[PATCH v2] libxl: put RSDP for PVH guest near 4GB

[PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

Instead of locating the RSDP table below 1MB put it just below 4GB
like the rest of the ACPI tables in case of PVH guests. This will
avoid punching more holes than necessary into the memory map.

Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Wei Liu <wei.liu2@citrix.com>
---
 tools/libxc/xc_dom_hvmloader.c | 2 +-
 tools/libxl/libxl_x86_acpi.c   | 5 ++---
 2 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
index 59f94e51e5..3f0bd65547 100644
--- a/tools/libxc/xc_dom_hvmloader.c
+++ b/tools/libxc/xc_dom_hvmloader.c
@@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
     struct xc_dom_seg seg;
     void *dest;
 
-    if ( module->length )
+    if ( module->length && !module->guest_addr_out )
     {
         if ( xc_dom_alloc_segment(dom, &seg, name, 0, module->length) )
             goto err;
diff --git a/tools/libxl/libxl_x86_acpi.c b/tools/libxl/libxl_x86_acpi.c
index 9a7c90467d..fe87418bc1 100644
--- a/tools/libxl/libxl_x86_acpi.c
+++ b/tools/libxl/libxl_x86_acpi.c
@@ -22,8 +22,6 @@
 
  /* Number of pages holding ACPI tables */
 #define NUM_ACPI_PAGES 16
-/* Store RSDP in the last 64 bytes of BIOS RO memory */
-#define RSDP_ADDRESS (0x100000 - 64)
 #define ACPI_INFO_PHYSICAL_ADDRESS 0xfc000000
 
 struct libxl_acpi_ctxt {
@@ -220,7 +218,8 @@ int libxl__dom_load_acpi(libxl__gc *gc,
 
     dom->acpi_modules[0].data = (void *)config.rsdp;
     dom->acpi_modules[0].length = 64;
-    dom->acpi_modules[0].guest_addr_out = RSDP_ADDRESS;
+    dom->acpi_modules[0].guest_addr_out = ACPI_INFO_PHYSICAL_ADDRESS +
+        (1 + acpi_pages_num) * libxl_ctxt.page_size;
 
     dom->acpi_modules[1].data = (void *)config.infop;
     dom->acpi_modules[1].length = 4096;
-- 
2.12.3


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

Ping?

On 01/12/17 15:14, Juergen Gross wrote:
> Instead of locating the RSDP table below 1MB put it just below 4GB
> like the rest of the ACPI tables in case of PVH guests. This will
> avoid punching more holes than necessary into the memory map.
> 
> Signed-off-by: Juergen Gross <jgross@suse.com>
> Acked-by: Wei Liu <wei.liu2@citrix.com>
> ---
>  tools/libxc/xc_dom_hvmloader.c | 2 +-
>  tools/libxl/libxl_x86_acpi.c   | 5 ++---
>  2 files changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
> index 59f94e51e5..3f0bd65547 100644
> --- a/tools/libxc/xc_dom_hvmloader.c
> +++ b/tools/libxc/xc_dom_hvmloader.c
> @@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
>      struct xc_dom_seg seg;
>      void *dest;
>  
> -    if ( module->length )
> +    if ( module->length && !module->guest_addr_out )
>      {
>          if ( xc_dom_alloc_segment(dom, &seg, name, 0, module->length) )
>              goto err;
> diff --git a/tools/libxl/libxl_x86_acpi.c b/tools/libxl/libxl_x86_acpi.c
> index 9a7c90467d..fe87418bc1 100644
> --- a/tools/libxl/libxl_x86_acpi.c
> +++ b/tools/libxl/libxl_x86_acpi.c
> @@ -22,8 +22,6 @@
>  
>   /* Number of pages holding ACPI tables */
>  #define NUM_ACPI_PAGES 16
> -/* Store RSDP in the last 64 bytes of BIOS RO memory */
> -#define RSDP_ADDRESS (0x100000 - 64)
>  #define ACPI_INFO_PHYSICAL_ADDRESS 0xfc000000
>  
>  struct libxl_acpi_ctxt {
> @@ -220,7 +218,8 @@ int libxl__dom_load_acpi(libxl__gc *gc,
>  
>      dom->acpi_modules[0].data = (void *)config.rsdp;
>      dom->acpi_modules[0].length = 64;
> -    dom->acpi_modules[0].guest_addr_out = RSDP_ADDRESS;
> +    dom->acpi_modules[0].guest_addr_out = ACPI_INFO_PHYSICAL_ADDRESS +
> +        (1 + acpi_pages_num) * libxl_ctxt.page_size;
>  
>      dom->acpi_modules[1].data = (void *)config.infop;
>      dom->acpi_modules[1].length = 4096;
> 


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Roger Pau Monné]

On Fri, Dec 01, 2017 at 03:14:07PM +0100, Juergen Gross wrote:
> Instead of locating the RSDP table below 1MB put it just below 4GB
> like the rest of the ACPI tables in case of PVH guests. This will
> avoid punching more holes than necessary into the memory map.
> 
> Signed-off-by: Juergen Gross <jgross@suse.com>
> Acked-by: Wei Liu <wei.liu2@citrix.com>
> ---
>  tools/libxc/xc_dom_hvmloader.c | 2 +-
>  tools/libxl/libxl_x86_acpi.c   | 5 ++---
>  2 files changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
> index 59f94e51e5..3f0bd65547 100644
> --- a/tools/libxc/xc_dom_hvmloader.c
> +++ b/tools/libxc/xc_dom_hvmloader.c
> @@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
>      struct xc_dom_seg seg;
>      void *dest;
>  
> -    if ( module->length )
> +    if ( module->length && !module->guest_addr_out )

Isn't that kind of a separate fix? AFAICT this just prevents
allocating memory if guest_addr_out is already set to a fixed
position.

The rest LGTM.

Thanks, Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 19/12/17 16:38, Roger Pau Monné wrote:
> On Fri, Dec 01, 2017 at 03:14:07PM +0100, Juergen Gross wrote:
>> Instead of locating the RSDP table below 1MB put it just below 4GB
>> like the rest of the ACPI tables in case of PVH guests. This will
>> avoid punching more holes than necessary into the memory map.
>>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>> ---
>>  tools/libxc/xc_dom_hvmloader.c | 2 +-
>>  tools/libxl/libxl_x86_acpi.c   | 5 ++---
>>  2 files changed, 3 insertions(+), 4 deletions(-)
>>
>> diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
>> index 59f94e51e5..3f0bd65547 100644
>> --- a/tools/libxc/xc_dom_hvmloader.c
>> +++ b/tools/libxc/xc_dom_hvmloader.c
>> @@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
>>      struct xc_dom_seg seg;
>>      void *dest;
>>  
>> -    if ( module->length )
>> +    if ( module->length && !module->guest_addr_out )
> 
> Isn't that kind of a separate fix? AFAICT this just prevents
> allocating memory if guest_addr_out is already set to a fixed
> position.

No, this is mandatory, as I have to skip the allocation for PVH, while
HVM guests really want the allocation to take place.


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Roger Pau Monné]

On Tue, Dec 19, 2017 at 04:46:37PM +0100, Juergen Gross wrote:
> On 19/12/17 16:38, Roger Pau Monné wrote:
> > On Fri, Dec 01, 2017 at 03:14:07PM +0100, Juergen Gross wrote:
> >> Instead of locating the RSDP table below 1MB put it just below 4GB
> >> like the rest of the ACPI tables in case of PVH guests. This will
> >> avoid punching more holes than necessary into the memory map.
> >>
> >> Signed-off-by: Juergen Gross <jgross@suse.com>
> >> Acked-by: Wei Liu <wei.liu2@citrix.com>
> >> ---
> >>  tools/libxc/xc_dom_hvmloader.c | 2 +-
> >>  tools/libxl/libxl_x86_acpi.c   | 5 ++---
> >>  2 files changed, 3 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
> >> index 59f94e51e5..3f0bd65547 100644
> >> --- a/tools/libxc/xc_dom_hvmloader.c
> >> +++ b/tools/libxc/xc_dom_hvmloader.c
> >> @@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
> >>      struct xc_dom_seg seg;
> >>      void *dest;
> >>  
> >> -    if ( module->length )
> >> +    if ( module->length && !module->guest_addr_out )
> > 
> > Isn't that kind of a separate fix? AFAICT this just prevents
> > allocating memory if guest_addr_out is already set to a fixed
> > position.
> 
> No, this is mandatory, as I have to skip the allocation for PVH, while
> HVM guests really want the allocation to take place.

Was this also a problem before? Other ACPI modules also set
guest_addr_out, and previously they would also get memory allocated.

Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 19/12/17 17:11, Roger Pau Monné wrote:
> On Tue, Dec 19, 2017 at 04:46:37PM +0100, Juergen Gross wrote:
>> On 19/12/17 16:38, Roger Pau Monné wrote:
>>> On Fri, Dec 01, 2017 at 03:14:07PM +0100, Juergen Gross wrote:
>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>> avoid punching more holes than necessary into the memory map.
>>>>
>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>> ---
>>>>  tools/libxc/xc_dom_hvmloader.c | 2 +-
>>>>  tools/libxl/libxl_x86_acpi.c   | 5 ++---
>>>>  2 files changed, 3 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
>>>> index 59f94e51e5..3f0bd65547 100644
>>>> --- a/tools/libxc/xc_dom_hvmloader.c
>>>> +++ b/tools/libxc/xc_dom_hvmloader.c
>>>> @@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
>>>>      struct xc_dom_seg seg;
>>>>      void *dest;
>>>>  
>>>> -    if ( module->length )
>>>> +    if ( module->length && !module->guest_addr_out )
>>>
>>> Isn't that kind of a separate fix? AFAICT this just prevents
>>> allocating memory if guest_addr_out is already set to a fixed
>>> position.
>>
>> No, this is mandatory, as I have to skip the allocation for PVH, while
>> HVM guests really want the allocation to take place.
> 
> Was this also a problem before? Other ACPI modules also set
> guest_addr_out, and previously they would also get memory allocated.

module_init_one() would only be called for the first ACPI module, which
happens to be the RSDP.


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Roger Pau Monné]

On Tue, Dec 19, 2017 at 05:20:41PM +0100, Juergen Gross wrote:
> On 19/12/17 17:11, Roger Pau Monné wrote:
> > On Tue, Dec 19, 2017 at 04:46:37PM +0100, Juergen Gross wrote:
> >> On 19/12/17 16:38, Roger Pau Monné wrote:
> >>> On Fri, Dec 01, 2017 at 03:14:07PM +0100, Juergen Gross wrote:
> >>>> Instead of locating the RSDP table below 1MB put it just below 4GB
> >>>> like the rest of the ACPI tables in case of PVH guests. This will
> >>>> avoid punching more holes than necessary into the memory map.
> >>>>
> >>>> Signed-off-by: Juergen Gross <jgross@suse.com>
> >>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
> >>>> ---
> >>>>  tools/libxc/xc_dom_hvmloader.c | 2 +-
> >>>>  tools/libxl/libxl_x86_acpi.c   | 5 ++---
> >>>>  2 files changed, 3 insertions(+), 4 deletions(-)
> >>>>
> >>>> diff --git a/tools/libxc/xc_dom_hvmloader.c b/tools/libxc/xc_dom_hvmloader.c
> >>>> index 59f94e51e5..3f0bd65547 100644
> >>>> --- a/tools/libxc/xc_dom_hvmloader.c
> >>>> +++ b/tools/libxc/xc_dom_hvmloader.c
> >>>> @@ -136,7 +136,7 @@ static int module_init_one(struct xc_dom_image *dom,
> >>>>      struct xc_dom_seg seg;
> >>>>      void *dest;
> >>>>  
> >>>> -    if ( module->length )
> >>>> +    if ( module->length && !module->guest_addr_out )
> >>>
> >>> Isn't that kind of a separate fix? AFAICT this just prevents
> >>> allocating memory if guest_addr_out is already set to a fixed
> >>> position.
> >>
> >> No, this is mandatory, as I have to skip the allocation for PVH, while
> >> HVM guests really want the allocation to take place.
> > 
> > Was this also a problem before? Other ACPI modules also set
> > guest_addr_out, and previously they would also get memory allocated.
> 
> module_init_one() would only be called for the first ACPI module, which
> happens to be the RSDP.

OK, I have to admit I find all this quite confusing. In any case, the
approach seems correct to me.

Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>

Thanks, Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Ian Jackson]

Juergen Gross writes ("[PATCH v2] libxl: put RSDP for PVH guest near 4GB"):
> Instead of locating the RSDP table below 1MB put it just below 4GB
> like the rest of the ACPI tables in case of PVH guests. This will
> avoid punching more holes than necessary into the memory map.
> 
> Signed-off-by: Juergen Gross <jgross@suse.com>
> Acked-by: Wei Liu <wei.liu2@citrix.com>

I had held off committing this despite it having Wei's ack because I
didn't feel confident about it (mostly, because it seems to be about
x86/ACPI things I don't feel I understand).

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

Wei,

On 01/12/17 15:14, Juergen Gross wrote:
> Instead of locating the RSDP table below 1MB put it just below 4GB
> like the rest of the ACPI tables in case of PVH guests. This will
> avoid punching more holes than necessary into the memory map.
> 
> Signed-off-by: Juergen Gross <jgross@suse.com>
> Acked-by: Wei Liu <wei.liu2@citrix.com>

Mind applying this one?


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Wei Liu]

On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
> Wei,
> 
> On 01/12/17 15:14, Juergen Gross wrote:
> > Instead of locating the RSDP table below 1MB put it just below 4GB
> > like the rest of the ACPI tables in case of PVH guests. This will
> > avoid punching more holes than necessary into the memory map.
> > 
> > Signed-off-by: Juergen Gross <jgross@suse.com>
> > Acked-by: Wei Liu <wei.liu2@citrix.com>
> 
> Mind applying this one?

Don't worry, it is in my queue.

Will come to this and other patches I accumulated soon.

Wei.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 01/18/2018 05:33 AM, Wei Liu wrote:
> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>> Wei,
>>
>> On 01/12/17 15:14, Juergen Gross wrote:
>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>> like the rest of the ACPI tables in case of PVH guests. This will
>>> avoid punching more holes than necessary into the memory map.
>>>
>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>> Mind applying this one?
> Don't worry, it is in my queue.
>
> Will come to this and other patches I accumulated soon.
>
> Wei.

This requires kernel changes, doesn't it?

https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html

And this series apparently never made it to the tree.

PVH guests are broken now on staging.

-boris



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[George Dunlap]

On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
<boris.ostrovsky@oracle.com> wrote:
> On 01/18/2018 05:33 AM, Wei Liu wrote:
>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>> Wei,
>>>
>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>> avoid punching more holes than necessary into the memory map.
>>>>
>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>> Mind applying this one?
>> Don't worry, it is in my queue.
>>
>> Will come to this and other patches I accumulated soon.
>>
>> Wei.
>
> This requires kernel changes, doesn't it?
>
> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>
> And this series apparently never made it to the tree.
>
> PVH guests are broken now on staging.

And the Linux side of PVH is officially supported now, right?  Which
means we need to have some way of being backwards compatible with
kernels without that patch, even if it does get  backported.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Roger Pau Monné]

On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
> <boris.ostrovsky@oracle.com> wrote:
> > On 01/18/2018 05:33 AM, Wei Liu wrote:
> >> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
> >>> Wei,
> >>>
> >>> On 01/12/17 15:14, Juergen Gross wrote:
> >>>> Instead of locating the RSDP table below 1MB put it just below 4GB
> >>>> like the rest of the ACPI tables in case of PVH guests. This will
> >>>> avoid punching more holes than necessary into the memory map.
> >>>>
> >>>> Signed-off-by: Juergen Gross <jgross@suse.com>
> >>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
> >>> Mind applying this one?
> >> Don't worry, it is in my queue.
> >>
> >> Will come to this and other patches I accumulated soon.
> >>
> >> Wei.
> >
> > This requires kernel changes, doesn't it?
> >
> > https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
> >
> > And this series apparently never made it to the tree.
> >
> > PVH guests are broken now on staging.
> 
> And the Linux side of PVH is officially supported now, right?  Which
> means we need to have some way of being backwards compatible with
> kernels without that patch, even if it does get  backported.

Linux had a bug that on PVH it expected to find the RSDP at the low
1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
should be below 1MiB. There's just one comment in the start_info
header that says:

"Xen on x86 will always try to place all the data below the 4GiB
boundary."

And the RSDP address must be fetched from the rsdp_paddr field of the
start_info.

I agree that the fact that Linux is broken is bad, but should not
force us to change the protocol. What's more, when booting a PVHv2
Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
going to find the firmware provided RSDP and everything is going to
blow up.

IMO, Linux must be fixed and the patch backported up to 4.11.

Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 24/01/18 11:54, Roger Pau Monné wrote:
> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>> <boris.ostrovsky@oracle.com> wrote:
>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>> Wei,
>>>>>
>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>
>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>> Mind applying this one?
>>>> Don't worry, it is in my queue.
>>>>
>>>> Will come to this and other patches I accumulated soon.
>>>>
>>>> Wei.
>>>
>>> This requires kernel changes, doesn't it?
>>>
>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>
>>> And this series apparently never made it to the tree.
>>>
>>> PVH guests are broken now on staging.
>>
>> And the Linux side of PVH is officially supported now, right?  Which
>> means we need to have some way of being backwards compatible with
>> kernels without that patch, even if it does get  backported.
> 
> Linux had a bug that on PVH it expected to find the RSDP at the low
> 1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
> should be below 1MiB. There's just one comment in the start_info
> header that says:
> 
> "Xen on x86 will always try to place all the data below the 4GiB
> boundary."
> 
> And the RSDP address must be fetched from the rsdp_paddr field of the
> start_info.
> 
> I agree that the fact that Linux is broken is bad, but should not
> force us to change the protocol. What's more, when booting a PVHv2
> Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
> going to find the firmware provided RSDP and everything is going to
> blow up.

Right.

My patches to repair this in the Linux kernel are on hold as they are
touching the boot parameters between grub2 and the kernel, too. And
this caused the kernel no longer booting in some cases as most distros
are shipping a grub2 breaking this interface via a non-upstream patch.

> IMO, Linux must be fixed and the patch backported up to 4.11.

I'll try soon.


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 01/24/2018 07:06 AM, Juergen Gross wrote:
> On 24/01/18 11:54, Roger Pau Monné wrote:
>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>> <boris.ostrovsky@oracle.com> wrote:
>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>> Wei,
>>>>>>
>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>
>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>> Mind applying this one?
>>>>> Don't worry, it is in my queue.
>>>>>
>>>>> Will come to this and other patches I accumulated soon.
>>>>>
>>>>> Wei.
>>>> This requires kernel changes, doesn't it?
>>>>
>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>
>>>> And this series apparently never made it to the tree.
>>>>
>>>> PVH guests are broken now on staging.
>>> And the Linux side of PVH is officially supported now, right?


AFAIK PVH is still considered a tech preview --- Linux or Xen.


>>>   Which
>>> means we need to have some way of being backwards compatible with
>>> kernels without that patch, even if it does get  backported.
>> Linux had a bug that on PVH it expected to find the RSDP at the low
>> 1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
>> should be below 1MiB. There's just one comment in the start_info
>> header that says:
>>
>> "Xen on x86 will always try to place all the data below the 4GiB
>> boundary."
>>
>> And the RSDP address must be fetched from the rsdp_paddr field of the
>> start_info.
>>
>> I agree that the fact that Linux is broken is bad, but should not
>> force us to change the protocol. What's more, when booting a PVHv2
>> Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
>> going to find the firmware provided RSDP and everything is going to
>> blow up.
> Right.
>
> My patches to repair this in the Linux kernel are on hold as they are
> touching the boot parameters between grub2 and the kernel, too. And
> this caused the kernel no longer booting in some cases as most distros
> are shipping a grub2 breaking this interface via a non-upstream patch.
>
>> IMO, Linux must be fixed and the patch backported up to 4.11.
> I'll try soon.

For backports we don't need bootparams changes, do we?

I suspect we can do as little as removing "#ifdef CONFIG_KEXEC" around
acpi_rsdp in drivers/acpi/osl.c and then assigning it the value in
pvh_start_info.rsdp_paddr. (I haven't tried it)

-boris




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 24/01/18 15:10, Boris Ostrovsky wrote:
> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>> Wei,
>>>>>>>
>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>
>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>> Mind applying this one?
>>>>>> Don't worry, it is in my queue.
>>>>>>
>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>
>>>>>> Wei.
>>>>> This requires kernel changes, doesn't it?
>>>>>
>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>
>>>>> And this series apparently never made it to the tree.
>>>>>
>>>>> PVH guests are broken now on staging.
>>>> And the Linux side of PVH is officially supported now, right?
> 
> 
> AFAIK PVH is still considered a tech preview --- Linux or Xen.
> 
> 
>>>>   Which
>>>> means we need to have some way of being backwards compatible with
>>>> kernels without that patch, even if it does get  backported.
>>> Linux had a bug that on PVH it expected to find the RSDP at the low
>>> 1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
>>> should be below 1MiB. There's just one comment in the start_info
>>> header that says:
>>>
>>> "Xen on x86 will always try to place all the data below the 4GiB
>>> boundary."
>>>
>>> And the RSDP address must be fetched from the rsdp_paddr field of the
>>> start_info.
>>>
>>> I agree that the fact that Linux is broken is bad, but should not
>>> force us to change the protocol. What's more, when booting a PVHv2
>>> Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
>>> going to find the firmware provided RSDP and everything is going to
>>> blow up.
>> Right.
>>
>> My patches to repair this in the Linux kernel are on hold as they are
>> touching the boot parameters between grub2 and the kernel, too. And
>> this caused the kernel no longer booting in some cases as most distros
>> are shipping a grub2 breaking this interface via a non-upstream patch.
>>
>>> IMO, Linux must be fixed and the patch backported up to 4.11.
>> I'll try soon.
> 
> For backports we don't need bootparams changes, do we?
> 
> I suspect we can do as little as removing "#ifdef CONFIG_KEXEC" around
> acpi_rsdp in drivers/acpi/osl.c and then assigning it the value in
> pvh_start_info.rsdp_paddr. (I haven't tried it)

That was the approach I started with (more or less). It was seen as a
layering violation.

I'd rather take the approach like in

https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?h=x86/boot&id=0c89cf36424f7c1177de8a5712514d7cc2eb369f

and have a PVH specific acpi_arch_get_root_pointer() instead one in
arch/x86/kernel/acpi/boot.c. This can later be modified easily to
take the RSDP address from boot params instead.


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 01/24/2018 09:25 AM, Juergen Gross wrote:
> On 24/01/18 15:10, Boris Ostrovsky wrote:
>>
>> I suspect we can do as little as removing "#ifdef CONFIG_KEXEC" around
>> acpi_rsdp in drivers/acpi/osl.c and then assigning it the value in
>> pvh_start_info.rsdp_paddr. (I haven't tried it)
> That was the approach I started with (more or less). It was seen as a
> layering violation.
>
> I'd rather take the approach like in
>
> https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?h=x86/boot&id=0c89cf36424f7c1177de8a5712514d7cc2eb369f
>
> and have a PVH specific acpi_arch_get_root_pointer() instead one in
> arch/x86/kernel/acpi/boot.c. This can later be modified easily to
> take the RSDP address from boot params instead.

Sure.

In the meantime I assume this patch will be reverted from staging until
Linux is taken care of?

-boris


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Wei Liu]

On Wed, Jan 24, 2018 at 09:42:57AM -0500, Boris Ostrovsky wrote:
> On 01/24/2018 09:25 AM, Juergen Gross wrote:
> > On 24/01/18 15:10, Boris Ostrovsky wrote:
> >>
> >> I suspect we can do as little as removing "#ifdef CONFIG_KEXEC" around
> >> acpi_rsdp in drivers/acpi/osl.c and then assigning it the value in
> >> pvh_start_info.rsdp_paddr. (I haven't tried it)
> > That was the approach I started with (more or less). It was seen as a
> > layering violation.
> >
> > I'd rather take the approach like in
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?h=x86/boot&id=0c89cf36424f7c1177de8a5712514d7cc2eb369f
> >
> > and have a PVH specific acpi_arch_get_root_pointer() instead one in
> > arch/x86/kernel/acpi/boot.c. This can later be modified easily to
> > take the RSDP address from boot params instead.
> 
> Sure.
> 
> In the meantime I assume this patch will be reverted from staging until
> Linux is taken care of?

Do you mean we need to wait until all stable kernels get the backport?

Wei.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[George Dunlap]

On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
<boris.ostrovsky@oracle.com> wrote:
> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>> Wei,
>>>>>>>
>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>
>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>> Mind applying this one?
>>>>>> Don't worry, it is in my queue.
>>>>>>
>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>
>>>>>> Wei.
>>>>> This requires kernel changes, doesn't it?
>>>>>
>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>
>>>>> And this series apparently never made it to the tree.
>>>>>
>>>>> PVH guests are broken now on staging.
>>>> And the Linux side of PVH is officially supported now, right?
>
>
> AFAIK PVH is still considered a tech preview --- Linux or Xen.

From SUPPORT.md:

### x86/PVH guest

    Status: Supported

I was under the impression that PVH guest in Linux was complete and
stable as of Linux 4.11.  If that's not true it should have been
brought up during the 4.10 development cycle, where we declared PVH
domUs as "supported".

>>>>   Which
>>>> means we need to have some way of being backwards compatible with
>>>> kernels without that patch, even if it does get  backported.
>>> Linux had a bug that on PVH it expected to find the RSDP at the low
>>> 1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
>>> should be below 1MiB. There's just one comment in the start_info
>>> header that says:
>>>
>>> "Xen on x86 will always try to place all the data below the 4GiB
>>> boundary."
>>>
>>> And the RSDP address must be fetched from the rsdp_paddr field of the
>>> start_info.
>>>
>>> I agree that the fact that Linux is broken is bad, but should not
>>> force us to change the protocol. What's more, when booting a PVHv2
>>> Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
>>> going to find the firmware provided RSDP and everything is going to
>>> blow up.

FWIW I can buy this argument.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 24/01/18 16:07, George Dunlap wrote:
> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
> <boris.ostrovsky@oracle.com> wrote:
>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>> Wei,
>>>>>>>>
>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>
>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>> Mind applying this one?
>>>>>>> Don't worry, it is in my queue.
>>>>>>>
>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>
>>>>>>> Wei.
>>>>>> This requires kernel changes, doesn't it?
>>>>>>
>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>
>>>>>> And this series apparently never made it to the tree.
>>>>>>
>>>>>> PVH guests are broken now on staging.
>>>>> And the Linux side of PVH is officially supported now, right?
>>
>>
>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
> 
> From SUPPORT.md:
> 
> ### x86/PVH guest
> 
>     Status: Supported
> 
> I was under the impression that PVH guest in Linux was complete and
> stable as of Linux 4.11.  If that's not true it should have been
> brought up during the 4.10 development cycle, where we declared PVH
> domUs as "supported".

So what is the problem here?

- current Linux can't be booted as PVH guest with xen-unstable due to
  a bug in Linux, patches for Linux are being worked on
- booting Linux as PVH guest with xen 4.10 is working

Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[George Dunlap]

On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
> On 24/01/18 16:07, George Dunlap wrote:
>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>> <boris.ostrovsky@oracle.com> wrote:
>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>> Wei,
>>>>>>>>>
>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>
>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>> Mind applying this one?
>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>
>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>
>>>>>>>> Wei.
>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>
>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>
>>>>>>> And this series apparently never made it to the tree.
>>>>>>>
>>>>>>> PVH guests are broken now on staging.
>>>>>> And the Linux side of PVH is officially supported now, right?
>>>
>>>
>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>
>> From SUPPORT.md:
>>
>> ### x86/PVH guest
>>
>>     Status: Supported
>>
>> I was under the impression that PVH guest in Linux was complete and
>> stable as of Linux 4.11.  If that's not true it should have been
>> brought up during the 4.10 development cycle, where we declared PVH
>> domUs as "supported".
>
> So what is the problem here?
>
> - current Linux can't be booted as PVH guest with xen-unstable due to
>   a bug in Linux, patches for Linux are being worked on
> - booting Linux as PVH guest with xen 4.10 is working

I was responding to Boris's claim that PVH is considered tech preview.
I can't say anything one way or the other about PVH in Linux, but PVH
in Xen is definitely now considered supported.

My subsequent response to Roger ("FWIW I can buy this argument") was
meant to indicate I didn't have any more objection to the approach you
guys were planning on taking.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 01/24/2018 10:26 AM, George Dunlap wrote:
> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>> On 24/01/18 16:07, George Dunlap wrote:
>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>> <boris.ostrovsky@oracle.com> wrote:
>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>> Wei,
>>>>>>>>>>
>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>
>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>> Mind applying this one?
>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>
>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>
>>>>>>>>> Wei.
>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>
>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>
>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>
>>>>>>>> PVH guests are broken now on staging.
>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>
>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>> From SUPPORT.md:
>>>
>>> ### x86/PVH guest
>>>
>>>     Status: Supported
>>>
>>> I was under the impression that PVH guest in Linux was complete and
>>> stable as of Linux 4.11.  If that's not true it should have been
>>> brought up during the 4.10 development cycle, where we declared PVH
>>> domUs as "supported".


OK, I missed that, sorry.

I though we were going to wait for PCI passthrough to become available
until declaring it supported.


>> So what is the problem here?
>>
>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>   a bug in Linux, patches for Linux are being worked on

I would prefer for patches to first show up in Linux. Otherwise we will
be adding Xen code without any means to test it with PVH guests. And
then, when Linux does get proper RSDP support for PVH and tries to boot,
we may discover that all sorts of unrelated things are broken.

-boris


>> - booting Linux as PVH guest with xen 4.10 is working
> I was responding to Boris's claim that PVH is considered tech preview.
> I can't say anything one way or the other about PVH in Linux, but PVH
> in Xen is definitely now considered supported.
>
> My subsequent response to Roger ("FWIW I can buy this argument") was
> meant to indicate I didn't have any more objection to the approach you
> guys were planning on taking.
>
>  -George


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 24/01/18 17:10, Boris Ostrovsky wrote:
> On 01/24/2018 10:26 AM, George Dunlap wrote:
>> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>>> On 24/01/18 16:07, George Dunlap wrote:
>>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>>> Wei,
>>>>>>>>>>>
>>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>>
>>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>>> Mind applying this one?
>>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>>
>>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>>
>>>>>>>>>> Wei.
>>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>>
>>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>>
>>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>>
>>>>>>>>> PVH guests are broken now on staging.
>>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>>
>>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>>> From SUPPORT.md:
>>>>
>>>> ### x86/PVH guest
>>>>
>>>>     Status: Supported
>>>>
>>>> I was under the impression that PVH guest in Linux was complete and
>>>> stable as of Linux 4.11.  If that's not true it should have been
>>>> brought up during the 4.10 development cycle, where we declared PVH
>>>> domUs as "supported".
> 
> 
> OK, I missed that, sorry.
> 
> I though we were going to wait for PCI passthrough to become available
> until declaring it supported.
> 
> 
>>> So what is the problem here?
>>>
>>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>>   a bug in Linux, patches for Linux are being worked on
> 
> I would prefer for patches to first show up in Linux. Otherwise we will
> be adding Xen code without any means to test it with PVH guests. And
> then, when Linux does get proper RSDP support for PVH and tries to boot,
> we may discover that all sorts of unrelated things are broken.

I wanted to do it this way.

My Linux patches were accepted for 4.16 and only then I sent the Xen
patches. Unfortunately the grub2 bug resulted in the Linux patches being
put on hold.


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 01/24/2018 11:15 AM, Juergen Gross wrote:
> On 24/01/18 17:10, Boris Ostrovsky wrote:
>
>>
>>>> So what is the problem here?
>>>>
>>>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>>>   a bug in Linux, patches for Linux are being worked on
>> I would prefer for patches to first show up in Linux. Otherwise we will
>> be adding Xen code without any means to test it with PVH guests. And
>> then, when Linux does get proper RSDP support for PVH and tries to boot,
>> we may discover that all sorts of unrelated things are broken.
> I wanted to do it this way.
>
> My Linux patches were accepted for 4.16 and only then I sent the Xen
> patches. Unfortunately the grub2 bug resulted in the Linux patches being
> put on hold.


And until this is all resolved we should revert this patch.

-boris

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Sander Eikelenboom]

On 24/01/18 16:26, George Dunlap wrote:
> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>> On 24/01/18 16:07, George Dunlap wrote:
>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>> <boris.ostrovsky@oracle.com> wrote:
>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>> Wei,
>>>>>>>>>>
>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>
>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>> Mind applying this one?
>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>
>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>
>>>>>>>>> Wei.
>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>
>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>
>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>
>>>>>>>> PVH guests are broken now on staging.
>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>
>>>>
>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>>
>>> From SUPPORT.md:
>>>
>>> ### x86/PVH guest
>>>
>>>     Status: Supported
>>>
>>> I was under the impression that PVH guest in Linux was complete and
>>> stable as of Linux 4.11.  If that's not true it should have been
>>> brought up during the 4.10 development cycle, where we declared PVH
>>> domUs as "supported".
>>
>> So what is the problem here?
>>
>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>   a bug in Linux, patches for Linux are being worked on
>> - booting Linux as PVH guest with xen 4.10 is working
> 
> I was responding to Boris's claim that PVH is considered tech preview.
> I can't say anything one way or the other about PVH in Linux, but PVH
> in Xen is definitely now considered supported.
> 
> My subsequent response to Roger ("FWIW I can buy this argument") was
> meant to indicate I didn't have any more objection to the approach you
> guys were planning on taking.
> 
>  -George

L.S.,

Seems I lost track, is there any progress on this issue ?
(doesn't seem a fix has landed in 4.16-rc2 yet).

--
Sander


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 19/02/18 10:47, Sander Eikelenboom wrote:
> On 24/01/18 16:26, George Dunlap wrote:
>> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>>> On 24/01/18 16:07, George Dunlap wrote:
>>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>>> Wei,
>>>>>>>>>>>
>>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>>
>>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>>> Mind applying this one?
>>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>>
>>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>>
>>>>>>>>>> Wei.
>>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>>
>>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>>
>>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>>
>>>>>>>>> PVH guests are broken now on staging.
>>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>>
>>>>>
>>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>>>
>>>> From SUPPORT.md:
>>>>
>>>> ### x86/PVH guest
>>>>
>>>>     Status: Supported
>>>>
>>>> I was under the impression that PVH guest in Linux was complete and
>>>> stable as of Linux 4.11.  If that's not true it should have been
>>>> brought up during the 4.10 development cycle, where we declared PVH
>>>> domUs as "supported".
>>>
>>> So what is the problem here?
>>>
>>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>>   a bug in Linux, patches for Linux are being worked on
>>> - booting Linux as PVH guest with xen 4.10 is working
>>
>> I was responding to Boris's claim that PVH is considered tech preview.
>> I can't say anything one way or the other about PVH in Linux, but PVH
>> in Xen is definitely now considered supported.
>>
>> My subsequent response to Roger ("FWIW I can buy this argument") was
>> meant to indicate I didn't have any more objection to the approach you
>> guys were planning on taking.
>>
>>  -George
> 
> L.S.,
> 
> Seems I lost track, is there any progress on this issue ?
> (doesn't seem a fix has landed in 4.16-rc2 yet).

Just sent a new patch series.


Juergen


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Sander Eikelenboom]

On 19/02/18 11:16, Juergen Gross wrote:
> On 19/02/18 10:47, Sander Eikelenboom wrote:
>> On 24/01/18 16:26, George Dunlap wrote:
>>> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>>>> On 24/01/18 16:07, George Dunlap wrote:
>>>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>>>> Wei,
>>>>>>>>>>>>
>>>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>>>> Mind applying this one?
>>>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>>>
>>>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>>>
>>>>>>>>>>> Wei.
>>>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>>>
>>>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>>>
>>>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>>>
>>>>>>>>>> PVH guests are broken now on staging.
>>>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>>>
>>>>>>
>>>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>>>>
>>>>> From SUPPORT.md:
>>>>>
>>>>> ### x86/PVH guest
>>>>>
>>>>>     Status: Supported
>>>>>
>>>>> I was under the impression that PVH guest in Linux was complete and
>>>>> stable as of Linux 4.11.  If that's not true it should have been
>>>>> brought up during the 4.10 development cycle, where we declared PVH
>>>>> domUs as "supported".
>>>>
>>>> So what is the problem here?
>>>>
>>>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>>>   a bug in Linux, patches for Linux are being worked on
>>>> - booting Linux as PVH guest with xen 4.10 is working
>>>
>>> I was responding to Boris's claim that PVH is considered tech preview.
>>> I can't say anything one way or the other about PVH in Linux, but PVH
>>> in Xen is definitely now considered supported.
>>>
>>> My subsequent response to Roger ("FWIW I can buy this argument") was
>>> meant to indicate I didn't have any more objection to the approach you
>>> guys were planning on taking.
>>>
>>>  -George
>>
>> L.S.,
>>
>> Seems I lost track, is there any progress on this issue ?
>> (doesn't seem a fix has landed in 4.16-rc2 yet).
> 
> Just sent a new patch series.

Just tested and it works fine here.

--
Sander

> 
> Juergen
> 


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Sander Eikelenboom]

On 19/02/18 22:13, Sander Eikelenboom wrote:
> On 19/02/18 11:16, Juergen Gross wrote:
>> On 19/02/18 10:47, Sander Eikelenboom wrote:
>>> On 24/01/18 16:26, George Dunlap wrote:
>>>> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>>>>> On 24/01/18 16:07, George Dunlap wrote:
>>>>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>>>>> Wei,
>>>>>>>>>>>>>
>>>>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>>>>> Mind applying this one?
>>>>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>>>>
>>>>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>>>>
>>>>>>>>>>>> Wei.
>>>>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>>>>
>>>>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>>>>
>>>>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>>>>
>>>>>>>>>>> PVH guests are broken now on staging.
>>>>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>>>>
>>>>>>>
>>>>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>>>>>
>>>>>> From SUPPORT.md:
>>>>>>
>>>>>> ### x86/PVH guest
>>>>>>
>>>>>>     Status: Supported
>>>>>>
>>>>>> I was under the impression that PVH guest in Linux was complete and
>>>>>> stable as of Linux 4.11.  If that's not true it should have been
>>>>>> brought up during the 4.10 development cycle, where we declared PVH
>>>>>> domUs as "supported".
>>>>>
>>>>> So what is the problem here?
>>>>>
>>>>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>>>>   a bug in Linux, patches for Linux are being worked on
>>>>> - booting Linux as PVH guest with xen 4.10 is working
>>>>
>>>> I was responding to Boris's claim that PVH is considered tech preview.
>>>> I can't say anything one way or the other about PVH in Linux, but PVH
>>>> in Xen is definitely now considered supported.
>>>>
>>>> My subsequent response to Roger ("FWIW I can buy this argument") was
>>>> meant to indicate I didn't have any more objection to the approach you
>>>> guys were planning on taking.
>>>>
>>>>  -George
>>>
>>> L.S.,
>>>
>>> Seems I lost track, is there any progress on this issue ?
>>> (doesn't seem a fix has landed in 4.16-rc2 yet).
>>
>> Just sent a new patch series.
> 
> Just tested and it works fine here.

Hi Juergen,

I don't know by which tree those patches should arrive at Linus,
so i can't check if they fell through the cracks somewhere, but 4.16-rc5
hasn't got them yet.

--
Sander


> 
> --
> Sander
> 
>>
>> Juergen
>>
> 


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 03/12/2018 03:26 PM, Sander Eikelenboom wrote:
> On 19/02/18 22:13, Sander Eikelenboom wrote:
>> On 19/02/18 11:16, Juergen Gross wrote:
>>> On 19/02/18 10:47, Sander Eikelenboom wrote:
>>>> On 24/01/18 16:26, George Dunlap wrote:
>>>>> On Wed, Jan 24, 2018 at 3:20 PM, Juergen Gross <jgross@suse.com> wrote:
>>>>>> On 24/01/18 16:07, George Dunlap wrote:
>>>>>>> On Wed, Jan 24, 2018 at 2:10 PM, Boris Ostrovsky
>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>> On 01/24/2018 07:06 AM, Juergen Gross wrote:
>>>>>>>>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>>>>>>>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>>>>>>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>>>>>>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>>>>>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>>>>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>>>>>>>>> Wei,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>>>>>>>>> Mind applying this one?
>>>>>>>>>>>>> Don't worry, it is in my queue.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Wei.
>>>>>>>>>>>> This requires kernel changes, doesn't it?
>>>>>>>>>>>>
>>>>>>>>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>>>>>>>>
>>>>>>>>>>>> And this series apparently never made it to the tree.
>>>>>>>>>>>>
>>>>>>>>>>>> PVH guests are broken now on staging.
>>>>>>>>>>> And the Linux side of PVH is officially supported now, right?
>>>>>>>>
>>>>>>>> AFAIK PVH is still considered a tech preview --- Linux or Xen.
>>>>>>> From SUPPORT.md:
>>>>>>>
>>>>>>> ### x86/PVH guest
>>>>>>>
>>>>>>>     Status: Supported
>>>>>>>
>>>>>>> I was under the impression that PVH guest in Linux was complete and
>>>>>>> stable as of Linux 4.11.  If that's not true it should have been
>>>>>>> brought up during the 4.10 development cycle, where we declared PVH
>>>>>>> domUs as "supported".
>>>>>> So what is the problem here?
>>>>>>
>>>>>> - current Linux can't be booted as PVH guest with xen-unstable due to
>>>>>>   a bug in Linux, patches for Linux are being worked on
>>>>>> - booting Linux as PVH guest with xen 4.10 is working
>>>>> I was responding to Boris's claim that PVH is considered tech preview.
>>>>> I can't say anything one way or the other about PVH in Linux, but PVH
>>>>> in Xen is definitely now considered supported.
>>>>>
>>>>> My subsequent response to Roger ("FWIW I can buy this argument") was
>>>>> meant to indicate I didn't have any more objection to the approach you
>>>>> guys were planning on taking.
>>>>>
>>>>>  -George
>>>> L.S.,
>>>>
>>>> Seems I lost track, is there any progress on this issue ?
>>>> (doesn't seem a fix has landed in 4.16-rc2 yet).
>>> Just sent a new patch series.
>> Just tested and it works fine here.
> Hi Juergen,
>
> I don't know by which tree those patches should arrive at Linus,
> so i can't check if they fell through the cracks somewhere, but 4.16-rc5
> hasn't got them yet.
>


I was just about to send this exact question. Last I saw a note from
Ingo that he put it in tip:x86/boot
(https://lkml.org/lkml/2017/12/11/358) but I don't see it there. (And
I'd think it would have been pulled by Linus by now anyway)

-boris

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 12/03/18 20:26, Sander Eikelenboom wrote:
> Hi Juergen,
> 
> I don't know by which tree those patches should arrive at Linus,
> so i can't check if they fell through the cracks somewhere, but 4.16-rc5
> hasn't got them yet.

They are queued for 4.17 in:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/boot


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Boris Ostrovsky]

On 03/13/2018 01:27 AM, Juergen Gross wrote:
> On 12/03/18 20:26, Sander Eikelenboom wrote:
>> Hi Juergen,
>>
>> I don't know by which tree those patches should arrive at Linus,
>> so i can't check if they fell through the cracks somewhere, but 4.16-rc5
>> hasn't got them yet.
> 
> They are queued for 4.17 in:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/boot


I may not be looking at the right place but I don't see it there:

https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/tree/arch/x86/xen/enlighten_pvh.c?h=x86/boot

-boris

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 13/03/18 18:51, Boris Ostrovsky wrote:
> 
> 
> On 03/13/2018 01:27 AM, Juergen Gross wrote:
>> On 12/03/18 20:26, Sander Eikelenboom wrote:
>>> Hi Juergen,
>>>
>>> I don't know by which tree those patches should arrive at Linus,
>>> so i can't check if they fell through the cracks somewhere, but 4.16-rc5
>>> hasn't got them yet.
>>
>> They are queued for 4.17 in:
>>
>> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/boot
> 
> 
> I may not be looking at the right place but I don't see it there:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/tree/arch/x86/xen/enlighten_pvh.c?h=x86/boot

Aah, seems as if Ingo has merged the patches into x86/mm and x86/boot
has been re-inited again. Look at:

https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/tree/arch/x86/xen/enlighten_pvh.c?h=x86/mm


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Hans van Kranenburg]

On 01/24/2018 01:06 PM, Juergen Gross wrote:
> On 24/01/18 11:54, Roger Pau Monné wrote:
>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>> <boris.ostrovsky@oracle.com> wrote:
>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>> Wei,
>>>>>>
>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>
>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>> Mind applying this one?
>>>>> Don't worry, it is in my queue.
>>>>>
>>>>> Will come to this and other patches I accumulated soon.
>>>>>
>>>>> Wei.
>>>>
>>>> This requires kernel changes, doesn't it?
>>>>
>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>
>>>> And this series apparently never made it to the tree.
>>>>
>>>> PVH guests are broken now on staging.
>>>
>>> And the Linux side of PVH is officially supported now, right?  Which
>>> means we need to have some way of being backwards compatible with
>>> kernels without that patch, even if it does get  backported.
>>
>> Linux had a bug that on PVH it expected to find the RSDP at the low
>> 1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
>> should be below 1MiB. There's just one comment in the start_info
>> header that says:
>>
>> "Xen on x86 will always try to place all the data below the 4GiB
>> boundary."
>>
>> And the RSDP address must be fetched from the rsdp_paddr field of the
>> start_info.
>>
>> I agree that the fact that Linux is broken is bad, but should not
>> force us to change the protocol. What's more, when booting a PVHv2
>> Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
>> going to find the firmware provided RSDP and everything is going to
>> blow up.
> 
> Right.
> 
> My patches to repair this in the Linux kernel are on hold as they are
> touching the boot parameters between grub2 and the kernel, too. And
> this caused the kernel no longer booting in some cases as most distros
> are shipping a grub2 breaking this interface via a non-upstream patch.

Linking the related thread:
https://www.spinics.net/lists/linux-acpi/msg80003.html

Out of interest, which "non-upstream patch" is this by the way? I didn't
see it mentioned yet.

Is it defining an extra custom field in that stuct yolo style? Because
from my limited understanding that would of course also need the
counterpart being implemented at kernel side.

Or (and I suspect it's this) is is not really using those bytes, but
leaving garbage behind, which forces you to implement the extra
validation checks?

Thanks,

>> IMO, Linux must be fixed and the patch backported up to 4.11.
> 
> I'll try soon.

Hans


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [PATCH v2] libxl: put RSDP for PVH guest near 4GB

[Juergen Gross]

On 25/01/18 11:38, Hans van Kranenburg wrote:
> On 01/24/2018 01:06 PM, Juergen Gross wrote:
>> On 24/01/18 11:54, Roger Pau Monné wrote:
>>> On Wed, Jan 24, 2018 at 10:42:39AM +0000, George Dunlap wrote:
>>>> On Wed, Jan 24, 2018 at 2:41 AM, Boris Ostrovsky
>>>> <boris.ostrovsky@oracle.com> wrote:
>>>>> On 01/18/2018 05:33 AM, Wei Liu wrote:
>>>>>> On Thu, Jan 18, 2018 at 11:31:32AM +0100, Juergen Gross wrote:
>>>>>>> Wei,
>>>>>>>
>>>>>>> On 01/12/17 15:14, Juergen Gross wrote:
>>>>>>>> Instead of locating the RSDP table below 1MB put it just below 4GB
>>>>>>>> like the rest of the ACPI tables in case of PVH guests. This will
>>>>>>>> avoid punching more holes than necessary into the memory map.
>>>>>>>>
>>>>>>>> Signed-off-by: Juergen Gross <jgross@suse.com>
>>>>>>>> Acked-by: Wei Liu <wei.liu2@citrix.com>
>>>>>>> Mind applying this one?
>>>>>> Don't worry, it is in my queue.
>>>>>>
>>>>>> Will come to this and other patches I accumulated soon.
>>>>>>
>>>>>> Wei.
>>>>>
>>>>> This requires kernel changes, doesn't it?
>>>>>
>>>>> https://lists.xenproject.org/archives/html/xen-devel/2017-12/msg00714.html
>>>>>
>>>>> And this series apparently never made it to the tree.
>>>>>
>>>>> PVH guests are broken now on staging.
>>>>
>>>> And the Linux side of PVH is officially supported now, right?  Which
>>>> means we need to have some way of being backwards compatible with
>>>> kernels without that patch, even if it does get  backported.
>>>
>>> Linux had a bug that on PVH it expected to find the RSDP at the low
>>> 1MiB. The PVHv2 boot protocol has _never_ mentioned that the RSDP
>>> should be below 1MiB. There's just one comment in the start_info
>>> header that says:
>>>
>>> "Xen on x86 will always try to place all the data below the 4GiB
>>> boundary."
>>>
>>> And the RSDP address must be fetched from the rsdp_paddr field of the
>>> start_info.
>>>
>>> I agree that the fact that Linux is broken is bad, but should not
>>> force us to change the protocol. What's more, when booting a PVHv2
>>> Dom0 the RSDP _must_ not be searched in the low 1MiB, or else Dom0 is
>>> going to find the firmware provided RSDP and everything is going to
>>> blow up.
>>
>> Right.
>>
>> My patches to repair this in the Linux kernel are on hold as they are
>> touching the boot parameters between grub2 and the kernel, too. And
>> this caused the kernel no longer booting in some cases as most distros
>> are shipping a grub2 breaking this interface via a non-upstream patch.
> 
> Linking the related thread:
> https://www.spinics.net/lists/linux-acpi/msg80003.html
> 
> Out of interest, which "non-upstream patch" is this by the way? I didn't
> see it mentioned yet.

That patch is adding another Linux loader for EFI environment.

> Is it defining an extra custom field in that stuct yolo style? Because
> from my limited understanding that would of course also need the
> counterpart being implemented at kernel side.
> 
> Or (and I suspect it's this) is is not really using those bytes, but
> leaving garbage behind, which forces you to implement the extra
> validation checks?

Right. And up to now this garbage hasn't been breaking anything (or
nobody noticed it).


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel