Xen devel - Oct 2006 - Why did Changeset 12028 remove the handling of ESR?

Hi Keir,
I found you did some clean-up in Changeset 12028.
I noticed you removed the handling of ESR(Error status register). This includes:
1) remove the handling of reading/writing APIC_ESR;
2) removing vlapic_check_vector(); when writing the entries in Local
APIC''s LVT, no calling vlapic_check_vector();
3) In cpu_get_apic_interrupt(), no checking whether the want-to-deliver vector
number < 0x10;
Could you explain the reason?
When installing x64 windows, we found without the handling of ESR, the
installation process would hang.
I think we might as well restore the handling of ESR?

Thanks!
 -- Dexuan

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

On 31/10/06 11:47, "Cui, Dexuan" <dexuan.cui@intel.com> wrote:
> I found you did some clean-up in Changeset 12028.
> I noticed you removed the handling of ESR(Error status register). This
> includes:
> 1) remove the handling of reading/writing APIC_ESR;
> 2) removing vlapic_check_vector(); when writing the entries in Local
APIC''s
> LVT, no calling vlapic_check_vector();
> 3) In cpu_get_apic_interrupt(), no checking whether the want-to-deliver
vector
> number < 0x10;
> Could you explain the reason?
> When installing x64 windows, we found without the handling of ESR, the
> installation process would hang.
> I think we might as well restore the handling of ESR?
I removed it because it was broken. The guest would always read err_status
as zero as it was never written to APIC_ESR in the vlapic register page. And
it did not seem worth fixing since I can hardly believe that an OS would
rely on defined ESR behaviour during boot (especially when there are various
errata relating to its behaviour). Surely there must be a deeper underlying
problem here?

 -- Keir



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