Re: xhci data toggle out of sync

On 4/18/22 18:44, Mahesh Vardhamanaiah wrote:
> Hi all,
> 
> 
>   Need your help regarding an urgent issue where we are observing an issue with Freebsd stable/12. The data toggles for BOT endpoint
>   (DATA0/DATA1) are out of sync with respect to EP and  the  system experiences the  READ(10) errors. The READ(10) error recovers within
>   couple of retries most of the times but few cases we have observed that the read retries gets exhausted  and  system moves to unusable state
> (continuous g_vfs_done() errors) . We are using Junos but  the  xhci driver etc.. are all pristine stable 12 drivers no Juniper specific  changes.
> This issue was never observed with Linux kernel 5.4.2 on the same HW suspecting if it's a xhci driver error recovery issue.
> 
>   Errors Seen on console
> 
>   (da0:umass-sim0:0:0:0): READ(10). CDB: 28 00 00 28 cf 28 00 00 40 00
> 
>   (da0:umass-sim0:0:0:0): CAM status: CCB request completed with an error
> 
>   (da0:umass-sim0:0:0:0): Retrying command, 3 more tries remain
> 
>   (da0:umass-sim0:0:0:0): READ(10). CDB: 28 00 00 28 cf 28 00 00 40 00
> 
>   (da0:umass-sim0:0:0:0): CAM status: CCB request completed with an error
> 
>   (da0:umass-sim0:0:0:0): Retrying command, 2 more tries remain
> 
>   FreeBSD/arm (Amnesiac) (ttyu0)
> 
>   login:
> 
> 
> Please let us know is there a place holder to upload the traces and provide a link here ?
> Thanks,
> Mahesh

Hi Mahesh,

Do you have wire recording using Ellisys or the Beagle, showing which 
side is at fault of the data toggle error?

Is this issue reproducible using the EHCI HC driver?

Is the device connected at USB 3.x or USB 2.0 speed?

--HPS