Re: [PATCH v3] mwifiex: fix sleep in atomic context bugs caused by dev_coredumpv

[Kalle Valo]

duoming@xxxxxxxxxx writes:

> Hello maintainers,
>
> Thank you for your time and suggestions! 
>
>> > There are sleep in atomic context bugs when uploading device dump
>> > data in mwifiex. The root cause is that dev_coredumpv could not
>> > be used in atomic contexts, because it calls dev_set_name which
>> > include operations that may sleep. The call tree shows execution
>> > paths that could lead to bugs:
>> >
>> >    (Interrupt context)
>> > fw_dump_timer_fn
>> >   mwifiex_upload_device_dump
>> >     dev_coredumpv(..., GFP_KERNEL)
>> >       dev_coredumpm()
>> >         kzalloc(sizeof(*devcd), gfp); //may sleep
>> >         dev_set_name
>> >           kobject_set_name_vargs
>> >             kvasprintf_const(GFP_KERNEL, ...); //may sleep
>> >             kstrdup(s, GFP_KERNEL); //may sleep
>> >
>> > In order to let dev_coredumpv support atomic contexts, this patch
>> > changes the gfp_t parameter of kvasprintf_const and kstrdup in
>> > kobject_set_name_vargs from GFP_KERNEL to GFP_ATOMIC. What's more,
>> > In order to mitigate bug, this patch changes the gfp_t parameter
>> > of dev_coredumpv from GFP_KERNEL to GFP_ATOMIC.
>> 
>> vmalloc in atomic context?
>> 
>> Not only does dev_coredumpm set a device name dev_coredumpm creates an
>> entire device to hold the device dump.
>> 
>> My sense is that either dev_coredumpm needs to be rebuilt on a
>> completely different principle that does not need a device to hold the
>> coredump (so that it can be called from interrupt context) or that
>> dev_coredumpm should never be called in an context that can not sleep.
>
> The following solution removes the gfp_t parameter of dev_coredumpv(), 
> dev_coredumpm() and dev_coredumpsg() and change the gfp_t parameter of 
> kzalloc() in dev_coredumpm() to GFP_KERNEL, in order to show that these 
> functions can not be used in atomic context.
>
> What's more, I move the operations that may sleep into a work item and use
> schedule_work() to call a kernel thread to do the operations that may sleep.
>

[...]

> --- a/drivers/net/wireless/marvell/mwifiex/init.c
> +++ b/drivers/net/wireless/marvell/mwifiex/init.c
> @@ -63,11 +63,19 @@ static void wakeup_timer_fn(struct timer_list *t)
>  		adapter->if_ops.card_reset(adapter);
>  }
>  
> +static void fw_dump_work(struct work_struct *work)
> +{
> +	struct mwifiex_adapter *adapter =
> +		container_of(work, struct mwifiex_adapter, devdump_work);
> +
> +	mwifiex_upload_device_dump(adapter);
> +}
> +
>  static void fw_dump_timer_fn(struct timer_list *t)
>  {
>  	struct mwifiex_adapter *adapter = from_timer(adapter, t, devdump_timer);
>  
> -	mwifiex_upload_device_dump(adapter);
> +	schedule_work(&adapter->devdump_work);
>  }
>  
>  /*
> @@ -321,6 +329,7 @@ static void mwifiex_init_adapter(struct mwifiex_adapter *adapter)
>  	adapter->active_scan_triggered = false;
>  	timer_setup(&adapter->wakeup_timer, wakeup_timer_fn, 0);
>  	adapter->devdump_len = 0;
> +	INIT_WORK(&adapter->devdump_work, fw_dump_work);
>  	timer_setup(&adapter->devdump_timer, fw_dump_timer_fn, 0);
>  }
>  
> @@ -401,6 +410,7 @@ mwifiex_adapter_cleanup(struct mwifiex_adapter *adapter)
>  {
>  	del_timer(&adapter->wakeup_timer);
>  	del_timer_sync(&adapter->devdump_timer);
> +	cancel_work_sync(&adapter->devdump_work);
>  	mwifiex_cancel_all_pending_cmd(adapter);
>  	wake_up_interruptible(&adapter->cmd_wait_q.wait);
>  	wake_up_interruptible(&adapter->hs_activate_wait_q);

In this patch please only do the API change in mwifiex. The change to
using a workqueue needs to be in separate patch so it can be properly
tested. I don't want a change like that going to the kernel without
testing on a real device.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches