Re: [lwip-users] LWIP never tries to send a packet?

Hi Chris,

If you are able to read/write PHY registers that means that the SMI interface is up and running.

Before digging into LwIP make sure your MII/RMII and DMA are configured correctly with the

driver. I am working with LwIP for last 7 years and I almost had no problems with LwIP.

If you are using your own hardware and not some EVB make sure that all your CPU pin-outs

are properly defined. Make sure that you are working in correct PHY mode (RMII or MII),

check that you defined all the hardware clocks and DMA...

I run LwIP on ST9, STM32F2/4 and now 7 and so far I am happy :-)

Hope that helped a bit.

BR,

Noam.

From: lwip-users <lwip-users-bounces+address@hidden> on behalf of Will Wykeham <address@hidden>
Sent: Thursday, January 18, 2018 7:48 PM
To: Mailing list for lwIP users
Subject: Re: [lwip-users] LWIP never tries to send a packet?

I had a similar looking issue last week bringing up lwip for the first time. In my case it was because I hadn't set both the interface and the link as being 'up'.

Will

On 18 Jan 2018 5:46 pm, "Chris Seto" <address@hidden> wrote:

Hi Noam, Thanks for the reply.

A few of the defs seemed to be wrong for my PHY. Specifically, the bits in MISR and BMSR. Slight differences. The bits existed, but not in the order the HAL config files indicated them to.I copied my config to the bottom of this message.

That said, I don't think that the PHY is the issue for this problem, since it does seem like the PHY is RX'ing, but the bigger issue I'm seeing is that I'm not even getting debug output from LwIP other than "etharp_timer" once every second. I can confirm that I can talk to the PHY and that it's up and running ok. When I mentioned that I can guarantee the link up after low_level_init runs, what I meant is that I can actually watch the registers change, and I don't let low_level_init run until the link is up, as per the PHY SR.

I'm guessing the issue here is some config value in LwIP.

I've also tried allowing a full hanging ASSERT, the code never ASSERTs.

I've played around with the debug message config, and I simply cannot tease any other messages out of the stack, even though at least DHCP is started. I do know that printf works called from LwIP, because I can at least see the etharp_timer message.

Thanks,

Chris

For the TLK110, here are my settings:

// TLK110 datasheet, page 5

#define PHY_ADDR 0x01

// TLK110 Registers

#define PHY_MISR1 0x0012

#define PHY_BMSR 0x0001

#define PHY_BCR ((uint16_t)0x00) /*!< Transceiver Basic Control Register */

#define PHY_BSR ((uint16_t)0x01) /*!< Transceiver Basic Status Register */

#define PHY_RESET ((uint16_t)(1 << 15)) /*!< PHY Reset */

#define PHY_LOOPBACK ((uint16_t)(1 << 14)) /*!< Select loop-back mode */

#define PHY_AUTONEGOTIATION ((uint16_t)(1 << 12)) /*!< Enable auto-negotiation function */

#define PHY_RESTART_AUTONEGOTIATION ((uint16_t)(1 << 9)) /*!< Restart auto-negotiation function */

#define PHY_POWERDOWN ((uint16_t)(1 << 11)) /*!< Select the power down mode */

#define PHY_ISOLATE ((uint16_t)(1 << 10)) /*!< Isolate PHY from MII */

#define PHY_AUTONEGO_COMPLETE ((uint16_t)(1 << 5)) /*!< Auto-Negotiation process completed */

#define PHY_LINKED_STATUS ((uint16_t)(1 << 2)) /*!< Valid link established */

#define PHY_JABBER_DETECTION ((uint16_t)(1 << 1)) /*!< Jabber condition detected */

#define PHY_SR ((uint16_t)0x10) /*!< PHY special control/ status register Offset */

#define PHY_SPEED_STATUS ((uint16_t)(1 << 1)) /*!< PHY Speed mask */

#define PHY_DUPLEX_STATUS ((uint16_t)(1 << 2)) /*!< PHY Duplex mask */

On Thu, Jan 18, 2018 at 11:31 AM, Noam Weissman <address@hidden> wrote:

Hi Chris,

I am not working with ST HAL, rather with the older SPL (standard peripheral library).

I do not know why you needed to change the PHY driver as all the standard PHY’s that are IEEE

compatible will work the same.

The ST driver has two portions:

1. MDIO/MDC connectivity (SMI)

2. RMII/MII data transfer etc…

Via SMI you control the PHY and or make queries. Meaning you can check if you have a link,

what is the speed etc…

Regarding MII/RMII you need to make sure that you defined the driver to work in the same mode

that the PHY is strapped. If the PHY is defined to work in RMII but your driver is defined to work in MII,

it will not work !

I suggest going over ST examples first and then try to see why you code is not working for you.

BT,

Noam.

From: lwip-users [mailto:lwip-users-bounces+noam=address@hidden] On Behalf Of Chris Seto
Sent: Thursday, January 18, 2018 6:47 PM
To: address@hidden
Subject: [lwip-users] LWIP never tries to send a packet?

Hi,

I'm using LwIP 2.0 running on an STM32F4 with a TI TLK110 ethernet PHY. I've written the driver for the PHY and corrected the definitions within the STM32 HAL such that the PHY is initialized correctly. When low_level_init() returns, the link is guaranteed physically up.

I'm having an issue where it doesn't seem like LwIP is ever trying to send packets. I put printf("RX\r\n"); and printf("TX\r\n"); calls in my ethernetif.c handlers, and while RX periodically is called, TX is /never/ called. Even if I try to open a TCP socket, and even if I try dhcp_start().

I'm sure this is some kind of minor config issue. Any advice?

Debugging is enabled, and I see a bunch of "etharp_timer" at about 2hz or so, but nothing else. Shouldn't I at least expect to see a DHCP discover packet sent?

Code here:

ethernetif.c

https://pastebin.com/eXGyDWJX

lwipopts.h

https://pastebin.com/ZTmmG43P

Networking.c

https://pastebin.com/20DN07dy

Main loop:

while (1)

{

HAL_Delay(50);

LedToggle(LED_1);

LedToggle(LED_2);

NetworkingUpdate();

}

Thanks!

_______________________________________________
lwip-users mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/lwip-users

_______________________________________________
lwip-users mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/lwip-users

From:	Noam Weissman
Subject:	Re: [lwip-users] LWIP never tries to send a packet?
Date:	Thu, 18 Jan 2018 18:30:25 +0000