[Xen-ia64-devel] RE: [PATCH] First round of cleanup CONFIG_VTI patch

["Tian, Kevin" <kevin.tian@xxxxxxxxx>]

>From: Magenheimer, Dan (HP Labs Fort Collins)
>
>Just getting started again, I grabbed -Intel tip (cset 5986)
>and tried to run it on the simulator.  It appears
>the code in construct_dom0 which calls ia64_pal_cache_flush
>crashes Xen.  Commenting out this call allows me
>to boot (at least into domain0).  Perhaps this should
>be "if (!running_on_sim)"?

Maybe ski doesn't support pal cache flush well? Anyway, it should be
safe to disable this line for ski in the start.

>
>I then tried it on my test machine.  I got no console output
>at all from Xen and then a machine check.  This both
>with and without the cache_flush code.

This is interesting... Did you have a clue where it hung? Actually last
weekend I did test with cset 5986 and everything works OK on Tiger4. Is
it possible that some changes are platform specific and then affects zx
machine, though I think not? Anyway, I'll make a test again, and better
if you can take some time to help review change in common side. Then you
may point out easily if obvious neglect exists. ;-)

Thanks,
Kevin
>
>Just to verify, I fell back to xen-ia64-unstable tip
>and it booted fine (with console output).
>
>Dan
>
>P.S. Note that this is testing many recent Intel changes
>at once so may not be due to your latest cset.
>
>> -----Original Message-----
>> From: Tian, Kevin [mailto:kevin.tian@xxxxxxxxx]
>> Sent: Friday, August 19, 2005 11:16 PM
>> To: Magenheimer, Dan (HP Labs Fort Collins)
>> Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
>> Subject: [PATCH] First round of cleanup CONFIG_VTI patch
>>
>> Hi, Dan,
>>      Following Fred's info about supporting xen0+xenU+xenVTI at
>> runtime, we've checked in first round of cleanup patch into Intel
>> staging tree with Rev. 5986. Could you take a look? No
>> functional level
>> change and most is code cleanup. We verified this patch not breaking
>> anything and xen0+xenU still working. If close to your
>> thought, can you
>> please take a pull? ;-)
>>
>> Thanks,
>> Kevin
>>
>>      Following is the description about the changeset:
>>
>> changeset:   5986:1ec2225aa8c696ca4e96e0fc27b4eafe36a9633f
>> tag:         tip
>> user:        fred@xxxxxxxxxxxxxxxxxxxxx
>> date:        Sat Aug 20 13:19:39 2005
>> summary:     First step to remove CONFIG_VTI for final supporting
>> xen0+xenU+xenVTI at runtime. This changeset mainly addresses
>> common code
>> like domain creation and rid allocation policy, including:
>>
>> - Boot time vti feature detection
>> - Uniform arch_do_createdomain, new_thread, arch_set_infoguest, and
>> construct_dom0. Now these function level CONFIG_VTIs have been
>removed
>> with several specific lines still protected by CONFIG_VTIs. With more
>> feature cleanup later, these lines will be free out grandually.
>> - Use same rid allocation policy including physical emulation
>> - Remove duplicated definition rr_t.
>>
>> Verified breaking nothing. ;-)
>>
>> Signed-off-by Kevin Tian <kevin.tian@xxxxxxxxx>
>>
>>

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel