From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: Re: [Xen-users] Hypervisor hangs on startup, 2.6.32-5-xen dom0 kernel panic Date: Mon, 27 Dec 2010 10:28:12 -0500 Message-ID: <20101227152811.GC3728@dumpdata.com> References: <4D0AD782.8030204@wavecon.de> <4D0B3DF2.5000100@wavecon.de> <20101217115132.GG2754@reaktio.net> <4D0B7E79.4010106@wavecon.de> <20101220201819.GN2754@reaktio.net> <1292937209.4500.1852.camel@zakaz.uk.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Return-path: Content-Disposition: inline In-Reply-To: <1292937209.4500.1852.camel@zakaz.uk.xensource.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Ian Campbell Cc: Arno Toell , "xen-devel@lists.xensource.com" , "xen-users@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org On Tue, Dec 21, 2010 at 01:13:29PM +0000, Ian Campbell wrote: > On Mon, 2010-12-20 at 20:18 +0000, Pasi K=E4rkk=E4inen wrote: > > > > [ 6.122370] Initializing CPU#0 > > > > [ 6.131080] DMA: Placing 64MB software IO TLB between ffff8800= 20000000 - ffff880024000000 > > > > [ 6.131087] DMA: software IO TLB at phys 0x20000000 - 0x240000= 00 > > > > [ 6.131090] xen_swiotlb_fixup: buf=3Dffff880020000000 size=3D6= 7108864 > > > > [ 6.139836] Kernel panic - not syncing: DMA(-12): Failed to ex= change pages allocated for DMA with Xen! We either don't have the permiss= ion or you do not ! >=20 > This suggests that the xen_exchange_memory hypercall has failed for som= e > reason. >=20 > -12 is -ENOMEM but size is only 64M and this is start of day so I can't > think why this would be the case (or at least why it wouldn't always be > the case if it was happening at all). >=20 > Might be worth trying the dom0_mem hypervisor parameter to limit the > initial size of dom0, perhaps to 32G. His 'node_to_dma32' ended being 0 and 62 respectivly. Not enough to alloc= ate the 64MB SWIOTLB buffer. Isn't there some logic in the Xen hypervisor to=20 leave some swath of memory for DMA32? >=20 > Ian. >=20 > > > > [ 6.139838] > > > > [ 6.139845] Pid: 0, comm: swapper Not tainted 2.6.32-5-xen-amd= 64 #1 > > > > [ 6.139847] Call Trace: > > > > [ 6.139856] [] ? panic+0x86/0x143 > > > > [ 6.139862] [] ? xen_restore_fl_direct_end+= 0x0/0x1 > > > > [ 6.139867] [] ? _spin_unlock_irqrestore+0x= d/0xe > > > > [ 6.139871] [] ? xen_create_contiguous_regi= on+0xc8/0xe1 > > > > [ 6.139878] [] ? xen_swiotlb_fixup+0x8b/0xb= a > > > > [ 6.139882] [] ? _etext+0x0/0x1f360d > > > > [ 6.139887] [] ? xen_swiotlb_init+0x5c/0x5f > > > > [ 6.139891] [] ? pci_xen_swiotlb_init+0x17/= 0x24 > > > > [ 6.139895] [] ? pci_iommu_alloc+0x72/0x78 > > > > [ 6.139900] [] ? mem_init+0x15/0xe5 > > > > [ 6.139904] [] ? start_kernel+0x202/0x3e8 > > > > [ 6.139908] [] ? xen_start_kernel+0x586/0x5= 8a > > > > (XEN) Domain 0 crashed: rebooting machine in 5 seconds.=20 >=20 >=20 > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel