Re: [PATCH V14 10/10] arm/arm64: KVM: add guest SEA support

From: Christoffer Dall
Date: Tue Mar 28 2017 - 18:19:13 EST

Next message: Steven Rostedt: "[for-next][PATCH 2/7] ftrace: Move ftrace_init() to right after memory initialization"
Previous message: Steven Rostedt: "[for-next][PATCH 5/7] ftrace: Allow for function tracing to record init functions on boot up"
In reply to: Baicar, Tyler: "Re: [PATCH V14 10/10] arm/arm64: KVM: add guest SEA support"
Next in thread: Tyler Baicar: "[PATCH V14 04/10] arm64: exception: handle Synchronous External Abort"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue, Mar 28, 2017 at 03:33:17PM -0600, Baicar, Tyler wrote:
> Hello Christoffer,
>
>
> On 3/28/2017 2:26 PM, Christoffer Dall wrote:
> >On Tue, Mar 28, 2017 at 01:30:40PM -0600, Tyler Baicar wrote:
> >>Currently external aborts are unsupported by the guest abort
> >>handling. Add handling for SEAs so that the host kernel reports
> >>SEAs which occur in the guest kernel.
> >>
> >>When an SEA occurs in the guest kernel, the guest exits and is
> >>routed to kvm_handle_guest_abort(). Prior to this patch, a print
> >>message of an unsupported FSC would be printed and nothing else
> >>would happen. With this patch, the code gets routed to the APEI
> >>handling of SEAs in the host kernel to report the SEA information.
> >>
> >>Signed-off-by: Tyler Baicar <tbaicar@xxxxxxxxxxxxxx>
> >>Acked-by: Catalin Marinas <catalin.marinas@xxxxxxx>
> >>Acked-by: Marc Zyngier <marc.zyngier@xxxxxxx>
> >>---
> >> arch/arm/include/asm/kvm_arm.h | 10 ++++++++++
> >> arch/arm/include/asm/system_misc.h | 5 +++++
> >> arch/arm/kvm/mmu.c | 34 +++++++++++++++++++++++++++++++---
> >> arch/arm64/include/asm/kvm_arm.h | 10 ++++++++++
> >> arch/arm64/include/asm/system_misc.h | 2 ++
> >> arch/arm64/mm/fault.c | 23 +++++++++++++++++++++--
> >> drivers/acpi/apei/ghes.c | 13 +++++++------
> >> include/acpi/ghes.h | 2 +-
> >> 8 files changed, 87 insertions(+), 12 deletions(-)
> >>
> >>diff --git a/arch/arm/include/asm/kvm_arm.h b/arch/arm/include/asm/kvm_arm.h
> >>index a3f0b3d..ebf020b 100644
> >>--- a/arch/arm/include/asm/kvm_arm.h
> >>+++ b/arch/arm/include/asm/kvm_arm.h
> >>@@ -187,6 +187,16 @@
> >> #define FSC_FAULT (0x04)
> >> #define FSC_ACCESS (0x08)
> >> #define FSC_PERM (0x0c)
> >>+#define FSC_SEA (0x10)
> >>+#define FSC_SEA_TTW0 (0x14)
> >>+#define FSC_SEA_TTW1 (0x15)
> >>+#define FSC_SEA_TTW2 (0x16)
> >>+#define FSC_SEA_TTW3 (0x17)
> >>+#define FSC_SECC (0x18)
> >>+#define FSC_SECC_TTW0 (0x1c)
> >>+#define FSC_SECC_TTW1 (0x1d)
> >>+#define FSC_SECC_TTW2 (0x1e)
> >>+#define FSC_SECC_TTW3 (0x1f)
> >> /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
> >> #define HPFAR_MASK (~0xf)
> >>diff --git a/arch/arm/include/asm/system_misc.h b/arch/arm/include/asm/system_misc.h
> >>index a3d61ad..8c4a89f 100644
> >>--- a/arch/arm/include/asm/system_misc.h
> >>+++ b/arch/arm/include/asm/system_misc.h
> >>@@ -22,6 +22,11 @@
> >> extern unsigned int user_debug;
> >>+static inline int handle_guest_sea(phys_addr_t addr, unsigned int esr)
> >>+{
> >>+ return -1;
> >>+}
> >>+
> >> #endif /* !__ASSEMBLY__ */
> >> #endif /* __ASM_ARM_SYSTEM_MISC_H */
> >>diff --git a/arch/arm/kvm/mmu.c b/arch/arm/kvm/mmu.c
> >>index 962616f..9a977c8 100644
> >>--- a/arch/arm/kvm/mmu.c
> >>+++ b/arch/arm/kvm/mmu.c
> >>@@ -29,6 +29,7 @@
> >> #include <asm/kvm_asm.h>
> >> #include <asm/kvm_emulate.h>
> >> #include <asm/virt.h>
> >>+#include <asm/system_misc.h>
> >> #include "trace.h"
> >>@@ -1406,6 +1407,24 @@ static void handle_access_fault(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa)
> >> kvm_set_pfn_accessed(pfn);
> >> }
> >>+static bool is_abort_sea(unsigned long fault_status) {
> >>+ switch (fault_status) {
> >>+ case FSC_SEA:
> >>+ case FSC_SEA_TTW0:
> >>+ case FSC_SEA_TTW1:
> >>+ case FSC_SEA_TTW2:
> >>+ case FSC_SEA_TTW3:
> >>+ case FSC_SECC:
> >>+ case FSC_SECC_TTW0:
> >>+ case FSC_SECC_TTW1:
> >>+ case FSC_SECC_TTW2:
> >>+ case FSC_SECC_TTW3:
> >>+ return true;
> >>+ default:
> >>+ return false;
> >>+ }
> >>+}
> >>+
> >> /**
> >> * kvm_handle_guest_abort - handles all 2nd stage aborts
> >> * @vcpu: the VCPU pointer
> >>@@ -1428,19 +1447,28 @@ int kvm_handle_guest_abort(struct kvm_vcpu *vcpu, struct kvm_run *run)
> >> gfn_t gfn;
> >> int ret, idx;
> >>+ fault_status = kvm_vcpu_trap_get_fault_type(vcpu);
> >>+
> >>+ fault_ipa = kvm_vcpu_get_fault_ipa(vcpu);
> >>+
> >>+ /*
> >>+ * The host kernel will handle the synchronous external abort. There
> >>+ * is no need to pass the error into the guest.
> >>+ */
> >>+ if (is_abort_sea(fault_status))
> >>+ if (!handle_guest_sea(fault_ipa, kvm_vcpu_get_hsr(vcpu)))
> >>+ return 1;
> >>+
> >So what's the logic in presenting this as a vabt to the guest when the
> >host couldn't handle it? What do you do in other parts of the kernel if
> >you see an abort that the host cannot handle?
> The only cases that handle_guest_sea() will not return zero are
> there is no error to process
> (firmware support isn't there) or the error populated is invalid. If
> there is no error to process,
> we want to continue here with the handling that exists. If the error
> is invalid then GHES will
> report that and since the error wasn't handled properly we should
> continue with existing
> handling.
>
> If there is an abort that GHES fails to handle in the host then GHES
> just reports that and we
> also continue with the abort handling.

Hmmm, I don't see how the guest can possibly handle the error better
than the host in this situation, where the host has already decided that
there is either no error or we have no way to obtain information about
that error.

Regardless of existing behavior that seems weird, but I suppose one
could argue that something strange happened while running the guest and
it should be notified somehow.

Anyway, I may just be failing to come up with a scenario where it makes
sense.

> >
> >nit: I'd prefer to see braces around the the multi-line block above.
> I will add braces.
> >
> >
> >> is_iabt = kvm_vcpu_trap_is_iabt(vcpu);
> >> if (unlikely(!is_iabt && kvm_vcpu_dabt_isextabt(vcpu))) {
> >> kvm_inject_vabt(vcpu);
> >> return 1;
> >> }
> >>- fault_ipa = kvm_vcpu_get_fault_ipa(vcpu);
> >>-
> >> trace_kvm_guest_fault(*vcpu_pc(vcpu), kvm_vcpu_get_hsr(vcpu),
> >> kvm_vcpu_get_hfar(vcpu), fault_ipa);
> >> /* Check the stage-2 fault is trans. fault or write fault */
> >>- fault_status = kvm_vcpu_trap_get_fault_type(vcpu);
> >> if (fault_status != FSC_FAULT && fault_status != FSC_PERM &&
> >> fault_status != FSC_ACCESS) {
> >> kvm_err("Unsupported FSC: EC=%#x xFSC=%#lx ESR_EL2=%#lx\n",
> >>diff --git a/arch/arm64/include/asm/kvm_arm.h b/arch/arm64/include/asm/kvm_arm.h
> >>index 6e99978..61d694c 100644
> >>--- a/arch/arm64/include/asm/kvm_arm.h
> >>+++ b/arch/arm64/include/asm/kvm_arm.h
> >>@@ -204,6 +204,16 @@
> >> #define FSC_FAULT ESR_ELx_FSC_FAULT
> >> #define FSC_ACCESS ESR_ELx_FSC_ACCESS
> >> #define FSC_PERM ESR_ELx_FSC_PERM
> >>+#define FSC_SEA ESR_ELx_FSC_EXTABT
> >>+#define FSC_SEA_TTW0 (0x14)
> >>+#define FSC_SEA_TTW1 (0x15)
> >>+#define FSC_SEA_TTW2 (0x16)
> >>+#define FSC_SEA_TTW3 (0x17)
> >>+#define FSC_SECC (0x18)
> >>+#define FSC_SECC_TTW0 (0x1c)
> >>+#define FSC_SECC_TTW1 (0x1d)
> >>+#define FSC_SECC_TTW2 (0x1e)
> >>+#define FSC_SECC_TTW3 (0x1f)
> >> /* Hyp Prefetch Fault Address Register (HPFAR/HDFAR) */
> >> #define HPFAR_MASK (~UL(0xf))
> >>diff --git a/arch/arm64/include/asm/system_misc.h b/arch/arm64/include/asm/system_misc.h
> >>index bc81243..95aa442 100644
> >>--- a/arch/arm64/include/asm/system_misc.h
> >>+++ b/arch/arm64/include/asm/system_misc.h
> >>@@ -56,6 +56,8 @@ void hook_debug_fault_code(int nr, int (*fn)(unsigned long, unsigned int,
> >> __show_ratelimited; \
> >> })
> >>+int handle_guest_sea(phys_addr_t addr, unsigned int esr);
> >>+
> >> #endif /* __ASSEMBLY__ */
> >> #endif /* __ASM_SYSTEM_MISC_H */
> >>diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
> >>index f7372ce..3abb367 100644
> >>--- a/arch/arm64/mm/fault.c
> >>+++ b/arch/arm64/mm/fault.c
> >>@@ -497,6 +497,7 @@ static int do_bad(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> >> static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> >> {
> >> struct siginfo info;
> >>+ int ret = 0;
> >> pr_err("Synchronous External Abort: %s (0x%08x) at 0x%016lx\n",
> >> fault_name(esr), esr, addr);
> >>@@ -508,7 +509,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> >> */
> >> if (IS_ENABLED(CONFIG_ACPI_APEI_SEA)) {
> >> nmi_enter();
> >>- ghes_notify_sea();
> >>+ ret = ghes_notify_sea();
> >> nmi_exit();
> >> }
> >>@@ -521,7 +522,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> >> info.si_addr = (void __user *)addr;
> >> arm64_notify_die("", regs, &info, esr);
> >>- return 0;
> >>+ return ret;
> >> }
> >> static const struct fault_info {
> >>@@ -603,6 +604,24 @@ static const char *fault_name(unsigned int esr)
> >> }
> >> /*
> >>+ * Handle Synchronous External Aborts that occur in a guest kernel.
> >>+ *
> >>+ * The return value will be zero if the SEA was successfully handled
> >>+ * and non-zero if there was an error processing the error or there was
> >>+ * no error to process.
> >>+ */
> >>+int handle_guest_sea(phys_addr_t addr, unsigned int esr)
> >>+{
> >>+ int ret = -ENOENT;
> >>+
> >>+ if(IS_ENABLED(CONFIG_ACPI_APEI_SEA)) {
> >nit: missing white space
> I will add the space.
> >
> >>+ ret = ghes_notify_sea();
> >>+ }
> >>+
> >>+ return ret;
> >>+}
> >>+
> >>+/*
> >> * Dispatch a data abort to the relevant handler.
> >> */
> >> asmlinkage void __exception do_mem_abort(unsigned long addr, unsigned int esr,
> >>diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
> >>index 230b095..81eabc6 100644
> >>--- a/drivers/acpi/apei/ghes.c
> >>+++ b/drivers/acpi/apei/ghes.c
> >>@@ -810,17 +810,18 @@ static int ghes_notify_sci(struct notifier_block *this,
> >> #ifdef CONFIG_ACPI_APEI_SEA
> >> static LIST_HEAD(ghes_sea);
> >>-void ghes_notify_sea(void)
> >>+int ghes_notify_sea(void)
> >> {
> >> struct ghes *ghes;
> >>+ int ret = -ENOENT;
> >>- /*
> >>- * synchronize_rcu() will wait for nmi_exit(), so no need to
> >>- * rcu_read_lock().
> >>- */
> >>+ rcu_read_lock();
> >> list_for_each_entry_rcu(ghes, &ghes_sea, list) {
> >>- ghes_proc(ghes);
> >>+ if(!ghes_proc(ghes))
> >>+ ret = 0;
> >> }
> >>+ rcu_read_unlock();
> >>+ return ret;
> >> }
> >> static void ghes_sea_add(struct ghes *ghes)
> >>diff --git a/include/acpi/ghes.h b/include/acpi/ghes.h
> >>index 18bc935..2a727dc 100644
> >>--- a/include/acpi/ghes.h
> >>+++ b/include/acpi/ghes.h
> >>@@ -99,6 +99,6 @@ static inline void *acpi_hest_generic_data_payload(struct acpi_hest_generic_data
> >> gdata + 1;
> >> }
> >>-void ghes_notify_sea(void);
> >>+int ghes_notify_sea(void);
> >> #endif /* GHES_H */
> >>--
> >I feel like the changes to acpi/arm64 in this patch could have been a
> >separate patch or folded into previous patches, but for the KVM part of
> >this patch:
> >
> >Acked-by: Christoffer Dall <cdall@xxxxxxxxxx>
> Thank you. I can break out the GHES changes if that is really desired.
>
That's not for me to say, so if Catalin is happy with the current split,
you don't have to respin just for that or for the nits I pointed out on
account of me.

Thanks,
-Christoffer

Next message: Steven Rostedt: "[for-next][PATCH 2/7] ftrace: Move ftrace_init() to right after memory initialization"
Previous message: Steven Rostedt: "[for-next][PATCH 5/7] ftrace: Allow for function tracing to record init functions on boot up"
In reply to: Baicar, Tyler: "Re: [PATCH V14 10/10] arm/arm64: KVM: add guest SEA support"
Next in thread: Tyler Baicar: "[PATCH V14 04/10] arm64: exception: handle Synchronous External Abort"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]