[PATCH v2 0/2] Preclude IOCCSZ/IORCSZ validation for admin controller connect

[PATCH v2 0/2] Preclude IOCCSZ/IORCSZ validation for admin controller connect

[Kamaljit Singh]

During negative testing it was observed that when ioccsz and/or iorcsz 
were cleared to zero on a Target, a new connection to an administrative 
controller failed to establish, due to driver validation of those fields. 
Since an administrative controller does not support I/O queues, these 
checks should be excluded for such a controller.

Kamaljit Singh (2):
  nvme-core: add method to check for an I/O controller
  nvme-core: do ioccsz/iorcsz validation only for I/O controllers

 drivers/nvme/host/core.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

-- 
2.43.0

[PATCH v2 1/2] nvme-core: add method to check for an I/O controller

[Kamaljit Singh]

Add nvme_is_io_ctrl() to check if the controller is of type I/O
controller. Uses negative logic by excluding an administrative
controller and a discovery controller.

Signed-off-by: Kamaljit Singh <kamaljit.singh@opensource.wdc.com>
---
 drivers/nvme/host/core.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index d67545bd7abb..fab672e50510 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -3167,6 +3167,11 @@ static inline bool nvme_admin_ctrl(struct nvme_ctrl *ctrl)
 	return ctrl->cntrltype == NVME_CTRL_ADMIN;
 }
 
+static inline bool nvme_is_io_ctrl(struct nvme_ctrl *ctrl)
+{
+	return !nvme_discovery_ctrl(ctrl) && !nvme_admin_ctrl(ctrl);
+}
+
 static bool nvme_validate_cntlid(struct nvme_subsystem *subsys,
 		struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id)
 {
-- 
2.43.0

Re: [PATCH v2 1/2] nvme-core: add method to check for an I/O controller

[Hannes Reinecke]

On 9/6/25 01:25, Kamaljit Singh wrote:
> Add nvme_is_io_ctrl() to check if the controller is of type I/O
> controller. Uses negative logic by excluding an administrative
> controller and a discovery controller.
> 
> Signed-off-by: Kamaljit Singh <kamaljit.singh@opensource.wdc.com>
> ---
>   drivers/nvme/host/core.c | 5 +++++
>   1 file changed, 5 insertions(+)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH v2 1/2] nvme-core: add method to check for an I/O controller

[Kamaljit Singh]

On Mon, Sep 08, 2025 at 02:56:35PM +0200, Hannes Reinecke wrote:
> On 9/6/25 01:25, Kamaljit Singh wrote:
> > Add nvme_is_io_ctrl() to check if the controller is of type I/O
> > controller. Uses negative logic by excluding an administrative
> > controller and a discovery controller.
> > 
> > Signed-off-by: Kamaljit Singh <kamaljit.singh@opensource.wdc.com>
> > ---
> >   drivers/nvme/host/core.c | 5 +++++
> >   1 file changed, 5 insertions(+)
> > 
> Reviewed-by: Hannes Reinecke <hare@suse.de>
> 
Please review and provide any feedback on these 2 patches.

Thanks,
Kamaljit Singh

[PATCH v2 2/2] nvme-core: do ioccsz/iorcsz validation only for I/O controllers

[Kamaljit Singh]

An administrative controller does not support I/O queues, hence it
should ignore existing checks for IOCCSZ/IORCSZ. Currently, these checks
only exclude a discovery controller but need to also exclude an
administrative controller.

Signed-off-by: Kamaljit Singh <kamaljit.singh@opensource.wdc.com>
---
 drivers/nvme/host/core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index fab672e50510..ef6f84d677d8 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -3496,14 +3496,14 @@ static int nvme_check_ctrl_fabric_info(struct nvme_ctrl *ctrl, struct nvme_id_ct
 		return -EINVAL;
 	}
 
-	if (!nvme_discovery_ctrl(ctrl) && ctrl->ioccsz < 4) {
+	if (nvme_is_io_ctrl(ctrl) && ctrl->ioccsz < 4) {
 		dev_err(ctrl->device,
 			"I/O queue command capsule supported size %d < 4\n",
 			ctrl->ioccsz);
 		return -EINVAL;
 	}
 
-	if (!nvme_discovery_ctrl(ctrl) && ctrl->iorcsz < 1) {
+	if (nvme_is_io_ctrl(ctrl) && ctrl->iorcsz < 1) {
 		dev_err(ctrl->device,
 			"I/O queue response capsule supported size %d < 1\n",
 			ctrl->iorcsz);
-- 
2.43.0

Re: [PATCH v2 2/2] nvme-core: do ioccsz/iorcsz validation only for I/O controllers

[Hannes Reinecke]

On 9/6/25 01:25, Kamaljit Singh wrote:
> An administrative controller does not support I/O queues, hence it
> should ignore existing checks for IOCCSZ/IORCSZ. Currently, these checks
> only exclude a discovery controller but need to also exclude an
> administrative controller.
> 
> Signed-off-by: Kamaljit Singh <kamaljit.singh@opensource.wdc.com>
> ---
>   drivers/nvme/host/core.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH v2 2/2] nvme-core: do ioccsz/iorcsz validation only for I/O controllers

[Kamaljit Singh]

On Mon, Sep 08, 2025 at 02:57:00PM +0200, Hannes Reinecke wrote:
> On 9/6/25 01:25, Kamaljit Singh wrote:
> > An administrative controller does not support I/O queues, hence it
> > should ignore existing checks for IOCCSZ/IORCSZ. Currently, these checks
> > only exclude a discovery controller but need to also exclude an
> > administrative controller.
> > 
> > Signed-off-by: Kamaljit Singh <kamaljit.singh@opensource.wdc.com>
> > ---
> >   drivers/nvme/host/core.c | 4 ++--
> >   1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> Reviewed-by: Hannes Reinecke <hare@suse.de>
> 
Please review and provide any feedback for both of these patches.

Thanks,
Kamaljit

Re: [PATCH v2 0/2] Preclude IOCCSZ/IORCSZ validation for admin controller connect

[Keith Busch]

On Fri, Sep 05, 2025 at 04:25:48PM -0700, Kamaljit Singh wrote:
> During negative testing it was observed that when ioccsz and/or iorcsz 
> were cleared to zero on a Target, a new connection to an administrative 
> controller failed to establish, due to driver validation of those fields. 
> Since an administrative controller does not support I/O queues, these 
> checks should be excluded for such a controller.

Thanks, applied to nvme-6.18.