Re: [Xen-devel] [xen-unstable] Commit 2ca9fbd739b8a72b16dd790d0fff7b75f5488fb8 AMD IOMMU: allocate IRTE entries instead of using a static mapping, makes dom0 boot process stall several times.

[Sander Eikelenboom <linux@xxxxxxxxxxxxxx>]

Friday, August 16, 2013, 10:03:31 AM, you wrote:

>>>> On 16.08.13 at 09:42, Sander Eikelenboom <linux@xxxxxxxxxxxxxx> wrote:

>> Friday, August 16, 2013, 9:21:01 AM, you wrote:
>> 
>>>>>> On 16.08.13 at 01:41, Sander Eikelenboom <linux@xxxxxxxxxxxxxx> wrote:
>>>> Ok i have to revise ... *sigh* ... it gets stuck with clocksource=pit as 
>>>> well (although not always it seems)
>> 
>>> Not that surprising. Doing it the other way round (suppressing
>>> use of cpuidle and cpufreq - first together, then separately -
>>> would seem like the more promising route with the information
>>> you gathered, with cpuidle being the more likely of the two.
>> 
>> Well not specifying "cpuidle" and "cpufreq=xen" on the xen boot line in 
>> grub, should not enable them for the clocksource=pit case right ?
>> I did exactly that on this boot.

> Depends on which cpuidle driver the system is using - at least in
> the mwait_idle case, omitting "cpuidle" doesn't help; you'd need
> to explicitly say "no-cpuidle" (or equivalent).

Hmm only the "no-cpuidle" is needed (cpufreq=xen can stay) to make the stalls 
disappear,
but makes me wonder how that is related to the commit the bisection found ..
machine has been running with cpuidle enabled for ages ..


> Jan




_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel