[CentOS-virt] USB card reader causing qemu-kvm SEGV's

Sat Mar 18 02:57:03 UTC 2017

It seem more likely related to specific USB driver your have installed and not compatible with your existing kernel.

Have you try other USB device and compare?

Xlord

From: CentOS-virt [mailto:centos-virt-bounces at centos.org] On Behalf Of Sandro Bonazzola
Sent: Friday, March 17, 2017 4:06 PM
To: Discussion about the virtualization on CentOS <centos-virt at centos.org>; Paolo Bonzini <pbonzini at redhat.com>; Miroslav Rezanina <mrezanin at redhat.com>
Subject: Re: [CentOS-virt] USB card reader causing qemu-kvm SEGV's

Adding Paolo and Miroslav.

On Sun, Mar 12, 2017 at 11:41 PM, Philip Prindeville <philipp_subx at redfish-solutions.com <mailto:philipp_subx at redfish-solutions.com> > wrote:

Hi.

I have a Supermicro 5018D-FN4T (Xeon D-1541 based SBC) that I use for virtualization.  I’m running Centos 7.3 on it (updated), with the CentOS-QEMU-EV.repo repository as the source for virtualization packages.

I run an Ubuntu 16.04-2 guest VM on it, which is ordinary enough.  What’s perhaps less ordinary is that I’ve attached a Lexar Media, Inc. “Lexar Professional Workflow CR1 CFast 2.0 USB 3.0 Reader” (LRWCR1TBNA) as well as a WEme Superspeed Aluminum USB 3.0 Multi-in-1 Card Reader … for CF/SD/TF Micro SD.  Here’s the relevant lsusb output:

Bus 004 Device 004: ID 0bda:0309 Realtek Semiconductor Corp.
Bus 004 Device 025: ID 05dc:ba04 Lexar Media, Inc.

and both are hanging off of an Amazon Basics 4-port USB 3.0 hub:

Bus 004 Device 002: ID 2109:8110 VIA Labs, Inc. Hub

The Ubuntu VM I use for doing buildroot cross builds of an embedded Linux environment, which I then burn the image of onto a CFast card, having attached the Lexar card reader to the guest VM.

Problem is that the card reader is extremely dodgy, and sometimes I guess a flurry of messages on the virtualization host complaining about the device:

Mar 11 15:52:46 kvm1 kernel: usb 4-2.2: Disable of device-initiated U1 failed.
Mar 11 15:52:46 kvm1 kernel: usb 4-2.2: Disable of device-initiated U2 failed.
Mar 11 15:52:46 kvm1 kernel: usb 4-2.2: Set SEL for device-initiated U1 failed.
Mar 11 15:52:46 kvm1 kernel: usb 4-2.2: Set SEL for device-initiated U2 failed.
Mar 11 15:52:47 kvm1 kernel: usb 4-2.2: reset SuperSpeed USB device number 25 using xhci_hcd
Mar 11 15:52:47 kvm1 kernel: scsi host13: uas
Mar 11 15:52:47 kvm1 kernel: scsi 13:0:0:0: Direct-Access     Lexar    WorkflowCR1      0    PQ: 0 ANSI: 6
Mar 11 15:52:47 kvm1 kernel: sd 13:0:0:0: [sda] 30198988 512-byte logical blocks: (15.4 GB/14.3 GiB)
Mar 11 15:52:47 kvm1 kernel: sd 13:0:0:0: Attached scsi generic sg0 type 0
Mar 11 15:52:47 kvm1 kernel: sd 13:0:0:0: [sda] Write Protect is off
Mar 11 15:52:47 kvm1 kernel: sd 13:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Mar 11 15:52:47 kvm1 kernel: sda: sda1 sda2
Mar 11 15:52:47 kvm1 kernel: sd 13:0:0:0: [sda] Attached SCSI removable disk
Mar 11 15:52:48 kvm1 kernel: usb 4-2.2: usbfs: process 4354 (CPU 3/KVM) did not claim interface 0 before use
Mar 11 15:52:48 kvm1 kernel: usb 4-2.2: usbfs: process 4354 (CPU 3/KVM) did not claim interface 0 before use
Mar 11 15:53:21 kvm1 kernel: usb 4-2.2: usbfs: process 4355 (CPU 4/KVM) did not claim interface 0 before use
Mar 11 15:53:21 kvm1 kernel: usb 4-2.2: reset SuperSpeed USB device number 25 using xhci_hcd
Mar 11 15:53:21 kvm1 kernel: usb 4-2.2: reset SuperSpeed USB device number 25 using xhci_hcd

but lately it just kills my VM when I’m writing to the device.  So from annoying to a serious potential for losing work.

I run “yum update” a couple of times a week but now I’m thinking that was a mistake, since this has gone from being ultra-reliable to being a hazard.

Is there a workaround?  Would I be better off plugging the USB 3.0 card-reader into a USB 2.0 hub and settling for slower throughput… but not crashing?

Below, for instance, I was doing a “dd if=xyzzy of=/dev/sda bs=1M” where /dev/sda is the device assigned to the USB card storage device.

Thanks,

-Philip

URL for ABRT report:  https://da.gd/emJO

_______________________________________________
CentOS-virt mailing list
CentOS-virt at centos.org <mailto:CentOS-virt at centos.org> 
https://lists.centos.org/mailman/listinfo/centos-virt

-- 

Sandro Bonazzola
Better technology. Faster innovation. Powered by community collaboration.
See how it works at redhat.com <http://redhat.com> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.centos.org/pipermail/centos-virt/attachments/20170318/3b73eacb/attachment.html>