[PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[Liu, Jinsong]

[-- Attachment #1: Type: text/plain, Size: 293 bytes --]

Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

Win2000 will BSOD if cpus > 15, in order to compatible support win2000 HVM, this patch change HVM_MAX_VCPUS to 15.
This patch also fix a small bug related to HVM_MAX_VCPUS.

Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>

[-- Attachment #2: ras_vcpuhotplug_2.patch --]
[-- Type: application/octet-stream, Size: 1802 bytes --]

Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

Win2000 will BSOD if cpus > 15, in order to compatible support win2000 HVM, this patch change HVM_MAX_VCPUS to 15.
This patch also fix a small bug related to HVM_MAX_VCPUS.

Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>

diff -r fa4d31ee800f tools/python/xen/lowlevel/xc/xc.c
--- a/tools/python/xen/lowlevel/xc/xc.c	Thu Feb 11 14:38:39 2010 +0800
+++ b/tools/python/xen/lowlevel/xc/xc.c	Thu Feb 11 21:44:13 2010 +0800
@@ -916,7 +916,7 @@ static PyObject *pyxc_hvm_build(XcObject
     char *image;
     int memsize, target=-1, vcpus = 1, acpi = 0, apic = 1;
     PyObject *vcpu_avail_handle = NULL;
-    uint8_t vcpu_avail[HVM_MAX_VCPUS/8];
+    uint8_t vcpu_avail[(HVM_MAX_VCPUS + 7)/8];
 
     static char *kwd_list[] = { "domid",
                                 "memsize", "image", "target", "vcpus", 
diff -r fa4d31ee800f xen/include/public/hvm/hvm_info_table.h
--- a/xen/include/public/hvm/hvm_info_table.h	Thu Feb 11 14:38:39 2010 +0800
+++ b/xen/include/public/hvm/hvm_info_table.h	Thu Feb 11 21:44:13 2010 +0800
@@ -29,8 +29,8 @@
 #define HVM_INFO_OFFSET      0x800
 #define HVM_INFO_PADDR       ((HVM_INFO_PFN << 12) + HVM_INFO_OFFSET)
 
-/* Maximum we can support with current vLAPIC ID mapping. */
-#define HVM_MAX_VCPUS        128
+/* Maximum vcpu supported, limited by win2000 */
+#define HVM_MAX_VCPUS        15
 
 struct hvm_info_table {
     char        signature[8]; /* "HVM INFO" */
@@ -69,7 +69,7 @@ struct hvm_info_table {
     uint32_t    high_mem_pgend;
 
     /* Bitmap of which CPUs are online at boot time. */
-    uint8_t     vcpu_online[HVM_MAX_VCPUS/8];
+    uint8_t     vcpu_online[(HVM_MAX_VCPUS + 7)/8];
 };
 
 #endif /* __XEN_PUBLIC_HVM_HVM_INFO_TABLE_H__ */

[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

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

Re: [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[Christoph Egger]

On Thursday 11 February 2010 18:40:58 Liu, Jinsong wrote:
> Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug
>
> Win2000 will BSOD if cpus > 15, in order to compatible support win2000 HVM,
> this patch change HVM_MAX_VCPUS to 15. This patch also fix a small bug
> related to HVM_MAX_VCPUS.
>
> Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>

This patch makes no sense to me.
Didn't win2000 crash before the number of HVM_MAX_VCPUS
has been increased from 32 to 128 ?

Christoph


-- 
---to satisfy European Law for business letters:
Advanced Micro Devices GmbH
Karl-Hammerschmidt-Str. 34, 85609 Dornach b. Muenchen
Geschaeftsfuehrer: Andrew Bowd, Thomas M. McCoy, Giuliano Meroni
Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen
Registergericht Muenchen, HRB Nr. 43632

RE: [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[Jiang, Yunhong]

I think this is more about a regression. Originally hvmloader only create 14 Processor definition in dsdt and the HVM_MAX_VCPUS is not used in that code.Later, when the vCPU hotplug added, that 14 is changed to HVM_MAX_VCPUS and cause regression.

Thanks
--jyh

>-----Original Message-----
>From: Christoph Egger [mailto:Christoph.Egger@amd.com]
>Sent: Friday, February 12, 2010 2:00 AM
>To: xen-devel@lists.xensource.com
>Cc: Liu, Jinsong; Jiang, Yunhong; Keir Fraser
>Subject: Re: [Xen-devel] [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15,
>and fix a small bug
>
>On Thursday 11 February 2010 18:40:58 Liu, Jinsong wrote:
>> Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug
>>
>> Win2000 will BSOD if cpus > 15, in order to compatible support win2000 HVM,
>> this patch change HVM_MAX_VCPUS to 15. This patch also fix a small bug
>> related to HVM_MAX_VCPUS.
>>
>> Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
>
>This patch makes no sense to me.
>Didn't win2000 crash before the number of HVM_MAX_VCPUS
>has been increased from 32 to 128 ?
>
>Christoph
>
>
>--
>---to satisfy European Law for business letters:
>Advanced Micro Devices GmbH
>Karl-Hammerschmidt-Str. 34, 85609 Dornach b. Muenchen
>Geschaeftsfuehrer: Andrew Bowd, Thomas M. McCoy, Giuliano Meroni
>Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen
>Registergericht Muenchen, HRB Nr. 43632

Re: [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[Christoph Egger]

Ah, I see. Thanks for clarification.

Christoph


On Friday 12 February 2010 02:51:03 Jiang, Yunhong wrote:
> I think this is more about a regression. Originally hvmloader only create
> 14 Processor definition in dsdt and the HVM_MAX_VCPUS is not used in that
> code.Later, when the vCPU hotplug added, that 14 is changed to
> HVM_MAX_VCPUS and cause regression.
>
> Thanks
> --jyh
>
> >-----Original Message-----
> >From: Christoph Egger [mailto:Christoph.Egger@amd.com]
> >Sent: Friday, February 12, 2010 2:00 AM
> >To: xen-devel@lists.xensource.com
> >Cc: Liu, Jinsong; Jiang, Yunhong; Keir Fraser
> >Subject: Re: [Xen-devel] [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to
> > 15, and fix a small bug
> >
> >On Thursday 11 February 2010 18:40:58 Liu, Jinsong wrote:
> >> Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug
> >>
> >> Win2000 will BSOD if cpus > 15, in order to compatible support win2000
> >> HVM, this patch change HVM_MAX_VCPUS to 15. This patch also fix a small
> >> bug related to HVM_MAX_VCPUS.
> >>
> >> Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
> >
> >This patch makes no sense to me.
> >Didn't win2000 crash before the number of HVM_MAX_VCPUS
> >has been increased from 32 to 128 ?
> >
> >Christoph
> >
> >
> >--
> >---to satisfy European Law for business letters:
> >Advanced Micro Devices GmbH
> >Karl-Hammerschmidt-Str. 34, 85609 Dornach b. Muenchen
> >Geschaeftsfuehrer: Andrew Bowd, Thomas M. McCoy, Giuliano Meroni
> >Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen
> >Registergericht Muenchen, HRB Nr. 43632



-- 
---to satisfy European Law for business letters:
Advanced Micro Devices GmbH
Karl-Hammerschmidt-Str. 34, 85609 Dornach b. Muenchen
Geschaeftsfuehrer: Andrew Bowd, Thomas M. McCoy, Giuliano Meroni
Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen
Registergericht Muenchen, HRB Nr. 43632

Re: [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[Keir Fraser]

On 11/02/2010 17:40, "Liu, Jinsong" <jinsong.liu@intel.com> wrote:

> Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug
> 
> Win2000 will BSOD if cpus > 15, in order to compatible support win2000 HVM,
> this patch change HVM_MAX_VCPUS to 15.
> This patch also fix a small bug related to HVM_MAX_VCPUS.

Yeah, it's not really on to limit HVM #vcpus for all time, to deal with one
legacy OS. We'll have to build a legacy DSDT to deploy when
hvm_info->nr_vcpus<=15.

I guess I can look into that, and also I'm going to stop checking in the
pre-built dsdt.c file. It's now really big and constantly changing, and
people who build the Xen tree are just going to have to sort out to have
iasl installed. In fact the Makefile already grabs it automatically...

 -- Keir

> Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>

Re: Re: [PATCH 2/2] Vcpu hotplug: change HVM_MAX_VCPUS to 15, and fix a small bug

[Keir Fraser]

On 11/02/2010 18:27, "Keir Fraser" <keir.fraser@eu.citrix.com> wrote:

> Yeah, it's not really on to limit HVM #vcpus for all time, to deal with one
> legacy OS. We'll have to build a legacy DSDT to deploy when
> hvm_info->nr_vcpus<=15.
> 
> I guess I can look into that, and also I'm going to stop checking in the
> pre-built dsdt.c file.

See xen-unstable:20933, which should give best of both worlds (working W2K
for up to 15 processors; other OSes can continue to have up to HVM_MAX_VCPUS
processors).

 -- Keir