Hi: Sorry again, The last email was wrong, fixed my email to not send html email and respond again. The drivers manufacturers advertised high speed support platform is Windows,MacOSX, not include linux 4.9 , this bug will affect the users normal use, run iozone is just one of the behaviors that trigger UAS crash ,We will handle that slow issue. Thanks! zhongling0719@xxxxxxx From: Greg KH Date: 2022-09-01 16:37 To: zenghongling CC: stern; linux-usb; usb-storage; zhongling0719 Subject: Re: [PATCH v4] uas: add no-uas quirk for Thinkplus and Hiksemi usb-storage On Thu, Sep 01, 2022 at 03:00:53PM +0800, zenghongling wrote: > UAS:if ignore uas feature for these drivers can fix the not working and > output error message bug. > > The UAS mode of Thinkplus and Hiksemi is reported to fail to work on > several platforms with the following error message,I tested these USB > disks on other architecture platforms as arm/loongson for different xHCI > controller,the same error occurred: > > [ 39.702439] xhci_hcd 0000:0c:00.3: ERROR Transfer event for disabled > endpoint or incorrect stream ring > [ 39.702442] xhci_hcd 0000:0c:00.3: @000000026c61f810 00000000 00000000 > 1b000000 05038000 > > [ 592.490369][ 1] xhci_hcd 0000:0c:00.3: Assuming host is dying, halting host. > [ 592.518442][ 2] sd 8:0:0:0: [sda] tag#17 uas_eh_abort_handler 0 uas-tag 18 > inflight: CMD > [ 592.527575][ 2] sd 8:0:0:0: [sda] tag#17 CDB: Write(10) 2a 00 03 6f 88 00 00 > 04 00 00 > [ 592.536330][ 2] sd 8:0:0:0: [sda] tag#0 uas_eh_abort_handler 0 uas-tag 1 > inflight: CMD > [ 592.545266][ 2] sd 8:0:0:0: [sda] tag#0 CDB: Write(10) 2a 00 07 44 1a 88 00 > 00 08 00 > > And when running iozone will disconnect from the USB controller, then after > re-connecting the device will be offlined and not working at all. > > We changed a lot of USB devices and contact relevant manufacturers to > confirm the USB disk is ok. I do not understand why there are failures on Linux, yet not on other operating systems. If you run iozone on OS-X, does the same errors happen? How about a storage stress-test on Windows? Why is Linux going to have to make these devices go very very slow and what will happen to the users that now can not operate their devices at the advertised speeds? Who will handle that support issue? thanks, greg k-h