kvm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [Bug 196527] New: Terrible KVM USB passthrough performance
@ 2017-07-28 22:42 bugzilla-daemon
  2017-08-10 12:34 ` [Bug 196527] " bugzilla-daemon
  2017-08-10 12:38 ` bugzilla-daemon
  0 siblings, 2 replies; 3+ messages in thread
From: bugzilla-daemon @ 2017-07-28 22:42 UTC (permalink / raw)
  To: kvm

https://bugzilla.kernel.org/show_bug.cgi?id=196527

            Bug ID: 196527
           Summary: Terrible KVM USB passthrough performance
           Product: Virtualization
           Version: unspecified
    Kernel Version: 4.9.32
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: kvm
          Assignee: virtualization_kvm@kernel-bugs.osdl.org
          Reporter: nazar@mokrynskyi.com
        Regression: No

Host system is Alpine Linux x86_64 with kernel 4.9.32 and QEMU 2.8.1-r1.
KVM guest is LEDE with Wi-Fi adapter attached to it, chipset used is
pc-q35-2.8.
Wi-Fi adapter is USB 2.0, but inserted into USB 3.0 port on the motherboard.
CPU is Intel Celeron N3150.

The problem I'm having is that it can barely cross 30Mbps speed with this Wi-Fi
adapter, while both adapter acting as an AP and client device can handle up to
300Mbps, well within USB 2.0 spec and well above what I'm currently seeing.

When USB 2.0 was used in VM download speed was only 10Mpbs, when switched to
USB 3.0 it increased to 14Mbps, upload in both cases is 30-32Mbps.

There is some serious bottleneck, is this an expected level of performance or
am I missing something?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* [Bug 196527] Terrible KVM USB passthrough performance
  2017-07-28 22:42 [Bug 196527] New: Terrible KVM USB passthrough performance bugzilla-daemon
@ 2017-08-10 12:34 ` bugzilla-daemon
  2017-08-10 12:38 ` bugzilla-daemon
  1 sibling, 0 replies; 3+ messages in thread
From: bugzilla-daemon @ 2017-08-10 12:34 UTC (permalink / raw)
  To: kvm

https://bugzilla.kernel.org/show_bug.cgi?id=196527

heyark (kvm-usb@suremail.info) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kvm-usb@suremail.info

--- Comment #1 from heyark (kvm-usb@suremail.info) ---

I have to agree here.
In my case, I tried to use two USB devices inside a KVM VM - and failed with
both due to timeout issues.
I pass them through to the VM using kvm parameters -usb -usbdevice
host:[vendorId]:[deviceId].

They are visible inside using lsusb. But trying to access them, both cannot be
used, while there is no problem using them directly on the host machine
(current Ubuntu 16.04.3 on a Core i5 7500).

One is a HomeMatic USB-sender
(http://www.fhemwiki.de/wiki/HM-CFG-USB_USB_Konfigurations-Adapter), the other
one is a smartcard reader.
In case of the Homematic, the driver (hmland) always says:
 usb-transfer took more than 100ms (125ms), this may lead to timing problems!
And the warning is right: nothing works, nothing gets received (using the stick
outside the VM works great).

Maybe the usb passthrough works generally with less sensible devices (did not
try an USB stick or something), but obviously the performance in handling this
is so bad, that any device that needs a proper timing is not usable.

Probably USB devices at servers are rare in the professional environments where
kvm is heavily used, nonetheless this is very disappointing. It definitely
limits the use of KVM virtualization.
Probably not a good sign too, that the following page is empty:
https://www.linux-kvm.org/page/UsbPassthrough

Is there any chance this will be cared for soon? Some kind of higher priority
for handling the USB passthrough?
Thanks!

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* [Bug 196527] Terrible KVM USB passthrough performance
  2017-07-28 22:42 [Bug 196527] New: Terrible KVM USB passthrough performance bugzilla-daemon
  2017-08-10 12:34 ` [Bug 196527] " bugzilla-daemon
@ 2017-08-10 12:38 ` bugzilla-daemon
  1 sibling, 0 replies; 3+ messages in thread
From: bugzilla-daemon @ 2017-08-10 12:38 UTC (permalink / raw)
  To: kvm

https://bugzilla.kernel.org/show_bug.cgi?id=196527

--- Comment #2 from heyark (kvm-usb@suremail.info) ---
(of course I meant the higher prio in a technical sense for handling the USB
communication of the VM - not an organizational prio that developers care for
this - would be great of course)

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2017-08-10 12:38 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-07-28 22:42 [Bug 196527] New: Terrible KVM USB passthrough performance bugzilla-daemon
2017-08-10 12:34 ` [Bug 196527] " bugzilla-daemon
2017-08-10 12:38 ` bugzilla-daemon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).