KVM/arm fixes for 5.7, take #2
Subsystems (4)
Marc Zyngier
maintainerPaolo,
This is the second batch of KVM/arm fixes for 5.7. A compilation fix, a GICv4.1 fix, plus a couple of sanity checks (SP_EL0 save/restore, and the sanitising of AArch32 registers).
Note that the pull request I sent a week ago[1] is still valid, and that this new series is built on top of the previous one.
Please pull,
M.
[1] https://lore.kernel.org/kvm/[email protected]/
The following changes since commit 446c0768f5509793a0e527a439d4866b24707b0e:
Merge branch 'kvm-arm64/vgic-fixes-5.7' into kvmarm-master/master (2020-04-23 16:27:33 +0100)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm.git tags/kvmarm-fixes-5.7-2
for you to fetch changes up to 0225fd5e0a6a32af7af0aefac45c8ebf19dc5183:
KVM: arm64: Fix 32bit PC wrap-around (2020-05-01 09:51:08 +0100)
----------------------------------------------------------------
KVM/arm fixes for Linux 5.7, take #2
- Fix compilation with Clang
- Correctly initialize GICv4.1 in the absence of a virtual ITS
- Move SP_EL0 save/restore to the guest entry/exit code
- Handle PC wrap around on 32bit guests, and narrow all 32bit registers on userspace access
----------------------------------------------------------------
Fangrui Song (1):
KVM: arm64: Delete duplicated label in invalid_vector
Marc Zyngier (3):
KVM: arm64: Save/restore sp_el0 as part of __guest_enter
KVM: arm64: vgic-v4: Initialize GICv4.1 even in the absence of a virtual ITS
KVM: arm64: Fix 32bit PC wrap-around
arch/arm64/kvm/guest.c | 7 +++++++ arch/arm64/kvm/hyp/entry.S | 23 +++++++++++++++++++++++ arch/arm64/kvm/hyp/hyp-entry.S | 1 - arch/arm64/kvm/hyp/sysreg-sr.c | 17 +++-------------- virt/kvm/arm/hyp/aarch32.c | 8 ++++++-- virt/kvm/arm/vgic/vgic-init.c | 9 ++++++++- virt/kvm/arm/vgic/vgic-mmio-v3.c | 3 ++- 7 files changed, 49 insertions(+), 19 deletions(-)
_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
Marc Zyngier
maintainerPATCH 1/4 KVM: arm64: Delete duplicated label in invalid_vector
From: Fangrui Song <[email protected]>
SYM_CODE_START defines \label , so it is redundant to define \label again. A redefinition at the same place is accepted by GNU as
(https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=159fbb6088f17a341bcaaac960623cab881b4981)
but rejected by the clang integrated assembler.
--- arch/arm64/kvm/hyp/hyp-entry.S | 1 - 1 file changed, 1 deletion(-)
arch/arm64/kvm/hyp/hyp-entry.S +0 -1
@@ -198,7 +198,6 @@ SYM_CODE_END(__hyp_panic) .macro invalid_vector label, target = __hyp_panic .align 2 SYM_CODE_START(\label) -\label: b \target SYM_CODE_END(\label) .endm
Signature · 1 line
2.26.2
Marc Zyngier
maintainerPATCH 2/4 KVM: arm64: Save/restore sp_el0 as part of __guest_enter
In reply to Marc ZyngierWe currently save/restore sp_el0 in C code. This is a bit unsafe, as a lot of the C code expects 'current' to be accessible from there (and the opportunity to run kernel code in HYP is specially great with VHE).
Instead, let's move the save/restore of sp_el0 to the assembly code (in __guest_enter), making sure that sp_el0 is correct very early on when we exit the guest, and is preserved as long as possible to its host value when we enter the guest.
--- arch/arm64/kvm/hyp/entry.S | 23 +++++++++++++++++++++++ arch/arm64/kvm/hyp/sysreg-sr.c | 17 +++-------------- 2 files changed, 26 insertions(+), 14 deletions(-)
arch/arm64/kvm/hyp/entry.S +23 -0
@@ -18,6 +18,7 @@ #define CPU_GP_REG_OFFSET(x) (CPU_GP_REGS + x) #define CPU_XREG_OFFSET(x) CPU_GP_REG_OFFSET(CPU_USER_PT_REGS + 8*x) +#define CPU_SP_EL0_OFFSET (CPU_XREG_OFFSET(30) + 8) .text .pushsection .hyp.text, "ax"
@@ -47,6 +48,16 @@ ldp x29, lr, [\ctxt, #CPU_XREG_OFFSET(29)] .endm +.macro save_sp_el0 ctxt, tmp + mrs \tmp, sp_el0 + str \tmp, [\ctxt, #CPU_SP_EL0_OFFSET] +.endm + +.macro restore_sp_el0 ctxt, tmp + ldr \tmp, [\ctxt, #CPU_SP_EL0_OFFSET] + msr sp_el0, \tmp +.endm + /* * u64 __guest_enter(struct kvm_vcpu *vcpu, * struct kvm_cpu_context *host_ctxt);
@@ -60,6 +71,9 @@ SYM_FUNC_START(__guest_enter) // Store the host regs save_callee_saved_regs x1 + // Save the host's sp_el0 + save_sp_el0 x1, x2 + // Now the host state is stored if we have a pending RAS SError it must // affect the host. If any asynchronous exception is pending we defer // the guest entry. The DSB isn't necessary before v8.2 as any SError
@@ -83,6 +97,9 @@ alternative_else_nop_endif // when this feature is enabled for kernel code. ptrauth_switch_to_guest x29, x0, x1, x2 + // Restore the guest's sp_el0 + restore_sp_el0 x29, x0 + // Restore guest regs x0-x17 ldp x0, x1, [x29, #CPU_XREG_OFFSET(0)] ldp x2, x3, [x29, #CPU_XREG_OFFSET(2)]
@@ -130,6 +147,9 @@ SYM_INNER_LABEL(__guest_exit, SYM_L_GLOBAL) // Store the guest regs x18-x29, lr save_callee_saved_regs x1 + // Store the guest's sp_el0 + save_sp_el0 x1, x2 + get_host_ctxt x2, x3 // Macro ptrauth_switch_to_guest format:
@@ -139,6 +159,9 @@ SYM_INNER_LABEL(__guest_exit, SYM_L_GLOBAL) // when this feature is enabled for kernel code. ptrauth_switch_to_host x1, x2, x3, x4, x5 + // Restore the hosts's sp_el0 + restore_sp_el0 x2, x3 + // Now restore the host regs restore_callee_saved_regs x2
arch/arm64/kvm/hyp/sysreg-sr.c +3 -14
@@ -15,8 +15,9 @@/* * Non-VHE: Both host and guest must save everything. * -* VHE: Host and guest must save mdscr_el1 and sp_el0 (and the PC and pstate, -* which are handled as part of the el2 return state) on every switch. +* VHE: Host and guest must save mdscr_el1 and sp_el0 (and the PC and +* pstate, which are handled as part of the el2 return state) on every +* switch (sp_el0 is being dealt with in the assembly code). * tpidr_el0 and tpidrro_el0 only need to be switched when going * to host userspace or a different VCPU. EL1 registers only need to be * switched when potentially going to run a different VCPU. The latter two
@@ -26,12 +27,6 @@ static void __hyp_text __sysreg_save_common_state(struct kvm_cpu_context *ctxt) { ctxt->sys_regs[MDSCR_EL1 ] =read_sysreg (mdscr_el1); - -/* -* The host arm64 Linux uses sp_el0 to point to 'current' and it must -* therefore be saved/restored on every entry/exit to/from the guest. -*/ - ctxt->gp_regs.regs.sp =read_sysreg (sp_el0); } static void __hyp_text __sysreg_save_user_state(struct kvm_cpu_context *ctxt)
@@ -99,12 +94,6 @@ NOKPROBE_SYMBOL(sysreg_save_guest_state_vhe); static void __hyp_text __sysreg_restore_common_state(struct kvm_cpu_context *ctxt) {write_sysreg (ctxt->sys_regs[MDSCR_EL1 ], mdscr_el1); - -/* -* The host arm64 Linux uses sp_el0 to point to 'current' and it must -* therefore be saved/restored on every entry/exit to/from the guest. -*/ -write_sysreg (ctxt->gp_regs.regs.sp, sp_el0); } static void __hyp_text __sysreg_restore_user_state(struct kvm_cpu_context *ctxt)
Signature · 1 line
2.26.2
Marc Zyngier
maintainerPATCH 3/4 KVM: arm64: vgic-v4: Initialize GICv4.1 even in the absence of a virtual ITS
In reply to Marc ZyngierKVM now expects to be able to use HW-accelerated delivery of vSGIs as soon as the guest has enabled thm. Unfortunately, we only initialize the GICv4 context if we have a virtual ITS exposed to the guest.
Fix it by always initializing the GICv4.1 context if it is available on the host.
--- virt/kvm/arm/vgic/vgic-init.c | 9 ++++++++- virt/kvm/arm/vgic/vgic-mmio-v3.c | 3 ++- 2 files changed, 10 insertions(+), 2 deletions(-)
virt/kvm/arm/vgic/vgic-init.c +8 -1
@@ -294,8 +294,15 @@ int vgic_init(struct kvm *kvm) } } -if (vgic_has_its (kvm)) { +if (vgic_has_its (kvm))vgic_lpi_translation_cache_init (kvm); + +/* +* If we have GICv4.1 enabled, unconditionnaly request enable the +* v4 support so that we get HW-accelerated vSGIs. Otherwise, only +* enable it if we present a virtual ITS to the guest. +*/ +if (vgic_supports_direct_msis (kvm)) { ret =vgic_v4_init (kvm);if (ret) goto out;
virt/kvm/arm/vgic/vgic-mmio-v3.c +2 -1
@@ -50,7 +50,8 @@ bool vgic_has_its(struct kvm *kvm)bool vgic_supports_direct_msis (struct kvm *kvm) { -return kvm_vgic_global_state.has_gicv4 &&vgic_has_its (kvm); +return (kvm_vgic_global_state.has_gicv4_1 || + (kvm_vgic_global_state.has_gicv4 &&vgic_has_its (kvm))); }/*
Signature · 1 line
2.26.2
Marc Zyngier
maintainerPATCH 4/4 KVM: arm64: Fix 32bit PC wrap-around
In reply to Marc ZyngierIn the unlikely event that a 32bit vcpu traps into the hypervisor on an instruction that is located right at the end of the 32bit range, the emulation of that instruction is going to increment PC past the 32bit range. This isn't great, as userspace can then observe this value and get a bit confused.
Conversly, userspace can do things like (in the context of a 64bit guest that is capable of 32bit EL0) setting PSTATE to AArch64-EL0, set PC to a 64bit value, change PSTATE to AArch32-USR, and observe that PC hasn't been truncated. More confusion.
Fix both by:
- truncating PC increments for 32bit guests
- sanitizing all 32bit regs every time a core reg is changed by userspace, and that PSTATE indicates a 32bit mode.
--- arch/arm64/kvm/guest.c | 7 +++++++ virt/kvm/arm/hyp/aarch32.c | 8 ++++++-- 2 files changed, 13 insertions(+), 2 deletions(-)
arch/arm64/kvm/guest.c +7 -0
@@ -200,6 +200,13 @@ static int set_core_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg) }memcpy ((u32 *)regs + off, valp,KVM_REG_SIZE (reg->id)); + +if (*vcpu_cpsr (vcpu) &PSR_MODE32_BIT ) { +int i; + +for (i =0 ; i <16 ; i++) + *vcpu_reg32 (vcpu, i) = (u32)*vcpu_reg32 (vcpu, i); + } out:return err; }
virt/kvm/arm/hyp/aarch32.c +6 -2
@@ -125,12 +125,16 @@ static void __hyp_text kvm_adjust_itstate(struct kvm_vcpu *vcpu)*/ void __hyp_text kvm_skip_instr32(struct kvm_vcpu *vcpu, bool is_wide_instr) { +u32 pc = *vcpu_pc (vcpu);bool is_thumb; is_thumb = !!(*vcpu_cpsr (vcpu) &PSR_AA32_T_BIT );if (is_thumb && !is_wide_instr) - *vcpu_pc (vcpu) +=2 ; + pc +=2 ;else - *vcpu_pc (vcpu) +=4 ; + pc +=4 ; + + *vcpu_pc (vcpu) = pc; +kvm_adjust_itstate (vcpu); }
Signature · 1 line
2.26.2
Sasha Levin
maintainerRe: [PATCH 4/4] KVM: arm64: Fix 32bit PC wrap-around
Hi
[This is an automated email]
This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all
The bot has tested the following trees: v5.6.8, v5.4.36, v4.19.119, v4.14.177, v4.9.220, v4.4.220.
v5.6.8: Build OK!
v5.4.36: Build OK!
v4.19.119: Build OK!
v4.14.177: Failed to apply! Possible dependencies:
00536ec47660 ("KVM: arm/arm64: Prepare to handle deferred save/restore of SPSR_EL1")
256c0960b7b6 ("kvm/arm: use PSR_AA32 definitions")
52f6c4f02164 ("KVM: arm64: Change 32-bit handling of VM system registers")
54ceb1bcf8d8 ("KVM: arm64: Move debug dirty flag calculation out of world switch")
623e1528d409 ("KVM: arm/arm64: Move cc/it checks under hyp's Makefile to avoid instrumentation")
74a64a981662 ("KVM: arm/arm64: Unify 32bit fault injection")
8d404c4c2461 ("KVM: arm64: Rewrite system register accessors to read/write functions")
a892819560c4 ("KVM: arm64: Prepare to handle deferred save/restore of 32-bit registers")
b220244d4179 ("arm64: vgic-v2: Fix proxying of cpuif access")v4.9.220: Failed to apply! Possible dependencies:
021234ef3752 ("KVM: arm64: Make kvm_condition_valid32() accessible from EL2")
256c0960b7b6 ("kvm/arm: use PSR_AA32 definitions")
3dbbdf78636e ("KVM: arm/arm64: Report PMU overflow interrupts to userspace irqchip")
52f6c4f02164 ("KVM: arm64: Change 32-bit handling of VM system registers")
54ceb1bcf8d8 ("KVM: arm64: Move debug dirty flag calculation out of world switch")
623e1528d409 ("KVM: arm/arm64: Move cc/it checks under hyp's Makefile to avoid instrumentation")
6c0070366dea ("arm64: KVM: PMU: Refactor pmu_*_el0_disabled")
74a64a981662 ("KVM: arm/arm64: Unify 32bit fault injection")
8d404c4c2461 ("KVM: arm64: Rewrite system register accessors to read/write functions")
9171fa2e0951 ("KVM: arm/arm64: Decouple kvm timer functions from virtual timer")
a5a1d1c2914b ("clocksource: Use a plain u64 instead of cycle_t")
b7484931e4a8 ("KVM: arm/arm64: PMU: remove request-less vcpu kick")
d9e139778376 ("KVM: arm/arm64: Support arch timers with a userspace gic")
d9f89b4e9290 ("KVM: arm/arm64: PMU: Fix overflow interrupt injection")
f85279b4bd48 ("arm64: KVM: Save/restore the host SPE state when entering/leaving a VM")v4.4.220: Failed to apply! Possible dependencies:
08dcbfda0774 ("ARM: KVM: Add a HYP-specific header file")
1d58d2cbf723 ("ARM: KVM: Add TLB invalidation code")
33280b4cd1dc ("ARM: KVM: Add banked registers save/restore")
3c29568768df ("ARM: KVM: Add system register accessor macros")
59cbcdb5d83b ("ARM: KVM: Add VFP save/restore")
623e1528d409 ("KVM: arm/arm64: Move cc/it checks under hyp's Makefile to avoid instrumentation") 68130cb5db09 ("ARM: KVM: Use common version of timer-sr.c")
b5fa5d3e628b ("ARM: KVM: Use common version of vgic-v2-sr.c")
c0c2cdbffef2 ("ARM: KVM: Add vgic v2 save/restore")
c7ce6c63a05f ("ARM: KVM: Add CP15 save/restore code")
e59bff9bf302 ("ARM: KVM: Add timer save/restore")
f1c9cad7c508 ("ARM: KVM: Move kvm/hyp/hyp.h to include/asm/kvm_hyp.h")
NOTE: The patch will not be queued to stable trees until it is upstream.
How should we proceed with this patch?
Signature · 2 lines
Thanks Sasha
Will Deacon
maintainerHi Marc, Paolo,
Quoted 8 lines · Marc Zyngier
This is the second batch of KVM/arm fixes for 5.7. A compilation fix, a GICv4.1 fix, plus a couple of sanity checks (
SP_EL0save/restore, and the sanitising of AArch32 registers).Note that the pull request I sent a week ago[1] is still valid, and that this new series is built on top of the previous one.
Please pull,
I don't see this queued up in the kvm tree, which appears to have been sitting dormant for 10 days. Consequently, there are fixes sitting in limbo and we /still/ don't have a sensible base for arm64/kvm patches targetting 5.8.
Paolo -- how can I help get this stuff moving again? I'm more than happy to send this lot up to Linus via arm64 if you're busy atm. Please just let me know.
Cheers,
Will
Paolo Bonzini
maintainerQuoted 8 lines · Will Deacon
I don't see this queued up in the kvm tree, which appears to have been sitting dormant for 10 days. Consequently, there are fixes sitting in limbo and we /still/ don't have a sensible base for arm64/kvm patches targetting 5.8.
Paolo -- how can I help get this stuff moving again? I'm more than happy to send this lot up to Linus via arm64 if you're busy atm. Please just let me know.
10 days is one week during which I could hardly work and the two adjacent weekends. So this is basically really bad timing in Marc's first pull request, that he couldn't have anticipated.
I have pulled both trees now, so you can base 5.8 development on kvm/master. It will get to Linus in a couple days.
Paolo
Will Deacon
maintainerQuoted 6 lines · Paolo Bonzini
Quoted 5 lines · Will Deacon
I don't see this queued up in the kvm tree, which appears to have been sitting dormant for 10 days. Consequently, there are fixes sitting in limbo and we /still/ don't have a sensible base for arm64/kvm patches targetting 5.8.
Quoted 7 lines
Quoted 3 lines
Paolo -- how can I help get this stuff moving again? I'm more than happy to send this lot up to Linus via arm64 if you're busy atm. Please just let me know.
10 days is one week during which I could hardly work and the two adjacent weekends. So this is basically really bad timing in Marc's first pull request, that he couldn't have anticipated.
Understood, and thanks for the quick reply. If possible, please just let us know in future as we can probably figure something out rather than having things sit in limbo.
Quoted 2 lines
I have pulled both trees now, so you can base 5.8 development on kvm/master. It will get to Linus in a couple days.
Thanks, Paolo!
Will
Paolo Bonzini
maintainerQuoted 7 lines · Will Deacon
Quoted 3 lines
10 days is one week during which I could hardly work and the two adjacent weekends. So this is basically really bad timing in Marc's first pull request, that he couldn't have anticipated.
Understood, and thanks for the quick reply. If possible, please just let us know in future as we can probably figure something out rather than having things sit in limbo.
Indeed, it was my fault. I got stuck in a "1: tomorrow should be better" / "no it was actually worse" / "goto 1b" loop.
Paolo