Mathias： Thank you very much! > This issue was resolved in 5.19 kernel with patch: > 25355e046d29 xhci: use generic command timer for stop endpoint commands. > After this patch the timer is started when xHC start processing the command, not when driver queues it. > I strongly recommend trying a more recent kernel where all these issues are fixed already. I see 5.19 kernel has done what you say. Because the kernel version of our company is 4.65 , I need to put the patch to our 4.65 kernel we used. Today compared to the latest kernel is 6.7, I found there are a lot of changes for usb module between kernel 4.65 and 6.7. Could you provide me for the issues patch. -----邮件原件----- 发件人: Mathias Nyman [mailto:mathias.nyman@xxxxxxxxxxxxxxx] 发送时间: 2024年1月24日 21:49 收件人: zhangzhansheng (操作系统开发部部门, RD) <zhang.zhansheng@xxxxxxx>; Ladislav Michl <oss-lists@xxxxxxxxx> 抄送: linux-usb@xxxxxxxxxxxxxxx; Sneeker Yeh <sneeker.yeh@xxxxxxxxx>; wangxiaoqing (操作系统开发部/OM, RD) <wangxiaoqing@xxxxxxx>; xinhaining (操作系统开发部/OM, RD) <xinhaining@xxxxxxx> 主题: Re: 答复: 答复: [Consulting about: xHCI host dies on device unplug] On 24.1.2024 13.42, Zhangzhansheng wrote: > Mathias： > > > In regard to the question “What the reason of the usb host controller handing the stop endpoint command timeout maybe two a) and b).”Just to add the follow one : > When the problem arising, the function of “xhci_halt “ operated the host controller successfully. So I think the usb host controller should be ok. > > So, I gussess that the usb host controller has no enough time to handle with a large number of xhci command. > This issue was resolved in 5.19 kernel with patch: 25355e046d29 xhci: use generic command timer for stop endpoint commands. After this patch the timer is started when xHC start processing the command, not when driver queues it. I strongly recommend trying a more recent kernel where all these issues are fixed already. Thanks Mathias
