Regressions introduced by ioemu-remote -- Xen #18039.

Regressions introduced by ioemu-remote -- Xen #18039.

[Li, Haicheng]

With ioemu-remote, still failed to create hvm guest with vncviewer
disabled by 'sdl=1' or 'vnc=1 & vncconsole=0'. Two issues were found
with Xen #18039 as log pasted below.

BTW, will xen-3.3 release use ioemu-remote as default qemu? Looks there
would be some compatibility issues between xm/xend and ioemu-remote.

Issue1: if no vif option set. qemu exits with following log:
domid: 14
qemu: the number of cpus is 1
Unknown network device: user

Issue2: with vif option set, qemu exits with following log:
domid: 18
qemu: the number of cpus is 1
config qemu network with xen bridge for  tap18.0 eth4
Strip off blktap sub-type prefix to
/share/imgs/hli22/xen-test/manual-test/ia32e_xp_smp.qcow (drv 'qcow')
qemu: could not open vbd '/local/domain/18/device/vbd/768/phantom_vbd'
or hard disk image
'/share/imgs/hli22/xen-test/manual-test/ia32e_xp_smp.qcow' (drv 'qcow'
format '0')
Watching /local/domain/0/device-model/18/logdirty/next-active
Watching /local/domain/0/device-model/18/command
char device redirected to /dev/pts/3
qemu_map_cache_init nr_buckets = 10000 size 3145728
shared page at pfn 7ffe
buffered io page at pfn 7ffc
Time offset set 0
Register xen platform.
Done register platform.
I/O request not ready: 0, ptr: 0, port: 0, data: 0, count: 0, size: 0
(qemu) Failed allocation for dom 18: 2048 extents of order 0
Failed to populate physmap
Failed set_mm_mapping
NULL vram_pointer


-- haicheng

Re: Regressions introduced by ioemu-remote -- Xen #18039.

[Yosuke Iwamatsu]

[-- Attachment #1: Type: text/plain, Size: 1836 bytes --]

Perhaps the attached patch will fix the Issue2?
cs 18025 changed the size of vga ram, so we should reduce VGA_RAM_SIZE
to 4MB for ioemu-remote tree likewise.

Thanks,
-- Yosuke

Li, Haicheng wrote:
> With ioemu-remote, still failed to create hvm guest with vncviewer
> disabled by 'sdl=1' or 'vnc=1 & vncconsole=0'. Two issues were found
> with Xen #18039 as log pasted below.
> 
> BTW, will xen-3.3 release use ioemu-remote as default qemu? Looks there
> would be some compatibility issues between xm/xend and ioemu-remote.
> 
> Issue1: if no vif option set. qemu exits with following log:
> domid: 14
> qemu: the number of cpus is 1
> Unknown network device: user
> 
> Issue2: with vif option set, qemu exits with following log:
> domid: 18
> qemu: the number of cpus is 1
> config qemu network with xen bridge for  tap18.0 eth4
> Strip off blktap sub-type prefix to
> /share/imgs/hli22/xen-test/manual-test/ia32e_xp_smp.qcow (drv 'qcow')
> qemu: could not open vbd '/local/domain/18/device/vbd/768/phantom_vbd'
> or hard disk image
> '/share/imgs/hli22/xen-test/manual-test/ia32e_xp_smp.qcow' (drv 'qcow'
> format '0')
> Watching /local/domain/0/device-model/18/logdirty/next-active
> Watching /local/domain/0/device-model/18/command
> char device redirected to /dev/pts/3
> qemu_map_cache_init nr_buckets = 10000 size 3145728
> shared page at pfn 7ffe
> buffered io page at pfn 7ffc
> Time offset set 0
> Register xen platform.
> Done register platform.
> I/O request not ready: 0, ptr: 0, port: 0, data: 0, count: 0, size: 0
> (qemu) Failed allocation for dom 18: 2048 extents of order 0
> Failed to populate physmap
> Failed set_mm_mapping
> NULL vram_pointer
> 
> 
> -- haicheng
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel


[-- Attachment #2: vga_ram_size.patch --]
[-- Type: all/allfiles, Size: 462 bytes --]

[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

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

Re: Regressions introduced by ioemu-remote -- Xen #18039.

[Keir Fraser]

Actually this is already applied to out private tree but hasn't been
automatically pushed publicly for some reason. I'm sure Ian will sort it out
today.

 -- Keir

On 14/7/08 09:38, "Yosuke Iwamatsu" <y-iwamatsu@ab.jp.nec.com> wrote:

> Perhaps the attached patch will fix the Issue2?
> cs 18025 changed the size of vga ram, so we should reduce VGA_RAM_SIZE
> to 4MB for ioemu-remote tree likewise.
> 
> Thanks,
> -- Yosuke
> 
> Li, Haicheng wrote:
>> With ioemu-remote, still failed to create hvm guest with vncviewer
>> disabled by 'sdl=1' or 'vnc=1 & vncconsole=0'. Two issues were found
>> with Xen #18039 as log pasted below.
>> 
>> BTW, will xen-3.3 release use ioemu-remote as default qemu? Looks there
>> would be some compatibility issues between xm/xend and ioemu-remote.
>> 
>> Issue1: if no vif option set. qemu exits with following log:
>> domid: 14
>> qemu: the number of cpus is 1
>> Unknown network device: user
>> 
>> Issue2: with vif option set, qemu exits with following log:
>> domid: 18
>> qemu: the number of cpus is 1
>> config qemu network with xen bridge for  tap18.0 eth4
>> Strip off blktap sub-type prefix to
>> /share/imgs/hli22/xen-test/manual-test/ia32e_xp_smp.qcow (drv 'qcow')
>> qemu: could not open vbd '/local/domain/18/device/vbd/768/phantom_vbd'
>> or hard disk image
>> '/share/imgs/hli22/xen-test/manual-test/ia32e_xp_smp.qcow' (drv 'qcow'
>> format '0')
>> Watching /local/domain/0/device-model/18/logdirty/next-active
>> Watching /local/domain/0/device-model/18/command
>> char device redirected to /dev/pts/3
>> qemu_map_cache_init nr_buckets = 10000 size 3145728
>> shared page at pfn 7ffe
>> buffered io page at pfn 7ffc
>> Time offset set 0
>> Register xen platform.
>> Done register platform.
>> I/O request not ready: 0, ptr: 0, port: 0, data: 0, count: 0, size: 0
>> (qemu) Failed allocation for dom 18: 2048 extents of order 0
>> Failed to populate physmap
>> Failed set_mm_mapping
>> NULL vram_pointer
>> 
>> 
>> -- haicheng
>> 
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re: Regressions introduced by ioemu-remote -- Xen #18039.

[Keir Fraser]

On 14/7/08 08:54, "Li, Haicheng" <haicheng.li@intel.com> wrote:

> With ioemu-remote, still failed to create hvm guest with vncviewer
> disabled by 'sdl=1' or 'vnc=1 & vncconsole=0'. Two issues were found
> with Xen #18039 as log pasted below.
> 
> BTW, will xen-3.3 release use ioemu-remote as default qemu? Looks there
> would be some compatibility issues between xm/xend and ioemu-remote.

We have every intention of shipping 3.3 with the new qemu merge enabled as
default. We still have a month to shake out the bugs.

 -- Keir

Re: Regressions introduced by ioemu-remote -- Xen #18039.

[Ian Jackson]

Keir Fraser writes ("Re: [Xen-devel] Regressions introduced by ioemu-remote -- Xen #18039."):
> Actually this is already applied to out private tree but hasn't been
> automatically pushed publicly for some reason. I'm sure Ian will sort it out
> today.

I've been investigating this.  A manual push worked, using the same
rune as the automatic system is using.  I haven't yet managed to
determine which the automatic patchman-controlled push didn't work.

I've added some more diagnostics and will keep an eye on it.

Ian.