[PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Gerd Hoffmann]

qemu uses the etc/e820 fw_cfg file not only for memory, but
also for reservations.  Handle reservations by adding resource
descriptor hobs for them.

A typical qemu configuration has a small reservation between
lapic and flash:

  # sudo cat /proc/iomem
  [ ... ]
  fee00000-fee00fff : Local APIC
  feffc000-feffffff : Reserved          <= HERE
  ffc00000-ffffffff : Reserved
  [ ... ]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 OvmfPkg/Library/PlatformInitLib/MemDetect.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/OvmfPkg/Library/PlatformInitLib/MemDetect.c b/OvmfPkg/Library/PlatformInitLib/MemDetect.c
index b8feae4309de..a7a4e028ad50 100644
--- a/OvmfPkg/Library/PlatformInitLib/MemDetect.c
+++ b/OvmfPkg/Library/PlatformInitLib/MemDetect.c
@@ -228,6 +228,22 @@ PlatformScanOrAdd64BitE820Ram (
             ));
         }
       }
+    } else if (E820Entry.Type == EfiAcpiAddressRangeReserved) {
+      if (AddHighHob) {
+        DEBUG ((
+          DEBUG_INFO,
+          "%a: Reserved: Base=0x%Lx Length=0x%Lx\n",
+          __FUNCTION__,
+          E820Entry.BaseAddr,
+          E820Entry.Length
+          ));
+        BuildResourceDescriptorHob (
+          EFI_RESOURCE_MEMORY_RESERVED,
+          0,
+          E820Entry.BaseAddr,
+          E820Entry.Length
+          );
+      }
     }
   }
 
-- 
2.38.1

Re: [edk2-devel] [PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Ard Biesheuvel]

On Wed, 7 Dec 2022 at 06:32, Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> qemu uses the etc/e820 fw_cfg file not only for memory, but
> also for reservations.  Handle reservations by adding resource
> descriptor hobs for them.
>
> A typical qemu configuration has a small reservation between
> lapic and flash:
>
>   # sudo cat /proc/iomem
>   [ ... ]
>   fee00000-fee00fff : Local APIC
>   feffc000-feffffff : Reserved          <= HERE
>   ffc00000-ffffffff : Reserved
>   [ ... ]
>
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Merged as #3828


> ---
>  OvmfPkg/Library/PlatformInitLib/MemDetect.c | 16 ++++++++++++++++
>  1 file changed, 16 insertions(+)
>
> diff --git a/OvmfPkg/Library/PlatformInitLib/MemDetect.c b/OvmfPkg/Library/PlatformInitLib/MemDetect.c
> index b8feae4309de..a7a4e028ad50 100644
> --- a/OvmfPkg/Library/PlatformInitLib/MemDetect.c
> +++ b/OvmfPkg/Library/PlatformInitLib/MemDetect.c
> @@ -228,6 +228,22 @@ PlatformScanOrAdd64BitE820Ram (
>              ));
>          }
>        }
> +    } else if (E820Entry.Type == EfiAcpiAddressRangeReserved) {
> +      if (AddHighHob) {
> +        DEBUG ((
> +          DEBUG_INFO,
> +          "%a: Reserved: Base=0x%Lx Length=0x%Lx\n",
> +          __FUNCTION__,
> +          E820Entry.BaseAddr,
> +          E820Entry.Length
> +          ));
> +        BuildResourceDescriptorHob (
> +          EFI_RESOURCE_MEMORY_RESERVED,
> +          0,
> +          E820Entry.BaseAddr,
> +          E820Entry.Length
> +          );
> +      }
>      }
>    }
>
> --
> 2.38.1
>
>
>
> 
>
>

Re: [edk2-devel] [PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Lendacky, Thomas]

On 12/6/22 23:32, Gerd Hoffmann via groups.io wrote:
> qemu uses the etc/e820 fw_cfg file not only for memory, but
> also for reservations.  Handle reservations by adding resource
> descriptor hobs for them.
> 
> A typical qemu configuration has a small reservation between
> lapic and flash:
> 
>    # sudo cat /proc/iomem
>    [ ... ]
>    fee00000-fee00fff : Local APIC
>    feffc000-feffffff : Reserved          <= HERE
>    ffc00000-ffffffff : Reserved
>    [ ... ]
> 
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

On newer versions of Qemu, this patch causes a boot failure on my
AMD system with the following output:

...
PciHostBridgeUtilityInitRootBridge: populated root bus 0, with room for 255 subordinate bus(es)
RootBridge: PciRoot(0x0)
   Support/Attr: 70069 / 70069
     DmaAbove4G: No
NoExtConfSpace: No
      AllocAttr: 3 (CombineMemPMem Mem64Decode)
            Bus: 0 - FF Translation=0
             Io: 6000 - FFFF Translation=0
            Mem: C0000000 - FBFFFFFF Translation=0
     MemAbove4G: E000000000 - FFFFFFFFFF Translation=0
           PMem: FFFFFFFFFFFFFFFF - 0 Translation=0
    PMemAbove4G: FFFFFFFFFFFFFFFF - 0 Translation=0
PciHostBridgeDxe: IntersectMemoryDescriptor: desc [FD00000000, 10000000000) type 1 cap 8000000000026000 conflicts with aperture [E000000000, 10000000000) cap 1

ASSERT_EFI_ERROR (Status = Invalid Parameter)
ASSERT [PciHostBridgeDxe] /root/kernels/ovmf-build-X64/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridge.c(550): !(((INTN)(RETURN_STATUS)(Status)) < 0)

Doing some bisecting, before Qemu patch

    8504f129450b ("i386/pc: relocate 4g start to 1T where applicable")

this issue wasn't observed. And before this OVMF patch, the newer version
of Qemu doesn't have an issue.

Also, if I add host-phys-bits=true to the Qemu -cpu parameter, then
the issue isn't observed regardless of Qemu/OVMF level.

I'm hoping someone has an understanding of what is going on and what
can be done to fix it.

Thanks,
Tom


> ---
>   OvmfPkg/Library/PlatformInitLib/MemDetect.c | 16 ++++++++++++++++
>   1 file changed, 16 insertions(+)
> 
> diff --git a/OvmfPkg/Library/PlatformInitLib/MemDetect.c b/OvmfPkg/Library/PlatformInitLib/MemDetect.c
> index b8feae4309de..a7a4e028ad50 100644
> --- a/OvmfPkg/Library/PlatformInitLib/MemDetect.c
> +++ b/OvmfPkg/Library/PlatformInitLib/MemDetect.c
> @@ -228,6 +228,22 @@ PlatformScanOrAdd64BitE820Ram (
>               ));
>           }
>         }
> +    } else if (E820Entry.Type == EfiAcpiAddressRangeReserved) {
> +      if (AddHighHob) {
> +        DEBUG ((
> +          DEBUG_INFO,
> +          "%a: Reserved: Base=0x%Lx Length=0x%Lx\n",
> +          __FUNCTION__,
> +          E820Entry.BaseAddr,
> +          E820Entry.Length
> +          ));
> +        BuildResourceDescriptorHob (
> +          EFI_RESOURCE_MEMORY_RESERVED,
> +          0,
> +          E820Entry.BaseAddr,
> +          E820Entry.Length
> +          );
> +      }
>       }
>     }
>   

Re: [edk2-devel] [PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Ard Biesheuvel]

On Thu, 12 Jan 2023 at 18:09, Tom Lendacky <thomas.lendacky@amd.com> wrote:
>
> On 12/6/22 23:32, Gerd Hoffmann via groups.io wrote:
> > qemu uses the etc/e820 fw_cfg file not only for memory, but
> > also for reservations.  Handle reservations by adding resource
> > descriptor hobs for them.
> >
> > A typical qemu configuration has a small reservation between
> > lapic and flash:
> >
> >    # sudo cat /proc/iomem
> >    [ ... ]
> >    fee00000-fee00fff : Local APIC
> >    feffc000-feffffff : Reserved          <= HERE
> >    ffc00000-ffffffff : Reserved
> >    [ ... ]
> >
> > Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
>
> On newer versions of Qemu, this patch causes a boot failure on my
> AMD system with the following output:
>
> ...
> PciHostBridgeUtilityInitRootBridge: populated root bus 0, with room for 255 subordinate bus(es)
> RootBridge: PciRoot(0x0)
>    Support/Attr: 70069 / 70069
>      DmaAbove4G: No
> NoExtConfSpace: No
>       AllocAttr: 3 (CombineMemPMem Mem64Decode)
>             Bus: 0 - FF Translation=0
>              Io: 6000 - FFFF Translation=0
>             Mem: C0000000 - FBFFFFFF Translation=0
>      MemAbove4G: E000000000 - FFFFFFFFFF Translation=0
>            PMem: FFFFFFFFFFFFFFFF - 0 Translation=0
>     PMemAbove4G: FFFFFFFFFFFFFFFF - 0 Translation=0
> PciHostBridgeDxe: IntersectMemoryDescriptor: desc [FD00000000, 10000000000) type 1 cap 8000000000026000 conflicts with aperture [E000000000, 10000000000) cap 1
>
> ASSERT_EFI_ERROR (Status = Invalid Parameter)
> ASSERT [PciHostBridgeDxe] /root/kernels/ovmf-build-X64/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridge.c(550): !(((INTN)(RETURN_STATUS)(Status)) < 0)
>
> Doing some bisecting, before Qemu patch
>
>     8504f129450b ("i386/pc: relocate 4g start to 1T where applicable")
>
> this issue wasn't observed. And before this OVMF patch, the newer version
> of Qemu doesn't have an issue.
>
> Also, if I add host-phys-bits=true to the Qemu -cpu parameter, then
> the issue isn't observed regardless of Qemu/OVMF level.
>
> I'm hoping someone has an understanding of what is going on and what
> can be done to fix it.
>

Could someone look into this please? Gerd?

Re: [edk2-devel] [PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Lendacky, Thomas]

On 1/16/23 11:09, Ard Biesheuvel wrote:
> On Thu, 12 Jan 2023 at 18:09, Tom Lendacky <thomas.lendacky@amd.com> wrote:
>>
>> On 12/6/22 23:32, Gerd Hoffmann via groups.io wrote:
>>> qemu uses the etc/e820 fw_cfg file not only for memory, but
>>> also for reservations.  Handle reservations by adding resource
>>> descriptor hobs for them.
>>>
>>> A typical qemu configuration has a small reservation between
>>> lapic and flash:
>>>
>>>     # sudo cat /proc/iomem
>>>     [ ... ]
>>>     fee00000-fee00fff : Local APIC
>>>     feffc000-feffffff : Reserved          <= HERE
>>>     ffc00000-ffffffff : Reserved
>>>     [ ... ]
>>>
>>> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
>>
>> On newer versions of Qemu, this patch causes a boot failure on my
>> AMD system with the following output:
>>
>> ...
>> PciHostBridgeUtilityInitRootBridge: populated root bus 0, with room for 255 subordinate bus(es)
>> RootBridge: PciRoot(0x0)
>>     Support/Attr: 70069 / 70069
>>       DmaAbove4G: No
>> NoExtConfSpace: No
>>        AllocAttr: 3 (CombineMemPMem Mem64Decode)
>>              Bus: 0 - FF Translation=0
>>               Io: 6000 - FFFF Translation=0
>>              Mem: C0000000 - FBFFFFFF Translation=0
>>       MemAbove4G: E000000000 - FFFFFFFFFF Translation=0
>>             PMem: FFFFFFFFFFFFFFFF - 0 Translation=0
>>      PMemAbove4G: FFFFFFFFFFFFFFFF - 0 Translation=0
>> PciHostBridgeDxe: IntersectMemoryDescriptor: desc [FD00000000, 10000000000) type 1 cap 8000000000026000 conflicts with aperture [E000000000, 10000000000) cap 1
>>
>> ASSERT_EFI_ERROR (Status = Invalid Parameter)
>> ASSERT [PciHostBridgeDxe] /root/kernels/ovmf-build-X64/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridge.c(550): !(((INTN)(RETURN_STATUS)(Status)) < 0)
>>
>> Doing some bisecting, before Qemu patch
>>
>>      8504f129450b ("i386/pc: relocate 4g start to 1T where applicable")
>>
>> this issue wasn't observed. And before this OVMF patch, the newer version
>> of Qemu doesn't have an issue.
>>
>> Also, if I add host-phys-bits=true to the Qemu -cpu parameter, then
>> the issue isn't observed regardless of Qemu/OVMF level.
>>
>> I'm hoping someone has an understanding of what is going on and what
>> can be done to fix it.
>>
> 
> Could someone look into this please? Gerd?

Looks like Gerd is already on it.

See https://bugzilla.tianocore.org/show_bug.cgi?id=4251

Thanks,
Tom

Re: [edk2-devel] [PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Ard Biesheuvel]

On Mon, 16 Jan 2023 at 21:10, Tom Lendacky <thomas.lendacky@amd.com> wrote:
>
> On 1/16/23 11:09, Ard Biesheuvel wrote:
> > On Thu, 12 Jan 2023 at 18:09, Tom Lendacky <thomas.lendacky@amd.com> wrote:
> >>
> >> On 12/6/22 23:32, Gerd Hoffmann via groups.io wrote:
> >>> qemu uses the etc/e820 fw_cfg file not only for memory, but
> >>> also for reservations.  Handle reservations by adding resource
> >>> descriptor hobs for them.
> >>>
> >>> A typical qemu configuration has a small reservation between
> >>> lapic and flash:
> >>>
> >>>     # sudo cat /proc/iomem
> >>>     [ ... ]
> >>>     fee00000-fee00fff : Local APIC
> >>>     feffc000-feffffff : Reserved          <= HERE
> >>>     ffc00000-ffffffff : Reserved
> >>>     [ ... ]
> >>>
> >>> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> >>
> >> On newer versions of Qemu, this patch causes a boot failure on my
> >> AMD system with the following output:
> >>
> >> ...
> >> PciHostBridgeUtilityInitRootBridge: populated root bus 0, with room for 255 subordinate bus(es)
> >> RootBridge: PciRoot(0x0)
> >>     Support/Attr: 70069 / 70069
> >>       DmaAbove4G: No
> >> NoExtConfSpace: No
> >>        AllocAttr: 3 (CombineMemPMem Mem64Decode)
> >>              Bus: 0 - FF Translation=0
> >>               Io: 6000 - FFFF Translation=0
> >>              Mem: C0000000 - FBFFFFFF Translation=0
> >>       MemAbove4G: E000000000 - FFFFFFFFFF Translation=0
> >>             PMem: FFFFFFFFFFFFFFFF - 0 Translation=0
> >>      PMemAbove4G: FFFFFFFFFFFFFFFF - 0 Translation=0
> >> PciHostBridgeDxe: IntersectMemoryDescriptor: desc [FD00000000, 10000000000) type 1 cap 8000000000026000 conflicts with aperture [E000000000, 10000000000) cap 1
> >>
> >> ASSERT_EFI_ERROR (Status = Invalid Parameter)
> >> ASSERT [PciHostBridgeDxe] /root/kernels/ovmf-build-X64/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridge.c(550): !(((INTN)(RETURN_STATUS)(Status)) < 0)
> >>
> >> Doing some bisecting, before Qemu patch
> >>
> >>      8504f129450b ("i386/pc: relocate 4g start to 1T where applicable")
> >>
> >> this issue wasn't observed. And before this OVMF patch, the newer version
> >> of Qemu doesn't have an issue.
> >>
> >> Also, if I add host-phys-bits=true to the Qemu -cpu parameter, then
> >> the issue isn't observed regardless of Qemu/OVMF level.
> >>
> >> I'm hoping someone has an understanding of what is going on and what
> >> can be done to fix it.
> >>
> >
> > Could someone look into this please? Gerd?
>
> Looks like Gerd is already on it.
>
> See https://bugzilla.tianocore.org/show_bug.cgi?id=4251
>

Right, I didn't realize these things were related.

In any case, those patches have been merged now - care to check
whether the issue has been resolved? Thanks.

Re: [edk2-devel] [PATCH 1/1] OvmfPkg/PlatformInitLib: pass through reservations from qemu

[Lendacky, Thomas]

On 1/17/23 10:49, Ard Biesheuvel wrote:
> On Mon, 16 Jan 2023 at 21:10, Tom Lendacky <thomas.lendacky@amd.com> wrote:
>>
>> On 1/16/23 11:09, Ard Biesheuvel wrote:
>>> On Thu, 12 Jan 2023 at 18:09, Tom Lendacky <thomas.lendacky@amd.com> wrote:
>>>>
>>>> On 12/6/22 23:32, Gerd Hoffmann via groups.io wrote:
>>>>> qemu uses the etc/e820 fw_cfg file not only for memory, but
>>>>> also for reservations.  Handle reservations by adding resource
>>>>> descriptor hobs for them.
>>>>>
>>>>> A typical qemu configuration has a small reservation between
>>>>> lapic and flash:
>>>>>
>>>>>      # sudo cat /proc/iomem
>>>>>      [ ... ]
>>>>>      fee00000-fee00fff : Local APIC
>>>>>      feffc000-feffffff : Reserved          <= HERE
>>>>>      ffc00000-ffffffff : Reserved
>>>>>      [ ... ]
>>>>>
>>>>> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
>>>>
>>>> On newer versions of Qemu, this patch causes a boot failure on my
>>>> AMD system with the following output:
>>>>
>>>> ...
>>>> PciHostBridgeUtilityInitRootBridge: populated root bus 0, with room for 255 subordinate bus(es)
>>>> RootBridge: PciRoot(0x0)
>>>>      Support/Attr: 70069 / 70069
>>>>        DmaAbove4G: No
>>>> NoExtConfSpace: No
>>>>         AllocAttr: 3 (CombineMemPMem Mem64Decode)
>>>>               Bus: 0 - FF Translation=0
>>>>                Io: 6000 - FFFF Translation=0
>>>>               Mem: C0000000 - FBFFFFFF Translation=0
>>>>        MemAbove4G: E000000000 - FFFFFFFFFF Translation=0
>>>>              PMem: FFFFFFFFFFFFFFFF - 0 Translation=0
>>>>       PMemAbove4G: FFFFFFFFFFFFFFFF - 0 Translation=0
>>>> PciHostBridgeDxe: IntersectMemoryDescriptor: desc [FD00000000, 10000000000) type 1 cap 8000000000026000 conflicts with aperture [E000000000, 10000000000) cap 1
>>>>
>>>> ASSERT_EFI_ERROR (Status = Invalid Parameter)
>>>> ASSERT [PciHostBridgeDxe] /root/kernels/ovmf-build-X64/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridge.c(550): !(((INTN)(RETURN_STATUS)(Status)) < 0)
>>>>
>>>> Doing some bisecting, before Qemu patch
>>>>
>>>>       8504f129450b ("i386/pc: relocate 4g start to 1T where applicable")
>>>>
>>>> this issue wasn't observed. And before this OVMF patch, the newer version
>>>> of Qemu doesn't have an issue.
>>>>
>>>> Also, if I add host-phys-bits=true to the Qemu -cpu parameter, then
>>>> the issue isn't observed regardless of Qemu/OVMF level.
>>>>
>>>> I'm hoping someone has an understanding of what is going on and what
>>>> can be done to fix it.
>>>>
>>>
>>> Could someone look into this please? Gerd?
>>
>> Looks like Gerd is already on it.
>>
>> See https://bugzilla.tianocore.org/show_bug.cgi?id=4251
>>
> 
> Right, I didn't realize these things were related.
> 
> In any case, those patches have been merged now - care to check
> whether the issue has been resolved? Thanks.

Yes, the issue is now resolved.

Thanks,
Tom