Date: Thu, 19 Apr 2007 20:52:55 +0200 (CEST) From: Juergen Lock <nox@jelal.kn-bremen.de> To: anderson@freebsd.org Cc: freebsd-emulation@freebsd.org Subject: Re: experimental qemu-devel port update, please test! Message-ID: <200704191852.l3JIqtaT062272@saturn.kn-bremen.de> In-Reply-To: <46237925.60309@freebsd.org> References: <20070405210341.GA25547@saturn.kn-bremen.de> <200704072004.l37K4rua048988@saturn.kn-bremen.de> <461FD9E7.2040201@freebsd.org> <20070414170037.GA9357@saturn.kn-bremen.de>
next in thread | previous in thread | raw e-mail | index | archive | help
In article <46237925.60309@freebsd.org> you write: >On 04/14/07 12:00, Juergen Lock wrote: >> On Fri, Apr 13, 2007 at 02:28:39PM -0500, Eric Anderson wrote: >>> On 04/07/07 15:04, Juergen Lock wrote: >>>> In article <4615DB05.8000608@freebsd.org> you write: >>>>> On 04/05/07 16:03, Juergen Lock wrote: >>>>>> Hi! >>>>>> >>>>>> There have been a few interesting commits to qemu cvs, like, >>>>>> a qcow2 corruption bugfix, vmware vga emulation (-vmwarevga), and >>>>>> several emulated eepro100 (fxp) nic variants (-net nic,model=i82557b >>>>>> -net user worked for me with FreeBSD and linux guests, and seemed even >>>>>> slightly faster than the rtl8139 that already outperforms the default >>>>>> ne2kpci), so I prepared another qemu-devel port update that I now want >>>>>> you all to test with your various guests before I commit it! >>>>>> >>>>>> TIA, and enjoy... >>>>>> Juergen >>>>>> >>>>>> Remove files: files/patch-audio::ossaudio.c files/patch-bh >>>>> Building it now. Quick question: why have some diffs below in the >>>>> patch, when you say remove those files? Just curious.. >>>> Oh, just for the benefit of users who might forget to rm them after >>>> applying the diff. :) (patch turning them into empty files...) >>>> >>>> Just committed this btw. >>>> >>>> Juergen >>> >>> Hmm.. Now, using the vmware driver makes my system either shutdown, or >>> insta-reboot. :( >> >> You mean -vmwarevga? Which guest? And, does the guest crash or the host? > >Yea, but I found out it does it whether or not I use the vmwarevga >option, and seems to be tied to kqemu kldloaded. I'm still looking at >it - kind of painful since it crashes my host system. The guest is a >etherbooted FreeBSD-current, but it doesn't get far enough to even boot >the OS in the guest. Ah so its a kqemu problem? You may want to post about that on the qemu list... (You can also disable kqemu by running qemu with -no-kqemu btw.) Or, did something in your kernel change maybe that necessitates a rebuild of the kqemu kld? > > >> You could try disabling the >> #define HW_RECT_ACCEL >> #define HW_FILL_ACCEL >> near the top of hw/vmware_vga.c, a range check is missing in the >> accellerated case, see >> http://lists.gnu.org/archive/html/qemu-devel/2007-04/msg00136.html >>> kqemu-kmod-1.3.0.p11 Kernel Acceralator for QEMU CPU Emulator >>> (development versi >>> qemu-devel-0.9.0s.20070405_2 QEMU CPU Emulator - development snapshot >> >> Yeah those are the lastest... > >I'll try that if I still suspect it. Although, I can say that without >kqemu loaded, but the vmwarevga option used, my ethernet device ends up >with a MAC of FF:FF:FF:FF:FF:FF. I can't override it either. Thats probably a result of io address overlaps with the emulated nic caused by the current version of the vmwarevga emulation (the author of that is currently away afaik, otherwise it probably would have been fixed already at least in qemu cvs.) You may be able to get around that by using the eepro nic emulation. (sound is also affected btw, see the mentioned -vmwarevga thread on the qemu list.) > >I'll try some more things, and report back. Ok. Juergen
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200704191852.l3JIqtaT062272>