[Bug 115781] New: Hyper-V video driver leaking memory

bugzilla-daemon at bugzilla.kernel.org bugzilla-daemon at bugzilla.kernel.org
Sat Apr 2 04:53:51 UTC 2016


https://bugzilla.kernel.org/show_bug.cgi?id=115781

            Bug ID: 115781
           Summary: Hyper-V video driver leaking memory
           Product: Drivers
           Version: 2.5
    Kernel Version: 4.4.5-1
          Hardware: Intel
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: Video(DRI - non Intel)
          Assignee: drivers_video-dri at kernel-bugs.osdl.org
          Reporter: feduzi at mailinator.com
        Regression: No

After some testing seems like hyper-v video driver is leaking memory.

The setup:
Host OS: Windows 10 + Hyper-V
Guest OS: a linux distro

Steps to reproduce (archlinux example):
1. Setup a minimal system (and xorg-server, xorg-server-utils, xterm)
2. Launch "top" and wait. (I waited for an hour, and have seen no changes in
memory allocation)
3. Use "startx" to launch basic X11 session.
4. Launch "top" and wait. (It takes 5-10 minutes for me, for memory to be
mostly occupied, about 20% is left)

Steps to reproduce (ubuntu server example):
1. Setup a minimal system (and xorg, xterm)
2. Launch "top" and wait. (I waited for an hour, and have seen no changes in
memory allocation)
3. Use "startx" to launch basic X11 session.
4. Launch "top" and wait. (It takes 5-10 minutes for me, for memory to be
mostly occupied, about 15% is left)

As seen from the 2 examples above, we have an issue across different distros.
Also I've set up the same 2 distros on VirtualBox VMs and had no issues with
memory allocation, which point into Hyper-V direction.
As seen previously the issue is not occurring until X11 is not launched, which
points into display-related direction.
Also to note, I tried to use top/htop tools to pinpoint a program, which leaks
memory. But the most used memory percentage is 0.3 at the time when 80% of RAM
is used, which points into non-program direction (I assumed driver).

Thanks.

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


More information about the dri-devel mailing list