qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] Re: [Try2][PATCH] Initial implementation of a mpeg1 lay


From: Anthony Liguori
Subject: Re: [Qemu-devel] Re: [Try2][PATCH] Initial implementation of a mpeg1 layer2 streaming audio driver.
Date: Tue, 16 Nov 2010 13:33:31 -0600
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.15) Gecko/20101027 Lightning/1.0b1 Thunderbird/3.0.10

On 11/16/2010 01:22 PM, Andreas Färber wrote:
Am 15.11.2010 um 22:53 schrieb François Revol:

I'm still having a hard time getting sound to work reliably in Haiku guests regardless the output or emulated card. OTH it works perfectly in VirtualBox with almost no cpu overhead. If anyone has a clue... It might be related to the use of RDTSC as primary time source, still it's the same in vbox.

Did you check the time drift command line options? Iirc Gleb needed some hacks for timer-related issues on Windows.

Time drift correction is only applicable to the RTC and PIT and deals with catching up the periodic timers.

The issue with the TSC as a time source is that it skips around during CPU migration. The easiest way to work around this is to simply pin your guest to a single physical CPU (assuming you're using KVM).

With TCG, it depends, but you can actually end up accessing the host TSC in which case, your SOL because the code in QEMU that uses the host TSC assumes a UP host which aren't all that common anymore :-)

Regards,

Anthony Liguori

(does [KVM] even work on OSX ?)

No.

Andreas




reply via email to

[Prev in Thread] Current Thread [Next in Thread]