powerpc/dt_cpu_ftrs: Set CPU_FTR_P11_PVR for Power11 and later processors
Subsystems (2)
PATCH 1/1 powerpc/dt_cpu_ftrs: Set CPU_FTR_P11_PVR for Power11 and later processors
When using device tree CPU features (dt-cpu-ftrs), the kernel bypasses the traditional cputable-based CPU identification and instead derives CPU features from the device tree's "ibm,powerpc-cpu-features" node provided by firmware.
However, CPU_FTR_P11_PVR is a kernel-internal feature flag used to identify Power11 and later processors, and is not represented in the device tree's ISA feature set. While ISA v3.1 support (indicated by CPU_FTR_ARCH_31) is present on both Power10 and Power11, the CPU_FTR_P11_PVR flag is specifically needed by code that must distinguish between Power10 and Power11 processors.
Without this flag set, code that checks for Power11 using cpu_has_feature(CPU_FTR_P11_PVR) will incorrectly return false on Power11+ systems using dt-cpu-ftrs, leading to incorrect behavior.
This issue manifests specifically in powernv environments (bare-metal or QEMU TCG with powernv machine type), where skiboot/OPAL firmware provides the "ibm,powerpc-cpu-features" node, causing the kernel to use dt-cpu-ftrs. The issue does not affect pseries guests, where SLOF firmware does not provide this node, causing the kernel to fall back to the traditional cputable path (identify_cpu) which correctly sets CPU_FTR_P11_PVR during PVR-based CPU identification.
In powernv TCG guests, the missing flag causes KVM code to trigger warnings when attempting to create KVM guests, as cpu_features shows 0x000c00eb8f4fb187 (missing bit 53) instead of the correct 0x002c00eb8f4fb187 (with bit 53 set).
Fix this by setting CPU_FTR_P11_PVR for all processors with PVR >= PVR_POWER11 when ISA v3.1 support is detected in cpufeatures_setup_start(). This approach ensures forward compatibility with future processor generations.
---
Related: https://lore.kernel.org/all/[email protected]/
---
arch/powerpc/kernel/dt_cpu_ftrs.c | 9 +++++++++ 1 file changed, 9 insertions(+)
arch/powerpc/kernel/dt_cpu_ftrs.c +9 -0
@@ -704,6 +704,15 @@ static void __init cpufeatures_setup_start(u32 isa)if (isa >=ISA_V3_1 ) { cur_cpu_spec->cpu_features |=CPU_FTR_ARCH_31 ; cur_cpu_spec->cpu_user_features2 |=PPC_FEATURE2_ARCH_3_1 ; + +/* +* CPU_FTR_P11_PVR is a kernel-internal flag to identify +* Power11 and later processors. While ISA v3.1 is supported +* by Power10+, this flag specifically indicates Power11+ +* for code that needs to distinguish between P10 and P11. +*/ +if (PVR_VER (mfspr (SPRN_PVR )) >=PVR_POWER11 ) + cur_cpu_spec->cpu_features |=CPU_FTR_P11_PVR ; } }
base-commit: 424280953322cf66314f3ba5e2d1ef345f21c770
Signature · 1 line
2.50.1 (Apple Git-155)
Quoted 65 lines · Amit Machhiwal
When using device tree CPU features (dt-cpu-ftrs), the kernel bypasses the traditional cputable-based CPU identification and instead derives CPU features from the device tree's "ibm,powerpc-cpu-features" node provided by firmware.
However,
CPU_FTR_P11_PVRis a kernel-internal feature flag used to identify Power11 and later processors, and is not represented in the device tree's ISA feature set. While ISA v3.1 support (indicated byCPU_FTR_ARCH_31) is present on both Power10 and Power11, theCPU_FTR_P11_PVRflag is specifically needed by code that must distinguish between Power10 and Power11 processors.Without this flag set, code that checks for Power11 using
cpu_has_feature(CPU_FTR_P11_PVR)will incorrectly return false on Power11+ systems using dt-cpu-ftrs, leading to incorrect behavior.This issue manifests specifically in powernv environments (bare-metal or QEMU TCG with powernv machine type), where skiboot/OPAL firmware provides the "ibm,powerpc-cpu-features" node, causing the kernel to use dt-cpu-ftrs. The issue does not affect pseries guests, where SLOF firmware does not provide this node, causing the kernel to fall back to the traditional cputable path (
identify_cpu) which correctly setsCPU_FTR_P11_PVRduring PVR-based CPU identification.In powernv TCG guests, the missing flag causes KVM code to trigger warnings when attempting to create KVM guests, as
cpu_featuresshows0x000c00eb8f4fb187(missing bit 53) instead of the correct0x002c00eb8f4fb187(with bit 53 set).Fix this by setting
CPU_FTR_P11_PVRfor all processors with PVR >=PVR_POWER11when ISA v3.1 support is detected incpufeatures_setup_start(). This approach ensures forward compatibility with future processor generations.Fixes 96e266e3bcd6 ("KVM: PPC: Book3S HV: Add Power11 capability support for Nested PAPR guests")Cc [email protected] # v6.13+Signed-off-by Amit Machhiwal <[email protected]>---
Related: https://lore.kernel.org/all/[email protected]/
---arch/powerpc/kernel/dt_cpu_ftrs.c | 9 +++++++++ 1 file changed, 9 insertions(+)arch/powerpc/kernel/dt_cpu_ftrs.c +9
-0@@ -704,6 +704,15 @@ static void __init cpufeatures_setup_start(u32 isa)if (isa >=ISA_V3_1 ) { cur_cpu_spec->cpu_features |=CPU_FTR_ARCH_31 ; cur_cpu_spec->cpu_user_features2 |=PPC_FEATURE2_ARCH_3_1 ; + +/* +* CPU_FTR_P11_PVR is a kernel-internal flag to identify +* Power11 and later processors. While ISA v3.1 is supported +* by Power10+, this flag specifically indicates Power11+ +* for code that needs to distinguish between P10 and P11. +*/ +if (PVR_VER (mfspr (SPRN_PVR )) >=PVR_POWER11 ) + cur_cpu_spec->cpu_features |=CPU_FTR_P11_PVR ; } }
Quoted 4 lines
base-commit:
424280953322cf66314f3ba5e2d1ef345f21c770Signature · 1 line
2.50.1 (Apple Git-155)
LGTM
Christophe Leroy (CS GROUP)
reviewerLe 14/06/2026 à 19:34, Amit Machhiwal a écrit :
Quoted 50 lines
When using device tree CPU features (dt-cpu-ftrs), the kernel bypasses the traditional cputable-based CPU identification and instead derives CPU features from the device tree's "ibm,powerpc-cpu-features" node provided by firmware.
However,
CPU_FTR_P11_PVRis a kernel-internal feature flag used to identify Power11 and later processors, and is not represented in the device tree's ISA feature set. While ISA v3.1 support (indicated byCPU_FTR_ARCH_31) is present on both Power10 and Power11, theCPU_FTR_P11_PVRflag is specifically needed by code that must distinguish between Power10 and Power11 processors.Without this flag set, code that checks for Power11 using
cpu_has_feature(CPU_FTR_P11_PVR)will incorrectly return false on Power11+ systems using dt-cpu-ftrs, leading to incorrect behavior.This issue manifests specifically in powernv environments (bare-metal or QEMU TCG with powernv machine type), where skiboot/OPAL firmware provides the "ibm,powerpc-cpu-features" node, causing the kernel to use dt-cpu-ftrs. The issue does not affect pseries guests, where SLOF firmware does not provide this node, causing the kernel to fall back to the traditional cputable path (
identify_cpu) which correctly setsCPU_FTR_P11_PVRduring PVR-based CPU identification.In powernv TCG guests, the missing flag causes KVM code to trigger warnings when attempting to create KVM guests, as
cpu_featuresshows0x000c00eb8f4fb187(missing bit 53) instead of the correct0x002c00eb8f4fb187(with bit 53 set).Fix this by setting
CPU_FTR_P11_PVRfor all processors with PVR >=PVR_POWER11when ISA v3.1 support is detected incpufeatures_setup_start(). This approach ensures forward compatibility with future processor generations.Fixes 96e266e3bcd6 ("KVM: PPC: Book3S HV: Add Power11 capability support for Nested PAPR guests")Cc [email protected] # v6.13+Signed-off-by Amit Machhiwal <[email protected]>---
Related: https://lore.kernel.org/all/[email protected]/
---arch/powerpc/kernel/dt_cpu_ftrs.c | 9 +++++++++ 1 file changed, 9 insertions(+)arch/powerpc/kernel/dt_cpu_ftrs.c +0
-0@@ -704,6 +704,15 @@ static void __init cpufeatures_setup_start(u32 isa)if (isa >=ISA_V3_1 ) {
arch/powerpc/kernel/dt_cpu_ftrs.c static void __init cpufeatures_setup_start(u32 isa) · line 714
Quoted 10 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +8
-0cur_cpu_spec->cpu_features |=CPU_FTR_ARCH_31 ; cur_cpu_spec->cpu_user_features2 |=PPC_FEATURE2_ARCH_3_1 ; + +/* +* CPU_FTR_P11_PVR is a kernel-internal flag to identify +* Power11 and later processors. While ISA v3.1 is supported +* by Power10+, this flag specifically indicates Power11+ +* for code that needs to distinguish between P10 and P11. +*/ +if (PVR_VER (mfspr (SPRN_PVR )) >=PVR_POWER11 )
Are we sure this test will always be correct ?
For instance PVR_PA6T is higher than PVR_POWER11 allthough it is not ISA 3.1
Wouldn't is be cleaner and safer to just do:
PVR_VER(mfspr(SPRN_PVR)) == PVR_POWER11
Quoted 6 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +1
-0+ cur_cpu_spec->cpu_features |=CPU_FTR_P11_PVR ; } }base-commit:
424280953322cf66314f3ba5e2d1ef345f21c770
Hi Christophe,
Thanks for reviewing this patch. Please find my response inline.
Quoted 66 lines · Christophe Leroy (CS GROUP)
Le 14/06/2026 à 19:34, Amit Machhiwal a écrit :
Quoted 50 lines
When using device tree CPU features (dt-cpu-ftrs), the kernel bypasses the traditional cputable-based CPU identification and instead derives CPU features from the device tree's "ibm,powerpc-cpu-features" node provided by firmware.
However,
CPU_FTR_P11_PVRis a kernel-internal feature flag used to identify Power11 and later processors, and is not represented in the device tree's ISA feature set. While ISA v3.1 support (indicated byCPU_FTR_ARCH_31) is present on both Power10 and Power11, theCPU_FTR_P11_PVRflag is specifically needed by code that must distinguish between Power10 and Power11 processors.Without this flag set, code that checks for Power11 using
cpu_has_feature(CPU_FTR_P11_PVR)will incorrectly return false on Power11+ systems using dt-cpu-ftrs, leading to incorrect behavior.This issue manifests specifically in powernv environments (bare-metal or QEMU TCG with powernv machine type), where skiboot/OPAL firmware provides the "ibm,powerpc-cpu-features" node, causing the kernel to use dt-cpu-ftrs. The issue does not affect pseries guests, where SLOF firmware does not provide this node, causing the kernel to fall back to the traditional cputable path (
identify_cpu) which correctly setsCPU_FTR_P11_PVRduring PVR-based CPU identification.In powernv TCG guests, the missing flag causes KVM code to trigger warnings when attempting to create KVM guests, as
cpu_featuresshows0x000c00eb8f4fb187(missing bit 53) instead of the correct0x002c00eb8f4fb187(with bit 53 set).Fix this by setting
CPU_FTR_P11_PVRfor all processors with PVR >=PVR_POWER11when ISA v3.1 support is detected incpufeatures_setup_start(). This approach ensures forward compatibility with future processor generations.Fixes 96e266e3bcd6 ("KVM: PPC: Book3S HV: Add Power11 capability support for Nested PAPR guests")Cc [email protected] # v6.13+Signed-off-by Amit Machhiwal <[email protected]>---
Related: https://lore.kernel.org/all/[email protected]/
---arch/powerpc/kernel/dt_cpu_ftrs.c | 9 +++++++++ 1 file changed, 9 insertions(+)arch/powerpc/kernel/dt_cpu_ftrs.c +0
-0@@ -704,6 +704,15 @@ static void __init cpufeatures_setup_start(u32 isa)if (isa >=ISA_V3_1 ) {arch/powerpc/kernel/dt_cpu_ftrs.c static void __init cpufeatures_setup_start(u32 isa) · line 714
Quoted 10 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +8
-0cur_cpu_spec->cpu_features |=CPU_FTR_ARCH_31 ; cur_cpu_spec->cpu_user_features2 |=PPC_FEATURE2_ARCH_3_1 ; + +/* +* CPU_FTR_P11_PVR is a kernel-internal flag to identify +* Power11 and later processors. While ISA v3.1 is supported +* by Power10+, this flag specifically indicates Power11+ +* for code that needs to distinguish between P10 and P11. +*/ +if (PVR_VER (mfspr (SPRN_PVR )) >=PVR_POWER11 )Are we sure this test will always be correct ?
Quoted 5 lines
For instance
PVR_PA6Tis higher thanPVR_POWER11allthough it is not ISA 3.1Wouldn't is be cleaner and safer to just do:
PVR_VER(mfspr(SPRN_PVR)) == PVR_POWER11
You're absolutely right to point out the PVR ordering concern. But PA6T cannot actually reach this path because we're already gated by:
if (isa >= ISA_V3_1)
and PA6T does not implement ISA v3.1.
My rationale for using >= PVR_POWER11 is that CPU_FTR_P11_PVR is intended to be included for Power11 and later processors, not just Power11 itself, as it identifies a CPU feature. Using == PVR_POWER11 would mean we'd need to revisit this code for every future generation.
This approach is consistent with existing kernel code. For example, in arch/powerpc/perf/hv-gpci.c:
/* sysinfo interface files are only available for power10 and above platforms */ if (PVR_VER(mfspr(SPRN_PVR)) >= PVR_POWER10)
add_sysinfo_interface_files();
Also, I couldn't find any current users of PVR_PA6T or PVR_BE in the kernel tree, so there doesn't appear to be a present-day ISA v3.1+ example where the comparison would misidentify a processor.
Please let me know your further thoughts on this.
Thanks,
Amit
Quoted 7 lines
Quoted 6 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +1
-0+ cur_cpu_spec->cpu_features |=CPU_FTR_P11_PVR ; } }base-commit:
424280953322cf66314f3ba5e2d1ef345f21c770
Hi Christophe,
<snip>
Quoted 42 lines
Quoted 19 lines
Quoted 6 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +0
-0@@ -704,6 +704,15 @@ static void __init cpufeatures_setup_start(u32 isa)if (isa >=ISA_V3_1 ) {arch/powerpc/kernel/dt_cpu_ftrs.c static void __init cpufeatures_setup_start(u32 isa) · line 714
Quoted 10 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +8
-0cur_cpu_spec->cpu_features |=CPU_FTR_ARCH_31 ; cur_cpu_spec->cpu_user_features2 |=PPC_FEATURE2_ARCH_3_1 ; + +/* +* CPU_FTR_P11_PVR is a kernel-internal flag to identify +* Power11 and later processors. While ISA v3.1 is supported +* by Power10+, this flag specifically indicates Power11+ +* for code that needs to distinguish between P10 and P11. +*/ +if (PVR_VER (mfspr (SPRN_PVR )) >=PVR_POWER11 )Are we sure this test will always be correct ?
Quoted 5 lines
For instance
PVR_PA6Tis higher thanPVR_POWER11allthough it is not ISA 3.1Wouldn't is be cleaner and safer to just do:
PVR_VER(mfspr(SPRN_PVR)) == PVR_POWER11You're absolutely right to point out the PVR ordering concern. But PA6T cannot actually reach this path because we're already gated by:
if (isa >=
ISA_V3_1)and PA6T does not implement ISA v3.1.
My rationale for using
>= PVR_POWER11is thatCPU_FTR_P11_PVRis intended to be included for Power11 and later processors, not just Power11 itself, as it identifies a CPU feature. Using== PVR_POWER11would mean we'd need to revisit this code for every future generation.This approach is consistent with existing kernel code. For example, in
arch/powerpc/perf/hv-gpci.c:/* sysinfo interface files are only available for power10 and above platforms */ if (PVR_VER(mfspr(SPRN_PVR)) >=
PVR_POWER10)
Quoted 7 lines
add_sysinfo_interface_files();Also, I couldn't find any current users of
PVR_PA6TorPVR_BEin the kernel tree, so there doesn't appear to be a present-day ISA v3.1+ example where the comparison would misidentify a processor.Please let me know your further thoughts on this.
Just checking in — did my previous response address your concern, or do you have further comments?
Thanks,
Amit
Le 25/06/2026 à 15:14, Amit Machhiwal a écrit :
Quoted 48 lines
Hi Christophe,
<snip>
Quoted 44 lines
Quoted 19 lines
Quoted 6 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +0
-0@@ -704,6 +704,15 @@ static void __init cpufeatures_setup_start(u32 isa)if (isa >=ISA_V3_1 ) {arch/powerpc/kernel/dt_cpu_ftrs.c static void __init cpufeatures_setup_start(u32 isa) · line 714
Quoted 10 lines
arch/powerpc/kernel/dt_cpu_ftrs.c +8
-0cur_cpu_spec->cpu_features |=CPU_FTR_ARCH_31 ; cur_cpu_spec->cpu_user_features2 |=PPC_FEATURE2_ARCH_3_1 ; + +/* +* CPU_FTR_P11_PVR is a kernel-internal flag to identify +* Power11 and later processors. While ISA v3.1 is supported +* by Power10+, this flag specifically indicates Power11+ +* for code that needs to distinguish between P10 and P11. +*/ +if (PVR_VER (mfspr (SPRN_PVR )) >=PVR_POWER11 )Are we sure this test will always be correct ?
Quoted 5 lines
For instance
PVR_PA6Tis higher thanPVR_POWER11allthough it is not ISA 3.1Wouldn't is be cleaner and safer to just do:
PVR_VER(mfspr(SPRN_PVR)) == PVR_POWER11You're absolutely right to point out the PVR ordering concern. But PA6T cannot actually reach this path because we're already gated by:
if (isa >= ISA_V3_1)and PA6T does not implement ISA v3.1.
My rationale for using
>= PVR_POWER11is thatCPU_FTR_P11_PVRis intended to be included for Power11 and later processors, not just Power11 itself, as it identifies a CPU feature. Using== PVR_POWER11would mean we'd need to revisit this code for every future generation.This approach is consistent with existing kernel code. For example, in
arch/powerpc/perf/hv-gpci.c:/* sysinfo interface files are only available for power10 and above platforms */ if (PVR_VER(mfspr(SPRN_PVR)) >= PVR_POWER10) add_sysinfo_interface_files();
Quoted 8 lines
Quoted 5 lines
Also, I couldn't find any current users of
PVR_PA6TorPVR_BEin the kernel tree, so there doesn't appear to be a present-day ISA v3.1+ example where the comparison would misidentify a processor.Please let me know your further thoughts on this.
Just checking in — did my previous response address your concern, or do you have further comments?
That ok, no more comments.
Quoted 5 lines · Amit Machhiwal
When using device tree CPU features (dt-cpu-ftrs), the kernel bypasses the traditional cputable-based CPU identification and instead derives CPU features from the device tree's "ibm,powerpc-cpu-features" node provided by firmware.
Quoted 8 lines
However,
CPU_FTR_P11_PVRis a kernel-internal feature flag used to identify Power11 and later processors, and is not represented in the device tree's ISA feature set. While ISA v3.1 support (indicated byCPU_FTR_ARCH_31) is present on both Power10 and Power11, theCPU_FTR_P11_PVRflag is specifically needed by code that must distinguish between Power10 and Power11 processors.[...]
Applied to powerpc/fixes.
[1/1] powerpc/dt_cpu_ftrs: Set CPU_FTR_P11_PVR for Power11 and later processors
https://git.kernel.org/powerpc/c/e4de1b9cb3b5c981e4fe9bca253a7fb9161f5acd
cheers