[PATCH v4 0/1] i40e: additional safety check

[PATCH v4 0/1] i40e: additional safety check

[gregory.herrero]

From: Gregory Herrero <gregory.herrero@oracle.com>

On code inspection, I realized we may want to check ring_len parameter
against hardware specific values in i40e_config_vsi_tx_queue() and
i40e_config_vsi_rx_queue().

v4:
- remove u32 cast in i40e_config_vsi_tx_queue() too and don't mention it
  anymore in commit description.
- wrap i40e_get_max_num_descriptors() description
v3:
- drop trailing period from the subject
- reword commit description
- remove u32 cast in i40e_config_vsi_rx_queue()
v2:
- make i40e_get_max_num_descriptors() 'pf' argument const.
- reword i40e_get_max_num_descriptors() description.
- modify commit description to explain potential behavior change.

Gregory Herrero (1):
  i40e: validate ring_len parameter against hardware-specific values

 drivers/net/ethernet/intel/i40e/i40e.h         | 18 ++++++++++++++++++
 drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
 .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
 3 files changed, 20 insertions(+), 14 deletions(-)

-- 
2.51.0

[PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[gregory.herrero]

From: Gregory Herrero <gregory.herrero@oracle.com>

The maximum number of descriptors supported by the hardware is hardware
dependent and can be retrieved using i40e_get_max_num_descriptors().
Move this function to a shared header and use it when checking for valid
ring_len parameter rather than using hardcoded value.

By fixing an over-acceptance issue, behavior change could be seen where
ring_len could now be rejected while configuring rx and tx queues if its
size is larger than the hardware-specific maximum number of descriptors.

Fixes: 55d225670def ("i40e: add validation for ring_len param")
Signed-off-by: Gregory Herrero <gregory.herrero@oracle.com>
---
 drivers/net/ethernet/intel/i40e/i40e.h         | 18 ++++++++++++++++++
 drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
 .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
 3 files changed, 20 insertions(+), 14 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 801a57a925da..5b367397ae43 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -1418,4 +1418,22 @@ static inline struct i40e_veb *i40e_pf_get_main_veb(struct i40e_pf *pf)
 	return (pf->lan_veb != I40E_NO_VEB) ? pf->veb[pf->lan_veb] : NULL;
 }
 
+/**
+ * i40e_get_max_num_descriptors - get maximum number of descriptors for this
+ * hardware.
+ * @pf: pointer to a PF
+ *
+ * Return: u32 value corresponding to the maximum number of descriptors.
+ **/
+static inline u32 i40e_get_max_num_descriptors(const struct i40e_pf *pf)
+{
+	const struct i40e_hw *hw = &pf->hw;
+
+	switch (hw->mac.type) {
+	case I40E_MAC_XL710:
+		return I40E_MAX_NUM_DESCRIPTORS_XL710;
+	default:
+		return I40E_MAX_NUM_DESCRIPTORS;
+	}
+}
 #endif /* _I40E_H_ */
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 86c72596617a..61c39e881b00 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -2013,18 +2013,6 @@ static void i40e_get_drvinfo(struct net_device *netdev,
 		drvinfo->n_priv_flags += I40E_GL_PRIV_FLAGS_STR_LEN;
 }
 
-static u32 i40e_get_max_num_descriptors(struct i40e_pf *pf)
-{
-	struct i40e_hw *hw = &pf->hw;
-
-	switch (hw->mac.type) {
-	case I40E_MAC_XL710:
-		return I40E_MAX_NUM_DESCRIPTORS_XL710;
-	default:
-		return I40E_MAX_NUM_DESCRIPTORS;
-	}
-}
-
 static void i40e_get_ringparam(struct net_device *netdev,
 			       struct ethtool_ringparam *ring,
 			       struct kernel_ethtool_ringparam *kernel_ring,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 081a4526a2f0..cf831c649c9c 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -656,7 +656,7 @@ static int i40e_config_vsi_tx_queue(struct i40e_vf *vf, u16 vsi_id,
 
 	/* ring_len has to be multiple of 8 */
 	if (!IS_ALIGNED(info->ring_len, 8) ||
-	    info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
+	    info->ring_len > i40e_get_max_num_descriptors(pf)) {
 		ret = -EINVAL;
 		goto error_context;
 	}
@@ -726,7 +726,7 @@ static int i40e_config_vsi_rx_queue(struct i40e_vf *vf, u16 vsi_id,
 
 	/* ring_len has to be multiple of 32 */
 	if (!IS_ALIGNED(info->ring_len, 32) ||
-	    info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
+	    info->ring_len > i40e_get_max_num_descriptors(pf)) {
 		ret = -EINVAL;
 		goto error_param;
 	}
-- 
2.51.0

RE: [PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[Loktionov, Aleksandr]

> -----Original Message-----
> From: gregory.herrero@oracle.com <gregory.herrero@oracle.com>
> Sent: Monday, November 17, 2025 9:33 AM
> To: Loktionov, Aleksandr <aleksandr.loktionov@intel.com>; Nguyen,
> Anthony L <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> <przemyslaw.kitszel@intel.com>; andrew+netdev@lunn.ch;
> davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> pabeni@redhat.com
> Cc: intel-wired-lan@lists.osuosl.org; netdev@vger.kernel.org; linux-
> kernel@vger.kernel.org; Gregory Herrero <gregory.herrero@oracle.com>
> Subject: [PATCH v4 1/1] i40e: validate ring_len parameter against
> hardware-specific values
> 
> From: Gregory Herrero <gregory.herrero@oracle.com>
> 
> The maximum number of descriptors supported by the hardware is
> hardware dependent and can be retrieved using
First paragraph uses “hardware dependent” (no hyphen) while later text uses “hardware‑specific” (hyphenated).
Prefer “hardware‑dependent” for consistency.

> i40e_get_max_num_descriptors().
> Move this function to a shared header and use it when checking for
> valid ring_len parameter rather than using hardcoded value.
> 
> By fixing an over-acceptance issue, behavior change could be seen
> where ring_len could now be rejected while configuring rx and tx
> queues if its size is larger than the hardware-specific maximum number
> of descriptors.
> 
The message explains the behavioral change but does not state how the change was tested
(e.g., which MAC types exercised, ethtool -G paths, VF configuration via virtchnl, acceptance/rejection boundaries). 
Netdev routinely asks for this when behavior changes.

> Fixes: 55d225670def ("i40e: add validation for ring_len param")
> Signed-off-by: Gregory Herrero <gregory.herrero@oracle.com>
> ---
>  drivers/net/ethernet/intel/i40e/i40e.h         | 18
> ++++++++++++++++++
>  drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
> .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
>  3 files changed, 20 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/i40e/i40e.h
> b/drivers/net/ethernet/intel/i40e/i40e.h
> index 801a57a925da..5b367397ae43 100644
> --- a/drivers/net/ethernet/intel/i40e/i40e.h
> +++ b/drivers/net/ethernet/intel/i40e/i40e.h
> @@ -1418,4 +1418,22 @@ static inline struct i40e_veb
> *i40e_pf_get_main_veb(struct i40e_pf *pf)
>  	return (pf->lan_veb != I40E_NO_VEB) ? pf->veb[pf->lan_veb] :
> NULL;  }

...

> --
> 2.51.0

RE: [PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[Romanowski, Rafal]

> -----Original Message-----
> From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
> Loktionov, Aleksandr
> Sent: Monday, November 17, 2025 12:59
> To: gregory.herrero@oracle.com; Nguyen, Anthony L
> <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> <przemyslaw.kitszel@intel.com>; andrew+netdev@lunn.ch;
> davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> pabeni@redhat.com
> Cc: intel-wired-lan@lists.osuosl.org; netdev@vger.kernel.org; linux-
> kernel@vger.kernel.org
> Subject: Re: [Intel-wired-lan] [PATCH v4 1/1] i40e: validate ring_len parameter
> against hardware-specific values
> 
> 
> 
> > -----Original Message-----
> > From: gregory.herrero@oracle.com <gregory.herrero@oracle.com>
> > Sent: Monday, November 17, 2025 9:33 AM
> > To: Loktionov, Aleksandr <aleksandr.loktionov@intel.com>; Nguyen,
> > Anthony L <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> > <przemyslaw.kitszel@intel.com>; andrew+netdev@lunn.ch;
> > davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> > pabeni@redhat.com
> > Cc: intel-wired-lan@lists.osuosl.org; netdev@vger.kernel.org; linux-
> > kernel@vger.kernel.org; Gregory Herrero <gregory.herrero@oracle.com>
> > Subject: [PATCH v4 1/1] i40e: validate ring_len parameter against
> > hardware-specific values
> >
> > From: Gregory Herrero <gregory.herrero@oracle.com>
> >
> > The maximum number of descriptors supported by the hardware is
> > hardware dependent and can be retrieved using
> First paragraph uses “hardware dependent” (no hyphen) while later text uses
> “hardware‑specific” (hyphenated).
> Prefer “hardware‑dependent” for consistency.
> 
> > i40e_get_max_num_descriptors().
> > Move this function to a shared header and use it when checking for
> > valid ring_len parameter rather than using hardcoded value.
> >
> > By fixing an over-acceptance issue, behavior change could be seen
> > where ring_len could now be rejected while configuring rx and tx
> > queues if its size is larger than the hardware-specific maximum number
> > of descriptors.
> >
> The message explains the behavioral change but does not state how the change
> was tested (e.g., which MAC types exercised, ethtool -G paths, VF configuration
> via virtchnl, acceptance/rejection boundaries).
> Netdev routinely asks for this when behavior changes.
> 
> > Fixes: 55d225670def ("i40e: add validation for ring_len param")
> > Signed-off-by: Gregory Herrero <gregory.herrero@oracle.com>
> > ---
> >  drivers/net/ethernet/intel/i40e/i40e.h         | 18
> > ++++++++++++++++++
> >  drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
> > .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
> >  3 files changed, 20 insertions(+), 14 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/intel/i40e/i40e.h
> > b/drivers/net/ethernet/intel/i40e/i40e.h
> > index 801a57a925da..5b367397ae43 100644
> > --- a/drivers/net/ethernet/intel/i40e/i40e.h
> > +++ b/drivers/net/ethernet/intel/i40e/i40e.h
> > @@ -1418,4 +1418,22 @@ static inline struct i40e_veb
> > *i40e_pf_get_main_veb(struct i40e_pf *pf)
> >  	return (pf->lan_veb != I40E_NO_VEB) ? pf->veb[pf->lan_veb] :
> > NULL;  }
> 
> ...
> 
> > --
> > 2.51.0


Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>

Re: [PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[Creeley, Brett]

On 11/17/2025 12:33 AM, gregory.herrero@oracle.com wrote:
> Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
>
>
> From: Gregory Herrero <gregory.herrero@oracle.com>
>
> The maximum number of descriptors supported by the hardware is hardware
> dependent and can be retrieved using i40e_get_max_num_descriptors().
> Move this function to a shared header and use it when checking for valid
> ring_len parameter rather than using hardcoded value.
>
> By fixing an over-acceptance issue, behavior change could be seen where
> ring_len could now be rejected while configuring rx and tx queues if its
> size is larger than the hardware-specific maximum number of descriptors.
>
> Fixes: 55d225670def ("i40e: add validation for ring_len param")
> Signed-off-by: Gregory Herrero <gregory.herrero@oracle.com>
> ---
>   drivers/net/ethernet/intel/i40e/i40e.h         | 18 ++++++++++++++++++
>   drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
>   .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
>   3 files changed, 20 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
> index 801a57a925da..5b367397ae43 100644
> --- a/drivers/net/ethernet/intel/i40e/i40e.h
> +++ b/drivers/net/ethernet/intel/i40e/i40e.h
> @@ -1418,4 +1418,22 @@ static inline struct i40e_veb *i40e_pf_get_main_veb(struct i40e_pf *pf)
>          return (pf->lan_veb != I40E_NO_VEB) ? pf->veb[pf->lan_veb] : NULL;
>   }
>
> +/**
> + * i40e_get_max_num_descriptors - get maximum number of descriptors for this
> + * hardware.
> + * @pf: pointer to a PF
> + *
> + * Return: u32 value corresponding to the maximum number of descriptors.
> + **/

Nit, but the function name is descriptive enough without the documentation.

I think the purpose of the function would be even more obvious if the 
argument was a pointer to the hw structure instead of a pointer to the 
pf since the max is based on the hw not the pf.

Brett

> +static inline u32 i40e_get_max_num_descriptors(const struct i40e_pf *pf)
> +{
> +       const struct i40e_hw *hw = &pf->hw;
> +
> +       switch (hw->mac.type) {
> +       case I40E_MAC_XL710:
> +               return I40E_MAX_NUM_DESCRIPTORS_XL710;
> +       default:
> +               return I40E_MAX_NUM_DESCRIPTORS;
> +       }
> +}
>   #endif /* _I40E_H_ */
> diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
> index 86c72596617a..61c39e881b00 100644
> --- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
> +++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
> @@ -2013,18 +2013,6 @@ static void i40e_get_drvinfo(struct net_device *netdev,
>                  drvinfo->n_priv_flags += I40E_GL_PRIV_FLAGS_STR_LEN;
>   }
>
> -static u32 i40e_get_max_num_descriptors(struct i40e_pf *pf)
> -{
> -       struct i40e_hw *hw = &pf->hw;
> -
> -       switch (hw->mac.type) {
> -       case I40E_MAC_XL710:
> -               return I40E_MAX_NUM_DESCRIPTORS_XL710;
> -       default:
> -               return I40E_MAX_NUM_DESCRIPTORS;
> -       }
> -}
> -
>   static void i40e_get_ringparam(struct net_device *netdev,
>                                 struct ethtool_ringparam *ring,
>                                 struct kernel_ethtool_ringparam *kernel_ring,
> diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
> index 081a4526a2f0..cf831c649c9c 100644
> --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
> +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
> @@ -656,7 +656,7 @@ static int i40e_config_vsi_tx_queue(struct i40e_vf *vf, u16 vsi_id,
>
>          /* ring_len has to be multiple of 8 */
>          if (!IS_ALIGNED(info->ring_len, 8) ||
> -           info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
> +           info->ring_len > i40e_get_max_num_descriptors(pf)) {
>                  ret = -EINVAL;
>                  goto error_context;
>          }
> @@ -726,7 +726,7 @@ static int i40e_config_vsi_rx_queue(struct i40e_vf *vf, u16 vsi_id,
>
>          /* ring_len has to be multiple of 32 */
>          if (!IS_ALIGNED(info->ring_len, 32) ||
> -           info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
> +           info->ring_len > i40e_get_max_num_descriptors(pf)) {
>                  ret = -EINVAL;
>                  goto error_param;
>          }
> --
> 2.51.0
>
>

Re: [PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[Gregory Herrero]

On Mon, Nov 17, 2025 at 11:58:48AM +0000, Loktionov, Aleksandr wrote:
> 
> 
> > -----Original Message-----
> > From: gregory.herrero@oracle.com <gregory.herrero@oracle.com>
> > Sent: Monday, November 17, 2025 9:33 AM
> > To: Loktionov, Aleksandr <aleksandr.loktionov@intel.com>; Nguyen,
> > Anthony L <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> > <przemyslaw.kitszel@intel.com>; andrew+netdev@lunn.ch;
> > davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> > pabeni@redhat.com
> > Cc: intel-wired-lan@lists.osuosl.org; netdev@vger.kernel.org; linux-
> > kernel@vger.kernel.org; Gregory Herrero <gregory.herrero@oracle.com>
> > Subject: [PATCH v4 1/1] i40e: validate ring_len parameter against
> > hardware-specific values
> > 
> > From: Gregory Herrero <gregory.herrero@oracle.com>
> > 
> > The maximum number of descriptors supported by the hardware is
> > hardware dependent and can be retrieved using
> First paragraph uses “hardware dependent” (no hyphen) while later text uses “hardware‑specific” (hyphenated).
> Prefer “hardware‑dependent” for consistency.
> 
I will address in v5
> > i40e_get_max_num_descriptors().
> > Move this function to a shared header and use it when checking for
> > valid ring_len parameter rather than using hardcoded value.
> > 
> > By fixing an over-acceptance issue, behavior change could be seen
> > where ring_len could now be rejected while configuring rx and tx
> > queues if its size is larger than the hardware-specific maximum number
> > of descriptors.
> > 
> The message explains the behavioral change but does not state how the change was tested
> (e.g., which MAC types exercised, ethtool -G paths, VF configuration via virtchnl, acceptance/rejection boundaries). 
> Netdev routinely asks for this when behavior changes.
> 
In the meantime, Rafal Romanowski tested it so I will add:
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>

Let me know if that's enough or if more details are required.

Thanks,
Gregory

Re: [PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[Gregory Herrero]

On Fri, Dec 12, 2025 at 09:37:20AM -0800, Creeley, Brett wrote:
> 
> On 11/17/2025 12:33 AM, gregory.herrero@oracle.com wrote:
> > Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
> > 
> > 
> > From: Gregory Herrero <gregory.herrero@oracle.com>
> > 
> > The maximum number of descriptors supported by the hardware is hardware
> > dependent and can be retrieved using i40e_get_max_num_descriptors().
> > Move this function to a shared header and use it when checking for valid
> > ring_len parameter rather than using hardcoded value.
> > 
> > By fixing an over-acceptance issue, behavior change could be seen where
> > ring_len could now be rejected while configuring rx and tx queues if its
> > size is larger than the hardware-specific maximum number of descriptors.
> > 
> > Fixes: 55d225670def ("i40e: add validation for ring_len param")
> > Signed-off-by: Gregory Herrero <gregory.herrero@oracle.com>
> > ---
> >   drivers/net/ethernet/intel/i40e/i40e.h         | 18 ++++++++++++++++++
> >   drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
> >   .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
> >   3 files changed, 20 insertions(+), 14 deletions(-)
> > 
> > diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
> > index 801a57a925da..5b367397ae43 100644
> > --- a/drivers/net/ethernet/intel/i40e/i40e.h
> > +++ b/drivers/net/ethernet/intel/i40e/i40e.h
> > @@ -1418,4 +1418,22 @@ static inline struct i40e_veb *i40e_pf_get_main_veb(struct i40e_pf *pf)
> >          return (pf->lan_veb != I40E_NO_VEB) ? pf->veb[pf->lan_veb] : NULL;
> >   }
> > 
> > +/**
> > + * i40e_get_max_num_descriptors - get maximum number of descriptors for this
> > + * hardware.
> > + * @pf: pointer to a PF
> > + *
> > + * Return: u32 value corresponding to the maximum number of descriptors.
> > + **/
> 
> Nit, but the function name is descriptive enough without the documentation.
> 
> I think the purpose of the function would be even more obvious if the
> argument was a pointer to the hw structure instead of a pointer to the pf
> since the max is based on the hw not the pf.
> 
I agree, it's just that it will require changing the 5 callers of this
function and invalidate the testing from Rafal Romanowski.
Please let me know what you think, I can wait before sending v5.

Thanks,
Gregory
> > +static inline u32 i40e_get_max_num_descriptors(const struct i40e_pf *pf)
> > +{
> > +       const struct i40e_hw *hw = &pf->hw;
> > +
> > +       switch (hw->mac.type) {
> > +       case I40E_MAC_XL710:
> > +               return I40E_MAX_NUM_DESCRIPTORS_XL710;
> > +       default:
> > +               return I40E_MAX_NUM_DESCRIPTORS;
> > +       }
> > +}
> >   #endif /* _I40E_H_ */
> > diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
> > index 86c72596617a..61c39e881b00 100644
> > --- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
> > +++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
> > @@ -2013,18 +2013,6 @@ static void i40e_get_drvinfo(struct net_device *netdev,
> >                  drvinfo->n_priv_flags += I40E_GL_PRIV_FLAGS_STR_LEN;
> >   }
> > 
> > -static u32 i40e_get_max_num_descriptors(struct i40e_pf *pf)
> > -{
> > -       struct i40e_hw *hw = &pf->hw;
> > -
> > -       switch (hw->mac.type) {
> > -       case I40E_MAC_XL710:
> > -               return I40E_MAX_NUM_DESCRIPTORS_XL710;
> > -       default:
> > -               return I40E_MAX_NUM_DESCRIPTORS;
> > -       }
> > -}
> > -
> >   static void i40e_get_ringparam(struct net_device *netdev,
> >                                 struct ethtool_ringparam *ring,
> >                                 struct kernel_ethtool_ringparam *kernel_ring,
> > diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
> > index 081a4526a2f0..cf831c649c9c 100644
> > --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
> > +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
> > @@ -656,7 +656,7 @@ static int i40e_config_vsi_tx_queue(struct i40e_vf *vf, u16 vsi_id,
> > 
> >          /* ring_len has to be multiple of 8 */
> >          if (!IS_ALIGNED(info->ring_len, 8) ||
> > -           info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
> > +           info->ring_len > i40e_get_max_num_descriptors(pf)) {
> >                  ret = -EINVAL;
> >                  goto error_context;
> >          }
> > @@ -726,7 +726,7 @@ static int i40e_config_vsi_rx_queue(struct i40e_vf *vf, u16 vsi_id,
> > 
> >          /* ring_len has to be multiple of 32 */
> >          if (!IS_ALIGNED(info->ring_len, 32) ||
> > -           info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
> > +           info->ring_len > i40e_get_max_num_descriptors(pf)) {
> >                  ret = -EINVAL;
> >                  goto error_param;
> >          }
> > --
> > 2.51.0
> > 
> > 

Re: [PATCH v4 1/1] i40e: validate ring_len parameter against hardware-specific values

[Creeley, Brett]

On 12/12/2025 12:22 PM, Gregory Herrero wrote:
> Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
>
>
> On Fri, Dec 12, 2025 at 09:37:20AM -0800, Creeley, Brett wrote:
>> On 11/17/2025 12:33 AM, gregory.herrero@oracle.com wrote:
>>> Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
>>>
>>>
>>> From: Gregory Herrero <gregory.herrero@oracle.com>
>>>
>>> The maximum number of descriptors supported by the hardware is hardware
>>> dependent and can be retrieved using i40e_get_max_num_descriptors().
>>> Move this function to a shared header and use it when checking for valid
>>> ring_len parameter rather than using hardcoded value.
>>>
>>> By fixing an over-acceptance issue, behavior change could be seen where
>>> ring_len could now be rejected while configuring rx and tx queues if its
>>> size is larger than the hardware-specific maximum number of descriptors.
>>>
>>> Fixes: 55d225670def ("i40e: add validation for ring_len param")
>>> Signed-off-by: Gregory Herrero <gregory.herrero@oracle.com>
>>> ---
>>>    drivers/net/ethernet/intel/i40e/i40e.h         | 18 ++++++++++++++++++
>>>    drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 ------------
>>>    .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  4 ++--
>>>    3 files changed, 20 insertions(+), 14 deletions(-)
>>>
>>> diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
>>> index 801a57a925da..5b367397ae43 100644
>>> --- a/drivers/net/ethernet/intel/i40e/i40e.h
>>> +++ b/drivers/net/ethernet/intel/i40e/i40e.h
>>> @@ -1418,4 +1418,22 @@ static inline struct i40e_veb *i40e_pf_get_main_veb(struct i40e_pf *pf)
>>>           return (pf->lan_veb != I40E_NO_VEB) ? pf->veb[pf->lan_veb] : NULL;
>>>    }
>>>
>>> +/**
>>> + * i40e_get_max_num_descriptors - get maximum number of descriptors for this
>>> + * hardware.
>>> + * @pf: pointer to a PF
>>> + *
>>> + * Return: u32 value corresponding to the maximum number of descriptors.
>>> + **/
>> Nit, but the function name is descriptive enough without the documentation.
>>
>> I think the purpose of the function would be even more obvious if the
>> argument was a pointer to the hw structure instead of a pointer to the pf
>> since the max is based on the hw not the pf.
>>
> I agree, it's just that it will require changing the 5 callers of this
> function and invalidate the testing from Rafal Romanowski.
> Please let me know what you think, I can wait before sending v5.

I'm fine with the PF being the parameter. However, the documentation 
should really be removed IMO.

Thanks,

Brett
>
> Thanks,
> Gregory
>>> +static inline u32 i40e_get_max_num_descriptors(const struct i40e_pf *pf)
>>> +{
>>> +       const struct i40e_hw *hw = &pf->hw;
>>> +
>>> +       switch (hw->mac.type) {
>>> +       case I40E_MAC_XL710:
>>> +               return I40E_MAX_NUM_DESCRIPTORS_XL710;
>>> +       default:
>>> +               return I40E_MAX_NUM_DESCRIPTORS;
>>> +       }
>>> +}
>>>    #endif /* _I40E_H_ */
>>> diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
>>> index 86c72596617a..61c39e881b00 100644
>>> --- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
>>> +++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
>>> @@ -2013,18 +2013,6 @@ static void i40e_get_drvinfo(struct net_device *netdev,
>>>                   drvinfo->n_priv_flags += I40E_GL_PRIV_FLAGS_STR_LEN;
>>>    }
>>>
>>> -static u32 i40e_get_max_num_descriptors(struct i40e_pf *pf)
>>> -{
>>> -       struct i40e_hw *hw = &pf->hw;
>>> -
>>> -       switch (hw->mac.type) {
>>> -       case I40E_MAC_XL710:
>>> -               return I40E_MAX_NUM_DESCRIPTORS_XL710;
>>> -       default:
>>> -               return I40E_MAX_NUM_DESCRIPTORS;
>>> -       }
>>> -}
>>> -
>>>    static void i40e_get_ringparam(struct net_device *netdev,
>>>                                  struct ethtool_ringparam *ring,
>>>                                  struct kernel_ethtool_ringparam *kernel_ring,
>>> diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
>>> index 081a4526a2f0..cf831c649c9c 100644
>>> --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
>>> +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
>>> @@ -656,7 +656,7 @@ static int i40e_config_vsi_tx_queue(struct i40e_vf *vf, u16 vsi_id,
>>>
>>>           /* ring_len has to be multiple of 8 */
>>>           if (!IS_ALIGNED(info->ring_len, 8) ||
>>> -           info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
>>> +           info->ring_len > i40e_get_max_num_descriptors(pf)) {
>>>                   ret = -EINVAL;
>>>                   goto error_context;
>>>           }
>>> @@ -726,7 +726,7 @@ static int i40e_config_vsi_rx_queue(struct i40e_vf *vf, u16 vsi_id,
>>>
>>>           /* ring_len has to be multiple of 32 */
>>>           if (!IS_ALIGNED(info->ring_len, 32) ||
>>> -           info->ring_len > I40E_MAX_NUM_DESCRIPTORS_XL710) {
>>> +           info->ring_len > i40e_get_max_num_descriptors(pf)) {
>>>                   ret = -EINVAL;
>>>                   goto error_param;
>>>           }
>>> --
>>> 2.51.0
>>>
>>>