>>>>> "Manuel" == Manuel Bouyer <bou...@antioche.eu.org> writes:
Manuel> On Wed, Feb 22, 2012 at 10:15:50AM +0530, Cherry G. Mathew wrote:
>> Hi Manuel,
>>
>> On 22 February 2012 00:40, Manuel Bouyer <bou...@netbsd.org> wrote:
>> > Module Name: src > Committed By: bouyer > Date:
>> Tue Feb 21 19:10:13 UTC 2012
>> >
>> > Modified Files: > src/sys/arch/x86/x86: pmap.c >
>> src/sys/arch/xen/x86: cpu.c
>> >
>> > Log Message: > Avoid early use of xen_kpm_sync(); locks are not
>> available at this time. > Don't call cpu_init() twice.
>> >
>> > Makes LOCKDEBUG kernels boot again
>> >
>> >
>> > To generate a diff of this commit: > cvs rdiff -u -r1.166
>> -r1.167 src/sys/arch/x86/x86/pmap.c
>>
>> @@ -4183,14 +4183,30 @@ pte = pmap_pa2pte(pa) | PG_k | PG_V |
>> PG_RW; #ifdef XEN xpq_queue_pte_update(xpmap_ptetomach(&pdep[i]),
>> pte); - if (level == PTP_LEVELS) { #if defined(PAE) ||
>> defined(__x86_64__) - if (i >= PDIR_SLOT_KERN) { + if (level ==
>> PTP_LEVELS && i >= PDIR_SLOT_KERN) { + if (__predict_true( +
>> cpu_info_primary.ci_flags & CPUF_PRESENT)) {
>>
>>
>> Can we use that as a canonical MI method in Xen to detect if the
>> tls (gs/fs) has been setup ?
Manuel> probably not; it works here because we want to know if we
Manuel> already probed CPUs; I'm not sure CPUF_PRESENT will garantee
Manuel> that curcpu() will work.
I meant we could make it work, (it would already for amd64/xen since
cpu_init_msrs() is called from cpu_hatch()) since xen has its own cpu.c
Cheers,
--
Cherry
