Re: problem using compiled kernel -- Sisu

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Tue, Aug 03, 2010 at 10:01:08AM -0700, Jeremy Fitzhardinge wrote:
>  On 08/02/2010 08:25 PM, Sisu Xi wrote:
> >
> >Fedora release 13 (Goddard)
> >
> >Kernel 2.6.32.16 on an x86_64 (/dev/hvc0)
> >
> >
> >
> >flamenco118.cse.wustl.edu login: (XEN) mm.c:3852:d0 Bad page
> >00000000000b918f: ed=ffff830124ff0000(0), sd=ffff830124ff0000,
> >caf=8000000000000002, taf=7400000000000001
> >(XEN) mm.c:3852:d0 Bad page 00000000000b6850: ed=ffff830124ff0000(0),
> >sd=ffff830124ff0000, caf=8000000000000002, taf=7400000000000001

That is interesting. I see those when I used PVGRUB and launched
a guest that would utilize Xen-SWIOTLB. Thought I wouldn't have the 'd0'
but rather the guest domain id.

I did found out that if I did not use PVGRUB but just used the kernel
directly:

kernel = "/mnt/tmp/vmlinuz-2.6.34NEB"
ramdisk = "/mnt/tmp/initramfs-2.6.34NEB.img"
extra ="ro root=/dev/mapper/vg_vs03-lv_root rd_LVM_LV=vg_vs03/lv_root rd_LVM_LV=vg_vs03/lv_swap rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us console=hvc0 debug init=/bin/bash"

#kernel = "/usr/lib/xen/boot/pv-grub-x86_64.gz"
#kernel = "/home/konrad/ssd/neb/xen/stubdom/mini-os-x86_64-grub/mini-os.gz"
#extra = "(hd0,0)/grub/menu.lst"

these errors would disappear. So I blamed it on MiniOS, but haven't gone
any further than that :-(

FYI, this is what I see on the serial console:
 
(XEN) mm.c:4147:d1 Bad page 00000000001321bb: ed=ffff830188ce0000(1),
sd=ffff830188ce0000, caf=8000000000000002, taf=7400000000000001
[    0.000000] Kernel panic - not syncing: DMA(-12): Failed to exchange
pages allocated for DMA with Xen! We either don't have the permission or
you do not have enoughfree memory under 4GB!
[    0.000000] 
[    0.000000] Pid: 0, comm: swapper Not tainted 2.6.34NEB #1
[    0.000000] Call Trace:
[    0.000000]  [<ffffffff814536ac>] panic+0x73/0xee
[    0.000000]  [<ffffffff810052c2>] ?
xen_create_contiguous_region+0xd3/0xee
[    0.000000]  [<ffffffff818c6946>] xen_swiotlb_init+0x10a/0x142
[    0.000000]  [<ffffffff8145ceb5>] ? _etext+0x0/0x3

I wonder if he is doing some sort of exchange MFNs and that triggers
this.

Sisu, can you make your Xen command line be as so:
 guest_loglvl=all loglvl=all console=com1,vga com1=115200,8n1 console_to_ring sync_console 

The 'console_to_ring' is of interest here.