Received-SPF: pass (google.com: domain of linux-wireless+bounces-6132-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Sender: Larry Finger <larry.finger@gmail.com>
Message-ID: <8909f667-96e2-484d-a0fd-a20847e42d27@lwfinger.net>
Date: Wed, 10 Apr 2024 14:03:58 -0500
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Content-Language: en-US
To: linux-wireless <linux-wireless@vger.kernel.org>
From: Larry Finger <Larry.Finger@lwfinger.net>
Subject: Need Help with SDIO device
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi,

I am helping a user of the RTW8821CS device, which fails on startup. The device 
is composed of an SDIO wifi unit, and a UART-based Bluetooth section. Our 
discussion is at https://github.com/lwfinger/rtw88/issues/187.

What an msleep(150) was added just before the SDIO probe routine proceeds, the 
success rate was about 60%. Examination of the dmesg logs of 10 runs showed that 
when the BT section completed its initialization before the SDIO probe routine 
was executed, wifi started without error. If the SDIO routine was entered before 
the BT driver completed initialization, wifi failed with error -110 for SDIO 
read/write operations.

What mechanism is available to ensure that one driver finishes initialization 
before another starts? Is there anything more sophisticated than an mdelay(500) 
at the start of the SDIO probe routine?

Thanks,

Larry