From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mario Subject: Re: PV networking issue in 2.6.38.6 and 2.6.39-rc7 pvops dom0 kernels Date: Tue, 17 May 2011 13:20:32 +0200 Message-ID: <4DD25A00.6090707@slackverse.org> References: <4DD104AB.8000401@slackverse.org> <1305550008.20907.8.camel@zakaz.uk.xensource.com> <4DD11EF5.1050203@slackverse.org> <1305550841.20907.10.camel@zakaz.uk.xensource.com> <4DD120EA.4060204@slackverse.org> <1305551216.20907.12.camel@zakaz.uk.xensource.com> <4DD12212.7020403@slackverse.org> <4DD12389.1020905@slackverse.org> <20110516132231.GS32595@reaktio.net> <4DD252AE.6080609@slackverse.org> <20110517110440.GU32595@reaktio.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <20110517110440.GU32595@reaktio.net> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-users-bounces@lists.xensource.com Errors-To: xen-users-bounces@lists.xensource.com To: =?ISO-8859-1?Q?Pasi_K=E4rkk=E4inen?= Cc: Ian Campbell , "todd.deshane@xen.org" , "xen-devel@lists.xensource.com" , "xen-users@lists.xensource.com" , Konrad Rzeszutek Wilk List-Id: xen-devel@lists.xenproject.org On 05/17/2011 01:04 PM, Pasi K=E4rkk=E4inen wrote: > On Tue, May 17, 2011 at 12:49:18PM +0200, Mario wrote: >> On 05/16/2011 03:22 PM, Pasi K=E4rkk=E4inen wrote: >>> On Mon, May 16, 2011 at 03:15:53PM +0200, Mario wrote: >>>> On 05/16/2011 03:09 PM, Mario wrote: >>>>> On 05/16/2011 03:06 PM, Ian Campbell wrote: >>>>>> On Mon, 2011-05-16 at 14:04 +0100, Mario wrote: >>>>>>> So you are telling me that 2.6.38 does not have a working network= ing on >>>>>>> PV guests, but HVM does? >>>>>> >>>>>> Yes, netback did not go upstream until 2.6.39 so PV networking (ei= ther >>>>>> for PV or HVM guests with PV drivers) will not work on mainline ke= rnels >>>>>> before that. >>>>>> >>>>>> Emulated HVM networking does not use netback so that will work on >>>>>> earlier kernels (it depends on qemu) >>>>>> >>>>>> Ian. >>>>>> >>>>> >>>>> Ok, thanks, it all makes more sense now. :-) >>>> >>>> >>>> Ian, while i have your attention, any idea why some framebuffer devi= ces >>>> blank screen when dom0 boots up. I have seen this happen mostly on o= lder >>>> vga cards but its not a rule. Should i just keep on blacklisting ker= nel >>>> modules for those cards, or is there some kind of a more general >>>> solution? >>> >>> You could try "nomodeset" parameter to disable KMS modesetting.. >>> (and possibly adding "console=3Dtty0") >>> >>> Xen patches for DRM/KMS haven't been merged to upstream Linux yet. >>> >>> -- Pasi >>> >> >> >> nomodeset and/or nofb do not seem to help. >> console=3Dtty0 also does not seem to have any effect. >> >> Only way to successfuly boot dom0 is with one of these: >> >> CONFIG_FB=3Dn >> >> or >> >> echo "blacklist s3fb"> /etc/modprobe.d/blacklist-s3fb.conf >> >> As soon as fb device is modprobed, booting stops, and server freezes. >> Can anything else be done? Perhaps some setting for xen in kernel..? > > Set up a serial console and configure Xen + dom0 kernel log to it.. > and see what goes wrong. http://wiki.xen.org/xenwiki/XenSerialConsole > > ie. does dom0 kernel crash while loading s3fb? > > -- Pasi > Well, i forgot to mention one more method, that is to add: fbcon=3Dmap:X Where X is something that is not in /proc/fb, basicly an invalid device. So, for example: fbcon=3Dmap:9 To make it short, dom0 freezes/crashes and it never reaches a full=20 booted up state, so it will not respond to pings. The last thing=20 displayed on the screen is s3fb being loaded and that /dev/fb0 was create= d. When i first experienced this problem, it was back when i first tried=20 Xen 4.0.1, however, i had just a blank screen with blinking cursor to=20 work with. So i had to set up a serial console to figure out what was=20 wrong. After seeing how it stops on framebuffer, i blacklisted the=20 module, and it worked. I kept on doing so on every other setup i have=20 made, but now, one year later, i was wondering if this is ever going to=20 be fixed, or if i can somehow help to get it fixed.. :-) m.