Re: [PATCH] brcmfmac: stop watchdog before detach and free everything

[Arend van Spriel <arend.vanspriel@broadcom.com>]

On 5/28/2018 5:33 PM, Michael Nazzareno Trimarchi wrote:
> Hi Andy
>
> The problem seems really easy to solve:
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> index 412a05b..ba60b151 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> @@ -4227,13 +4227,7 @@ struct brcmf_sdio *brcmf_sdio_probe(struct
> brcmf_sdio_dev *sdiodev)
>          timer_setup(&bus->timer, brcmf_sdio_watchdog, 0);
>          /* Initialize watchdog thread */
>          init_completion(&bus->watchdog_wait);
> -       bus->watchdog_tsk = kthread_run(brcmf_sdio_watchdog_thread,
> -                                       bus, "brcmf_wdog/%s",
> -                                       dev_name(&sdiodev->func1->dev));
> -       if (IS_ERR(bus->watchdog_tsk)) {
> -               pr_warn("brcmf_watchdog thread failed to start\n");
> -               bus->watchdog_tsk = NULL;
> -       }
> +
>          /* Initialize DPC thread */
>          bus->dpc_triggered = false;
>          bus->dpc_running = false;
> @@ -4281,6 +4275,14 @@ struct brcmf_sdio *brcmf_sdio_probe(struct
> brcmf_sdio_dev *sdiodev)
>                  goto fail;
>          }
>
> +       bus->watchdog_tsk = kthread_run(brcmf_sdio_watchdog_thread,
> +                                       bus, "brcmf_wdog/%s",
> +                                       dev_name(&sdiodev->func1->dev));
> +       if (IS_ERR(bus->watchdog_tsk)) {
> +               pr_warn("brcmf_watchdog thread failed to start\n");
> +               bus->watchdog_tsk = NULL;
> +       }
> +
>          return bus;

Hi Michael,

That makes no sense. Or are you saying the function 
brcmf_fw_get_firmwares() fails for you? Oh wait, you mentioned the 
driver was built-in, right? The above change does not solve the issue. 
It just makes it less likely to occur. So I think your initial fix is 
the best solution for this.

Regards,
Arend