Re: [EXT] Re: RIP: 0010:qede_load+0x128d/0x13b0 [qede] - 5.19.0

[Jakub Kicinski]

On Thu, 18 Aug 2022 17:55:28 +0000 Manish Chopra wrote:
> 3. You mentioned about commit 3aa6bce9af0e ("net: watchdog: hold device global xmit lock during tx disable")

FWIW that was just my guess based on the stack trace, Bruno posted the
stacktraces with line numbers decoded here:

https://lore.kernel.org/all/CA+QYu4ob4cbh3Vnh9DWgaPpyw8nTLFG__TbBpBsYg1tWJPxygg@xxxxxxxxxxxxxx/

>     Do you mean issue started surfacing only after this commit ? Driver calls netif_tx_disable() from these two relevant contexts -
> 
>     a. One in ndo_stop() flow 
> 
>       	        /* Close OS Tx */
>         netif_tx_disable(edev->ndev);
>         netif_carrier_off(edev->ndev);
>    
>    b. Other in LINK events handling from the hard IRQ context
> 
>         DP_NOTICE(edev, "Link is down\n");
>         netif_tx_disable(edev->ndev);
>         netif_carrier_off(edev->ndev);