Skip to contentSkip to conversation

igc: Add netdev

PATCH Superseded Sasha Neftin 3 participants 4 messages Aug 23, 2018 – Aug 28, 2018

Subsystems (2)

v6 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

PATCH v6 3/11 igc: Add netdev

Now that we have the ability to configure the basic settings on the device we can start allocating and configuring a netdev for the interface.

Sasha Neftin (v2):
add description

Sasha Neftin (v3):
minor cosmetic changes

Sasha Neftin (v4):
address comments
fix code indentation
replace e1000_ prefix with igc_ prefix
optimize pcie access

Sasha Neftin (v5):
change return value of igc_up() method to void

Sasha Neftin (v6):
fix define of MAX_Q_VECTORS
fix code indentation
minor cosmetic changes

---
 drivers/net/ethernet/intel/igc/igc.h         |  48 ++
 drivers/net/ethernet/intel/igc/igc_defines.h |  15 +
 drivers/net/ethernet/intel/igc/igc_hw.h      |   1 +
 drivers/net/ethernet/intel/igc/igc_main.c    | 475 ++++++++++++++++++-
 4 files changed, 537 insertions(+), 2 deletions(-)
drivers/net/ethernet/intel/igc/igc.h +48 -0
index 481b2ee694fa..cccc02b999c1 100644
@@ -28,15 +28,63 @@
 extern char igc_driver_name[];
 extern char igc_driver_version[];
 
+/* Transmit and receive queues */
+#define IGC_MAX_RX_QUEUES		4
+#define IGC_MAX_TX_QUEUES		4
+
+#define MAX_Q_VECTORS			8
+#define MAX_STD_JUMBO_FRAME_SIZE	9216
+
+enum igc_state_t {
+	__IGC_TESTING,
+	__IGC_RESETTING,
+	__IGC_DOWN,
+	__IGC_PTP_TX_IN_PROGRESS,
+};
+
+struct igc_q_vector {
+	struct igc_adapter *adapter;    /* backlink */
+
+	struct napi_struct napi;
+};
+
+struct igc_mac_addr {
+	u8 addr[ETH_ALEN];
+	u8 queue;
+	u8 state; /* bitmask */
+};
+
+#define IGC_MAC_STATE_DEFAULT   0x1
+#define IGC_MAC_STATE_MODIFIED  0x2
+#define IGC_MAC_STATE_IN_USE    0x4
+
 /* Board specific private data structure */
 struct igc_adapter {
+	struct net_device *netdev;
+
+	unsigned long state;
+	unsigned int flags;
+	unsigned int num_q_vectors;
+	u16 link_speed;
+	u16 link_duplex;
+
+	u8 port_num;
+
 	u8 __iomem *io_addr;
+	struct work_struct watchdog_task;
+
+	int msg_enable;
+	u32 max_frame_size;
 
 	/* OS defined structs */
 	struct pci_dev *pdev;
 
 	/* structs defined in igc_hw.h */
 	struct igc_hw hw;
+
+	struct igc_q_vector *q_vector[MAX_Q_VECTORS];
+
+	struct igc_mac_addr *mac_table;
 };
 
 #endif /* _IGC_H_ */
drivers/net/ethernet/intel/igc/igc_defines.h +15 -0
index 05bdba94b045..3817cbc0a6ad 100644
@@ -4,6 +4,8 @@
 #ifndef _IGC_DEFINES_H_
 #define _IGC_DEFINES_H_
 
+#define IGC_CTRL_EXT_DRV_LOAD         0x10000000 /* Drv loaded bit for FW */
+
 /* PCI Function Mask */
 #define IGC_STATUS_FUNC_MASK			0x0000000C
 
@@ -11,6 +13,16 @@
 #define PCIE_DEVICE_CONTROL2		0x28
 #define PCIE_DEVICE_CONTROL2_16ms	0x0005
 
+/* Receive Address
+ * Number of high/low register pairs in the RAR. The RAR (Receive Address
+ * Registers) holds the directed and multicast addresses that we monitor.
+ * Technically, we have 16 spots.  However, we reserve one of these spots
+ * (RAR[15]) for our directed address used by controllers with
+ * manageability enabled, allowing us room for 15 multicast addresses.
+ */
+#define IGC_RAH_AV		0x80000000 /* Receive descriptor valid */
+#define IGC_RAH_POOL_1		0x00040000
+
 /* Error Codes */
 #define IGC_SUCCESS			0
 #define IGC_ERR_NVM			1
@@ -20,6 +32,9 @@
 #define IGC_ERR_MAC_INIT		5
 #define IGC_ERR_RESET			9
 
+/* PBA constants */
+#define IGC_PBA_34K		0x0022
+
 /* Device Status */
 #define IGC_STATUS_FD		0x00000001      /* Full duplex.0=half,1=full */
 #define IGC_STATUS_LU		0x00000002      /* Link up.0=no,1=link */
drivers/net/ethernet/intel/igc/igc_hw.h +1 -0
index f115ef08f8ac..caed846fdb14 100644
@@ -84,6 +84,7 @@ struct igc_mac_info {
 
 	bool autoneg;
 	bool autoneg_failed;
+	bool get_link_status;
 };
 
 struct igc_bus_info {
drivers/net/ethernet/intel/igc/igc_main.c +473 -2

553 lines not on this page · raw message

Signature · 1 line
2.18.0
#1 lore raw

PATCH v6 3/11 igc: Add netdev

On 8/23/2018 12:05 AM, Sasha Neftin wrote:

[...]

Quoted 1 line
drivers/net/ethernet/intel/igc/igc_main.c +1 -0
+	strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1);

[...]

drivers/net/ethernet/intel/igc/igc_main.c static int igc_probe(struct pci_dev *pdev, · line 578

Quoted 1 line
drivers/net/ethernet/intel/igc/igc_main.c +1 -0
+	strncpy(netdev->name, "eth%d", IFNAMSIZ);

I've mentioned this to you before: why are you setting the netdev->name twice, and to different values? Besides, the OS's udev rules will likely change the name to something else.

drivers/net/ethernet/intel/igc/igc_main.c MODULE_AUTHOR("Intel Corporation, <[email protected]>"); · line 718

Quoted 4 lines
drivers/net/ethernet/intel/igc/igc_main.c +3 -0
+
+module_param(debug, int, 0);
+MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)");
  /* igc_main.c */

It would be nice to have this up above, next to the definition of int debug.

sln

#2 lore raw

PATCH v6 3/11 igc: Add netdev

Quoted 6 lines · Shannon Nelson

On 8/23/2018 12:05 AM, Sasha Neftin wrote:

[...]

Quoted 1 line

+??? strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1);

Quoted 8 lines

[...]

Quoted 1 line

+??? strncpy(netdev->name, "eth%d", IFNAMSIZ);

I've mentioned this to you before: why are you setting the netdev->name twice, and to different values?? Besides, the OS's udev rules will likely change the name to something else.

I believe I already answered in previous review (v3 or v4). Since I remove second settings of netdev->name: "strncpy(netdev->name, "eth%d", IFNAMSIZ)" I got error in probe: "probe of 0000:05:00.0 failed with error -22". I agree with you that we can do some optimization here - I will look later again in this code.

Quoted 9 lines
Quoted 5 lines
drivers/net/ethernet/intel/igc/igc_main.c +3 -0
+
+module_param(debug, int, 0);
+MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)");

? /* igc_main.c */

It would be nice to have this up above, next to the definition of int debug.

so, may be all code start from MODULE_AUTHOR... move to begin of the file. it is not urgent. I will look at other examples. what is benefit from this?

Quoted 2 lines

sln

Sasha

#3 lore raw

PATCH v6 3/11 igc: Add netdev

Quoted 15 lines · Sasha
Quoted 6 lines · Shannon Nelson

On 8/23/2018 12:05 AM, Sasha Neftin wrote:

[...]

Quoted 1 line

+??? strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1);

Quoted 8 lines

[...]

Quoted 1 line

+??? strncpy(netdev->name, "eth%d", IFNAMSIZ);

I've mentioned this to you before: why are you setting the netdev->name twice, and to different values?? Besides, the OS's udev rules will likely change the name to something else.

Quoted 6 lines

I believe I already answered in previous review (v3 or v4). Since I remove second settings of netdev->name: "strncpy(netdev->name, "eth%d", IFNAMSIZ)" I got error in probe: "probe of 0000:05:00.0 failed with error -22". I agree with you that we can do some optimization here - I will look later again in this code.

Quoted 1 line

Thanks that you point me on this. I figure out that 'strncpy(netdev->name, pci_name(pdev)...' line have no point and should be removed. Since v6 should be final I will consult with Jeff how to submit change, probably in different patch. Also, we need run validation cycle.

Quoted 18 lines
Quoted 8 lines
Quoted 4 lines
drivers/net/ethernet/intel/igc/igc_main.c +3 -0
+
+module_param(debug, int, 0);
+MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)");

? /* igc_main.c */

It would be nice to have this up above, next to the definition of int debug.

so, may be all code start from MODULE_AUTHOR... move to begin of the file. it is not urgent. I will look at other examples. what is benefit from this?

Quoted 2 lines

sln

Sasha
_______________________________________________
Intel-wired-lan mailing list
Intel-wired-lan at osuosl.org
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan