mbox series

[0/5] xen: fix dom0 PV boot on some AMD machines

Message ID 20240807104110.18117-1-jgross@suse.com (mailing list archive)
Headers show
Series xen: fix dom0 PV boot on some AMD machines | expand

Message

Jürgen Groß Aug. 7, 2024, 10:41 a.m. UTC
There have been reports of failed boots with Xen due to an overlap of
the kernel's memory with ACPI NVS reported in the E820 map of the host.

This series fixes this issue by moving the NVS area in dom0 to some
higher address.

Juergen Gross (5):
  xen: use correct end address of kernel for conflict checking
  xen: introduce generic helper checking for memory map conflicts
  xen: move checks for e820 conflicts further up
  xen: move max_pfn in xen_memory_setup() out of function scope
  xen: tolerate ACPI NVS memory overlapping with Xen allocated memory

 arch/x86/xen/mmu_pv.c  |   5 +-
 arch/x86/xen/setup.c   | 242 +++++++++++++++++++++++++++++++++--------
 arch/x86/xen/xen-ops.h |   3 +-
 3 files changed, 201 insertions(+), 49 deletions(-)

Comments

Jürgen Groß Aug. 15, 2024, 2:34 p.m. UTC | #1
On 07.08.24 12:41, Juergen Gross wrote:
> There have been reports of failed boots with Xen due to an overlap of
> the kernel's memory with ACPI NVS reported in the E820 map of the host.
> 
> This series fixes this issue by moving the NVS area in dom0 to some
> higher address.
> 
> Juergen Gross (5):
>    xen: use correct end address of kernel for conflict checking
>    xen: introduce generic helper checking for memory map conflicts
>    xen: move checks for e820 conflicts further up
>    xen: move max_pfn in xen_memory_setup() out of function scope
>    xen: tolerate ACPI NVS memory overlapping with Xen allocated memory
> 
>   arch/x86/xen/mmu_pv.c  |   5 +-
>   arch/x86/xen/setup.c   | 242 +++++++++++++++++++++++++++++++++--------
>   arch/x86/xen/xen-ops.h |   3 +-
>   3 files changed, 201 insertions(+), 49 deletions(-)
> 

While testing V2 of this series another thought came to my mind:

What about PV guests on such a system with e820_host=1? I guess those
will have a similar problem as a PV dom0, as the ACPI NVS entry in the
memory map will be at the same position as on the host.

Shouldn't we get rid of the ACPI related e820 entries in the guests?


Juergen