Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <d0642a21-4675-76f7-a470-b0dc1ef37cec@gmail.com>
Date:   Wed, 5 Apr 2023 23:19:25 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.9.0
Subject: Re: [BUG] staging: rtl8192e: oops occurs when finding hardware
 rtl8192se
To:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc:     linux-staging@lists.linux.dev, linux-kernel@vger.kernel.org
References: <39dc735c-fd6d-e405-856c-788a52704d63@gmail.com>
 <2023040521-angelic-emptier-1367@gregkh>
Content-Language: en-US
From:   Philipp Hortmann <philipp.g.hortmann@gmail.com>
In-Reply-To: <2023040521-angelic-emptier-1367@gregkh>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 4/5/23 16:37, Greg Kroah-Hartman wrote:
> On Sun, Apr 02, 2023 at 05:00:13PM +0200, Philipp Hortmann wrote:
>> Hi,
>>
>> when I use the hardware rtl8192se the driver
>> drivers/staging/rtl8192e/rtl8192e/r8192e_pci.ko detects that it should not
>> run on this hardware and aborts.
>> But when the driver is freeing the resources an oops occures. Find oops at
>> the end of this Email.
>>
>> When I comment out the following lines those errors disappear:
>> cancel_delayed_work_sync(&ieee->hw_wakeup_wq);
>> cancel_delayed_work_sync(&ieee->hw_sleep_wq);
>> cancel_work_sync(&ieee->ips_leave_wq);
>>
>> When I do an init before the cancel:
>> INIT_DELAYED_WORK(&priv->rtllib->hw_wakeup_wq, (void *)rtl92e_hw_wakeup_wq);
>> The oops are gone as well.
>>
>> When I use cancel_delayed_work() instead of cancel_delayed_work_sync() it
>> also works.
>>
>> Can somebody give me a hint what the expected way is to solve this?
> 
> Is this a new thing, or has it always been there?
I would need to check in several places.
It seems to me that it was an issue in kernel 4.0 already.


> 
> Why is the driver loading if you don't have hardware for it?  Or are you
> manually loading it?
Reason why they two drivers are loaded is that realtek managed to have
two different devices with the same vid and did.

from rtl8192se/sw.c:
static const struct pci_device_id rtl92se_pci_ids[] = {
	{RTL_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8192, rtl92se_hal_cfg)},
	...

from rtl8192e/rtl8192e/rtl_core.c
static struct pci_device_id rtl8192_pci_id_tbl[] = {
	{RTL_PCI_DEVICE(0x10ec, 0x8192, rtl819xp_ops)},

#define PCI_VENDOR_ID_REALTEK		0x10ec

The drivers are loaded both and find out with the pci revision_id if 
they are in charge.

> 
> thanks,
> 
> greg k-h

May be I should mention that I do have an rtl8192se. But it does use a 
different device id.

I am simulating the driver that it is rtl8192se hardware by changing the 
revision_id.


Thanks for your response.

Bye Philipp