nvme: make providing NGUID as UUID usage less scary
PATCH 1/1 nvme: make providing NGUID as UUID usage less scary
Hi,
Quoted 5 lines
From: Keith Busch <[email protected]>
The warning is a bit alarming, and it only prints for the very first non-sgl capable device that receives a passthrough command. Just log an informational message on initial discovery for every device.
Since we decide to show this warning at device initial discovery stage, here are some similar warnings that I think are essentially the same.
grep -C3 -rn "dev_warn_once" ./drivers/nvme/*
```
./drivers/nvme/host/sysfs.c-147- * we have no UUID set
./drivers/nvme/host/sysfs.c-148- */
./drivers/nvme/host/sysfs.c-149- if (uuid_is_null(&ids->uuid)) {
./drivers/nvme/host/sysfs.c:150: dev_warn_once(dev,
./drivers/nvme/host/sysfs.c-151- "No UUID available providing old NGUID\n");
./drivers/nvme/host/sysfs.c-152- return sysfs_emit(buf, "%pU\n", ids->nguid);
./drivers/nvme/host/sysfs.c-153- }
```This warning will only complain for the first partition of the first device also. And for NVMe devices, according to the NVM-Express-1_4 specification p.175 Figure 251:
"Bit 9 (UUID List): ...", UUID is not mandatory i guess. So let's degrade it into a informational message on initial discovery for every device.
---
drivers/nvme/host/core.c +4 -0
@@ -4296,6 +4296,10 @@ static void nvme_validate_ns(struct nvme_ns *ns, struct nvme_ns_info *info) goto out; } +if (uuid_is_null (&info->ids.uuid)) { +dev_info (ns->ctrl->device,"No UUID available, uuid_show providing old NGUID\n" ); + } + ret =nvme_update_ns_info (ns, info); out:/*
drivers/nvme/host/sysfs.c +0 -2
@@ -147,8 +147,6 @@ static ssize_t uuid_show(struct device *dev, struct device_attribute *attr,* we have no UUID set */ if (uuid_is_null (&ids->uuid)) { -dev_warn_once (dev, -"No UUID available providing old NGUID\n" );return sysfs_emit (buf,"%pU\n" , ids->nguid); }return sysfs_emit (buf,"%pU\n" , &ids->uuid);
--
2.54.0
PATCH v2 1/1 nvme: make providing NGUID as UUID usage less scary
Since we decide to show the prp passthrough usage warning at device initial discovery stage and degrade it to a informational message, here is a warning that I think are essentially the same.
grep -C3 -rn "dev_warn_once" ./drivers/nvme/*
```
./drivers/nvme/host/sysfs.c-147- * we have no UUID set
./drivers/nvme/host/sysfs.c-148- */
./drivers/nvme/host/sysfs.c-149- if (uuid_is_null(&ids->uuid)) {
./drivers/nvme/host/sysfs.c:150: dev_warn_once(dev,
./drivers/nvme/host/sysfs.c-151- "No UUID available providing old NGUID\n");
./drivers/nvme/host/sysfs.c-152- return sysfs_emit(buf, "%pU\n", ids->nguid);
./drivers/nvme/host/sysfs.c-153- }
```This warning will only complain for the first partition of the first device also. And for NVMe devices, according to the NVM-Express-1_4 specification p.175 Figure 251:
"Bit 9 (UUID List): ...", UUID is not mandatory. So let's degrade it into a informational message on initial discovery for every device.
---
v2: move the hint to nvme_init_ns_head instead of nvme_validate_ns
v2: clean the commit message lines
In my understanding, the nvme_init_ns_head will be called each time a namespace created, and it does some checking on ids so we also place the hint there, and here is the result of dmesg applied current patch.
[ 4.502052] nvme nvme1: pci function 0000:05:00.0 [ 4.502060] nvme nvme0: pci function 0000:01:00.0 [ 4.508544] nvme nvme1: missing or invalid SUBNQN field. [ 4.508664] nvme nvme1: D3 entry latency set to 8 seconds [ 4.512422] nvme nvme0: missing or invalid SUBNQN field. [ 4.517581] nvme nvme0: allocated 32 MiB host memory buffer (1 segment). [ 4.523519] nvme nvme0: 8/0/0 default/read/poll queues [ 4.524761] nvme nvme1: 16/0/0 default/read/poll queues [ 4.525438] nvme nvme0: No UUID available, uuid_show providing old NGUID [ 4.526683] nvme nvme1: No UUID available, uuid_show providing old NGUID [ 4.527243] nvme0n1: p1 p2 p3 p4 p5 p6 [ 4.528351] nvme1n1: p1
drivers/nvme/host/core.c +4 -0
@@ -4007,6 +4007,10 @@ static int nvme_init_ns_head(struct nvme_ns *ns, struct nvme_ns_info *info) ctrl->quirks |=NVME_QUIRK_BOGUS_NID ; } +if (uuid_is_null (&info->ids.uuid)) { +dev_info (ctrl->device,"No UUID available, uuid_show providing old NGUID\n" ); + } +mutex_lock (&ctrl->subsys->lock); head =nvme_find_ns_head (ctrl, info->nsid);if (!head) {
drivers/nvme/host/sysfs.c +0 -2
@@ -147,8 +147,6 @@ static ssize_t uuid_show(struct device *dev, struct device_attribute *attr,* we have no UUID set */ if (uuid_is_null (&ids->uuid)) { -dev_warn_once (dev, -"No UUID available providing old NGUID\n" );return sysfs_emit (buf,"%pU\n" , ids->nguid); }return sysfs_emit (buf,"%pU\n" , &ids->uuid);
--
2.54.0
Sagi Grimberg
maintainerQuoted 48 lines · On 07/05/2026 19:36, AlanCui4080 wrote:
Hi,
Quoted 5 lines
From: Keith Busch <[email protected]>
The warning is a bit alarming, and it only prints for the very first non-sgl capable device that receives a passthrough command. Just log an informational message on initial discovery for every device.
Since we decide to show this warning at device initial discovery stage, here are some similar warnings that I think are essentially the same.
grep -C3 -rn "
dev_warn_once"./drivers/nvme/*``` ./drivers/nvme/host/sysfs.c-147- * we have no UUID set ./drivers/nvme/host/sysfs.c-148- */ ./drivers/nvme/host/sysfs.c-149- if (uuid_is_null(&ids->uuid)) { ./drivers/nvme/host/sysfs.c:150: dev_warn_once(dev, ./drivers/nvme/host/sysfs.c-151- "No UUID available providing old NGUID\n"); ./drivers/nvme/host/sysfs.c-152- return sysfs_emit(buf, "%pU\n", ids->nguid); ./drivers/nvme/host/sysfs.c-153- } ```This warning will only complain for the first partition of the first device also. And for NVMe devices, according to the NVM-Express-1_4 specification p.175 Figure 251:
"Bit 9 (UUID List): ...", UUID is not mandatory i guess. So let's degrade it into a informational message on initial discovery for every device.Signed-off-by Alan Cui <[email protected]>---
drivers/nvme/host/core.c +4
-0@@ -4296,6 +4296,10 @@ static void nvme_validate_ns(struct nvme_ns *ns, struct nvme_ns_info *info) goto out; } +if (uuid_is_null (&info->ids.uuid)) { +dev_info (ns->ctrl->device,"No UUID available, uuid_show providing old NGUID\n" ); + } + ret =nvme_update_ns_info (ns, info); out:/* drivers/nvme/host/sysfs.c +0
-0
drivers/nvme/host/sysfs.c static ssize_t uuid_show(struct device *dev, struct device_attribute *attr, · line 152
Quoted 8 lines
drivers/nvme/host/sysfs.c +0
-2@@ -147,8 +147,6 @@ static ssize_t uuid_show(struct device *dev, struct device_attribute *attr,* we have no UUID set */ if (uuid_is_null (&ids->uuid)) { -dev_warn_once (dev, -"No UUID available providing old NGUID\n" );return sysfs_emit (buf,"%pU\n" , ids->nguid); }return sysfs_emit (buf,"%pU\n" , &ids->uuid);
TBH, I'm not sure why should this be a warning or an info log. This looks like a debug print to me.
Keith WDYT?
Hi,
On Monday, 11 May 2026 06:19,you wrote:
Quoted 4 lines
TBH, I'm not sure why should this be a warning or an info log. This looks like a debug print to me.
Keith WDYT?
By the way, current way how kernel generate UUID by NGUID is not compliant with RFC 4122 and NVMe spec.
[alan@AlanArchDesktop ~]$ cat /sys/block/nvme0n1/uuid 00000000-0000-0000-a428-xxxxxxxxxxxx [alan@AlanArchDesktop ~]$ cat /sys/block/nvme0n1/nguid 00000000-0000-0000-a428-xxxxxxxxxxxx
And according to NVMe 1.4 spec 7.10.6 UUID
The Universally Unique Identifier is defined in RFC4122 and con- tained in the Namespace Identification Descriptor
So,
According to RFC 4122 4.3 Algorithm for Creating a Name-Based UUID:
Quoted 9 lines
The version 3 or 5 UUID is meant for generating UUIDs from "names" that are drawn from, and unique within, some "name space". The concept of name and name space should be broadly construed, and not limited to textual names. For example, some name spaces are the domain name system, URLs, ISO Object IDs (OIDs), X.500 Distinguished Names (DNs), and reserved words in a programming language. The mechanisms or conventions used for allocating names and ensuring their uniqueness within their name spaces are beyond the scope of this specification.
...
Quoted 19 lines
The algorithm for generating a UUID from a name and a name space are as follows:
o Allocate a UUID to use as a "name space ID" for all UUIDs generated from names in that name space; see Appendix C for some pre-defined values.
o Choose either MD5 [4] or SHA-1 [8] as the hash algorithm; If backward compatibility is not an issue, SHA-1 is preferred.
o Convert the name to a canonical sequence of octets (as defined by the standards or conventions of its name space); put the name space ID in network byte order.
o Compute the hash of the name space ID concatenated with the name.
o Set octts zero through 3 of thetime_lowfield to octets zero through 3 of the hash.
o Set octets zero and one of thetime_midfield to octets 4 and 5 of the hash.
o Set octets zero and one of thetime_hi_and_versionfield to octets 6 and 7 of the hash.
o Set the four most significant bits (bits 12 through 15) of thetime_hi_and_versionfield to the appropriate 4-bit version number
Quoted 8 lines
from Section 4.1.3.
o Set theclock_seq_hi_and_reservedfield to octet 8 of the hash.
o Set the two most significant bits (bits 6 and 7) of theclock_seq_hi_and_reservedto zero and one, respectively.
o Set theclock_seq_lowfield to octet 9 of the hash.
o Set octets zero through five of the node field to octets 10 through 15 of the hash.
o Convert the resulting UUID to local byte order.
However, the NVMe spec does not specified a namespace ID, so,
I selected a stirng "nqn.2014-08.org.nvmexpress:nguid",
then do SHA1 on it, get 4dead072-6b40-f992-9447-40f235d55fee,
then replace the version to '5' and variant to '1' to follow RFC 4122, get 4dead072-6b40-1992-8447-40f235d55fee
So for the NGUID A42812345678ABCD, a way to generate a canonnical UUID is:
```
import uuid
namespace = uuid.UUID('4dead072-6b40-1992-8447-40f235d55fee')
nguid_bytes = bytes.fromhex('A42812345678ABCD')
result_uuid = uuid.uuid5(namespace, nguid_bytes)
print(result_uuid)
```Resulting 9f4344db-6bf0-5782-b8a9-a08b6aa55169.
It's too complicated, right? RFC9562 provides us UUIDv8,
Quoted 3 lines
UUIDv8 provides a format for experimental or vendor-specific
use cases. The only requirement is that the variant and version bits MUST be set as defined in Sections 4.1 and 4.2.
Let's set the first 6 avalible custom bytes to "NGUID:", move the first 4bits of NGUID before the variant 4bits .
So for the NGUID A42812345678ABCD, it's 78718573-6858-800A-8428-12345678abcd.
Alan.
[the patch seem to be missing some maintainer Ccs]
Quoted 2 lines · Sagi Grimberg
TBH, I'm not sure why should this be a warning or an info log. This looks like a debug print to me.
It is worse. For some reason the commit seems to imply the uuid replaces the nguid and they can be intermixed.
We really need separate sysfs files for the different ids. I gues we need to keep the existing one someone, but it is a mess.
On Monday, 11 May 2026 16:16,Christoph Hellwig wrote:
Quoted 1 line
[the patch seem to be missing some maintainer Ccs]
Sorry, I've resent it.
Quoted 6 lines · Sagi Grimberg
Quoted 2 lines
TBH, I'm not sure why should this be a warning or an info log. This looks like a debug print to me.
It is worse. For some reason the commit seems to imply the uuid replaces the nguid and they can be intermixed.
Which you refer to?
If it's my patch, My patch is actually intended to fix the issue where "the kernel uses dev_warn_once, thus only reporting warnings for the first block or drive where this problem occurs". So the key is "This asymmetry can be misleading to users. If all devices in the system report the same issue, it might not be a problem, but if only one device reports it, it might (especially since I have two identical drives). ".
Futher more, should it be reported as a warning? The answer is No:
According to the NVM-Express-1_4 specification p.175 Figure 251:
"Bit 9 (UUID List): ...", UUID is not mandatory. And for devices below spec revision 1.3, there is no UUID yet.
And yes?? Because according to NVMe 1.4 spec 7.10.6 UUID:
Quoted 2 lines
The Universally Unique Identifier is defined in RFC4122 and con- tained in the Namespace Identification Descriptor
Current way how kernel generate UUID by NGUID is not compliant with RFC 4122 and NVMe spec.
[alan@AlanArchDesktop ~]$ cat /sys/block/nvme0n1/uuid 00000000-0000-0000-a428-xxxxxxxxxxxx
[alan@AlanArchDesktop ~]$ cat /sys/block/nvme0n1/nguid 00000000-0000-0000-a428-xxxxxxxxxxxx
Quoted 2 lines
We really need separate sysfs files for the different ids. I gues we need to keep the existing one someone, but it is a mess.
I don't really know how and why user needs this UUID, because kernel said:
/* For backward compatibility expose the NGUID to userspace if
* we have no UUID set
*/
[alan@AlanArchDesktop ~]$ ls /sys/block/nvme0n1/uuid (NVMe 1.4) /sys/block/nvme0n1/uuid
[alan@AlanArchDesktop ~]$ cat /sys/block/nvme0n1/uuid 00000000-0000-0000-a428-xxxxxxxxxxxx
[alan@AlanArchDesktop ~]$ cat /sys/block/nvme0n1/nguid 00000000-0000-0000-a428-xxxxxxxxxxxx
[alan@AlanArchDesktop ~]$ ls /sys/block/nvme2n1/uuid (NVMe 1.3)
ls: cannot access '/sys/block/nvme2n1/uuid': No such file or directory [alan@AlanArchDesktop ~]$ cat /sys/block/nvme2n1/nguid
cat: /sys/block/nvme2n1/nguid: No such file or directory [alan@AlanArchDesktop ~]$ cat /sys/block/nvme2n1/uuid
cat: /sys/block/nvme2n1/uuid: No such file or directory
If it's mandatory for userspace, why the nvme2n1 has not, if it's not mandatory, why should we create the UUID sysfd attr. then put NGUID in it? But for a NVMe 1.3 device there is both no UUID and no NGUID, while kernel is not reporting warnings at all?
Quoted 5 lines · The NVMe 1.4 spec Figure. 253 said:
If the namespace does not support an IEEE Extended Unigue ldentifier (i.e., EUI64 field is cleared to 0h) and does not support a Namespace Globally Unigue ldentifier (i.e., the NGUID field is cleared to Oh), then the namespace shall report a Namespace ldentification Descriptor with a value of type 3h.
So it's really a optional parameter right?
Alan
Quoted 7 lines · On Monday, 11 May 2026 16:59,I wrote:
For NVMe spec revision 1.3, there is no UUID yet.
If it's mandatory for userspace, why the nvme2n1 has not, if it's not mandatory, why should we create the UUID sysfd attr. then put NGUID in it? But for a NVMe 1.3 device there is both no UUID and no NGUID, while kernel is not reporting warnings at all?
Sorry, I checked the 1.3 specification and found that there is NGUID and UUID in it.
The only difference between revision 1.3 and 1.4 is Revision 1.4 states that UUIDs should be reported when neither EUI64 nor NGUID is supported.
It seems that for an NVMe 1.4 device, its UUID is mandatory when EUI64/NGUID is not supported. However, the kernel chooses to expose both the UUID and NGUID of a device with an NGUID but no UUID to sysfs, then warns us every time we read the device that an NGUID is being used to impersonate a UUID.
But for an NVMe 1.3 device, its UUID is optional when EUI64/NGUID is not present. However, the kernel chooses not to expose either the UUID or NGUID of a device with an NGUID but no UUID to sysfs.
This is incredibly strange.
Alan.
Quoted 4 lines · On Mon, May 11, 2026 at 04:59:19PM +0800, AlanCui4080 wrote:
Quoted 2 lines
It is worse. For some reason the commit seems to imply the uuid replaces the nguid and they can be intermixed.
Which you refer to?
The original patch adding the warning.
Keith Busch
maintainerQuoted 11 lines · Christoph Hellwig
[the patch seem to be missing some maintainer Ccs]
Quoted 2 lines · Sagi Grimberg
TBH, I'm not sure why should this be a warning or an info log. This looks like a debug print to me.
It is worse. For some reason the commit seems to imply the uuid replaces the nguid and they can be intermixed.
We really need separate sysfs files for the different ids. I gues we need to keep the existing one someone, but it is a mess.
Yes, sorry for the unpleasant situation this is. The attributes were rushed into the driver at the time.
Quoted 7 lines · Keith Busch
Quoted 6 lines · Christoph Hellwig
[the patch seem to be missing some maintainer Ccs]
Quoted 2 lines · Sagi Grimberg
TBH, I'm not sure why should this be a warning or an info log. This looks like a debug print to me.
Quoted 8 lines
Quoted 5 lines
It is worse. For some reason the commit seems to imply the uuid replaces the nguid and they can be intermixed.
We really need separate sysfs files for the different ids. I gues we need to keep the existing one someone, but it is a mess.
Yes, sorry for the unpleasant situation this is. The attributes were rushed into the driver at the time.
I'm not trying to put blame. But I think we need to sort this out somehow - plenty of devices ship without a uuid and a system log warning is not useful. So I guess we need to slowly deprecate this attribute and add proper ones for nguid/uuid. But I can't really come up with good names for them.
RFC PATCH v2 1/1 nvme: make providing NGUID as UUID usage less scary
In reply to AlanCui4080Since we decide to show the prp passthrough usage warning at device initial discovery stage and degrade it to a informational message, here is a warning that I think are essentially the same.
grep -C3 -rn "dev_warn_once" ./drivers/nvme/*
```
./drivers/nvme/host/sysfs.c-147- * we have no UUID set
./drivers/nvme/host/sysfs.c-148- */
./drivers/nvme/host/sysfs.c-149- if (uuid_is_null(&ids->uuid)) {
./drivers/nvme/host/sysfs.c:150: dev_warn_once(dev,
./drivers/nvme/host/sysfs.c-151- "No UUID available providing old NGUID\n");
./drivers/nvme/host/sysfs.c-152- return sysfs_emit(buf, "%pU\n", ids->nguid);
./drivers/nvme/host/sysfs.c-153- }
```This warning will only complain for the first partition of the first device also. And for NVMe devices, according to the NVM-Express-1_4 specification p.175 Figure 251:
"Bit 9 (UUID List): ...", UUID is not mandatory. So let's degrade it into a informational message on initial discovery for every device.
---
v2: move the hint to nvme_init_ns_head instead of nvme_validate_ns
v2: clean the commit message lines
In my understanding, the nvme_init_ns_head will be called each time a namespace created, and it does some checking on ids so we also place the hint there, and here is the result of dmesg applied current patch.
resend v2: add Ccs to all maintainers
resend v2: see https://lists.infradead.org/pipermail/linux-nvme/2026-May/062831.html
After discussion, I begin to think we should treat this as a warning,
since the way kernel generate UUID by NGUID is not a valid RFC4122 UUID, and NVMe specification requires "The Universally Unique Identifier is defined in RFC4122"
However the "No UUID available providing old NGUID" is should be reported per device anyway.
[ 4.502052] nvme nvme1: pci function 0000:05:00.0 [ 4.502060] nvme nvme0: pci function 0000:01:00.0 [ 4.508544] nvme nvme1: missing or invalid SUBNQN field. [ 4.508664] nvme nvme1: D3 entry latency set to 8 seconds [ 4.512422] nvme nvme0: missing or invalid SUBNQN field. [ 4.517581] nvme nvme0: allocated 32 MiB host memory buffer (1 segment). [ 4.523519] nvme nvme0: 8/0/0 default/read/poll queues [ 4.524761] nvme nvme1: 16/0/0 default/read/poll queues [ 4.525438] nvme nvme0: No UUID available, uuid_show providing old NGUID [ 4.526683] nvme nvme1: No UUID available, uuid_show providing old NGUID [ 4.527243] nvme0n1: p1 p2 p3 p4 p5 p6 [ 4.528351] nvme1n1: p1
drivers/nvme/host/core.c +4 -0
@@ -4007,6 +4007,10 @@ static int nvme_init_ns_head(struct nvme_ns *ns, struct nvme_ns_info *info) ctrl->quirks |=NVME_QUIRK_BOGUS_NID ; } +if (uuid_is_null (&info->ids.uuid)) { +dev_info (ctrl->device,"No UUID available, uuid_show providing old NGUID\n" ); + } +mutex_lock (&ctrl->subsys->lock); head =nvme_find_ns_head (ctrl, info->nsid);if (!head) {
drivers/nvme/host/sysfs.c +0 -2
@@ -147,8 +147,6 @@ static ssize_t uuid_show(struct device *dev, struct device_attribute *attr,* we have no UUID set */ if (uuid_is_null (&ids->uuid)) { -dev_warn_once (dev, -"No UUID available providing old NGUID\n" );return sysfs_emit (buf,"%pU\n" , ids->nguid); }return sysfs_emit (buf,"%pU\n" , &ids->uuid);
--
2.54.0
[see https://lists.infradead.org/pipermail/linux-nvme/2026-May/062762.html for full thread]
Hi everyone,
About this patch and the discussion above, I just did a survey on kernel commit history:
Long ago, the kernel sees NGUID as UUID.
2b9b6e86bca7209de02754fc84acf7ab3e78734e (NVMe: Export namespace attributes to sysfs)
Quoted 2 lines
if (ns->ctrl->vs >= NVME_VS(1, 2)) memcpy(ns->uuid, id->nguid, sizeof(ns->uuid));
Then, the kernel begins by retrieving the UUID from the driver and, in a backward compatible manner, reports the NGUID as the UUID when the UUID is unavailable, and throws a warning.
d934f9848a77be4afe0ca336ea419dd066c934f3 (nvme: provide UUID value to userspace Now that we have a way for getting the UUID from a target, provide it to userspace as well.
Unfortunately there is already a sysfs attribute called UUID which is a misnomer as it holds the NGUID value. So instead of creating yet another wrong name, create a new 'nguid' sysfs attribute for the NGUID. For the UUID attribute add a check wheter the namespace has a UUID assigned to it and return this or return the NGUID to maintain backwards compatibility. This should give userspace a chance to catch up.)
Quoted 10 lines
static ssize_t uuid_show(...)
/* For backward compatibility expose the NGUID to userspace if
* we have no UUID set
*/
if (uuid_is_null(&ns->uuid)) {
printk_ratelimited(KERN_WARNING
"No UUID available providing old NGUID\n");
return sprintf(buf, "%pU\n", ns->nguid);
}
return sprintf(buf, "%pU\n", &ns->uuid);And at d934f984:
Quoted 13 lines
static
umode_tnvme_ns_attrs_are_visible(struct kobject *kobj,
struct attribute *a, int n){ struct device *dev = container_of(kobj, struct device, kobj); struct nvme_ns *ns = nvme_get_ns_from_dev(dev); if (a == &dev_attr_uuid.attr) { if (uuid_is_null(&ns->uuid) || !memchr_inv(ns->nguid, 0, sizeof(ns->nguid))) // ??????? Why return 0; } if (a == &dev_attr_nguid.attr) { if (!memchr_inv(ns->nguid, 0, sizeof(ns->nguid)))
Quoted 8 lines
return 0;
}
if (a == &dev_attr_eui.attr) {
if (!memchr_inv(ns->eui, 0, sizeof(ns->eui)))
return 0;
}Therefore, my suggestion is to directly remove the UUID that is copied from NGUID.
return a->mode;
}The visibility of the uuid attr is depends on if NGUID exists, I'dont know why. I believe that in this scenario, if NGUID doesn't exist, then UUID won't exist either. Therefore, the user-space assumption that UUID exists has always been a bug. Consequently, we don't need to assume any responsibility for backward compatibility.
I have hard drives that can demonstrate this behavior.
``` sudo nvme id-ctrl /dev/nvme1n1 vid : 0x144d mn : Samsung SSD 970 EVO Plus 500GB fr : 2B2QEXM7 ieee : 002538 ver : 0x10300 cat /sys/block/nvme1n1/eui 00 25 ** ** ** ** ** 2a cat /sys/block/nvme1n1/nguid cat: /sys/block/nvme1n1/nguid: No such file or directory cat /sys/block/nvme1n1/uuid cat: /sys/block/nvme1n1/uuid: No such file or directory cat /sys/block/nvme1n1/wwid eui.0025**********a2 ```
As you can see, this NVMe 1.3 compatible device have no nguid, so it has no both nguid and uuid node.
``` sudo nvme id-ctrl /dev/nvme0n1 NVME Identify Controller: vid : 0x1e49 mn : ZHITAI TiPlus7100 2TB fr : ZTA22002 ieee : 000000 ver : 0x10400 cat /sys/block/nvme0n1/eui a4 ** ** ** ** ** ** c3 cat /sys/block/nvme0n1/nguid 00000000-0000-0000-a428-**********c3 cat /sys/block/nvme0n1/uuid 00000000-0000-0000-a428-**********c3 cat /sys/block/nvme0n1/wwid eui.0000000000000000a428**********c3 ```
As you can see, this NVMe 1.4 compatible device have a nguid, its nguid is derived from eui64, then because it has an nguid, its uuid is visible and is copied from nguid.
According to the spec., both devices are compliant.
Therefore, I believe the core issue is no longer just the level or location of the warning. As you can see, in the current version of the kernel, the visibility of each attr is highly device-dependent. I think we shouldn't retain a backward-compatible UUID attr since no userspace software should assume the existence of any attr and The current kernel code is not backward compatible in essence! (How can userspace determine whether a device has an NGUID through means other than the kernel sysfs? If a program in a previous version assumed the existence of /sys/block/nvmexx/uuid, then that assumption is incorrect, because NGUID itself may not exist. Only any one of EUI, NGUID, and UUID needs to exist.)
Another issue is Why WWIDs generated from NGUIDs have prefix "eui."?
Seems come from:118472ab8532e55f48395ef5764b354fe48b1d73 (NVMe: Expose ns wwid through single sysfs entry)
Quoted 2 lines
if (
memchr_inv(ns->uuid, 0, sizeof(ns->uuid)))return sprintf(buf, "eui.%16phN\n", ns->uuid);
This should be a naming error.
Then, when the kernel changed the UUID to a more appropriate name, the eui. prefix was also carried over. And the commit said "This patch has the driver figure this out and exports the unique string through a single 'wwid' attribute so the user doesn't have this burden.”
Since /dev/disk/by-id/ already uses wwid, I don't think there's much we can do, otherwise this would destroy compatibility. This seems to indirectly confirm that the kernel is requiring users to check the existence of /sys/block/nvmexx/[eui,nguid,uuid] before using it.
Therefore, my suggestion is to directly remove the UUID that is copied from NGUID. Make the visibility of UUID depend on whether the device has a UUID, rather than an NGUID to completely eliminates this strange warning and backward compatibility path.
Thank you for your attention to this matter, I'm look forward to your thoughts.
Alan.