As per DDI 0487 RWTXBY we need to manage PSTATE.EXLOCK when entering an
exception, when the exception is entered from a lower EL the bit is cleared
while if entering from the same EL it is set to GCSCR_ELx.EXLOCKEN.
Implement this behaviour in enter_exception64().
Signed-off-by: Mark Brown <broonie@kernel.org>
---
arch/arm64/include/uapi/asm/ptrace.h | 2 ++
arch/arm64/kvm/hyp/exception.c | 10 ++++++++++
2 files changed, 12 insertions(+)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index 0f39ba4f3efd4a8760f0fca0fbf1a2563b191c7d..9987957f4f7137bf107653b817885bb976853a83 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -37,6 +37,7 @@
#define PSR_MODE_EL3t 0x0000000c
#define PSR_MODE_EL3h 0x0000000d
#define PSR_MODE_MASK 0x0000000f
+#define PSR_EL_MASK 0x0000000c
/* AArch32 CPSR bits */
#define PSR_MODE32_BIT 0x00000010
@@ -56,6 +57,7 @@
#define PSR_C_BIT 0x20000000
#define PSR_Z_BIT 0x40000000
#define PSR_N_BIT 0x80000000
+#define PSR_EXLOCK_BIT 0x400000000
#define PSR_BTYPE_SHIFT 10
diff --git a/arch/arm64/kvm/hyp/exception.c b/arch/arm64/kvm/hyp/exception.c
index 424a5107cddb5e1cdd75ef3581adef03aaadabb7..0d41b9b75cf83250b2c0d20cd82c153869efb0e4 100644
--- a/arch/arm64/kvm/hyp/exception.c
+++ b/arch/arm64/kvm/hyp/exception.c
@@ -160,6 +160,16 @@ static void enter_exception64(struct kvm_vcpu *vcpu, unsigned long target_mode,
// PSTATE.BTYPE is set to zero upon any exception to AArch64
// See ARM DDI 0487E.a, pages D1-2293 to D1-2294.
+ // PSTATE.EXLOCK is set to 0 upon any exception to a higher
+ // EL, or to GCSCR_ELx.EXLOCKEN for an exception to the same
+ // exception level. See ARM DDI 0487 RWTXBY, D.1.3.2 in K.a.
+ if (kvm_has_gcs(vcpu->kvm) &&
+ (target_mode & PSR_EL_MASK) == (mode & PSR_EL_MASK)) {
+ u64 gcscr = __vcpu_read_sys_reg(vcpu, GCSCR_EL1);
+ if (gcscr & GCSCR_ELx_EXLOCKEN)
+ new |= PSR_EXLOCK_BIT;
+ }
+
new |= PSR_D_BIT;
new |= PSR_A_BIT;
new |= PSR_I_BIT;
--
2.39.2
On Sat, 05 Oct 2024 11:37:31 +0100,
Mark Brown <broonie@kernel.org> wrote:
>
> As per DDI 0487 RWTXBY we need to manage PSTATE.EXLOCK when entering an
> exception, when the exception is entered from a lower EL the bit is cleared
> while if entering from the same EL it is set to GCSCR_ELx.EXLOCKEN.
> Implement this behaviour in enter_exception64().
>
> Signed-off-by: Mark Brown <broonie@kernel.org>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 2 ++
> arch/arm64/kvm/hyp/exception.c | 10 ++++++++++
> 2 files changed, 12 insertions(+)
>
> diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> index 0f39ba4f3efd4a8760f0fca0fbf1a2563b191c7d..9987957f4f7137bf107653b817885bb976853a83 100644
> --- a/arch/arm64/include/uapi/asm/ptrace.h
> +++ b/arch/arm64/include/uapi/asm/ptrace.h
> @@ -37,6 +37,7 @@
> #define PSR_MODE_EL3t 0x0000000c
> #define PSR_MODE_EL3h 0x0000000d
> #define PSR_MODE_MASK 0x0000000f
> +#define PSR_EL_MASK 0x0000000c
>
> /* AArch32 CPSR bits */
> #define PSR_MODE32_BIT 0x00000010
> @@ -56,6 +57,7 @@
> #define PSR_C_BIT 0x20000000
> #define PSR_Z_BIT 0x40000000
> #define PSR_N_BIT 0x80000000
> +#define PSR_EXLOCK_BIT 0x400000000
>
> #define PSR_BTYPE_SHIFT 10
>
> diff --git a/arch/arm64/kvm/hyp/exception.c b/arch/arm64/kvm/hyp/exception.c
> index 424a5107cddb5e1cdd75ef3581adef03aaadabb7..0d41b9b75cf83250b2c0d20cd82c153869efb0e4 100644
> --- a/arch/arm64/kvm/hyp/exception.c
> +++ b/arch/arm64/kvm/hyp/exception.c
> @@ -160,6 +160,16 @@ static void enter_exception64(struct kvm_vcpu *vcpu, unsigned long target_mode,
> // PSTATE.BTYPE is set to zero upon any exception to AArch64
> // See ARM DDI 0487E.a, pages D1-2293 to D1-2294.
>
> + // PSTATE.EXLOCK is set to 0 upon any exception to a higher
> + // EL, or to GCSCR_ELx.EXLOCKEN for an exception to the same
> + // exception level. See ARM DDI 0487 RWTXBY, D.1.3.2 in K.a.
> + if (kvm_has_gcs(vcpu->kvm) &&
> + (target_mode & PSR_EL_MASK) == (mode & PSR_EL_MASK)) {
> + u64 gcscr = __vcpu_read_sys_reg(vcpu, GCSCR_EL1);
No, please. This only works by luck when a guest has AArch32 EL0, and
creates more havoc on a NV guest. In general, this PSR_EL_MASK creates
more problem than anything else, and doesn't fit the rest of the code.
So this needs to:
- explicitly only apply to exceptions from AArch64
- handle exception from EL2, since this helper already deals with that
The latter point of course means introducing GCSCR_EL2 (and everything
that depends on it, such as the trap handling).
M.
--
Without deviation from the norm, progress is not possible.
On Sat, Oct 05, 2024 at 01:36:09PM +0100, Marc Zyngier wrote:
> Mark Brown <broonie@kernel.org> wrote:
> > + // PSTATE.EXLOCK is set to 0 upon any exception to a higher
> > + // EL, or to GCSCR_ELx.EXLOCKEN for an exception to the same
> > + // exception level. See ARM DDI 0487 RWTXBY, D.1.3.2 in K.a.
> > + if (kvm_has_gcs(vcpu->kvm) &&
> > + (target_mode & PSR_EL_MASK) == (mode & PSR_EL_MASK)) {
> > + u64 gcscr = __vcpu_read_sys_reg(vcpu, GCSCR_EL1);
> No, please. This only works by luck when a guest has AArch32 EL0, and
> creates more havoc on a NV guest. In general, this PSR_EL_MASK creates
> more problem than anything else, and doesn't fit the rest of the code.
You say luck, I say careful architecture definition but sure.
> So this needs to:
> - explicitly only apply to exceptions from AArch64
> - handle exception from EL2, since this helper already deals with that
> The latter point of course means introducing GCSCR_EL2 (and everything
> that depends on it, such as the trap handling).
For clarity, which trap handling specifically?
On Sat, 05 Oct 2024 15:14:21 +0100,
Mark Brown <broonie@kernel.org> wrote:
>
> On Sat, Oct 05, 2024 at 01:36:09PM +0100, Marc Zyngier wrote:
> > Mark Brown <broonie@kernel.org> wrote:
>
> > > + // PSTATE.EXLOCK is set to 0 upon any exception to a higher
> > > + // EL, or to GCSCR_ELx.EXLOCKEN for an exception to the same
> > > + // exception level. See ARM DDI 0487 RWTXBY, D.1.3.2 in K.a.
> > > + if (kvm_has_gcs(vcpu->kvm) &&
> > > + (target_mode & PSR_EL_MASK) == (mode & PSR_EL_MASK)) {
> > > + u64 gcscr = __vcpu_read_sys_reg(vcpu, GCSCR_EL1);
>
> > No, please. This only works by luck when a guest has AArch32 EL0, and
> > creates more havoc on a NV guest. In general, this PSR_EL_MASK creates
> > more problem than anything else, and doesn't fit the rest of the code.
>
> You say luck, I say careful architecture definition but sure.
I wasn't talking about the architecture, but sure.
>
> > So this needs to:
> > - explicitly only apply to exceptions from AArch64
> > - handle exception from EL2, since this helper already deals with that
>
> > The latter point of course means introducing GCSCR_EL2 (and everything
> > that depends on it, such as the trap handling).
>
> For clarity, which trap handling specifically?
All the traps described in the GCSCR_EL2 documentation -- I see two
control bits described in K.a, all of which needs to be propagated and
their effects handled. Similarly, GCSPR_EL2 needs to be defined.
M.
--
Without deviation from the norm, progress is not possible.
© 2016 - 2026 Red Hat, Inc.