[Xen-bugs] [Bug 1779] pv-grub in xen-4.1.0 cannot boot kernel in graphical mode (vfb)

[bugzilla-daemon@xxxxxxxxxxxxxxxxxxx]

http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1779





------- Comment #12 from johneed@xxxxxxxxxxx  2011-09-05 23:35 -------
yes, I think so.

I have kernel-3.0.0 compiled and booted as a dom0, also 2.6.38-xen-gentoo.

With both kernels I get the same result.
2.6.38 has a problem with connecting to the inet, so don't be concerned re
errors for that.  I rely on 3.0.0 anyway, being more current.

Both manage to make an sdl window.

Now It has two kernels.  ie guest kernels.  

2.6.38 is configure to boot as hvm with /dev/sda2 as root.
2.6.39 is configured as XenPVonHVM, so it is the kernel of choice for this.
Selecting the 2.6.39 kernel in the grub list of the sdl window yields killing
or xm destroying the vm, it just vanishes the moment the boot touches the
kernel.  Selecting the 2.6.38 it just freezes or hangs.
I have used console=tty0 and console=hvc0 on selecting the 2.6.39 kernel. 
Makes no diff, just disappears.

Commenting out the selection of vfb entirely allows it to boot in the host
console, xm create -c genpv.cfg.

This yields;

gentoo64 linux # xm create genpv.cfg -c
Using config file "/etc/xen/genpv.cfg".
Started domain gentoopv (id=10)
                               Xen Minimal OS!
  start_info: 0xc8e000(VA)
    nr_pages: 0x5dc00
  shared_inf: 0xcf8fe000(MA)
     pt_base: 0xc91000(VA)
nr_pt_frames: 0xb
    mfn_list: 0x9a0000(VA)
   mod_start: 0x0(VA)
     mod_len: 0
       flags: 0x0
    cmd_line: (hd0,0)/grub/menu.lst console=hvc0
  stack:      0x95f9e0-0x97f9e0
MM: Init
      _text: 0x0(VA)
     _etext: 0x6f209(VA)
   _erodata: 0x89000(VA)
     _edata: 0x91c60(VA)
stack start: 0x95f9e0(VA)
       _end: 0x99ffe0(VA)
  start_pfn: c9f
    max_pfn: 5dc00
Mapping memory range 0x1000000 - 0x5dc00000
setting 0x0-0x89000 readonly
skipped 0x1000
MM: Initialise page allocator for f86000(f86000)-5dc00000(5dc00000)
MM: done
Demand map pfns at 5dc01000-205dc01000.
Heap resides at 205dc02000-405dc02000.
Initialising timer interface
Initialising console ... done.
gnttab_table mapped at 0x5dc01000.
Initialising scheduler
Thread "Idle": pointer: 0x205dc02050, stack: 0x5da90000
Initialising xenbus
Thread "xenstore": pointer: 0x205dc02800, stack: 0x5daa0000
Dummy main: start_info=0x97fae0
Thread "main": pointer: 0x205dc02fb0, stack: 0x5dab0000
Thread "pcifront": pointer: 0x205dc03760, stack: 0x5dac0000
"main" "(hd0,0)/grub/menu.lst" "console=hvc0" 
pcifront_watches: waiting for backend path to appear device/pci/0/backend

    GNU GRUB  version 0.97  (1536000K lower / 0K upper memory)

 +-------------------------------------------------------------------------+
 | Gentoo linux kernel 2.6.38-r6                                           |  
 | Gentoo linux 2.6.39-r3                                                  |

Selecting the 2.6.39 and editing to console=hvc0, it happily boot.

This is gentoo64-vm.unknown_domain (Linux x86_64 2.6.39-gentoo-r3) 14:16:58

gentoo64-vm login: 

You ask for kernel dumps.  Well here is my best interpretation.
You may need to clarify or specify just how to do so.
xen kernel stuff often does not go to dmesg.  

I deleted the xend.log & xend-debug.log on booting into kernel-3.0 so as to
acquire info related only to the attempt to boot in pvgrub.

idella@gentoo64 ~ $ ls -ld /var/log/xen/xend-debug.log
ls: cannot access /var/log/xen/xend-debug.log: No such file or directory
It made no record.

idella@gentoo64 ~ $ ls -ld /var/log/xen/xend.log
-rw-r--r-- 1 root root 0 Sep  6 13:44 /var/log/xen/xend.log
It made the file, no content.

dmesg tail shows a clear citation if i2c, which I don't fully comprehend. A
series of repeated

[ 3153.312154] i2c i2c-1: master_xfer[0] W, addr=0x50, len=1
[ 3153.312158] i2c i2c-1: master_xfer[1] R, addr=0x50, len=1
[ 3153.315452] i2c i2c-1: master_xfer[0] W, addr=0x50, len=1
[ 3153.315455] i2c i2c-1: master_xfer[1] R, addr=0x50, len=1
[ 3153.318733] i2c i2c-1: master_xfer[0] W, addr=0x50, len=1
[ 3153.318736] i2c i2c-1: master_xfer[1] R, addr=0x50, len=128
[ 3153.416872] i2c i2c-2: master_xfer[0] W, addr=0x50, len=1
[ 3153.416875] i2c i2c-2: master_xfer[1] R, addr=0x50, len=1
[ 3153.420544] i2c i2c-2: NAK from device addr 0x50 msg #0

That is my strongest new clue.
View the couple of attachments from the run in the xen kernel 2.6.38
Feel free to prompt further as to how to execute this desired dump.

Oh this may help a little.


    GNU GRUB  version 0.97  (1536000K lower / 0K upper memory)

       [ Minimal BASH-like line editing is supported.   For
         the   first   word,  TAB  lists  possible  command
         completions.  Anywhere else TAB lists the possible
         completions of a device/filename.  ESC at any time
         exits. ]

grubdom> root (hd0,0)                                                          
Error ENOENT when reading the backend path device/vkbd/0/backend
Thread "kbdfront" exited.
grubdom> root (hd0,0)                                                          
 Filesystem type is ext2fs, partition type 0x83
grubdom>                                                                       
grubdom>                                                                       

grubdom> kernel /vmlinuz-2.6.39-gentoo-r3 ro console=hvc0                      

<xvda2 video=cirrusfb                                                          

grubdom> kernel /vmlinuz-2.6.39-gentoo-r3 ro console=hvc0 root=/dev/xvda2 vide>

grubdom>                                                                       

grubdom>   

Note Thread "kbdfront" exited.  Note on citing the kernel, it does not display
a loading address msg.  Just a blank. but in the host console, it does boot the
selected kernel.


-- 
Configure bugmail: 
http://bugzilla.xensource.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

_______________________________________________
Xen-bugs mailing list
Xen-bugs@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-bugs