Re: [PATCH kvm-unit-tests] x86: fix last commit

[Bandan Das <bsd@redhat.com>]

Shih-Wei Li <shihwei@cs.columbia.edu> writes:

> Hi Paolo,
>
> I've tried to apply the patch, and found that it passed most of the
> problematic tests I mentioned earlier (IPI related, kvmclock_test).
> However, it stopped still at "s3" and couldn't finish it. Do you know
> what might go wrong?

Nothing is wrong, that's the way the test is. You need to resume from
qemu for it to proceed and it should quit with 1 for error or 0 for
success.

> Thanks,
> Shih-Wei
>
> On Thu, Jul 30, 2015 at 9:38 AM, Paolo Bonzini <pbonzini@redhat.com> wrote:
>> Keeping the selector that was loaded from the 32-bit GDT is okay, because
>> only code segment descriptors differ between 32- and 64-bit mode.  In
>> fact the same is true for %ss as well, so let's just remove the whole
>> segment loading from load_tss.
>>
>> Thanks to Bandan Das for debugging.
>>
>> Reported-by: Shih-Wei Li <shihwei@cs.columbia.edu>
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> ---
>>  x86/cstart64.S | 6 ------
>>  1 file changed, 6 deletions(-)
>>
>> diff --git a/x86/cstart64.S b/x86/cstart64.S
>> index 8d5ee2d..e947888 100644
>> --- a/x86/cstart64.S
>> +++ b/x86/cstart64.S
>> @@ -213,12 +213,6 @@ idt_descr:
>>
>>  load_tss:
>>         lidtq idt_descr
>> -       mov $0x10, %eax
>> -       mov %ax, %ds
>> -       mov %ax, %es
>> -       mov %ax, %fs
>> -       mov %ax, %gs
>> -       mov %ax, %ss
>>         mov $(APIC_DEFAULT_PHYS_BASE + APIC_ID), %eax
>>         mov (%rax), %eax
>>         shr $24, %eax
>> --
>> 1.8.3.1
>>