Skip to contentSkip to conversation

KVM: arm64: NV trap forwarding infrastructure

PATCH Superseded Marc Zyngier 6 participants 68 messages Jul 28, 2023 – Aug 7, 2023

Subsystems (3)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Marc Zyngier

author maintainer
#0 lore raw

v2 cover KVM: arm64: NV trap forwarding infrastructure

As people are getting tired of seeing the full NV series, I've extracted some of the easy stuff which I'm targeting for 6.6.

This implements the so called "trap forwarding" infrastructure, which gets used when we take a trap from an L2 guest and that the L1 guest wants to see the trap for itself.

Most of the series is pretty boring stuff, mostly a long list of encodings which are mapped to a set of trap bits. I swear they are correct. Sort of (then are much more correct now that Eric has gone through them all with a fine comb).

The interesting bit is around how we compute the trap result, which is pretty complex due to the layers of crap the architecture has piled over the years (a single op can be trapped by multiple coarse grained trap bits, or a fine grained trap bit, which may itself be conditioned by another control bit -- madness).

This also results in some rework of both the FGT stuff (for which I carry a patch from Mark) and newly introduced the HCRX support.

With that (and the rest of the NV series[0]), FGT gets exposed to guests and the trapping seems to work as expected.

* From v1 [1]:

- Lots of fixups all over the map (too many to mention) after Eric's fantastic reviewing effort. Hopefully the result is easier to understand and less wrong

- Amended Mark's patch to use the ARM64_CPUID_FIELDS() macro

- Collected RBs, with thanks.

[0] https://git.kernel.org/pub/scm/linux/kernel/git/maz/arm-platforms.git/log/?h=kvm-arm64/nv-6.6-WIP
[1] https://lore.kernel.org/all/[email protected]/

Marc Zyngier (25):
arm64: Add missing VA CMO encodings
arm64: Add missing ERX*_EL1 encodings
arm64: Add missing DC ZVA/GVA/GZVA encodings
arm64: Add TLBI operation encodings
arm64: Add AT operation encodings
arm64: Add debug registers affected by HDFGxTR_EL2
arm64: Add missing BRB/CFP/DVP/CPP instructions
arm64: Add HDFGRTR_EL2 and HDFGWTR_EL2 layouts
KVM: arm64: Correctly handle ACCDATA_EL1 traps
KVM: arm64: Add missing HCR_EL2 trap bits
KVM: arm64: nv: Add FGT registers
KVM: arm64: Restructure FGT register switching
KVM: arm64: nv: Add trap forwarding infrastructure
KVM: arm64: nv: Add trap forwarding for HCR_EL2
KVM: arm64: nv: Expose FEAT_EVT to nested guests
KVM: arm64: nv: Add trap forwarding for MDCR_EL2
KVM: arm64: nv: Add trap forwarding for CNTHCTL_EL2
KVM: arm64: nv: Add trap forwarding for HFGxTR_EL2
KVM: arm64: nv: Add trap forwarding for HFGITR_EL2
KVM: arm64: nv: Add trap forwarding for HDFGxTR_EL2
KVM: arm64: nv: Add SVC trap forwarding
KVM: arm64: nv: Add switching support for HFGxTR/HDFGxTR
KVM: arm64: nv: Expose FGT to nested guests
KVM: arm64: Move HCRX_EL2 switch to load/put on VHE systems
KVM: arm64: nv: Add support for HCRX_EL2

Mark Brown (1):
arm64: Add feature detection for fine grained traps

 arch/arm64/include/asm/kvm_arm.h        |   50 +
 arch/arm64/include/asm/kvm_host.h       |    7 +
 arch/arm64/include/asm/kvm_nested.h     |    2 +
 arch/arm64/include/asm/sysreg.h         |  270 +++-
 arch/arm64/kernel/cpufeature.c          |    7 +
 arch/arm64/kvm/arm.c                    |    4 +
 arch/arm64/kvm/emulate-nested.c         | 1731 +++++++++++++++++++++++
 arch/arm64/kvm/handle_exit.c            |   12 +
 arch/arm64/kvm/hyp/include/hyp/switch.h |  127 +-
 arch/arm64/kvm/nested.c                 |   11 +-
 arch/arm64/kvm/sys_regs.c               |   15 +
 arch/arm64/kvm/trace_arm.h              |   19 +
 arch/arm64/tools/cpucaps                |    1 +
 arch/arm64/tools/sysreg                 |  129 ++
 14 files changed, 2345 insertions(+), 40 deletions(-)
Signature · 1 line
2.34.1
author maintainer
#1 lore raw

PATCH v2 1/26 arm64: Add missing VA CMO encodings

Add the missing VA-based CMOs encodings.

---
 arch/arm64/include/asm/sysreg.h | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)
arch/arm64/include/asm/sysreg.h +26 -0
index b481935e9314..85447e68951a 100644
@@ -124,6 +124,32 @@
 #define SYS_DC_CIGSW			sys_insn(1, 0, 7, 14, 4)
 #define SYS_DC_CIGDSW			sys_insn(1, 0, 7, 14, 6)
 
+#define SYS_IC_IALLUIS			sys_insn(1, 0, 7, 1, 0)
+#define SYS_IC_IALLU			sys_insn(1, 0, 7, 5, 0)
+#define SYS_IC_IVAU			sys_insn(1, 3, 7, 5, 1)
+
+#define SYS_DC_IVAC			sys_insn(1, 0, 7, 6, 1)
+#define SYS_DC_IGVAC			sys_insn(1, 0, 7, 6, 3)
+#define SYS_DC_IGDVAC			sys_insn(1, 0, 7, 6, 5)
+
+#define SYS_DC_CVAC			sys_insn(1, 3, 7, 10, 1)
+#define SYS_DC_CGVAC			sys_insn(1, 3, 7, 10, 3)
+#define SYS_DC_CGDVAC			sys_insn(1, 3, 7, 10, 5)
+
+#define SYS_DC_CVAU			sys_insn(1, 3, 7, 11, 1)
+
+#define SYS_DC_CVAP			sys_insn(1, 3, 7, 12, 1)
+#define SYS_DC_CGVAP			sys_insn(1, 3, 7, 12, 3)
+#define SYS_DC_CGDVAP			sys_insn(1, 3, 7, 12, 5)
+
+#define SYS_DC_CVADP			sys_insn(1, 3, 7, 13, 1)
+#define SYS_DC_CGVADP			sys_insn(1, 3, 7, 13, 3)
+#define SYS_DC_CGDVADP			sys_insn(1, 3, 7, 13, 5)
+
+#define SYS_DC_CIVAC			sys_insn(1, 3, 7, 14, 1)
+#define SYS_DC_CIGVAC			sys_insn(1, 3, 7, 14, 3)
+#define SYS_DC_CIGDVAC			sys_insn(1, 3, 7, 14, 5)
+
 /*
  * Automatically generated definitions for system registers, the
  * manual encodings below are in the process of being converted to
Signature · 1 line
2.34.1
Reviewed-by on v2 1/26
#2 lore raw

Re: [PATCH v2 01/26] arm64: Add missing VA CMO encodings

Hi Marc,

Quoted 34 lines · Marc Zyngier

Add the missing VA-based CMOs encodings.

---
arch/arm64/include/asm/sysreg.h | 26 ++++++++++++++++++++++++++

1 file changed, 26 insertions(+)
arch/arm64/include/asm/sysreg.h +16 -0
index b481935e9314..85447e68951a 100644
@@ -124,6 +124,32 @@
 #define SYS_DC_CIGSW sys_insn(1, 0, 7, 14, 4)
 #define SYS_DC_CIGDSW sys_insn(1, 0, 7, 14, 6)
 
+#define SYS_IC_IALLUIS sys_insn(1, 0, 7, 1, 0)
+#define SYS_IC_IALLU sys_insn(1, 0, 7, 5, 0)
+#define SYS_IC_IVAU sys_insn(1, 3, 7, 5, 1)
+
+#define SYS_DC_IVAC sys_insn(1, 0, 7, 6, 1)
+#define SYS_DC_IGVAC sys_insn(1, 0, 7, 6, 3)
+#define SYS_DC_IGDVAC sys_insn(1, 0, 7, 6, 5)
+
+#define SYS_DC_CVAC sys_insn(1, 3, 7, 10, 1)
+#define SYS_DC_CGVAC sys_insn(1, 3, 7, 10, 3)
+#define SYS_DC_CGDVAC sys_insn(1, 3, 7, 10, 5)
+
+#define SYS_DC_CVAU sys_insn(1, 3, 7, 11, 1)
+
+#define SYS_DC_CVAP sys_insn(1, 3, 7, 12, 1)
+#define SYS_DC_CGVAP sys_insn(1, 3, 7, 12, 3)

arch/arm64/include/asm/sysreg.h line 152

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define SYS_DC_CGDVAP sys_insn(1, 3, 7, 12, 5)
+
+#define SYS_DC_CVADP sys_insn(1, 3, 7, 13, 1)
+#define SYS_DC_CGVADP sys_insn(1, 3, 7, 13, 3)
+#define SYS_DC_CGDVADP sys_insn(1, 3, 7, 13, 5)
+
+#define SYS_DC_CIVAC sys_insn(1, 3, 7, 14, 1)
+#define SYS_DC_CIGVAC sys_insn(1, 3, 7, 14, 3)
+#define SYS_DC_CIGDVAC sys_insn(1, 3, 7, 14, 5)
+

Thanks
Miguel

Quoted 6 lines
arch/arm64/include/asm/sysreg.h +0 -0
 /*
 * Automatically generated definitions for system registers, the
 * manual encodings below are in the process of being converted to
Signature · 1 line
2.34.1
author maintainer
#5 lore raw

PATCH v2 2/26 arm64: Add missing ERX*_EL1 encodings

In reply to Marc Zyngier

We only describe a few of the ERX*_EL1 registers. Add the missing ones (ERXPFGF_EL1, ERXPFGCTL_EL1, ERXPFGCDN_EL1, ERXMISC2_EL1 and ERXMISC3_EL1).

---
 arch/arm64/include/asm/sysreg.h | 5 +++++
 1 file changed, 5 insertions(+)
arch/arm64/include/asm/sysreg.h +5 -0
index 85447e68951a..ed2739897859 100644
@@ -229,8 +229,13 @@
 #define SYS_ERXCTLR_EL1			sys_reg(3, 0, 5, 4, 1)
 #define SYS_ERXSTATUS_EL1		sys_reg(3, 0, 5, 4, 2)
 #define SYS_ERXADDR_EL1			sys_reg(3, 0, 5, 4, 3)
+#define SYS_ERXPFGF_EL1			sys_reg(3, 0, 5, 4, 4)
+#define SYS_ERXPFGCTL_EL1		sys_reg(3, 0, 5, 4, 5)
+#define SYS_ERXPFGCDN_EL1		sys_reg(3, 0, 5, 4, 6)
 #define SYS_ERXMISC0_EL1		sys_reg(3, 0, 5, 5, 0)
 #define SYS_ERXMISC1_EL1		sys_reg(3, 0, 5, 5, 1)
+#define SYS_ERXMISC2_EL1		sys_reg(3, 0, 5, 5, 2)
+#define SYS_ERXMISC3_EL1		sys_reg(3, 0, 5, 5, 3)
 #define SYS_TFSR_EL1			sys_reg(3, 0, 5, 6, 0)
 #define SYS_TFSRE0_EL1			sys_reg(3, 0, 5, 6, 1)
 
Signature · 1 line
2.34.1
Reviewed-by on v2 2/26
#6 lore raw

Re: [PATCH v2 02/26] arm64: Add missing ERX*_EL1 encodings

Hi Marc,

Quoted 20 lines · Marc Zyngier

We only describe a few of the ERX*_EL1 registers. Add the missing ones (ERXPFGF_EL1, ERXPFGCTL_EL1, ERXPFGCDN_EL1, ERXMISC2_EL1 and ERXMISC3_EL1).

---
arch/arm64/include/asm/sysreg.h | 5 +++++

1 file changed, 5 insertions(+)
arch/arm64/include/asm/sysreg.h +0 -0
index 85447e68951a..ed2739897859 100644
@@ -229,8 +229,13 @@
 #define SYS_ERXCTLR_EL1 sys_reg(3, 0, 5, 4, 1)
 #define SYS_ERXSTATUS_EL1 sys_reg(3, 0, 5, 4, 2)
 #define SYS_ERXADDR_EL1 sys_reg(3, 0, 5, 4, 3)

arch/arm64/include/asm/sysreg.h line 241

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +5 -0
+#define SYS_ERXPFGF_EL1 sys_reg(3, 0, 5, 4, 4)
+#define SYS_ERXPFGCTL_EL1 sys_reg(3, 0, 5, 4, 5)
+#define SYS_ERXPFGCDN_EL1 sys_reg(3, 0, 5, 4, 6)
 #define SYS_ERXMISC0_EL1 sys_reg(3, 0, 5, 5, 0)
 #define SYS_ERXMISC1_EL1 sys_reg(3, 0, 5, 5, 1)
+#define SYS_ERXMISC2_EL1 sys_reg(3, 0, 5, 5, 2)
+#define SYS_ERXMISC3_EL1 sys_reg(3, 0, 5, 5, 3)
 #define SYS_TFSR_EL1 sys_reg(3, 0, 5, 6, 0)
 #define SYS_TFSRE0_EL1 sys_reg(3, 0, 5, 6, 1)
 

Thanks
Miguel

Quoted 4 lines
Signature · 1 line
2.34.1
author maintainer
#9 lore raw

PATCH v2 3/26 arm64: Add missing DC ZVA/GVA/GZVA encodings

In reply to Marc Zyngier

Add the missing DC *VA encodings.

---
 arch/arm64/include/asm/sysreg.h | 5 +++++
 1 file changed, 5 insertions(+)
arch/arm64/include/asm/sysreg.h +5 -0
index ed2739897859..5084add86897 100644
@@ -150,6 +150,11 @@
 #define SYS_DC_CIGVAC			sys_insn(1, 3, 7, 14, 3)
 #define SYS_DC_CIGDVAC			sys_insn(1, 3, 7, 14, 5)
 
+/* Data cache zero operations */
+#define SYS_DC_ZVA			sys_insn(1, 3, 7, 4, 1)
+#define SYS_DC_GVA			sys_insn(1, 3, 7, 4, 3)
+#define SYS_DC_GZVA			sys_insn(1, 3, 7, 4, 4)
+
 /*
  * Automatically generated definitions for system registers, the
  * manual encodings below are in the process of being converted to
Signature · 1 line
2.34.1
Reviewed-by on v2 3/26
#10 lore raw

Re: [PATCH v2 03/26] arm64: Add missing DC ZVA/GVA/GZVA encodings

Hi Marc,

Quoted 15 lines · Marc Zyngier

Add the missing DC *VA encodings.

---
arch/arm64/include/asm/sysreg.h | 5 +++++

1 file changed, 5 insertions(+)
arch/arm64/include/asm/sysreg.h +0 -0
index ed2739897859..5084add86897 100644

arch/arm64/include/asm/sysreg.h line 157

Quoted 8 lines
arch/arm64/include/asm/sysreg.h +5 -0
@@ -150,6 +150,11 @@
 #define SYS_DC_CIGVAC sys_insn(1, 3, 7, 14, 3)
 #define SYS_DC_CIGDVAC sys_insn(1, 3, 7, 14, 5)
 
+/* Data cache zero operations */
+#define SYS_DC_ZVA sys_insn(1, 3, 7, 4, 1)
+#define SYS_DC_GVA sys_insn(1, 3, 7, 4, 3)
+#define SYS_DC_GZVA sys_insn(1, 3, 7, 4, 4)
+

Thanks
Miguel

Quoted 6 lines
arch/arm64/include/asm/sysreg.h +0 -0
 /*
 * Automatically generated definitions for system registers, the
 * manual encodings below are in the process of being converted to
Signature · 1 line
2.34.1
author maintainer
#13 lore raw

PATCH v2 4/26 arm64: Add TLBI operation encodings

In reply to Marc Zyngier

Add all the TLBI encodings that are usable from Non-Secure.

---
 arch/arm64/include/asm/sysreg.h | 128 ++++++++++++++++++++++++++++++++
 1 file changed, 128 insertions(+)
arch/arm64/include/asm/sysreg.h +128 -0
index 5084add86897..72e18480ce62 100644
@@ -514,6 +514,134 @@
 
 #define SYS_SP_EL2			sys_reg(3, 6,  4, 1, 0)
 
+/* TLBI instructions */
+#define OP_TLBI_VMALLE1OS		sys_insn(1, 0, 8, 1, 0)
+#define OP_TLBI_VAE1OS			sys_insn(1, 0, 8, 1, 1)
+#define OP_TLBI_ASIDE1OS		sys_insn(1, 0, 8, 1, 2)
+#define OP_TLBI_VAAE1OS			sys_insn(1, 0, 8, 1, 3)
+#define OP_TLBI_VALE1OS			sys_insn(1, 0, 8, 1, 5)
+#define OP_TLBI_VAALE1OS		sys_insn(1, 0, 8, 1, 7)
+#define OP_TLBI_RVAE1IS			sys_insn(1, 0, 8, 2, 1)
+#define OP_TLBI_RVAAE1IS		sys_insn(1, 0, 8, 2, 3)
+#define OP_TLBI_RVALE1IS		sys_insn(1, 0, 8, 2, 5)
+#define OP_TLBI_RVAALE1IS		sys_insn(1, 0, 8, 2, 7)
+#define OP_TLBI_VMALLE1IS		sys_insn(1, 0, 8, 3, 0)
+#define OP_TLBI_VAE1IS			sys_insn(1, 0, 8, 3, 1)
+#define OP_TLBI_ASIDE1IS		sys_insn(1, 0, 8, 3, 2)
+#define OP_TLBI_VAAE1IS			sys_insn(1, 0, 8, 3, 3)
+#define OP_TLBI_VALE1IS			sys_insn(1, 0, 8, 3, 5)
+#define OP_TLBI_VAALE1IS		sys_insn(1, 0, 8, 3, 7)
+#define OP_TLBI_RVAE1OS			sys_insn(1, 0, 8, 5, 1)
+#define OP_TLBI_RVAAE1OS		sys_insn(1, 0, 8, 5, 3)
+#define OP_TLBI_RVALE1OS		sys_insn(1, 0, 8, 5, 5)
+#define OP_TLBI_RVAALE1OS		sys_insn(1, 0, 8, 5, 7)
+#define OP_TLBI_RVAE1			sys_insn(1, 0, 8, 6, 1)
+#define OP_TLBI_RVAAE1			sys_insn(1, 0, 8, 6, 3)
+#define OP_TLBI_RVALE1			sys_insn(1, 0, 8, 6, 5)
+#define OP_TLBI_RVAALE1			sys_insn(1, 0, 8, 6, 7)
+#define OP_TLBI_VMALLE1			sys_insn(1, 0, 8, 7, 0)
+#define OP_TLBI_VAE1			sys_insn(1, 0, 8, 7, 1)
+#define OP_TLBI_ASIDE1			sys_insn(1, 0, 8, 7, 2)
+#define OP_TLBI_VAAE1			sys_insn(1, 0, 8, 7, 3)
+#define OP_TLBI_VALE1			sys_insn(1, 0, 8, 7, 5)
+#define OP_TLBI_VAALE1			sys_insn(1, 0, 8, 7, 7)
+#define OP_TLBI_VMALLE1OSNXS		sys_insn(1, 0, 9, 1, 0)
+#define OP_TLBI_VAE1OSNXS		sys_insn(1, 0, 9, 1, 1)
+#define OP_TLBI_ASIDE1OSNXS		sys_insn(1, 0, 9, 1, 2)
+#define OP_TLBI_VAAE1OSNXS		sys_insn(1, 0, 9, 1, 3)
+#define OP_TLBI_VALE1OSNXS		sys_insn(1, 0, 9, 1, 5)
+#define OP_TLBI_VAALE1OSNXS		sys_insn(1, 0, 9, 1, 7)
+#define OP_TLBI_RVAE1ISNXS		sys_insn(1, 0, 9, 2, 1)
+#define OP_TLBI_RVAAE1ISNXS		sys_insn(1, 0, 9, 2, 3)
+#define OP_TLBI_RVALE1ISNXS		sys_insn(1, 0, 9, 2, 5)
+#define OP_TLBI_RVAALE1ISNXS		sys_insn(1, 0, 9, 2, 7)
+#define OP_TLBI_VMALLE1ISNXS		sys_insn(1, 0, 9, 3, 0)
+#define OP_TLBI_VAE1ISNXS		sys_insn(1, 0, 9, 3, 1)
+#define OP_TLBI_ASIDE1ISNXS		sys_insn(1, 0, 9, 3, 2)
+#define OP_TLBI_VAAE1ISNXS		sys_insn(1, 0, 9, 3, 3)
+#define OP_TLBI_VALE1ISNXS		sys_insn(1, 0, 9, 3, 5)
+#define OP_TLBI_VAALE1ISNXS		sys_insn(1, 0, 9, 3, 7)
+#define OP_TLBI_RVAE1OSNXS		sys_insn(1, 0, 9, 5, 1)
+#define OP_TLBI_RVAAE1OSNXS		sys_insn(1, 0, 9, 5, 3)
+#define OP_TLBI_RVALE1OSNXS		sys_insn(1, 0, 9, 5, 5)
+#define OP_TLBI_RVAALE1OSNXS		sys_insn(1, 0, 9, 5, 7)
+#define OP_TLBI_RVAE1NXS		sys_insn(1, 0, 9, 6, 1)
+#define OP_TLBI_RVAAE1NXS		sys_insn(1, 0, 9, 6, 3)
+#define OP_TLBI_RVALE1NXS		sys_insn(1, 0, 9, 6, 5)
+#define OP_TLBI_RVAALE1NXS		sys_insn(1, 0, 9, 6, 7)
+#define OP_TLBI_VMALLE1NXS		sys_insn(1, 0, 9, 7, 0)
+#define OP_TLBI_VAE1NXS			sys_insn(1, 0, 9, 7, 1)
+#define OP_TLBI_ASIDE1NXS		sys_insn(1, 0, 9, 7, 2)
+#define OP_TLBI_VAAE1NXS		sys_insn(1, 0, 9, 7, 3)
+#define OP_TLBI_VALE1NXS		sys_insn(1, 0, 9, 7, 5)
+#define OP_TLBI_VAALE1NXS		sys_insn(1, 0, 9, 7, 7)
+#define OP_TLBI_IPAS2E1IS		sys_insn(1, 4, 8, 0, 1)
+#define OP_TLBI_RIPAS2E1IS		sys_insn(1, 4, 8, 0, 2)
+#define OP_TLBI_IPAS2LE1IS		sys_insn(1, 4, 8, 0, 5)
+#define OP_TLBI_RIPAS2LE1IS		sys_insn(1, 4, 8, 0, 6)
+#define OP_TLBI_ALLE2OS			sys_insn(1, 4, 8, 1, 0)
+#define OP_TLBI_VAE2OS			sys_insn(1, 4, 8, 1, 1)
+#define OP_TLBI_ALLE1OS			sys_insn(1, 4, 8, 1, 4)
+#define OP_TLBI_VALE2OS			sys_insn(1, 4, 8, 1, 5)
+#define OP_TLBI_VMALLS12E1OS		sys_insn(1, 4, 8, 1, 6)
+#define OP_TLBI_RVAE2IS			sys_insn(1, 4, 8, 2, 1)
+#define OP_TLBI_RVALE2IS		sys_insn(1, 4, 8, 2, 5)
+#define OP_TLBI_ALLE2IS			sys_insn(1, 4, 8, 3, 0)
+#define OP_TLBI_VAE2IS			sys_insn(1, 4, 8, 3, 1)
+#define OP_TLBI_ALLE1IS			sys_insn(1, 4, 8, 3, 4)
+#define OP_TLBI_VALE2IS			sys_insn(1, 4, 8, 3, 5)
+#define OP_TLBI_VMALLS12E1IS		sys_insn(1, 4, 8, 3, 6)
+#define OP_TLBI_IPAS2E1OS		sys_insn(1, 4, 8, 4, 0)
+#define OP_TLBI_IPAS2E1			sys_insn(1, 4, 8, 4, 1)
+#define OP_TLBI_RIPAS2E1		sys_insn(1, 4, 8, 4, 2)
+#define OP_TLBI_RIPAS2E1OS		sys_insn(1, 4, 8, 4, 3)
+#define OP_TLBI_IPAS2LE1OS		sys_insn(1, 4, 8, 4, 4)
+#define OP_TLBI_IPAS2LE1		sys_insn(1, 4, 8, 4, 5)
+#define OP_TLBI_RIPAS2LE1		sys_insn(1, 4, 8, 4, 6)
+#define OP_TLBI_RIPAS2LE1OS		sys_insn(1, 4, 8, 4, 7)
+#define OP_TLBI_RVAE2OS			sys_insn(1, 4, 8, 5, 1)
+#define OP_TLBI_RVALE2OS		sys_insn(1, 4, 8, 5, 5)
+#define OP_TLBI_RVAE2			sys_insn(1, 4, 8, 6, 1)
+#define OP_TLBI_RVALE2			sys_insn(1, 4, 8, 6, 5)
+#define OP_TLBI_ALLE2			sys_insn(1, 4, 8, 7, 0)
+#define OP_TLBI_VAE2			sys_insn(1, 4, 8, 7, 1)
+#define OP_TLBI_ALLE1			sys_insn(1, 4, 8, 7, 4)
+#define OP_TLBI_VALE2			sys_insn(1, 4, 8, 7, 5)
+#define OP_TLBI_VMALLS12E1		sys_insn(1, 4, 8, 7, 6)
+#define OP_TLBI_IPAS2E1ISNXS		sys_insn(1, 4, 9, 0, 1)
+#define OP_TLBI_RIPAS2E1ISNXS		sys_insn(1, 4, 9, 0, 2)
+#define OP_TLBI_IPAS2LE1ISNXS		sys_insn(1, 4, 9, 0, 5)
+#define OP_TLBI_RIPAS2LE1ISNXS		sys_insn(1, 4, 9, 0, 6)
+#define OP_TLBI_ALLE2OSNXS		sys_insn(1, 4, 9, 1, 0)
+#define OP_TLBI_VAE2OSNXS		sys_insn(1, 4, 9, 1, 1)
+#define OP_TLBI_ALLE1OSNXS		sys_insn(1, 4, 9, 1, 4)
+#define OP_TLBI_VALE2OSNXS		sys_insn(1, 4, 9, 1, 5)
+#define OP_TLBI_VMALLS12E1OSNXS		sys_insn(1, 4, 9, 1, 6)
+#define OP_TLBI_RVAE2ISNXS		sys_insn(1, 4, 9, 2, 1)
+#define OP_TLBI_RVALE2ISNXS		sys_insn(1, 4, 9, 2, 5)
+#define OP_TLBI_ALLE2ISNXS		sys_insn(1, 4, 9, 3, 0)
+#define OP_TLBI_VAE2ISNXS		sys_insn(1, 4, 9, 3, 1)
+#define OP_TLBI_ALLE1ISNXS		sys_insn(1, 4, 9, 3, 4)
+#define OP_TLBI_VALE2ISNXS		sys_insn(1, 4, 9, 3, 5)
+#define OP_TLBI_VMALLS12E1ISNXS		sys_insn(1, 4, 9, 3, 6)
+#define OP_TLBI_IPAS2E1OSNXS		sys_insn(1, 4, 9, 4, 0)
+#define OP_TLBI_IPAS2E1NXS		sys_insn(1, 4, 9, 4, 1)
+#define OP_TLBI_RIPAS2E1NXS		sys_insn(1, 4, 9, 4, 2)
+#define OP_TLBI_RIPAS2E1OSNXS		sys_insn(1, 4, 9, 4, 3)
+#define OP_TLBI_IPAS2LE1OSNXS		sys_insn(1, 4, 9, 4, 4)
+#define OP_TLBI_IPAS2LE1NXS		sys_insn(1, 4, 9, 4, 5)
+#define OP_TLBI_RIPAS2LE1NXS		sys_insn(1, 4, 9, 4, 6)
+#define OP_TLBI_RIPAS2LE1OSNXS		sys_insn(1, 4, 9, 4, 7)
+#define OP_TLBI_RVAE2OSNXS		sys_insn(1, 4, 9, 5, 1)
+#define OP_TLBI_RVALE2OSNXS		sys_insn(1, 4, 9, 5, 5)
+#define OP_TLBI_RVAE2NXS		sys_insn(1, 4, 9, 6, 1)
+#define OP_TLBI_RVALE2NXS		sys_insn(1, 4, 9, 6, 5)
+#define OP_TLBI_ALLE2NXS		sys_insn(1, 4, 9, 7, 0)
+#define OP_TLBI_VAE2NXS			sys_insn(1, 4, 9, 7, 1)
+#define OP_TLBI_ALLE1NXS		sys_insn(1, 4, 9, 7, 4)
+#define OP_TLBI_VALE2NXS		sys_insn(1, 4, 9, 7, 5)
+#define OP_TLBI_VMALLS12E1NXS		sys_insn(1, 4, 9, 7, 6)
+
 /* Common SCTLR_ELx flags. */
 #define SCTLR_ELx_ENTP2	(BIT(60))
 #define SCTLR_ELx_DSSBS	(BIT(44))
Signature · 1 line
2.34.1
author maintainer
#15 lore raw

Re: [PATCH v2 04/26] arm64: Add TLBI operation encodings

Quoted 9 lines · Catalin Marinas
Quoted 4 lines · Marc Zyngier

Add all the TLBI encodings that are usable from Non-Secure.

I trust you they are all correct, haven't checked against the Arm ARM ;).

The XML is my forever friend... ;-)

Quoted 1 line

Thanks!

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
Reviewed-by on v2 4/26
#16 lore raw

Re: [PATCH v2 04/26] arm64: Add TLBI operation encodings

In reply to Marc Zyngier

Hi Marc,

Quoted 136 lines · Marc Zyngier

Add all the TLBI encodings that are usable from Non-Secure.

---
arch/arm64/include/asm/sysreg.h | 128 ++++++++++++++++++++++++++++++++

1 file changed, 128 insertions(+)
arch/arm64/include/asm/sysreg.h +118 -0
index 5084add86897..72e18480ce62 100644
@@ -514,6 +514,134 @@
 
 #define SYS_SP_EL2 sys_reg(3, 6,  4, 1, 0)
 
+/* TLBI instructions */
+#define OP_TLBI_VMALLE1OS sys_insn(1, 0, 8, 1, 0)
+#define OP_TLBI_VAE1OS sys_insn(1, 0, 8, 1, 1)
+#define OP_TLBI_ASIDE1OS sys_insn(1, 0, 8, 1, 2)
+#define OP_TLBI_VAAE1OS sys_insn(1, 0, 8, 1, 3)
+#define OP_TLBI_VALE1OS sys_insn(1, 0, 8, 1, 5)
+#define OP_TLBI_VAALE1OS sys_insn(1, 0, 8, 1, 7)
+#define OP_TLBI_RVAE1IS sys_insn(1, 0, 8, 2, 1)
+#define OP_TLBI_RVAAE1IS sys_insn(1, 0, 8, 2, 3)
+#define OP_TLBI_RVALE1IS sys_insn(1, 0, 8, 2, 5)
+#define OP_TLBI_RVAALE1IS sys_insn(1, 0, 8, 2, 7)
+#define OP_TLBI_VMALLE1IS sys_insn(1, 0, 8, 3, 0)
+#define OP_TLBI_VAE1IS sys_insn(1, 0, 8, 3, 1)
+#define OP_TLBI_ASIDE1IS sys_insn(1, 0, 8, 3, 2)
+#define OP_TLBI_VAAE1IS sys_insn(1, 0, 8, 3, 3)
+#define OP_TLBI_VALE1IS sys_insn(1, 0, 8, 3, 5)
+#define OP_TLBI_VAALE1IS sys_insn(1, 0, 8, 3, 7)
+#define OP_TLBI_RVAE1OS sys_insn(1, 0, 8, 5, 1)
+#define OP_TLBI_RVAAE1OS sys_insn(1, 0, 8, 5, 3)
+#define OP_TLBI_RVALE1OS sys_insn(1, 0, 8, 5, 5)
+#define OP_TLBI_RVAALE1OS sys_insn(1, 0, 8, 5, 7)
+#define OP_TLBI_RVAE1 sys_insn(1, 0, 8, 6, 1)
+#define OP_TLBI_RVAAE1 sys_insn(1, 0, 8, 6, 3)
+#define OP_TLBI_RVALE1 sys_insn(1, 0, 8, 6, 5)
+#define OP_TLBI_RVAALE1 sys_insn(1, 0, 8, 6, 7)
+#define OP_TLBI_VMALLE1 sys_insn(1, 0, 8, 7, 0)
+#define OP_TLBI_VAE1 sys_insn(1, 0, 8, 7, 1)
+#define OP_TLBI_ASIDE1 sys_insn(1, 0, 8, 7, 2)
+#define OP_TLBI_VAAE1 sys_insn(1, 0, 8, 7, 3)
+#define OP_TLBI_VALE1 sys_insn(1, 0, 8, 7, 5)
+#define OP_TLBI_VAALE1 sys_insn(1, 0, 8, 7, 7)
+#define OP_TLBI_VMALLE1OSNXS sys_insn(1, 0, 9, 1, 0)
+#define OP_TLBI_VAE1OSNXS sys_insn(1, 0, 9, 1, 1)
+#define OP_TLBI_ASIDE1OSNXS sys_insn(1, 0, 9, 1, 2)
+#define OP_TLBI_VAAE1OSNXS sys_insn(1, 0, 9, 1, 3)
+#define OP_TLBI_VALE1OSNXS sys_insn(1, 0, 9, 1, 5)
+#define OP_TLBI_VAALE1OSNXS sys_insn(1, 0, 9, 1, 7)
+#define OP_TLBI_RVAE1ISNXS sys_insn(1, 0, 9, 2, 1)
+#define OP_TLBI_RVAAE1ISNXS sys_insn(1, 0, 9, 2, 3)
+#define OP_TLBI_RVALE1ISNXS sys_insn(1, 0, 9, 2, 5)
+#define OP_TLBI_RVAALE1ISNXS sys_insn(1, 0, 9, 2, 7)
+#define OP_TLBI_VMALLE1ISNXS sys_insn(1, 0, 9, 3, 0)
+#define OP_TLBI_VAE1ISNXS sys_insn(1, 0, 9, 3, 1)
+#define OP_TLBI_ASIDE1ISNXS sys_insn(1, 0, 9, 3, 2)
+#define OP_TLBI_VAAE1ISNXS sys_insn(1, 0, 9, 3, 3)
+#define OP_TLBI_VALE1ISNXS sys_insn(1, 0, 9, 3, 5)
+#define OP_TLBI_VAALE1ISNXS sys_insn(1, 0, 9, 3, 7)
+#define OP_TLBI_RVAE1OSNXS sys_insn(1, 0, 9, 5, 1)
+#define OP_TLBI_RVAAE1OSNXS sys_insn(1, 0, 9, 5, 3)
+#define OP_TLBI_RVALE1OSNXS sys_insn(1, 0, 9, 5, 5)
+#define OP_TLBI_RVAALE1OSNXS sys_insn(1, 0, 9, 5, 7)
+#define OP_TLBI_RVAE1NXS sys_insn(1, 0, 9, 6, 1)
+#define OP_TLBI_RVAAE1NXS sys_insn(1, 0, 9, 6, 3)
+#define OP_TLBI_RVALE1NXS sys_insn(1, 0, 9, 6, 5)
+#define OP_TLBI_RVAALE1NXS sys_insn(1, 0, 9, 6, 7)
+#define OP_TLBI_VMALLE1NXS sys_insn(1, 0, 9, 7, 0)
+#define OP_TLBI_VAE1NXS sys_insn(1, 0, 9, 7, 1)
+#define OP_TLBI_ASIDE1NXS sys_insn(1, 0, 9, 7, 2)
+#define OP_TLBI_VAAE1NXS sys_insn(1, 0, 9, 7, 3)
+#define OP_TLBI_VALE1NXS sys_insn(1, 0, 9, 7, 5)
+#define OP_TLBI_VAALE1NXS sys_insn(1, 0, 9, 7, 7)
+#define OP_TLBI_IPAS2E1IS sys_insn(1, 4, 8, 0, 1)
+#define OP_TLBI_RIPAS2E1IS sys_insn(1, 4, 8, 0, 2)
+#define OP_TLBI_IPAS2LE1IS sys_insn(1, 4, 8, 0, 5)
+#define OP_TLBI_RIPAS2LE1IS sys_insn(1, 4, 8, 0, 6)
+#define OP_TLBI_ALLE2OS sys_insn(1, 4, 8, 1, 0)
+#define OP_TLBI_VAE2OS sys_insn(1, 4, 8, 1, 1)
+#define OP_TLBI_ALLE1OS sys_insn(1, 4, 8, 1, 4)
+#define OP_TLBI_VALE2OS sys_insn(1, 4, 8, 1, 5)
+#define OP_TLBI_VMALLS12E1OS sys_insn(1, 4, 8, 1, 6)
+#define OP_TLBI_RVAE2IS sys_insn(1, 4, 8, 2, 1)
+#define OP_TLBI_RVALE2IS sys_insn(1, 4, 8, 2, 5)
+#define OP_TLBI_ALLE2IS sys_insn(1, 4, 8, 3, 0)
+#define OP_TLBI_VAE2IS sys_insn(1, 4, 8, 3, 1)
+#define OP_TLBI_ALLE1IS sys_insn(1, 4, 8, 3, 4)
+#define OP_TLBI_VALE2IS sys_insn(1, 4, 8, 3, 5)
+#define OP_TLBI_VMALLS12E1IS sys_insn(1, 4, 8, 3, 6)
+#define OP_TLBI_IPAS2E1OS sys_insn(1, 4, 8, 4, 0)
+#define OP_TLBI_IPAS2E1 sys_insn(1, 4, 8, 4, 1)
+#define OP_TLBI_RIPAS2E1 sys_insn(1, 4, 8, 4, 2)
+#define OP_TLBI_RIPAS2E1OS sys_insn(1, 4, 8, 4, 3)
+#define OP_TLBI_IPAS2LE1OS sys_insn(1, 4, 8, 4, 4)
+#define OP_TLBI_IPAS2LE1 sys_insn(1, 4, 8, 4, 5)
+#define OP_TLBI_RIPAS2LE1 sys_insn(1, 4, 8, 4, 6)
+#define OP_TLBI_RIPAS2LE1OS sys_insn(1, 4, 8, 4, 7)
+#define OP_TLBI_RVAE2OS sys_insn(1, 4, 8, 5, 1)
+#define OP_TLBI_RVALE2OS sys_insn(1, 4, 8, 5, 5)
+#define OP_TLBI_RVAE2 sys_insn(1, 4, 8, 6, 1)
+#define OP_TLBI_RVALE2 sys_insn(1, 4, 8, 6, 5)
+#define OP_TLBI_ALLE2 sys_insn(1, 4, 8, 7, 0)
+#define OP_TLBI_VAE2 sys_insn(1, 4, 8, 7, 1)
+#define OP_TLBI_ALLE1 sys_insn(1, 4, 8, 7, 4)
+#define OP_TLBI_VALE2 sys_insn(1, 4, 8, 7, 5)
+#define OP_TLBI_VMALLS12E1 sys_insn(1, 4, 8, 7, 6)
+#define OP_TLBI_IPAS2E1ISNXS sys_insn(1, 4, 9, 0, 1)
+#define OP_TLBI_RIPAS2E1ISNXS sys_insn(1, 4, 9, 0, 2)
+#define OP_TLBI_IPAS2LE1ISNXS sys_insn(1, 4, 9, 0, 5)
+#define OP_TLBI_RIPAS2LE1ISNXS sys_insn(1, 4, 9, 0, 6)
+#define OP_TLBI_ALLE2OSNXS sys_insn(1, 4, 9, 1, 0)
+#define OP_TLBI_VAE2OSNXS sys_insn(1, 4, 9, 1, 1)
+#define OP_TLBI_ALLE1OSNXS sys_insn(1, 4, 9, 1, 4)
+#define OP_TLBI_VALE2OSNXS sys_insn(1, 4, 9, 1, 5)
+#define OP_TLBI_VMALLS12E1OSNXS sys_insn(1, 4, 9, 1, 6)
+#define OP_TLBI_RVAE2ISNXS sys_insn(1, 4, 9, 2, 1)
+#define OP_TLBI_RVALE2ISNXS sys_insn(1, 4, 9, 2, 5)
+#define OP_TLBI_ALLE2ISNXS sys_insn(1, 4, 9, 3, 0)
+#define OP_TLBI_VAE2ISNXS sys_insn(1, 4, 9, 3, 1)
+#define OP_TLBI_ALLE1ISNXS sys_insn(1, 4, 9, 3, 4)
+#define OP_TLBI_VALE2ISNXS sys_insn(1, 4, 9, 3, 5)
+#define OP_TLBI_VMALLS12E1ISNXS sys_insn(1, 4, 9, 3, 6)
+#define OP_TLBI_IPAS2E1OSNXS sys_insn(1, 4, 9, 4, 0)
+#define OP_TLBI_IPAS2E1NXS sys_insn(1, 4, 9, 4, 1)
+#define OP_TLBI_RIPAS2E1NXS sys_insn(1, 4, 9, 4, 2)
+#define OP_TLBI_RIPAS2E1OSNXS sys_insn(1, 4, 9, 4, 3)
+#define OP_TLBI_IPAS2LE1OSNXS sys_insn(1, 4, 9, 4, 4)
+#define OP_TLBI_IPAS2LE1NXS sys_insn(1, 4, 9, 4, 5)
+#define OP_TLBI_RIPAS2LE1NXS sys_insn(1, 4, 9, 4, 6)
+#define OP_TLBI_RIPAS2LE1OSNXS sys_insn(1, 4, 9, 4, 7)

arch/arm64/include/asm/sysreg.h line 644

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define OP_TLBI_RVAE2OSNXS sys_insn(1, 4, 9, 5, 1)
+#define OP_TLBI_RVALE2OSNXS sys_insn(1, 4, 9, 5, 5)
+#define OP_TLBI_RVAE2NXS sys_insn(1, 4, 9, 6, 1)
+#define OP_TLBI_RVALE2NXS sys_insn(1, 4, 9, 6, 5)
+#define OP_TLBI_ALLE2NXS sys_insn(1, 4, 9, 7, 0)
+#define OP_TLBI_VAE2NXS sys_insn(1, 4, 9, 7, 1)
+#define OP_TLBI_ALLE1NXS sys_insn(1, 4, 9, 7, 4)
+#define OP_TLBI_VALE2NXS sys_insn(1, 4, 9, 7, 5)
+#define OP_TLBI_VMALLS12E1NXS sys_insn(1, 4, 9, 7, 6)
+

As documented in DDI0487J.a.

FWIW, the remaining 34 TLBI instructions I’ve found in the same document that are not in this patch had access ‘UNDEFINED’ from Non-Secure.

Thanks
Miguel

Quoted 6 lines
arch/arm64/include/asm/sysreg.h +0 -0
 /* Common SCTLR_ELx flags. */
 #define SCTLR_ELx_ENTP2 (BIT(60))
 #define SCTLR_ELx_DSSBS (BIT(44))
Signature · 1 line
2.34.1
author maintainer
#18 lore raw

PATCH v2 5/26 arm64: Add AT operation encodings

In reply to Marc Zyngier

Add the encodings for the AT operation that are usable from NS.

---
 arch/arm64/include/asm/sysreg.h | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)
arch/arm64/include/asm/sysreg.h +17 -0
index 72e18480ce62..76289339b43b 100644
@@ -514,6 +514,23 @@
 
 #define SYS_SP_EL2			sys_reg(3, 6,  4, 1, 0)
 
+/* AT instructions */
+#define AT_Op0 1
+#define AT_CRn 7
+
+#define OP_AT_S1E1R	sys_insn(AT_Op0, 0, AT_CRn, 8, 0)
+#define OP_AT_S1E1W	sys_insn(AT_Op0, 0, AT_CRn, 8, 1)
+#define OP_AT_S1E0R	sys_insn(AT_Op0, 0, AT_CRn, 8, 2)
+#define OP_AT_S1E0W	sys_insn(AT_Op0, 0, AT_CRn, 8, 3)
+#define OP_AT_S1E1RP	sys_insn(AT_Op0, 0, AT_CRn, 9, 0)
+#define OP_AT_S1E1WP	sys_insn(AT_Op0, 0, AT_CRn, 9, 1)
+#define OP_AT_S1E2R	sys_insn(AT_Op0, 4, AT_CRn, 8, 0)
+#define OP_AT_S1E2W	sys_insn(AT_Op0, 4, AT_CRn, 8, 1)
+#define OP_AT_S12E1R	sys_insn(AT_Op0, 4, AT_CRn, 8, 4)
+#define OP_AT_S12E1W	sys_insn(AT_Op0, 4, AT_CRn, 8, 5)
+#define OP_AT_S12E0R	sys_insn(AT_Op0, 4, AT_CRn, 8, 6)
+#define OP_AT_S12E0W	sys_insn(AT_Op0, 4, AT_CRn, 8, 7)
+
 /* TLBI instructions */
 #define OP_TLBI_VMALLE1OS		sys_insn(1, 0, 8, 1, 0)
 #define OP_TLBI_VAE1OS			sys_insn(1, 0, 8, 1, 1)
Signature · 1 line
2.34.1
Reviewed-by on v2 5/26
#20 lore raw

Re: [PATCH v2 05/26] arm64: Add AT operation encodings

In reply to Marc Zyngier

Hi Marc,

Quoted 25 lines · Marc Zyngier

Add the encodings for the AT operation that are usable from NS.

---
arch/arm64/include/asm/sysreg.h | 17 +++++++++++++++++

1 file changed, 17 insertions(+)
arch/arm64/include/asm/sysreg.h +7 -0
index 72e18480ce62..76289339b43b 100644
@@ -514,6 +514,23 @@
 
 #define SYS_SP_EL2 sys_reg(3, 6,  4, 1, 0)
 
+/* AT instructions */
+#define AT_Op0 1
+#define AT_CRn 7
+
+#define OP_AT_S1E1R sys_insn(AT_Op0, 0, AT_CRn, 8, 0)
+#define OP_AT_S1E1W sys_insn(AT_Op0, 0, AT_CRn, 8, 1)
+#define OP_AT_S1E0R sys_insn(AT_Op0, 0, AT_CRn, 8, 2)

arch/arm64/include/asm/sysreg.h line 533

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define OP_AT_S1E0W sys_insn(AT_Op0, 0, AT_CRn, 8, 3)
+#define OP_AT_S1E1RP sys_insn(AT_Op0, 0, AT_CRn, 9, 0)
+#define OP_AT_S1E1WP sys_insn(AT_Op0, 0, AT_CRn, 9, 1)
+#define OP_AT_S1E2R sys_insn(AT_Op0, 4, AT_CRn, 8, 0)
+#define OP_AT_S1E2W sys_insn(AT_Op0, 4, AT_CRn, 8, 1)
+#define OP_AT_S12E1R sys_insn(AT_Op0, 4, AT_CRn, 8, 4)
+#define OP_AT_S12E1W sys_insn(AT_Op0, 4, AT_CRn, 8, 5)
+#define OP_AT_S12E0R sys_insn(AT_Op0, 4, AT_CRn, 8, 6)
+#define OP_AT_S12E0W sys_insn(AT_Op0, 4, AT_CRn, 8, 7)
+

Thanks

Miguel

Quoted 6 lines
arch/arm64/include/asm/sysreg.h +0 -0
 /* TLBI instructions */
 #define OP_TLBI_VMALLE1OS sys_insn(1, 0, 8, 1, 0)
 #define OP_TLBI_VAE1OS sys_insn(1, 0, 8, 1, 1)
Signature · 1 line
2.34.1
author maintainer
#22 lore raw

PATCH v2 6/26 arm64: Add debug registers affected by HDFGxTR_EL2

In reply to Marc Zyngier

The HDFGxTR_EL2 registers trap a (huge) set of debug and trace related registers. Add their encodings (and only that, because we really don't care about what these registers actually do at this stage).

---
 arch/arm64/include/asm/sysreg.h | 78 +++++++++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)
arch/arm64/include/asm/sysreg.h +78 -0
index 76289339b43b..9dfd127be55a 100644
@@ -194,6 +194,84 @@
 #define SYS_DBGDTRTX_EL0		sys_reg(2, 3, 0, 5, 0)
 #define SYS_DBGVCR32_EL2		sys_reg(2, 4, 0, 7, 0)
 
+#define SYS_BRBINF_EL1(n)		sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 0))
+#define SYS_BRBINFINJ_EL1		sys_reg(2, 1, 9, 1, 0)
+#define SYS_BRBSRC_EL1(n)		sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 1))
+#define SYS_BRBSRCINJ_EL1		sys_reg(2, 1, 9, 1, 1)
+#define SYS_BRBTGT_EL1(n)		sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 2))
+#define SYS_BRBTGTINJ_EL1		sys_reg(2, 1, 9, 1, 2)
+#define SYS_BRBTS_EL1			sys_reg(2, 1, 9, 0, 2)
+
+#define SYS_BRBCR_EL1			sys_reg(2, 1, 9, 0, 0)
+#define SYS_BRBFCR_EL1			sys_reg(2, 1, 9, 0, 1)
+#define SYS_BRBIDR0_EL1			sys_reg(2, 1, 9, 2, 0)
+
+#define SYS_TRCITECR_EL1		sys_reg(3, 0, 1, 2, 3)
+#define SYS_TRCITECR_EL1		sys_reg(3, 0, 1, 2, 3)
+#define SYS_TRCACATR(m)			sys_reg(2, 1, 2, ((m & 7) << 1), (2 | (m >> 3)))
+#define SYS_TRCACVR(m)			sys_reg(2, 1, 2, ((m & 7) << 1), (0 | (m >> 3)))
+#define SYS_TRCAUTHSTATUS		sys_reg(2, 1, 7, 14, 6)
+#define SYS_TRCAUXCTLR			sys_reg(2, 1, 0, 6, 0)
+#define SYS_TRCBBCTLR			sys_reg(2, 1, 0, 15, 0)
+#define SYS_TRCCCCTLR			sys_reg(2, 1, 0, 14, 0)
+#define SYS_TRCCIDCCTLR0		sys_reg(2, 1, 3, 0, 2)
+#define SYS_TRCCIDCCTLR1		sys_reg(2, 1, 3, 1, 2)
+#define SYS_TRCCIDCVR(m)		sys_reg(2, 1, 3, ((m & 7) << 1), 0)
+#define SYS_TRCCLAIMCLR			sys_reg(2, 1, 7, 9, 6)
+#define SYS_TRCCLAIMSET			sys_reg(2, 1, 7, 8, 6)
+#define SYS_TRCCNTCTLR(m)		sys_reg(2, 1, 0, (4 | (m & 3)), 5)
+#define SYS_TRCCNTRLDVR(m)		sys_reg(2, 1, 0, (0 | (m & 3)), 5)
+#define SYS_TRCCNTVR(m)			sys_reg(2, 1, 0, (8 | (m & 3)), 5)
+#define SYS_TRCCONFIGR			sys_reg(2, 1, 0, 4, 0)
+#define SYS_TRCDEVARCH			sys_reg(2, 1, 7, 15, 6)
+#define SYS_TRCDEVID			sys_reg(2, 1, 7, 2, 7)
+#define SYS_TRCEVENTCTL0R		sys_reg(2, 1, 0, 8, 0)
+#define SYS_TRCEVENTCTL1R		sys_reg(2, 1, 0, 9, 0)
+#define SYS_TRCEXTINSELR(m)		sys_reg(2, 1, 0, (8 | (m & 3)), 4)
+#define SYS_TRCIDR0			sys_reg(2, 1, 0, 8, 7)
+#define SYS_TRCIDR10			sys_reg(2, 1, 0, 2, 6)
+#define SYS_TRCIDR11			sys_reg(2, 1, 0, 3, 6)
+#define SYS_TRCIDR12			sys_reg(2, 1, 0, 4, 6)
+#define SYS_TRCIDR13			sys_reg(2, 1, 0, 5, 6)
+#define SYS_TRCIDR1			sys_reg(2, 1, 0, 9, 7)
+#define SYS_TRCIDR2			sys_reg(2, 1, 0, 10, 7)
+#define SYS_TRCIDR3			sys_reg(2, 1, 0, 11, 7)
+#define SYS_TRCIDR4			sys_reg(2, 1, 0, 12, 7)
+#define SYS_TRCIDR5			sys_reg(2, 1, 0, 13, 7)
+#define SYS_TRCIDR6			sys_reg(2, 1, 0, 14, 7)
+#define SYS_TRCIDR7			sys_reg(2, 1, 0, 15, 7)
+#define SYS_TRCIDR8			sys_reg(2, 1, 0, 0, 6)
+#define SYS_TRCIDR9			sys_reg(2, 1, 0, 1, 6)
+#define SYS_TRCIMSPEC0			sys_reg(2, 1, 0, 0, 7)
+#define SYS_TRCIMSPEC(m)		sys_reg(2, 1, 0, (m & 7), 7)
+#define SYS_TRCITEEDCR			sys_reg(2, 1, 0, 2, 1)
+#define SYS_TRCOSLSR			sys_reg(2, 1, 1, 1, 4)
+#define SYS_TRCPRGCTLR			sys_reg(2, 1, 0, 1, 0)
+#define SYS_TRCQCTLR			sys_reg(2, 1, 0, 1, 1)
+#define SYS_TRCRSCTLR(m)		sys_reg(2, 1, 1, (m & 15), (0 | (m >> 4)))
+#define SYS_TRCRSR			sys_reg(2, 1, 0, 10, 0)
+#define SYS_TRCSEQEVR(m)		sys_reg(2, 1, 0, (m & 3), 4)
+#define SYS_TRCSEQRSTEVR		sys_reg(2, 1, 0, 6, 4)
+#define SYS_TRCSEQSTR			sys_reg(2, 1, 0, 7, 4)
+#define SYS_TRCSSCCR(m)			sys_reg(2, 1, 1, (m & 7), 2)
+#define SYS_TRCSSCSR(m)			sys_reg(2, 1, 1, (8 | (m & 7)), 2)
+#define SYS_TRCSSPCICR(m)		sys_reg(2, 1, 1, (m & 7), 3)
+#define SYS_TRCSTALLCTLR		sys_reg(2, 1, 0, 11, 0)
+#define SYS_TRCSTATR			sys_reg(2, 1, 0, 3, 0)
+#define SYS_TRCSYNCPR			sys_reg(2, 1, 0, 13, 0)
+#define SYS_TRCTRACEIDR			sys_reg(2, 1, 0, 0, 1)
+#define SYS_TRCTSCTLR			sys_reg(2, 1, 0, 12, 0)
+#define SYS_TRCVICTLR			sys_reg(2, 1, 0, 0, 2)
+#define SYS_TRCVIIECTLR			sys_reg(2, 1, 0, 1, 2)
+#define SYS_TRCVIPCSSCTLR		sys_reg(2, 1, 0, 3, 2)
+#define SYS_TRCVISSCTLR			sys_reg(2, 1, 0, 2, 2)
+#define SYS_TRCVMIDCCTLR0		sys_reg(2, 1, 3, 2, 2)
+#define SYS_TRCVMIDCCTLR1		sys_reg(2, 1, 3, 3, 2)
+#define SYS_TRCVMIDCVR(m)		sys_reg(2, 1, 3, ((m & 7) << 1), 1)
+
+/* ETM */
+#define SYS_TRCOSLAR			sys_reg(2, 1, 1, 0, 4)
+
 #define SYS_MIDR_EL1			sys_reg(3, 0, 0, 0, 0)
 #define SYS_MPIDR_EL1			sys_reg(3, 0, 0, 0, 5)
 #define SYS_REVIDR_EL1			sys_reg(3, 0, 0, 0, 6)
Signature · 1 line
2.34.1
maintainer Acked-by on v2 6/26
#23 lore raw

Re: [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

Quoted 6 lines · Marc Zyngier

The HDFGxTR_EL2 registers trap a (huge) set of debug and trace related registers. Add their encodings (and only that, because we really don't care about what these registers actually do at this stage).

#24 lore raw

Re: [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

In reply to Marc Zyngier

Hi Marc,

A few comments on this one, please see below.

Quoted 24 lines · Marc Zyngier

The HDFGxTR_EL2 registers trap a (huge) set of debug and trace related registers. Add their encodings (and only that, because we really don't care about what these registers actually do at this stage).

---
arch/arm64/include/asm/sysreg.h | 78 +++++++++++++++++++++++++++++++++

1 file changed, 78 insertions(+)
arch/arm64/include/asm/sysreg.h +4 -0
index 76289339b43b..9dfd127be55a 100644
@@ -194,6 +194,84 @@
 #define SYS_DBGDTRTX_EL0 sys_reg(2, 3, 0, 5, 0)
 #define SYS_DBGVCR32_EL2 sys_reg(2, 4, 0, 7, 0)
 
+#define SYS_BRBINF_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 0))
+#define SYS_BRBINFINJ_EL1 sys_reg(2, 1, 9, 1, 0)
+#define SYS_BRBSRC_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 1))
+#define SYS_BRBSRCINJ_EL1 sys_reg(2, 1, 9, 1, 1)

arch/arm64/include/asm/sysreg.h line 210

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define SYS_BRBTGT_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 2))
+#define SYS_BRBTGTINJ_EL1 sys_reg(2, 1, 9, 1, 2)
+#define SYS_BRBTS_EL1 sys_reg(2, 1, 9, 0, 2)
+
+#define SYS_BRBCR_EL1 sys_reg(2, 1, 9, 0, 0)
+#define SYS_BRBFCR_EL1 sys_reg(2, 1, 9, 0, 1)
+#define SYS_BRBIDR0_EL1 sys_reg(2, 1, 9, 2, 0)
+
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)

SYS_TRCITECR_EL1 shows up twice.

arch/arm64/include/asm/sysreg.h line 211

Quoted 1 line
arch/arm64/include/asm/sysreg.h +1 -0
+#define SYS_TRCACATR(m) sys_reg(2, 1, 2, ((m & 7) << 1), (2 | (m >> 3)))

Besides m’s restrictions it could be sanitised in op2 to consider only bit m[3]. Suggestion for op2: (2 | ((m & 8) >> 3)))

arch/arm64/include/asm/sysreg.h line 212

Quoted 1 line
arch/arm64/include/asm/sysreg.h +1 -0
+#define SYS_TRCACVR(m) sys_reg(2, 1, 2, ((m & 7) << 1), (0 | (m >> 3)))

Same for SYS_TRCACVR(m) op2: (0 | ((m & 8) >> 3) )

Quoted 24 lines
arch/arm64/include/asm/sysreg.h +24 -0
+#define SYS_TRCAUTHSTATUS sys_reg(2, 1, 7, 14, 6)
+#define SYS_TRCAUXCTLR sys_reg(2, 1, 0, 6, 0)
+#define SYS_TRCBBCTLR sys_reg(2, 1, 0, 15, 0)
+#define SYS_TRCCCCTLR sys_reg(2, 1, 0, 14, 0)
+#define SYS_TRCCIDCCTLR0 sys_reg(2, 1, 3, 0, 2)
+#define SYS_TRCCIDCCTLR1 sys_reg(2, 1, 3, 1, 2)
+#define SYS_TRCCIDCVR(m) sys_reg(2, 1, 3, ((m & 7) << 1), 0)
+#define SYS_TRCCLAIMCLR sys_reg(2, 1, 7, 9, 6)
+#define SYS_TRCCLAIMSET sys_reg(2, 1, 7, 8, 6)
+#define SYS_TRCCNTCTLR(m) sys_reg(2, 1, 0, (4 | (m & 3)), 5)
+#define SYS_TRCCNTRLDVR(m) sys_reg(2, 1, 0, (0 | (m & 3)), 5)
+#define SYS_TRCCNTVR(m) sys_reg(2, 1, 0, (8 | (m & 3)), 5)
+#define SYS_TRCCONFIGR sys_reg(2, 1, 0, 4, 0)
+#define SYS_TRCDEVARCH sys_reg(2, 1, 7, 15, 6)
+#define SYS_TRCDEVID sys_reg(2, 1, 7, 2, 7)
+#define SYS_TRCEVENTCTL0R sys_reg(2, 1, 0, 8, 0)
+#define SYS_TRCEVENTCTL1R sys_reg(2, 1, 0, 9, 0)
+#define SYS_TRCEXTINSELR(m) sys_reg(2, 1, 0, (8 | (m & 3)), 4)
+#define SYS_TRCIDR0 sys_reg(2, 1, 0, 8, 7)
+#define SYS_TRCIDR10 sys_reg(2, 1, 0, 2, 6)
+#define SYS_TRCIDR11 sys_reg(2, 1, 0, 3, 6)
+#define SYS_TRCIDR12 sys_reg(2, 1, 0, 4, 6)
+#define SYS_TRCIDR13 sys_reg(2, 1, 0, 5, 6)
+#define SYS_TRCIDR1 sys_reg(2, 1, 0, 9, 7)

arch/arm64/include/asm/sysreg.h line 246

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define SYS_TRCIDR2 sys_reg(2, 1, 0, 10, 7)
+#define SYS_TRCIDR3 sys_reg(2, 1, 0, 11, 7)
+#define SYS_TRCIDR4 sys_reg(2, 1, 0, 12, 7)
+#define SYS_TRCIDR5 sys_reg(2, 1, 0, 13, 7)
+#define SYS_TRCIDR6 sys_reg(2, 1, 0, 14, 7)
+#define SYS_TRCIDR7 sys_reg(2, 1, 0, 15, 7)
+#define SYS_TRCIDR8 sys_reg(2, 1, 0, 0, 6)
+#define SYS_TRCIDR9 sys_reg(2, 1, 0, 1, 6)
+#define SYS_TRCIMSPEC0 sys_reg(2, 1, 0, 0, 7)
+#define SYS_TRCIMSPEC(m) sys_reg(2, 1, 0, (m & 7), 7)

Since ‘m' restrictions are not being considered, when m = 0 it clashes with SYS_TRCIMSPEC0 so overall SYS_TRCIMSPEC(m) already covers SYS_TRCIMSPEC0. Is this redundancy needed?

arch/arm64/include/asm/sysreg.h line 251

Quoted 5 lines
arch/arm64/include/asm/sysreg.h +5 -0
+#define SYS_TRCITEEDCR sys_reg(2, 1, 0, 2, 1)
+#define SYS_TRCOSLSR sys_reg(2, 1, 1, 1, 4)
+#define SYS_TRCPRGCTLR sys_reg(2, 1, 0, 1, 0)
+#define SYS_TRCQCTLR sys_reg(2, 1, 0, 1, 1)
+#define SYS_TRCRSCTLR(m) sys_reg(2, 1, 1, (m & 15), (0 | (m >> 4)))

Suggestion for SYS_TRCRSCTLR(m) op2: (0 | ((m & 16) >> 4))

arch/arm64/include/asm/sysreg.h line 253

Quoted 2 lines
arch/arm64/include/asm/sysreg.h +2 -0
+#define SYS_TRCRSR sys_reg(2, 1, 0, 10, 0)
+#define SYS_TRCSEQEVR(m) sys_reg(2, 1, 0, (m & 3), 4)

Without considering ‘m' restrictions CRm can clash with 3.

Please let me know if I missed something.

Thanks
Miguel

Quoted 27 lines
arch/arm64/include/asm/sysreg.h +21 -0
+#define SYS_TRCSEQRSTEVR sys_reg(2, 1, 0, 6, 4)
+#define SYS_TRCSEQSTR sys_reg(2, 1, 0, 7, 4)
+#define SYS_TRCSSCCR(m) sys_reg(2, 1, 1, (m & 7), 2)
+#define SYS_TRCSSCSR(m) sys_reg(2, 1, 1, (8 | (m & 7)), 2)
+#define SYS_TRCSSPCICR(m) sys_reg(2, 1, 1, (m & 7), 3)
+#define SYS_TRCSTALLCTLR sys_reg(2, 1, 0, 11, 0)
+#define SYS_TRCSTATR sys_reg(2, 1, 0, 3, 0)
+#define SYS_TRCSYNCPR sys_reg(2, 1, 0, 13, 0)
+#define SYS_TRCTRACEIDR sys_reg(2, 1, 0, 0, 1)
+#define SYS_TRCTSCTLR sys_reg(2, 1, 0, 12, 0)
+#define SYS_TRCVICTLR sys_reg(2, 1, 0, 0, 2)
+#define SYS_TRCVIIECTLR sys_reg(2, 1, 0, 1, 2)
+#define SYS_TRCVIPCSSCTLR sys_reg(2, 1, 0, 3, 2)
+#define SYS_TRCVISSCTLR sys_reg(2, 1, 0, 2, 2)
+#define SYS_TRCVMIDCCTLR0 sys_reg(2, 1, 3, 2, 2)
+#define SYS_TRCVMIDCCTLR1 sys_reg(2, 1, 3, 3, 2)
+#define SYS_TRCVMIDCVR(m) sys_reg(2, 1, 3, ((m & 7) << 1), 1)
+
+/* ETM */
+#define SYS_TRCOSLAR sys_reg(2, 1, 1, 0, 4)
+
 #define SYS_MIDR_EL1 sys_reg(3, 0, 0, 0, 0)
 #define SYS_MPIDR_EL1 sys_reg(3, 0, 0, 0, 5)
 #define SYS_REVIDR_EL1 sys_reg(3, 0, 0, 0, 6)
Signature · 1 line
2.34.1
author maintainer
#25 lore raw

Re: [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

Quoted 25 lines · Miguel Luis

Hi Marc,

A few comments on this one, please see below.

Quoted 20 lines · Marc Zyngier

The HDFGxTR_EL2 registers trap a (huge) set of debug and trace related registers. Add their encodings (and only that, because we really don't care about what these registers actually do at this stage).

---
arch/arm64/include/asm/sysreg.h | 78 +++++++++++++++++++++++++++++++++

1 file changed, 78 insertions(+)
arch/arm64/include/asm/sysreg.h +0 -0
index 76289339b43b..9dfd127be55a 100644
@@ -194,6 +194,84 @@
 #define SYS_DBGDTRTX_EL0 sys_reg(2, 3, 0, 5, 0)
 #define SYS_DBGVCR32_EL2 sys_reg(2, 4, 0, 7, 0)
 
Quoted 16 lines
Quoted 4 lines
arch/arm64/include/asm/sysreg.h +4 -0
+#define SYS_BRBINF_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 0))
+#define SYS_BRBINFINJ_EL1 sys_reg(2, 1, 9, 1, 0)
+#define SYS_BRBSRC_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 1))
+#define SYS_BRBSRCINJ_EL1 sys_reg(2, 1, 9, 1, 1)

arch/arm64/include/asm/sysreg.h line 210

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define SYS_BRBTGT_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 2))
+#define SYS_BRBTGTINJ_EL1 sys_reg(2, 1, 9, 1, 2)
+#define SYS_BRBTS_EL1 sys_reg(2, 1, 9, 0, 2)
+
+#define SYS_BRBCR_EL1 sys_reg(2, 1, 9, 0, 0)
+#define SYS_BRBFCR_EL1 sys_reg(2, 1, 9, 0, 1)
+#define SYS_BRBIDR0_EL1 sys_reg(2, 1, 9, 2, 0)
+
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)

SYS_TRCITECR_EL1 shows up twice.

Ah, nice one. Too many registers.

Quoted 5 lines

arch/arm64/include/asm/sysreg.h line 211

Quoted 2 lines
arch/arm64/include/asm/sysreg.h +1 -0
+#define SYS_TRCACATR(m) sys_reg(2, 1, 2, ((m & 7) << 1), (2 | (m >> 3)))

Besides m’s restrictions it could be sanitised in op2 to consider only bit m[3]. Suggestion for op2: (2 | ((m & 8) >> 3)))

It is fully expected that 'm' will be in the 0-15 range, as per the architecture (D19.4.8), and the tables only use that exact range.

Do you see an actual bug, or is this defensive programming?

Thanks,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
#26 lore raw

Re: [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

Hi Marc,

Quoted 51 lines · Marc Zyngier
Quoted 25 lines · Miguel Luis

Hi Marc,

A few comments on this one, please see below.

Quoted 20 lines · Marc Zyngier

The HDFGxTR_EL2 registers trap a (huge) set of debug and trace related registers. Add their encodings (and only that, because we really don't care about what these registers actually do at this stage).

---
arch/arm64/include/asm/sysreg.h | 78 +++++++++++++++++++++++++++++++++

1 file changed, 78 insertions(+)
arch/arm64/include/asm/sysreg.h +0 -0
index 76289339b43b..9dfd127be55a 100644
@@ -194,6 +194,84 @@
 #define SYS_DBGDTRTX_EL0 sys_reg(2, 3, 0, 5, 0)
 #define SYS_DBGVCR32_EL2 sys_reg(2, 4, 0, 7, 0)
 
Quoted 16 lines
Quoted 4 lines
arch/arm64/include/asm/sysreg.h +4 -0
+#define SYS_BRBINF_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 0))
+#define SYS_BRBINFINJ_EL1 sys_reg(2, 1, 9, 1, 0)
+#define SYS_BRBSRC_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 1))
+#define SYS_BRBSRCINJ_EL1 sys_reg(2, 1, 9, 1, 1)

arch/arm64/include/asm/sysreg.h line 210

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define SYS_BRBTGT_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 2))
+#define SYS_BRBTGTINJ_EL1 sys_reg(2, 1, 9, 1, 2)
+#define SYS_BRBTS_EL1 sys_reg(2, 1, 9, 0, 2)
+
+#define SYS_BRBCR_EL1 sys_reg(2, 1, 9, 0, 0)
+#define SYS_BRBFCR_EL1 sys_reg(2, 1, 9, 0, 1)
+#define SYS_BRBIDR0_EL1 sys_reg(2, 1, 9, 2, 0)
+
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)

SYS_TRCITECR_EL1 shows up twice.

Ah, nice one. Too many registers.

Quoted 3 lines
Quoted 2 lines
arch/arm64/include/asm/sysreg.h +1 -0
+#define SYS_TRCACATR(m) sys_reg(2, 1, 2, ((m & 7) << 1), (2 | (m >> 3)))
Quoted 8 lines
Quoted 2 lines

Besides m’s restrictions it could be sanitised in op2 to consider only bit m[3]. Suggestion for op2: (2 | ((m & 8) >> 3)))

It is fully expected that 'm' will be in the 0-15 range, as per the architecture (D19.4.8), and the tables only use that exact range.

Do you see an actual bug, or is this defensive programming?

I was seeing a problem when m[5]=1 then Op2 of 0b01:m[3] isn’t guaranteed anymore overridden with 0b11:m[3].

Clearly ‘m’ would be outside the range but not messing with Op2 fixed bits 0b01. Not a problem for patch 21 though.

Due to the uncertainty if this can bite later, hence the suggestion and also open to advices.

Thank you,

Miguel

Quoted 6 lines

Thanks,

M.

Signature · 1 line
Without deviation from the norm, progress is not possible.
author maintainer
#27 lore raw

Re: [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

Quoted 57 lines · Miguel Luis

Hi Marc,

Quoted 54 lines · Marc Zyngier
Quoted 25 lines · Miguel Luis

Hi Marc,

A few comments on this one, please see below.

Quoted 20 lines · Marc Zyngier

The HDFGxTR_EL2 registers trap a (huge) set of debug and trace related registers. Add their encodings (and only that, because we really don't care about what these registers actually do at this stage).

---
arch/arm64/include/asm/sysreg.h | 78 +++++++++++++++++++++++++++++++++

1 file changed, 78 insertions(+)
arch/arm64/include/asm/sysreg.h +0 -0
index 76289339b43b..9dfd127be55a 100644
@@ -194,6 +194,84 @@
 #define SYS_DBGDTRTX_EL0 sys_reg(2, 3, 0, 5, 0)
 #define SYS_DBGVCR32_EL2 sys_reg(2, 4, 0, 7, 0)
 
Quoted 16 lines
Quoted 4 lines
arch/arm64/include/asm/sysreg.h +4 -0
+#define SYS_BRBINF_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 0))
+#define SYS_BRBINFINJ_EL1 sys_reg(2, 1, 9, 1, 0)
+#define SYS_BRBSRC_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 1))
+#define SYS_BRBSRCINJ_EL1 sys_reg(2, 1, 9, 1, 1)

arch/arm64/include/asm/sysreg.h line 210

Quoted 10 lines
arch/arm64/include/asm/sysreg.h +10 -0
+#define SYS_BRBTGT_EL1(n) sys_reg(2, 1, 8, (n & 15), (((n & 16) >> 2) | 2))
+#define SYS_BRBTGTINJ_EL1 sys_reg(2, 1, 9, 1, 2)
+#define SYS_BRBTS_EL1 sys_reg(2, 1, 9, 0, 2)
+
+#define SYS_BRBCR_EL1 sys_reg(2, 1, 9, 0, 0)
+#define SYS_BRBFCR_EL1 sys_reg(2, 1, 9, 0, 1)
+#define SYS_BRBIDR0_EL1 sys_reg(2, 1, 9, 2, 0)
+
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)
+#define SYS_TRCITECR_EL1 sys_reg(3, 0, 1, 2, 3)

SYS_TRCITECR_EL1 shows up twice.

Ah, nice one. Too many registers.

Quoted 5 lines

arch/arm64/include/asm/sysreg.h line 211

Quoted 2 lines
arch/arm64/include/asm/sysreg.h +1 -0
+#define SYS_TRCACATR(m) sys_reg(2, 1, 2, ((m & 7) << 1), (2 | (m >> 3)))

Besides m’s restrictions it could be sanitised in op2 to consider only bit m[3]. Suggestion for op2: (2 | ((m & 8) >> 3)))

Quoted 8 lines
Quoted 5 lines

It is fully expected that 'm' will be in the 0-15 range, as per the architecture (D19.4.8), and the tables only use that exact range.

Do you see an actual bug, or is this defensive programming?

I was seeing a problem when m[5]=1 then Op2 of 0b01:m[3] isn’t guaranteed anymore overridden with 0b11:m[3].

But that'd be wrong anyway. Now, we'd have an encoding that possibly aliases to something else, and we don't know about it.

Quoted 2 lines

Clearly ‘m’ would be outside the range but not messing with Op2 fixed bits 0b01. Not a problem for patch 21 though.

All these macros have as the basis for their use that you use *valid* input.

Quoted 2 lines

Due to the uncertainty if this can bite later, hence the suggestion and also open to advices.

If you really want some defensive programming on that front, then you should make sure we detect the issue at compile time, rather than silently changing the encoding.

Thanks,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
author maintainer
#28 lore raw

PATCH v2 7/26 arm64: Add missing BRB/CFP/DVP/CPP instructions

In reply to Marc Zyngier

HFGITR_EL2 traps a bunch of instructions for which we don't have encodings yet. Add them.

---
 arch/arm64/include/asm/sysreg.h | 7 +++++++
 1 file changed, 7 insertions(+)
arch/arm64/include/asm/sysreg.h +7 -0
index 9dfd127be55a..e2357529c633 100644
@@ -737,6 +737,13 @@
 #define OP_TLBI_VALE2NXS		sys_insn(1, 4, 9, 7, 5)
 #define OP_TLBI_VMALLS12E1NXS		sys_insn(1, 4, 9, 7, 6)
 
+/* Misc instructions */
+#define OP_BRB_IALL			sys_insn(1, 1, 7, 2, 4)
+#define OP_BRB_INJ			sys_insn(1, 1, 7, 2, 5)
+#define OP_CFP_RCTX			sys_insn(1, 3, 7, 3, 4)
+#define OP_DVP_RCTX			sys_insn(1, 3, 7, 3, 5)
+#define OP_CPP_RCTX			sys_insn(1, 3, 7, 3, 7)
+
 /* Common SCTLR_ELx flags. */
 #define SCTLR_ELx_ENTP2	(BIT(60))
 #define SCTLR_ELx_DSSBS	(BIT(44))
Signature · 1 line
2.34.1
author maintainer
#30 lore raw

PATCH v2 8/26 arm64: Add HDFGRTR_EL2 and HDFGWTR_EL2 layouts

In reply to Marc Zyngier

As we're about to implement full support for FEAT_FGT, add the full HDFGRTR_EL2 and HDFGWTR_EL2 layouts.

---
 arch/arm64/include/asm/sysreg.h |   2 -
 arch/arm64/tools/sysreg         | 129 ++++++++++++++++++++++++++++++++
 2 files changed, 129 insertions(+), 2 deletions(-)
arch/arm64/include/asm/sysreg.h +0 -2
index e2357529c633..6d3d16fac227 100644
@@ -497,8 +497,6 @@
 #define SYS_VTCR_EL2			sys_reg(3, 4, 2, 1, 2)
 
 #define SYS_TRFCR_EL2			sys_reg(3, 4, 1, 2, 1)
-#define SYS_HDFGRTR_EL2			sys_reg(3, 4, 3, 1, 4)
-#define SYS_HDFGWTR_EL2			sys_reg(3, 4, 3, 1, 5)
 #define SYS_HAFGRTR_EL2			sys_reg(3, 4, 3, 1, 6)
 #define SYS_SPSR_EL2			sys_reg(3, 4, 4, 0, 0)
 #define SYS_ELR_EL2			sys_reg(3, 4, 4, 0, 1)
arch/arm64/tools/sysreg +129 -0

135 lines not on this page · raw message

Signature · 1 line
2.34.1
Reviewed-by on v2 8/26
#32 lore raw

Re: [PATCH v2 08/26] arm64: Add HDFGRTR_EL2 and HDFGWTR_EL2 layouts

In reply to Marc Zyngier

Hi Marc,

Quoted 152 lines · Marc Zyngier

As we're about to implement full support for FEAT_FGT, add the full HDFGRTR_EL2 and HDFGWTR_EL2 layouts.

---
arch/arm64/include/asm/sysreg.h | 2 -
arch/arm64/tools/sysreg | 129 ++++++++++++++++++++++++++++++++

2 files changed, 129 insertions(+), 2 deletions(-)
arch/arm64/include/asm/sysreg.h +0 -2
index e2357529c633..6d3d16fac227 100644
@@ -497,8 +497,6 @@
 #define SYS_VTCR_EL2 sys_reg(3, 4, 2, 1, 2)
 
 #define SYS_TRFCR_EL2 sys_reg(3, 4, 1, 2, 1)
-#define SYS_HDFGRTR_EL2 sys_reg(3, 4, 3, 1, 4)
-#define SYS_HDFGWTR_EL2 sys_reg(3, 4, 3, 1, 5)
 #define SYS_HAFGRTR_EL2 sys_reg(3, 4, 3, 1, 6)
 #define SYS_SPSR_EL2 sys_reg(3, 4, 4, 0, 0)
 #define SYS_ELR_EL2 sys_reg(3, 4, 4, 0, 1)
arch/arm64/tools/sysreg +0 -0
index 65866bf819c3..2517ef7c21cf 100644
@@ -2156,6 +2156,135 @@ Field 1 ICIALLU

Field 0 ICIALLUIS
EndSysreg

arch/arm64/tools/sysreg +119 -0
+Sysreg HDFGRTR_EL2 3 4 3 1 4
+Field 63 PMBIDR_EL1
+Field 62 nPMSNEVFR_EL1
+Field 61 nBRBDATA
+Field 60 nBRBCTL
+Field 59 nBRBIDR
+Field 58 PMCEIDn_EL0
+Field 57 PMUSERENR_EL0
+Field 56 TRBTRG_EL1
+Field 55 TRBSR_EL1
+Field 54 TRBPTR_EL1
+Field 53 TRBMAR_EL1
+Field 52 TRBLIMITR_EL1
+Field 51 TRBIDR_EL1
+Field 50 TRBBASER_EL1
+Res0 49
+Field 48 TRCVICTLR
+Field 47 TRCSTATR
+Field 46 TRCSSCSRn
+Field 45 TRCSEQSTR
+Field 44 TRCPRGCTLR
+Field 43 TRCOSLSR
+Res0 42
+Field 41 TRCIMSPECn
+Field 40 TRCID
+Res0 39:38
+Field 37 TRCCNTVRn
+Field 36 TRCCLAIM
+Field 35 TRCAUXCTLR
+Field 34 TRCAUTHSTATUS
+Field 33 TRC
+Field 32 PMSLATFR_EL1
+Field 31 PMSIRR_EL1
+Field 30 PMSIDR_EL1
+Field 29 PMSICR_EL1
+Field 28 PMSFCR_EL1
+Field 27 PMSEVFR_EL1
+Field 26 PMSCR_EL1
+Field 25 PMBSR_EL1
+Field 24 PMBPTR_EL1
+Field 23 PMBLIMITR_EL1
+Field 22 PMMIR_EL1
+Res0 21:20
+Field 19 PMSELR_EL0
+Field 18 PMOVS
+Field 17 PMINTEN
+Field 16 PMCNTEN
+Field 15 PMCCNTR_EL0
+Field 14 PMCCFILTR_EL0
+Field 13 PMEVTYPERn_EL0
+Field 12 PMEVCNTRn_EL0
+Field 11 OSDLR_EL1
+Field 10 OSECCR_EL1
+Field 9 OSLSR_EL1
+Res0 8
+Field 7 DBGPRCR_EL1
+Field 6 DBGAUTHSTATUS_EL1
+Field 5 DBGCLAIM
+Field 4 MDSCR_EL1
+Field 3 DBGWVRn_EL1
+Field 2 DBGWCRn_EL1
+Field 1 DBGBVRn_EL1
+Field 0 DBGBCRn_EL1
+EndSysreg
+
+Sysreg HDFGWTR_EL2 3 4 3 1 5
+Res0 63
+Field 62 nPMSNEVFR_EL1
+Field 61 nBRBDATA
+Field 60 nBRBCTL
+Res0 59:58
+Field 57 PMUSERENR_EL0
+Field 56 TRBTRG_EL1
+Field 55 TRBSR_EL1
+Field 54 TRBPTR_EL1
+Field 53 TRBMAR_EL1
+Field 52 TRBLIMITR_EL1
+Res0 51
+Field 50 TRBBASER_EL1
+Field 49 TRFCR_EL1
+Field 48 TRCVICTLR
+Res0 47
+Field 46 TRCSSCSRn
+Field 45 TRCSEQSTR
+Field 44 TRCPRGCTLR
+Res0 43
+Field 42 TRCOSLAR
+Field 41 TRCIMSPECn
+Res0 40:38
+Field 37 TRCCNTVRn
+Field 36 TRCCLAIM
+Field 35 TRCAUXCTLR
+Res0 34
+Field 33 TRC
+Field 32 PMSLATFR_EL1
+Field 31 PMSIRR_EL1
+Res0 30
+Field 29 PMSICR_EL1
+Field 28 PMSFCR_EL1
+Field 27 PMSEVFR_EL1
+Field 26 PMSCR_EL1
+Field 25 PMBSR_EL1
+Field 24 PMBPTR_EL1
+Field 23 PMBLIMITR_EL1
+Res0 22
+Field 21 PMCR_EL0
+Field 20 PMSWINC_EL0
+Field 19 PMSELR_EL0
+Field 18 PMOVS
+Field 17 PMINTEN
+Field 16 PMCNTEN
+Field 15 PMCCNTR_EL0
+Field 14 PMCCFILTR_EL0
+Field 13 PMEVTYPERn_EL0
+Field 12 PMEVCNTRn_EL0
+Field 11 OSDLR_EL1
+Field 10 OSECCR_EL1
+Res0 9
+Field 8 OSLAR_EL1

arch/arm64/tools/sysreg Field 1 ICIALLU · line 2287

Quoted 10 lines
arch/arm64/tools/sysreg +10 -0
+Field 7 DBGPRCR_EL1
+Res0 6
+Field 5 DBGCLAIM
+Field 4 MDSCR_EL1
+Field 3 DBGWVRn_EL1
+Field 2 DBGWCRn_EL1
+Field 1 DBGBVRn_EL1
+Field 0 DBGBCRn_EL1
+EndSysreg
+

Thanks

Miguel

Quoted 6 lines

Sysreg ZCR_EL2 3 4 1 2 0
Fields ZCR_ELx
EndSysreg

Signature · 1 line
2.34.1
author maintainer
#33 lore raw

PATCH v2 9/26 arm64: Add feature detection for fine grained traps

In reply to Marc Zyngier

From: Mark Brown <[email protected]>

In order to allow us to have shared code for managing fine grained traps for KVM guests add it as a detected feature rather than relying on it being a dependency of other features.

---
 arch/arm64/kernel/cpufeature.c | 7 +++++++
 arch/arm64/tools/cpucaps       | 1 +
 2 files changed, 8 insertions(+)
arch/arm64/kernel/cpufeature.c +7 -0
index f9d456fe132d..668e2872a086 100644
@@ -2627,6 +2627,13 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
 		.matches = has_cpuid_feature,
 		ARM64_CPUID_FIELDS(ID_AA64ISAR1_EL1, LRCPC, IMP)
 	},
+	{
+		.desc = "Fine Grained Traps",
+		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
+		.capability = ARM64_HAS_FGT,
+		.matches = has_cpuid_feature,
+		ARM64_CPUID_FIELDS(ID_AA64MMFR0_EL1, FGT, IMP)
+	},
 #ifdef CONFIG_ARM64_SME
 	{
 		.desc = "Scalable Matrix Extension",
arch/arm64/tools/cpucaps +1 -0
index c80ed4f3cbce..c3f06fdef609 100644
@@ -26,6 +26,7 @@ HAS_ECV
 HAS_ECV_CNTPOFF
 HAS_EPAN
 HAS_EVT
+HAS_FGT
 HAS_GENERIC_AUTH
 HAS_GENERIC_AUTH_ARCH_QARMA3
 HAS_GENERIC_AUTH_ARCH_QARMA5
Signature · 1 line
2.34.1
reviewer Reviewed-by on v2 9/26
#34 lore raw

Re: [PATCH v2 09/26] arm64: Add feature detection for fine grained traps

Quoted 12 lines · Marc Zyngier

From: Mark Brown <[email protected]>

In order to allow us to have shared code for managing fine grained traps for KVM guests add it as a detected feature rather than relying on it being a dependency of other features.

Quoted 4 lines
---
 arch/arm64/kernel/cpufeature.c | 7 +++++++
 arch/arm64/tools/cpucaps       | 1 +
 2 files changed, 8 insertions(+)
Reviewed-by on v2 9/26
#35 lore raw

Re: [PATCH v2 09/26] arm64: Add feature detection for fine grained traps

In reply to Marc Zyngier

Hi Marc,

Quoted 42 lines · Marc Zyngier

From: Mark Brown <[email protected]>

In order to allow us to have shared code for managing fine grained traps for KVM guests add it as a detected feature rather than relying on it being a dependency of other features.

---
arch/arm64/kernel/cpufeature.c | 7 +++++++
arch/arm64/tools/cpucaps | 1 +

2 files changed, 8 insertions(+)
arch/arm64/kernel/cpufeature.c +7 -0
index f9d456fe132d..668e2872a086 100644
@@ -2627,6 +2627,13 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
 .matches = has_cpuid_feature,
 ARM64_CPUID_FIELDS(ID_AA64ISAR1_EL1, LRCPC, IMP)
 },
+ {
+ .desc = "Fine Grained Traps",
+ .type = ARM64_CPUCAP_SYSTEM_FEATURE,
+ .capability = ARM64_HAS_FGT,
+ .matches = has_cpuid_feature,
+ ARM64_CPUID_FIELDS(ID_AA64MMFR0_EL1, FGT, IMP)
+ },
 #ifdef CONFIG_ARM64_SME
 {
 .desc = "Scalable Matrix Extension",
arch/arm64/tools/cpucaps +0 -0
index c80ed4f3cbce..c3f06fdef609 100644
@@ -26,6 +26,7 @@ HAS_ECV
Quoted 4 lines

HAS_ECV_CNTPOFF
HAS_EPAN
HAS_EVT
+HAS_FGT

Thanks

Miguel

Quoted 6 lines

HAS_GENERIC_AUTH
HAS_GENERIC_AUTH_ARCH_QARMA3
HAS_GENERIC_AUTH_ARCH_QARMA5

Signature · 1 line
2.34.1
author maintainer
#36 lore raw

PATCH v2 10/26 KVM: arm64: Correctly handle ACCDATA_EL1 traps

In reply to Marc Zyngier

As we blindly reset some HFGxTR_EL2 bits to 0, we also randomly trap unsuspecting sysregs that have their trap bits with a negative polarity.

ACCDATA_EL1 is one such register that can be accessed by the guest, causing a splat on the host as we don't have a proper handler for it.

Adding such handler addresses the issue, though there are a number of other registers missing as the current architecture documentation doesn't describe them yet.

---
 arch/arm64/include/asm/sysreg.h | 2 ++
 arch/arm64/kvm/sys_regs.c       | 2 ++
 2 files changed, 4 insertions(+)
arch/arm64/include/asm/sysreg.h +2 -0
index 6d3d16fac227..d528a79417a0 100644
@@ -389,6 +389,8 @@
 #define SYS_ICC_IGRPEN0_EL1		sys_reg(3, 0, 12, 12, 6)
 #define SYS_ICC_IGRPEN1_EL1		sys_reg(3, 0, 12, 12, 7)
 
+#define SYS_ACCDATA_EL1			sys_reg(3, 0, 13, 0, 5)
+
 #define SYS_CNTKCTL_EL1			sys_reg(3, 0, 14, 1, 0)
 
 #define SYS_AIDR_EL1			sys_reg(3, 1, 0, 0, 7)
arch/arm64/kvm/sys_regs.c +2 -0
index 2ca2973abe66..38f221f9fc98 100644
@@ -2151,6 +2151,8 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 	{ SYS_DESC(SYS_CONTEXTIDR_EL1), access_vm_reg, reset_val, CONTEXTIDR_EL1, 0 },
 	{ SYS_DESC(SYS_TPIDR_EL1), NULL, reset_unknown, TPIDR_EL1 },
 
+	{ SYS_DESC(SYS_ACCDATA_EL1), undef_access },
+
 	{ SYS_DESC(SYS_SCXTNUM_EL1), undef_access },
 
 	{ SYS_DESC(SYS_CNTKCTL_EL1), NULL, reset_val, CNTKCTL_EL1, 0},
Signature · 1 line
2.34.1
Reviewed-by on v2 10/26
#37 lore raw

Re: [PATCH v2 10/26] KVM: arm64: Correctly handle ACCDATA_EL1 traps

Hi Marc,

Quoted 39 lines · Marc Zyngier

As we blindly reset some HFGxTR_EL2 bits to 0, we also randomly trap unsuspecting sysregs that have their trap bits with a negative polarity.

ACCDATA_EL1 is one such register that can be accessed by the guest, causing a splat on the host as we don't have a proper handler for it.

Adding such handler addresses the issue, though there are a number of other registers missing as the current architecture documentation doesn't describe them yet.

---
arch/arm64/include/asm/sysreg.h | 2 ++
arch/arm64/kvm/sys_regs.c | 2 ++

2 files changed, 4 insertions(+)
arch/arm64/include/asm/sysreg.h +2 -0
index 6d3d16fac227..d528a79417a0 100644
@@ -389,6 +389,8 @@
 #define SYS_ICC_IGRPEN0_EL1 sys_reg(3, 0, 12, 12, 6)
 #define SYS_ICC_IGRPEN1_EL1 sys_reg(3, 0, 12, 12, 7)
 
+#define SYS_ACCDATA_EL1 sys_reg(3, 0, 13, 0, 5)
+
 #define SYS_CNTKCTL_EL1 sys_reg(3, 0, 14, 1, 0)
 
 #define SYS_AIDR_EL1 sys_reg(3, 1, 0, 0, 7)
arch/arm64/kvm/sys_regs.c +0 -0
index 2ca2973abe66..38f221f9fc98 100644

arch/arm64/kvm/sys_regs.c static const struct sys_reg_desc sys_reg_descs[] = { · line 2155

Quoted 5 lines
arch/arm64/kvm/sys_regs.c +2 -0
@@ -2151,6 +2151,8 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 { SYS_DESC(SYS_CONTEXTIDR_EL1), access_vm_reg, reset_val, CONTEXTIDR_EL1, 0 },
 { SYS_DESC(SYS_TPIDR_EL1), NULL, reset_unknown, TPIDR_EL1 },
 
+ { SYS_DESC(SYS_ACCDATA_EL1), undef_access },
+

Per DDI0487J.a, adding ACCDATA_EL1 completes the list of sysregs that would trap with a negative polarity bit on HFGxTR_EL2.

Thanks

Miguel

Quoted 6 lines
arch/arm64/kvm/sys_regs.c +0 -0
 { SYS_DESC(SYS_SCXTNUM_EL1), undef_access },
 
 { SYS_DESC(SYS_CNTKCTL_EL1), NULL, reset_val, CNTKCTL_EL1, 0},
Signature · 1 line
2.34.1
author maintainer
#38 lore raw

PATCH v2 11/26 KVM: arm64: Add missing HCR_EL2 trap bits

In reply to Marc Zyngier

We're still missing a handfull of HCR_EL2 trap bits. Add them.

---
 arch/arm64/include/asm/kvm_arm.h | 9 +++++++++
 1 file changed, 9 insertions(+)
arch/arm64/include/asm/kvm_arm.h +9 -0
index 58e5eb27da68..028049b147df 100644
@@ -18,10 +18,19 @@
 #define HCR_DCT		(UL(1) << 57)
 #define HCR_ATA_SHIFT	56
 #define HCR_ATA		(UL(1) << HCR_ATA_SHIFT)
+#define HCR_TTLBOS	(UL(1) << 55)
+#define HCR_TTLBIS	(UL(1) << 54)
+#define HCR_ENSCXT	(UL(1) << 53)
+#define HCR_TOCU	(UL(1) << 52)
 #define HCR_AMVOFFEN	(UL(1) << 51)
+#define HCR_TICAB	(UL(1) << 50)
 #define HCR_TID4	(UL(1) << 49)
 #define HCR_FIEN	(UL(1) << 47)
 #define HCR_FWB		(UL(1) << 46)
+#define HCR_NV2		(UL(1) << 45)
+#define HCR_AT		(UL(1) << 44)
+#define HCR_NV1		(UL(1) << 43)
+#define HCR_NV		(UL(1) << 42)
 #define HCR_API		(UL(1) << 41)
 #define HCR_APK		(UL(1) << 40)
 #define HCR_TEA		(UL(1) << 37)
Signature · 1 line
2.34.1
Reviewed-by on v2 11/26
#39 lore raw

Re: [PATCH v2 11/26] KVM: arm64: Add missing HCR_EL2 trap bits

Hi Marc,

Quoted 21 lines · Marc Zyngier

We're still missing a handfull of HCR_EL2 trap bits. Add them.

---
arch/arm64/include/asm/kvm_arm.h | 9 +++++++++

1 file changed, 9 insertions(+)
arch/arm64/include/asm/kvm_arm.h +3 -0
index 58e5eb27da68..028049b147df 100644
@@ -18,10 +18,19 @@
 #define HCR_DCT (UL(1) << 57)
 #define HCR_ATA_SHIFT 56
 #define HCR_ATA (UL(1) << HCR_ATA_SHIFT)
+#define HCR_TTLBOS (UL(1) << 55)
+#define HCR_TTLBIS (UL(1) << 54)
+#define HCR_ENSCXT (UL(1) << 53)

arch/arm64/include/asm/kvm_arm.h line 33

Quoted 10 lines
arch/arm64/include/asm/kvm_arm.h +6 -0
+#define HCR_TOCU (UL(1) << 52)
 #define HCR_AMVOFFEN (UL(1) << 51)
+#define HCR_TICAB (UL(1) << 50)
 #define HCR_TID4 (UL(1) << 49)
 #define HCR_FIEN (UL(1) << 47)
 #define HCR_FWB (UL(1) << 46)
+#define HCR_NV2 (UL(1) << 45)
+#define HCR_AT (UL(1) << 44)
+#define HCR_NV1 (UL(1) << 43)
+#define HCR_NV (UL(1) << 42)

Thanks
Miguel

Quoted 6 lines
arch/arm64/include/asm/kvm_arm.h +0 -0
 #define HCR_API (UL(1) << 41)
 #define HCR_APK (UL(1) << 40)
 #define HCR_TEA (UL(1) << 37)
Signature · 1 line
2.34.1
author maintainer
#40 lore raw

PATCH v2 12/26 KVM: arm64: nv: Add FGT registers

In reply to Marc Zyngier

Add the 5 registers covering FEAT_FGT. The AMU-related registers are currently left out as we don't have a plan for them. Yet.

---
 arch/arm64/include/asm/kvm_host.h | 5 +++++
 arch/arm64/kvm/sys_regs.c         | 5 +++++
 2 files changed, 10 insertions(+)
arch/arm64/include/asm/kvm_host.h +5 -0
index d3dd05bbfe23..721680da1011 100644
@@ -400,6 +400,11 @@ enum vcpu_sysreg {
 	TPIDR_EL2,	/* EL2 Software Thread ID Register */
 	CNTHCTL_EL2,	/* Counter-timer Hypervisor Control register */
 	SP_EL2,		/* EL2 Stack Pointer */
+	HFGRTR_EL2,
+	HFGWTR_EL2,
+	HFGITR_EL2,
+	HDFGRTR_EL2,
+	HDFGWTR_EL2,
 	CNTHP_CTL_EL2,
 	CNTHP_CVAL_EL2,
 	CNTHV_CTL_EL2,
arch/arm64/kvm/sys_regs.c +5 -0
index 38f221f9fc98..f5baaa508926 100644
@@ -2367,6 +2367,9 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 	EL2_REG(MDCR_EL2, access_rw, reset_val, 0),
 	EL2_REG(CPTR_EL2, access_rw, reset_val, CPTR_NVHE_EL2_RES1),
 	EL2_REG(HSTR_EL2, access_rw, reset_val, 0),
+	EL2_REG(HFGRTR_EL2, access_rw, reset_val, 0),
+	EL2_REG(HFGWTR_EL2, access_rw, reset_val, 0),
+	EL2_REG(HFGITR_EL2, access_rw, reset_val, 0),
 	EL2_REG(HACR_EL2, access_rw, reset_val, 0),
 
 	EL2_REG(TTBR0_EL2, access_rw, reset_val, 0),
@@ -2376,6 +2379,8 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 	EL2_REG(VTCR_EL2, access_rw, reset_val, 0),
 
 	{ SYS_DESC(SYS_DACR32_EL2), NULL, reset_unknown, DACR32_EL2 },
+	EL2_REG(HDFGRTR_EL2, access_rw, reset_val, 0),
+	EL2_REG(HDFGWTR_EL2, access_rw, reset_val, 0),
 	EL2_REG(SPSR_EL2, access_rw, reset_val, 0),
 	EL2_REG(ELR_EL2, access_rw, reset_val, 0),
 	{ SYS_DESC(SYS_SP_EL1), access_sp_el1},
Signature · 1 line
2.34.1
Reviewed-by on v2 12/26
#41 lore raw

Re: [PATCH v2 12/26] KVM: arm64: nv: Add FGT registers

Hi Marc,

Quoted 43 lines · Marc Zyngier

Add the 5 registers covering FEAT_FGT. The AMU-related registers are currently left out as we don't have a plan for them. Yet.

---
arch/arm64/include/asm/kvm_host.h | 5 +++++
arch/arm64/kvm/sys_regs.c | 5 +++++

2 files changed, 10 insertions(+)
arch/arm64/include/asm/kvm_host.h +0 -0
index d3dd05bbfe23..721680da1011 100644
@@ -400,6 +400,11 @@ enum vcpu_sysreg {

TPIDR_EL2, /* EL2 Software Thread ID Register */ CNTHCTL_EL2, /* Counter-timer Hypervisor Control register */ SP_EL2, /* EL2 Stack Pointer */
+ HFGRTR_EL2,
+ HFGWTR_EL2,
+ HFGITR_EL2,
+ HDFGRTR_EL2,
+ HDFGWTR_EL2,
CNTHP_CTL_EL2,
CNTHP_CVAL_EL2,
CNTHV_CTL_EL2,

arch/arm64/kvm/sys_regs.c +3 -0
index 38f221f9fc98..f5baaa508926 100644
@@ -2367,6 +2367,9 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 EL2_REG(MDCR_EL2, access_rw, reset_val, 0),
 EL2_REG(CPTR_EL2, access_rw, reset_val, CPTR_NVHE_EL2_RES1),
 EL2_REG(HSTR_EL2, access_rw, reset_val, 0),
+ EL2_REG(HFGRTR_EL2, access_rw, reset_val, 0),
+ EL2_REG(HFGWTR_EL2, access_rw, reset_val, 0),
+ EL2_REG(HFGITR_EL2, access_rw, reset_val, 0),
 EL2_REG(HACR_EL2, access_rw, reset_val, 0),
 
 EL2_REG(TTBR0_EL2, access_rw, reset_val, 0),

arch/arm64/kvm/sys_regs.c static const struct sys_reg_desc sys_reg_descs[] = { · line 2383

Quoted 5 lines
arch/arm64/kvm/sys_regs.c +2 -0
@@ -2376,6 +2379,8 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 EL2_REG(VTCR_EL2, access_rw, reset_val, 0),
 
 { SYS_DESC(SYS_DACR32_EL2), NULL, reset_unknown, DACR32_EL2 },
+ EL2_REG(HDFGRTR_EL2, access_rw, reset_val, 0),
+ EL2_REG(HDFGWTR_EL2, access_rw, reset_val, 0),

Thanks
Miguel

Quoted 6 lines
arch/arm64/kvm/sys_regs.c +0 -0
 EL2_REG(SPSR_EL2, access_rw, reset_val, 0),
 EL2_REG(ELR_EL2, access_rw, reset_val, 0),
 { SYS_DESC(SYS_SP_EL1), access_sp_el1},
Signature · 1 line
2.34.1
author maintainer
#42 lore raw

PATCH v2 13/26 KVM: arm64: Restructure FGT register switching

In reply to Marc Zyngier

As we're about to majorly extend the handling of FGT registers, restructure the code to actually save/restore the registers as required. This is made easy thanks to the previous addition of the EL2 registers, allowing us to use the host context for this purpose.

---
 arch/arm64/include/asm/kvm_arm.h        | 21 ++++++++++
 arch/arm64/kvm/hyp/include/hyp/switch.h | 56 +++++++++++++------------
 2 files changed, 50 insertions(+), 27 deletions(-)
arch/arm64/include/asm/kvm_arm.h +21 -0
index 028049b147df..85908aa18908 100644
@@ -333,6 +333,27 @@
 				 BIT(18) |		\
 				 GENMASK(16, 15))
 
+/*
+ * FGT register definitions
+ *
+ * RES0 and polarity masks as of DDI0487J.a, to be updated as needed.
+ * We're not using the generated masks as they are usually ahead of
+ * the published ARM ARM, which we use as a reference.
+ *
+ * Once we get to a point where the two describe the same thing, we'll
+ * merge the definitions. One day.
+ */
+#define __HFGRTR_EL2_RES0	(GENMASK(63, 56) | GENMASK(53, 51))
+#define __HFGRTR_EL2_MASK	GENMASK(49, 0)
+#define __HFGRTR_EL2_nMASK	(GENMASK(55, 54) | BIT(50))
+
+#define __HFGWTR_EL2_RES0	(GENMASK(63, 56) | GENMASK(53, 51) |	\
+				 BIT(46) | BIT(42) | BIT(40) | BIT(28) | \
+				 GENMASK(26, 25) | BIT(21) | BIT(18) |	\
+				 GENMASK(15, 14) | GENMASK(10, 9) | BIT(2))
+#define __HFGWTR_EL2_MASK	GENMASK(49, 0)
+#define __HFGWTR_EL2_nMASK	(GENMASK(55, 54) | BIT(50))
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/kvm/hyp/include/hyp/switch.h +29 -27

91 lines not on this page · raw message

Signature · 1 line
2.34.1
Reviewed-by on v2 13/26
#43 lore raw

Re: [PATCH v2 13/26] KVM: arm64: Restructure FGT register switching

Hi Marc,

Quoted 134 lines · Marc Zyngier

As we're about to majorly extend the handling of FGT registers, restructure the code to actually save/restore the registers as required. This is made easy thanks to the previous addition of the EL2 registers, allowing us to use the host context for this purpose.

---
 arch/arm64/include/asm/kvm_arm.h        | 21 ++++++++++
 arch/arm64/kvm/hyp/include/hyp/switch.h | 56 +++++++++++++------------
 2 files changed, 50 insertions(+), 27 deletions(-)
arch/arm64/include/asm/kvm_arm.h +21 -0
index 028049b147df..85908aa18908 100644
@@ -333,6 +333,27 @@
 				 BIT(18) |		\
 				 GENMASK(16, 15))
 
+/*
+ * FGT register definitions
+ *
+ * RES0 and polarity masks as of DDI0487J.a, to be updated as needed.
+ * We're not using the generated masks as they are usually ahead of
+ * the published ARM ARM, which we use as a reference.
+ *
+ * Once we get to a point where the two describe the same thing, we'll
+ * merge the definitions. One day.
+ */
+#define __HFGRTR_EL2_RES0	(GENMASK(63, 56) | GENMASK(53, 51))
+#define __HFGRTR_EL2_MASK	GENMASK(49, 0)
+#define __HFGRTR_EL2_nMASK	(GENMASK(55, 54) | BIT(50))
+
+#define __HFGWTR_EL2_RES0	(GENMASK(63, 56) | GENMASK(53, 51) |	\
+				 BIT(46) | BIT(42) | BIT(40) | BIT(28) | \
+				 GENMASK(26, 25) | BIT(21) | BIT(18) |	\
+				 GENMASK(15, 14) | GENMASK(10, 9) | BIT(2))
+#define __HFGWTR_EL2_MASK	GENMASK(49, 0)
+#define __HFGWTR_EL2_nMASK	(GENMASK(55, 54) | BIT(50))
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/kvm/hyp/include/hyp/switch.h +28 -25
index 4bddb8541bec..966295178aee 100644
@@ -70,20 +70,19 @@ static inline void __activate_traps_fpsimd32(struct kvm_vcpu *vcpu)
 	}
 }
 
-static inline bool __hfgxtr_traps_required(void)
-{
-	if (cpus_have_final_cap(ARM64_SME))
-		return true;
-
-	if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
-		return true;
 
-	return false;
-}
 
-static inline void __activate_traps_hfgxtr(void)
+static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 {
+	struct kvm_cpu_context *hctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	u64 r_clr = 0, w_clr = 0, r_set = 0, w_set = 0, tmp;
+	u64 r_val, w_val;
+
+	if (!cpus_have_final_cap(ARM64_HAS_FGT))
+		return;
+
+	ctxt_sys_reg(hctxt, HFGRTR_EL2) = read_sysreg_s(SYS_HFGRTR_EL2);
+	ctxt_sys_reg(hctxt, HFGWTR_EL2) = read_sysreg_s(SYS_HFGWTR_EL2);
 
 	if (cpus_have_final_cap(ARM64_SME)) {
 		tmp = HFGxTR_EL2_nSMPRI_EL1_MASK | HFGxTR_EL2_nTPIDR2_EL0_MASK;
@@ -98,26 +97,31 @@ static inline void __activate_traps_hfgxtr(void)
 	if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
 		w_set |= HFGxTR_EL2_TCR_EL1_MASK;
 
-	sysreg_clear_set_s(SYS_HFGRTR_EL2, r_clr, r_set);
-	sysreg_clear_set_s(SYS_HFGWTR_EL2, w_clr, w_set);
+
+	/* The default is not to trap amything but ACCDATA_EL1 */
+	r_val = __HFGRTR_EL2_nMASK & ~HFGxTR_EL2_nACCDATA_EL1;
+	r_val |= r_set;
+	r_val &= ~r_clr;
+
+	w_val = __HFGWTR_EL2_nMASK & ~HFGxTR_EL2_nACCDATA_EL1;
+	w_val |= w_set;
+	w_val &= ~w_clr;
+
+	write_sysreg_s(r_val, SYS_HFGRTR_EL2);
+	write_sysreg_s(w_val, SYS_HFGWTR_EL2);
 }
 
-static inline void __deactivate_traps_hfgxtr(void)
+static inline void __deactivate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 {
-	u64 r_clr = 0, w_clr = 0, r_set = 0, w_set = 0, tmp;
+	struct kvm_cpu_context *hctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 
-	if (cpus_have_final_cap(ARM64_SME)) {
-		tmp = HFGxTR_EL2_nSMPRI_EL1_MASK | HFGxTR_EL2_nTPIDR2_EL0_MASK;
+	if (!cpus_have_final_cap(ARM64_HAS_FGT))
+		return;
 
-		r_set |= tmp;
-		w_set |= tmp;
-	}
+	write_sysreg_s(ctxt_sys_reg(hctxt, HFGRTR_EL2), SYS_HFGRTR_EL2);
+	write_sysreg_s(ctxt_sys_reg(hctxt, HFGWTR_EL2), SYS_HFGWTR_EL2);
 
-	if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
-		w_clr |= HFGxTR_EL2_TCR_EL1_MASK;
 
-	sysreg_clear_set_s(SYS_HFGRTR_EL2, r_clr, r_set);
-	sysreg_clear_set_s(SYS_HFGWTR_EL2, w_clr, w_set);
 }
 
 static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
@@ -145,8 +149,7 @@ static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
 	vcpu->arch.mdcr_el2_host = read_sysreg(mdcr_el2);
 	write_sysreg(vcpu->arch.mdcr_el2, mdcr_el2);
 
-	if (__hfgxtr_traps_required())
-		__activate_traps_hfgxtr();
+	__activate_traps_hfgxtr(vcpu);
 }
 
 static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu)

arch/arm64/kvm/hyp/include/hyp/switch.h static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu) · line 171

Quoted 9 lines
arch/arm64/kvm/hyp/include/hyp/switch.h +1 -2
@@ -162,8 +165,7 @@ static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu)
 		vcpu_clear_flag(vcpu, PMUSERENR_ON_CPU);
 	}
 
-	if (__hfgxtr_traps_required())
-		__deactivate_traps_hfgxtr();
+	__deactivate_traps_hfgxtr(vcpu);
 }
 
 static inline void ___activate_traps(struct kvm_vcpu *vcpu)

Thanks

Eric

Reviewed-by on v2 13/26
#44 lore raw

Re: [PATCH v2 13/26] KVM: arm64: Restructure FGT register switching

In reply to Marc Zyngier

Hey Marc,

Looks good, just a typo to fix.

Quoted 7 lines · Marc Zyngier

As we're about to majorly extend the handling of FGT registers, restructure the code to actually save/restore the registers as required. This is made easy thanks to the previous addition of the EL2 registers, allowing us to use the host context for this purpose.

Quoted 16 lines
arch/arm64/kvm/hyp/include/hyp/switch.h +9 -1
-static inline void __activate_traps_hfgxtr(void)
+static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 {
+	struct kvm_cpu_context *hctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	u64 r_clr = 0, w_clr = 0, r_set = 0, w_set = 0, tmp;
+	u64 r_val, w_val;
+
+	if (!cpus_have_final_cap(ARM64_HAS_FGT))
+		return;
+
+	ctxt_sys_reg(hctxt, HFGRTR_EL2) = read_sysreg_s(SYS_HFGRTR_EL2);
+	ctxt_sys_reg(hctxt, HFGWTR_EL2) = read_sysreg_s(SYS_HFGWTR_EL2);
 
 	if (cpus_have_final_cap(ARM64_SME)) {
 		tmp = HFGxTR_EL2_nSMPRI_EL1_MASK | HFGxTR_EL2_nTPIDR2_EL0_MASK;

arch/arm64/kvm/hyp/include/hyp/switch.h static inline void __activate_traps_hfgxtr(void) · line 101

Quoted 7 lines
arch/arm64/kvm/hyp/include/hyp/switch.h +2 -2
@@ -98,26 +97,31 @@ static inline void __activate_traps_hfgxtr(void)
 	if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
 		w_set |= HFGxTR_EL2_TCR_EL1_MASK;
 
-	sysreg_clear_set_s(SYS_HFGRTR_EL2, r_clr, r_set);
-	sysreg_clear_set_s(SYS_HFGWTR_EL2, w_clr, w_set);
+
+	/* The default is not to trap amything but ACCDATA_EL1 */

typo: anything

Signature · 2 lines
Thanks,
Oliver
Reviewed-by on v2 13/26
#45 lore raw

Re: [PATCH v2 13/26] KVM: arm64: Restructure FGT register switching

In reply to Marc Zyngier

Hi Marc,

Quoted 136 lines · Marc Zyngier

As we're about to majorly extend the handling of FGT registers, restructure the code to actually save/restore the registers as required. This is made easy thanks to the previous addition of the EL2 registers, allowing us to use the host context for this purpose.

---
arch/arm64/include/asm/kvm_arm.h | 21 ++++++++++
arch/arm64/kvm/hyp/include/hyp/switch.h | 56 +++++++++++++------------

2 files changed, 50 insertions(+), 27 deletions(-)
arch/arm64/include/asm/kvm_arm.h +21 -0
index 028049b147df..85908aa18908 100644
@@ -333,6 +333,27 @@
 BIT(18) | \
 GENMASK(16, 15))
 
+/*
+ * FGT register definitions
+ *
+ * RES0 and polarity masks as of DDI0487J.a, to be updated as needed.
+ * We're not using the generated masks as they are usually ahead of
+ * the published ARM ARM, which we use as a reference.
+ *
+ * Once we get to a point where the two describe the same thing, we'll
+ * merge the definitions. One day.
+ */
+#define __HFGRTR_EL2_RES0 (GENMASK(63, 56) | GENMASK(53, 51))
+#define __HFGRTR_EL2_MASK GENMASK(49, 0)
+#define __HFGRTR_EL2_nMASK (GENMASK(55, 54) | BIT(50))
+
+#define __HFGWTR_EL2_RES0 (GENMASK(63, 56) | GENMASK(53, 51) | \
+ BIT(46) | BIT(42) | BIT(40) | BIT(28) | \
+ GENMASK(26, 25) | BIT(21) | BIT(18) | \
+ GENMASK(15, 14) | GENMASK(10, 9) | BIT(2))
+#define __HFGWTR_EL2_MASK GENMASK(49, 0)
+#define __HFGWTR_EL2_nMASK (GENMASK(55, 54) | BIT(50))
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK (~UL(0xf))
 /*
arch/arm64/kvm/hyp/include/hyp/switch.h +28 -25
index 4bddb8541bec..966295178aee 100644
@@ -70,20 +70,19 @@ static inline void __activate_traps_fpsimd32(struct kvm_vcpu *vcpu)
 }
 }
 
-static inline bool __hfgxtr_traps_required(void)
-{
- if (cpus_have_final_cap(ARM64_SME))
- return true;
-
- if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
- return true;
 
- return false;
-}
 
-static inline void __activate_traps_hfgxtr(void)
+static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 {
+ struct kvm_cpu_context *hctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 u64 r_clr = 0, w_clr = 0, r_set = 0, w_set = 0, tmp;
+ u64 r_val, w_val;
+
+ if (!cpus_have_final_cap(ARM64_HAS_FGT))
+ return;
+
+ ctxt_sys_reg(hctxt, HFGRTR_EL2) = read_sysreg_s(SYS_HFGRTR_EL2);
+ ctxt_sys_reg(hctxt, HFGWTR_EL2) = read_sysreg_s(SYS_HFGWTR_EL2);
 
 if (cpus_have_final_cap(ARM64_SME)) {
 tmp = HFGxTR_EL2_nSMPRI_EL1_MASK | HFGxTR_EL2_nTPIDR2_EL0_MASK;
@@ -98,26 +97,31 @@ static inline void __activate_traps_hfgxtr(void)
 if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
 w_set |= HFGxTR_EL2_TCR_EL1_MASK;
 
- sysreg_clear_set_s(SYS_HFGRTR_EL2, r_clr, r_set);
- sysreg_clear_set_s(SYS_HFGWTR_EL2, w_clr, w_set);
+
+ /* The default is not to trap amything but ACCDATA_EL1 */
+ r_val = __HFGRTR_EL2_nMASK & ~HFGxTR_EL2_nACCDATA_EL1;
+ r_val |= r_set;
+ r_val &= ~r_clr;
+
+ w_val = __HFGWTR_EL2_nMASK & ~HFGxTR_EL2_nACCDATA_EL1;
+ w_val |= w_set;
+ w_val &= ~w_clr;
+
+ write_sysreg_s(r_val, SYS_HFGRTR_EL2);
+ write_sysreg_s(w_val, SYS_HFGWTR_EL2);
 }
 
-static inline void __deactivate_traps_hfgxtr(void)
+static inline void __deactivate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 {
- u64 r_clr = 0, w_clr = 0, r_set = 0, w_set = 0, tmp;
+ struct kvm_cpu_context *hctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 
- if (cpus_have_final_cap(ARM64_SME)) {
- tmp = HFGxTR_EL2_nSMPRI_EL1_MASK | HFGxTR_EL2_nTPIDR2_EL0_MASK;
+ if (!cpus_have_final_cap(ARM64_HAS_FGT))
+ return;
 
- r_set |= tmp;
- w_set |= tmp;
- }
+ write_sysreg_s(ctxt_sys_reg(hctxt, HFGRTR_EL2), SYS_HFGRTR_EL2);
+ write_sysreg_s(ctxt_sys_reg(hctxt, HFGWTR_EL2), SYS_HFGWTR_EL2);
 
- if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
- w_clr |= HFGxTR_EL2_TCR_EL1_MASK;
 
- sysreg_clear_set_s(SYS_HFGRTR_EL2, r_clr, r_set);
- sysreg_clear_set_s(SYS_HFGWTR_EL2, w_clr, w_set);
 }
 
 static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
@@ -145,8 +149,7 @@ static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
 vcpu->arch.mdcr_el2_host = read_sysreg(mdcr_el2);
 write_sysreg(vcpu->arch.mdcr_el2, mdcr_el2);
 
- if (__hfgxtr_traps_required())
- __activate_traps_hfgxtr();
+ __activate_traps_hfgxtr(vcpu);
 }
 
 static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu)

arch/arm64/kvm/hyp/include/hyp/switch.h static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu) · line 168

Quoted 6 lines
arch/arm64/kvm/hyp/include/hyp/switch.h +1 -2
@@ -162,8 +165,7 @@ static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu)
 vcpu_clear_flag(vcpu, PMUSERENR_ON_CPU);
 }
 
- if (__hfgxtr_traps_required())
- __deactivate_traps_hfgxtr();
+ __deactivate_traps_hfgxtr(vcpu);

Thanks

Miguel

Quoted 6 lines
arch/arm64/kvm/hyp/include/hyp/switch.h +0 -0
 }
 
 static inline void ___activate_traps(struct kvm_vcpu *vcpu)
Signature · 1 line
2.34.1
author maintainer
#46 lore raw

PATCH v2 14/26 KVM: arm64: nv: Add trap forwarding infrastructure

In reply to Marc Zyngier

A significant part of what a NV hypervisor needs to do is to decide whether a trap from a L2+ guest has to be forwarded to a L1 guest or handled locally. This is done by checking for the trap bits that the guest hypervisor has set and acting accordingly, as described by the architecture.

A previous approach was to sprinkle a bunch of checks in all the system register accessors, but this is pretty error prone and doesn't help getting an overview of what is happening.

Instead, implement a set of global tables that describe a trap bit, combinations of trap bits, behaviours on trap, and what bits must be evaluated on a system register trap.

Although this is painful to describe, this allows to specify each and every control bit in a static manner. To make it efficient, the table is inserted in an xarray that is global to the system, and checked each time we trap a system register while running a L2 guest.

Add the basic infrastructure for now, while additional patches will implement configuration registers.

---
 arch/arm64/include/asm/kvm_host.h   |   1 +
 arch/arm64/include/asm/kvm_nested.h |   2 +
 arch/arm64/kvm/emulate-nested.c     | 211 ++++++++++++++++++++++++++++
 arch/arm64/kvm/sys_regs.c           |   6 +
 arch/arm64/kvm/trace_arm.h          |  19 +++
 5 files changed, 239 insertions(+)
arch/arm64/include/asm/kvm_host.h +1 -0
index 721680da1011..d44575604e7e 100644
@@ -988,6 +988,7 @@ int kvm_handle_cp10_id(struct kvm_vcpu *vcpu);
 void kvm_reset_sys_regs(struct kvm_vcpu *vcpu);
 
 int __init kvm_sys_reg_table_init(void);
+void __init populate_nv_trap_config(void);
 
 bool lock_all_vcpus(struct kvm *kvm);
 void unlock_all_vcpus(struct kvm *kvm);
arch/arm64/include/asm/kvm_nested.h +2 -0
index 8fb67f032fd1..fa23cc9c2adc 100644
@@ -11,6 +11,8 @@ static inline bool vcpu_has_nv(const struct kvm_vcpu *vcpu)
 		test_bit(KVM_ARM_VCPU_HAS_EL2, vcpu->arch.features));
 }
 
+extern bool __check_nv_sr_forward(struct kvm_vcpu *vcpu);
+
 struct sys_reg_params;
 struct sys_reg_desc;
 
arch/arm64/kvm/emulate-nested.c +211 -0

217 lines not on this page · raw message

arch/arm64/kvm/sys_regs.c +6 -0
index f5baaa508926..dfd72b3a625f 100644
@@ -3177,6 +3177,9 @@ int kvm_handle_sys_reg(struct kvm_vcpu *vcpu)
 
 	trace_kvm_handle_sys_reg(esr);
 
+	if (__check_nv_sr_forward(vcpu))
+		return 1;
+
 	params = esr_sys64_to_params(esr);
 	params.regval = vcpu_get_reg(vcpu, Rt);
 
@@ -3594,5 +3597,8 @@ int __init kvm_sys_reg_table_init(void)
 	if (!first_idreg)
 		return -EINVAL;
 
+	if (kvm_get_mode() == KVM_MODE_NV)
+		populate_nv_trap_config();
+
 	return 0;
 }
arch/arm64/kvm/trace_arm.h +19 -0
index 6ce5c025218d..1f0f3f653606 100644
@@ -364,6 +364,25 @@ TRACE_EVENT(kvm_inject_nested_exception,
 		  __entry->hcr_el2)
 );
 
+TRACE_EVENT(kvm_forward_sysreg_trap,
+	    TP_PROTO(struct kvm_vcpu *vcpu, u32 sysreg, bool is_read),
+	    TP_ARGS(vcpu, sysreg, is_read),
+
+	    TP_STRUCT__entry(
+		__field(struct kvm_vcpu *, vcpu)
+		__field(u32,		   sysreg)
+		__field(bool,		   is_read)
+	    ),
+
+	    TP_fast_assign(
+		__entry->vcpu = vcpu;
+		__entry->sysreg = sysreg;
+		__entry->is_read = is_read;
+	    ),
+
+	    TP_printk("%c %x", __entry->is_read ? 'R' : 'W', __entry->sysreg)
+);
+
 #endif /* _TRACE_ARM_ARM64_KVM_H */
 
 #undef TRACE_INCLUDE_PATH
Signature · 1 line
2.34.1
#47 lore raw

Re: [PATCH v2 14/26] KVM: arm64: nv: Add trap forwarding infrastructure

On Fri, Jul 28, 2023 at 09:29:40AM +0100, Marc Zyngier wrote:

[...]

Quoted 12 lines
arch/arm64/kvm/emulate-nested.c +12 -0
+/*
+ * Bit assignment for the trap controls. We use a 64bit word with the
+ * following layout for each trapped sysreg:
+ *
+ * [9:0]	enum trap_group (10 bits)
+ * [13:10]	enum fgt_group_id (4 bits)
+ * [19:14]	bit number in the FGT register (6 bits)
+ * [20]		trap polarity (1 bit)
+ * [62:21]	Unused (42 bits)
+ * [63]		RES0 - Must be zero, as lost on insertion in the xarray
+ */
+union trap_config {

arch/arm64/kvm/emulate-nested.c line 95

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+	u64	val;
+	struct {
+		unsigned long	cgt:10;	/* Coarse trap id */
+		unsigned long	fgt:4;	/* Fing Grained Trap id */
+		unsigned long	bit:6;	/* Bit number */
+		unsigned long	pol:1;	/* Polarity */
+		unsigned long	unk:42;	/* Unknown */
+		unsigned long	mbz:1;	/* Must Be Zero */
+	};
+};

Correct me if I'm wrong, but I don't think the compiler is going to whine if any of these bitfields are initialized with a larger value than can be represented... Do you think some BUILD_BUG_ON() is in order to ensure that trap_group fits in ::cgt?

	BUILD_BUG_ON(__NR_TRAP_IDS__ >= BIT(10));
Quoted 24 lines
arch/arm64/kvm/emulate-nested.c +24 -0
+struct encoding_to_trap_config {
+	const u32			encoding;
+	const u32			end;
+	const union trap_config		tc;
+};
+
+#define SR_RANGE_TRAP(sr_start, sr_end, trap_id)			\
+	{								\
+		.encoding	= sr_start,				\
+		.end		= sr_end,				\
+		.tc		= {					\
+			.cgt		= trap_id,			\
+		},							\
+	}
+
+#define SR_TRAP(sr, trap_id)		SR_RANGE_TRAP(sr, sr, trap_id)
+
+/*
+ * Map encoding to trap bits for exception reported with EC=0x18.
+ * These must only be evaluated when running a nested hypervisor, but
+ * that the current context is not a hypervisor context. When the
+ * trapped access matches one of the trap controls, the exception is
+ * re-injected in the nested hypervisor.
+ */

arch/arm64/kvm/emulate-nested.c line 130

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
+};
+
+static DEFINE_XARRAY(sr_forward_xa);
+
+static union trap_config get_trap_config(u32 sysreg)
+{
+	return (union trap_config) {
+		.val = xa_to_value(xa_load(&sr_forward_xa, sysreg)),
+	};

Should we be checking for NULL here? AFAICT, the use of sentinel values in the trap_group enum would effectively guarantee each trap_config has a nonzero value.

arch/arm64/kvm/emulate-nested.c line 141

Quoted 11 lines
arch/arm64/kvm/emulate-nested.c +11 -0
+}
+
+void __init populate_nv_trap_config(void)
+{
+	for (int i = 0; i < ARRAY_SIZE(encoding_to_cgt); i++) {
+		const struct encoding_to_trap_config *cgt = &encoding_to_cgt[i];
+		void *prev;
+
+		prev = xa_store_range(&sr_forward_xa, cgt->encoding, cgt->end,
+				      xa_mk_value(cgt->tc.val), GFP_KERNEL);
+		WARN_ON(prev);

Returning the error here and failing the overall KVM initialization seems to be the safest option. The WARN is still handy, though.

Quoted 26 lines
arch/arm64/kvm/emulate-nested.c +26 -0
+	}
+
+	kvm_info("nv: %ld coarse grained trap handlers\n",
+		 ARRAY_SIZE(encoding_to_cgt));
+
+}
+
+static enum trap_behaviour get_behaviour(struct kvm_vcpu *vcpu,
+					 const struct trap_bits *tb)
+{
+	enum trap_behaviour b = BEHAVE_HANDLE_LOCALLY;
+	u64 val;
+
+	val = __vcpu_sys_reg(vcpu, tb->index);
+	if ((val & tb->mask) == tb->value)
+		b |= tb->behaviour;
+
+	return b;
+}
+
+static enum trap_behaviour __do_compute_trap_behaviour(struct kvm_vcpu *vcpu,
+						       const enum trap_group id,
+						       enum trap_behaviour b)
+{
+	switch (id) {
+		const enum trap_group *cgids;

arch/arm64/kvm/emulate-nested.c line 177

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+
+	case __RESERVED__ ... __MULTIPLE_CONTROL_BITS__ - 1:
+		if (likely(id != __RESERVED__))
+			b |= get_behaviour(vcpu, &coarse_trap_bits[id]);
+		break;
+	case __MULTIPLE_CONTROL_BITS__ ... __COMPLEX_CONDITIONS__ - 1:
+		/* Yes, this is recursive. Don't do anything stupid. */
+		cgids = coarse_control_combo[id - __MULTIPLE_CONTROL_BITS__];
+		for (int i = 0; cgids[i] != __RESERVED__; i++)
+			b |= __do_compute_trap_behaviour(vcpu, cgids[i], b);

Would it make sense to WARN here if one of the child trap ids was in the recursive range?

Signature · 2 lines
Thanks,
Oliver
author maintainer
#48 lore raw

Re: [PATCH v2 14/26] KVM: arm64: nv: Add trap forwarding infrastructure

Quoted 33 lines · Oliver Upton

On Fri, Jul 28, 2023 at 09:29:40AM +0100, Marc Zyngier wrote:

[...]

Quoted 12 lines
arch/arm64/kvm/emulate-nested.c +12 -0
+/*
+ * Bit assignment for the trap controls. We use a 64bit word with the
+ * following layout for each trapped sysreg:
+ *
+ * [9:0]	enum trap_group (10 bits)
+ * [13:10]	enum fgt_group_id (4 bits)
+ * [19:14]	bit number in the FGT register (6 bits)
+ * [20]		trap polarity (1 bit)
+ * [62:21]	Unused (42 bits)
+ * [63]		RES0 - Must be zero, as lost on insertion in the xarray
+ */
+union trap_config {

arch/arm64/kvm/emulate-nested.c line 95

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+	u64	val;
+	struct {
+		unsigned long	cgt:10;	/* Coarse trap id */
+		unsigned long	fgt:4;	/* Fing Grained Trap id */
+		unsigned long	bit:6;	/* Bit number */
+		unsigned long	pol:1;	/* Polarity */
+		unsigned long	unk:42;	/* Unknown */
+		unsigned long	mbz:1;	/* Must Be Zero */
+	};
+};

Correct me if I'm wrong, but I don't think the compiler is going to whine if any of these bitfields are initialized with a larger value than can be represented... Do you think some BUILD_BUG_ON() is in order to ensure that trap_group fits in ::cgt?

Quoted 1 line
	BUILD_BUG_ON(__NR_TRAP_IDS__ >= BIT(10));

Indeed. This might also apply to ::fgt, and I want to add some sanity checks to verify that the whole union isn't larger than a 'void *', as we rely on that.

Quoted 31 lines
Quoted 31 lines
arch/arm64/kvm/emulate-nested.c +30 -0
+struct encoding_to_trap_config {
+	const u32			encoding;
+	const u32			end;
+	const union trap_config		tc;
+};
+
+#define SR_RANGE_TRAP(sr_start, sr_end, trap_id)			\
+	{								\
+		.encoding	= sr_start,				\
+		.end		= sr_end,				\
+		.tc		= {					\
+			.cgt		= trap_id,			\
+		},							\
+	}
+
+#define SR_TRAP(sr, trap_id)		SR_RANGE_TRAP(sr, sr, trap_id)
+
+/*
+ * Map encoding to trap bits for exception reported with EC=0x18.
+ * These must only be evaluated when running a nested hypervisor, but
+ * that the current context is not a hypervisor context. When the
+ * trapped access matches one of the trap controls, the exception is
+ * re-injected in the nested hypervisor.
+ */
+static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
+};
+
+static DEFINE_XARRAY(sr_forward_xa);
+
+static union trap_config get_trap_config(u32 sysreg)
Quoted 8 lines

arch/arm64/kvm/emulate-nested.c line 130

Quoted 4 lines
arch/arm64/kvm/emulate-nested.c +4 -0
+{
+	return (union trap_config) {
+		.val = xa_to_value(xa_load(&sr_forward_xa, sysreg)),
+	};

Should we be checking for NULL here? AFAICT, the use of sentinel values in the trap_group enum would effectively guarantee each trap_config has a nonzero value.

if xa_load() returns NULL, xa_to_value() will still give us a 0, which is an indication of a sysreg that isn't present in the trap configuration. This can happen if we trap something that isn't yet supported in NV, which is quite common. This allows us to use features on the host without having to immediately write the same support for NV guests.

But this is obviously a temporary situation. At some point, I'll become a total bastard and demand that people treat NV as a first class citizen. One day ;-).

Quoted 15 lines

arch/arm64/kvm/emulate-nested.c line 141

Quoted 12 lines
arch/arm64/kvm/emulate-nested.c +11 -0
+}
+
+void __init populate_nv_trap_config(void)
+{
+	for (int i = 0; i < ARRAY_SIZE(encoding_to_cgt); i++) {
+		const struct encoding_to_trap_config *cgt = &encoding_to_cgt[i];
+		void *prev;
+
+		prev = xa_store_range(&sr_forward_xa, cgt->encoding, cgt->end,
+				      xa_mk_value(cgt->tc.val), GFP_KERNEL);
+		WARN_ON(prev);

Returning the error here and failing the overall KVM initialization seems to be the safest option. The WARN is still handy, though.

Yeah, this has found a number of bugs initially. Happy to fail the initialisation after we've iterated over all the array (nothing is more annoying than fixing a bunch of errors iteratively).

Quoted 32 lines
Quoted 32 lines
arch/arm64/kvm/emulate-nested.c +31 -0
+	}
+
+	kvm_info("nv: %ld coarse grained trap handlers\n",
+		 ARRAY_SIZE(encoding_to_cgt));
+
+}
+
+static enum trap_behaviour get_behaviour(struct kvm_vcpu *vcpu,
+					 const struct trap_bits *tb)
+{
+	enum trap_behaviour b = BEHAVE_HANDLE_LOCALLY;
+	u64 val;
+
+	val = __vcpu_sys_reg(vcpu, tb->index);
+	if ((val & tb->mask) == tb->value)
+		b |= tb->behaviour;
+
+	return b;
+}
+
+static enum trap_behaviour __do_compute_trap_behaviour(struct kvm_vcpu *vcpu,
+						       const enum trap_group id,
+						       enum trap_behaviour b)
+{
+	switch (id) {
+		const enum trap_group *cgids;
+
+	case __RESERVED__ ... __MULTIPLE_CONTROL_BITS__ - 1:
+		if (likely(id != __RESERVED__))
+			b |= get_behaviour(vcpu, &coarse_trap_bits[id]);
+		break;
Quoted 8 lines

arch/arm64/kvm/emulate-nested.c line 177

Quoted 5 lines
arch/arm64/kvm/emulate-nested.c +5 -0
+	case __MULTIPLE_CONTROL_BITS__ ... __COMPLEX_CONDITIONS__ - 1:
+		/* Yes, this is recursive. Don't do anything stupid. */
+		cgids = coarse_control_combo[id - __MULTIPLE_CONTROL_BITS__];
+		for (int i = 0; cgids[i] != __RESERVED__; i++)
+			b |= __do_compute_trap_behaviour(vcpu, cgids[i], b);

Would it make sense to WARN here if one of the child trap ids was in the recursive range?

This might be needed at some point, but we can probably tighten it for now.

Thanks,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
#49 lore raw

Re: [PATCH v2 14/26] KVM: arm64: nv: Add trap forwarding infrastructure

Quoted 7 lines · Marc Zyngier
Quoted 5 lines · Oliver Upton

Correct me if I'm wrong, but I don't think the compiler is going to whine if any of these bitfields are initialized with a larger value than can be represented... Do you think some BUILD_BUG_ON() is in order to ensure that trap_group fits in ::cgt?

Quoted 5 lines
Quoted 1 line
	BUILD_BUG_ON(__NR_TRAP_IDS__ >= BIT(10));

Indeed. This might also apply to ::fgt, and I want to add some sanity checks to verify that the whole union isn't larger than a 'void *', as we rely on that.

Agreed on both.

Quoted 22 lines
Quoted 13 lines
Quoted 13 lines
arch/arm64/kvm/emulate-nested.c +13 -0
+/*
+ * Map encoding to trap bits for exception reported with EC=0x18.
+ * These must only be evaluated when running a nested hypervisor, but
+ * that the current context is not a hypervisor context. When the
+ * trapped access matches one of the trap controls, the exception is
+ * re-injected in the nested hypervisor.
+ */
+static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
+};
+
+static DEFINE_XARRAY(sr_forward_xa);
+
+static union trap_config get_trap_config(u32 sysreg)
Quoted 8 lines

arch/arm64/kvm/emulate-nested.c line 130

Quoted 4 lines
arch/arm64/kvm/emulate-nested.c +4 -0
+{
+	return (union trap_config) {
+		.val = xa_to_value(xa_load(&sr_forward_xa, sysreg)),
+	};

Should we be checking for NULL here? AFAICT, the use of sentinel values in the trap_group enum would effectively guarantee each trap_config has a nonzero value.

Quoted 6 lines

if xa_load() returns NULL, xa_to_value() will still give us a 0, which is an indication of a sysreg that isn't present in the trap configuration. This can happen if we trap something that isn't yet supported in NV, which is quite common. This allows us to use features on the host without having to immediately write the same support for NV guests.

That's fine by me, I couldn't piece it together where a value of 0 is explicitly handled as having no trap config.

Quoted 3 lines

But this is obviously a temporary situation. At some point, I'll become a total bastard and demand that people treat NV as a first class citizen. One day ;-).

No time like the present!

--
Thanks,
Oliver

author maintainer
#50 lore raw

PATCH v2 15/26 KVM: arm64: nv: Add trap forwarding for HCR_EL2

In reply to Marc Zyngier

Describe the HCR_EL2 register, and associate it with all the sysregs it allows to trap.

---
 arch/arm64/kvm/emulate-nested.c | 486 ++++++++++++++++++++++++++++++++
 1 file changed, 486 insertions(+)
arch/arm64/kvm/emulate-nested.c +486 -0
index 7d4288de1df6..094534aa0f72 100644
@@ -37,12 +37,48 @@ enum trap_group {
 	 * on their own instead of being part of a combination of
 	 * trap controls.
 	 */
+	CGT_HCR_TID1,
+	CGT_HCR_TID2,
+	CGT_HCR_TID3,
+	CGT_HCR_IMO,
+	CGT_HCR_FMO,
+	CGT_HCR_TIDCP,
+	CGT_HCR_TACR,
+	CGT_HCR_TSW,
+	CGT_HCR_TPC,
+	CGT_HCR_TPU,
+	CGT_HCR_TTLB,
+	CGT_HCR_TVM,
+	CGT_HCR_TDZ,
+	CGT_HCR_TRVM,
+	CGT_HCR_TLOR,
+	CGT_HCR_TERR,
+	CGT_HCR_APK,
+	CGT_HCR_NV,
+	CGT_HCR_NV_nNV2,
+	CGT_HCR_NV1_nNV2,
+	CGT_HCR_AT,
+	CGT_HCR_FIEN,
+	CGT_HCR_TID4,
+	CGT_HCR_TICAB,
+	CGT_HCR_TOCU,
+	CGT_HCR_ENSCXT,
+	CGT_HCR_TTLBIS,
+	CGT_HCR_TTLBOS,
 
 	/*
 	 * Anything after this point is a combination of trap controls,
 	 * which all must be evaluated to decide what to do.
 	 */
 	__MULTIPLE_CONTROL_BITS__,
+	CGT_HCR_IMO_FMO = __MULTIPLE_CONTROL_BITS__,
+	CGT_HCR_TID2_TID4,
+	CGT_HCR_TTLB_TTLBIS,
+	CGT_HCR_TTLB_TTLBOS,
+	CGT_HCR_TVM_TRVM,
+	CGT_HCR_TPU_TICAB,
+	CGT_HCR_TPU_TOCU,
+	CGT_HCR_NV1_nNV2_ENSCXT,
 
 	/*
 	 * Anything after this point requires a callback evaluating a
@@ -52,6 +88,174 @@ enum trap_group {
 };
 
 static const struct trap_bits coarse_trap_bits[] = {
+	[CGT_HCR_TID1] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_TID1,
+		.mask		= HCR_TID1,
+		.behaviour	= BEHAVE_FORWARD_READ,
+	},
+	[CGT_HCR_TID2] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_TID2,
+		.mask		= HCR_TID2,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TID3] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_TID3,
+		.mask		= HCR_TID3,
+		.behaviour	= BEHAVE_FORWARD_READ,
+	},
+	[CGT_HCR_IMO] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_IMO,
+		.mask		= HCR_IMO,
+		.behaviour	= BEHAVE_FORWARD_WRITE,
+	},
+	[CGT_HCR_FMO] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_FMO,
+		.mask		= HCR_FMO,
+		.behaviour	= BEHAVE_FORWARD_WRITE,
+	},
+	[CGT_HCR_TIDCP] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TIDCP,
+		.mask		= HCR_TIDCP,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TACR] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TACR,
+		.mask		= HCR_TACR,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TSW] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TSW,
+		.mask		= HCR_TSW,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TPC] = { /* Also called TCPC when FEAT_DPB is implemented */
+		.index		= HCR_EL2,
+		.value		= HCR_TPC,
+		.mask		= HCR_TPC,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TPU] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TPU,
+		.mask		= HCR_TPU,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TTLB] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TTLB,
+		.mask		= HCR_TTLB,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TVM] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TVM,
+		.mask		= HCR_TVM,
+		.behaviour	= BEHAVE_FORWARD_WRITE,
+	},
+	[CGT_HCR_TDZ] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TDZ,
+		.mask		= HCR_TDZ,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TRVM] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TRVM,
+		.mask		= HCR_TRVM,
+		.behaviour	= BEHAVE_FORWARD_READ,
+	},
+	[CGT_HCR_TLOR] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TLOR,
+		.mask		= HCR_TLOR,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TERR] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TERR,
+		.mask		= HCR_TERR,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_APK] = {
+		.index		= HCR_EL2,
+		.value		= 0,
+		.mask		= HCR_APK,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_NV] = {
+		.index		= HCR_EL2,
+		.value		= HCR_NV,
+		.mask		= HCR_NV,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_NV_nNV2] = {
+		.index		= HCR_EL2,
+		.value		= HCR_NV,
+		.mask		= HCR_NV | HCR_NV2,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_NV1_nNV2] = {
+		.index		= HCR_EL2,
+		.value		= HCR_NV | HCR_NV1,
+		.mask		= HCR_NV | HCR_NV1 | HCR_NV2,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_AT] = {
+		.index		= HCR_EL2,
+		.value		= HCR_AT,
+		.mask		= HCR_AT,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_FIEN] = {
+		.index		= HCR_EL2,
+		.value		= HCR_FIEN,
+		.mask		= HCR_FIEN,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TID4] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_TID4,
+		.mask		= HCR_TID4,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TICAB] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_TICAB,
+		.mask		= HCR_TICAB,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TOCU] = {
+		.index		= HCR_EL2,
+		.value 		= HCR_TOCU,
+		.mask		= HCR_TOCU,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_ENSCXT] = {
+		.index		= HCR_EL2,
+		.value 		= 0,
+		.mask		= HCR_ENSCXT,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TTLBIS] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TTLBIS,
+		.mask		= HCR_TTLBIS,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_HCR_TTLBOS] = {
+		.index		= HCR_EL2,
+		.value		= HCR_TTLBOS,
+		.mask		= HCR_TTLBOS,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
 };
 
 #define MCB(id, ...)					\
@@ -61,6 +265,14 @@ static const struct trap_bits coarse_trap_bits[] = {
 		}
 
 static const enum trap_group *coarse_control_combo[] = {
+	MCB(CGT_HCR_IMO_FMO,		CGT_HCR_IMO, CGT_HCR_FMO),
+	MCB(CGT_HCR_TID2_TID4,		CGT_HCR_TID2, CGT_HCR_TID4),
+	MCB(CGT_HCR_TTLB_TTLBIS,	CGT_HCR_TTLB, CGT_HCR_TTLBIS),
+	MCB(CGT_HCR_TTLB_TTLBOS,	CGT_HCR_TTLB, CGT_HCR_TTLBOS),
+	MCB(CGT_HCR_TVM_TRVM,		CGT_HCR_TVM, CGT_HCR_TRVM),
+	MCB(CGT_HCR_TPU_TICAB,		CGT_HCR_TPU, CGT_HCR_TICAB),
+	MCB(CGT_HCR_TPU_TOCU,		CGT_HCR_TPU, CGT_HCR_TOCU),
+	MCB(CGT_HCR_NV1_nNV2_ENSCXT,	CGT_HCR_NV1_nNV2, CGT_HCR_ENSCXT),
 };
 
 typedef enum trap_behaviour (*complex_condition_check)(struct kvm_vcpu *);
@@ -119,6 +331,280 @@ struct encoding_to_trap_config {
  * re-injected in the nested hypervisor.
  */
 static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
+	SR_TRAP(SYS_REVIDR_EL1,		CGT_HCR_TID1),
+	SR_TRAP(SYS_AIDR_EL1,		CGT_HCR_TID1),
+	SR_TRAP(SYS_SMIDR_EL1,		CGT_HCR_TID1),
+	SR_TRAP(SYS_CTR_EL0,		CGT_HCR_TID2),
+	SR_TRAP(SYS_CCSIDR_EL1,		CGT_HCR_TID2_TID4),
+	SR_TRAP(SYS_CCSIDR2_EL1,	CGT_HCR_TID2_TID4),
+	SR_TRAP(SYS_CLIDR_EL1,		CGT_HCR_TID2_TID4),
+	SR_TRAP(SYS_CSSELR_EL1,		CGT_HCR_TID2_TID4),
+	SR_RANGE_TRAP(SYS_ID_PFR0_EL1,
+		      sys_reg(3, 0, 0, 7, 7), CGT_HCR_TID3),
+	SR_TRAP(SYS_ICC_SGI0R_EL1,	CGT_HCR_IMO_FMO),
+	SR_TRAP(SYS_ICC_ASGI1R_EL1,	CGT_HCR_IMO_FMO),
+	SR_TRAP(SYS_ICC_SGI1R_EL1,	CGT_HCR_IMO_FMO),
+	SR_RANGE_TRAP(sys_reg(3, 0, 11, 0, 0),
+		      sys_reg(3, 0, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 1, 11, 0, 0),
+		      sys_reg(3, 1, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 2, 11, 0, 0),
+		      sys_reg(3, 2, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 3, 11, 0, 0),
+		      sys_reg(3, 3, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 4, 11, 0, 0),
+		      sys_reg(3, 4, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 5, 11, 0, 0),
+		      sys_reg(3, 5, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 6, 11, 0, 0),
+		      sys_reg(3, 6, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 7, 11, 0, 0),
+		      sys_reg(3, 7, 11, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 0, 15, 0, 0),
+		      sys_reg(3, 0, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 1, 15, 0, 0),
+		      sys_reg(3, 1, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 2, 15, 0, 0),
+		      sys_reg(3, 2, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 3, 15, 0, 0),
+		      sys_reg(3, 3, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 4, 15, 0, 0),
+		      sys_reg(3, 4, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 5, 15, 0, 0),
+		      sys_reg(3, 5, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 6, 15, 0, 0),
+		      sys_reg(3, 6, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_RANGE_TRAP(sys_reg(3, 7, 15, 0, 0),
+		      sys_reg(3, 7, 15, 15, 7), CGT_HCR_TIDCP),
+	SR_TRAP(SYS_ACTLR_EL1,		CGT_HCR_TACR),
+	SR_TRAP(SYS_DC_ISW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CISW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_IGSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_IGDSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CGSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CGDSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CIGSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CIGDSW,		CGT_HCR_TSW),
+	SR_TRAP(SYS_DC_CIVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CVAP,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CVADP,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_IVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CIGVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CIGDVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_IGVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_IGDVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CGVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CGDVAC,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CGVAP,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CGDVAP,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CGVADP,		CGT_HCR_TPC),
+	SR_TRAP(SYS_DC_CGDVADP,		CGT_HCR_TPC),
+	SR_TRAP(SYS_IC_IVAU,		CGT_HCR_TPU_TOCU),
+	SR_TRAP(SYS_IC_IALLU,		CGT_HCR_TPU_TOCU),
+	SR_TRAP(SYS_IC_IALLUIS,		CGT_HCR_TPU_TICAB),
+	SR_TRAP(SYS_DC_CVAU,		CGT_HCR_TPU_TOCU),
+	SR_TRAP(OP_TLBI_RVAE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVAAE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVALE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVAALE1,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VMALLE1,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VAE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_ASIDE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VAAE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VALE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VAALE1,		CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVAE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVAAE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVALE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVAALE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VMALLE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VAE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_ASIDE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VAAE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VALE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_VAALE1NXS,	CGT_HCR_TTLB),
+	SR_TRAP(OP_TLBI_RVAE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVAAE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVALE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVAALE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VMALLE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VAE1IS,		CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_ASIDE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VAAE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VALE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VAALE1IS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVAE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVAAE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVALE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_RVAALE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VMALLE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VAE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_ASIDE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VAAE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VALE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VAALE1ISNXS,	CGT_HCR_TTLB_TTLBIS),
+	SR_TRAP(OP_TLBI_VMALLE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VAE1OS,		CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_ASIDE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VAAE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VALE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VAALE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVAE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVAAE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVALE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVAALE1OS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VMALLE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VAE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_ASIDE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VAAE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VALE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_VAALE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVAE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVAAE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVALE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(OP_TLBI_RVAALE1OSNXS,	CGT_HCR_TTLB_TTLBOS),
+	SR_TRAP(SYS_SCTLR_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_TTBR0_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_TTBR1_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_TCR_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_ESR_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_FAR_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_AFSR0_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_AFSR1_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_MAIR_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_AMAIR_EL1,		CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_CONTEXTIDR_EL1,	CGT_HCR_TVM_TRVM),
+	SR_TRAP(SYS_DC_ZVA,		CGT_HCR_TDZ),
+	SR_TRAP(SYS_DC_GVA,		CGT_HCR_TDZ),
+	SR_TRAP(SYS_DC_GZVA,		CGT_HCR_TDZ),
+	SR_TRAP(SYS_LORSA_EL1,		CGT_HCR_TLOR),
+	SR_TRAP(SYS_LOREA_EL1, 		CGT_HCR_TLOR),
+	SR_TRAP(SYS_LORN_EL1, 		CGT_HCR_TLOR),
+	SR_TRAP(SYS_LORC_EL1, 		CGT_HCR_TLOR),
+	SR_TRAP(SYS_LORID_EL1,		CGT_HCR_TLOR),
+	SR_TRAP(SYS_ERRIDR_EL1,		CGT_HCR_TERR),
+	SR_TRAP(SYS_ERRSELR_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXADDR_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXCTLR_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXFR_EL1,		CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXMISC0_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXMISC1_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXMISC2_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXMISC3_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_ERXSTATUS_EL1,	CGT_HCR_TERR),
+	SR_TRAP(SYS_APIAKEYLO_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APIAKEYHI_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APIBKEYLO_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APIBKEYHI_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APDAKEYLO_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APDAKEYHI_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APDBKEYLO_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APDBKEYHI_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APGAKEYLO_EL1,	CGT_HCR_APK),
+	SR_TRAP(SYS_APGAKEYHI_EL1,	CGT_HCR_APK),
+	/* All _EL2 registers */
+	SR_RANGE_TRAP(sys_reg(3, 4, 0, 0, 0),
+		      sys_reg(3, 4, 10, 15, 7), CGT_HCR_NV),
+	SR_RANGE_TRAP(sys_reg(3, 4, 12, 0, 0),
+		      sys_reg(3, 4, 14, 15, 7), CGT_HCR_NV),
+	/* All _EL02, _EL12 registers */
+	SR_RANGE_TRAP(sys_reg(3, 5, 0, 0, 0),
+		      sys_reg(3, 5, 10, 15, 7), CGT_HCR_NV),
+	SR_RANGE_TRAP(sys_reg(3, 5, 12, 0, 0),
+		      sys_reg(3, 5, 14, 15, 7), CGT_HCR_NV),
+	SR_TRAP(OP_AT_S1E2R,		CGT_HCR_NV),
+	SR_TRAP(OP_AT_S1E2W,		CGT_HCR_NV),
+	SR_TRAP(OP_AT_S12E1R,		CGT_HCR_NV),
+	SR_TRAP(OP_AT_S12E1W,		CGT_HCR_NV),
+	SR_TRAP(OP_AT_S12E0R,		CGT_HCR_NV),
+	SR_TRAP(OP_AT_S12E0W,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2E1,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2E1,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2LE1,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2LE1,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVAE2,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVALE2,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE2,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VAE2,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE1,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VALE2,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VMALLS12E1,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2E1NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2E1NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2LE1NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2LE1NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVAE2NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVALE2NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE2NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VAE2NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE1NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VALE2NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VMALLS12E1NXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2E1IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2E1IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2LE1IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2LE1IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVAE2IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVALE2IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE2IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VAE2IS,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE1IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VALE2IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VMALLS12E1IS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2E1ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2E1ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2LE1ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2LE1ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVAE2ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVALE2ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE2ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VAE2ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE1ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VALE2ISNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VMALLS12E1ISNXS,CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE2OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VAE2OS,		CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE1OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VALE2OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VMALLS12E1OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2E1OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2E1OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2LE1OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2LE1OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVAE2OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVALE2OS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE2OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VAE2OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_ALLE1OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VALE2OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_VMALLS12E1OSNXS,CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2E1OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2E1OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_IPAS2LE1OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RIPAS2LE1OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVAE2OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_TLBI_RVALE2OSNXS,	CGT_HCR_NV),
+	SR_TRAP(OP_CPP_RCTX, 		CGT_HCR_NV),
+	SR_TRAP(OP_DVP_RCTX, 		CGT_HCR_NV),
+	SR_TRAP(OP_CFP_RCTX, 		CGT_HCR_NV),
+	SR_TRAP(SYS_SP_EL1,		CGT_HCR_NV_nNV2),
+	SR_TRAP(SYS_VBAR_EL1,		CGT_HCR_NV1_nNV2),
+	SR_TRAP(SYS_ELR_EL1,		CGT_HCR_NV1_nNV2),
+	SR_TRAP(SYS_SPSR_EL1,		CGT_HCR_NV1_nNV2),
+	SR_TRAP(SYS_SCXTNUM_EL1,	CGT_HCR_NV1_nNV2_ENSCXT),
+	SR_TRAP(SYS_SCXTNUM_EL0,	CGT_HCR_ENSCXT),
+	SR_TRAP(OP_AT_S1E1R, 		CGT_HCR_AT),
+	SR_TRAP(OP_AT_S1E1W, 		CGT_HCR_AT),
+	SR_TRAP(OP_AT_S1E0R, 		CGT_HCR_AT),
+	SR_TRAP(OP_AT_S1E0W, 		CGT_HCR_AT),
+	SR_TRAP(OP_AT_S1E1RP, 		CGT_HCR_AT),
+	SR_TRAP(OP_AT_S1E1WP, 		CGT_HCR_AT),
+	SR_TRAP(SYS_ERXPFGF_EL1,	CGT_HCR_FIEN),
+	SR_TRAP(SYS_ERXPFGCTL_EL1,	CGT_HCR_FIEN),
+	SR_TRAP(SYS_ERXPFGCDN_EL1,	CGT_HCR_FIEN),
+	SR_TRAP(SYS_SCXTNUM_EL0,	CGT_HCR_ENSCXT),
 };
 
 static DEFINE_XARRAY(sr_forward_xa);
Signature · 1 line
2.34.1
author maintainer
#51 lore raw

PATCH v2 16/26 KVM: arm64: nv: Expose FEAT_EVT to nested guests

In reply to Marc Zyngier

Now that we properly implement FEAT_EVT (as we correctly forward traps), expose it to guests.

---
 arch/arm64/kvm/nested.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
arch/arm64/kvm/nested.c +1 -2
index 315354d27978..7f80f385d9e8 100644
@@ -124,8 +124,7 @@ void access_nested_id_reg(struct kvm_vcpu *v, struct sys_reg_params *p,
 		break;
 
 	case SYS_ID_AA64MMFR2_EL1:
-		val &= ~(NV_FTR(MMFR2, EVT)	|
-			 NV_FTR(MMFR2, BBM)	|
+		val &= ~(NV_FTR(MMFR2, BBM)	|
 			 NV_FTR(MMFR2, TTL)	|
 			 GENMASK_ULL(47, 44)	|
 			 NV_FTR(MMFR2, ST)	|
Signature · 1 line
2.34.1
author maintainer
#52 lore raw

PATCH v2 17/26 KVM: arm64: nv: Add trap forwarding for MDCR_EL2

In reply to Marc Zyngier

Describe the MDCR_EL2 register, and associate it with all the sysregs it allows to trap.

---
 arch/arm64/kvm/emulate-nested.c | 265 ++++++++++++++++++++++++++++++++
 1 file changed, 265 insertions(+)
arch/arm64/kvm/emulate-nested.c +265 -0
index 094534aa0f72..339976764ac5 100644
@@ -66,6 +66,18 @@ enum trap_group {
 	CGT_HCR_TTLBIS,
 	CGT_HCR_TTLBOS,
 
+	CGT_MDCR_TPMCR,
+	CGT_MDCR_TPM,
+	CGT_MDCR_TDE,
+	CGT_MDCR_TDA,
+	CGT_MDCR_TDOSA,
+	CGT_MDCR_TDRA,
+	CGT_MDCR_E2PB,
+	CGT_MDCR_TPMS,
+	CGT_MDCR_TTRF,
+	CGT_MDCR_E2TB,
+	CGT_MDCR_TDCC,
+
 	/*
 	 * Anything after this point is a combination of trap controls,
 	 * which all must be evaluated to decide what to do.
@@ -79,6 +91,11 @@ enum trap_group {
 	CGT_HCR_TPU_TICAB,
 	CGT_HCR_TPU_TOCU,
 	CGT_HCR_NV1_nNV2_ENSCXT,
+	CGT_MDCR_TPM_TPMCR,
+	CGT_MDCR_TDE_TDA,
+	CGT_MDCR_TDE_TDOSA,
+	CGT_MDCR_TDE_TDRA,
+	CGT_MDCR_TDCC_TDE_TDA,
 
 	/*
 	 * Anything after this point requires a callback evaluating a
@@ -256,6 +273,72 @@ static const struct trap_bits coarse_trap_bits[] = {
 		.mask		= HCR_TTLBOS,
 		.behaviour	= BEHAVE_FORWARD_ANY,
 	},
+	[CGT_MDCR_TPMCR] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TPMCR,
+		.mask		= MDCR_EL2_TPMCR,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TPM] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TPM,
+		.mask		= MDCR_EL2_TPM,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TDE] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TDE,
+		.mask		= MDCR_EL2_TDE,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TDA] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TDA,
+		.mask		= MDCR_EL2_TDA,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TDOSA] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TDOSA,
+		.mask		= MDCR_EL2_TDOSA,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TDRA] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TDRA,
+		.mask		= MDCR_EL2_TDRA,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_E2PB] = {
+		.index		= MDCR_EL2,
+		.value		= 0,
+		.mask		= BIT(MDCR_EL2_E2PB_SHIFT),
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TPMS] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TPMS,
+		.mask		= MDCR_EL2_TPMS,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TTRF] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TTRF,
+		.mask		= MDCR_EL2_TTRF,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_E2TB] = {
+		.index		= MDCR_EL2,
+		.value		= 0,
+		.mask		= BIT(MDCR_EL2_E2TB_SHIFT),
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
+	[CGT_MDCR_TDCC] = {
+		.index		= MDCR_EL2,
+		.value		= MDCR_EL2_TDCC,
+		.mask		= MDCR_EL2_TDCC,
+		.behaviour	= BEHAVE_FORWARD_ANY,
+	},
 };
 
 #define MCB(id, ...)					\
@@ -273,6 +356,11 @@ static const enum trap_group *coarse_control_combo[] = {
 	MCB(CGT_HCR_TPU_TICAB,		CGT_HCR_TPU, CGT_HCR_TICAB),
 	MCB(CGT_HCR_TPU_TOCU,		CGT_HCR_TPU, CGT_HCR_TOCU),
 	MCB(CGT_HCR_NV1_nNV2_ENSCXT,	CGT_HCR_NV1_nNV2, CGT_HCR_ENSCXT),
+	MCB(CGT_MDCR_TPM_TPMCR,		CGT_MDCR_TPM, CGT_MDCR_TPMCR),
+	MCB(CGT_MDCR_TDE_TDA,		CGT_MDCR_TDE, CGT_MDCR_TDA),
+	MCB(CGT_MDCR_TDE_TDOSA,		CGT_MDCR_TDE, CGT_MDCR_TDOSA),
+	MCB(CGT_MDCR_TDE_TDRA,		CGT_MDCR_TDE, CGT_MDCR_TDRA),
+	MCB(CGT_MDCR_TDCC_TDE_TDA,	CGT_MDCR_TDCC, CGT_MDCR_TDE_TDA),
 };
 
 typedef enum trap_behaviour (*complex_condition_check)(struct kvm_vcpu *);
@@ -605,6 +693,183 @@ static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
 	SR_TRAP(SYS_ERXPFGCTL_EL1,	CGT_HCR_FIEN),
 	SR_TRAP(SYS_ERXPFGCDN_EL1,	CGT_HCR_FIEN),
 	SR_TRAP(SYS_SCXTNUM_EL0,	CGT_HCR_ENSCXT),
+	SR_TRAP(SYS_PMCR_EL0,		CGT_MDCR_TPM_TPMCR),
+	SR_TRAP(SYS_PMCNTENSET_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMCNTENCLR_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMOVSSET_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMOVSCLR_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMCEID0_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMCEID1_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMXEVTYPER_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMSWINC_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMSELR_EL0,		CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMXEVCNTR_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMCCNTR_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMUSERENR_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMINTENSET_EL1,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMINTENCLR_EL1,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMMIR_EL1,		CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(0),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(1),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(2),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(3),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(4),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(5),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(6),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(7),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(8),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(9),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(10),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(11),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(12),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(13),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(14),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(15),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(16),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(17),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(18),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(19),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(20),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(21),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(22),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(23),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(24),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(25),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(26),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(27),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(28),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(29),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVCNTRn_EL0(30),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(0),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(1),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(2),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(3),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(4),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(5),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(6),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(7),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(8),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(9),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(10),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(11),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(12),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(13),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(14),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(15),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(16),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(17),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(18),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(19),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(20),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(21),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(22),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(23),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(24),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(25),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(26),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(27),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(28),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(29),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMEVTYPERn_EL0(30),	CGT_MDCR_TPM),
+	SR_TRAP(SYS_PMCCFILTR_EL0,	CGT_MDCR_TPM),
+	SR_TRAP(SYS_MDCCSR_EL0,		CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_MDCCINT_EL1,	CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_OSDTRRX_EL1,	CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_OSDTRTX_EL1,	CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_DBGDTR_EL0,		CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_DBGDTRRX_EL0,	CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_DBGDTRTX_EL0,	CGT_MDCR_TDCC_TDE_TDA),
+	SR_TRAP(SYS_MDSCR_EL1,		CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_OSECCR_EL1,		CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(0),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(1),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(2),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(3),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(4),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(5),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(6),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(7),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(8),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(9),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(10),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(11),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(12),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(13),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(14),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBVRn_EL1(15),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(0),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(1),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(2),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(3),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(4),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(5),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(6),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(7),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(8),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(9),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(10),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(11),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(12),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(13),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(14),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGBCRn_EL1(15),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(0),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(1),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(2),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(3),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(4),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(5),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(6),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(7),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(8),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(9),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(10),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(11),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(12),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(13),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(14),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWVRn_EL1(15),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(0),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(1),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(2),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(3),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(4),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(5),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(6),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(7),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(8),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(9),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(10),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(11),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(12),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(13),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGWCRn_EL1(14),	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGCLAIMSET_EL1,	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGCLAIMCLR_EL1,	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_DBGAUTHSTATUS_EL1,	CGT_MDCR_TDE_TDA),
+	SR_TRAP(SYS_OSLAR_EL1,		CGT_MDCR_TDE_TDOSA),
+	SR_TRAP(SYS_OSLSR_EL1,		CGT_MDCR_TDE_TDOSA),
+	SR_TRAP(SYS_OSDLR_EL1,		CGT_MDCR_TDE_TDOSA),
+	SR_TRAP(SYS_DBGPRCR_EL1,	CGT_MDCR_TDE_TDOSA),
+	SR_TRAP(SYS_MDRAR_EL1,		CGT_MDCR_TDE_TDRA),
+	SR_TRAP(SYS_PMBLIMITR_EL1,	CGT_MDCR_E2PB),
+	SR_TRAP(SYS_PMBPTR_EL1,		CGT_MDCR_E2PB),
+	SR_TRAP(SYS_PMBSR_EL1,		CGT_MDCR_E2PB),
+	SR_TRAP(SYS_PMSCR_EL1,		CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSEVFR_EL1,	CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSFCR_EL1,		CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSICR_EL1,		CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSIDR_EL1,		CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSIRR_EL1,		CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSLATFR_EL1,	CGT_MDCR_TPMS),
+	SR_TRAP(SYS_PMSNEVFR_EL1,	CGT_MDCR_TPMS),
+	SR_TRAP(SYS_TRFCR_EL1,		CGT_MDCR_TTRF),
+	SR_TRAP(SYS_TRBBASER_EL1,	CGT_MDCR_E2TB),
+	SR_TRAP(SYS_TRBLIMITR_EL1,	CGT_MDCR_E2TB),
+	SR_TRAP(SYS_TRBMAR_EL1, 	CGT_MDCR_E2TB),
+	SR_TRAP(SYS_TRBPTR_EL1, 	CGT_MDCR_E2TB),
+	SR_TRAP(SYS_TRBSR_EL1, 		CGT_MDCR_E2TB),
+	SR_TRAP(SYS_TRBTRG_EL1,		CGT_MDCR_E2TB),
 };
 
 static DEFINE_XARRAY(sr_forward_xa);
Signature · 1 line
2.34.1
author maintainer
#53 lore raw

PATCH v2 18/26 KVM: arm64: nv: Add trap forwarding for CNTHCTL_EL2

In reply to Marc Zyngier

Describe the CNTHCTL_EL2 register, and associate it with all the sysregs it allows to trap.

---
 arch/arm64/kvm/emulate-nested.c | 50 ++++++++++++++++++++++++++++++++-
 1 file changed, 49 insertions(+), 1 deletion(-)
arch/arm64/kvm/emulate-nested.c +49 -1
index 339976764ac5..74b67895c791 100644
@@ -99,9 +99,11 @@ enum trap_group {
 
 	/*
 	 * Anything after this point requires a callback evaluating a
-	 * complex trap condition. Hopefully we'll never need this...
+	 * complex trap condition. Ugly stuff.
 	 */
 	__COMPLEX_CONDITIONS__,
+	CGT_CNTHCTL_EL1PCTEN = __COMPLEX_CONDITIONS__,
+	CGT_CNTHCTL_EL1PTEN,
 };
 
 static const struct trap_bits coarse_trap_bits[] = {
@@ -365,10 +367,51 @@ static const enum trap_group *coarse_control_combo[] = {
 
 typedef enum trap_behaviour (*complex_condition_check)(struct kvm_vcpu *);
 
+/*
+ * Warning, maximum confusion ahead.
+ *
+ * When E2H=0, CNTHCTL_EL2[1:0] are defined as EL1PCEN:EL1PCTEN
+ * When E2H=1, CNTHCTL_EL2[11:10] are defined as EL1PTEN:EL1PCTEN
+ *
+ * Note the single letter difference? Yet, the bits have the same
+ * function despite a different layout and a different name.
+ *
+ * We don't try to reconcile this mess. We just use the E2H=0 bits
+ * to generate something that is in the E2H=1 format, and live with
+ * it. You're welcome.
+ */
+static u64 get_sanitized_cnthctl(struct kvm_vcpu *vcpu)
+{
+	u64 val = __vcpu_sys_reg(vcpu, CNTHCTL_EL2);
+
+	if (!vcpu_el2_e2h_is_set(vcpu))
+		val = (val & (CNTHCTL_EL1PCEN | CNTHCTL_EL1PCTEN)) << 10;
+
+	return val & ((CNTHCTL_EL1PCEN | CNTHCTL_EL1PCTEN) << 10);
+}
+
+static enum trap_behaviour check_cnthctl_el1pcten(struct kvm_vcpu *vcpu)
+{
+	if (get_sanitized_cnthctl(vcpu) & (CNTHCTL_EL1PCTEN << 10))
+		return BEHAVE_HANDLE_LOCALLY;
+
+	return BEHAVE_FORWARD_ANY;
+}
+
+static enum trap_behaviour check_cnthctl_el1pten(struct kvm_vcpu *vcpu)
+{
+	if (get_sanitized_cnthctl(vcpu) & (CNTHCTL_EL1PCEN << 10))
+		return BEHAVE_HANDLE_LOCALLY;
+
+	return BEHAVE_FORWARD_ANY;
+}
+
 #define CCC(id, fn)				\
 	[id - __COMPLEX_CONDITIONS__] = fn
 
 static const complex_condition_check ccc[] = {
+	CCC(CGT_CNTHCTL_EL1PCTEN, check_cnthctl_el1pcten),
+	CCC(CGT_CNTHCTL_EL1PTEN, check_cnthctl_el1pten),
 };
 
 /*
@@ -870,6 +913,11 @@ static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
 	SR_TRAP(SYS_TRBPTR_EL1, 	CGT_MDCR_E2TB),
 	SR_TRAP(SYS_TRBSR_EL1, 		CGT_MDCR_E2TB),
 	SR_TRAP(SYS_TRBTRG_EL1,		CGT_MDCR_E2TB),
+	SR_TRAP(SYS_CNTP_TVAL_EL0,	CGT_CNTHCTL_EL1PTEN),
+	SR_TRAP(SYS_CNTP_CVAL_EL0,	CGT_CNTHCTL_EL1PTEN),
+	SR_TRAP(SYS_CNTP_CTL_EL0,	CGT_CNTHCTL_EL1PTEN),
+	SR_TRAP(SYS_CNTPCT_EL0,		CGT_CNTHCTL_EL1PCTEN),
+	SR_TRAP(SYS_CNTPCTSS_EL0,	CGT_CNTHCTL_EL1PCTEN),
 };
 
 static DEFINE_XARRAY(sr_forward_xa);
Signature · 1 line
2.34.1
author maintainer
#54 lore raw

PATCH v2 19/26 KVM: arm64: nv: Add trap forwarding for HFGxTR_EL2

In reply to Marc Zyngier

Fine Grained Traps are fun. Not.

Implement the trap forwarding for traps describer by HFGxTR_EL2, reusing the Coarse Grained Traps infrastructure previously implemented.

Each sysreg/instruction inserted in the xarray gets a FGT group (vaguely equivalent to a register number), a bit number in that register, and a polarity.

It is then pretty easy to check the FGT state at handling time, just like we do for the coarse version (it is just faster).

---
 arch/arm64/kvm/emulate-nested.c | 133 +++++++++++++++++++++++++++++++-
 1 file changed, 132 insertions(+), 1 deletion(-)
arch/arm64/kvm/emulate-nested.c +132 -1
index 74b67895c791..5f4cf824eadc 100644
@@ -922,6 +922,88 @@ static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
 
 static DEFINE_XARRAY(sr_forward_xa);
 
+enum fgt_group_id {
+	__NO_FGT_GROUP__,
+	HFGxTR_GROUP,
+};
+
+#define SR_FGT(sr, g, b, p)					\
+	{							\
+		.encoding	= sr,				\
+		.end		= sr,				\
+		.tc		= {				\
+			.fgt = g ## _GROUP,			\
+			.bit = g ## _EL2_ ## b ## _SHIFT,	\
+			.pol = p,				\
+		},						\
+	}
+
+static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = {
+	/* HFGTR_EL2, HFGWTR_EL2 */
+	SR_FGT(SYS_TPIDR2_EL0,		HFGxTR, nTPIDR2_EL0, 0),
+	SR_FGT(SYS_SMPRI_EL1,		HFGxTR, nSMPRI_EL1, 0),
+	SR_FGT(SYS_ACCDATA_EL1,		HFGxTR, nACCDATA_EL1, 0),
+	SR_FGT(SYS_ERXADDR_EL1,		HFGxTR, ERXADDR_EL1, 1),
+	SR_FGT(SYS_ERXPFGCDN_EL1,	HFGxTR, ERXPFGCDN_EL1, 1),
+	SR_FGT(SYS_ERXPFGCTL_EL1,	HFGxTR, ERXPFGCTL_EL1, 1),
+	SR_FGT(SYS_ERXPFGF_EL1,		HFGxTR, ERXPFGF_EL1, 1),
+	SR_FGT(SYS_ERXMISC0_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXMISC1_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXMISC2_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXMISC3_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXSTATUS_EL1,	HFGxTR, ERXSTATUS_EL1, 1),
+	SR_FGT(SYS_ERXCTLR_EL1,		HFGxTR, ERXCTLR_EL1, 1),
+	SR_FGT(SYS_ERXFR_EL1,		HFGxTR, ERXFR_EL1, 1),
+	SR_FGT(SYS_ERRSELR_EL1,		HFGxTR, ERRSELR_EL1, 1),
+	SR_FGT(SYS_ERRIDR_EL1,		HFGxTR, ERRIDR_EL1, 1),
+	SR_FGT(SYS_ICC_IGRPEN0_EL1,	HFGxTR, ICC_IGRPENn_EL1, 1),
+	SR_FGT(SYS_ICC_IGRPEN1_EL1,	HFGxTR, ICC_IGRPENn_EL1, 1),
+	SR_FGT(SYS_VBAR_EL1,		HFGxTR, VBAR_EL1, 1),
+	SR_FGT(SYS_TTBR1_EL1,		HFGxTR, TTBR1_EL1, 1),
+	SR_FGT(SYS_TTBR0_EL1,		HFGxTR, TTBR0_EL1, 1),
+	SR_FGT(SYS_TPIDR_EL0,		HFGxTR, TPIDR_EL0, 1),
+	SR_FGT(SYS_TPIDRRO_EL0,		HFGxTR, TPIDRRO_EL0, 1),
+	SR_FGT(SYS_TPIDR_EL1,		HFGxTR, TPIDR_EL1, 1),
+	SR_FGT(SYS_TCR_EL1,		HFGxTR, TCR_EL1, 1),
+	SR_FGT(SYS_SCXTNUM_EL0,		HFGxTR, SCXTNUM_EL0, 1),
+	SR_FGT(SYS_SCXTNUM_EL1, 	HFGxTR, SCXTNUM_EL1, 1),
+	SR_FGT(SYS_SCTLR_EL1, 		HFGxTR, SCTLR_EL1, 1),
+	SR_FGT(SYS_REVIDR_EL1, 		HFGxTR, REVIDR_EL1, 1),
+	SR_FGT(SYS_PAR_EL1, 		HFGxTR, PAR_EL1, 1),
+	SR_FGT(SYS_MPIDR_EL1, 		HFGxTR, MPIDR_EL1, 1),
+	SR_FGT(SYS_MIDR_EL1, 		HFGxTR, MIDR_EL1, 1),
+	SR_FGT(SYS_MAIR_EL1, 		HFGxTR, MAIR_EL1, 1),
+	SR_FGT(SYS_LORSA_EL1, 		HFGxTR, LORSA_EL1, 1),
+	SR_FGT(SYS_LORN_EL1, 		HFGxTR, LORN_EL1, 1),
+	SR_FGT(SYS_LORID_EL1, 		HFGxTR, LORID_EL1, 1),
+	SR_FGT(SYS_LOREA_EL1, 		HFGxTR, LOREA_EL1, 1),
+	SR_FGT(SYS_LORC_EL1, 		HFGxTR, LORC_EL1, 1),
+	SR_FGT(SYS_ISR_EL1, 		HFGxTR, ISR_EL1, 1),
+	SR_FGT(SYS_FAR_EL1, 		HFGxTR, FAR_EL1, 1),
+	SR_FGT(SYS_ESR_EL1, 		HFGxTR, ESR_EL1, 1),
+	SR_FGT(SYS_DCZID_EL0, 		HFGxTR, DCZID_EL0, 1),
+	SR_FGT(SYS_CTR_EL0, 		HFGxTR, CTR_EL0, 1),
+	SR_FGT(SYS_CSSELR_EL1, 		HFGxTR, CSSELR_EL1, 1),
+	SR_FGT(SYS_CPACR_EL1, 		HFGxTR, CPACR_EL1, 1),
+	SR_FGT(SYS_CONTEXTIDR_EL1, 	HFGxTR, CONTEXTIDR_EL1, 1),
+	SR_FGT(SYS_CLIDR_EL1, 		HFGxTR, CLIDR_EL1, 1),
+	SR_FGT(SYS_CCSIDR_EL1, 		HFGxTR, CCSIDR_EL1, 1),
+	SR_FGT(SYS_APIBKEYLO_EL1, 	HFGxTR, APIBKey, 1),
+	SR_FGT(SYS_APIBKEYHI_EL1, 	HFGxTR, APIBKey, 1),
+	SR_FGT(SYS_APIAKEYLO_EL1, 	HFGxTR, APIAKey, 1),
+	SR_FGT(SYS_APIAKEYHI_EL1, 	HFGxTR, APIAKey, 1),
+	SR_FGT(SYS_APGAKEYLO_EL1, 	HFGxTR, APGAKey, 1),
+	SR_FGT(SYS_APGAKEYHI_EL1, 	HFGxTR, APGAKey, 1),
+	SR_FGT(SYS_APDBKEYLO_EL1, 	HFGxTR, APDBKey, 1),
+	SR_FGT(SYS_APDBKEYHI_EL1, 	HFGxTR, APDBKey, 1),
+	SR_FGT(SYS_APDAKEYLO_EL1, 	HFGxTR, APDAKey, 1),
+	SR_FGT(SYS_APDAKEYHI_EL1, 	HFGxTR, APDAKey, 1),
+	SR_FGT(SYS_AMAIR_EL1, 		HFGxTR, AMAIR_EL1, 1),
+	SR_FGT(SYS_AIDR_EL1, 		HFGxTR, AIDR_EL1, 1),
+	SR_FGT(SYS_AFSR1_EL1, 		HFGxTR, AFSR1_EL1, 1),
+	SR_FGT(SYS_AFSR0_EL1, 		HFGxTR, AFSR0_EL1, 1),
+};
+
 static union trap_config get_trap_config(u32 sysreg)
 {
 	return (union trap_config) {
@@ -943,6 +1025,27 @@ void __init populate_nv_trap_config(void)
 	kvm_info("nv: %ld coarse grained trap handlers\n",
 		 ARRAY_SIZE(encoding_to_cgt));
 
+	for (int i = 0; i < ARRAY_SIZE(encoding_to_fgt); i++) {
+		const struct encoding_to_trap_config *fgt = &encoding_to_fgt[i];
+		union trap_config tc;
+
+		tc = get_trap_config(fgt->encoding);
+
+		WARN(tc.fgt,
+		     "Duplicate FGT for sys_reg(%d, %d, %d, %d, %d)\n",
+		     sys_reg_Op0(fgt->encoding),
+		     sys_reg_Op1(fgt->encoding),
+		     sys_reg_CRn(fgt->encoding),
+		     sys_reg_CRm(fgt->encoding),
+		     sys_reg_Op2(fgt->encoding));
+
+		tc.val |= fgt->tc.val;
+		xa_store(&sr_forward_xa, fgt->encoding,
+			 xa_mk_value(tc.val), GFP_KERNEL);
+	}
+
+	kvm_info("nv: %ld fine grained trap handlers\n",
+		 ARRAY_SIZE(encoding_to_fgt));
 }
 
 static enum trap_behaviour get_behaviour(struct kvm_vcpu *vcpu,
@@ -992,13 +1095,26 @@ static enum trap_behaviour compute_trap_behaviour(struct kvm_vcpu *vcpu,
 	return __do_compute_trap_behaviour(vcpu, tc.cgt, b);
 }
 
+static bool check_fgt_bit(u64 val, const union trap_config tc)
+{
+	return ((val >> tc.bit) & 1) == tc.pol;
+}
+
+#define sanitised_sys_reg(vcpu, reg)			\
+	({						\
+		u64 __val;				\
+		__val = __vcpu_sys_reg(vcpu, reg);	\
+		__val &= ~__ ## reg ## _RES0;		\
+		(__val);				\
+	})
+
 bool __check_nv_sr_forward(struct kvm_vcpu *vcpu)
 {
 	union trap_config tc;
 	enum trap_behaviour b;
 	bool is_read;
 	u32 sysreg;
-	u64 esr;
+	u64 esr, val;
 
 	if (!vcpu_has_nv(vcpu) || is_hyp_ctxt(vcpu))
 		return false;
@@ -1009,6 +1125,21 @@ bool __check_nv_sr_forward(struct kvm_vcpu *vcpu)
 
 	tc = get_trap_config(sysreg);
 
+	switch ((enum fgt_group_id)tc.fgt) {
+	case __NO_FGT_GROUP__:
+		break;
+
+	case HFGxTR_GROUP:
+		if (is_read)
+			val = sanitised_sys_reg(vcpu, HFGRTR_EL2);
+		else
+			val = sanitised_sys_reg(vcpu, HFGWTR_EL2);
+		break;
+	}
+
+	if (tc.fgt != __NO_FGT_GROUP__ && check_fgt_bit(val, tc))
+		goto inject;
+
 	b = compute_trap_behaviour(vcpu, tc);
 
 	if (((b & BEHAVE_FORWARD_READ) && is_read) ||
Signature · 1 line
2.34.1
#55 lore raw

Re: [PATCH v2 19/26] KVM: arm64: nv: Add trap forwarding for HFGxTR_EL2

Hey Marc,

On Fri, Jul 28, 2023 at 09:29:45AM +0100, Marc Zyngier wrote:

[...]

arch/arm64/kvm/emulate-nested.c void __init populate_nv_trap_config(void) · line 1026

Quoted 3 lines
arch/arm64/kvm/emulate-nested.c +0 -0
@@ -943,6 +1025,27 @@ void __init populate_nv_trap_config(void)
 	kvm_info("nv: %ld coarse grained trap handlers\n",
 		 ARRAY_SIZE(encoding_to_cgt));

It might make sense to skip insertion of the FGT trap controls if the system doesn't have FGT in the first place.

arch/arm64/kvm/emulate-nested.c void __init populate_nv_trap_config(void) · line 1040

Quoted 13 lines
arch/arm64/kvm/emulate-nested.c +13 -0
+	for (int i = 0; i < ARRAY_SIZE(encoding_to_fgt); i++) {
+		const struct encoding_to_trap_config *fgt = &encoding_to_fgt[i];
+		union trap_config tc;
+
+		tc = get_trap_config(fgt->encoding);
+
+		WARN(tc.fgt,
+		     "Duplicate FGT for sys_reg(%d, %d, %d, %d, %d)\n",
+		     sys_reg_Op0(fgt->encoding),
+		     sys_reg_Op1(fgt->encoding),
+		     sys_reg_CRn(fgt->encoding),
+		     sys_reg_CRm(fgt->encoding),
+		     sys_reg_Op2(fgt->encoding));

Same comment here, we should just bail.

Quoted 8 lines
arch/arm64/kvm/emulate-nested.c +7 -0
+		tc.val |= fgt->tc.val;
+		xa_store(&sr_forward_xa, fgt->encoding,
+			 xa_mk_value(tc.val), GFP_KERNEL);
+	}
+
+	kvm_info("nv: %ld fine grained trap handlers\n",
+		 ARRAY_SIZE(encoding_to_fgt));
 }
Signature · 2 lines
Thanks,
Oliver
author maintainer
#56 lore raw

Re: [PATCH v2 19/26] KVM: arm64: nv: Add trap forwarding for HFGxTR_EL2

Quoted 5 lines · Oliver Upton

Hey Marc,

On Fri, Jul 28, 2023 at 09:29:45AM +0100, Marc Zyngier wrote:

Quoted 8 lines

[...]

arch/arm64/kvm/emulate-nested.c void __init populate_nv_trap_config(void) · line 1026

Quoted 3 lines
arch/arm64/kvm/emulate-nested.c +0 -0
@@ -943,6 +1025,27 @@ void __init populate_nv_trap_config(void)
 	kvm_info("nv: %ld coarse grained trap handlers\n",
 		 ARRAY_SIZE(encoding_to_cgt));

It might make sense to skip insertion of the FGT trap controls if the system doesn't have FGT in the first place.

Yup, good point.

Quoted 16 lines
Quoted 4 lines
arch/arm64/kvm/emulate-nested.c +3 -0
+	for (int i = 0; i < ARRAY_SIZE(encoding_to_fgt); i++) {
+		const struct encoding_to_trap_config *fgt = &encoding_to_fgt[i];
+		union trap_config tc;

arch/arm64/kvm/emulate-nested.c void __init populate_nv_trap_config(void) · line 1040

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+
+		tc = get_trap_config(fgt->encoding);
+
+		WARN(tc.fgt,
+		     "Duplicate FGT for sys_reg(%d, %d, %d, %d, %d)\n",
+		     sys_reg_Op0(fgt->encoding),
+		     sys_reg_Op1(fgt->encoding),
+		     sys_reg_CRn(fgt->encoding),
+		     sys_reg_CRm(fgt->encoding),
+		     sys_reg_Op2(fgt->encoding));

Same comment here, we should just bail.

Yup. I'll also turn it into a kvm_err() instead, as WARN() is too noisy (nobody needs the same stack trace 200 times).

Thanks,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
Reviewed-by on v2 19/26
#57 lore raw

Re: [PATCH v2 19/26] KVM: arm64: nv: Add trap forwarding for HFGxTR_EL2

In reply to Marc Zyngier

Hi Marc,

Quoted 3 lines · Marc Zyngier

Fine Grained Traps are fun. Not.

Implement the trap forwarding for traps describer by HFGxTR_EL2,

described

Quoted 30 lines

reusing the Coarse Grained Traps infrastructure previously implemented.

Each sysreg/instruction inserted in the xarray gets a FGT group (vaguely equivalent to a register number), a bit number in that register, and a polarity.

It is then pretty easy to check the FGT state at handling time, just like we do for the coarse version (it is just faster).

---
 arch/arm64/kvm/emulate-nested.c | 133 +++++++++++++++++++++++++++++++-
 1 file changed, 132 insertions(+), 1 deletion(-)
arch/arm64/kvm/emulate-nested.c +8 -0
index 74b67895c791..5f4cf824eadc 100644
@@ -922,6 +922,88 @@ static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
 
 static DEFINE_XARRAY(sr_forward_xa);
 
+enum fgt_group_id {
+	__NO_FGT_GROUP__,
+	HFGxTR_GROUP,
+};
+
+#define SR_FGT(sr, g, b, p)					\
+	{							\
+		.encoding	= sr,				\

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = { · line 942

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+		.end		= sr,				\
+		.tc		= {				\
+			.fgt = g ## _GROUP,			\
+			.bit = g ## _EL2_ ## b ## _SHIFT,	\
+			.pol = p,				\
+		},						\
+	}
+
+static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = {
+	/* HFGTR_EL2, HFGWTR_EL2 */

HFGRTR_EL2

Quoted 135 lines
arch/arm64/kvm/emulate-nested.c +107 -1
+	SR_FGT(SYS_TPIDR2_EL0,		HFGxTR, nTPIDR2_EL0, 0),
+	SR_FGT(SYS_SMPRI_EL1,		HFGxTR, nSMPRI_EL1, 0),
+	SR_FGT(SYS_ACCDATA_EL1,		HFGxTR, nACCDATA_EL1, 0),
+	SR_FGT(SYS_ERXADDR_EL1,		HFGxTR, ERXADDR_EL1, 1),
+	SR_FGT(SYS_ERXPFGCDN_EL1,	HFGxTR, ERXPFGCDN_EL1, 1),
+	SR_FGT(SYS_ERXPFGCTL_EL1,	HFGxTR, ERXPFGCTL_EL1, 1),
+	SR_FGT(SYS_ERXPFGF_EL1,		HFGxTR, ERXPFGF_EL1, 1),
+	SR_FGT(SYS_ERXMISC0_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXMISC1_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXMISC2_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXMISC3_EL1,	HFGxTR, ERXMISCn_EL1, 1),
+	SR_FGT(SYS_ERXSTATUS_EL1,	HFGxTR, ERXSTATUS_EL1, 1),
+	SR_FGT(SYS_ERXCTLR_EL1,		HFGxTR, ERXCTLR_EL1, 1),
+	SR_FGT(SYS_ERXFR_EL1,		HFGxTR, ERXFR_EL1, 1),
+	SR_FGT(SYS_ERRSELR_EL1,		HFGxTR, ERRSELR_EL1, 1),
+	SR_FGT(SYS_ERRIDR_EL1,		HFGxTR, ERRIDR_EL1, 1),
+	SR_FGT(SYS_ICC_IGRPEN0_EL1,	HFGxTR, ICC_IGRPENn_EL1, 1),
+	SR_FGT(SYS_ICC_IGRPEN1_EL1,	HFGxTR, ICC_IGRPENn_EL1, 1),
+	SR_FGT(SYS_VBAR_EL1,		HFGxTR, VBAR_EL1, 1),
+	SR_FGT(SYS_TTBR1_EL1,		HFGxTR, TTBR1_EL1, 1),
+	SR_FGT(SYS_TTBR0_EL1,		HFGxTR, TTBR0_EL1, 1),
+	SR_FGT(SYS_TPIDR_EL0,		HFGxTR, TPIDR_EL0, 1),
+	SR_FGT(SYS_TPIDRRO_EL0,		HFGxTR, TPIDRRO_EL0, 1),
+	SR_FGT(SYS_TPIDR_EL1,		HFGxTR, TPIDR_EL1, 1),
+	SR_FGT(SYS_TCR_EL1,		HFGxTR, TCR_EL1, 1),
+	SR_FGT(SYS_SCXTNUM_EL0,		HFGxTR, SCXTNUM_EL0, 1),
+	SR_FGT(SYS_SCXTNUM_EL1, 	HFGxTR, SCXTNUM_EL1, 1),
+	SR_FGT(SYS_SCTLR_EL1, 		HFGxTR, SCTLR_EL1, 1),
+	SR_FGT(SYS_REVIDR_EL1, 		HFGxTR, REVIDR_EL1, 1),
+	SR_FGT(SYS_PAR_EL1, 		HFGxTR, PAR_EL1, 1),
+	SR_FGT(SYS_MPIDR_EL1, 		HFGxTR, MPIDR_EL1, 1),
+	SR_FGT(SYS_MIDR_EL1, 		HFGxTR, MIDR_EL1, 1),
+	SR_FGT(SYS_MAIR_EL1, 		HFGxTR, MAIR_EL1, 1),
+	SR_FGT(SYS_LORSA_EL1, 		HFGxTR, LORSA_EL1, 1),
+	SR_FGT(SYS_LORN_EL1, 		HFGxTR, LORN_EL1, 1),
+	SR_FGT(SYS_LORID_EL1, 		HFGxTR, LORID_EL1, 1),
+	SR_FGT(SYS_LOREA_EL1, 		HFGxTR, LOREA_EL1, 1),
+	SR_FGT(SYS_LORC_EL1, 		HFGxTR, LORC_EL1, 1),
+	SR_FGT(SYS_ISR_EL1, 		HFGxTR, ISR_EL1, 1),
+	SR_FGT(SYS_FAR_EL1, 		HFGxTR, FAR_EL1, 1),
+	SR_FGT(SYS_ESR_EL1, 		HFGxTR, ESR_EL1, 1),
+	SR_FGT(SYS_DCZID_EL0, 		HFGxTR, DCZID_EL0, 1),
+	SR_FGT(SYS_CTR_EL0, 		HFGxTR, CTR_EL0, 1),
+	SR_FGT(SYS_CSSELR_EL1, 		HFGxTR, CSSELR_EL1, 1),
+	SR_FGT(SYS_CPACR_EL1, 		HFGxTR, CPACR_EL1, 1),
+	SR_FGT(SYS_CONTEXTIDR_EL1, 	HFGxTR, CONTEXTIDR_EL1, 1),
+	SR_FGT(SYS_CLIDR_EL1, 		HFGxTR, CLIDR_EL1, 1),
+	SR_FGT(SYS_CCSIDR_EL1, 		HFGxTR, CCSIDR_EL1, 1),
+	SR_FGT(SYS_APIBKEYLO_EL1, 	HFGxTR, APIBKey, 1),
+	SR_FGT(SYS_APIBKEYHI_EL1, 	HFGxTR, APIBKey, 1),
+	SR_FGT(SYS_APIAKEYLO_EL1, 	HFGxTR, APIAKey, 1),
+	SR_FGT(SYS_APIAKEYHI_EL1, 	HFGxTR, APIAKey, 1),
+	SR_FGT(SYS_APGAKEYLO_EL1, 	HFGxTR, APGAKey, 1),
+	SR_FGT(SYS_APGAKEYHI_EL1, 	HFGxTR, APGAKey, 1),
+	SR_FGT(SYS_APDBKEYLO_EL1, 	HFGxTR, APDBKey, 1),
+	SR_FGT(SYS_APDBKEYHI_EL1, 	HFGxTR, APDBKey, 1),
+	SR_FGT(SYS_APDAKEYLO_EL1, 	HFGxTR, APDAKey, 1),
+	SR_FGT(SYS_APDAKEYHI_EL1, 	HFGxTR, APDAKey, 1),
+	SR_FGT(SYS_AMAIR_EL1, 		HFGxTR, AMAIR_EL1, 1),
+	SR_FGT(SYS_AIDR_EL1, 		HFGxTR, AIDR_EL1, 1),
+	SR_FGT(SYS_AFSR1_EL1, 		HFGxTR, AFSR1_EL1, 1),
+	SR_FGT(SYS_AFSR0_EL1, 		HFGxTR, AFSR0_EL1, 1),
+};
+
 static union trap_config get_trap_config(u32 sysreg)
 {
 	return (union trap_config) {
@@ -943,6 +1025,27 @@ void __init populate_nv_trap_config(void)
 	kvm_info("nv: %ld coarse grained trap handlers\n",
 		 ARRAY_SIZE(encoding_to_cgt));
 
+	for (int i = 0; i < ARRAY_SIZE(encoding_to_fgt); i++) {
+		const struct encoding_to_trap_config *fgt = &encoding_to_fgt[i];
+		union trap_config tc;
+
+		tc = get_trap_config(fgt->encoding);
+
+		WARN(tc.fgt,
+		     "Duplicate FGT for sys_reg(%d, %d, %d, %d, %d)\n",
+		     sys_reg_Op0(fgt->encoding),
+		     sys_reg_Op1(fgt->encoding),
+		     sys_reg_CRn(fgt->encoding),
+		     sys_reg_CRm(fgt->encoding),
+		     sys_reg_Op2(fgt->encoding));
+
+		tc.val |= fgt->tc.val;
+		xa_store(&sr_forward_xa, fgt->encoding,
+			 xa_mk_value(tc.val), GFP_KERNEL);
+	}
+
+	kvm_info("nv: %ld fine grained trap handlers\n",
+		 ARRAY_SIZE(encoding_to_fgt));
 }
 
 static enum trap_behaviour get_behaviour(struct kvm_vcpu *vcpu,
@@ -992,13 +1095,26 @@ static enum trap_behaviour compute_trap_behaviour(struct kvm_vcpu *vcpu,
 	return __do_compute_trap_behaviour(vcpu, tc.cgt, b);
 }
 
+static bool check_fgt_bit(u64 val, const union trap_config tc)
+{
+	return ((val >> tc.bit) & 1) == tc.pol;
+}
+
+#define sanitised_sys_reg(vcpu, reg)			\
+	({						\
+		u64 __val;				\
+		__val = __vcpu_sys_reg(vcpu, reg);	\
+		__val &= ~__ ## reg ## _RES0;		\
+		(__val);				\
+	})
+
 bool __check_nv_sr_forward(struct kvm_vcpu *vcpu)
 {
 	union trap_config tc;
 	enum trap_behaviour b;
 	bool is_read;
 	u32 sysreg;
-	u64 esr;
+	u64 esr, val;
 
 	if (!vcpu_has_nv(vcpu) || is_hyp_ctxt(vcpu))
 		return false;
@@ -1009,6 +1125,21 @@ bool __check_nv_sr_forward(struct kvm_vcpu *vcpu)
 
 	tc = get_trap_config(sysreg);
 
+	switch ((enum fgt_group_id)tc.fgt) {
+	case __NO_FGT_GROUP__:
+		break;
+
+	case HFGxTR_GROUP:
+		if (is_read)
+			val = sanitised_sys_reg(vcpu, HFGRTR_EL2);
+		else

arch/arm64/kvm/emulate-nested.c bool __check_nv_sr_forward(struct kvm_vcpu *vcpu) · line 1145

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +7 -0
+			val = sanitised_sys_reg(vcpu, HFGWTR_EL2);
+		break;
+	}
+
+	if (tc.fgt != __NO_FGT_GROUP__ && check_fgt_bit(val, tc))
+		goto inject;
+
 	b = compute_trap_behaviour(vcpu, tc);
 
 	if (((b & BEHAVE_FORWARD_READ) && is_read) ||

With Oliver's comments

Eric

author maintainer
#58 lore raw

PATCH v2 20/26 KVM: arm64: nv: Add trap forwarding for HFGITR_EL2

In reply to Marc Zyngier

Similarly, implement the trap forwarding for instructions affected by HFGITR_EL2.

Note that the TLBI*nXS instructions should be affected by HCRX_EL2, which will be dealt with down the line.

---
 arch/arm64/include/asm/kvm_arm.h |   4 ++
 arch/arm64/kvm/emulate-nested.c  | 109 +++++++++++++++++++++++++++++++
 2 files changed, 113 insertions(+)
arch/arm64/include/asm/kvm_arm.h +4 -0
index 85908aa18908..809bc86acefd 100644
@@ -354,6 +354,10 @@
 #define __HFGWTR_EL2_MASK	GENMASK(49, 0)
 #define __HFGWTR_EL2_nMASK	(GENMASK(55, 54) | BIT(50))
 
+#define __HFGITR_EL2_RES0	GENMASK(63, 57)
+#define __HFGITR_EL2_MASK	GENMASK(54, 0)
+#define __HFGITR_EL2_nMASK	GENMASK(56, 55)
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/kvm/emulate-nested.c +109 -0

127 lines not on this page · raw message

Signature · 1 line
2.34.1
Reviewed-by on v2 20/26
#59 lore raw

Re: [PATCH v2 20/26] KVM: arm64: nv: Add trap forwarding for HFGITR_EL2

Hi Marc,

Quoted 5 lines · Marc Zyngier

Similarly, implement the trap forwarding for instructions affected by HFGITR_EL2.

Note that the TLBI*nXS instructions should be affected by HCRX_EL2, which will be dealt with down the line.

I think you should also add a comment about the fact SVC_EL1/0 and ERET are not dealt with in this patch.

Quoted 146 lines
---
 arch/arm64/include/asm/kvm_arm.h |   4 ++
 arch/arm64/kvm/emulate-nested.c  | 109 +++++++++++++++++++++++++++++++
 2 files changed, 113 insertions(+)
arch/arm64/include/asm/kvm_arm.h +4 -0
index 85908aa18908..809bc86acefd 100644
@@ -354,6 +354,10 @@
 #define __HFGWTR_EL2_MASK	GENMASK(49, 0)
 #define __HFGWTR_EL2_nMASK	(GENMASK(55, 54) | BIT(50))
 
+#define __HFGITR_EL2_RES0	GENMASK(63, 57)
+#define __HFGITR_EL2_MASK	GENMASK(54, 0)
+#define __HFGITR_EL2_nMASK	GENMASK(56, 55)
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/kvm/emulate-nested.c +105 -0
index 5f4cf824eadc..72619d845cc8 100644
@@ -925,6 +925,7 @@ static DEFINE_XARRAY(sr_forward_xa);
 enum fgt_group_id {
 	__NO_FGT_GROUP__,
 	HFGxTR_GROUP,
+	HFGITR_GROUP,
 };
 
 #define SR_FGT(sr, g, b, p)					\
@@ -1002,6 +1003,110 @@ static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = {
 	SR_FGT(SYS_AIDR_EL1, 		HFGxTR, AIDR_EL1, 1),
 	SR_FGT(SYS_AFSR1_EL1, 		HFGxTR, AFSR1_EL1, 1),
 	SR_FGT(SYS_AFSR0_EL1, 		HFGxTR, AFSR0_EL1, 1),
+	/* HFGITR_EL2 */
+	SR_FGT(OP_BRB_IALL, 		HFGITR, nBRBIALL, 0),
+	SR_FGT(OP_BRB_INJ, 		HFGITR, nBRBINJ, 0),
+	SR_FGT(SYS_DC_CVAC, 		HFGITR, DCCVAC, 1),
+	SR_FGT(SYS_DC_CGVAC, 		HFGITR, DCCVAC, 1),
+	SR_FGT(SYS_DC_CGDVAC, 		HFGITR, DCCVAC, 1),
+	SR_FGT(OP_CPP_RCTX, 		HFGITR, CPPRCTX, 1),
+	SR_FGT(OP_DVP_RCTX, 		HFGITR, DVPRCTX, 1),
+	SR_FGT(OP_CFP_RCTX, 		HFGITR, CFPRCTX, 1),
+	SR_FGT(OP_TLBI_VAALE1, 		HFGITR, TLBIVAALE1, 1),y
+	SR_FGT(OP_TLBI_VALE1, 		HFGITR, TLBIVALE1, 1),y
+	SR_FGT(OP_TLBI_VAAE1, 		HFGITR, TLBIVAAE1, 1),y
+	SR_FGT(OP_TLBI_ASIDE1, 		HFGITR, TLBIASIDE1, 1),y
+	SR_FGT(OP_TLBI_VAE1, 		HFGITR, TLBIVAE1, 1),y
+	SR_FGT(OP_TLBI_VMALLE1, 	HFGITR, TLBIVMALLE1, 1),y
+	SR_FGT(OP_TLBI_RVAALE1, 	HFGITR, TLBIRVAALE1, 1),y
+	SR_FGT(OP_TLBI_RVALE1, 		HFGITR, TLBIRVALE1, 1),y
+	SR_FGT(OP_TLBI_RVAAE1, 		HFGITR, TLBIRVAAE1, 1),y
+	SR_FGT(OP_TLBI_RVAE1, 		HFGITR, TLBIRVAE1, 1),y
+	SR_FGT(OP_TLBI_RVAALE1IS, 	HFGITR, TLBIRVAALE1IS, 1),y
+	SR_FGT(OP_TLBI_RVALE1IS, 	HFGITR, TLBIRVALE1IS, 1),y
+	SR_FGT(OP_TLBI_RVAAE1IS, 	HFGITR, TLBIRVAAE1IS, 1),y
+	SR_FGT(OP_TLBI_RVAE1IS, 	HFGITR, TLBIRVAE1IS, 1),y
+	SR_FGT(OP_TLBI_VAALE1IS, 	HFGITR, TLBIVAALE1IS, 1),y
+	SR_FGT(OP_TLBI_VALE1IS, 	HFGITR, TLBIVALE1IS, 1),y
+	SR_FGT(OP_TLBI_VAAE1IS, 	HFGITR, TLBIVAAE1IS, 1),y
+	SR_FGT(OP_TLBI_ASIDE1IS, 	HFGITR, TLBIASIDE1IS, 1),y
+	SR_FGT(OP_TLBI_VAE1IS, 		HFGITR, TLBIVAE1IS, 1),y
+	SR_FGT(OP_TLBI_VMALLE1IS, 	HFGITR, TLBIVMALLE1IS, 1),y
+	SR_FGT(OP_TLBI_RVAALE1OS, 	HFGITR, TLBIRVAALE1OS, 1),y
+	SR_FGT(OP_TLBI_RVALE1OS, 	HFGITR, TLBIRVALE1OS, 1),y
+	SR_FGT(OP_TLBI_RVAAE1OS, 	HFGITR, TLBIRVAAE1OS, 1),y
+	SR_FGT(OP_TLBI_RVAE1OS, 	HFGITR, TLBIRVAE1OS, 1),y
+	SR_FGT(OP_TLBI_VAALE1OS, 	HFGITR, TLBIVAALE1OS, 1),y
+	SR_FGT(OP_TLBI_VALE1OS, 	HFGITR, TLBIVALE1OS, 1),y
+	SR_FGT(OP_TLBI_VAAE1OS, 	HFGITR, TLBIVAAE1OS, 1),y
+	SR_FGT(OP_TLBI_ASIDE1OS, 	HFGITR, TLBIASIDE1OS, 1),y
+	SR_FGT(OP_TLBI_VAE1OS, 		HFGITR, TLBIVAE1OS, 1),y
+	SR_FGT(OP_TLBI_VMALLE1OS, 	HFGITR, TLBIVMALLE1OS, 1),y
+	/* FIXME: nXS variants must be checked against HCRX_EL2.FGTnXS */
+	SR_FGT(OP_TLBI_VAALE1NXS, 	HFGITR, TLBIVAALE1, 1),y
+	SR_FGT(OP_TLBI_VALE1NXS, 	HFGITR, TLBIVALE1, 1),y
+	SR_FGT(OP_TLBI_VAAE1NXS, 	HFGITR, TLBIVAAE1, 1),y
+	SR_FGT(OP_TLBI_ASIDE1NXS, 	HFGITR, TLBIASIDE1, 1),y
+	SR_FGT(OP_TLBI_VAE1NXS, 	HFGITR, TLBIVAE1, 1),y
+	SR_FGT(OP_TLBI_VMALLE1NXS, 	HFGITR, TLBIVMALLE1, 1),y
+	SR_FGT(OP_TLBI_RVAALE1NXS, 	HFGITR, TLBIRVAALE1, 1),y
+	SR_FGT(OP_TLBI_RVALE1NXS, 	HFGITR, TLBIRVALE1, 1),y
+	SR_FGT(OP_TLBI_RVAAE1NXS, 	HFGITR, TLBIRVAAE1, 1),y
+	SR_FGT(OP_TLBI_RVAE1NXS, 	HFGITR, TLBIRVAE1, 1),y
+	SR_FGT(OP_TLBI_RVAALE1ISNXS, 	HFGITR, TLBIRVAALE1IS, 1),y
+	SR_FGT(OP_TLBI_RVALE1ISNXS, 	HFGITR, TLBIRVALE1IS, 1),y
+	SR_FGT(OP_TLBI_RVAAE1ISNXS, 	HFGITR, TLBIRVAAE1IS, 1),y
+	SR_FGT(OP_TLBI_RVAE1ISNXS, 	HFGITR, TLBIRVAE1IS, 1),y
+	SR_FGT(OP_TLBI_VAALE1ISNXS, 	HFGITR, TLBIVAALE1IS, 1),y
+	SR_FGT(OP_TLBI_VALE1ISNXS, 	HFGITR, TLBIVALE1IS, 1),y
+	SR_FGT(OP_TLBI_VAAE1ISNXS, 	HFGITR, TLBIVAAE1IS, 1),
+	SR_FGT(OP_TLBI_ASIDE1ISNXS, 	HFGITR, TLBIASIDE1IS, 1),y
+	SR_FGT(OP_TLBI_VAE1ISNXS, 	HFGITR, TLBIVAE1IS, 1),y
+	SR_FGT(OP_TLBI_VMALLE1ISNXS, 	HFGITR, TLBIVMALLE1IS, 1),y
+	SR_FGT(OP_TLBI_RVAALE1OSNXS, 	HFGITR, TLBIRVAALE1OS, 1),y
+	SR_FGT(OP_TLBI_RVALE1OSNXS, 	HFGITR, TLBIRVALE1OS, 1),y
+	SR_FGT(OP_TLBI_RVAAE1OSNXS, 	HFGITR, TLBIRVAAE1OS, 1),y
+	SR_FGT(OP_TLBI_RVAE1OSNXS, 	HFGITR, TLBIRVAE1OS, 1),y
+	SR_FGT(OP_TLBI_VAALE1OSNXS, 	HFGITR, TLBIVAALE1OS, 1),y
+	SR_FGT(OP_TLBI_VALE1OSNXS, 	HFGITR, TLBIVALE1OS, 1),y
+	SR_FGT(OP_TLBI_VAAE1OSNXS, 	HFGITR, TLBIVAAE1OS, 1),y
+	SR_FGT(OP_TLBI_ASIDE1OSNXS, 	HFGITR, TLBIASIDE1OS, 1),y
+	SR_FGT(OP_TLBI_VAE1OSNXS, 	HFGITR, TLBIVAE1OS, 1),y
+	SR_FGT(OP_TLBI_VMALLE1OSNXS, 	HFGITR, TLBIVMALLE1OS, 1),y
+	SR_FGT(OP_AT_S1E1WP, 		HFGITR, ATS1E1WP, 1),
+	SR_FGT(OP_AT_S1E1RP, 		HFGITR, ATS1E1RP, 1),
+	SR_FGT(OP_AT_S1E0W, 		HFGITR, ATS1E0W, 1),
+	SR_FGT(OP_AT_S1E0R, 		HFGITR, ATS1E0R, 1),
+	SR_FGT(OP_AT_S1E1W, 		HFGITR, ATS1E1W, 1),
+	SR_FGT(OP_AT_S1E1R, 		HFGITR, ATS1E1R, 1),
+	SR_FGT(SYS_DC_ZVA, 		HFGITR, DCZVA, 1),
+	SR_FGT(SYS_DC_GVA, 		HFGITR, DCZVA, 1),
+	SR_FGT(SYS_DC_GZVA, 		HFGITR, DCZVA, 1),
+	SR_FGT(SYS_DC_CIVAC, 		HFGITR, DCCIVAC, 1),
+	SR_FGT(SYS_DC_CIGVAC, 		HFGITR, DCCIVAC, 1),
+	SR_FGT(SYS_DC_CIGDVAC, 		HFGITR, DCCIVAC, 1),
+	SR_FGT(SYS_DC_CVADP, 		HFGITR, DCCVADP, 1),
+	SR_FGT(SYS_DC_CGVADP, 		HFGITR, DCCVADP, 1),
+	SR_FGT(SYS_DC_CGDVADP, 		HFGITR, DCCVADP, 1),
+	SR_FGT(SYS_DC_CVAP, 		HFGITR, DCCVAP, 1),
+	SR_FGT(SYS_DC_CGVAP, 		HFGITR, DCCVAP, 1),
+	SR_FGT(SYS_DC_CGDVAP, 		HFGITR, DCCVAP, 1),
+	SR_FGT(SYS_DC_CVAU, 		HFGITR, DCCVAU, 1),
+	SR_FGT(SYS_DC_CISW, 		HFGITR, DCCISW, 1),
+	SR_FGT(SYS_DC_CIGSW, 		HFGITR, DCCISW, 1),
+	SR_FGT(SYS_DC_CIGDSW, 		HFGITR, DCCISW, 1),
+	SR_FGT(SYS_DC_CSW, 		HFGITR, DCCSW, 1),
+	SR_FGT(SYS_DC_CGSW, 		HFGITR, DCCSW, 1),
+	SR_FGT(SYS_DC_CGDSW, 		HFGITR, DCCSW, 1),
+	SR_FGT(SYS_DC_ISW, 		HFGITR, DCISW, 1),
+	SR_FGT(SYS_DC_IGSW, 		HFGITR, DCISW, 1),
+	SR_FGT(SYS_DC_IGDSW, 		HFGITR, DCISW, 1),
+	SR_FGT(SYS_DC_IVAC, 		HFGITR, DCIVAC, 1),
+	SR_FGT(SYS_DC_IGVAC, 		HFGITR, DCIVAC, 1),
+	SR_FGT(SYS_DC_IGDVAC, 		HFGITR, DCIVAC, 1),
+	SR_FGT(SYS_IC_IVAU, 		HFGITR, ICIVAU, 1),
+	SR_FGT(SYS_IC_IALLU, 		HFGITR, ICIALLU, 1),
+	SR_FGT(SYS_IC_IALLUIS, 		HFGITR, ICIALLUIS, 1),
 };
 
 static union trap_config get_trap_config(u32 sysreg)

arch/arm64/kvm/emulate-nested.c bool __check_nv_sr_forward(struct kvm_vcpu *vcpu) · line 1249

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +4 -0
@@ -1135,6 +1240,10 @@ bool __check_nv_sr_forward(struct kvm_vcpu *vcpu)
 		else
 			val = sanitised_sys_reg(vcpu, HFGWTR_EL2);
 		break;
+
+	case HFGITR_GROUP:
+		val = sanitised_sys_reg(vcpu, HFGITR_EL2);
+		break;
 	}
 
 	if (tc.fgt != __NO_FGT_GROUP__ && check_fgt_bit(val, tc))

Eric

author maintainer
#60 lore raw

PATCH v2 21/26 KVM: arm64: nv: Add trap forwarding for HDFGxTR_EL2

In reply to Marc Zyngier

... and finally, the Debug version of FGT, with its *enormous* list of trapped registers.

---
 arch/arm64/include/asm/kvm_arm.h |  11 +
 arch/arm64/kvm/emulate-nested.c  | 460 +++++++++++++++++++++++++++++++
 2 files changed, 471 insertions(+)
arch/arm64/include/asm/kvm_arm.h +11 -0
index 809bc86acefd..d229f238c3b6 100644
@@ -358,6 +358,17 @@
 #define __HFGITR_EL2_MASK	GENMASK(54, 0)
 #define __HFGITR_EL2_nMASK	GENMASK(56, 55)
 
+#define __HDFGRTR_EL2_RES0	(BIT(49) | BIT(42) | GENMASK(39, 38) |	\
+				 GENMASK(21, 20) | BIT(8))
+#define __HDFGRTR_EL2_MASK	~__HDFGRTR_EL2_nMASK
+#define __HDFGRTR_EL2_nMASK	GENMASK(62, 59)
+
+#define __HDFGWTR_EL2_RES0	(BIT(63) | GENMASK(59, 58) | BIT(51) | BIT(47) | \
+				 BIT(43) | GENMASK(40, 38) | BIT(34) | BIT(30) | \
+				 BIT(22) | BIT(9) | BIT(6))
+#define __HDFGWTR_EL2_MASK	~__HDFGWTR_EL2_nMASK
+#define __HDFGWTR_EL2_nMASK	GENMASK(62, 60)
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/kvm/emulate-nested.c +460 -0

478 lines not on this page · raw message

Signature · 1 line
2.34.1
#61 lore raw

Re: [PATCH v2 21/26] KVM: arm64: nv: Add trap forwarding for HDFGxTR_EL2

Hi Marc,

Quoted 5 lines · Marc Zyngier

... and finally, the Debug version of FGT, with its *enormous* list of trapped registers.

Hi Marc, I think you mixed up with the R-b's sent on [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

Quoted 165 lines
---
 arch/arm64/include/asm/kvm_arm.h |  11 +
 arch/arm64/kvm/emulate-nested.c  | 460 +++++++++++++++++++++++++++++++
 2 files changed, 471 insertions(+)
arch/arm64/include/asm/kvm_arm.h +11 -0
index 809bc86acefd..d229f238c3b6 100644
@@ -358,6 +358,17 @@
 #define __HFGITR_EL2_MASK	GENMASK(54, 0)
 #define __HFGITR_EL2_nMASK	GENMASK(56, 55)
 
+#define __HDFGRTR_EL2_RES0	(BIT(49) | BIT(42) | GENMASK(39, 38) |	\
+				 GENMASK(21, 20) | BIT(8))
+#define __HDFGRTR_EL2_MASK	~__HDFGRTR_EL2_nMASK
+#define __HDFGRTR_EL2_nMASK	GENMASK(62, 59)
+
+#define __HDFGWTR_EL2_RES0	(BIT(63) | GENMASK(59, 58) | BIT(51) | BIT(47) | \
+				 BIT(43) | GENMASK(40, 38) | BIT(34) | BIT(30) | \
+				 BIT(22) | BIT(9) | BIT(6))
+#define __HDFGWTR_EL2_MASK	~__HDFGWTR_EL2_nMASK
+#define __HDFGWTR_EL2_nMASK	GENMASK(62, 60)
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/kvm/emulate-nested.c +122 -0
index 72619d845cc8..e3bfb2440801 100644
@@ -925,6 +925,8 @@ static DEFINE_XARRAY(sr_forward_xa);
 enum fgt_group_id {
 	__NO_FGT_GROUP__,
 	HFGxTR_GROUP,
+	HDFGRTR_GROUP,
+	HDFGWTR_GROUP,
 	HFGITR_GROUP,
 };
 
@@ -1107,6 +1109,456 @@ static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = {
 	SR_FGT(SYS_IC_IVAU, 		HFGITR, ICIVAU, 1),
 	SR_FGT(SYS_IC_IALLU, 		HFGITR, ICIALLU, 1),
 	SR_FGT(SYS_IC_IALLUIS, 		HFGITR, ICIALLUIS, 1),
+	/* HDFGRTR_EL2 */
+	SR_FGT(SYS_PMBIDR_EL1, 		HDFGRTR, PMBIDR_EL1, 1),y
+	SR_FGT(SYS_PMSNEVFR_EL1, 	HDFGRTR, nPMSNEVFR_EL1, 0),y
+	SR_FGT(SYS_BRBINF_EL1(0), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(1), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(2), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(3), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(4), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(5), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(6), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(7), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(8), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(9), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(10), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(11), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(12), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(13), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(14), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(15), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(16), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(17), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(18), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(19), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(20), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(21), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(22), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(23), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(24), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(25), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(26), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(27), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(28), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(29), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(30), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINF_EL1(31), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBINFINJ_EL1, 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(0), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(1), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(2), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(3), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(4), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(5), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(6), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(7), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(8), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(9), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(10), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(11), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(12), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(13), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(14), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(15), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(16), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(17), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(18), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(19), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(20), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(21), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(22), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(23), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(24), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(25), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(26), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(27), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(28), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(29), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(30), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRC_EL1(31), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBSRCINJ_EL1, 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(0), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(1), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(2), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(3), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(4), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(5), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(6), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(7), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(8), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(9), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(10), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(11), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(12), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(13), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(14), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(15), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(16), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(17), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(18), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(19), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(20), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(21), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(22), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(23), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(24), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(25), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(26), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(27), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(28), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(29), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(30), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGT_EL1(31), 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTGTINJ_EL1, 	HDFGRTR, nBRBDATA, 0),
+	SR_FGT(SYS_BRBTS_EL1, 		HDFGRTR, nBRBDATA, 0),y
+	SR_FGT(SYS_BRBCR_EL1, 		HDFGRTR, nBRBCTL, 0),
+	SR_FGT(SYS_BRBFCR_EL1, 		HDFGRTR, nBRBCTL, 0),
+	SR_FGT(SYS_BRBIDR0_EL1, 	HDFGRTR, nBRBIDR, 0),
+	SR_FGT(SYS_PMCEID0_EL0, 	HDFGRTR, PMCEIDn_EL0, 1),y
+	SR_FGT(SYS_PMCEID1_EL0, 	HDFGRTR, PMCEIDn_EL0, 1),y
+	SR_FGT(SYS_PMUSERENR_EL0, 	HDFGRTR, PMUSERENR_EL0, 1),y
+	SR_FGT(SYS_TRBTRG_EL1, 		HDFGRTR, TRBTRG_EL1, 1),y
+	SR_FGT(SYS_TRBSR_EL1, 		HDFGRTR, TRBSR_EL1, 1),y
+	SR_FGT(SYS_TRBPTR_EL1, 		HDFGRTR, TRBPTR_EL1, 1),y
+	SR_FGT(SYS_TRBMAR_EL1, 		HDFGRTR, TRBMAR_EL1, 1),y
+	SR_FGT(SYS_TRBLIMITR_EL1, 	HDFGRTR, TRBLIMITR_EL1, 1),y
+	SR_FGT(SYS_TRBIDR_EL1, 		HDFGRTR, TRBIDR_EL1, 1),y
+	SR_FGT(SYS_TRBBASER_EL1, 	HDFGRTR, TRBBASER_EL1, 1),y
+	SR_FGT(SYS_TRCVICTLR, 		HDFGRTR, TRCVICTLR, 1),y
+	SR_FGT(SYS_TRCSTATR, 		HDFGRTR, TRCSTATR, 1),y
+	SR_FGT(SYS_TRCSSCSR(0), 	HDFGRTR, TRCSSCSRn, 1),y
+	SR_FGT(SYS_TRCSSCSR(1), 	HDFGRTR, TRCSSCSRn, 1),

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1241

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+	SR_FGT(SYS_TRCSSCSR(2), 	HDFGRTR, TRCSSCSRn, 1),
+	SR_FGT(SYS_TRCSSCSR(3), 	HDFGRTR, TRCSSCSRn, 1),
+	SR_FGT(SYS_TRCSSCSR(4), 	HDFGRTR, TRCSSCSRn, 1),
+	SR_FGT(SYS_TRCSSCSR(5), 	HDFGRTR, TRCSSCSRn, 1),
+	SR_FGT(SYS_TRCSSCSR(6), 	HDFGRTR, TRCSSCSRn, 1),
+	SR_FGT(SYS_TRCSSCSR(7), 	HDFGRTR, TRCSSCSRn, 1),y
+	SR_FGT(SYS_TRCSEQSTR, 		HDFGRTR, TRCSEQSTR, 1),y
+	SR_FGT(SYS_TRCPRGCTLR, 		HDFGRTR, TRCPRGCTLR, 1),y
+	SR_FGT(SYS_TRCOSLSR, 		HDFGRTR, TRCOSLSR, 1),y
+	SR_FGT(SYS_TRCIMSPEC0, 		HDFGRTR, TRCIMSPECn, 1),

why not SYS_TRCIMSPEC(0)?

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1250

Quoted 9 lines
arch/arm64/kvm/emulate-nested.c +9 -0
+	SR_FGT(SYS_TRCIMSPEC(1), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(2), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(3), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(4), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(5), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(6), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(7), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCDEVARCH, 		HDFGRTR, TRCID, 1),
+	SR_FGT(SYS_TRCDEVID, 		HDFGRTR, TRCID, 1),

what about all of the TRCIDR regs refered to in the spec?

Quoted 88 lines
arch/arm64/kvm/emulate-nested.c +88 -0
+	SR_FGT(SYS_TRCCNTVR(0), 	HDFGRTR, TRCCNTVRn, 1),
+	SR_FGT(SYS_TRCCNTVR(1), 	HDFGRTR, TRCCNTVRn, 1),
+	SR_FGT(SYS_TRCCNTVR(2), 	HDFGRTR, TRCCNTVRn, 1),
+	SR_FGT(SYS_TRCCNTVR(3), 	HDFGRTR, TRCCNTVRn, 1),y
+	SR_FGT(SYS_TRCCLAIMCLR, 	HDFGRTR, TRCCLAIM, 1),y
+	SR_FGT(SYS_TRCCLAIMSET, 	HDFGRTR, TRCCLAIM, 1),y
+	SR_FGT(SYS_TRCAUXCTLR, 		HDFGRTR, TRCAUXCTLR, 1),y
+	SR_FGT(SYS_TRCAUTHSTATUS, 	HDFGRTR, TRCAUTHSTATUS, 1),y
+	SR_FGT(SYS_TRCACATR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCACATR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(7), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(8), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(9), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(10), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(11), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(12), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(13), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(14), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACATR(15), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCACVR(0), 		HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCACVR(1), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(2), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(3), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(4), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(5), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(6), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(7), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(8), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(9), 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(10), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(11), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(12), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(13), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(14), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCACVR(15), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCBBCTLR, 		HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCCCTLR, 		HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCIDCCTLR0, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCIDCCTLR1, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCIDCVR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCIDCVR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCIDCVR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCIDCVR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCIDCVR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCIDCVR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCIDCVR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCIDCVR(7), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCNTCTLR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCNTCTLR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCNTCTLR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCNTCTLR(3), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCNTRLDVR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCNTRLDVR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCNTRLDVR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCCNTRLDVR(3), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCCONFIGR, 		HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCEVENTCTL0R, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCEVENTCTL1R, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCEXTINSELR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCEXTINSELR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCEXTINSELR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCEXTINSELR(3), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCRSCTLR(2), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCRSCTLR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(7), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(8), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(9), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(10), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(11), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(12), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(13), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(14), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(15), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(16), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(17), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(18), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(19), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(20), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(21), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(22), 	HDFGRTR, TRC, 1),

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1348

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+	SR_FGT(SYS_TRCRSCTLR(23), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(24), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(25), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(26), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(27), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(28), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(29), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(30), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(31), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCQCTLR, 		HDFGRTR, TRC, 1),

nit: maybe put this one before the

SYS_TRCRSCTLR(n) series to follow the spec order

Quoted 196 lines
arch/arm64/kvm/emulate-nested.c +196 -0
+	SR_FGT(SYS_TRCRSR, 		HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSEQEVR(0), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSEQEVR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSEQEVR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSEQRSTEVR, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCSSCCR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCSSCCR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSCCR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSCCR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSCCR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSCCR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSCCR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSCCR(7), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCSSPCICR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCSSPCICR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSPCICR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSPCICR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSPCICR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSPCICR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSPCICR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCSSPCICR(7), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCSTALLCTLR, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCSYNCPR, 		HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCTRACEIDR, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCTSCTLR, 		HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVIIECTLR, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVIPCSSCTLR, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVISSCTLR, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVMIDCCTLR0, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVMIDCCTLR1, 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVMIDCVR(0), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCVMIDCVR(1), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCVMIDCVR(2), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCVMIDCVR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCVMIDCVR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCVMIDCVR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCVMIDCVR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCVMIDCVR(7), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_PMSLATFR_EL1, 	HDFGRTR, PMSLATFR_EL1, 1),y
+	SR_FGT(SYS_PMSIRR_EL1, 		HDFGRTR, PMSIRR_EL1, 1),y
+	SR_FGT(SYS_PMSIDR_EL1, 		HDFGRTR, PMSIDR_EL1, 1),y
+	SR_FGT(SYS_PMSICR_EL1, 		HDFGRTR, PMSICR_EL1, 1),y
+	SR_FGT(SYS_PMSFCR_EL1, 		HDFGRTR, PMSFCR_EL1, 1),y
+	SR_FGT(SYS_PMSEVFR_EL1, 	HDFGRTR, PMSEVFR_EL1, 1),y
+	SR_FGT(SYS_PMSCR_EL1, 		HDFGRTR, PMSCR_EL1, 1),y
+	SR_FGT(SYS_PMBSR_EL1, 		HDFGRTR, PMBSR_EL1, 1),y
+	SR_FGT(SYS_PMBPTR_EL1, 		HDFGRTR, PMBPTR_EL1, 1),y
+	SR_FGT(SYS_PMBLIMITR_EL1, 	HDFGRTR, PMBLIMITR_EL1, 1),y
+	SR_FGT(SYS_PMMIR_EL1, 		HDFGRTR, PMMIR_EL1, 1),y
+	SR_FGT(SYS_PMSELR_EL0, 		HDFGRTR, PMSELR_EL0, 1),y
+	SR_FGT(SYS_PMOVSCLR_EL0, 	HDFGRTR, PMOVS, 1),y
+	SR_FGT(SYS_PMOVSSET_EL0, 	HDFGRTR, PMOVS, 1),y
+	SR_FGT(SYS_PMINTENCLR_EL1, 	HDFGRTR, PMINTEN, 1),y
+	SR_FGT(SYS_PMINTENSET_EL1, 	HDFGRTR, PMINTEN, 1),y
+	SR_FGT(SYS_PMCNTENCLR_EL0, 	HDFGRTR, PMCNTEN, 1),y
+	SR_FGT(SYS_PMCNTENSET_EL0, 	HDFGRTR, PMCNTEN, 1),y
+	SR_FGT(SYS_PMCCNTR_EL0, 	HDFGRTR, PMCCNTR_EL0, 1),y
+	SR_FGT(SYS_PMCCFILTR_EL0, 	HDFGRTR, PMCCFILTR_EL0, 1),y
+	SR_FGT(SYS_PMEVTYPERn_EL0(0), 	HDFGRTR, PMEVTYPERn_EL0, 1),y
+	SR_FGT(SYS_PMEVTYPERn_EL0(1), 	HDFGRTR, PMEVTYPERn_EL0, 1),y
+	SR_FGT(SYS_PMEVTYPERn_EL0(2), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(3), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(4), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(5), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(6), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(7), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(8), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(9), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(10), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(11), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(12), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(13), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(14), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(15), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(16), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(17), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(18), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(19), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(20), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(21), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(22), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(23), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(24), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(25), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(26), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(27), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(28), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(29), 	HDFGRTR, PMEVTYPERn_EL0, 1),
+	SR_FGT(SYS_PMEVTYPERn_EL0(30), 	HDFGRTR, PMEVTYPERn_EL0, 1),y
+	SR_FGT(SYS_PMEVCNTRn_EL0(0), 	HDFGRTR, PMEVCNTRn_EL0, 1),y
+	SR_FGT(SYS_PMEVCNTRn_EL0(1), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(2), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(3), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(4), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(5), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(6), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(7), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(8), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(9), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(10), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(11), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(12), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(13), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(14), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(15), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(16), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(17), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(18), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(19), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(20), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(21), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(22), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(23), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(24), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(25), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(26), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(27), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(28), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(29), 	HDFGRTR, PMEVCNTRn_EL0, 1),
+	SR_FGT(SYS_PMEVCNTRn_EL0(30), 	HDFGRTR, PMEVCNTRn_EL0, 1),y
+	SR_FGT(SYS_OSDLR_EL1, 		HDFGRTR, OSDLR_EL1, 1),y
+	SR_FGT(SYS_OSECCR_EL1, 		HDFGRTR, OSECCR_EL1, 1),y
+	SR_FGT(SYS_OSLSR_EL1, 		HDFGRTR, OSLSR_EL1, 1),y
+	SR_FGT(SYS_DBGPRCR_EL1, 	HDFGRTR, DBGPRCR_EL1, 1),y
+	SR_FGT(SYS_DBGAUTHSTATUS_EL1, 	HDFGRTR, DBGAUTHSTATUS_EL1, 1),y
+	SR_FGT(SYS_DBGCLAIMSET_EL1, 	HDFGRTR, DBGCLAIM, 1),y
+	SR_FGT(SYS_DBGCLAIMCLR_EL1, 	HDFGRTR, DBGCLAIM, 1),y
+	SR_FGT(SYS_MDSCR_EL1, 		HDFGRTR, MDSCR_EL1, 1),y
+	/*
+	 * The trap bits capture *64* debug registers per bit, but the
+	 * ARM ARM only describes the encoding for the first 16, and
+	 * we don't really support more than that anyway.
+	 */
+	SR_FGT(SYS_DBGWVRn_EL1(0), 	HDFGRTR, DBGWVRn_EL1, 1),y
+	SR_FGT(SYS_DBGWVRn_EL1(1), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(2), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(3), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(4), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(5), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(6), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(7), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(8), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(9), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(10), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(11), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(12), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(13), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(14), 	HDFGRTR, DBGWVRn_EL1, 1),
+	SR_FGT(SYS_DBGWVRn_EL1(15), 	HDFGRTR, DBGWVRn_EL1, 1),y
+	SR_FGT(SYS_DBGWCRn_EL1(0), 	HDFGRTR, DBGWCRn_EL1, 1),y
+	SR_FGT(SYS_DBGWCRn_EL1(1), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(2), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(3), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(4), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(5), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(6), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(7), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(8), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(9), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(10), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(11), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(12), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(13), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(14), 	HDFGRTR, DBGWCRn_EL1, 1),
+	SR_FGT(SYS_DBGWCRn_EL1(15), 	HDFGRTR, DBGWCRn_EL1, 1),y
+	SR_FGT(SYS_DBGBVRn_EL1(0), 	HDFGRTR, DBGBVRn_EL1, 1),y
+	SR_FGT(SYS_DBGBVRn_EL1(1), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(2), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(3), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(4), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(5), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(6), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(7), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(8), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(9), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(10), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(11), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(12), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(13), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(14), 	HDFGRTR, DBGBVRn_EL1, 1),
+	SR_FGT(SYS_DBGBVRn_EL1(15), 	HDFGRTR, DBGBVRn_EL1, 1),y
+	SR_FGT(SYS_DBGBCRn_EL1(0), 	HDFGRTR, DBGBCRn_EL1, 1),y
+	SR_FGT(SYS_DBGBCRn_EL1(1), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(2), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(3), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(4), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(5), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(6), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(7), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(8), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(9), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(10), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(11), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(12), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(13), 	HDFGRTR, DBGBCRn_EL1, 1),
+	SR_FGT(SYS_DBGBCRn_EL1(14), 	HDFGRTR, DBGBCRn_EL1, 1),

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1554

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+	SR_FGT(SYS_DBGBCRn_EL1(15), 	HDFGRTR, DBGBCRn_EL1, 1),y
+	/*
+	 * HDFGWTR_EL2
+	 *
+	 * Although HDFGRTR_EL2 and HDFGWTR_EL2 registers largely
+	 * overlap in their bit assignment, there are a number of bits
+	 * that are RES0 on one side, and an actual trap bit on the
+	 * other.  The policy chosen here is to describe all the
+	 * read-side mappings, and only the write-side mappings that
+	 * differ from the write side, and the trap handler will pick

differ from the read side?

Quoted 15 lines
arch/arm64/kvm/emulate-nested.c +8 -0
+	 * the correct shadow register based on the access type.
+	 */
+	SR_FGT(SYS_TRFCR_EL1,		HDFGWTR, TRFCR_EL1, 1),
+	SR_FGT(SYS_TRCOSLAR,		HDFGWTR, TRCOSLAR, 1),
+	SR_FGT(SYS_PMCR_EL0,		HDFGWTR, PMCR_EL0, 1),
+	SR_FGT(SYS_PMSWINC_EL0,		HDFGWTR, PMSWINC_EL0, 1),
+	SR_FGT(SYS_OSLAR_EL1,		HDFGWTR, OSLAR_EL1, 1),
 };
 
 static union trap_config get_trap_config(u32 sysreg)
@@ -1241,6 +1693,14 @@ bool __check_nv_sr_forward(struct kvm_vcpu *vcpu)
 			val = sanitised_sys_reg(vcpu, HFGWTR_EL2);
 		break;
 
+	case HDFGRTR_GROUP:

arch/arm64/kvm/emulate-nested.c bool __check_nv_sr_forward(struct kvm_vcpu *vcpu) · line 1706

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +7 -0
+	case HDFGWTR_GROUP:
+		if (is_read)
+			val = sanitised_sys_reg(vcpu, HDFGRTR_EL2);
+		else
+			val = sanitised_sys_reg(vcpu, HDFGWTR_EL2);
+		break;
+
 	case HFGITR_GROUP:
 		val = sanitised_sys_reg(vcpu, HFGITR_EL2);
 		break;

Thanks

Eric

author maintainer
#62 lore raw

Re: [PATCH v2 21/26] KVM: arm64: nv: Add trap forwarding for HDFGxTR_EL2

Hi Eric,

Quoted 10 lines · Eric Auger

Hi Marc,

Quoted 5 lines · Marc Zyngier

... and finally, the Debug version of FGT, with its *enormous* list of trapped registers.

Hi Marc, I think you mixed up with the R-b's sent on [PATCH v2 06/26] arm64: Add debug registers affected by HDFGxTR_EL2

Most probably. Really sorry about that. I'll fix that right away.

Quoted 4 lines
Quoted 4 lines
---
 arch/arm64/include/asm/kvm_arm.h |  11 +
 arch/arm64/kvm/emulate-nested.c  | 460 +++++++++++++++++++++++++++++++

[...]

Quoted 2 lines

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1241

Quoted 1 line
arch/arm64/kvm/emulate-nested.c +1 -0
+	SR_FGT(SYS_TRCIMSPEC0, 		HDFGRTR, TRCIMSPECn, 1),

why not SYS_TRCIMSPEC(0)?

Yup, it's been mentioned a couple of time already, now fixed.

Quoted 10 lines

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1250

Quoted 9 lines
arch/arm64/kvm/emulate-nested.c +9 -0
+	SR_FGT(SYS_TRCIMSPEC(1), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(2), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(3), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(4), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(5), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(6), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCIMSPEC(7), 	HDFGRTR, TRCIMSPECn, 1),
+	SR_FGT(SYS_TRCDEVARCH, 		HDFGRTR, TRCID, 1),
+	SR_FGT(SYS_TRCDEVID, 		HDFGRTR, TRCID, 1),

what about all of the TRCIDR regs refered to in the spec?

Ah, nothing escapes you! Yup totally missed it, now added.

[...]

Quoted 33 lines
Quoted 21 lines
arch/arm64/kvm/emulate-nested.c +21 -0
+	SR_FGT(SYS_TRCRSCTLR(2), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCRSCTLR(3), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(4), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(5), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(6), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(7), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(8), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(9), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(10), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(11), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(12), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(13), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(14), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(15), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(16), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(17), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(18), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(19), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(20), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(21), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(22), 	HDFGRTR, TRC, 1),

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1348

Quoted 10 lines
arch/arm64/kvm/emulate-nested.c +10 -0
+	SR_FGT(SYS_TRCRSCTLR(23), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(24), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(25), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(26), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(27), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(28), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(29), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(30), 	HDFGRTR, TRC, 1),
+	SR_FGT(SYS_TRCRSCTLR(31), 	HDFGRTR, TRC, 1),y
+	SR_FGT(SYS_TRCQCTLR, 		HDFGRTR, TRC, 1),

nit: maybe put this one before the

Quoted 1 line

SYS_TRCRSCTLR(n) series to follow the spec order

Done.

[...]

Quoted 10 lines

arch/arm64/kvm/emulate-nested.c static const struct encoding_to_trap_config encoding_to_fgt[] __initconst = { · line 1554

Quoted 9 lines
arch/arm64/kvm/emulate-nested.c +9 -0
+	/*
+	 * HDFGWTR_EL2
+	 *
+	 * Although HDFGRTR_EL2 and HDFGWTR_EL2 registers largely
+	 * overlap in their bit assignment, there are a number of bits
+	 * that are RES0 on one side, and an actual trap bit on the
+	 * other.  The policy chosen here is to describe all the
+	 * read-side mappings, and only the write-side mappings that
+	 * differ from the write side, and the trap handler will pick

differ from the read side?

Well spotted! Now fixed.

Thanks again!

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
author maintainer
#63 lore raw

PATCH v2 22/26 KVM: arm64: nv: Add SVC trap forwarding

In reply to Marc Zyngier

HFGITR_EL2 allows the trap of SVC instructions to EL2. Allow these traps to be forwarded. Take this opportunity to deny any 32bit activity when NV is enabled.

---
 arch/arm64/kvm/arm.c         |  4 ++++
 arch/arm64/kvm/handle_exit.c | 12 ++++++++++++
 2 files changed, 16 insertions(+)
arch/arm64/kvm/arm.c +4 -0
index 72dc53a75d1c..8b51570a76f8 100644
@@ -36,6 +36,7 @@
 #include <asm/kvm_arm.h>
 #include <asm/kvm_asm.h>
 #include <asm/kvm_mmu.h>
+#include <asm/kvm_nested.h>
 #include <asm/kvm_pkvm.h>
 #include <asm/kvm_emulate.h>
 #include <asm/sections.h>
@@ -818,6 +819,9 @@ static bool vcpu_mode_is_bad_32bit(struct kvm_vcpu *vcpu)
 	if (likely(!vcpu_mode_is_32bit(vcpu)))
 		return false;
 
+	if (vcpu_has_nv(vcpu))
+		return true;
+
 	return !kvm_supports_32bit_el0();
 }
 
arch/arm64/kvm/handle_exit.c +12 -0
index 6dcd6604b6bc..3b86d534b995 100644
@@ -226,6 +226,17 @@ static int kvm_handle_eret(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
+static int handle_svc(struct kvm_vcpu *vcpu)
+{
+	/*
+	 * So far, SVC traps only for NV via HFGITR_EL2. A SVC from a
+	 * 32bit guest would be caught by vpcu_mode_is_bad_32bit(), so
+	 * we should only have to deal with a 64 bit exception.
+	 */
+	kvm_inject_nested_sync(vcpu, kvm_vcpu_get_esr(vcpu));
+	return 1;
+}
+
 static exit_handle_fn arm_exit_handlers[] = {
 	[0 ... ESR_ELx_EC_MAX]	= kvm_handle_unknown_ec,
 	[ESR_ELx_EC_WFx]	= kvm_handle_wfx,
@@ -239,6 +250,7 @@ static exit_handle_fn arm_exit_handlers[] = {
 	[ESR_ELx_EC_SMC32]	= handle_smc,
 	[ESR_ELx_EC_HVC64]	= handle_hvc,
 	[ESR_ELx_EC_SMC64]	= handle_smc,
+	[ESR_ELx_EC_SVC64]	= handle_svc,
 	[ESR_ELx_EC_SYS64]	= kvm_handle_sys_reg,
 	[ESR_ELx_EC_SVE]	= handle_sve,
 	[ESR_ELx_EC_ERET]	= kvm_handle_eret,
Signature · 1 line
2.34.1
author maintainer
#64 lore raw

PATCH v2 23/26 KVM: arm64: nv: Add switching support for HFGxTR/HDFGxTR

In reply to Marc Zyngier

Now that we can evaluate the FGT registers, allow them to be merged with the hypervisor's own configuration (in the case of HFG{RW}TR_EL2) or simply set for HFGITR_EL2, HDGFRTR_EL2 and HDFGWTR_EL2.

---
 arch/arm64/kvm/hyp/include/hyp/switch.h | 48 +++++++++++++++++++++++++
 1 file changed, 48 insertions(+)
arch/arm64/kvm/hyp/include/hyp/switch.h +48 -0
index 966295178aee..ad9c730b7d1a 100644
@@ -70,6 +70,13 @@ static inline void __activate_traps_fpsimd32(struct kvm_vcpu *vcpu)
 	}
 }
 
+#define compute_clr_set(vcpu, reg, clr, set)				\
+	do {								\
+		u64 hfg;						\
+		hfg = __vcpu_sys_reg(vcpu, reg) & ~__ ## reg ## _RES0;	\
+		set |= hfg & __ ## reg ## _MASK; 			\
+		clr |= ~hfg & __ ## reg ## _nMASK; 			\
+	} while(0)
 
 
 static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu)
@@ -97,6 +104,10 @@ static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 	if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38))
 		w_set |= HFGxTR_EL2_TCR_EL1_MASK;
 
+	if (vcpu_has_nv(vcpu) && !is_hyp_ctxt(vcpu)) {
+		compute_clr_set(vcpu, HFGRTR_EL2, r_clr, r_set);
+		compute_clr_set(vcpu, HFGWTR_EL2, w_clr, w_set);
+	}
 
 	/* The default is not to trap amything but ACCDATA_EL1 */
 	r_val = __HFGRTR_EL2_nMASK & ~HFGxTR_EL2_nACCDATA_EL1;
@@ -109,6 +120,38 @@ static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 
 	write_sysreg_s(r_val, SYS_HFGRTR_EL2);
 	write_sysreg_s(w_val, SYS_HFGWTR_EL2);
+
+	if (!vcpu_has_nv(vcpu) || is_hyp_ctxt(vcpu))
+		return;
+
+	ctxt_sys_reg(hctxt, HFGITR_EL2) = read_sysreg_s(SYS_HFGITR_EL2);
+
+	r_set = r_clr = 0;
+	compute_clr_set(vcpu, HFGITR_EL2, r_clr, r_set);
+	r_val = __HFGITR_EL2_nMASK;
+	r_val |= r_set;
+	r_val &= ~r_clr;
+
+	write_sysreg_s(r_val, SYS_HFGITR_EL2);
+
+	ctxt_sys_reg(hctxt, HDFGRTR_EL2) = read_sysreg_s(SYS_HDFGRTR_EL2);
+	ctxt_sys_reg(hctxt, HDFGWTR_EL2) = read_sysreg_s(SYS_HDFGWTR_EL2);
+
+	r_clr = r_set = w_clr = w_set = 0;
+
+	compute_clr_set(vcpu, HDFGRTR_EL2, r_clr, r_set);
+	compute_clr_set(vcpu, HDFGWTR_EL2, w_clr, w_set);
+
+	r_val = __HDFGRTR_EL2_nMASK;
+	r_val |= r_set;
+	r_val &= ~r_clr;
+
+	w_val = __HDFGWTR_EL2_nMASK;
+	w_val |= w_set;
+	w_val &= ~w_clr;
+
+	write_sysreg_s(r_val, SYS_HDFGRTR_EL2);
+	write_sysreg_s(w_val, SYS_HDFGWTR_EL2);
 }
 
 static inline void __deactivate_traps_hfgxtr(struct kvm_vcpu *vcpu)
@@ -121,7 +164,12 @@ static inline void __deactivate_traps_hfgxtr(struct kvm_vcpu *vcpu)
 	write_sysreg_s(ctxt_sys_reg(hctxt, HFGRTR_EL2), SYS_HFGRTR_EL2);
 	write_sysreg_s(ctxt_sys_reg(hctxt, HFGWTR_EL2), SYS_HFGWTR_EL2);
 
+	if (!vcpu_has_nv(vcpu) || is_hyp_ctxt(vcpu))
+		return;
 
+	write_sysreg_s(ctxt_sys_reg(hctxt, HFGITR_EL2), SYS_HFGITR_EL2);
+	write_sysreg_s(ctxt_sys_reg(hctxt, HDFGRTR_EL2), SYS_HDFGRTR_EL2);
+	write_sysreg_s(ctxt_sys_reg(hctxt, HDFGWTR_EL2), SYS_HDFGWTR_EL2);
 }
 
 static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
Signature · 1 line
2.34.1
author maintainer
#65 lore raw

PATCH v2 24/26 KVM: arm64: nv: Expose FGT to nested guests

In reply to Marc Zyngier

Now that we have FGT support, expose the feature to NV guests.

---
 arch/arm64/kvm/nested.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
arch/arm64/kvm/nested.c +3 -2
index 7f80f385d9e8..3facd8918ae3 100644
@@ -71,8 +71,9 @@ void access_nested_id_reg(struct kvm_vcpu *v, struct sys_reg_params *p,
 		break;
 
 	case SYS_ID_AA64MMFR0_EL1:
-		/* Hide ECV, FGT, ExS, Secure Memory */
-		val &= ~(GENMASK_ULL(63, 43)		|
+		/* Hide ECV, ExS, Secure Memory */
+		val &= ~(NV_FTR(MMFR0, ECV)		|
+			 NV_FTR(MMFR0, EXS)		|
 			 NV_FTR(MMFR0, TGRAN4_2)	|
 			 NV_FTR(MMFR0, TGRAN16_2)	|
 			 NV_FTR(MMFR0, TGRAN64_2)	|
Signature · 1 line
2.34.1
author maintainer
#66 lore raw

PATCH v2 25/26 KVM: arm64: Move HCRX_EL2 switch to load/put on VHE systems

In reply to Marc Zyngier

Although the nVHE behaviour requires HCRX_EL2 to be switched on each switch between host and guest, there is nothing in this register that would affect a VHE host.

It is thus possible to save/restore this register on load/put on VHE systems, avoiding unnecessary sysreg access on the hot path. Additionally, it avoids unnecessary traps when running with NV.

To achieve this, simply move the read/writes to the *_common() helpers, which are called on load/put on VHE, and more eagerly on nVHE.

---
 arch/arm64/kvm/hyp/include/hyp/switch.h | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
arch/arm64/kvm/hyp/include/hyp/switch.h +6 -6
index ad9c730b7d1a..5bcffd1906ce 100644
@@ -197,6 +197,9 @@ static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
 	vcpu->arch.mdcr_el2_host = read_sysreg(mdcr_el2);
 	write_sysreg(vcpu->arch.mdcr_el2, mdcr_el2);
 
+	if (cpus_have_final_cap(ARM64_HAS_HCX))
+		write_sysreg_s(HCRX_GUEST_FLAGS, SYS_HCRX_EL2);
+
 	__activate_traps_hfgxtr(vcpu);
 }
 
@@ -213,6 +216,9 @@ static inline void __deactivate_traps_common(struct kvm_vcpu *vcpu)
 		vcpu_clear_flag(vcpu, PMUSERENR_ON_CPU);
 	}
 
+	if (cpus_have_final_cap(ARM64_HAS_HCX))
+		write_sysreg_s(HCRX_HOST_FLAGS, SYS_HCRX_EL2);
+
 	__deactivate_traps_hfgxtr(vcpu);
 }
 
@@ -227,9 +233,6 @@ static inline void ___activate_traps(struct kvm_vcpu *vcpu)
 
 	if (cpus_have_final_cap(ARM64_HAS_RAS_EXTN) && (hcr & HCR_VSE))
 		write_sysreg_s(vcpu->arch.vsesr_el2, SYS_VSESR_EL2);
-
-	if (cpus_have_final_cap(ARM64_HAS_HCX))
-		write_sysreg_s(HCRX_GUEST_FLAGS, SYS_HCRX_EL2);
 }
 
 static inline void ___deactivate_traps(struct kvm_vcpu *vcpu)
@@ -244,9 +247,6 @@ static inline void ___deactivate_traps(struct kvm_vcpu *vcpu)
 		vcpu->arch.hcr_el2 &= ~HCR_VSE;
 		vcpu->arch.hcr_el2 |= read_sysreg(hcr_el2) & HCR_VSE;
 	}
-
-	if (cpus_have_final_cap(ARM64_HAS_HCX))
-		write_sysreg_s(HCRX_HOST_FLAGS, SYS_HCRX_EL2);
 }
 
 static inline bool __populate_fault_info(struct kvm_vcpu *vcpu)
Signature · 1 line
2.34.1
author maintainer
#67 lore raw

PATCH v2 26/26 KVM: arm64: nv: Add support for HCRX_EL2

In reply to Marc Zyngier

HCRX_EL2 has an interesting effect on HFGITR_EL2, as it conditions the traps of TLBI*nXS.

Expand the FGT support to add a new Fine Grained Filter that will get checked when the instruction gets trapped, allowing the shadow register to override the trap as needed.

---
 arch/arm64/include/asm/kvm_arm.h        |  5 ++
 arch/arm64/include/asm/kvm_host.h       |  1 +
 arch/arm64/kvm/emulate-nested.c         | 89 +++++++++++++++----------
 arch/arm64/kvm/hyp/include/hyp/switch.h | 15 ++++-
 arch/arm64/kvm/nested.c                 |  3 +-
 arch/arm64/kvm/sys_regs.c               |  2 +
 6 files changed, 78 insertions(+), 37 deletions(-)
arch/arm64/include/asm/kvm_arm.h +5 -0
index d229f238c3b6..137f732789c9 100644
@@ -369,6 +369,11 @@
 #define __HDFGWTR_EL2_MASK	~__HDFGWTR_EL2_nMASK
 #define __HDFGWTR_EL2_nMASK	GENMASK(62, 60)
 
+/* Similar definitions for HCRX_EL2 */
+#define __HCRX_EL2_RES0		(GENMASK(63, 16) | GENMASK(13, 12))
+#define __HCRX_EL2_MASK		(0)
+#define __HCRX_EL2_nMASK	(GENMASK(15, 14) | GENMASK(4, 0))
+
 /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
 #define HPFAR_MASK	(~UL(0xf))
 /*
arch/arm64/include/asm/kvm_host.h +1 -0
index d44575604e7e..dec1de1bb503 100644
@@ -380,6 +380,7 @@ enum vcpu_sysreg {
 	CPTR_EL2,	/* Architectural Feature Trap Register (EL2) */
 	HSTR_EL2,	/* Hypervisor System Trap Register */
 	HACR_EL2,	/* Hypervisor Auxiliary Control Register */
+	HCRX_EL2,	/* Extended Hypervisor Configuration Register */
 	TTBR0_EL2,	/* Translation Table Base Register 0 (EL2) */
 	TTBR1_EL2,	/* Translation Table Base Register 1 (EL2) */
 	TCR_EL2,	/* Translation Control Register (EL2) */
arch/arm64/kvm/emulate-nested.c +55 -34

128 lines not on this page · raw message

arch/arm64/kvm/hyp/include/hyp/switch.h +13 -2
index 5bcffd1906ce..0ce72af656b3 100644
@@ -197,8 +197,19 @@ static inline void __activate_traps_common(struct kvm_vcpu *vcpu)
 	vcpu->arch.mdcr_el2_host = read_sysreg(mdcr_el2);
 	write_sysreg(vcpu->arch.mdcr_el2, mdcr_el2);
 
-	if (cpus_have_final_cap(ARM64_HAS_HCX))
-		write_sysreg_s(HCRX_GUEST_FLAGS, SYS_HCRX_EL2);
+	if (cpus_have_final_cap(ARM64_HAS_HCX)) {
+		u64 hcrx = HCRX_GUEST_FLAGS;
+		if (vcpu_has_nv(vcpu) && !is_hyp_ctxt(vcpu)) {
+			u64 clr = 0, set = 0;
+
+			compute_clr_set(vcpu, HCRX_EL2, clr, set);
+
+			hcrx |= set;
+			hcrx &= ~clr;
+		}
+
+		write_sysreg_s(hcrx, SYS_HCRX_EL2);
+	}
 
 	__activate_traps_hfgxtr(vcpu);
 }
arch/arm64/kvm/nested.c +2 -1
index 3facd8918ae3..042695a210ce 100644
@@ -117,7 +117,8 @@ void access_nested_id_reg(struct kvm_vcpu *v, struct sys_reg_params *p,
 		break;
 
 	case SYS_ID_AA64MMFR1_EL1:
-		val &= (NV_FTR(MMFR1, PAN)	|
+		val &= (NV_FTR(MMFR1, HCX)	|
+			NV_FTR(MMFR1, PAN)	|
 			NV_FTR(MMFR1, LO)	|
 			NV_FTR(MMFR1, HPDS)	|
 			NV_FTR(MMFR1, VH)	|
arch/arm64/kvm/sys_regs.c +2 -0
index dfd72b3a625f..374b21f08fc3 100644
@@ -2372,6 +2372,8 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 	EL2_REG(HFGITR_EL2, access_rw, reset_val, 0),
 	EL2_REG(HACR_EL2, access_rw, reset_val, 0),
 
+	EL2_REG(HCRX_EL2, access_rw, reset_val, 0),
+
 	EL2_REG(TTBR0_EL2, access_rw, reset_val, 0),
 	EL2_REG(TTBR1_EL2, access_rw, reset_val, 0),
 	EL2_REG(TCR_EL2, access_rw, reset_val, TCR_EL2_RES1),
Signature · 1 line
2.34.1