue for 4.11? 2. Is there any patches NIFT/NVDIMM beyond 4.11 that's needed to get Persistent Memory type supported? 3. In this case, is there anyway to debug this further?
When you say Persistent Memory Type, which type are you referring. The spec compliant one is E820-type-7 or EFI- type-15. Linux also understands E820-type-12 which went upstream before the official types were defined.
E820-Type-12 is the problematic one, the others are equivalent to Reserved (E820-Type-2).
All that said we have had functional support for all these types since the 4.2 kernel, no kernel changes should be needed.
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm
A message that is not in the archive (sent off-list or not yet received): CAPcyv4juhhxZN9b1yVnFjboJUNMXJzLL8k5yf-+GXNE+aj86kw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org
I am sorry for not being more clear.
In ACPI spec 2.6, the Table 15-317 has all the UEFI Memory Types and mapping to ACPI address range types.
Here are the actual types that I referred to:
“Persistent”-> EfiPersistentMemory (14)“Reserved” -> Reserved type 15
Apparently, only my private kernel works as expected when dealing with Reserved type (15) but cannot boot with any EfiPersistentMemory device. I did see a EfiPersistentMemory device working with Ubuntu OS. I am suspecting it's my environment issues (kconfig flags or others), I am curious about anyone has any idea on how to debug this kind of "does not boot issue" or anyone has seen EfiPersistentMemory like this?
ThanksCheng-mean
On Thursday, July 20, 2017 6:03 AM, Dan Williams <[email protected]> wrote:
Quoted 5 lines · Soccer Liu
Hi:
I am trying to debug an issue related to my private built kernel's (4.11 from upstream with ACPI/NFIT/NVDIMM enabled) interaction with a persistent memory device's type.
We have a way to emulate a software based NVDIMM device into a guest, which is running my private built kernel. When the host set this emulated NVDIMM device as Reserved type, I was able to see a pmem0 block device exposed and everything seems to function correctly.however, when it's exposed as a Persistent Memory type, the kernel does not boot and it has a black screen and no usefully logging could help root cause this. My questions are:
1. Is this a known issue for 4.11? 2. Is there any patches NIFT/NVDIMM beyond 4.11 that's needed to get Persistent Memory type supported? 3. In this case, is there anyway to debug this further?
When you say Persistent Memory Type, which type are you referring. The spec compliant one is E820-type-7 or EFI- type-15. Linux also understands E820-type-12 which went upstream before the official types were defined.
E820-Type-12 is the problematic one, the others are equivalent to Reserved (E820-Type-2).
All that said we have had functional support for all these types since the 4.2 kernel, no kernel changes should be needed.
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm
A message that is not in the archive (sent off-list or not yet received): 2095070705.1732096.1500573379101-sAHhhX/[email protected]
Slightly correction:
“Persistent”-> EfiPersistentMemory (14)
“Reserved” -> EfiReservedMemoryType(0)
On Thursday, July 20, 2017 10:56 AM, Soccer Liu <[email protected]> wrote:
I am sorry for not being more clear.
In ACPI spec 2.6, the Table 15-317 has all the UEFI Memory Types and mapping to ACPI address range types.
Here are the actual types that I referred to:
“Persistent”-> EfiPersistentMemory (14)“Reserved” -> Reserved type 15
Apparently, only my private kernel works as expected when dealing with Reserved type (15) but cannot boot with any EfiPersistentMemory device. I did see a EfiPersistentMemory device working with Ubuntu OS. I am suspecting it's my environment issues (kconfig flags or others), I am curious about anyone has any idea on how to debug this kind of "does not boot issue" or anyone has seen EfiPersistentMemory like this?
ThanksCheng-mean
On Thursday, July 20, 2017 6:03 AM, Dan Williams <[email protected]> wrote:
Quoted 5 lines · Soccer Liu
Hi:
I am trying to debug an issue related to my private built kernel's (4.11 from upstream with ACPI/NFIT/NVDIMM enabled) interaction with a persistent memory device's type.
We have a way to emulate a software based NVDIMM device into a guest, which is running my private built kernel. When the host set this emulated NVDIMM device as Reserved type, I was able to see a pmem0 block device exposed and everything seems to function correctly.however, when it's exposed as a Persistent Memory type, the kernel does not boot and it has a black screen and no usefully logging could help root cause this. My questions are:
1. Is this a known issue for 4.11? 2. Is there any patches NIFT/NVDIMM beyond 4.11 that's needed to get Persistent Memory type supported? 3. In this case, is there anyway to debug this further?
When you say Persistent Memory Type, which type are you referring. The spec compliant one is E820-type-7 or EFI- type-15. Linux also understands E820-type-12 which went upstream before the official types were defined.
E820-Type-12 is the problematic one, the others are equivalent to Reserved (E820-Type-2).
All that said we have had functional support for all these types since the 4.2 kernel, no kernel changes should be needed.
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm
Quoted 5 lines · Soccer Liu
I am sorry for not being more clear.
In ACPI spec 2.6, the Table 15-317 has all the UEFI Memory Types and mapping to ACPI address range types.
Here are the actual types that I referred to:
“Persistent”-> EfiPersistentMemory (14)“Reserved” -> Reserved type
Quoted 8 lines
15
Apparently, only my private kernel works as expected when dealing with Reserved type (15) but cannot boot with any EfiPersistentMemory device. I did see a EfiPersistentMemory device working with Ubuntu OS. I am suspecting it's my environment issues (kconfig flags or others), I am curious about anyone has any idea on how to debug this kind of "does not boot issue" or anyone has seen EfiPersistentMemory like this?
I'd recommend you enable serial console on your system, and add the following boot options. Note, you would need to change the parameter of console= option according to your setup. This should help you to identify where the problem is.
console=ttyS1,115200 console=tty1,115200 libnvdimm.dyndbg nfit.dyndbg nd_pmem.dyndbg nd_btt.dyndbg efi=debug ignore_loglevel
-Toshi
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm