kvmppc: Paravirtualize KVM to support ultravisor
Subsystems (2)
v4 cover kvmppc: Paravirtualize KVM to support ultravisor
POWER platforms that supports the Protected Execution Facility (PEF) introduce features that combine hardware facilities and firmware to enable secure virtual machines. That includes a new processor mode (ultravisor mode) and the ultravisor firmware.
In PEF enabled systems, the ultravisor firmware runs at a privilege level above the hypervisor and also takes control over some system resources. The hypervisor, though, can make system calls to access these resources. Such system calls, a.k.a. ucalls, are handled by the ultravisor firmware.
The processor allows part of the system memory to be configured as secure memory, and introduces a a new mode, called secure mode, where any software entity in that mode can access secure memory. The hypervisor doesn't (and can't) run in secure mode, but a secure guest and the ultravisor firmware do.
This patch set adds support for ultravisor calls and do some preparation for running secure guests.
---
Changelog:
---
v3->v4:
- Patch "KVM: PPC: Ultravisor: Add PPC_UV config option":
- Moved to the patchset "kvmppc: HMM driver to manage pages of secure guest" v5 that will be posted by Bharata Rao.
- Patch "powerpc: Introduce FW_FEATURE_ULTRAVISOR":
- Changed to depend only on CONFIG_PPC_POWERNV.
- Patch "KVM: PPC: Ultravisor: Add generic ultravisor call handler":
- Fixed whitespaces in ucall.S and in ultravisor-api.h.
- Changed to depend only on CONFIG_PPC_POWERNV.
- Changed the ucall wrapper to pass the ucall number in R3.
- Patch "KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE:
- Changed to depend only on CONFIG_PPC_POWERNV.
- Patch "KVM: PPC: Ultravisor: Restrict LDBAR access":
- Fixed comment in opal-imc.c to be "Disable IMC devices, when Ultravisor is enabled.
- Fixed signed-off-by.
- Patch "KVM: PPC: Ultravisor: Enter a secure guest":
- Changed the UV_RETURN assembly call to save the actual R3 in R0 for the ultravisor and pass the UV_RETURN call number in R3.
- Patch "KVM: PPC: Ultravisor: Check for MSR_S during hv_reset_msr":
- Fixed commit message.
- Rebased to powerpc/next.
v2->v3:
- Squashed patches:
- "KVM: PPC: Ultravisor: Return to UV for hcalls from SVM"
- "KVM: PPC: Book3S HV: Fixed for running secure guests"
- Renamed patch from/to:
- "KVM: PPC: Ultravisor: Return to UV for hcalls from SVM"
- "KVM: PPC: Ultravisor: Enter a secure guest
- Rebased
- Addressed comments from Paul Mackerras
- Dropped ultravisor checks made in power8 code
- Updated the commit message for:
"KVM: PPC: Ultravisor: Enter a secure guest"
- Addressed comments from Maddy
- Dropped imc-pmu.c changes
- Changed opal-imc.c to fail the probe when the ultravisor is enabled
- Fixed "ucall defined but not used" issue when CONFIG_PPC_UV not set
v1->v2:
- Addressed comments from Paul Mackerras:
- Write the pate in HV's table before doing that in UV's
- Renamed and better documented the ultravisor header files. Also added all possible return codes for each ucall
- Updated the commit message that introduces the MSR_S bit
- Moved ultravisor.c and ucall.S to arch/powerpc/kernel
- Changed ucall.S to not save CR
- Rebased
- Changed the patches order
- Updated several commit messages
- Added FW_FEATURE_ULTRAVISOR to enable use of firmware_has_feature()
- Renamed CONFIG_PPC_KVM_UV to CONFIG_PPC_UV and used it to ifdef the ucall handler and the code that populates the powerpc_firmware_features for ultravisor
- Exported the ucall symbol. KVM may be built as module.
- Restricted LDBAR access if the ultravisor firmware is available
- Dropped patches:
- "[PATCH 06/13] KVM: PPC: Ultravisor: UV_RESTRICTED_SPR_WRITE ucall"
- "[PATCH 07/13] KVM: PPC: Ultravisor: UV_RESTRICTED_SPR_READ ucall"
- "[PATCH 08/13] KVM: PPC: Ultravisor: fix mtspr and mfspr"
- Squashed patches:
- "[PATCH 09/13] KVM: PPC: Ultravisor: Return to UV for hcalls from SVM"
- "[PATCH 13/13] KVM: PPC: UV: Have fast_guest_return check secure_guest"
Claudio Carvalho (2):
powerpc: Introduce FW_FEATURE_ULTRAVISOR
KVM: PPC: Ultravisor: Restrict LDBAR access
Michael Anderson (2):
KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
KVM: PPC: Ultravisor: Check for MSR_S during hv_reset_msr
Ram Pai (2):
KVM: PPC: Ultravisor: Add generic ultravisor call handler
KVM: PPC: Ultravisor: Restrict flush of the partition tlb cache
Sukadev Bhattiprolu (2):
KVM: PPC: Ultravisor: Introduce the MSR_S bit
KVM: PPC: Ultravisor: Enter a secure guest
arch/powerpc/include/asm/firmware.h | 5 +- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/reg.h | 3 ++ arch/powerpc/include/asm/ultravisor-api.h | 24 +++++++++ arch/powerpc/include/asm/ultravisor.h | 49 +++++++++++++++++ arch/powerpc/kernel/Makefile | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kernel/prom.c | 4 ++ arch/powerpc/kernel/ucall.S | 30 +++++++++++ arch/powerpc/kernel/ultravisor.c | 28 ++++++++++ arch/powerpc/kvm/book3s_64_mmu_hv.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 42 ++++++++++++--- arch/powerpc/mm/book3s64/hash_utils.c | 3 +- arch/powerpc/mm/book3s64/pgtable.c | 65 +++++++++++++++++------ arch/powerpc/mm/book3s64/radix_pgtable.c | 9 ++-- arch/powerpc/platforms/powernv/idle.c | 6 ++- arch/powerpc/platforms/powernv/opal-imc.c | 4 ++ 17 files changed, 246 insertions(+), 30 deletions(-) create mode 100644 arch/powerpc/include/asm/ultravisor-api.h create mode 100644 arch/powerpc/include/asm/ultravisor.h create mode 100644 arch/powerpc/kernel/ucall.S create mode 100644 arch/powerpc/kernel/ultravisor.c
Signature · 1 line
2.20.1
PATCH v4 1/8 KVM: PPC: Ultravisor: Introduce the MSR_S bit
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the MSR_S bit is used to indicate if the thread is in secure state. With the MSR_S bit, the privilege state of the thread is now determined by MSR_S, MSR_HV and MSR_PR, as follows:
S HV PR
-----------------------
0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reserved
The hypervisor doesn't (and can't) run with the MSR_S bit set, but a secure guest and the ultravisor firmware do.
--- arch/powerpc/include/asm/reg.h | 3 +++ 1 file changed, 3 insertions(+)
arch/powerpc/include/asm/reg.h +3 -0
@@ -38,6 +38,7 @@#define MSR_TM_LG 32/* Trans Mem Available */ #define MSR_VEC_LG 25/* Enable AltiVec */ #define MSR_VSX_LG 23/* Enable VSX */ +#define MSR_S_LG 22/* Secure VM bit */ #define MSR_POW_LG 18/* Enable Power Management */ #define MSR_WE_LG 18/* Wait State Enable */ #define MSR_TGPR_LG 17/* TLB Update registers in use */
@@ -71,11 +72,13 @@#define MSR_SF __MASK(MSR_SF_LG)/* Enable 64 bit mode */ #define MSR_ISF __MASK(MSR_ISF_LG)/* Interrupt 64b mode valid on 630 */ #define MSR_HV __MASK(MSR_HV_LG)/* Hypervisor state */ +#define MSR_S __MASK(MSR_S_LG)/* Secure state */ #else /* so tests for these bits fail on 32-bit */ #define MSR_SF 0#define MSR_ISF 0#define MSR_HV 0 +#define MSR_S 0#endif /*
Signature · 1 line
2.20.1
Re: [PATCH v4 1/8] KVM: PPC: Ultravisor: Introduce the MSR_S bit
Quoted 19 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the
MSR_Sbit is used to indicate if the thread is in secure state. With theMSR_Sbit, the privilege state of the thread is now determined byMSR_S,MSR_HVandMSR_PR, as follows:S HV PR
-----------------------0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reserved
Quoted 7 lines
The hypervisor doesn't (and can't) run with the
MSR_Sbit set, but a secure guest and the ultravisor firmware do.Signed-off-by Sukadev Bhattiprolu <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Update the commit message ]Signed-off-by Claudio Carvalho <[email protected]>
Quoted 31 lines
--- arch/powerpc/include/asm/reg.h | 3 +++ 1 file changed, 3 insertions(+)arch/powerpc/include/asm/reg.h +3
-0@@ -38,6 +38,7 @@#define MSR_TM_LG 32/* Trans Mem Available */ #define MSR_VEC_LG 25/* Enable AltiVec */ #define MSR_VSX_LG 23/* Enable VSX */ +#define MSR_S_LG 22/* Secure VM bit */ #define MSR_POW_LG 18/* Enable Power Management */ #define MSR_WE_LG 18/* Wait State Enable */ #define MSR_TGPR_LG 17/* TLB Update registers in use */ @@ -71,11 +72,13 @@#define MSR_SF __MASK(MSR_SF_LG)/* Enable 64 bit mode */ #define MSR_ISF __MASK(MSR_ISF_LG)/* Interrupt 64b mode valid on 630 */ #define MSR_HV __MASK(MSR_HV_LG)/* Hypervisor state */ +#define MSR_S __MASK(MSR_S_LG)/* Secure state */ #else /* so tests for these bits fail on 32-bit */ #define MSR_SF 0#define MSR_ISF 0#define MSR_HV 0 +#define MSR_S 0#endif /*
Michael Ellerman
reviewerRe: [PATCH v4 1/8] KVM: PPC: Ultravisor: Introduce the MSR_S bit
In reply to Claudio CarvalhoQuoted 10 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the
MSR_Sbit is used to indicate if the thread is in secure state. With theMSR_Sbit, the privilege state of the thread is now determined byMSR_S,MSR_HVandMSR_PR, as follows:
Quoted 8 lines
S HV PR
-----------------------0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reserved
What are you trying to express with the 'x' value?
I guess you mean it as "either" or "don't care" - but then you have cases where it could only have one value, such as hypervisor. I think it would be clearer if you spelled it out more explicitly.
Quoted 2 lines
The hypervisor doesn't (and can't) run with the
MSR_Sbit set, but a secure guest and the ultravisor firmware do.
I know you're trying to be helpful, but this comment is really just confusing to someone who doesn't have all the documentation.
I'd really like to see something in Documentation/powerpc describing at least the outline of how the system works. I'm pretty sure most of that is public, so even if it's mostly a list of references to other documentations or presentations that would be fine. But I'm not really happy with a whole new processor mode appearing with zero documentation in the tree.
Quoted 3 lines
Signed-off-by Sukadev Bhattiprolu <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Update the commit message ]
It's normal to prefix these comments with your handle to make it clear who is saying it.
Quoted 10 lines
Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/reg.h | 3 +++ 1 file changed, 3 insertions(+)arch/powerpc/include/asm/reg.h +0
-0
arch/powerpc/include/asm/reg.h line 41
Quoted 4 lines
arch/powerpc/include/asm/reg.h +1
-0@@ -38,6 +38,7 @@#define MSR_TM_LG 32/* Trans Mem Available */ #define MSR_VEC_LG 25/* Enable AltiVec */ #define MSR_VSX_LG 23/* Enable VSX */ +#define MSR_S_LG 22/* Secure VM bit */
I don't think that's the best description, because it's also the Ultravisor bit when MSR[HV] = 1.
So "Secure state" as you have below would be better IMO.
cheers
Quoted 19 lines
arch/powerpc/include/asm/reg.h +2
-0#define MSR_POW_LG 18/* Enable Power Management */ #define MSR_WE_LG 18/* Wait State Enable */ #define MSR_TGPR_LG 17/* TLB Update registers in use */ @@ -71,11 +72,13 @@#define MSR_SF __MASK(MSR_SF_LG)/* Enable 64 bit mode */ #define MSR_ISF __MASK(MSR_ISF_LG)/* Interrupt 64b mode valid on 630 */ #define MSR_HV __MASK(MSR_HV_LG)/* Hypervisor state */ +#define MSR_S __MASK(MSR_S_LG)/* Secure state */ #else /* so tests for these bits fail on 32-bit */ #define MSR_SF 0#define MSR_ISF 0#define MSR_HV 0 +#define MSR_S 0#endif /* Signature · 1 line
2.20.1
Nicholas Piggin
reviewerRe: [PATCH v4 1/8] KVM: PPC: Ultravisor: Introduce the MSR_S bit
Michael Ellerman's on July 11, 2019 10:57 pm:
Quoted 64 lines
Quoted 10 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the
MSR_Sbit is used to indicate if the thread is in secure state. With theMSR_Sbit, the privilege state of the thread is now determined byMSR_S,MSR_HVandMSR_PR, as follows:Quoted 8 lines
S HV PR
-----------------------0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reservedWhat are you trying to express with the 'x' value?
I guess you mean it as "either" or "don't care" - but then you have cases where it could only have one value, such as hypervisor. I think it would be clearer if you spelled it out more explicitly.
Quoted 2 lines
The hypervisor doesn't (and can't) run with the
MSR_Sbit set, but a secure guest and the ultravisor firmware do.I know you're trying to be helpful, but this comment is really just confusing to someone who doesn't have all the documentation.
I'd really like to see something in
Documentation/powerpcdescribing at least the outline of how the system works. I'm pretty sure most of that is public, so even if it's mostly a list of references to other documentations or presentations that would be fine. But I'm not really happy with a whole new processor mode appearing with zero documentation in the tree.Quoted 3 lines
Signed-off-by Sukadev Bhattiprolu <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Update the commit message ]It's normal to prefix these comments with your handle to make it clear who is saying it.
Quoted 10 lines
Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/reg.h | 3 +++ 1 file changed, 3 insertions(+)arch/powerpc/include/asm/reg.h +0
-0arch/powerpc/include/asm/reg.h line 41
Quoted 4 lines
arch/powerpc/include/asm/reg.h +1
-0@@ -38,6 +38,7 @@#define MSR_TM_LG 32/* Trans Mem Available */ #define MSR_VEC_LG 25/* Enable AltiVec */ #define MSR_VSX_LG 23/* Enable VSX */ +#define MSR_S_LG 22/* Secure VM bit */ I don't think that's the best description, because it's also the Ultravisor bit when MSR[HV] = 1.
Quoted 1 line
So "Secure state" as you have below would be better IMO.
Ooops I see Michael covered everything I wrote, sorry for the noise I missed the thread.
Thanks,
Nick
Nicholas Piggin
reviewerRe: [PATCH v4 1/8] KVM: PPC: Ultravisor: Introduce the MSR_S bit
In reply to Claudio CarvalhoClaudio Carvalho's on June 29, 2019 6:08 am:
Quoted 10 lines
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the
MSR_Sbit is used to indicate if the thread is in secure state. With theMSR_Sbit, the privilege state of the thread is now determined byMSR_S,MSR_HVandMSR_PR, as follows:
Quoted 8 lines
S HV PR
-----------------------0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reserved
What does this table mean? I thought 'x' meant either, but in that case there are several states that can apply to the same combination of bits.
Would it be clearer to rearrange the table so the columns are the HV and PR bits we know and love, plus the effect of S=1 on each of them?
HV PR S=0 S=1
---------------------------------------------
0 0 privileged privileged (secure guest kernel)
0 1 problem problem (secure guest userspace)
1 0 hypervisor ultravisor
1 1 problem reserved
Is that accurate?
Quoted 25 lines
The hypervisor doesn't (and can't) run with the
MSR_Sbit set, but a secure guest and the ultravisor firmware do.Signed-off-by Sukadev Bhattiprolu <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Update the commit message ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/reg.h | 3 +++ 1 file changed, 3 insertions(+)arch/powerpc/include/asm/reg.h +1
-0@@ -38,6 +38,7 @@#define MSR_TM_LG 32/* Trans Mem Available */ #define MSR_VEC_LG 25/* Enable AltiVec */ #define MSR_VSX_LG 23/* Enable VSX */ +#define MSR_S_LG 22/* Secure VM bit */ #define MSR_POW_LG 18/* Enable Power Management */ #define MSR_WE_LG 18/* Wait State Enable */ #define MSR_TGPR_LG 17/* TLB Update registers in use */
arch/powerpc/include/asm/reg.h line 75
Quoted 4 lines
arch/powerpc/include/asm/reg.h +1
-0@@ -71,11 +72,13 @@#define MSR_SF __MASK(MSR_SF_LG)/* Enable 64 bit mode */ #define MSR_ISF __MASK(MSR_ISF_LG)/* Interrupt 64b mode valid on 630 */ #define MSR_HV __MASK(MSR_HV_LG)/* Hypervisor state */ +#define MSR_S __MASK(MSR_S_LG)/* Secure state */
This is a real nitpick, but why two different comments for the bit number and the mask?
Michael Ellerman
reviewerRe: [PATCH v4 1/8] KVM: PPC: Ultravisor: Introduce the MSR_S bit
Quoted 27 lines · Nicholas Piggin
Claudio Carvalho's on June 29, 2019 6:08 am:
Quoted 10 lines
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the
MSR_Sbit is used to indicate if the thread is in secure state. With theMSR_Sbit, the privilege state of the thread is now determined byMSR_S,MSR_HVandMSR_PR, as follows:Quoted 8 lines
S HV PR
-----------------------0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reservedWhat does this table mean? I thought 'x' meant either, but in that case there are several states that can apply to the same combination of bits.
Would it be clearer to rearrange the table so the columns are the HV and PR bits we know and love, plus the effect of S=1 on each of them?
Quoted 8 lines
HV PR S=0 S=1
---------------------------------------------
0 0 privileged privileged (secure guest kernel)
0 1 problem problem (secure guest userspace)
1 0 hypervisor ultravisor
1 1 problem reservedIs that accurate?
I like that format.
cheers
Re: [PATCH v4 1/8] KVM: PPC: Ultravisor: Introduce the MSR_S bit
In reply to Nicholas PigginQuoted 19 lines · Nicholas Piggin
Claudio Carvalho's on June 29, 2019 6:08 am:
Quoted 10 lines
From: Sukadev Bhattiprolu <[email protected]>
The ultravisor processor mode is introduced in POWER platforms that supports the Protected Execution Facility (PEF). Ultravisor is higher privileged than hypervisor mode.
In PEF enabled platforms, the
MSR_Sbit is used to indicate if the thread is in secure state. With theMSR_Sbit, the privilege state of the thread is now determined byMSR_S,MSR_HVandMSR_PR, as follows:Quoted 8 lines
S HV PR
-----------------------0 x 1 problem 1 0 1 problem x x 0 privileged x 1 0 hypervisor 1 1 0 ultravisor 1 1 1 reserved
Quoted 1 line
What does this table mean? I thought 'x' meant either
Yes, it means either. The table was arranged that way to say that:
- hypervisor state is also a privileged state,
- ultravisor state is also a hypervisor state.
Quoted 7 lines
, but in that
case there are several states that can apply to the same combination of bits.Would it be clearer to rearrange the table so the columns are the HV and PR bits we know and love, plus the effect of S=1 on each of them?
Quoted 8 lines
HV PR S=0 S=1
---------------------------------------------
0 0 privileged privileged (secure guest kernel)
0 1 problem problem (secure guest userspace)
1 0 hypervisor ultravisor
1 1 problem reservedIs that accurate?
Yes, it is. I also like this format. I will consider it.
Quoted 24 lines
Quoted 24 lines
The hypervisor doesn't (and can't) run with the
MSR_Sbit set, but a secure guest and the ultravisor firmware do.Signed-off-by Sukadev Bhattiprolu <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Update the commit message ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/reg.h | 3 +++ 1 file changed, 3 insertions(+)arch/powerpc/include/asm/reg.h +1
-0@@ -38,6 +38,7 @@#define MSR_TM_LG 32/* Trans Mem Available */ #define MSR_VEC_LG 25/* Enable AltiVec */ #define MSR_VSX_LG 23/* Enable VSX */ +#define MSR_S_LG 22/* Secure VM bit */ #define MSR_POW_LG 18/* Enable Power Management */ #define MSR_WE_LG 18/* Wait State Enable */
Quoted 8 lines
arch/powerpc/include/asm/reg.h line 75
Quoted 6 lines
arch/powerpc/include/asm/reg.h +1
-0#define MSR_TGPR_LG 17/* TLB Update registers in use */ @@ -71,11 +72,13 @@#define MSR_SF __MASK(MSR_SF_LG)/* Enable 64 bit mode */ #define MSR_ISF __MASK(MSR_ISF_LG)/* Interrupt 64b mode valid on 630 */ #define MSR_HV __MASK(MSR_HV_LG)/* Hypervisor state */ +#define MSR_S __MASK(MSR_S_LG)/* Secure state */ This is a real nitpick, but why two different comments for the bit number and the mask?
Fixed for the next version. Both comments will be /* Secure state */
Thanks
Claudio
PATCH v4 2/8 powerpc: Introduce FW_FEATURE_ULTRAVISOR
In reply to Claudio CarvalhoThis feature tells if the ultravisor firmware is available to handle ucalls.
--- arch/powerpc/include/asm/firmware.h | 5 +++-- arch/powerpc/include/asm/ultravisor.h | 15 +++++++++++++++ arch/powerpc/kernel/Makefile | 1 + arch/powerpc/kernel/prom.c | 4 ++++ arch/powerpc/kernel/ultravisor.c | 24 ++++++++++++++++++++++++ 5 files changed, 47 insertions(+), 2 deletions(-) create mode 100644 arch/powerpc/include/asm/ultravisor.h create mode 100644 arch/powerpc/kernel/ultravisor.c
arch/powerpc/include/asm/firmware.h +3 -2
@@ -54,6 +54,7 @@#define FW_FEATURE_DRC_INFO ASM_CONST(0x0000000800000000)#define FW_FEATURE_BLOCK_REMOVE ASM_CONST(0x0000001000000000)#define FW_FEATURE_PAPR_SCM ASM_CONST(0x0000002000000000) +#define FW_FEATURE_ULTRAVISOR ASM_CONST(0x0000004000000000)#ifndef __ASSEMBLY__
@@ -72,9 +73,9 @@ enum {FW_FEATURE_TYPE1_AFFINITY |FW_FEATURE_PRRN |FW_FEATURE_HPT_RESIZE |FW_FEATURE_DRMEM_V2 |FW_FEATURE_DRC_INFO |FW_FEATURE_BLOCK_REMOVE | -FW_FEATURE_PAPR_SCM , +FW_FEATURE_PAPR_SCM |FW_FEATURE_ULTRAVISOR ,FW_FEATURE_PSERIES_ALWAYS =0 , -FW_FEATURE_POWERNV_POSSIBLE =FW_FEATURE_OPAL , +FW_FEATURE_POWERNV_POSSIBLE =FW_FEATURE_OPAL |FW_FEATURE_ULTRAVISOR ,FW_FEATURE_POWERNV_ALWAYS =0 ,FW_FEATURE_PS3_POSSIBLE =FW_FEATURE_LPAR |FW_FEATURE_PS3_LV1 , FW_FEATURE_PS3_ALWAYS = FW_FEATURE_LPAR | FW_FEATURE_PS3_LV1,
arch/powerpc/include/asm/ultravisor.h new file +15 -0
@@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor definitions +* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_H +#define _ASM_POWERPC_ULTRAVISOR_H + +/* Internal functions */ +extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data); + +#endif /* _ASM_POWERPC_ULTRAVISOR_H */
arch/powerpc/kernel/Makefile +1 -0
@@ -154,6 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o # Disable GCOV, KCOV & sanitizers in odd or sensitive code GCOV_PROFILE_prom_init.o := n
arch/powerpc/kernel/prom.c +4 -0
@@ -59,6 +59,7 @@#include <asm/firmware.h> #include <asm/dt_cpu_ftrs.h> #include <asm/drmem.h> +#include <asm/ultravisor.h> #include <mm/mmu_decl.h>
@@ -706,6 +707,9 @@ void __init early_init_devtree(void *params)#ifdef CONFIG_PPC_POWERNV /* Some machines might need OPAL info for debugging, grab it now. */ of_scan_flat_dt (early_init_dt_scan_opal,NULL ); + +/* Scan tree for ultravisor feature */ +of_scan_flat_dt (early_init_dt_scan_ultravisor,NULL );#endif #ifdef CONFIG_FA_DUMP
arch/powerpc/kernel/ultravisor.c new file +24 -0
@@ -0,0 +1,24 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Ultravisor high level interfaces +* +* Copyright 2019, IBM Corporation. +* +*/ +#include <linux/init.h> +#include <linux/printk.h> +#include <linux/string.h> + +#include <asm/ultravisor.h> +#include <asm/firmware.h> + +int __init early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data) +{ +if (depth !=1 ||strcmp (uname,"ibm,ultravisor" ) !=0 ) +return 0 ; + + powerpc_firmware_features |=FW_FEATURE_ULTRAVISOR ; +pr_debug ("Ultravisor detected!\n" ); +return 1 ; +}
Signature · 1 line
2.20.1
Re: [PATCH v4 2/8] powerpc: Introduce FW_FEATURE_ULTRAVISOR
Quoted 6 lines · Claudio Carvalho
This feature tells if the ultravisor firmware is available to handle ucalls.
Signed-off-by Claudio Carvalho <[email protected]>[ Device node name to "ibm,ultravisor" ]Signed-off-by Michael Anderson <[email protected]>
Quoted 124 lines
--- arch/powerpc/include/asm/firmware.h | 5 +++-- arch/powerpc/include/asm/ultravisor.h | 15 +++++++++++++++ arch/powerpc/kernel/Makefile | 1 + arch/powerpc/kernel/prom.c | 4 ++++ arch/powerpc/kernel/ultravisor.c | 24 ++++++++++++++++++++++++ 5 files changed, 47 insertions(+), 2 deletions(-) create mode 100644 arch/powerpc/include/asm/ultravisor.h create mode 100644 arch/powerpc/kernel/ultravisor.carch/powerpc/include/asm/firmware.h +3
-2@@ -54,6 +54,7 @@#define FW_FEATURE_DRC_INFO ASM_CONST(0x0000000800000000)#define FW_FEATURE_BLOCK_REMOVE ASM_CONST(0x0000001000000000)#define FW_FEATURE_PAPR_SCM ASM_CONST(0x0000002000000000) +#define FW_FEATURE_ULTRAVISOR ASM_CONST(0x0000004000000000)#ifndef __ASSEMBLY__@@ -72,9 +73,9 @@ enum {FW_FEATURE_TYPE1_AFFINITY |FW_FEATURE_PRRN |FW_FEATURE_HPT_RESIZE |FW_FEATURE_DRMEM_V2 |FW_FEATURE_DRC_INFO |FW_FEATURE_BLOCK_REMOVE | -FW_FEATURE_PAPR_SCM , +FW_FEATURE_PAPR_SCM |FW_FEATURE_ULTRAVISOR ,FW_FEATURE_PSERIES_ALWAYS =0 , -FW_FEATURE_POWERNV_POSSIBLE =FW_FEATURE_OPAL , +FW_FEATURE_POWERNV_POSSIBLE =FW_FEATURE_OPAL |FW_FEATURE_ULTRAVISOR ,FW_FEATURE_POWERNV_ALWAYS =0 , FW_FEATURE_PS3_POSSIBLE = FW_FEATURE_LPAR | FW_FEATURE_PS3_LV1,arch/powerpc/include/asm/firmware.h +0
-0FW_FEATURE_PS3_ALWAYS = FW_FEATURE_LPAR | FW_FEATURE_PS3_LV1,arch/powerpc/include/asm/ultravisor.h new file +12
-0@@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor definitions +* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_H +#define _ASM_POWERPC_ULTRAVISOR_H + +/* Internal functions */ +extern int early_init_dt_scan_ultravisor(unsigned long node, constchar *uname,
arch/powerpc/include/asm/ultravisor.h new file +3
-0+ int depth, void *data); + +#endif /* _ASM_POWERPC_ULTRAVISOR_H */ arch/powerpc/kernel/Makefile +1
-0@@ -154,6 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o # Disable GCOV, KCOV & sanitizers in odd or sensitive code GCOV_PROFILE_prom_init.o := narch/powerpc/kernel/prom.c +4
-0@@ -59,6 +59,7 @@#include <asm/firmware.h> #include <asm/dt_cpu_ftrs.h> #include <asm/drmem.h> +#include <asm/ultravisor.h> #include <mm/mmu_decl.h> @@ -706,6 +707,9 @@ void __init early_init_devtree(void *params)#ifdef CONFIG_PPC_POWERNV /* Some machines might need OPAL info for debugging, grab it now. */ of_scan_flat_dt (early_init_dt_scan_opal,NULL ); + +/* Scan tree for ultravisor feature */ +of_scan_flat_dt (early_init_dt_scan_ultravisor,NULL );#endif #ifdef CONFIG_FA_DUMP arch/powerpc/kernel/ultravisor.c new file +15
-0@@ -0,0 +1,24 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Ultravisor high level interfaces +* +* Copyright 2019, IBM Corporation. +* +*/ +#include <linux/init.h> +#include <linux/printk.h> +#include <linux/string.h> + +#include <asm/ultravisor.h> +#include <asm/firmware.h> + +int __init early_init_dt_scan_ultravisor(unsigned long node, constchar *uname,
arch/powerpc/kernel/ultravisor.c new file +9
-0+ int depth, void *data) +{ +if (depth !=1 ||strcmp (uname,"ibm,ultravisor" ) !=0 ) +return 0 ; + + powerpc_firmware_features |=FW_FEATURE_ULTRAVISOR ; +pr_debug ("Ultravisor detected!\n" ); +return 1 ; +}
Michael Ellerman
reviewerRe: [PATCH v4 2/8] powerpc: Introduce FW_FEATURE_ULTRAVISOR
In reply to Claudio CarvalhoQuoted 9 lines · Claudio Carvalho
arch/powerpc/include/asm/ultravisor.h new file +3
-0@@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor definitions
arch/powerpc/include/asm/ultravisor.h line 13
Quoted 10 lines
arch/powerpc/include/asm/ultravisor.h new file +10
-0+* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_H +#define _ASM_POWERPC_ULTRAVISOR_H + +/* Internal functions */ +extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data);
Please don't use extern in new headers.
arch/powerpc/kernel/ultravisor.c
Quoted 5 lines
arch/powerpc/kernel/ultravisor.c new file +0
-0
Is there a reason this (and other later files) aren't in platforms/powernv ?
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c new file +9
-0@@ -0,0 +1,24 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Ultravisor high level interfaces +* +* Copyright 2019, IBM Corporation. +* +*/ +#include <linux/init.h> +#include <linux/printk.h>
arch/powerpc/kernel/ultravisor.c line 19
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c new file +10
-0+#include <linux/string.h> + +#include <asm/ultravisor.h> +#include <asm/firmware.h> + +int __init early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data) +{ +if (depth !=1 ||strcmp (uname,"ibm,ultravisor" ) !=0 ) +return 0 ;
I know you're following the example of OPAL, but this is not the best way to search for the ultravisor node.
It makes the location and name of the node part of the ABI, when there's no need for it to be.
If instead you just scan the tree for a node that is *compatible* with "ibm,ultravisor" (or whatever compatible string) then the node can be placed any where in the tree and have any name, which gives us the most flexibility in future to change the location of the device tree node.
cheers
Re: [PATCH v4 2/8] powerpc: Introduce FW_FEATURE_ULTRAVISOR
Quoted 22 lines · Michael Ellerman
Quoted 9 lines · Claudio Carvalho
arch/powerpc/include/asm/ultravisor.h new file +3
-0@@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor definitions arch/powerpc/include/asm/ultravisor.h line 13
Quoted 10 lines
arch/powerpc/include/asm/ultravisor.h new file +10
-0+* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_H +#define _ASM_POWERPC_ULTRAVISOR_H + +/* Internal functions */ +extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data);Please don't use extern in new headers.
Quoted 6 lines
arch/powerpc/kernel/ultravisor.c
Quoted 5 lines
arch/powerpc/kernel/ultravisor.c new file +0
-0Is there a reason this (and other later files) aren't in platforms/powernv ?
Yes, there is.
https://www.spinics.net/lists/kvm-ppc/msg14998.html
We also need to do ucalls from a secure guest and its kernel may not have CONFIG_PPC_POWERNV=y. I can make it clear in the commit message.
Quoted 24 lines
Quoted 11 lines
arch/powerpc/kernel/ultravisor.c new file +9
-0@@ -0,0 +1,24 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Ultravisor high level interfaces +* +* Copyright 2019, IBM Corporation. +* +*/ +#include <linux/init.h> +#include <linux/printk.h> arch/powerpc/kernel/ultravisor.c line 19
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c new file +10
-0+#include <linux/string.h> + +#include <asm/ultravisor.h> +#include <asm/firmware.h> + +int __init early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data) +{ +if (depth !=1 ||strcmp (uname,"ibm,ultravisor" ) !=0 ) +return 0 ;I know you're following the example of OPAL, but this is not the best way to search for the ultravisor node.
Quoted 7 lines
It makes the location and name of the node part of the ABI, when there's no need for it to be.
If instead you just scan the tree for a node that is *compatible* with "ibm,ultravisor" (or whatever compatible string) then the node can be placed any where in the tree and have any name, which gives us the most flexibility in future to change the location of the device tree node.
I will do that in the next version.
Thanks,
Claudio
Quoted 3 lines
cheers
Michael Ellerman
reviewerRe: [PATCH v4 2/8] powerpc: Introduce FW_FEATURE_ULTRAVISOR
Quoted 30 lines · Claudio Carvalho
Quoted 22 lines · Michael Ellerman
Quoted 9 lines · Claudio Carvalho
arch/powerpc/include/asm/ultravisor.h new file +3
-0@@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor definitions arch/powerpc/include/asm/ultravisor.h line 13
Quoted 10 lines
arch/powerpc/include/asm/ultravisor.h new file +10
-0+* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_H +#define _ASM_POWERPC_ULTRAVISOR_H + +/* Internal functions */ +extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data);Please don't use extern in new headers.
Quoted 6 lines
arch/powerpc/kernel/ultravisor.c
Quoted 5 lines
arch/powerpc/kernel/ultravisor.c new file +0
-0Is there a reason this (and other later files) aren't in platforms/powernv ?
Quoted 5 lines
Yes, there is.
https://www.spinics.net/lists/kvm-ppc/msg14998.htmlWe also need to do ucalls from a secure guest and its kernel may not have
CONFIG_PPC_POWERNV=y. I can make it clear in the commit message.
OK, sorry I missed Paul's comment. Yeah that is obviously a valid reason :)
Quoted 26 lines
Quoted 26 lines
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c new file +9
-0@@ -0,0 +1,24 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Ultravisor high level interfaces +* +* Copyright 2019, IBM Corporation. +* +*/ +#include <linux/init.h> +#include <linux/printk.h> arch/powerpc/kernel/ultravisor.c line 19
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c new file +10
-0+#include <linux/string.h> + +#include <asm/ultravisor.h> +#include <asm/firmware.h> + +int __init early_init_dt_scan_ultravisor (unsigned long node,const char *uname, +int depth,void *data) +{ +if (depth !=1 ||strcmp (uname,"ibm,ultravisor" ) !=0 ) +return 0 ;I know you're following the example of OPAL, but this is not the best way to search for the ultravisor node.
It makes the location and name of the node part of the ABI, when there's no need for it to be.
Quoted 6 lines
Quoted 4 lines
If instead you just scan the tree for a node that is *compatible* with "ibm,ultravisor" (or whatever compatible string) then the node can be placed any where in the tree and have any name, which gives us the most flexibility in future to change the location of the device tree node.
I will do that in the next version.
Excellent, thanks.
cheers
PATCH v4 3/8 KVM: PPC: Ultravisor: Add generic ultravisor call handler
In reply to Claudio CarvalhoFrom: Ram Pai <[email protected]>
Add the ucall() function, which can be used to make ultravisor calls with varied number of in and out arguments. Ultravisor calls can be made from the host or guests.
This copies the implementation of plpar_hcall().
[ Change ucall.S to not save CR, rename and move headers, build ucall.S
if CONFIG_PPC_POWERNV set, use R3 for the ucall number and add some comments in the code ]
--- arch/powerpc/include/asm/ultravisor-api.h | 20 +++++++++++++++ arch/powerpc/include/asm/ultravisor.h | 20 +++++++++++++++ arch/powerpc/kernel/Makefile | 2 +- arch/powerpc/kernel/ucall.S | 30 +++++++++++++++++++++++ arch/powerpc/kernel/ultravisor.c | 4 +++ 5 files changed, 75 insertions(+), 1 deletion(-) create mode 100644 arch/powerpc/include/asm/ultravisor-api.h create mode 100644 arch/powerpc/kernel/ucall.S
arch/powerpc/include/asm/ultravisor-api.h new file +20 -0
@@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor API. +* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_API_H +#define _ASM_POWERPC_ULTRAVISOR_API_H + +#include <asm/hvcall.h> + +/* Return codes */ +#define U_NOT_AVAILABLE H_NOT_AVAILABLE +#define U_SUCCESS H_SUCCESS +#define U_FUNCTION H_FUNCTION +#define U_PARAMETER H_PARAMETER + +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ +
arch/powerpc/include/asm/ultravisor.h +20 -0
@@ -8,8 +8,28 @@#ifndef _ASM_POWERPC_ULTRAVISOR_H#define _ASM_POWERPC_ULTRAVISOR_H +#include <asm/ultravisor-api.h> + +#if !defined(__ASSEMBLY__) +/* Internal functions */ extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data); +/* API functions */ +#define UCALL_BUFSIZE 4 +/** +* ucall: Make a powerpc ultravisor call. +* @opcode: The ultravisor call to make. +* @retbuf: Buffer to store up to 4 return arguments in. +* +* This call supports up to 6 arguments and 4 return arguments. Use +* UCALL_BUFSIZE to size the return argument buffer. +*/ +#if defined(CONFIG_PPC_POWERNV ) +long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#endif + +#endif /* !__ASSEMBLY__ */ +#endif /* _ASM_POWERPC_ULTRAVISOR_H */
arch/powerpc/kernel/Makefile +1 -1
@@ -154,7 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o -obj-$(CONFIG_PPC_POWERNV) += ultravisor.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o ucall.o # Disable GCOV, KCOV & sanitizers in odd or sensitive code GCOV_PROFILE_prom_init.o := n
arch/powerpc/kernel/ucall.S new file +30 -0
@@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Generic code to perform an ultravisor call. + * + * Copyright 2019, IBM Corporation. + * + */ +#include <asm/ppc_asm.h> + +/* + * This function is based on the plpar_hcall() + */ +_GLOBAL_TOC(ucall) + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + + sc 2 /* Invoke the ultravisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + blr /* Return r3 = status */
arch/powerpc/kernel/ultravisor.c +4 -0
@@ -8,10 +8,14 @@#include <linux/init.h> #include <linux/printk.h> #include <linux/string.h> +#include <linux/export.h> #include <asm/ultravisor.h> #include <asm/firmware.h> +/* in ucall.S */ +EXPORT_SYMBOL_GPL (ucall); +int __init early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data) {
Signature · 1 line
2.20.1
Re: [PATCH v4 3/8] KVM: PPC: Ultravisor: Add generic ultravisor call handler
Quoted 7 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Add the
ucall()function, which can be used to make ultravisor calls with varied number of in and out arguments. Ultravisor calls can be made
from the host or guests.
Quoted 7 lines
This copies the implementation of
plpar_hcall().Signed-off-by Ram Pai <[email protected]>[ Change
ucall.Sto not save CR, rename and move headers, builducall.S
ifCONFIG_PPC_POWERNVset, use R3 for the ucall number and add some comments in the code ]Signed-off-by Claudio Carvalho <[email protected]>
Quoted 143 lines
--- arch/powerpc/include/asm/ultravisor-api.h | 20 +++++++++++++++ arch/powerpc/include/asm/ultravisor.h | 20 +++++++++++++++ arch/powerpc/kernel/Makefile | 2 +- arch/powerpc/kernel/ucall.S | 30 +++++++++++++++++++++++ arch/powerpc/kernel/ultravisor.c | 4 +++ 5 files changed, 75 insertions(+), 1 deletion(-) create mode 100644 arch/powerpc/include/asm/ultravisor-api.h create mode 100644 arch/powerpc/kernel/ucall.Sarch/powerpc/include/asm/ultravisor-api.h new file +20
-0@@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor API. +* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_API_H +#define _ASM_POWERPC_ULTRAVISOR_API_H + +#include <asm/hvcall.h> + +/* Return codes */ +#define U_NOT_AVAILABLE H_NOT_AVAILABLE +#define U_SUCCESS H_SUCCESS +#define U_FUNCTION H_FUNCTION +#define U_PARAMETER H_PARAMETER + +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ +arch/powerpc/include/asm/ultravisor.h +20
-0@@ -8,8 +8,28 @@#ifndef _ASM_POWERPC_ULTRAVISOR_H#define _ASM_POWERPC_ULTRAVISOR_H +#include <asm/ultravisor-api.h> + +#if !defined(__ASSEMBLY__) +/* Internal functions */ extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data); +/* API functions */ +#define UCALL_BUFSIZE 4 +/** +* ucall: Make a powerpc ultravisor call. +* @opcode: The ultravisor call to make. +* @retbuf: Buffer to store up to 4 return arguments in. +* +* This call supports up to 6 arguments and 4 return arguments. Use +* UCALL_BUFSIZE to size the return argument buffer. +*/ +#if defined(CONFIG_PPC_POWERNV ) +long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#endif + +#endif /* !__ASSEMBLY__ */ +#endif /*
_ASM_POWERPC_ULTRAVISOR_H*/arch/powerpc/kernel/Makefile +1
-1@@ -154,7 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o -obj-$(CONFIG_PPC_POWERNV) += ultravisor.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o ucall.o # Disable GCOV, KCOV & sanitizers in odd or sensitive code GCOV_PROFILE_prom_init.o := narch/powerpc/kernel/ucall.S new file +30
-0@@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Generic code to perform an ultravisor call. + * + * Copyright 2019, IBM Corporation. + * + */ +#include <asm/ppc_asm.h> + +/* + * This function is based on the plpar_hcall() + */ +_GLOBAL_TOC(ucall) + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + + sc 2 /* Invoke the ultravisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + blr /* Return r3 = status */arch/powerpc/kernel/ultravisor.c +4
-0@@ -8,10 +8,14 @@#include <linux/init.h> #include <linux/printk.h> #include <linux/string.h> +#include <linux/export.h> #include <asm/ultravisor.h> #include <asm/firmware.h> +/* in ucall.S */ +EXPORT_SYMBOL_GPL (ucall); +int __init early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data){
Michael Ellerman
reviewerRe: [PATCH v4 3/8] KVM: PPC: Ultravisor: Add generic ultravisor call handler
In reply to Claudio CarvalhoQuoted 7 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Add the
ucall()function, which can be used to make ultravisor calls with varied number of in and out arguments. Ultravisor calls can be made from the host or guests.This copies the implementation of
plpar_hcall().
.. with quite a few changes?
This is one of the things I'd like to see in a Documentation file, so that people can review the implementation vs the specification.
Quoted 4 lines
Signed-off-by Ram Pai <[email protected]>[ Change
ucall.Sto not save CR, rename and move headers, builducall.S
ifCONFIG_PPC_POWERNVset, use R3 for the ucall number and add some comments in the code ]
Why are we not saving CR? See previous comment about Documentation :)
Quoted 24 lines
Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/ultravisor-api.h | 20 +++++++++++++++ arch/powerpc/include/asm/ultravisor.h | 20 +++++++++++++++ arch/powerpc/kernel/Makefile | 2 +- arch/powerpc/kernel/ucall.S | 30 +++++++++++++++++++++++ arch/powerpc/kernel/ultravisor.c | 4 +++ 5 files changed, 75 insertions(+), 1 deletion(-) create mode 100644 arch/powerpc/include/asm/ultravisor-api.h create mode 100644 arch/powerpc/kernel/ucall.Sarch/powerpc/include/asm/ultravisor-api.h new file +7
-0@@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor API. +* +* Copyright 2019, IBM Corporation. +* +*/
arch/powerpc/include/asm/ultravisor-api.h line 17
Quoted 10 lines
arch/powerpc/include/asm/ultravisor-api.h new file +10
-0+#ifndef _ASM_POWERPC_ULTRAVISOR_API_H +#define _ASM_POWERPC_ULTRAVISOR_API_H + +#include <asm/hvcall.h> + +/* Return codes */ +#define U_NOT_AVAILABLE H_NOT_AVAILABLE +#define U_SUCCESS H_SUCCESS +#define U_FUNCTION H_FUNCTION +#define U_PARAMETER H_PARAMETER
Is there any benefit in redefining these?
arch/powerpc/include/asm/ultravisor.h line 13
Quoted 11 lines
arch/powerpc/include/asm/ultravisor.h +3
-0@@ -8,8 +8,28 @@#ifndef _ASM_POWERPC_ULTRAVISOR_H#define _ASM_POWERPC_ULTRAVISOR_H +#include <asm/ultravisor-api.h> + +#if !defined(__ASSEMBLY__)
Just #ifndef is fine.
arch/powerpc/include/asm/ultravisor.h line 15
Quoted 1 line
arch/powerpc/include/asm/ultravisor.h +0
-0/* Internal functions */
How is it internal?
arch/powerpc/include/asm/ultravisor.h line 20
Quoted 5 lines
arch/powerpc/include/asm/ultravisor.h +2
-0extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data); +/* API functions */ +#define UCALL_BUFSIZE 4
Please don't copy this design from the hcall code, it has led to bugs in the past.
See my (still unmerged) attempt to fix this for the hcall case:
https://patchwork.ozlabs.org/patch/683577/
Basically instead of asking callers nicely to define a certain sized buffer, and them forgetting, define a proper type that has the right size.
arch/powerpc/include/asm/ultravisor.h line 29
Quoted 9 lines
arch/powerpc/include/asm/ultravisor.h +9
-0+/** +* ucall: Make a powerpc ultravisor call. +* @opcode: The ultravisor call to make. +* @retbuf: Buffer to store up to 4 return arguments in. +* +* This call supports up to 6 arguments and 4 return arguments. Use +* UCALL_BUFSIZE to size the return argument buffer. +*/ +#if defined(CONFIG_PPC_POWERNV )
#ifdef
Quoted 11 lines
arch/powerpc/include/asm/ultravisor.h +5
-0+long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#endif + +#endif /* !__ASSEMBLY__ */ +#endif /* _ASM_POWERPC_ULTRAVISOR_H */ arch/powerpc/kernel/Makefile +0
-0
arch/powerpc/kernel/Makefile endif · line 157
Quoted 5 lines
arch/powerpc/kernel/Makefile +1
-1@@ -154,7 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o -obj-$(CONFIG_PPC_POWERNV) += ultravisor.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o ucall.o
Same comment about being platforms/powernv ?
Quoted 8 lines
arch/powerpc/kernel/ucall.S new file +1
-0@@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: GPL-2.0 */
arch/powerpc/kernel/ucall.S line 11
Quoted 10 lines
arch/powerpc/kernel/ucall.S new file +10
-0+/* + * Generic code to perform an ultravisor call. + * + * Copyright 2019, IBM Corporation. + * + */ +#include <asm/ppc_asm.h> + +/* + * This function is based on the plpar_hcall()
I don't think it meaningfully is any more.
arch/powerpc/kernel/ucall.S line 13
Quoted 2 lines
arch/powerpc/kernel/ucall.S new file +2
-0+ */ +_GLOBAL_TOC(ucall)
You don't need the TOC setup here (unless a later patch does?).
arch/powerpc/kernel/ucall.S line 20
Quoted 7 lines
arch/powerpc/kernel/ucall.S new file +7
-0+ std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10
Below you space the arguments, here you don't. Pick one or the other please.
Quoted 15 lines
arch/powerpc/kernel/ucall.S new file +10
-0+ + sc 2 /* Invoke the ultravisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + blr /* Return r3 = status */arch/powerpc/kernel/ultravisor.c +0
-0
arch/powerpc/kernel/ultravisor.c line 17
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c +3
-0@@ -8,10 +8,14 @@#include <linux/init.h> #include <linux/printk.h> #include <linux/string.h> +#include <linux/export.h> #include <asm/ultravisor.h> #include <asm/firmware.h> +/* in ucall.S */ +EXPORT_SYMBOL_GPL (ucall);
This should be in ucall.S
cheers
Re: [PATCH v4 3/8] KVM: PPC: Ultravisor: Add generic ultravisor call handler
Quoted 7 lines · Michael Ellerman
Quoted 6 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Add the
ucall()function, which can be used to make ultravisor calls with varied number of in and out arguments. Ultravisor calls can be made from the host or guests.
Quoted 5 lines
Quoted 1 line
This copies the implementation of
plpar_hcall()... with quite a few changes?
This is one of the things I'd like to see in a Documentation file, so that people can review the implementation vs the specification.
I will document this (and other things) in a file under Documentation/powerpc.
Quoted 67 lines
Quoted 5 lines
Signed-off-by Ram Pai <[email protected]>[ Change
ucall.Sto not save CR, rename and move headers, builducall.S
ifCONFIG_PPC_POWERNVset, use R3 for the ucall number and add some comments in the code ]Why are we not saving CR? See previous comment about Documentation :)
Quoted 24 lines
Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/ultravisor-api.h | 20 +++++++++++++++ arch/powerpc/include/asm/ultravisor.h | 20 +++++++++++++++ arch/powerpc/kernel/Makefile | 2 +- arch/powerpc/kernel/ucall.S | 30 +++++++++++++++++++++++ arch/powerpc/kernel/ultravisor.c | 4 +++ 5 files changed, 75 insertions(+), 1 deletion(-) create mode 100644 arch/powerpc/include/asm/ultravisor-api.h create mode 100644 arch/powerpc/kernel/ucall.Sarch/powerpc/include/asm/ultravisor-api.h new file +7
-0@@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor API. +* +* Copyright 2019, IBM Corporation. +* +*/ arch/powerpc/include/asm/ultravisor-api.h line 17
Quoted 10 lines
arch/powerpc/include/asm/ultravisor-api.h new file +10
-0+#ifndef _ASM_POWERPC_ULTRAVISOR_API_H +#define _ASM_POWERPC_ULTRAVISOR_API_H + +#include <asm/hvcall.h> + +/* Return codes */ +#define U_NOT_AVAILABLE H_NOT_AVAILABLE +#define U_SUCCESS H_SUCCESS +#define U_FUNCTION H_FUNCTION +#define U_PARAMETER H_PARAMETERIs there any benefit in redefining these?
arch/powerpc/include/asm/ultravisor.h line 13
Quoted 11 lines
arch/powerpc/include/asm/ultravisor.h +3
-0@@ -8,8 +8,28 @@#ifndef _ASM_POWERPC_ULTRAVISOR_H#define _ASM_POWERPC_ULTRAVISOR_H +#include <asm/ultravisor-api.h> + +#if !defined(__ASSEMBLY__)Just #ifndef is fine.
arch/powerpc/include/asm/ultravisor.h line 15
Quoted 1 line
arch/powerpc/include/asm/ultravisor.h +0
-0/* Internal functions */ How is it internal?
arch/powerpc/include/asm/ultravisor.h line 20
Quoted 5 lines
arch/powerpc/include/asm/ultravisor.h +2
-0extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data); +/* API functions */ +#define UCALL_BUFSIZE 4Please don't copy this design from the hcall code, it has led to bugs in the past.
Quoted 5 lines
See my (still unmerged) attempt to fix this for the hcall case:
https://patchwork.ozlabs.org/patch/683577/Basically instead of asking callers nicely to define a certain sized buffer, and them forgetting, define a proper type that has the right size.
I will keep that in mind. For now I think we don't need that since the v5 will have ucall_norets() instead.
Quoted 81 lines
arch/powerpc/include/asm/ultravisor.h line 29
Quoted 10 lines
arch/powerpc/include/asm/ultravisor.h +9
-0+/** +* ucall: Make a powerpc ultravisor call. +* @opcode: The ultravisor call to make. +* @retbuf: Buffer to store up to 4 return arguments in. +* +* This call supports up to 6 arguments and 4 return arguments. Use +* UCALL_BUFSIZE to size the return argument buffer. +*/ +#if defined(CONFIG_PPC_POWERNV )#ifdef
Quoted 11 lines
arch/powerpc/include/asm/ultravisor.h +5
-0+long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#endif + +#endif /* !__ASSEMBLY__ */ +#endif /* _ASM_POWERPC_ULTRAVISOR_H */ arch/powerpc/kernel/Makefile +0
-0arch/powerpc/kernel/Makefile endif · line 157
Quoted 5 lines
arch/powerpc/kernel/Makefile +1
-1@@ -154,7 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o -obj-$(CONFIG_PPC_POWERNV) += ultravisor.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o ucall.oSame comment about being platforms/powernv ?
Quoted 7 lines
arch/powerpc/kernel/ucall.S new file +1
-0@@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: GPL-2.0 */arch/powerpc/kernel/ucall.S line 11
Quoted 10 lines
arch/powerpc/kernel/ucall.S new file +10
-0+/* + * Generic code to perform an ultravisor call. + * + * Copyright 2019, IBM Corporation. + * + */ +#include <asm/ppc_asm.h> + +/* + * This function is based on the plpar_hcall()I don't think it meaningfully is any more.
arch/powerpc/kernel/ucall.S line 13
Quoted 2 lines
arch/powerpc/kernel/ucall.S new file +2
-0+ */ +_GLOBAL_TOC(ucall)You don't need the TOC setup here (unless a later patch does?).
arch/powerpc/kernel/ucall.S line 20
Quoted 7 lines
arch/powerpc/kernel/ucall.S new file +7
-0+ std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10Below you space the arguments, here you don't. Pick one or the other please.
Quoted 20 lines
arch/powerpc/kernel/ucall.S new file +10
-0+ + sc 2 /* Invoke the ultravisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + blr /* Return r3 = status */arch/powerpc/kernel/ultravisor.c +1
-0@@ -8,10 +8,14 @@#include <linux/init.h> #include <linux/printk.h> #include <linux/string.h> +#include <linux/export.h>
Quoted 6 lines
arch/powerpc/kernel/ultravisor.c line 17
Quoted 5 lines
arch/powerpc/kernel/ultravisor.c +2
-0#include <asm/ultravisor.h> #include <asm/firmware.h> +/* in ucall.S */ +EXPORT_SYMBOL_GPL (ucall);This should be in
ucall.S
Here I'm following the same way that hypercall wrapper symbols are exported.
Last time I tried to export that in ucall.S the linker complained that the ucall
symbol will not be versioned. Something like this:
https://patchwork.kernel.org/patch/9452759/
Thanks,
Claudio
Quoted 2 lines
cheers
Michael Ellerman
reviewerRe: [PATCH v4 3/8] KVM: PPC: Ultravisor: Add generic ultravisor call handler
Quoted 8 lines · Claudio Carvalho
Quoted 7 lines · Michael Ellerman
Quoted 6 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Add the
ucall()function, which can be used to make ultravisor calls with varied number of in and out arguments. Ultravisor calls can be made from the host or guests.
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
This copies the implementation of
plpar_hcall()... with quite a few changes?
This is one of the things I'd like to see in a Documentation file, so that people can review the implementation vs the specification.
I will document this (and other things) in a file under
Documentation/powerpc.
Thanks.
Quoted 5 lines
Quoted 5 lines
arch/powerpc/include/asm/ultravisor.h line 20
Quoted 2 lines
arch/powerpc/include/asm/ultravisor.h +2
-0+/* API functions */ +#define UCALL_BUFSIZE 4Please don't copy this design from the hcall code, it has led to bugs in the past.
Quoted 8 lines
Quoted 5 lines
See my (still unmerged) attempt to fix this for the hcall case:
https://patchwork.ozlabs.org/patch/683577/Basically instead of asking callers nicely to define a certain sized buffer, and them forgetting, define a proper type that has the right size.
I will keep that in mind. For now I think we don't need that since the v5 will have
ucall_norets()instead.
OK.
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
arch/powerpc/kernel/ultravisor.c +1
-0@@ -8,10 +8,14 @@#include <linux/init.h> #include <linux/printk.h> #include <linux/string.h> +#include <linux/export.h>
Quoted 8 lines
Quoted 6 lines
arch/powerpc/kernel/ultravisor.c line 17
Quoted 5 lines
arch/powerpc/kernel/ultravisor.c +2
-0#include <asm/ultravisor.h> #include <asm/firmware.h> +/* in ucall.S */ +EXPORT_SYMBOL_GPL (ucall);This should be in
ucall.SHere I'm following the same way that hypercall wrapper symbols are exported.
Yeah we used to not be able to export from .S but that was fixed a few years ago.
Quoted 4 lines
Last time I tried to export that in
ucall.Sthe linker complained that the ucall
symbol will not be versioned. Something like this:
https://patchwork.kernel.org/patch/9452759/
I think you just need to include <asm/export.h> in the .S ?
cheers
Nicholas Piggin
reviewerRe: [PATCH v4 3/8] KVM: PPC: Ultravisor: Add generic ultravisor call handler
In reply to Claudio CarvalhoClaudio Carvalho's on June 29, 2019 6:08 am:
Quoted 102 lines
From: Ram Pai <[email protected]>
Add the
ucall()function, which can be used to make ultravisor calls with varied number of in and out arguments. Ultravisor calls can be made from the host or guests.This copies the implementation of
plpar_hcall().Signed-off-by Ram Pai <[email protected]>[ Change
ucall.Sto not save CR, rename and move headers, builducall.S
ifCONFIG_PPC_POWERNVset, use R3 for the ucall number and add some comments in the code ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/ultravisor-api.h | 20 +++++++++++++++ arch/powerpc/include/asm/ultravisor.h | 20 +++++++++++++++ arch/powerpc/kernel/Makefile | 2 +- arch/powerpc/kernel/ucall.S | 30 +++++++++++++++++++++++ arch/powerpc/kernel/ultravisor.c | 4 +++ 5 files changed, 75 insertions(+), 1 deletion(-) create mode 100644 arch/powerpc/include/asm/ultravisor-api.h create mode 100644 arch/powerpc/kernel/ucall.Sarch/powerpc/include/asm/ultravisor-api.h new file +20
-0@@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Ultravisor API. +* +* Copyright 2019, IBM Corporation. +* +*/ +#ifndef _ASM_POWERPC_ULTRAVISOR_API_H +#define _ASM_POWERPC_ULTRAVISOR_API_H + +#include <asm/hvcall.h> + +/* Return codes */ +#define U_NOT_AVAILABLE H_NOT_AVAILABLE +#define U_SUCCESS H_SUCCESS +#define U_FUNCTION H_FUNCTION +#define U_PARAMETER H_PARAMETER + +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ +arch/powerpc/include/asm/ultravisor.h +20
-0@@ -8,8 +8,28 @@#ifndef _ASM_POWERPC_ULTRAVISOR_H#define _ASM_POWERPC_ULTRAVISOR_H +#include <asm/ultravisor-api.h> + +#if !defined(__ASSEMBLY__) +/* Internal functions */ extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data); +/* API functions */ +#define UCALL_BUFSIZE 4 +/** +* ucall: Make a powerpc ultravisor call. +* @opcode: The ultravisor call to make. +* @retbuf: Buffer to store up to 4 return arguments in. +* +* This call supports up to 6 arguments and 4 return arguments. Use +* UCALL_BUFSIZE to size the return argument buffer. +*/ +#if defined(CONFIG_PPC_POWERNV ) +long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#endif + +#endif /* !__ASSEMBLY__ */ +#endif /* _ASM_POWERPC_ULTRAVISOR_H */ arch/powerpc/kernel/Makefile +1
-1@@ -154,7 +154,7 @@ endif obj-$(CONFIG_EPAPR_PARAVIRT) += epapr_paravirt.o epapr_hcalls.o obj-$(CONFIG_KVM_GUEST) += kvm.o kvm_emul.o -obj-$(CONFIG_PPC_POWERNV) += ultravisor.o +obj-$(CONFIG_PPC_POWERNV) += ultravisor.o ucall.o # Disable GCOV, KCOV & sanitizers in odd or sensitive code GCOV_PROFILE_prom_init.o := narch/powerpc/kernel/ucall.S new file +1
-0@@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: GPL-2.0 */
arch/powerpc/kernel/ucall.S line 11
Quoted 10 lines
arch/powerpc/kernel/ucall.S new file +10
-0+/* + * Generic code to perform an ultravisor call. + * + * Copyright 2019, IBM Corporation. + * + */ +#include <asm/ppc_asm.h> + +/* + * This function is based on the plpar_hcall()
A better comment might be to specify calling convention.
arch/powerpc/kernel/ucall.S line 20
Quoted 9 lines
arch/powerpc/kernel/ucall.S new file +9
-0+ */ +_GLOBAL_TOC(ucall) + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10
I guess this is fine, is there any particular reason that you made the call convention this way rather than leaving regs in place and r4 is a scratch?
Quoted 9 lines
arch/powerpc/kernel/ucall.S new file +9
-0+ + sc 2 /* Invoke the ultravisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) +
PATCH v4 4/8 KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
In reply to Claudio CarvalhoFrom: Michael Anderson <[email protected]>
When running under an ultravisor, the ultravisor controls the real partition table and has it in secure memory where the hypervisor can't access it, and therefore we (the HV) have to do a ucall whenever we want to update an entry.
The HV still keeps a copy of its view of the partition table in normal memory so that the nest MMU can access it.
Both partition tables will have PATE entries for HV and normal virtual machines.
--- arch/powerpc/include/asm/ultravisor-api.h | 5 +++- arch/powerpc/include/asm/ultravisor.h | 14 ++++++++++ arch/powerpc/mm/book3s64/hash_utils.c | 3 +- arch/powerpc/mm/book3s64/pgtable.c | 34 +++++++++++++++++++++-- arch/powerpc/mm/book3s64/radix_pgtable.c | 9 ++++-- 5 files changed, 57 insertions(+), 8 deletions(-)
arch/powerpc/include/asm/ultravisor-api.h +4 -1
@@ -15,6 +15,9 @@#define U_SUCCESS H_SUCCESS#define U_FUNCTION H_FUNCTION#define U_PARAMETER H_PARAMETER +#define U_PERMISSION H_PERMISSION -#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ +/* opcodes */ +#define UV_WRITE_PATE 0xF104 +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */
arch/powerpc/include/asm/ultravisor.h +14 -0
@@ -12,6 +12,8 @@#if !defined(__ASSEMBLY__) +#include <linux/types.h> +/* Internal functions */ extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data);
@@ -28,8 +30,20 @@ extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname,*/ #if defined(CONFIG_PPC_POWERNV )long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#else +static long ucall (unsigned long opcode,unsigned long *retbuf, ...) +{ +return U_NOT_AVAILABLE ; +}#endif +static inline int uv_register_pate (u64 lpid,u64 dw0,u64 dw1) +{ +unsigned long retbuf[UCALL_BUFSIZE ]; + +return ucall (UV_WRITE_PATE , retbuf, lpid, dw0, dw1); +} +#endif /* !__ASSEMBLY__ */ #endif /* _ASM_POWERPC_ULTRAVISOR_H */
arch/powerpc/mm/book3s64/hash_utils.c +2 -1
@@ -1080,9 +1080,10 @@ void hash__early_init_mmu_secondary(void)if (!cpu_has_feature (CPU_FTR_ARCH_300 ))mtspr (SPRN_SDR1 , _SDR1); -else +else if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR ))mtspr (SPRN_PTCR ,__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); + }/* Initialize SLB */ slb_initialize ();
arch/powerpc/mm/book3s64/pgtable.c +31 -3
@@ -16,6 +16,8 @@#include <asm/tlb.h> #include <asm/trace.h> #include <asm/powernv.h> +#include <asm/firmware.h> +#include <asm/ultravisor.h> #include <mm/mmu_decl.h> #include <trace/events/thp.h>
@@ -206,12 +208,25 @@ void __init mmu_partition_table_init(void)* 64 K size. */ ptcr =__pa (partition_tb) | (PATB_SIZE_SHIFT -12 ); -mtspr (SPRN_PTCR , ptcr); +/* +* If ultravisor is available, it is responsible for creating and +* managing partition table +*/ +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR , ptcr); + +/* +* Since nestMMU cannot access secure memory. Create +* and manage our own partition table. This table +* contains entries for nonsecure and hypervisor +* partition. +*/ powernv_set_nmmu_ptcr (ptcr); } -void mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0, -unsigned long dw1) +static void __mmu_partition_table_set_entry (unsigned int lpid, +unsigned long dw0, +unsigned long dw1) {unsigned long old =be64_to_cpu (partition_tb[lpid].patb0);
@@ -238,6 +253,19 @@ void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,/* do we need fixup here ?*/ asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); } + +void mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0, +unsigned long dw1) +{ +__mmu_partition_table_set_entry (lpid, dw0, dw1); + +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) { +uv_register_pate (lpid, dw0, dw1); +pr_info ("PATE registered by ultravisor: dw0 = 0x%lx, dw1 = 0x%lx\n" , + dw0, dw1); + } +} +EXPORT_SYMBOL_GPL (mmu_partition_table_set_entry);static pmd_t *get_pmd_from_cache (struct mm_struct *mm)
arch/powerpc/mm/book3s64/radix_pgtable.c +6 -3
@@ -656,8 +656,10 @@ void radix__early_init_mmu_secondary(void) lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr |LPCR_UPRT |LPCR_HR ); -mtspr (SPRN_PTCR , -__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,__pa (partition_tb) | + (PATB_SIZE_SHIFT -12 )); +radix_init_amor (); }
@@ -673,7 +675,8 @@ void radix__mmu_cleanup_all(void)if (!firmware_has_feature (FW_FEATURE_LPAR )) { lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr & ~LPCR_UPRT ); -mtspr (SPRN_PTCR ,0 ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,0 );powernv_set_nmmu_ptcr (0 );radix__flush_tlb_all (); }
Signature · 1 line
2.20.1
Re: [PATCH v4 4/8] KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
Quoted 14 lines · Claudio Carvalho
From: Michael Anderson <[email protected]>
When running under an ultravisor, the ultravisor controls the real partition table and has it in secure memory where the hypervisor can't access it, and therefore we (the HV) have to do a ucall whenever we want
to update an entry.The HV still keeps a copy of its view of the partition table in normal memory so that the nest MMU can access it.
Both partition tables will have PATE entries for HV and normal virtual machines.
Quoted 6 lines
Suggested-by Ryan Grimm <[email protected]>Signed-off-by Michael Anderson <[email protected]>Signed-off-by Madhavan Srinivasan <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Write the pate in HV's table before doing that in UV's ]Signed-off-by Claudio Carvalho <[email protected]>
Quoted 172 lines
--- arch/powerpc/include/asm/ultravisor-api.h | 5 +++- arch/powerpc/include/asm/ultravisor.h | 14 ++++++++++ arch/powerpc/mm/book3s64/hash_utils.c | 3 +- arch/powerpc/mm/book3s64/pgtable.c | 34 +++++++++++++++++++++-- arch/powerpc/mm/book3s64/radix_pgtable.c | 9 ++++-- 5 files changed, 57 insertions(+), 8 deletions(-)arch/powerpc/include/asm/ultravisor-api.h +4
-1@@ -15,6 +15,9 @@#define U_SUCCESS H_SUCCESS#define U_FUNCTION H_FUNCTION#define U_PARAMETER H_PARAMETER +#define U_PERMISSION H_PERMISSION -#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ +/* opcodes */ +#define UV_WRITE_PATE 0xF104 +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ arch/powerpc/include/asm/ultravisor.h +2
-0@@ -12,6 +12,8 @@#if !defined(__ASSEMBLY__) +#include <linux/types.h> +/* Internal functions */ extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname,arch/powerpc/include/asm/ultravisor.h +12
-0int depth, void *data);@@ -28,8 +30,20 @@ extern int early_init_dt_scan_ultravisor(unsignedlong node, const char *uname, */ #if defined(CONFIG_PPC_POWERNV )long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#else +static long ucall (unsigned long opcode,unsigned long *retbuf, ...) +{ +return U_NOT_AVAILABLE ; +}#endif +static inline int uv_register_pate (u64 lpid,u64 dw0,u64 dw1) +{ +unsigned long retbuf[UCALL_BUFSIZE ]; + +return ucall (UV_WRITE_PATE , retbuf, lpid, dw0, dw1); +} +#endif /* !__ASSEMBLY__ */ #endif /*
_ASM_POWERPC_ULTRAVISOR_H*/arch/powerpc/mm/book3s64/hash_utils.c +2
-1@@ -1080,9 +1080,10 @@ void hash__early_init_mmu_secondary(void)if (!cpu_has_feature (CPU_FTR_ARCH_300 ))mtspr (SPRN_SDR1 , _SDR1); -else +else if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR ))mtspr (SPRN_PTCR ,__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); + }/* Initialize SLB */ slb_initialize ();arch/powerpc/mm/book3s64/pgtable.c +31
-3@@ -16,6 +16,8 @@#include <asm/tlb.h> #include <asm/trace.h> #include <asm/powernv.h> +#include <asm/firmware.h> +#include <asm/ultravisor.h> #include <mm/mmu_decl.h> #include <trace/events/thp.h> @@ -206,12 +208,25 @@ void __init mmu_partition_table_init(void)* 64 K size. */ ptcr =__pa (partition_tb) | (PATB_SIZE_SHIFT -12 ); -mtspr (SPRN_PTCR , ptcr); +/* +* If ultravisor is available, it is responsible for creating and +* managing partition table +*/ +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR , ptcr); + +/* +* Since nestMMU cannot access secure memory. Create +* and manage our own partition table. This table +* contains entries for nonsecure and hypervisor +* partition. +*/ powernv_set_nmmu_ptcr (ptcr); } -void mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0, -unsigned long dw1) +static void __mmu_partition_table_set_entry(unsigned int lpid, +unsigned long dw0, +unsigned long dw1) {unsigned long old =be64_to_cpu (partition_tb[lpid].patb0);@@ -238,6 +253,19 @@ void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,/* do we need fixup here ?*/ asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); } + +void mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0, +unsigned long dw1) +{ +__mmu_partition_table_set_entry (lpid, dw0, dw1); + +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) { +uv_register_pate (lpid, dw0, dw1); +pr_info ("PATE registered by ultravisor: dw0 = 0x%lx, dw1 = 0x%lx\n" , + dw0, dw1); + } +} +EXPORT_SYMBOL_GPL (mmu_partition_table_set_entry);static
pmd_t*get_pmd_from_cache(structmm_struct*mm)arch/powerpc/mm/book3s64/radix_pgtable.c +6
-3@@ -656,8 +656,10 @@ void radix__early_init_mmu_secondary(void) lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr |LPCR_UPRT |LPCR_HR ); -mtspr (SPRN_PTCR , -__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,__pa (partition_tb) | + (PATB_SIZE_SHIFT -12 )); +radix_init_amor (); }@@ -673,7 +675,8 @@ void radix__mmu_cleanup_all(void)if (!firmware_has_feature (FW_FEATURE_LPAR )) { lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr & ~LPCR_UPRT ); -mtspr (SPRN_PTCR ,0 ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,0 );powernv_set_nmmu_ptcr (0 );radix__flush_tlb_all (); }
Michael Ellerman
reviewerRe: [PATCH v4 4/8] KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
In reply to Claudio CarvalhoQuoted 11 lines · Claudio Carvalho
From: Michael Anderson <[email protected]>
When running under an ultravisor, the ultravisor controls the real partition table and has it in secure memory where the hypervisor can't access it, and therefore we (the HV) have to do a ucall whenever we want to update an entry.
The HV still keeps a copy of its view of the partition table in normal memory so that the nest MMU can access it.
Both partition tables will have PATE entries for HV and normal virtual
Can you expand novel acronyms on their first usage please, ie. PATE.
Quoted 3 lines
machines.
Suggested-by Ryan Grimm <[email protected]>
"Suggested" implies this is optional, but it's not :)
I'm not sure exactly what Ryan's input was, but feel free to make up a better tag :)
Quoted 18 lines
Signed-off-by Michael Anderson <[email protected]>Signed-off-by Madhavan Srinivasan <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Write the pate in HV's table before doing that in UV's ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/ultravisor-api.h | 5 +++- arch/powerpc/include/asm/ultravisor.h | 14 ++++++++++ arch/powerpc/mm/book3s64/hash_utils.c | 3 +- arch/powerpc/mm/book3s64/pgtable.c | 34 +++++++++++++++++++++-- arch/powerpc/mm/book3s64/radix_pgtable.c | 9 ++++-- 5 files changed, 57 insertions(+), 8 deletions(-)arch/powerpc/include/asm/ultravisor-api.h +0
-0
arch/powerpc/include/asm/ultravisor-api.h line 19
Quoted 6 lines
arch/powerpc/include/asm/ultravisor-api.h +1
-1@@ -15,6 +15,9 @@#define U_SUCCESS H_SUCCESS#define U_FUNCTION H_FUNCTION#define U_PARAMETER H_PARAMETER +#define U_PERMISSION H_PERMISSION -#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */
What happened there? Just a diff artifact?
Quoted 18 lines
arch/powerpc/include/asm/ultravisor-api.h +3
-0+/* opcodes */ +#define UV_WRITE_PATE 0xF104 +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ arch/powerpc/include/asm/ultravisor.h +2
-0@@ -12,6 +12,8 @@#if !defined(__ASSEMBLY__) +#include <linux/types.h> +/* Internal functions */ extern int early_init_dt_scan_ultravisor (unsigned long node,const char *uname,int depth,void *data);
arch/powerpc/include/asm/ultravisor.h extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname, · line 34
Quoted 5 lines
arch/powerpc/include/asm/ultravisor.h +2
-0@@ -28,8 +30,20 @@ extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname,*/ #if defined(CONFIG_PPC_POWERNV )long ucall (unsigned long opcode,unsigned long *retbuf, ...); +#else +static long ucall (unsigned long opcode,unsigned long *retbuf, ...)
^
inlinearch/powerpc/include/asm/ultravisor.h extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname, · line 38
Quoted 4 lines
arch/powerpc/include/asm/ultravisor.h +3
-0+{ +return U_NOT_AVAILABLE ; +}#endif
That should have been in the previous patch.
arch/powerpc/include/asm/ultravisor.h extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname, · line 44
Quoted 5 lines
arch/powerpc/include/asm/ultravisor.h +5
-0+static inline int uv_register_pate (u64 lpid,u64 dw0,u64 dw1) +{ +unsigned long retbuf[UCALL_BUFSIZE ]; + +return ucall (UV_WRITE_PATE , retbuf, lpid, dw0, dw1);
You probably want a ucall_norets().
Quoted 10 lines
arch/powerpc/include/asm/ultravisor.h +2
-0+} +#endif /* !__ASSEMBLY__ */ #endif /* _ASM_POWERPC_ULTRAVISOR_H */ arch/powerpc/mm/book3s64/hash_utils.c +0
-0
arch/powerpc/mm/book3s64/hash_utils.c void hash__early_init_mmu_secondary(void) · line 1085
Quoted 7 lines
arch/powerpc/mm/book3s64/hash_utils.c +1
-1@@ -1080,9 +1080,10 @@ void hash__early_init_mmu_secondary(void)if (!cpu_has_feature (CPU_FTR_ARCH_300 ))mtspr (SPRN_SDR1 , _SDR1); -else +else if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR ))mtspr (SPRN_PTCR ,__pa (partition_tb) | (PATB_SIZE_SHIFT -12 ));
Needs a comment for the else case.
Quoted 17 lines
arch/powerpc/mm/book3s64/hash_utils.c +0
-0}/* Initialize SLB */ slb_initialize ();arch/powerpc/mm/book3s64/pgtable.c +2
-0@@ -16,6 +16,8 @@#include <asm/tlb.h> #include <asm/trace.h> #include <asm/powernv.h> +#include <asm/firmware.h> +#include <asm/ultravisor.h> #include <mm/mmu_decl.h> #include <trace/events/thp.h>
arch/powerpc/mm/book3s64/pgtable.c void __init mmu_partition_table_init(void) · line 214
Quoted 8 lines
arch/powerpc/mm/book3s64/pgtable.c +4
-1@@ -206,12 +208,25 @@ void __init mmu_partition_table_init(void)* 64 K size. */ ptcr =__pa (partition_tb) | (PATB_SIZE_SHIFT -12 ); -mtspr (SPRN_PTCR , ptcr); +/* +* If ultravisor is available, it is responsible for creating and +* managing partition table +*/
But we just created the partition table?!
This comment and the one below would probably make more sense if they were merged and appeared further up, where we allocate the partition table.
arch/powerpc/mm/book3s64/pgtable.c void __init mmu_partition_table_init(void) · line 220
Quoted 6 lines
arch/powerpc/mm/book3s64/pgtable.c +6
-0+if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR , ptcr); + +/* +* Since nestMMU cannot access secure memory. Create +* and manage our own partition table. This table
But we just said the UV was responsible for it :)
Quoted 20 lines
arch/powerpc/mm/book3s64/pgtable.c +8
-2+* contains entries for nonsecure and hypervisor +* partition. +*/ powernv_set_nmmu_ptcr (ptcr); } -void mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0, -unsigned long dw1) +static void __mmu_partition_table_set_entry (unsigned int lpid, +unsigned long dw0, +unsigned long dw1) {unsigned long old =be64_to_cpu (partition_tb[lpid].patb0);@@ -238,6 +253,19 @@ void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,/* do we need fixup here ?*/ asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); } + +void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,
arch/powerpc/mm/book3s64/pgtable.c void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0, · line 267
Quoted 10 lines
arch/powerpc/mm/book3s64/pgtable.c +10
-0+ unsigned long dw1) +{ +__mmu_partition_table_set_entry (lpid, dw0, dw1); + +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) { +uv_register_pate (lpid, dw0, dw1); +pr_info ("PATE registered by ultravisor: dw0 = 0x%lx, dw1 = 0x%lx\n" , + dw0, dw1); + } +}
I agree with Alexey that this patch and the next should be merged and the result cleaned up a bit.
arch/powerpc/mm/book3s64/pgtable.c void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0, · line 268
Quoted 1 line
arch/powerpc/mm/book3s64/pgtable.c +1
-0+
No extra blank please.
Quoted 22 lines
arch/powerpc/mm/book3s64/pgtable.c +0
-0EXPORT_SYMBOL_GPL (mmu_partition_table_set_entry);static
pmd_t*get_pmd_from_cache(structmm_struct*mm)arch/powerpc/mm/book3s64/radix_pgtable.c +4
-2@@ -656,8 +656,10 @@ void radix__early_init_mmu_secondary(void) lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr |LPCR_UPRT |LPCR_HR ); -mtspr (SPRN_PTCR , -__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,__pa (partition_tb) | + (PATB_SIZE_SHIFT -12 )); +radix_init_amor (); }
arch/powerpc/mm/book3s64/radix_pgtable.c void radix__mmu_cleanup_all(void) · line 682
Quoted 9 lines
arch/powerpc/mm/book3s64/radix_pgtable.c +2
-1@@ -673,7 +675,8 @@ void radix__mmu_cleanup_all(void)if (!firmware_has_feature (FW_FEATURE_LPAR )) { lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr & ~LPCR_UPRT ); -mtspr (SPRN_PTCR ,0 ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,0 );powernv_set_nmmu_ptcr (0 );radix__flush_tlb_all (); }
There's four of these case where we skip touching the PTCR, which is right on the borderline of warranting an accessor. I guess we can do it as a cleanup later.
cheers
Ryan Grimm
Reviewed-byRe: [PATCH v4 4/8] KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
Quoted 10 lines · Michael Ellerman
Quoted 9 lines · Claudio Carvalho
From: Michael Anderson <[email protected]>
When running under an ultravisor, the ultravisor controls the real partition table and has it in secure memory where the hypervisor can't
access it, and therefore we (the HV) have to do a ucall whenever we want
to update an entry.
Quoted 9 lines
Quoted 6 lines
The HV still keeps a copy of its view of the partition table in normal
memory so that the nest MMU can access it.Both partition tables will have PATE entries for HV and normal virtual
Can you expand novel acronyms on their first usage please, ie. PATE.
Agreed. This confused me a while ago. It is "Partition Table Entry", correct?
Quoted 4 lines
Quoted 3 lines
machines.
Suggested-by Ryan Grimm <[email protected]>
Please remove this and add
Quoted 228 lines
"Suggested" implies this is optional, but it's not :)
I'm not sure exactly what Ryan's input was, but feel free to make up a
better tag :)Quoted 20 lines
Signed-off-by Michael Anderson <[email protected]>Signed-off-by Madhavan Srinivasan <[email protected]>Signed-off-by Ram Pai <[email protected]>[ Write the pate in HV's table before doing that in UV's ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/ultravisor-api.h | 5 +++- arch/powerpc/include/asm/ultravisor.h | 14 ++++++++++ arch/powerpc/mm/book3s64/hash_utils.c | 3 +- arch/powerpc/mm/book3s64/pgtable.c | 34 +++++++++++++++++++++-- arch/powerpc/mm/book3s64/radix_pgtable.c | 9 ++++-- 5 files changed, 57 insertions(+), 8 deletions(-)arch/powerpc/include/asm/ultravisor-api.h +0
-0arch/powerpc/include/asm/ultravisor-api.h line 19
Quoted 6 lines
arch/powerpc/include/asm/ultravisor-api.h +1
-1@@ -15,6 +15,9 @@#define U_SUCCESS H_SUCCESS#define U_FUNCTION H_FUNCTION#define U_PARAMETER H_PARAMETER +#define U_PERMISSION H_PERMISSION -#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ What happened there? Just a diff artifact?
Quoted 27 lines
arch/powerpc/include/asm/ultravisor-api.h +3
-0+/* opcodes */ +#define UV_WRITE_PATE 0xF104 +#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ arch/powerpc/include/asm/ultravisor.h +2
-0@@ -12,6 +12,8 @@#if !defined(__ASSEMBLY__) +#include <linux/types.h> +/* Internal functions */ extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname,arch/powerpc/include/asm/ultravisor.h +0
-0int depth, void *data);@@ -28,8 +30,20 @@ extern intearly_init_dt_scan_ultravisor(unsigned long node, const char *uname,
*/
#ifdefined(CONFIG_PPC_POWERNV)
longucall(unsigned long opcode, unsigned long *retbuf, ...);arch/powerpc/include/asm/ultravisor.h +2
-0+#else +static long ucall(unsigned long opcode, unsigned long *retbuf,Quoted 1 line
...)
^ inlinearch/powerpc/include/asm/ultravisor.h extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname, · line 38
Quoted 4 lines
arch/powerpc/include/asm/ultravisor.h +3
-0+{ +return U_NOT_AVAILABLE ; +}#endif That should have been in the previous patch.
arch/powerpc/include/asm/ultravisor.h extern int early_init_dt_scan_ultravisor(unsigned long node, const char *uname, · line 44
Quoted 5 lines
arch/powerpc/include/asm/ultravisor.h +5
-0+static inline int uv_register_pate (u64 lpid,u64 dw0,u64 dw1) +{ +unsigned long retbuf[UCALL_BUFSIZE ]; + +return ucall (UV_WRITE_PATE , retbuf, lpid, dw0, dw1);You probably want a
ucall_norets().Quoted 11 lines
arch/powerpc/include/asm/ultravisor.h +2
-0+} +#endif /* !__ASSEMBLY__ */ #endif /* _ASM_POWERPC_ULTRAVISOR_H */ arch/powerpc/mm/book3s64/hash_utils.c +0
-0arch/powerpc/mm/book3s64/hash_utils.c void hash__early_init_mmu_secondary(void) · line 1086
Quoted 8 lines
arch/powerpc/mm/book3s64/hash_utils.c +1
-1@@ -1080,9 +1080,10 @@ void hash__early_init_mmu_secondary(void)if (!cpu_has_feature (CPU_FTR_ARCH_300 ))mtspr (SPRN_SDR1 , _SDR1); -else +else if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR ))mtspr (SPRN_PTCR ,__pa (partition_tb) | (PATB_SIZE_SHIFT -12 ));Needs a comment for the else case.
Quoted 25 lines
arch/powerpc/mm/book3s64/hash_utils.c +0
-0}/* Initialize SLB */ slb_initialize ();arch/powerpc/mm/book3s64/pgtable.c +4
-1@@ -16,6 +16,8 @@#include <asm/tlb.h> #include <asm/trace.h> #include <asm/powernv.h> +#include <asm/firmware.h> +#include <asm/ultravisor.h> #include <mm/mmu_decl.h> #include <trace/events/thp.h> @@ -206,12 +208,25 @@ void __init mmu_partition_table_init(void)* 64 K size. */ ptcr =__pa (partition_tb) | (PATB_SIZE_SHIFT -12 ); -mtspr (SPRN_PTCR , ptcr); +/* +* If ultravisor is available, it is responsible for creating and
arch/powerpc/mm/book3s64/pgtable.c
Quoted 2 lines
arch/powerpc/mm/book3s64/pgtable.c +2
-0+* managing partition table +*/ But we just created the partition table?!
This comment and the one below would probably make more sense if they were merged and appeared further up, where we allocate the partition table.
arch/powerpc/mm/book3s64/pgtable.c void __init mmu_partition_table_init(void) · line 220
Quoted 6 lines
arch/powerpc/mm/book3s64/pgtable.c +6
-0+if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR , ptcr); + +/* +* Since nestMMU cannot access secure memory. Create +* and manage our own partition table. This table But we just said the UV was responsible for it :)
Quoted 24 lines
arch/powerpc/mm/book3s64/pgtable.c +3
-1+* contains entries for nonsecure and hypervisor +* partition. +*/ powernv_set_nmmu_ptcr (ptcr); } -void mmu_partition_table_set_entry(unsigned int lpid, unsignedlong dw0,
arch/powerpc/mm/book3s64/pgtable.c +6
-1- unsigned long dw1) +static void __mmu_partition_table_set_entry (unsigned int lpid, +unsigned long dw0, +unsigned long dw1) {unsigned long old =be64_to_cpu (partition_tb[lpid].patb0);@@ -238,6 +253,19 @@ void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0, /* do we need fixup here ?*/ asm volatile("eieio; tlbsync; ptesync" : : : "memory"); } + +void mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0, +unsigned long dw1)arch/powerpc/mm/book3s64/pgtable.c void mmu_partition_table_set_entry(unsigned · line 270
Quoted 10 lines
arch/powerpc/mm/book3s64/pgtable.c +9
-0+{ +__mmu_partition_table_set_entry (lpid, dw0, dw1); + +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) { +uv_register_pate (lpid, dw0, dw1); + pr_info("PATE registered by ultravisor: dw0 = 0x%lx, dw1 = 0x%lx\n", + dw0, dw1); + } +}I agree with Alexey that this patch and the next should be merged and the result cleaned up a bit.
arch/powerpc/mm/book3s64/pgtable.c void mmu_partition_table_set_entry(unsigned · line 271
Quoted 1 line
arch/powerpc/mm/book3s64/pgtable.c +1
-0+No extra blank please.
Quoted 22 lines
EXPORT_SYMBOL_GPL(mmu_partition_table_set_entry);static
pmd_t*get_pmd_from_cache(structmm_struct*mm)arch/powerpc/mm/book3s64/radix_pgtable.c +4
-2@@ -656,8 +656,10 @@ void radix__early_init_mmu_secondary(void) lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr |LPCR_UPRT |LPCR_HR ); -mtspr (SPRN_PTCR , -__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,__pa (partition_tb) | + (PATB_SIZE_SHIFT -12 )); +radix_init_amor (); }arch/powerpc/mm/book3s64/radix_pgtable.c void radix__mmu_cleanup_all(void) · line 682
Quoted 9 lines
arch/powerpc/mm/book3s64/radix_pgtable.c +2
-1@@ -673,7 +675,8 @@ void radix__mmu_cleanup_all(void)if (!firmware_has_feature (FW_FEATURE_LPAR )) { lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr & ~LPCR_UPRT ); -mtspr (SPRN_PTCR ,0 ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,0 );powernv_set_nmmu_ptcr (0 );radix__flush_tlb_all (); }There's four of these case where we skip touching the PTCR, which is right on the borderline of warranting an accessor. I guess we can do it
as a cleanup later.cheers
Re: [PATCH v4 4/8] KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
In reply to Michael EllermanQuoted 20 lines · Michael Ellerman
Quoted 20 lines
static
pmd_t*get_pmd_from_cache(structmm_struct*mm)arch/powerpc/mm/book3s64/radix_pgtable.c +4
-2@@ -656,8 +656,10 @@ void radix__early_init_mmu_secondary(void) lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr |LPCR_UPRT |LPCR_HR ); -mtspr (SPRN_PTCR , -__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,__pa (partition_tb) | + (PATB_SIZE_SHIFT -12 )); +radix_init_amor (); }
Quoted 13 lines
arch/powerpc/mm/book3s64/radix_pgtable.c void radix__mmu_cleanup_all(void) · line 682
Quoted 10 lines
arch/powerpc/mm/book3s64/radix_pgtable.c +2
-1@@ -673,7 +675,8 @@ void radix__mmu_cleanup_all(void)if (!firmware_has_feature (FW_FEATURE_LPAR )) { lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr & ~LPCR_UPRT ); -mtspr (SPRN_PTCR ,0 ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,0 );powernv_set_nmmu_ptcr (0 );radix__flush_tlb_all (); }There's four of these case where we skip touching the PTCR, which is right on the borderline of warranting an accessor. I guess we can do it as a cleanup later.
I agree.
Since the kernel doesn't need to access a big number of ultravisor privileged registers, maybe we can define mtspr_<reg> and mfspr_<reg> inline functions that in ultravisor.h that skip touching the register if an ultravisor is present and and the register is ultravisor privileged. Thus, we don't need to replicate comments and that also would make it easier for developers to know what are the ultravisor privileged registers.
Something like this:
arch/powerpc/include/asm/ultravisor.h +0 -0
@@ -10,10 +10,21 @@
#include <asm/ultravisor-api.h>
#include <asm/asm-prototypes.h>
+#include <asm/reg.h>
int early_init_dt_scan_ultravisor(unsigned long node, const char *uname, int depth, void *data);
+static inline void mtspr_ptcr(unsigned long val)
+{
+ /*
+ * If the ultravisor firmware is present, it maintains the partition + * table. PTCR becomes ultravisor privileged only for writing. + */
+ if (!firmware_has_feature(FW_FEATURE_ULTRAVISOR))
+ mtspr(SPRN_PTCR, val);
+}
+
static inline int uv_register_pate(u64 lpid, u64 dw0, u64 dw1) {
return ucall_norets(UV_WRITE_PATE, lpid, dw0, dw1);
(quoted diff) +0 -0
b/arch/powerpc/mm/book3s64/pgtable.c
index e1bbc48e730f..25156f9dfde8 100644
arch/powerpc/mm/book3s64/pgtable.c +0 -0
@@ -220,7 +220,7 @@ void __init mmu_partition_table_init(void)
* 64 K size.
*/
ptcr = __pa(partition_tb) | (PATB_SIZE_SHIFT - 12); - mtspr(SPRN_PTCR, ptcr);
+ mtspr_ptcr(ptcr); powernv_set_nmmu_ptcr(ptcr);
}
What do you think?
An alternative could be to change the mtspr() and mfspr() macros as we proposed in the v1, but access to non-ultravisor privileged registers would be performance impacted because we always would need to check if the register is one of the few ultravisor registers that the kernel needs to access.
Thanks,
Claudio
Quoted 2 lines
cheers
Michael Ellerman
reviewerRe: [PATCH v4 4/8] KVM: PPC: Ultravisor: Use UV_WRITE_PATE ucall to register a PATE
Quoted 55 lines · Claudio Carvalho
Quoted 18 lines · Michael Ellerman
Quoted 18 lines
static
pmd_t*get_pmd_from_cache(structmm_struct*mm)arch/powerpc/mm/book3s64/radix_pgtable.c +4
-2@@ -656,8 +656,10 @@ void radix__early_init_mmu_secondary(void) lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr |LPCR_UPRT |LPCR_HR ); -mtspr (SPRN_PTCR , -__pa (partition_tb) | (PATB_SIZE_SHIFT -12 )); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,__pa (partition_tb) | + (PATB_SIZE_SHIFT -12 )); +radix_init_amor (); }Quoted 13 lines
arch/powerpc/mm/book3s64/radix_pgtable.c void radix__mmu_cleanup_all(void) · line 682
Quoted 10 lines
arch/powerpc/mm/book3s64/radix_pgtable.c +2
-1@@ -673,7 +675,8 @@ void radix__mmu_cleanup_all(void)if (!firmware_has_feature (FW_FEATURE_LPAR )) { lpcr =mfspr (SPRN_LPCR );mtspr (SPRN_LPCR , lpcr & ~LPCR_UPRT ); -mtspr (SPRN_PTCR ,0 ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_PTCR ,0 );powernv_set_nmmu_ptcr (0 );radix__flush_tlb_all (); }There's four of these case where we skip touching the PTCR, which is right on the borderline of warranting an accessor. I guess we can do it as a cleanup later.
I agree.
Since the kernel doesn't need to access a big number of ultravisor privileged registers, maybe we can define mtspr_<reg> and mfspr_<reg> inline functions that in
ultravisor.hthat skip touching the register if an ultravisor is present and and the register is ultravisor privileged. Thus, we don't need to replicate comments and that also would make it easier for developers to know what are the ultravisor privileged registers.Something like this:
arch/powerpc/include/asm/ultravisor.h +0
-0@@ -10,10 +10,21 @@
#include <asm/ultravisor-api.h>
#include <asm/asm-prototypes.h>
+#include <asm/reg.h>
intearly_init_dt_scan_ultravisor(unsigned long node, const char *uname, int depth, void *data);
arch/powerpc/include/asm/ultravisor.h
Quoted 9 lines
arch/powerpc/include/asm/ultravisor.h +9
-0+static inline void mtspr_ptcr (unsigned long val) +{ + /* + * If the ultravisor firmware is present, it maintains the partition + * table. PTCR becomes ultravisor privileged only for writing. + */ + if (!firmware_has_feature(FW_FEATURE_ULTRAVISOR)) + mtspr(SPRN_PTCR, val); +}
+
Quoted 17 lines
static inline int
uv_register_pate(u64 lpid, u64 dw0, u64 dw1){
returnucall_norets(UV_WRITE_PATE, lpid, dw0, dw1);arch/powerpc/mm/book3s64/pgtable.c +1
-1@@ -220,7 +220,7 @@ void __init mmu_partition_table_init(void)* 64 K size. */ ptcr = __pa(partition_tb) | (PATB_SIZE_SHIFT - 12); - mtspr(SPRN_PTCR, ptcr); + mtspr_ptcr(ptcr); powernv_set_nmmu_ptcr(ptcr); }
Quoted 1 line
What do you think?
I don't think that's actually clearer.
If the logic was always:
if (ultravisor)
do_ucall()
else
mtspr()Then a wrapper called eg. set_ptcr() would make sense.
But because in some cases you do a ucall and some you don't, I don't think it helps to hide that in an accessor like above.
That is confusing to a reader who sees all this code to setup a value and then the write to PTCR does nothing.
And in fact you didn't explain why it's OK for those cases to not do the write at all.
Quoted 5 lines
An alternative could be to change the
mtspr()andmfspr()macros as we proposed in the v1, but access to non-ultravisor privileged registers would be performance impacted because we always would need to check if the register is one of the few ultravisor registers that the kernel needs to access.
Yeah that and it would be very confusing to a reader who sees:
ptcr = ...;
mtspr(SPRN_PTCR, ptcr);
...And then they discover the mtspr does *nothing* when the Ultravisor is enabled.
cheers
PATCH v4 5/8 KVM: PPC: Ultravisor: Restrict flush of the partition tlb cache
In reply to Claudio CarvalhoFrom: Ram Pai <[email protected]>
Ultravisor is responsible for flushing the tlb cache, since it manages the PATE entries. Hence skip tlb flush, if the ultravisor firmware is available.
--- arch/powerpc/mm/book3s64/pgtable.c | 33 +++++++++++++++++------------- 1 file changed, 19 insertions(+), 14 deletions(-)
arch/powerpc/mm/book3s64/pgtable.c +19 -14
@@ -224,6 +224,23 @@ void __init mmu_partition_table_init(void)powernv_set_nmmu_ptcr (ptcr); } +static void flush_partition (unsigned int lpid,unsigned long dw0) +{ +if (dw0 &PATB_HR ) { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); + asm volatile(PPC_TLBIE_5(%0, %1, 2, 1, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); + }else { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 0) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); + } +/* do we need fixup here ?*/ + asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +} +static void __mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0,unsigned long dw1)
@@ -238,20 +255,8 @@ static void __mmu_partition_table_set_entry(unsigned int lpid,* The type of flush (hash or radix) depends on what the previous * use of this partition ID was, not the new use. */ - asmvolatile ("ptesync" : : :"memory" ); -if (old &PATB_HR ) { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,1) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); - asm volatile(PPC_TLBIE_5(%0,%1,2,1,1) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); - }else { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,0) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); - } -/* do we need fixup here ?*/ - asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +flush_partition (lpid, old); } void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,
Signature · 1 line
2.20.1
Re: [PATCH v4 5/8] KVM: PPC: Ultravisor: Restrict flush of the partition tlb cache
Quoted 49 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Ultravisor is responsible for flushing the tlb cache, since it manages the PATE entries. Hence skip tlb flush, if the ultravisor firmware is available.
Signed-off-by Ram Pai <[email protected]>Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/mm/book3s64/pgtable.c | 33 +++++++++++++++++------------- 1 file changed, 19 insertions(+), 14 deletions(-)arch/powerpc/mm/book3s64/pgtable.c +17
-5@@ -224,6 +224,23 @@ void __init mmu_partition_table_init(void)powernv_set_nmmu_ptcr (ptcr); } +static void flush_partition (unsigned int lpid,unsigned long dw0) +{ +if (dw0 &PATB_HR ) { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); + asm volatile(PPC_TLBIE_5(%0, %1, 2, 1, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); + }else { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 0) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); + } +/* do we need fixup here ?*/ + asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +} +static void __mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0,unsigned long dw1)@@ -238,20 +255,8 @@ static void __mmu_partition_table_set_entry(unsigned int lpid,* The type of flush (hash or radix) depends on what the previous * use of this partition ID was, not the new use. */ - asmvolatile ("ptesync" : : :"memory" ); -if (old &PATB_HR ) { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,1) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); - asm volatile(PPC_TLBIE_5(%0,%1,2,1,1) : :
arch/powerpc/mm/book3s64/pgtable.c static void __mmu_partition_table_set_entry(unsigned int lpid, · line 258
Quoted 10 lines
arch/powerpc/mm/book3s64/pgtable.c +1
-9- "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); - }else { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,0) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); - } -/* do we need fixup here ?*/ - asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR ))
__mmu_partition_table_set_entry() checks for UV and mmu_partition_table_set_entry() (the caller) checks for UV and the whole point of having separate flush_partition() and __mmu_partition_table_set_entry() is not really clear.
4/8 and 5/8 make more sense as one patch imho.
Quoted 5 lines
arch/powerpc/mm/book3s64/pgtable.c +1
-0+flush_partition (lpid, old); } void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,
Signature · 1 line
Alexey
Re: [PATCH v4 5/8] KVM: PPC: Ultravisor: Restrict flush of the partition tlb cache
Quoted 68 lines · Alexey Kardashevskiy
Quoted 49 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Ultravisor is responsible for flushing the tlb cache, since it manages the PATE entries. Hence skip tlb flush, if the ultravisor firmware is available.
Signed-off-by Ram Pai <[email protected]>Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/mm/book3s64/pgtable.c | 33 +++++++++++++++++------------- 1 file changed, 19 insertions(+), 14 deletions(-)arch/powerpc/mm/book3s64/pgtable.c +17
-5@@ -224,6 +224,23 @@ void __init mmu_partition_table_init(void)powernv_set_nmmu_ptcr (ptcr); } +static void flush_partition (unsigned int lpid,unsigned long dw0) +{ +if (dw0 &PATB_HR ) { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); + asm volatile(PPC_TLBIE_5(%0, %1, 2, 1, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); + }else { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 0) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); + } +/* do we need fixup here ?*/ + asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +} +static void __mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0,unsigned long dw1)@@ -238,20 +255,8 @@ static void __mmu_partition_table_set_entry(unsigned int lpid,* The type of flush (hash or radix) depends on what the previous * use of this partition ID was, not the new use. */ - asmvolatile ("ptesync" : : :"memory" ); -if (old &PATB_HR ) { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,1) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); - asm volatile(PPC_TLBIE_5(%0,%1,2,1,1) : :arch/powerpc/mm/book3s64/pgtable.c static void __mmu_partition_table_set_entry(unsigned int lpid, · line 258
Quoted 10 lines
arch/powerpc/mm/book3s64/pgtable.c +1
-9- "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); - }else { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,0) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); - } -/* do we need fixup here ?*/ - asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR ))
__mmu_partition_table_set_entry()checks for UV andmmu_partition_table_set_entry()(the caller) checks for UV and the whole point of having separateflush_partition()and__mmu_partition_table_set_entry()is not really clear.
Quoted 1 line
4/8 and 5/8 make more sense as one patch imho.
Makes sense. I merged them in the next version. Thanks.
Claudio
Quoted 7 lines
Quoted 7 lines
arch/powerpc/mm/book3s64/pgtable.c +1
-0+flush_partition (lpid, old); } void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,
Re: [PATCH v4 5/8] KVM: PPC: Ultravisor: Restrict flush of the partition tlb cache
In reply to Claudio CarvalhoQuoted 42 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Ultravisor is responsible for flushing the tlb cache, since it manages the PATE entries. Hence skip tlb flush, if the ultravisor firmware is available.
Signed-off-by Ram Pai <[email protected]>Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/mm/book3s64/pgtable.c | 33 +++++++++++++++++------------- 1 file changed, 19 insertions(+), 14 deletions(-)arch/powerpc/mm/book3s64/pgtable.c +17
-0@@ -224,6 +224,23 @@ void __init mmu_partition_table_init(void)powernv_set_nmmu_ptcr (ptcr); } +static void flush_partition (unsigned int lpid,unsigned long dw0) +{ +if (dw0 &PATB_HR ) { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); + asm volatile(PPC_TLBIE_5(%0, %1, 2, 1, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); + }else { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 0) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); + } +/* do we need fixup here ?*/ + asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +} +static void __mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0,unsigned long dw1)
arch/powerpc/mm/book3s64/pgtable.c static void · line 242
Quoted 5 lines
arch/powerpc/mm/book3s64/pgtable.c +0
-1@@ -238,20 +255,8 @@ static void__mmu_partition_table_set_entry(unsigned int lpid, * The type of flush (hash or radix) depends on what the previous * use of this partition ID was, not the new use. */ - asmvolatile ("ptesync" : : :"memory" );
Doesn't the line above that was deleted need to be added to the
beginning of flush_partition()
Quoted 19 lines
arch/powerpc/mm/book3s64/pgtable.c +2
-13-if (old &PATB_HR ) { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,1) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); - asm volatile(PPC_TLBIE_5(%0,%1,2,1,1) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); - }else { - asm volatile(PPC_TLBIE_5(%0,%1,2,0,0) : : - "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); -trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); - } -/* do we need fixup here ?*/ - asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +flush_partition (lpid, old); }void mmu_partition_table_set_entry(unsigned int lpid, unsigned long
dw0,
Re: [PATCH v4 5/8] KVM: PPC: Ultravisor: Restrict flush of the partition tlb cache
Quoted 42 lines · janani
Quoted 41 lines · Claudio Carvalho
From: Ram Pai <[email protected]>
Ultravisor is responsible for flushing the tlb cache, since it manages the PATE entries. Hence skip tlb flush, if the ultravisor firmware is available.
Signed-off-by Ram Pai <[email protected]>Signed-off-by Claudio Carvalho <[email protected]>---
arch/powerpc/mm/book3s64/pgtable.c| 33 +++++++++++++++++-------------1 file changed, 19 insertions(+), 14 deletions(-)arch/powerpc/mm/book3s64/pgtable.c +17
-0@@ -224,6 +224,23 @@ void __init mmu_partition_table_init(void)powernv_set_nmmu_ptcr (ptcr); } +static void flush_partition (unsigned int lpid,unsigned long dw0) +{ +if (dw0 &PATB_HR ) { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); + asm volatile(PPC_TLBIE_5(%0, %1, 2, 1, 1) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,1 ); + }else { + asm volatile(PPC_TLBIE_5(%0, %1, 2, 0, 0) : : + "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid)); +trace_tlbie (lpid,0 ,TLBIEL_INVAL_SET_LPID , lpid,2 ,0 ,0 ); + } +/* do we need fixup here ?*/ + asmvolatile ("eieio; tlbsync; ptesync" : : :"memory" ); +} +static void __mmu_partition_table_set_entry (unsigned int lpid,unsigned long dw0,unsigned long dw1)
Quoted 8 lines
arch/powerpc/mm/book3s64/pgtable.c static void · line 242
Quoted 6 lines
arch/powerpc/mm/book3s64/pgtable.c +0
-1@@ -238,20 +255,8 @@ static void__mmu_partition_table_set_entry(unsigned int lpid, * The type of flush (hash or radix) depends on what the previous * use of this partition ID was, not the new use. */ - asmvolatile ("ptesync" : : :"memory" );Doesn't the line above that was deleted need to be added to the
beginning offlush_partition()
It has to. It got dropped erroneously.
This is a good catch!
Thanks,
RP
PATCH v4 6/8 KVM: PPC: Ultravisor: Restrict LDBAR access
In reply to Claudio CarvalhoWhen the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
--- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 2 ++ arch/powerpc/platforms/powernv/idle.c | 6 ++++-- arch/powerpc/platforms/powernv/opal-imc.c | 4 ++++ 3 files changed, 10 insertions(+), 2 deletions(-)
arch/powerpc/kvm/book3s_hv_rmhandlers.S +2 -0
@@ -375,8 +375,10 @@ BEGIN_FTR_SECTION mtspr SPRN_RPR, r0 ld r0, KVM_SPLIT_PMMAR(r6) mtspr SPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70) ld r0, KVM_SPLIT_LDBAR(r6) mtspr SPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70) isync FTR_SECTION_ELSE /* On P9 we use the split_info for coordinating LPCR changes */
arch/powerpc/platforms/powernv/idle.c +4 -2
@@ -679,7 +679,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on) sprs.ptcr =mfspr (SPRN_PTCR ); sprs.rpr =mfspr (SPRN_RPR ); sprs.tscr =mfspr (SPRN_TSCR ); - sprs.ldbar =mfspr (SPRN_LDBAR ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) + sprs.ldbar =mfspr (SPRN_LDBAR ); sprs_saved = true;
@@ -762,7 +763,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on)mtspr (SPRN_PTCR , sprs.ptcr);mtspr (SPRN_RPR , sprs.rpr);mtspr (SPRN_TSCR , sprs.tscr); -mtspr (SPRN_LDBAR , sprs.ldbar); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_LDBAR , sprs.ldbar);if (pls >= pnv_first_tb_loss_level) {/* TB loss */
arch/powerpc/platforms/powernv/opal-imc.c +4 -0
@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)bool core_imc_reg = false, thread_imc_reg = false;u32 type; +/* Disable IMC devices, when Ultravisor is enabled. */ +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +return -EACCES ; +/* * Check whether this is kdump kernel. If yes, force the engines to * stop and return.
Signature · 1 line
2.20.1
Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
Quoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
What does LDBAR do? "Power ISA™ Version 3.0 B" or "User’s Manual POWER9 Processor" do not tell.
Quoted 67 lines
Signed-off-by Claudio Carvalho <[email protected]>Reviewed-by Ram Pai <[email protected]>Reviewed-by Ryan Grimm <[email protected]>Acked-by Paul Mackerras <[email protected]>--- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 2 ++ arch/powerpc/platforms/powernv/idle.c | 6 ++++-- arch/powerpc/platforms/powernv/opal-imc.c | 4 ++++ 3 files changed, 10 insertions(+), 2 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +2
-0@@ -375,8 +375,10 @@ BEGIN_FTR_SECTION mtspr SPRN_RPR, r0 ld r0, KVM_SPLIT_PMMAR(r6) mtspr SPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70) ld r0, KVM_SPLIT_LDBAR(r6) mtspr SPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70) isync FTR_SECTION_ELSE /* On P9 we use the split_info for coordinating LPCR changes */arch/powerpc/platforms/powernv/idle.c +4
-2@@ -679,7 +679,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on) sprs.ptcr =mfspr (SPRN_PTCR ); sprs.rpr =mfspr (SPRN_RPR ); sprs.tscr =mfspr (SPRN_TSCR ); - sprs.ldbar =mfspr (SPRN_LDBAR ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) + sprs.ldbar =mfspr (SPRN_LDBAR ); sprs_saved = true;@@ -762,7 +763,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on)mtspr (SPRN_PTCR , sprs.ptcr);mtspr (SPRN_RPR , sprs.rpr);mtspr (SPRN_TSCR , sprs.tscr); -mtspr (SPRN_LDBAR , sprs.ldbar); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_LDBAR , sprs.ldbar);if (pls >= pnv_first_tb_loss_level) {/* TB loss */ arch/powerpc/platforms/powernv/opal-imc.c +4
-0@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)bool core_imc_reg = false, thread_imc_reg = false;u32 type; +/* Disable IMC devices, when Ultravisor is enabled. */ +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +return -EACCES ; +/* * Check whether this is kdump kernel. If yes, force the engines to * stop and return.
Signature · 1 line
Alexey
Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
Quoted 7 lines · Alexey Kardashevskiy
Quoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
What does LDBAR do? "Power ISA™ Version 3.0 B" or "User’s Manual POWER9 Processor" do not tell.
LDBAR is a per-thread SPR used by thread-imc pmu to dump the counter data into memory. LDBAR contains memory address along with few other configuration bits (it is populated by the thread-imc pmu driver). It is populated and enabled only when any of the thread imc pmu events are monitored.
Maddy
Quoted 68 lines
Quoted 68 lines
Signed-off-by Claudio Carvalho <[email protected]>Reviewed-by Ram Pai <[email protected]>Reviewed-by Ryan Grimm <[email protected]>Acked-by Paul Mackerras <[email protected]>--- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 2 ++ arch/powerpc/platforms/powernv/idle.c | 6 ++++-- arch/powerpc/platforms/powernv/opal-imc.c | 4 ++++ 3 files changed, 10 insertions(+), 2 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +2
-0@@ -375,8 +375,10 @@ BEGIN_FTR_SECTION mtspr SPRN_RPR, r0 ld r0, KVM_SPLIT_PMMAR(r6) mtspr SPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70) ld r0, KVM_SPLIT_LDBAR(r6) mtspr SPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70) isync FTR_SECTION_ELSE /* On P9 we use the split_info for coordinating LPCR changes */arch/powerpc/platforms/powernv/idle.c +4
-2@@ -679,7 +679,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on) sprs.ptcr =mfspr (SPRN_PTCR ); sprs.rpr =mfspr (SPRN_RPR ); sprs.tscr =mfspr (SPRN_TSCR ); - sprs.ldbar =mfspr (SPRN_LDBAR ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) + sprs.ldbar =mfspr (SPRN_LDBAR ); sprs_saved = true;@@ -762,7 +763,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on)mtspr (SPRN_PTCR , sprs.ptcr);mtspr (SPRN_RPR , sprs.rpr);mtspr (SPRN_TSCR , sprs.tscr); -mtspr (SPRN_LDBAR , sprs.ldbar); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_LDBAR , sprs.ldbar);if (pls >= pnv_first_tb_loss_level) {/* TB loss */ arch/powerpc/platforms/powernv/opal-imc.c +4
-0@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)bool core_imc_reg = false, thread_imc_reg = false;u32 type; +/* Disable IMC devices, when Ultravisor is enabled. */ +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +return -EACCES ; +/* * Check whether this is kdump kernel. If yes, force the engines to * stop and return.
Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
Quoted 9 lines · maddy
Quoted 7 lines · Alexey Kardashevskiy
Quoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
What does LDBAR do? "Power ISA™ Version 3.0 B" or "User’s Manual POWER9 Processor" do not tell.
Quoted 7 lines
LDBAR is a per-thread SPR used by thread-imc pmu to dump the counter data into memory.
LDBAR contains memory address along with few other configuration bits (it is populated
by the thread-imc pmu driver). It is populated and enabled only when any of the thread
imc pmu events are monitored.
I was actually looking for a spec for this register, what is the document name?
Quoted 74 lines
Maddy
Quoted 71 lines
Quoted 71 lines
Signed-off-by Claudio Carvalho <[email protected]>Reviewed-by Ram Pai <[email protected]>Reviewed-by Ryan Grimm <[email protected]>Acked-by Paul Mackerras <[email protected]>---
arch/powerpc/kvm/book3s_hv_rmhandlers.S | 2 ++ arch/powerpc/platforms/powernv/idle.c | 6 ++++--arch/powerpc/platforms/powernv/opal-imc.c| 4 ++++3 files changed, 10 insertions(+), 2 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +0
-0@@ -375,8 +375,10 @@ BEGIN_FTR_SECTIONmtspr
SPRN_RPR, r0
ld r0,KVM_SPLIT_PMMAR(r6)
mtsprSPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70)
ld r0,KVM_SPLIT_LDBAR(r6)
mtsprSPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70) isync
FTR_SECTION_ELSE
/* On P9 we use thesplit_infofor coordinating LPCR changes */arch/powerpc/platforms/powernv/idle.c +2
-1@@ -679,7 +679,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on) sprs.ptcr = mfspr(SPRN_PTCR); sprs.rpr = mfspr(SPRN_RPR); sprs.tscr = mfspr(SPRN_TSCR); - sprs.ldbar = mfspr(SPRN_LDBAR); + if (!firmware_has_feature(FW_FEATURE_ULTRAVISOR)) + sprs.ldbar = mfspr(SPRN_LDBAR); sprs_saved = true;@@ -762,7 +763,8 @@ static unsigned long
power9_idle_stop(unsigned long psscr, bool mmu_on)
mtspr(SPRN_PTCR, sprs.ptcr);
mtspr(SPRN_RPR, sprs.rpr);
mtspr(SPRN_TSCR, sprs.tscr);
-mtspr(SPRN_LDBAR, sprs.ldbar);
+ if (!firmware_has_feature(FW_FEATURE_ULTRAVISOR))
+mtspr(SPRN_LDBAR, sprs.ldbar);
if (pls >=pnv_first_tb_loss_level) {
/* TB loss */arch/powerpc/platforms/powernv/opal-imc.c +0
-0@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev) bool core_imc_reg = false, thread_imc_reg = false; u32 type;+ /* Disable IMC devices, when Ultravisor is enabled. */ + if (
firmware_has_feature(FW_FEATURE_ULTRAVISOR)) + return-EACCES;
+
/*
* Check whether this is kdump kernel. If yes, force the engines to
* stop and return.
Signature · 1 line
Alexey
Re: Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
Quoted 21 lines · Alexey Kardashevskiy
Quoted 9 lines · maddy
Quoted 7 lines · Alexey Kardashevskiy
Quoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
What does LDBAR do? "Power ISA™ Version 3.0 B" or "User’s Manual POWER9 Processor" do not tell.
Quoted 7 lines
LDBAR is a per-thread SPR used by thread-imc pmu to dump the counter data into memory.
LDBAR contains memory address along with few other configuration bits (it is populated
by the thread-imc pmu driver). It is populated and enabled only when any of the thread
imc pmu events are monitored.
Quoted 2 lines
I was actually looking for a spec for this register, what is the document name?
Its not a architected register. Its documented in the Power9 workbook.
RP
Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
Quoted 18 lines · Ram Pai
Quoted 17 lines · Alexey Kardashevskiy
Quoted 14 lines · maddy
Quoted 6 lines · Alexey Kardashevskiy
Quoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
What does LDBAR do? "Power ISA™ Version 3.0 B" or "User’s Manual POWER9 Processor" do not tell.
LDBAR is a per-thread SPR used by thread-imc pmu to dump the counter data into memory.
LDBAR contains memory address along with few other configuration bits (it is populated
by the thread-imc pmu driver). It is populated and enabled only when any of the thread
imc pmu events are monitored.
Quoted 4 lines
Quoted 2 lines
I was actually looking for a spec for this register, what is the document name?
Its not a architected register. Its documented in the Power9 workbook.
I also found some information about the LDBAR inarch/powerpc/perf/imc-pmu.c
Claudio
Quoted 3 lines
RP
Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
In reply to Claudio CarvalhoQuoted 9 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
Signed-off-by Claudio Carvalho <[email protected]>Reviewed-by Ram Pai <[email protected]>Reviewed-by Ryan Grimm <[email protected]>Acked-by Paul Mackerras <[email protected]>
Quoted 66 lines
--- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 2 ++ arch/powerpc/platforms/powernv/idle.c | 6 ++++-- arch/powerpc/platforms/powernv/opal-imc.c | 4 ++++ 3 files changed, 10 insertions(+), 2 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +2
-0@@ -375,8 +375,10 @@ BEGIN_FTR_SECTION mtspr SPRN_RPR, r0 ld r0, KVM_SPLIT_PMMAR(r6) mtspr SPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70) ld r0, KVM_SPLIT_LDBAR(r6) mtspr SPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70) isync FTR_SECTION_ELSE /* On P9 we use the split_info for coordinating LPCR changes */arch/powerpc/platforms/powernv/idle.c +4
-2@@ -679,7 +679,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on) sprs.ptcr = mfspr(SPRN_PTCR); sprs.rpr =mfspr (SPRN_RPR ); sprs.tscr =mfspr (SPRN_TSCR ); - sprs.ldbar =mfspr (SPRN_LDBAR ); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) + sprs.ldbar =mfspr (SPRN_LDBAR ); sprs_saved = true;@@ -762,7 +763,8 @@ static unsigned long power9_idle_stop(unsigned long psscr, bool mmu_on) mtspr(SPRN_PTCR, sprs.ptcr);mtspr (SPRN_RPR , sprs.rpr);mtspr (SPRN_TSCR , sprs.tscr); -mtspr (SPRN_LDBAR , sprs.ldbar); +if (!firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +mtspr (SPRN_LDBAR , sprs.ldbar);if (pls >= pnv_first_tb_loss_level) {arch/powerpc/platforms/powernv/idle.c +0
-0/* TB loss */ arch/powerpc/platforms/powernv/opal-imc.c +4
-0@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)bool core_imc_reg = false, thread_imc_reg = false;u32 type; +/* Disable IMC devices, when Ultravisor is enabled. */ +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +return -EACCES ; +/* * Check whether this is kdump kernel. If yes, force the engines to * stop and return.
Michael Ellerman
reviewerRe: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
In reply to Claudio CarvalhoQuoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
Please roll up the replies to Alexey's question about LDBAR into the change log.
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 381
Quoted 12 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +2
-0@@ -375,8 +375,10 @@ BEGIN_FTR_SECTION mtspr SPRN_RPR, r0 ld r0, KVM_SPLIT_PMMAR(r6) mtspr SPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70) ld r0, KVM_SPLIT_LDBAR(r6) mtspr SPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70)
That's in Power8 code isn't it? Which will never have an ultravisor.
arch/powerpc/platforms/powernv/opal-imc.c static int opal_imc_counters_probe(struct platform_device *pdev) · line 259
Quoted 11 lines
arch/powerpc/platforms/powernv/opal-imc.c +3
-0@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)bool core_imc_reg = false, thread_imc_reg = false;u32 type; +/* Disable IMC devices, when Ultravisor is enabled. */ +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +return -EACCES ;
I don't mind taking this change. But at the same time should the IMC stuff just be omitted from the device tree when we're in ultravisor mode?
cheers
Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access
Quoted 7 lines · Michael Ellerman
Quoted 3 lines · Claudio Carvalho
When the ultravisor firmware is available, it takes control over the LDBAR register. In this case, thread-imc updates and save/restore operations on the LDBAR register are handled by ultravisor.
Please roll up the replies to Alexey's question about LDBAR into the change log.
Quoted 13 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 381
Quoted 12 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +2
-0@@ -375,8 +375,10 @@ BEGIN_FTR_SECTION mtspr SPRN_RPR, r0 ld r0, KVM_SPLIT_PMMAR(r6) mtspr SPRN_PMMAR, r0 +BEGIN_FW_FTR_SECTION_NESTED(70) ld r0, KVM_SPLIT_LDBAR(r6) mtspr SPRN_LDBAR, r0 +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70)That's in Power8 code isn't it? Which will never have an ultravisor.
IIUC, it might be executed in Power9 as well, but I can double check that.
Quoted 9 lines
Quoted 9 lines
arch/powerpc/platforms/powernv/opal-imc.c +0
-0@@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)bool core_imc_reg = false, thread_imc_reg = false;u32 type;
Quoted 5 lines
arch/powerpc/platforms/powernv/opal-imc.c static int opal_imc_counters_probe(struct platform_device *pdev) · line 259
Quoted 3 lines
arch/powerpc/platforms/powernv/opal-imc.c +3
-0+/* Disable IMC devices, when Ultravisor is enabled. */ +if (firmware_has_feature (FW_FEATURE_ULTRAVISOR )) +return -EACCES ;I don't mind taking this change. But at the same time should the IMC stuff just be omitted from the device tree when we're in ultravisor mode?
Yes. Maddy said that he will patch skiboot to remove the IMC nodes if ultravisor is present.
I added this check just to protect the kernel in case skiboot is not in the right level for some
reason.
Thanks,
Claudio
Quoted 3 lines
cheers
PATCH v4 7/8 KVM: PPC: Ultravisor: Enter a secure guest
In reply to Claudio CarvalhoFrom: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the UV_RETURN ucall number. R6,7 were used as temporary registers, hence we restore them.
Have fast_guest_return check the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).
Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
[ Pass SRR1 in r11 for UV_RETURN, fix kvmppc_msr_interrupt to preserve
the MSR_S bit ]
[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURN ucall number. Update commit message and ret_to_ultra comment ]
--- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/ultravisor-api.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++---- 4 files changed, 37 insertions(+), 6 deletions(-)
arch/powerpc/include/asm/kvm_host.h +1 -0
@@ -294,6 +294,7 @@ struct kvm_arch {cpumask_t cpu_in_guest;u8 radix;u8 fwnmi_enabled; +u8 secure_guest;bool threads_indep;bool nested_enable;pgd_t *pgtable;
arch/powerpc/include/asm/ultravisor-api.h +1 -0
@@ -19,5 +19,6 @@/* opcodes */ #define UV_WRITE_PATE 0xF104 +#define UV_RETURN 0xF11C#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */
arch/powerpc/kernel/asm-offsets.c +1 -0
@@ -508,6 +508,7 @@ int main(void)OFFSET (KVM_VRMA_SLB_V , kvm, arch.vrma_slb_v);OFFSET (KVM_RADIX , kvm, arch.radix);OFFSET (KVM_FWNMI , kvm, arch.fwnmi_enabled); +OFFSET (KVM_SECURE_GUEST , kvm, arch.secure_guest);OFFSET (VCPU_DSISR , kvm_vcpu, arch.shregs.dsisr);OFFSET (VCPU_DAR , kvm_vcpu, arch.shregs.dar);OFFSET (VCPU_VPA , kvm_vcpu, arch.vpa.pinned_addr);
arch/powerpc/kvm/book3s_hv_rmhandlers.S +34 -6
@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \
@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4) mtlr r5 - mtcr r6 ld r1, VCPU_GPR(R1)(r4) ld r2, VCPU_GPR(R2)(r4) ld r3, VCPU_GPR(R3)(r4) ld r5, VCPU_GPR(R5)(r4) - ld r6, VCPU_GPR(R6)(r4) - ld r7, VCPU_GPR(R7)(r4) ld r8, VCPU_GPR(R8)(r4) ld r9, VCPU_GPR(R9)(r4) ld r10, VCPU_GPR(R10)(r4)
@@ -1119,10 +1116,38 @@ BEGIN_FTR_SECTION mtspr SPRN_HDSISR, r0 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300) + ld r6, VCPU_KVM(r4) + lbz r7, KVM_SECURE_GUEST(r6) + cmpdi r7, 0 + bne ret_to_ultra + + lwz r6, VCPU_CR(r4) + mtcr r6 + + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) ld r0, VCPU_GPR(R0)(r4) ld r4, VCPU_GPR(R4)(r4) HRFI_TO_GUEST b . +/* + * We are entering a secure guest, so we have to invoke the ultravisor to do + * that. If we are returning from a hcall, the results are already in the + * appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of + * the reflected hcall, therefore we move it to R0 for the ultravisor and set + * R3 to the UV_RETURN ucall number. R6,7 were used as temporary registers + * above, hence we restore them. + */ +ret_to_ultra: + lwz r6, VCPU_CR(r4) + mtcr r6 + mfspr r11, SPRN_SRR1 + mr r0, r3 + LOAD_REG_IMMEDIATE(r3, UV_RETURN) + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) + ld r4, VCPU_GPR(R4)(r4) + sc 2 /* * Enter the guest on a P9 or later system where we have exactly
@@ -3318,13 +3343,16 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) * r0 is used as a scratch register */ kvmppc_msr_interrupt: + andis. r0, r11, MSR_S@h rldicl r0, r11, 64 - MSR_TS_S_LG, 62 - cmpwi r0, 2 /* Check if we are in transactional state.. */ + cmpwi cr1, r0, 2 /* Check if we are in transactional state.. */ ld r11, VCPU_INTR_MSR(r9) - bne 1f + bne cr1, 1f /* ... if transactional, change to suspended */ li r0, 1 1: rldimi r11, r0, MSR_TS_S_LG, 63 - MSR_TS_T_LG + beqlr + oris r11, r11, MSR_S@h /* preserve MSR_S bit setting */ blr /*
Signature · 1 line
2.20.1
Re: [PATCH v4 7/8] KVM: PPC: Ultravisor: Enter a secure guest
Quoted 23 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the
UV_RETURNucall number. R6,7 were used as temporary registers, hence we restore them.Have
fast_guest_returncheck the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
Signed-off-by Sukadev Bhattiprolu <[email protected]>[ Pass SRR1 in r11 for
UV_RETURN, fixkvmppc_msr_interruptto preserve
theMSR_Sbit ]
Quoted 7 lines
Signed-off-by Paul Mackerras <[email protected]>[ Fix UV_RETURN ucall number and arch.secure_guest check ]Signed-off-by Ram Pai <[email protected]>[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURNucall number. Update commit message andret_to_ultra
comment ]Signed-off-by Claudio Carvalho <[email protected]>
Quoted 138 lines
--- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/ultravisor-api.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++---- 4 files changed, 37 insertions(+), 6 deletions(-)arch/powerpc/include/asm/kvm_host.h +1
-0@@ -294,6 +294,7 @@ struct kvm_arch {cpumask_t cpu_in_guest;u8 radix;u8 fwnmi_enabled; +u8 secure_guest;bool threads_indep;bool nested_enable;pgd_t *pgtable;arch/powerpc/include/asm/ultravisor-api.h +1
-0@@ -19,5 +19,6 @@/* opcodes */ #define UV_WRITE_PATE 0xF104 +#define UV_RETURN 0xF11C#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ arch/powerpc/kernel/asm-offsets.c +1
-0@@ -508,6 +508,7 @@ int main(void)OFFSET (KVM_VRMA_SLB_V , kvm, arch.vrma_slb_v);OFFSET (KVM_RADIX , kvm, arch.radix);OFFSET (KVM_FWNMI , kvm, arch.fwnmi_enabled); +OFFSET (KVM_SECURE_GUEST , kvm, arch.secure_guest);OFFSET (VCPU_DSISR , kvm_vcpu, arch.shregs.dsisr);OFFSET (VCPU_DAR , kvm_vcpu, arch.shregs.dar);OFFSET (VCPU_VPA , kvm_vcpu, arch.vpa.pinned_addr);arch/powerpc/kvm/book3s_hv_rmhandlers.S +13
-4@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4) mtlr r5 - mtcr r6 ld r1, VCPU_GPR(R1)(r4) ld r2, VCPU_GPR(R2)(r4) ld r3, VCPU_GPR(R3)(r4) ld r5, VCPU_GPR(R5)(r4) - ld r6, VCPU_GPR(R6)(r4) - ld r7, VCPU_GPR(R7)(r4) ld r8, VCPU_GPR(R8)(r4) ld r9, VCPU_GPR(R9)(r4) ld r10, VCPU_GPR(R10)(r4)@@ -1119,10 +1116,38 @@ BEGIN_FTR_SECTION mtspr SPRN_HDSISR, r0 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300) + ld r6, VCPU_KVM(r4) + lbz r7, KVM_SECURE_GUEST(r6) + cmpdi r7, 0 + bne ret_to_ultra + + lwz r6, VCPU_CR(r4) + mtcr r6 + + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) ld r0, VCPU_GPR(R0)(r4) ld r4, VCPU_GPR(R4)(r4) HRFI_TO_GUEST b . +/* + * We are entering a secure guest, so we have to invoke the ultravisorto do
+ * that. If we are returning from a hcall, the results are already in the
+ * appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of
+ * the reflected hcall, therefore we move it to R0 for the ultravisor and set
+ * R3 to theUV_RETURNucall number. R6,7 were used as temporary registers
+ * above, hence we restore them.
+ */arch/powerpc/kvm/book3s_hv_rmhandlers.S +15
-2+ret_to_ultra: + lwz r6, VCPU_CR(r4) + mtcr r6 + mfspr r11, SPRN_SRR1 + mr r0, r3 + LOAD_REG_IMMEDIATE(r3, UV_RETURN) + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) + ld r4, VCPU_GPR(R4)(r4) + sc 2 /* * Enter the guest on a P9 or later system where we have exactly@@ -3318,13 +3343,16 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) * r0 is used as a scratch register */ kvmppc_msr_interrupt: + andis. r0, r11, MSR_S@h rldicl r0, r11, 64 - MSR_TS_S_LG, 62 - cmpwi r0, 2 /* Check if we are in transactional state.. */ + cmpwi cr1, r0, 2 /* Check if we are in transactional state.. */ ld r11, VCPU_INTR_MSR(r9) - bne 1f + bne cr1, 1f /* ... if transactional, change to suspended */ li r0, 1 1: rldimi r11, r0, MSR_TS_S_LG, 63 - MSR_TS_T_LG + beqlr + oris r11, r11, MSR_S@h /* preserve MSR_S bit setting */ blr /*
Re: [PATCH v4 7/8] KVM: PPC: Ultravisor: Enter a secure guest
Quoted 30 lines · janani
Quoted 21 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the
UV_RETURNucall number. R6,7 were used as temporary registers, hence we restore them.Have
fast_guest_returncheck the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
Signed-off-by Sukadev Bhattiprolu <[email protected]>Quoted 8 lines
[ Pass SRR1 in r11 for
UV_RETURN, fixkvmppc_msr_interruptto preserve theMSR_Sbit ]Signed-off-by Paul Mackerras <[email protected]>[ Fix UV_RETURN ucall number and arch.secure_guest check ]Signed-off-by Ram Pai <[email protected]>[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURNucall number. Update commit message andret_to_ultracomment ]Signed-off-by Claudio Carvalho <[email protected]>
Quoted 1 line
Reviewed-by Janani Janakiraman <[email protected]>
Thanks Janani for reviewing the patchset.
Claudio
Quoted 137 lines
Quoted 136 lines
---
arch/powerpc/include/asm/kvm_host.h | 1 +arch/powerpc/include/asm/ultravisor-api.h| 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++----4 files changed, 37 insertions(+), 6 deletions(-)arch/powerpc/include/asm/kvm_host.h +1
-0@@ -294,6 +294,7 @@ struct kvm_arch { cpumask_t cpu_in_guest; u8 radix; u8 fwnmi_enabled; + u8 secure_guest; bool threads_indep; bool nested_enable; pgd_t *pgtable;arch/powerpc/include/asm/ultravisor-api.h +1
-0@@ -19,5 +19,6 @@ /* opcodes */ #define UV_WRITE_PATE 0xF104 +#define UV_RETURN 0xF11C #endif /* _ASM_POWERPC_ULTRAVISOR_API_H */arch/powerpc/kernel/asm-offsets.c +1
-0@@ -508,6 +508,7 @@ int main(void) OFFSET(KVM_VRMA_SLB_V, kvm, arch.vrma_slb_v); OFFSET(KVM_RADIX, kvm, arch.radix); OFFSET(KVM_FWNMI, kvm, arch.fwnmi_enabled); + OFFSET(KVM_SECURE_GUEST, kvm, arch.secure_guest); OFFSET(VCPU_DSISR, kvm_vcpu, arch.shregs.dsisr); OFFSET(VCPU_DAR, kvm_vcpu, arch.shregs.dar); OFFSET(VCPU_VPA, kvm_vcpu, arch.vpa.pinned_addr);arch/powerpc/kvm/book3s_hv_rmhandlers.S +1
-1@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4)mtlr r5
- mtcr r6ld r1, VCPU_GPR(R1)(r4)
ld r2, VCPU_GPR(R2)(r4)
ld r3, VCPU_GPR(R3)(r4)
ld r5, VCPU_GPR(R5)(r4)
- ld r6, VCPU_GPR(R6)(r4)
- ld r7, VCPU_GPR(R7)(r4)
ld r8, VCPU_GPR(R8)(r4)
ld r9, VCPU_GPR(R9)(r4)
ld r10, VCPU_GPR(R10)(r4)arch/powerpc/kvm/book3s_hv_rmhandlers.S +0
-0@@ -1119,10 +1116,38 @@ BEGIN_FTR_SECTIONmtspr
SPRN_HDSISR, r0
END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)+ ld r6,
VCPU_KVM(r4)
+ lbz r7,KVM_SECURE_GUEST(r6)
+ cmpdi r7, 0
+ bneret_to_ultraarch/powerpc/kvm/book3s_hv_rmhandlers.S +6
-0+ + lwz r6, VCPU_CR(r4) + mtcr r6 + + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4)ld r0, VCPU_GPR(R0)(r4)
ld r4, VCPU_GPR(R4)(r4)
HRFI_TO_GUEST
b .
+/*
+ * We are entering a secure guest, so we have to invoke the ultravisor to do
+ * that. If we are returning from a hcall, the results are already in the
+ * appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of
+ * the reflected hcall, therefore we move it to R0 for the ultravisor and set
+ * R3 to theUV_RETURNucall number. R6,7 were used as temporary registers
+ * above, hence we restore them.
+ */
+ret_to_ultra:
+ lwz r6,VCPU_CR(r4)
+ mtcr r6
+ mfspr r11,SPRN_SRR1
+ mr r0, r3
+LOAD_REG_IMMEDIATE(r3, UV_RETURN)
+ ld r7, VCPU_GPR(R7)(r4)
+ ld r6, VCPU_GPR(R6)(r4)
+ ld r4, VCPU_GPR(R4)(r4)
+ sc 2/*
* Enter the guest on a P9 or later system where we have exactlyarch/powerpc/kvm/book3s_hv_rmhandlers.S +0
-0@@ -3318,13 +3343,16 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) * r0 is used as a scratch register */kvmppc_msr_interrupt:
+ andis. r0, r11, MSR_S@h
rldicl r0, r11, 64 -MSR_TS_S_LG, 62
- cmpwi r0, 2 /* Check if we are in transactional state.. */ + cmpwi cr1, r0, 2 /* Check if we are in transactional state.. */ ld r11,VCPU_INTR_MSR(r9)
- bne 1f
+ bne cr1, 1f
/* ... if transactional, change to suspended */
li r0, 1
1: rldimi r11, r0,MSR_TS_S_LG, 63 -MSR_TS_T_LG
+ beqlr
+ oris r11, r11, MSR_S@h /* preserveMSR_Sbit setting */ blr/*
Michael Ellerman
reviewerRe: [PATCH v4 7/8] KVM: PPC: Ultravisor: Enter a secure guest
In reply to Claudio CarvalhoQuoted 9 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
Quoted 5 lines
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the
UV_RETURNucall number. R6,7 were used as temporary registers, hence we restore them.
This is another case where some documentation would help people to review the code.
Quoted 51 lines
Have
fast_guest_returncheck the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
Signed-off-by Sukadev Bhattiprolu <[email protected]>[ Pass SRR1 in r11 for
UV_RETURN, fixkvmppc_msr_interruptto preserve
theMSR_Sbit ]Signed-off-by Paul Mackerras <[email protected]>[ Fix UV_RETURN ucall number and arch.secure_guest check ]Signed-off-by Ram Pai <[email protected]>[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURNucall number. Update commit message andret_to_ultracomment ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/ultravisor-api.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++---- 4 files changed, 37 insertions(+), 6 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +1
-4@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4) mtlr r5 - mtcr r6 ld r1, VCPU_GPR(R1)(r4) ld r2, VCPU_GPR(R2)(r4) ld r3, VCPU_GPR(R3)(r4) ld r5, VCPU_GPR(R5)(r4) - ld r6, VCPU_GPR(R6)(r4) - ld r7, VCPU_GPR(R7)(r4) ld r8, VCPU_GPR(R8)(r4) ld r9, VCPU_GPR(R9)(r4) ld r10, VCPU_GPR(R10)(r4)
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1121
Quoted 6 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +3
-0@@ -1119,10 +1116,38 @@ BEGIN_FTR_SECTION mtspr SPRN_HDSISR, r0 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300) + ld r6, VCPU_KVM(r4) + lbz r7, KVM_SECURE_GUEST(r6) + cmpdi r7, 0
You could hoist the load of r6 and r7 to here?
Quoted 15 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +11
-0+ bne ret_to_ultra + + lwz r6, VCPU_CR(r4) + mtcr r6 + + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) ld r0, VCPU_GPR(R0)(r4) ld r4, VCPU_GPR(R4)(r4) HRFI_TO_GUEST b . +/* + * We are entering a secure guest, so we have to invoke the ultravisor to do + * that. If we are returning from a hcall, the results are already in the + * appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1146
Quoted 10 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +10
-0+ * the reflected hcall, therefore we move it to R0 for the ultravisor and set + * R3 to the UV_RETURN ucall number. R6,7 were used as temporary registers + * above, hence we restore them. + */ +ret_to_ultra: + lwz r6, VCPU_CR(r4) + mtcr r6 + mfspr r11, SPRN_SRR1 + mr r0, r3 + LOAD_REG_IMMEDIATE(r3, UV_RETURN)
Worth open coding to save three instructions?
Quoted 14 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +5
-1+ ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) + ld r4, VCPU_GPR(R4)(r4) + sc 2 /* * Enter the guest on a P9 or later system where we have exactly@@ -3318,13 +3343,16 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) * r0 is used as a scratch register */ kvmppc_msr_interrupt: + andis. r0, r11, MSR_S@h rldicl r0, r11, 64 - MSR_TS_S_LG, 62 - cmpwi r0, 2 /* Check if we are in transactional state.. */
arch/powerpc/kvm/book3s_hv_rmhandlers.S END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) · line 3356
Quoted 10 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +4
-1+ cmpwi cr1, r0, 2 /* Check if we are in transactional state.. */ ld r11, VCPU_INTR_MSR(r9) - bne 1f + bne cr1, 1f /* ... if transactional, change to suspended */ li r0, 1 1: rldimi r11, r0, MSR_TS_S_LG, 63 - MSR_TS_T_LG + beqlr + oris r11, r11, MSR_S@h /* preserve MSR_S bit setting */ blr
I don't see this part mentioned in the change log?
It's also pretty subtle, a comment might be helpful.
cheers
Re: [PATCH v4 7/8] KVM: PPC: Ultravisor: Enter a secure guest
Quoted 73 lines · Michael Ellerman [[email protected]] wrote:
Quoted 9 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
Quoted 5 lines
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the
UV_RETURNucall number. R6,7 were used as temporary registers, hence we restore them.This is another case where some documentation would help people to review the code.
Quoted 54 lines
Have
fast_guest_returncheck the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
Signed-off-by Sukadev Bhattiprolu <[email protected]>[ Pass SRR1 in r11 for
UV_RETURN, fixkvmppc_msr_interruptto preserve
theMSR_Sbit ]Signed-off-by Paul Mackerras <[email protected]>[ Fix UV_RETURN ucall number and arch.secure_guest check ]Signed-off-by Ram Pai <[email protected]>[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURNucall number. Update commit message andret_to_ultracomment ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/ultravisor-api.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++---- 4 files changed, 37 insertions(+), 6 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +1
-4@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4) mtlr r5 - mtcr r6 ld r1, VCPU_GPR(R1)(r4) ld r2, VCPU_GPR(R2)(r4) ld r3, VCPU_GPR(R3)(r4) ld r5, VCPU_GPR(R5)(r4) - ld r6, VCPU_GPR(R6)(r4) - ld r7, VCPU_GPR(R7)(r4) ld r8, VCPU_GPR(R8)(r4) ld r9, VCPU_GPR(R9)(r4) ld r10, VCPU_GPR(R10)(r4)@@ -1119,10 +1116,38 @@ BEGIN_FTR_SECTION mtspr SPRN_HDSISR, r0 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
Quoted 5 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1121
Quoted 3 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +3
-0+ ld r6, VCPU_KVM(r4) + lbz r7, KVM_SECURE_GUEST(r6) + cmpdi r7, 0You could hoist the load of r6 and r7 to here?
we could move 'ld r7' here. r6 is used to restore CR below so it (r6) has to stay there?
Quoted 20 lines
Quoted 20 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +15
-0+ bne ret_to_ultra + + lwz r6, VCPU_CR(r4) + mtcr r6 + + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) ld r0, VCPU_GPR(R0)(r4) ld r4, VCPU_GPR(R4)(r4) HRFI_TO_GUEST b . +/* + * We are entering a secure guest, so we have to invoke the ultravisor to do + * that. If we are returning from a hcall, the results are already in the + * appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of + * the reflected hcall, therefore we move it to R0 for the ultravisor and set + * R3 to the UV_RETURN ucall number. R6,7 were used as temporary registers + * above, hence we restore them. + */
Quoted 8 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1146
Quoted 6 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +6
-0+ret_to_ultra: + lwz r6, VCPU_CR(r4) + mtcr r6 + mfspr r11, SPRN_SRR1 + mr r0, r3 + LOAD_REG_IMMEDIATE(r3, UV_RETURN)Worth open coding to save three instructions?
Yes, good point:
- LOAD_REG_IMMEDIATE(r3, UV_RETURN)
+
+ li r3, 0 + oris r3, r3, (UV_RETURN)@__AS_ATHIGH + ori r3, r3, (UV_RETURN)@l
+
Quoted 31 lines
Quoted 15 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +5
-1+ ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) + ld r4, VCPU_GPR(R4)(r4) + sc 2 /* * Enter the guest on a P9 or later system where we have exactly@@ -3318,13 +3343,16 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) * r0 is used as a scratch register */ kvmppc_msr_interrupt: + andis. r0, r11, MSR_S@h rldicl r0, r11, 64 - MSR_TS_S_LG, 62 - cmpwi r0, 2 /* Check if we are in transactional state.. */arch/powerpc/kvm/book3s_hv_rmhandlers.S END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX) · line 3356
Quoted 10 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +4
-1+ cmpwi cr1, r0, 2 /* Check if we are in transactional state.. */ ld r11, VCPU_INTR_MSR(r9) - bne 1f + bne cr1, 1f /* ... if transactional, change to suspended */ li r0, 1 1: rldimi r11, r0, MSR_TS_S_LG, 63 - MSR_TS_T_LG + beqlr + oris r11, r11, MSR_S@h /* preserve MSR_S bit setting */ blrI don't see this part mentioned in the change log?
It's also pretty subtle, a comment might be helpful.
cheers
Michael Ellerman
reviewerRe: [PATCH v4 7/8] KVM: PPC: Ultravisor: Enter a secure guest
Quoted 74 lines · Sukadev Bhattiprolu
Quoted 73 lines · Michael Ellerman [[email protected]] wrote:
Quoted 9 lines · Claudio Carvalho
From: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
Quoted 5 lines
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the
UV_RETURNucall number. R6,7 were used as temporary registers, hence we restore them.This is another case where some documentation would help people to review the code.
Quoted 54 lines
Have
fast_guest_returncheck the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
Signed-off-by Sukadev Bhattiprolu <[email protected]>[ Pass SRR1 in r11 for
UV_RETURN, fixkvmppc_msr_interruptto preserve
theMSR_Sbit ]Signed-off-by Paul Mackerras <[email protected]>[ Fix UV_RETURN ucall number and arch.secure_guest check ]Signed-off-by Ram Pai <[email protected]>[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURNucall number. Update commit message andret_to_ultracomment ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/ultravisor-api.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++---- 4 files changed, 37 insertions(+), 6 deletions(-)arch/powerpc/kvm/book3s_hv_rmhandlers.S +1
-4@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4) mtlr r5 - mtcr r6 ld r1, VCPU_GPR(R1)(r4) ld r2, VCPU_GPR(R2)(r4) ld r3, VCPU_GPR(R3)(r4) ld r5, VCPU_GPR(R5)(r4) - ld r6, VCPU_GPR(R6)(r4) - ld r7, VCPU_GPR(R7)(r4) ld r8, VCPU_GPR(R8)(r4) ld r9, VCPU_GPR(R9)(r4) ld r10, VCPU_GPR(R10)(r4)@@ -1119,10 +1116,38 @@ BEGIN_FTR_SECTION mtspr SPRN_HDSISR, r0 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
Quoted 8 lines
Quoted 5 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1121
Quoted 3 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +3
-0+ ld r6, VCPU_KVM(r4) + lbz r7, KVM_SECURE_GUEST(r6) + cmpdi r7, 0You could hoist the load of r6 and r7 to here?
we could move 'ld r7' here. r6 is used to restore CR below so it (r6) has to stay there?
It's used to restore CR in both paths, so both paths load VCPU_CR(r4) into r6. So we could instead do that load once, before the branch?
Quoted 30 lines
Quoted 19 lines
Quoted 19 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +15
-0+ bne ret_to_ultra + + lwz r6, VCPU_CR(r4) + mtcr r6 + + ld r7, VCPU_GPR(R7)(r4) + ld r6, VCPU_GPR(R6)(r4) ld r0, VCPU_GPR(R0)(r4) ld r4, VCPU_GPR(R4)(r4) HRFI_TO_GUEST b . +/* + * We are entering a secure guest, so we have to invoke the ultravisor to do + * that. If we are returning from a hcall, the results are already in the + * appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of + * the reflected hcall, therefore we move it to R0 for the ultravisor and set + * R3 to the UV_RETURN ucall number. R6,7 were used as temporary registers + * above, hence we restore them. + */Quoted 8 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1146
Quoted 6 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +6
-0+ret_to_ultra: + lwz r6, VCPU_CR(r4) + mtcr r6 + mfspr r11, SPRN_SRR1 + mr r0, r3 + LOAD_REG_IMMEDIATE(r3, UV_RETURN)Worth open coding to save three instructions?
Yes, good point:
Review
Quoted 5 lines
(quoted diff) +4
-1- LOAD_REG_IMMEDIATE(r3, UV_RETURN) + + li r3, 0 + oris r3, r3, (UV_RETURN)@__AS_ATHIGH + ori r3, r3, (UV_RETURN)@l
This should do it no?
li r3, 0
oris r3, r3, UV_RETURN
cheers
Nicholas Piggin
reviewerRe: [PATCH v4 7/8] KVM: PPC: Ultravisor: Enter a secure guest
In reply to Claudio CarvalhoClaudio Carvalho's on June 29, 2019 6:08 am:
Quoted 90 lines
From: Sukadev Bhattiprolu <[email protected]>
To enter a secure guest, we have to go through the ultravisor, therefore we do a ucall when we are entering a secure guest.
This change is needed for any sort of entry to the secure guest from the hypervisor, whether it is a return from an hcall, a return from a hypervisor interrupt, or the first time that a secure guest vCPU is run.
If we are returning from an hcall, the results are already in the appropriate registers R3:12, except for R3, R6 and R7. R3 has the status of the reflected hcall, therefore we move it to R0 for the ultravisor and set R3 to the
UV_RETURNucall number. R6,7 were used as temporary registers, hence we restore them.Have
fast_guest_returncheck the kvm_arch.secure_guest field so that a new CPU enters UV when started (in response to a RTAS start-cpu call).Thanks to input from Paul Mackerras, Ram Pai and Mike Anderson.
Signed-off-by Sukadev Bhattiprolu <[email protected]>[ Pass SRR1 in r11 for
UV_RETURN, fixkvmppc_msr_interruptto preserve
theMSR_Sbit ]Signed-off-by Paul Mackerras <[email protected]>[ Fix UV_RETURN ucall number and arch.secure_guest check ]Signed-off-by Ram Pai <[email protected]>[ Save the actual R3 in R0 for the ultravisor and use R3 for the
UV_RETURNucall number. Update commit message andret_to_ultracomment ]Signed-off-by Claudio Carvalho <[email protected]>--- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/include/asm/ultravisor-api.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/book3s_hv_rmhandlers.S | 40 +++++++++++++++++++---- 4 files changed, 37 insertions(+), 6 deletions(-)arch/powerpc/include/asm/kvm_host.h +1
-0@@ -294,6 +294,7 @@ struct kvm_arch {cpumask_t cpu_in_guest;u8 radix;u8 fwnmi_enabled; +u8 secure_guest;bool threads_indep;bool nested_enable;pgd_t *pgtable;arch/powerpc/include/asm/ultravisor-api.h +1
-0@@ -19,5 +19,6 @@/* opcodes */ #define UV_WRITE_PATE 0xF104 +#define UV_RETURN 0xF11C#endif /* _ASM_POWERPC_ULTRAVISOR_API_H */ arch/powerpc/kernel/asm-offsets.c +1
-0@@ -508,6 +508,7 @@ int main(void)OFFSET (KVM_VRMA_SLB_V , kvm, arch.vrma_slb_v);OFFSET (KVM_RADIX , kvm, arch.radix);OFFSET (KVM_FWNMI , kvm, arch.fwnmi_enabled); +OFFSET (KVM_SECURE_GUEST , kvm, arch.secure_guest);OFFSET (VCPU_DSISR , kvm_vcpu, arch.shregs.dsisr);OFFSET (VCPU_DAR , kvm_vcpu, arch.shregs.dar);OFFSET (VCPU_VPA , kvm_vcpu, arch.vpa.pinned_addr);arch/powerpc/kvm/book3s_hv_rmhandlers.S +1
-2@@ -36,6 +36,7 @@ #include <asm/asm-compat.h> #include <asm/feature-fixups.h> #include <asm/cpuidle.h> +#include <asm/ultravisor-api.h> /* Sign-extend HDEC if not on POWER9 */ #define EXTEND_HDEC(reg) \@@ -1092,16 +1093,12 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) ld r5, VCPU_LR(r4) - ld r6, VCPU_CR(r4) mtlr r5 - mtcr r6
arch/powerpc/kvm/book3s_hv_rmhandlers.S BEGIN_FTR_SECTION · line 1104
Quoted 10 lines
arch/powerpc/kvm/book3s_hv_rmhandlers.S +0
-2ld r1, VCPU_GPR(R1)(r4) ld r2, VCPU_GPR(R2)(r4) ld r3, VCPU_GPR(R3)(r4) ld r5, VCPU_GPR(R5)(r4) - ld r6, VCPU_GPR(R6)(r4) - ld r7, VCPU_GPR(R7)(r4) ld r8, VCPU_GPR(R8)(r4) ld r9, VCPU_GPR(R9)(r4) ld r10, VCPU_GPR(R10)(r4)
Just to try to be less arbitrary about things, could you use regs adjacent to r4? Generally good because then it has a chance to get our loads paired up (which may not help some CPUs).
Thanks,
Nick
PATCH v4 8/8 KVM: PPC: Ultravisor: Check for MSR_S during hv_reset_msr
In reply to Claudio CarvalhoFrom: Michael Anderson <[email protected]>
- Check for MSR_S so that kvmppc_set_msr will include it. Prior to this change return to guest would not have the S bit set.
- Patch based on comment from Paul Mackerras <[email protected]>
--- arch/powerpc/kvm/book3s_64_mmu_hv.c | 1 + 1 file changed, 1 insertion(+)
arch/powerpc/kvm/book3s_64_mmu_hv.c +1 -0
@@ -295,6 +295,7 @@ static void kvmppc_mmu_book3s_64_hv_reset_msr(struct kvm_vcpu *vcpu) msr |=MSR_TS_S ; else msr |= vcpu->arch.shregs.msr & MSR_TS_MASK; + msr |= vcpu->arch.shregs.msr &MSR_S ;kvmppc_set_msr (vcpu, msr); }
Signature · 1 line
2.20.1
Re: [PATCH v4 8/8] KVM: PPC: Ultravisor: Check for MSR_S during hv_reset_msr
Quoted 11 lines · Claudio Carvalho
From: Michael Anderson <[email protected]>
- Check for
MSR_Sso thatkvmppc_set_msrwill include it. Prior to
this
change return to guest would not have the S bit set.- Patch based on comment from Paul Mackerras <[email protected]>
Signed-off-by Michael Anderson <[email protected]>Signed-off-by Claudio Carvalho <[email protected]>Acked-by Paul Mackerras <[email protected]>
Quoted 17 lines
--- arch/powerpc/kvm/book3s_64_mmu_hv.c | 1 + 1 file changed, 1 insertion(+)arch/powerpc/kvm/book3s_64_mmu_hv.c +1
-0@@ -295,6 +295,7 @@ static void kvmppc_mmu_book3s_64_hv_reset_msr(struct kvm_vcpu *vcpu) msr |= MSR_TS_S; else msr |= vcpu->arch.shregs.msr & MSR_TS_MASK; + msr |= vcpu->arch.shregs.msr &MSR_S ;kvmppc_set_msr (vcpu, msr); }
Michael Ellerman
reviewerRe: [PATCH v4 8/8] KVM: PPC: Ultravisor: Check for MSR_S during hv_reset_msr
In reply to Claudio CarvalhoQuoted 4 lines · Claudio Carvalho
From: Michael Anderson <[email protected]>
- Check for
MSR_Sso thatkvmppc_set_msrwill include it. Prior to this change return to guest would not have the S bit set.
That sounds like it would be bad?
Please spell out what the practical impact of the patch is, ie. somewhere on the spectrum from "without this patch everything catches fire", to "this is not a bug but makes things clearer because ..."
cheers
Quoted 23 lines
- Patch based on comment from Paul Mackerras <[email protected]>
Signed-off-by Michael Anderson <[email protected]>Signed-off-by Claudio Carvalho <[email protected]>Acked-by Paul Mackerras <[email protected]>--- arch/powerpc/kvm/book3s_64_mmu_hv.c | 1 + 1 file changed, 1 insertion(+)arch/powerpc/kvm/book3s_64_mmu_hv.c +1
-0@@ -295,6 +295,7 @@ static void kvmppc_mmu_book3s_64_hv_reset_msr(struct kvm_vcpu *vcpu) msr |=MSR_TS_S ; else msr |= vcpu->arch.shregs.msr & MSR_TS_MASK; + msr |= vcpu->arch.shregs.msr &MSR_S ;kvmppc_set_msr (vcpu, msr); }Signature · 1 line
2.20.1