netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
@ 2023-01-02 15:02 Lukasz Majewski
  2023-01-02 15:02 ` [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch Lukasz Majewski
                   ` (3 more replies)
  0 siblings, 4 replies; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-02 15:02 UTC (permalink / raw)
  To: Andrew Lunn, Vladimir Oltean
  Cc: Eric Dumazet, Florian Fainelli, David S. Miller, Jakub Kicinski,
	Russell King, Paolo Abeni, Alexander Duyck, netdev, linux-kernel,
	Lukasz Majewski

Different Marvell DSA switches support different size of max frame
bytes to be sent.

For example mv88e6185 supports max 1632 bytes, which is now in-driver
standard value. On the other hand - mv88e6250 supports 2048 bytes.

As this value is internal and may be different for each switch IC,
new entry in struct mv88e6xxx_info has been added to store it.

Signed-off-by: Lukasz Majewski <lukma@denx.de>

---
Changes for v2:
- Define max_frame_size with default value of 1632 bytes,
- Set proper value for the mv88e6250 switch SoC (linkstreet) family

Changes for v3:
- Add default value for 1632B of the max frame size (to avoid problems
  with not defined values)
---
 drivers/net/dsa/mv88e6xxx/chip.c | 13 ++++++++++++-
 drivers/net/dsa/mv88e6xxx/chip.h |  1 +
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
index 242b8b325504..19668e549391 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.c
+++ b/drivers/net/dsa/mv88e6xxx/chip.c
@@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct dsa_switch *ds, int port)
 	if (chip->info->ops->port_set_jumbo_size)
 		return 10240 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
 	else if (chip->info->ops->set_max_frame_size)
-		return 1632 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
+		return (max_t(int, chip->info->max_frame_size, 1632)
+			- VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN);
+
 	return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
 }
 
@@ -4955,6 +4957,7 @@ static const struct mv88e6xxx_ops mv88e6250_ops = {
 	.avb_ops = &mv88e6352_avb_ops,
 	.ptp_ops = &mv88e6250_ptp_ops,
 	.phylink_get_caps = mv88e6250_phylink_get_caps,
+	.set_max_frame_size = mv88e6185_g1_set_max_frame_size,
 };
 
 static const struct mv88e6xxx_ops mv88e6290_ops = {
@@ -5574,6 +5577,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.atu_move_port_mask = 0xf,
 		.multi_chip = true,
 		.ops = &mv88e6095_ops,
+		.max_frame_size = 1632,
 	},
 
 	[MV88E6097] = {
@@ -5598,6 +5602,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.multi_chip = true,
 		.edsa_support = MV88E6XXX_EDSA_SUPPORTED,
 		.ops = &mv88e6097_ops,
+		.max_frame_size = 1632,
 	},
 
 	[MV88E6123] = {
@@ -5622,6 +5627,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.multi_chip = true,
 		.edsa_support = MV88E6XXX_EDSA_SUPPORTED,
 		.ops = &mv88e6123_ops,
+		.max_frame_size = 1632,
 	},
 
 	[MV88E6131] = {
@@ -5692,6 +5698,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.edsa_support = MV88E6XXX_EDSA_SUPPORTED,
 		.ptp_support = true,
 		.ops = &mv88e6161_ops,
+		.max_frame_size = 1632,
 	},
 
 	[MV88E6165] = {
@@ -5716,6 +5723,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.multi_chip = true,
 		.ptp_support = true,
 		.ops = &mv88e6165_ops,
+		.max_frame_size = 1632,
 	},
 
 	[MV88E6171] = {
@@ -5835,6 +5843,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.multi_chip = true,
 		.edsa_support = MV88E6XXX_EDSA_SUPPORTED,
 		.ops = &mv88e6185_ops,
+		.max_frame_size = 1632,
 	},
 
 	[MV88E6190] = {
@@ -5968,6 +5977,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.num_internal_phys = 2,
 		.invalid_port_mask = BIT(2) | BIT(3) | BIT(4),
 		.max_vid = 4095,
+		.max_frame_size = 2048,
 		.port_base_addr = 0x08,
 		.phy_base_addr = 0x00,
 		.global1_addr = 0x0f,
@@ -6015,6 +6025,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.num_ports = 7,
 		.num_internal_phys = 5,
 		.max_vid = 4095,
+		.max_frame_size = 2048,
 		.port_base_addr = 0x08,
 		.phy_base_addr = 0x00,
 		.global1_addr = 0x0f,
diff --git a/drivers/net/dsa/mv88e6xxx/chip.h b/drivers/net/dsa/mv88e6xxx/chip.h
index e693154cf803..55948ef56cd0 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.h
+++ b/drivers/net/dsa/mv88e6xxx/chip.h
@@ -132,6 +132,7 @@ struct mv88e6xxx_info {
 	unsigned int num_gpio;
 	unsigned int max_vid;
 	unsigned int max_sid;
+	unsigned int max_frame_size;
 	unsigned int port_base_addr;
 	unsigned int phy_base_addr;
 	unsigned int global1_addr;
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 12+ messages in thread

* [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch
  2023-01-02 15:02 [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Lukasz Majewski
@ 2023-01-02 15:02 ` Lukasz Majewski
  2023-01-02 20:00   ` Andrew Lunn
  2023-01-02 15:02 ` [PATCH v3 3/3] net: dsa: mv88e6xxx: add support for MV88E6071 switch Lukasz Majewski
                   ` (2 subsequent siblings)
  3 siblings, 1 reply; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-02 15:02 UTC (permalink / raw)
  To: Andrew Lunn, Vladimir Oltean
  Cc: Eric Dumazet, Florian Fainelli, David S. Miller, Jakub Kicinski,
	Russell King, Paolo Abeni, Alexander Duyck, netdev, linux-kernel,
	Matthias Schiffer, Lukasz Majewski

From: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>

A mv88e6250 family (i.e. LinkStreet) switch with 2 PHY and RMII ports
and no PTP support.

Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Signed-off-by: Lukasz Majewski <lukma@denx.de>
---
Changes for v2:
- Add S-o-B
- Update commit message
- Add information about max packet size (2048 B)

Changes for v3:
- None
---
 drivers/net/dsa/mv88e6xxx/chip.c | 21 +++++++++++++++++++++
 drivers/net/dsa/mv88e6xxx/chip.h |  1 +
 drivers/net/dsa/mv88e6xxx/port.h |  1 +
 3 files changed, 23 insertions(+)

diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
index 19668e549391..be563b6e5a89 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.c
+++ b/drivers/net/dsa/mv88e6xxx/chip.c
@@ -5536,6 +5536,27 @@ static const struct mv88e6xxx_ops mv88e6393x_ops = {
 };
 
 static const struct mv88e6xxx_info mv88e6xxx_table[] = {
+	[MV88E6020] = {
+		.prod_num = MV88E6XXX_PORT_SWITCH_ID_PROD_6020,
+		.family = MV88E6XXX_FAMILY_6250,
+		.name = "Marvell 88E6020",
+		.num_databases = 64,
+		.num_ports = 7,
+		.num_internal_phys = 5,
+		.max_vid = 4095,
+		.max_frame_size = 2048,
+		.port_base_addr = 0x8,
+		.phy_base_addr = 0x0,
+		.global1_addr = 0xf,
+		.global2_addr = 0x7,
+		.age_time_coeff = 15000,
+		.g1_irqs = 9,
+		.g2_irqs = 5,
+		.atu_move_port_mask = 0xf,
+		.dual_chip = true,
+		.ops = &mv88e6250_ops,
+	},
+
 	[MV88E6085] = {
 		.prod_num = MV88E6XXX_PORT_SWITCH_ID_PROD_6085,
 		.family = MV88E6XXX_FAMILY_6097,
diff --git a/drivers/net/dsa/mv88e6xxx/chip.h b/drivers/net/dsa/mv88e6xxx/chip.h
index 55948ef56cd0..0d838e1c8996 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.h
+++ b/drivers/net/dsa/mv88e6xxx/chip.h
@@ -54,6 +54,7 @@ enum mv88e6xxx_frame_mode {
 
 /* List of supported models */
 enum mv88e6xxx_model {
+	MV88E6020,
 	MV88E6085,
 	MV88E6095,
 	MV88E6097,
diff --git a/drivers/net/dsa/mv88e6xxx/port.h b/drivers/net/dsa/mv88e6xxx/port.h
index aec9d4fd20e3..169ce5b6fa31 100644
--- a/drivers/net/dsa/mv88e6xxx/port.h
+++ b/drivers/net/dsa/mv88e6xxx/port.h
@@ -111,6 +111,7 @@
 /* Offset 0x03: Switch Identifier Register */
 #define MV88E6XXX_PORT_SWITCH_ID		0x03
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_MASK	0xfff0
+#define MV88E6XXX_PORT_SWITCH_ID_PROD_6020	0x0200
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6085	0x04a0
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6095	0x0950
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6097	0x0990
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 12+ messages in thread

* [PATCH v3 3/3] net: dsa: mv88e6xxx: add support for MV88E6071 switch
  2023-01-02 15:02 [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Lukasz Majewski
  2023-01-02 15:02 ` [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch Lukasz Majewski
@ 2023-01-02 15:02 ` Lukasz Majewski
  2023-01-02 19:58 ` [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Andrew Lunn
  2023-01-02 20:29 ` Andrew Lunn
  3 siblings, 0 replies; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-02 15:02 UTC (permalink / raw)
  To: Andrew Lunn, Vladimir Oltean
  Cc: Eric Dumazet, Florian Fainelli, David S. Miller, Jakub Kicinski,
	Russell King, Paolo Abeni, Alexander Duyck, netdev, linux-kernel,
	Lukasz Majewski

A mv88e6250 family (i.e. LinkStreet) switch with 5 internal PHYs,
2 RMIIs and no PTP support.

Signed-off-by: Lukasz Majewski <lukma@denx.de>
---
Changes for v2:
- Update commit message
- Add information about max frame size

Changes for v3:
- None
---
 drivers/net/dsa/mv88e6xxx/chip.c | 21 +++++++++++++++++++++
 drivers/net/dsa/mv88e6xxx/chip.h |  1 +
 drivers/net/dsa/mv88e6xxx/port.h |  1 +
 3 files changed, 23 insertions(+)

diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
index be563b6e5a89..bedd23145f2a 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.c
+++ b/drivers/net/dsa/mv88e6xxx/chip.c
@@ -5557,6 +5557,27 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = {
 		.ops = &mv88e6250_ops,
 	},
 
+	[MV88E6071] = {
+		.prod_num = MV88E6XXX_PORT_SWITCH_ID_PROD_6071,
+		.family = MV88E6XXX_FAMILY_6250,
+		.name = "Marvell 88E6071",
+		.num_databases = 64,
+		.num_ports = 7,
+		.num_internal_phys = 5,
+		.max_vid = 4095,
+		.max_frame_size = 2048,
+		.port_base_addr = 0x08,
+		.phy_base_addr = 0x00,
+		.global1_addr = 0x0f,
+		.global2_addr = 0x07,
+		.age_time_coeff = 15000,
+		.g1_irqs = 9,
+		.g2_irqs = 5,
+		.atu_move_port_mask = 0xf,
+		.dual_chip = true,
+		.ops = &mv88e6250_ops,
+	},
+
 	[MV88E6085] = {
 		.prod_num = MV88E6XXX_PORT_SWITCH_ID_PROD_6085,
 		.family = MV88E6XXX_FAMILY_6097,
diff --git a/drivers/net/dsa/mv88e6xxx/chip.h b/drivers/net/dsa/mv88e6xxx/chip.h
index 0d838e1c8996..6a4ed41dfcd6 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.h
+++ b/drivers/net/dsa/mv88e6xxx/chip.h
@@ -55,6 +55,7 @@ enum mv88e6xxx_frame_mode {
 /* List of supported models */
 enum mv88e6xxx_model {
 	MV88E6020,
+	MV88E6071,
 	MV88E6085,
 	MV88E6095,
 	MV88E6097,
diff --git a/drivers/net/dsa/mv88e6xxx/port.h b/drivers/net/dsa/mv88e6xxx/port.h
index 169ce5b6fa31..494a221c9d9a 100644
--- a/drivers/net/dsa/mv88e6xxx/port.h
+++ b/drivers/net/dsa/mv88e6xxx/port.h
@@ -112,6 +112,7 @@
 #define MV88E6XXX_PORT_SWITCH_ID		0x03
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_MASK	0xfff0
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6020	0x0200
+#define MV88E6XXX_PORT_SWITCH_ID_PROD_6071	0x0710
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6085	0x04a0
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6095	0x0950
 #define MV88E6XXX_PORT_SWITCH_ID_PROD_6097	0x0990
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-02 15:02 [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Lukasz Majewski
  2023-01-02 15:02 ` [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch Lukasz Majewski
  2023-01-02 15:02 ` [PATCH v3 3/3] net: dsa: mv88e6xxx: add support for MV88E6071 switch Lukasz Majewski
@ 2023-01-02 19:58 ` Andrew Lunn
  2023-01-02 20:29 ` Andrew Lunn
  3 siblings, 0 replies; 12+ messages in thread
From: Andrew Lunn @ 2023-01-02 19:58 UTC (permalink / raw)
  To: Lukasz Majewski
  Cc: Vladimir Oltean, Eric Dumazet, Florian Fainelli, David S. Miller,
	Jakub Kicinski, Russell King, Paolo Abeni, Alexander Duyck,
	netdev, linux-kernel

On Mon, Jan 02, 2023 at 04:02:07PM +0100, Lukasz Majewski wrote:
> Different Marvell DSA switches support different size of max frame
> bytes to be sent.
> 
> For example mv88e6185 supports max 1632 bytes, which is now in-driver
> standard value. On the other hand - mv88e6250 supports 2048 bytes.
> 
> As this value is internal and may be different for each switch IC,
> new entry in struct mv88e6xxx_info has been added to store it.
> 
> Signed-off-by: Lukasz Majewski <lukma@denx.de>
> 
> ---
> Changes for v2:
> - Define max_frame_size with default value of 1632 bytes,
> - Set proper value for the mv88e6250 switch SoC (linkstreet) family
> 
> Changes for v3:
> - Add default value for 1632B of the max frame size (to avoid problems
>   with not defined values)

I would add a WARN_ON_ONCE(!chip->info->max_frame_size) so a missing
value is detected. You can then skip the complexity of a default.

	Andrew

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch
  2023-01-02 15:02 ` [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch Lukasz Majewski
@ 2023-01-02 20:00   ` Andrew Lunn
  2023-01-03  8:46     ` Lukasz Majewski
  0 siblings, 1 reply; 12+ messages in thread
From: Andrew Lunn @ 2023-01-02 20:00 UTC (permalink / raw)
  To: Lukasz Majewski
  Cc: Vladimir Oltean, Eric Dumazet, Florian Fainelli, David S. Miller,
	Jakub Kicinski, Russell King, Paolo Abeni, Alexander Duyck,
	netdev, linux-kernel, Matthias Schiffer

On Mon, Jan 02, 2023 at 04:02:08PM +0100, Lukasz Majewski wrote:
> From: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
> 
> A mv88e6250 family (i.e. LinkStreet) switch with 2 PHY and RMII ports
> and no PTP support.

>  static const struct mv88e6xxx_info mv88e6xxx_table[] = {
> +	[MV88E6020] = {
> +		.prod_num = MV88E6XXX_PORT_SWITCH_ID_PROD_6020,
> +		.family = MV88E6XXX_FAMILY_6250,
> +		.name = "Marvell 88E6020",
> +		.num_databases = 64,
> +		.num_ports = 7,
> +		.num_internal_phys = 5,

You say in the commit message there are two PHYs, yet you have 5 here?

    Andrew

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-02 15:02 [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Lukasz Majewski
                   ` (2 preceding siblings ...)
  2023-01-02 19:58 ` [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Andrew Lunn
@ 2023-01-02 20:29 ` Andrew Lunn
  2023-01-03  9:02   ` Lukasz Majewski
  3 siblings, 1 reply; 12+ messages in thread
From: Andrew Lunn @ 2023-01-02 20:29 UTC (permalink / raw)
  To: Lukasz Majewski
  Cc: Vladimir Oltean, Eric Dumazet, Florian Fainelli, David S. Miller,
	Jakub Kicinski, Russell King, Paolo Abeni, Alexander Duyck,
	netdev, linux-kernel

> @@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct dsa_switch *ds, int port)
>  	if (chip->info->ops->port_set_jumbo_size)
>  		return 10240 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
>  	else if (chip->info->ops->set_max_frame_size)
> -		return 1632 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
> +		return (max_t(int, chip->info->max_frame_size, 1632)
> +			- VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN);
> +
>  	return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;

I would also prefer if all this if/else logic is removed, and the code
simply returned chip->info->max_frame_size - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;

> +++ b/drivers/net/dsa/mv88e6xxx/chip.h
> @@ -132,6 +132,7 @@ struct mv88e6xxx_info {
>  	unsigned int num_gpio;
>  	unsigned int max_vid;
>  	unsigned int max_sid;
> +	unsigned int max_frame_size;

It might be worth adding a comment here what this value actually
represents. We don't want any mixups where the value already has the
frame checksum removed for example.

      Andrew

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch
  2023-01-02 20:00   ` Andrew Lunn
@ 2023-01-03  8:46     ` Lukasz Majewski
  0 siblings, 0 replies; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-03  8:46 UTC (permalink / raw)
  To: Andrew Lunn
  Cc: Vladimir Oltean, Eric Dumazet, Florian Fainelli, David S. Miller,
	Jakub Kicinski, Russell King, Paolo Abeni, Alexander Duyck,
	netdev, linux-kernel, Matthias Schiffer

[-- Attachment #1: Type: text/plain, Size: 1259 bytes --]

Hi Andrew,

> On Mon, Jan 02, 2023 at 04:02:08PM +0100, Lukasz Majewski wrote:
> > From: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
> > 
> > A mv88e6250 family (i.e. LinkStreet) switch with 2 PHY and RMII
> > ports and no PTP support.  
> 
> >  static const struct mv88e6xxx_info mv88e6xxx_table[] = {
> > +	[MV88E6020] = {
> > +		.prod_num = MV88E6XXX_PORT_SWITCH_ID_PROD_6020,
> > +		.family = MV88E6XXX_FAMILY_6250,
> > +		.name = "Marvell 88E6020",
> > +		.num_databases = 64,
> > +		.num_ports = 7,
> > +		.num_internal_phys = 5,  
> 
> You say in the commit message there are two PHYs, yet you have 5 here?
> 

It looks like mine copy-paste error.

In the documentation there is stated that the 88E6020 device contains
two 10BASE-T/100BASE-TX transceivers (PHYs) and four independent Fast
Ethernet MACs (so it is a 2+2 device).

The 88E6071 is 5+2 device, so I'm going to correct this value to 2.
Also num_ports needs to be updated to 4.

>     Andrew


Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-02 20:29 ` Andrew Lunn
@ 2023-01-03  9:02   ` Lukasz Majewski
  2023-01-03 12:54     ` Andrew Lunn
  2023-01-05 10:37     ` Lukasz Majewski
  0 siblings, 2 replies; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-03  9:02 UTC (permalink / raw)
  To: Andrew Lunn, Alexander Duyck
  Cc: Vladimir Oltean, Eric Dumazet, Florian Fainelli, David S. Miller,
	Jakub Kicinski, Russell King, Paolo Abeni, netdev, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 2314 bytes --]

Hi Andrew,

> > @@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct
> > dsa_switch *ds, int port) if (chip->info->ops->port_set_jumbo_size)
> >  		return 10240 - VLAN_ETH_HLEN - EDSA_HLEN -
> > ETH_FCS_LEN; else if (chip->info->ops->set_max_frame_size)
> > -		return 1632 - VLAN_ETH_HLEN - EDSA_HLEN -
> > ETH_FCS_LEN;
> > +		return (max_t(int, chip->info->max_frame_size,
> > 1632)
> > +			- VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN);
> > +
> >  	return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;  
> 
> I would also prefer if all this if/else logic is removed, and the code
> simply returned chip->info->max_frame_size - VLAN_ETH_HLEN -
> EDSA_HLEN - ETH_FCS_LEN;
> 

So then the mv88e6xxx_get_max_mtu shall look like:

WARN_ON_ONCE(!chip->info->max_frame_size)

if (chip->info->ops->port_set_jumbo_size)
...
else 
    return chip->info->max_frame_size - VLAN_ETH_HLEN -
	EDSA_HLEN - ETH_FCS_LEN;


Or shall I put WARN_ON_ONCE to the mv88e6xxx_probe() function?


The above approach is contrary to one proposed by Alexander, who wanted
to improve the defensive approach in this driver (to avoid situation
where the max_frame_size callback is not defined and max_frame_size
member of *_info struct is not added by developer).

Which approach is the recommended one for this driver?

> > +++ b/drivers/net/dsa/mv88e6xxx/chip.h
> > @@ -132,6 +132,7 @@ struct mv88e6xxx_info {
> >  	unsigned int num_gpio;
> >  	unsigned int max_vid;
> >  	unsigned int max_sid;
> > +	unsigned int max_frame_size;  
> 
> It might be worth adding a comment here what this value actually
> represents.

Ok. I will add proper comment.

> We don't want any mixups where the value already has the
> frame checksum removed for example.

Could you be more specific here about this use case?

The max_frame_size is the maximal size of the ethernet frame for which
the IC designer provided specified amount of RAM (it is a different
value for different SoCs in the Link Street family).

> 
>       Andrew


Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-03  9:02   ` Lukasz Majewski
@ 2023-01-03 12:54     ` Andrew Lunn
  2023-01-05 10:37     ` Lukasz Majewski
  1 sibling, 0 replies; 12+ messages in thread
From: Andrew Lunn @ 2023-01-03 12:54 UTC (permalink / raw)
  To: Lukasz Majewski
  Cc: Alexander Duyck, Vladimir Oltean, Eric Dumazet, Florian Fainelli,
	David S. Miller, Jakub Kicinski, Russell King, Paolo Abeni,
	netdev, linux-kernel

On Tue, Jan 03, 2023 at 10:02:51AM +0100, Lukasz Majewski wrote:
> Hi Andrew,
> 
> > > @@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct
> > > dsa_switch *ds, int port) if (chip->info->ops->port_set_jumbo_size)
> > >  		return 10240 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > ETH_FCS_LEN; else if (chip->info->ops->set_max_frame_size)
> > > -		return 1632 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > ETH_FCS_LEN;
> > > +		return (max_t(int, chip->info->max_frame_size,
> > > 1632)
> > > +			- VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN);
> > > +
> > >  	return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;  
> > 
> > I would also prefer if all this if/else logic is removed, and the code
> > simply returned chip->info->max_frame_size - VLAN_ETH_HLEN -
> > EDSA_HLEN - ETH_FCS_LEN;
> > 
> 
> So then the mv88e6xxx_get_max_mtu shall look like:
> 
> WARN_ON_ONCE(!chip->info->max_frame_size)
> 
> if (chip->info->ops->port_set_jumbo_size)
> ...
> else 
>     return chip->info->max_frame_size - VLAN_ETH_HLEN -
> 	EDSA_HLEN - ETH_FCS_LEN;

I think it should go even further:

{
	WARN_ON_ONCE(!chip->info->max_frame_size)

	return chip->info->max_frame_size - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
}

If we are going to use info->max_frame_size, we should always use
info->max_frame_size.

	Andrew

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-03  9:02   ` Lukasz Majewski
  2023-01-03 12:54     ` Andrew Lunn
@ 2023-01-05 10:37     ` Lukasz Majewski
  2023-01-05 16:13       ` Alexander Duyck
  1 sibling, 1 reply; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-05 10:37 UTC (permalink / raw)
  To: Andrew Lunn, Alexander Duyck
  Cc: Vladimir Oltean, Eric Dumazet, Florian Fainelli, David S. Miller,
	Jakub Kicinski, Russell King, Paolo Abeni, netdev, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 2858 bytes --]

Hi Andrew, Alexander,

> Hi Andrew,
> 
> > > @@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct
> > > dsa_switch *ds, int port) if
> > > (chip->info->ops->port_set_jumbo_size) return 10240 -
> > > VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN; else if
> > > (chip->info->ops->set_max_frame_size)
> > > -		return 1632 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > ETH_FCS_LEN;
> > > +		return (max_t(int, chip->info->max_frame_size,
> > > 1632)
> > > +			- VLAN_ETH_HLEN - EDSA_HLEN -
> > > ETH_FCS_LEN); +
> > >  	return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
> > >  
> > 
> > I would also prefer if all this if/else logic is removed, and the
> > code simply returned chip->info->max_frame_size - VLAN_ETH_HLEN -
> > EDSA_HLEN - ETH_FCS_LEN;
> >   
> 
> So then the mv88e6xxx_get_max_mtu shall look like:
> 
> WARN_ON_ONCE(!chip->info->max_frame_size)
> 
> if (chip->info->ops->port_set_jumbo_size)
> ...
> else 
>     return chip->info->max_frame_size - VLAN_ETH_HLEN -
> 	EDSA_HLEN - ETH_FCS_LEN;
> 
> 
> Or shall I put WARN_ON_ONCE to the mv88e6xxx_probe() function?
> 
> 
> The above approach is contrary to one proposed by Alexander, who
> wanted to improve the defensive approach in this driver (to avoid
> situation where the max_frame_size callback is not defined and
> max_frame_size member of *_info struct is not added by developer).
> 
> Which approach is the recommended one for this driver?

Is there any decision regarding the preferred approach to rewrite this
code?

> 
> > > +++ b/drivers/net/dsa/mv88e6xxx/chip.h
> > > @@ -132,6 +132,7 @@ struct mv88e6xxx_info {
> > >  	unsigned int num_gpio;
> > >  	unsigned int max_vid;
> > >  	unsigned int max_sid;
> > > +	unsigned int max_frame_size;    
> > 
> > It might be worth adding a comment here what this value actually
> > represents.  
> 
> Ok. I will add proper comment.
> 
> > We don't want any mixups where the value already has the
> > frame checksum removed for example.  
> 
> Could you be more specific here about this use case?
> 
> The max_frame_size is the maximal size of the ethernet frame for which
> the IC designer provided specified amount of RAM (it is a different
> value for different SoCs in the Link Street family).
> 
> > 
> >       Andrew  
> 
> 
> Best regards,
> 
> Lukasz Majewski
> 
> --
> 
> DENX Software Engineering GmbH,      Managing Director: Erika Unter
> HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
> Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
> lukma@denx.de




Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-05 10:37     ` Lukasz Majewski
@ 2023-01-05 16:13       ` Alexander Duyck
  2023-01-05 17:44         ` Lukasz Majewski
  0 siblings, 1 reply; 12+ messages in thread
From: Alexander Duyck @ 2023-01-05 16:13 UTC (permalink / raw)
  To: Lukasz Majewski
  Cc: Andrew Lunn, Vladimir Oltean, Eric Dumazet, Florian Fainelli,
	David S. Miller, Jakub Kicinski, Russell King, Paolo Abeni,
	netdev, linux-kernel

On Thu, Jan 5, 2023 at 2:37 AM Lukasz Majewski <lukma@denx.de> wrote:
>
> Hi Andrew, Alexander,
>
> > Hi Andrew,
> >
> > > > @@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct
> > > > dsa_switch *ds, int port) if
> > > > (chip->info->ops->port_set_jumbo_size) return 10240 -
> > > > VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN; else if
> > > > (chip->info->ops->set_max_frame_size)
> > > > -         return 1632 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > > ETH_FCS_LEN;
> > > > +         return (max_t(int, chip->info->max_frame_size,
> > > > 1632)
> > > > +                 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > > ETH_FCS_LEN); +
> > > >   return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
> > > >
> > >
> > > I would also prefer if all this if/else logic is removed, and the
> > > code simply returned chip->info->max_frame_size - VLAN_ETH_HLEN -
> > > EDSA_HLEN - ETH_FCS_LEN;
> > >
> >
> > So then the mv88e6xxx_get_max_mtu shall look like:
> >
> > WARN_ON_ONCE(!chip->info->max_frame_size)
> >
> > if (chip->info->ops->port_set_jumbo_size)
> > ...
> > else
> >     return chip->info->max_frame_size - VLAN_ETH_HLEN -
> >       EDSA_HLEN - ETH_FCS_LEN;
> >
> >
> > Or shall I put WARN_ON_ONCE to the mv88e6xxx_probe() function?
> >
> >
> > The above approach is contrary to one proposed by Alexander, who
> > wanted to improve the defensive approach in this driver (to avoid
> > situation where the max_frame_size callback is not defined and
> > max_frame_size member of *_info struct is not added by developer).
> >
> > Which approach is the recommended one for this driver?
>
> Is there any decision regarding the preferred approach to rewrite this
> code?

I would defer to what Andrew proposed since he has more experience
with the DSA code than I do.

Thanks,

- Alex

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size
  2023-01-05 16:13       ` Alexander Duyck
@ 2023-01-05 17:44         ` Lukasz Majewski
  0 siblings, 0 replies; 12+ messages in thread
From: Lukasz Majewski @ 2023-01-05 17:44 UTC (permalink / raw)
  To: Alexander Duyck
  Cc: Andrew Lunn, Vladimir Oltean, Eric Dumazet, Florian Fainelli,
	David S. Miller, Jakub Kicinski, Russell King, Paolo Abeni,
	netdev, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 2331 bytes --]

Hi Alexander

> On Thu, Jan 5, 2023 at 2:37 AM Lukasz Majewski <lukma@denx.de> wrote:
> >
> > Hi Andrew, Alexander,
> >  
> > > Hi Andrew,
> > >  
> > > > > @@ -3548,7 +3548,9 @@ static int mv88e6xxx_get_max_mtu(struct
> > > > > dsa_switch *ds, int port) if
> > > > > (chip->info->ops->port_set_jumbo_size) return 10240 -
> > > > > VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN; else if
> > > > > (chip->info->ops->set_max_frame_size)
> > > > > -         return 1632 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > > > ETH_FCS_LEN;
> > > > > +         return (max_t(int, chip->info->max_frame_size,
> > > > > 1632)
> > > > > +                 - VLAN_ETH_HLEN - EDSA_HLEN -
> > > > > ETH_FCS_LEN); +
> > > > >   return 1522 - VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
> > > > >  
> > > >
> > > > I would also prefer if all this if/else logic is removed, and
> > > > the code simply returned chip->info->max_frame_size -
> > > > VLAN_ETH_HLEN - EDSA_HLEN - ETH_FCS_LEN;
> > > >  
> > >
> > > So then the mv88e6xxx_get_max_mtu shall look like:
> > >
> > > WARN_ON_ONCE(!chip->info->max_frame_size)
> > >
> > > if (chip->info->ops->port_set_jumbo_size)
> > > ...
> > > else
> > >     return chip->info->max_frame_size - VLAN_ETH_HLEN -
> > >       EDSA_HLEN - ETH_FCS_LEN;
> > >
> > >
> > > Or shall I put WARN_ON_ONCE to the mv88e6xxx_probe() function?
> > >
> > >
> > > The above approach is contrary to one proposed by Alexander, who
> > > wanted to improve the defensive approach in this driver (to avoid
> > > situation where the max_frame_size callback is not defined and
> > > max_frame_size member of *_info struct is not added by developer).
> > >
> > > Which approach is the recommended one for this driver?  
> >
> > Is there any decision regarding the preferred approach to rewrite
> > this code?  
> 
> I would defer to what Andrew proposed since he has more experience
> with the DSA code than I do.
> 

Ok, then I will prepare v4 according to Andrew suggestions.

Thanks for the update :-)

> Thanks,
> 
> - Alex




Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread

end of thread, other threads:[~2023-01-05 17:44 UTC | newest]

Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-01-02 15:02 [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Lukasz Majewski
2023-01-02 15:02 ` [PATCH v3 2/3] net: dsa: mv88e6xxx: add support for MV88E6020 switch Lukasz Majewski
2023-01-02 20:00   ` Andrew Lunn
2023-01-03  8:46     ` Lukasz Majewski
2023-01-02 15:02 ` [PATCH v3 3/3] net: dsa: mv88e6xxx: add support for MV88E6071 switch Lukasz Majewski
2023-01-02 19:58 ` [PATCH v3 1/3] dsa: marvell: Provide per device information about max frame size Andrew Lunn
2023-01-02 20:29 ` Andrew Lunn
2023-01-03  9:02   ` Lukasz Majewski
2023-01-03 12:54     ` Andrew Lunn
2023-01-05 10:37     ` Lukasz Majewski
2023-01-05 16:13       ` Alexander Duyck
2023-01-05 17:44         ` Lukasz Majewski

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).