[PATCH 0/4] multipath-tools: add new NVMe devices to hwtable

[PATCH 0/4] multipath-tools: add new NVMe devices to hwtable

[Xose Vazquez Perez]

Even if they work with the generic NVME config. It is convenient to have
them as an independent config to tweak their options, compare dm-multipath
against the NVMe subsystem, ...

Only two of them are missing ("Fujitsu ETERNUS AB/HB Series" and "NetApp E-Series")
as requested by its vendor: https://marc.info/?t=165256933900004

Xose Vazquez Perez (4):
 multipath-tools: remove checker from Huawei OceanStor NVMe
 multipath-tools: add HITACHI VSP NVMe to hwtable
 multipath-tools: add Infinidat InfiniBox NVMe to hwtable
 multipath-tools: add VASTData Block Subsystem NVMe to hwtable

 libmultipath/hwtable.c | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

Cc: Steven Schremmer <Steve.Schremmer@netapp.com>
Cc: NetApp RDAC team <ng-eseries-upstream-maintainers@netapp.com>
Cc: Martin Wilck <mwilck@suse.com>
Cc: Benjamin Marzinski <bmarzins@redhat.com>
Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>
Cc: DM_DEVEL-ML <dm-devel@lists.linux.dev>
Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>
-- 
2.51.0

[PATCH v2 1/4] multipath-tools: remove checker from Huawei OceanStor NVMe

[Xose Vazquez Perez]

Based on info from fbe89f4. It was deleted from default NVMe config:
"hwtable: set 'none' as default checker for NVMe

We shouldn't be using the directio checker for NVMe, as this might
cause spurious path failures under high load.
As the 'state' attribute provides reliable information about the link
status it's sufficient to just check this and don't call any
path checkers."

v2: add info from fbe89f4

Cc: Hannes Reinecke <hare@suse.de>
Cc: Martin Wilck <mwilck@suse.com>
Cc: Benjamin Marzinski <bmarzins@redhat.com>
Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>
Cc: DM_DEVEL-ML <dm-devel@lists.linux.dev>
Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>
---
 libmultipath/hwtable.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/libmultipath/hwtable.c b/libmultipath/hwtable.c
index d9723f87..399545ae 100644
--- a/libmultipath/hwtable.c
+++ b/libmultipath/hwtable.c
@@ -1121,7 +1121,6 @@ static struct hwentry default_hw[] = {
 		/* OceanStor NVMe */
 		.vendor        = "NVME",
 		.product       = "Huawei-XSG1",
-		.checker_name  = DIRECTIO,
 		.no_path_retry = 12,
 	},
 	/*
-- 
2.51.0

[PATCH 2/4] multipath-tools: add HITACHI VSP NVMe to hwtable

[Xose Vazquez Perez]

Cc: Martin Wilck <mwilck@suse.com>
Cc: Benjamin Marzinski <bmarzins@redhat.com>
Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>
Cc: DM_DEVEL-ML <dm-devel@lists.linux.dev>
Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>
---
 libmultipath/hwtable.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/libmultipath/hwtable.c b/libmultipath/hwtable.c
index 399545ae..fb446856 100644
--- a/libmultipath/hwtable.c
+++ b/libmultipath/hwtable.c
@@ -510,6 +510,12 @@ static struct hwentry default_hw[] = {
 		.checker_name  = DIRECTIO,
 		.detect_checker = DETECT_CHECKER_OFF,
 	},
+	{
+		// VSP NVMe
+		.vendor        = "NVME",
+		.product       = "HITACHI SVOS-RF-System",
+		.no_path_retry = NO_PATH_RETRY_QUEUE,
+	},
 	/*
 	 * IBM
 	 */
-- 
2.51.0

[PATCH 3/4] multipath-tools: add Infinidat InfiniBox NVMe to hwtable

[Xose Vazquez Perez]

Cc: Martin Wilck <mwilck@suse.com>
Cc: Benjamin Marzinski <bmarzins@redhat.com>
Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>
Cc: DM_DEVEL-ML <dm-devel@lists.linux.dev>
Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>
---
 libmultipath/hwtable.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/libmultipath/hwtable.c b/libmultipath/hwtable.c
index fb446856..dd3bd30c 100644
--- a/libmultipath/hwtable.c
+++ b/libmultipath/hwtable.c
@@ -1154,6 +1154,12 @@ static struct hwentry default_hw[] = {
 		.dev_loss      = MAX_DEV_LOSS_TMO,
 		.detect_prio   = DETECT_PRIO_OFF,
 	},
+	{
+		// InfiniBox NVMe
+		.vendor        = "NVME",
+		.product       = "InfiniBox",
+		.no_path_retry = NO_PATH_RETRY_QUEUE,
+	},
 	/*
 	 * Kaminario
 	 */
-- 
2.51.0

[PATCH 4/4] multipath-tools: add VASTData Block Subsystem NVMe to hwtable

[Xose Vazquez Perez]

Cc: Martin Wilck <mwilck@suse.com>
Cc: Benjamin Marzinski <bmarzins@redhat.com>
Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>
Cc: DM_DEVEL-ML <dm-devel@lists.linux.dev>
Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>
---
 libmultipath/hwtable.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/libmultipath/hwtable.c b/libmultipath/hwtable.c
index dd3bd30c..2b0f17f5 100644
--- a/libmultipath/hwtable.c
+++ b/libmultipath/hwtable.c
@@ -1416,6 +1416,15 @@ static struct hwentry default_hw[] = {
 		.no_path_retry = 30,
 		.prio_args     = "exclusive_pref_bit",
 	},
+	/*
+	 * VAST Data
+	 */
+	{
+		// Block Subsystem NVMe
+		.vendor        = "NVME",
+		.product       = "VASTData",
+		.no_path_retry = NO_PATH_RETRY_QUEUE,
+	},
 	/*
 	 * EOL
 	 */
-- 
2.51.0

Re: [PATCH 0/4] multipath-tools: add new NVMe devices to hwtable

[Martin Wilck]

On Sat, 2025-09-13 at 15:02 +0200, Xose Vazquez Perez wrote:
> Even if they work with the generic NVME config. It is convenient to
> have
> them as an independent config to tweak their options, compare dm-
> multipath
> against the NVMe subsystem, ..
> 
> Only two of them are missing ("Fujitsu ETERNUS AB/HB Series" and
> "NetApp E-Series")
> as requested by its vendor: https://marc.info/?t=165256933900004
> 
> Xose Vazquez Perez (4):
>  multipath-tools: remove checker from Huawei OceanStor NVMe
>  multipath-tools: add HITACHI VSP NVMe to hwtable
>  multipath-tools: add Infinidat InfiniBox NVMe to hwtable
>  multipath-tools: add VASTData Block Subsystem NVMe to hwtable
> 
>  libmultipath/hwtable.c | 22 +++++++++++++++++++++-
>  1 file changed, 21 insertions(+), 1 deletion(-)
> 
> Cc: Steven Schremmer <Steve.Schremmer@netapp.com>
> Cc: NetApp RDAC team <ng-eseries-upstream-maintainers@netapp.com>
> Cc: Martin Wilck <mwilck@suse.com>
> Cc: Benjamin Marzinski <bmarzins@redhat.com>
> Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>
> Cc: DM_DEVEL-ML <dm-devel@lists.linux.dev>
> Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>

For the series:
Reviewed-by: Martin Wilck <mwilck@suse.com>

I realize you set .no_path_retry = NO_PATH_RETRY_QUEUE for all devices.
Should we use this for the generic NVMe device, too?

Martin

Re: [PATCH 0/4] multipath-tools: add new NVMe devices to hwtable

[Xose Vazquez Perez]

On 9/22/25 8:56 PM, Martin Wilck wrote:

> I realize you set .no_path_retry = NO_PATH_RETRY_QUEUE for all devices.
> Should we use this for the generic NVMe device, too?

I played it safe, because most of them are using that value.
"generic" is fine as it is.