4.9-stable review patch. If anyone has any objections, please let me know. ------------------ From: Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx> commit 84d582d236dc1f9085e741affc72e9ba061a67c2 upstream. Recent discussion (http://marc.info/?l=xen-devel&m=149192184523741) established that commit 72a9b186292d ("xen: Remove event channel notification through Xen PCI platform device") (and thus commit da72ff5bfcb0 ("partially revert "xen: Remove event channel notification through Xen PCI platform device"")) are unnecessary and, in fact, prevent HVM guests from booting on Xen releases prior to 4.0 Therefore we revert both of those commits. The summary of that discussion is below: Here is the brief summary of the current situation: Before the offending commit (72a9b186292): 1) INTx does not work because of the reset_watches path. 2) The reset_watches path is only taken if you have Xen > 4.0 3) The Linux Kernel by default will use vector inject if the hypervisor support. So even INTx does not work no body running the kernel with Xen > 4.0 would notice. Unless he explicitly disabled this feature either in the kernel or in Xen (and this can only be disabled by modifying the code, not user-supported way to do it). After the offending commit (+ partial revert): 1) INTx is no longer support for HVM (only for PV guests). 2) Any HVM guest The kernel will not boot on Xen < 4.0 which does not have vector injection support. Since the only other mode supported is INTx which. So based on this summary, I think before commit (72a9b186292) we were in much better position from a user point of view. Signed-off-by: Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx> Reviewed-by: Juergen Gross <jgross@xxxxxxxx> Cc: Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx> Cc: Ingo Molnar <mingo@xxxxxxxxxx> Cc: "H. Peter Anvin" <hpa@xxxxxxxxx> Cc: x86@xxxxxxxxxx Cc: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx> Cc: Bjorn Helgaas <bhelgaas@xxxxxxxxxx> Cc: Stefano Stabellini <sstabellini@xxxxxxxxxx> Cc: Julien Grall <julien.grall@xxxxxxx> Cc: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx> Cc: Paul Gortmaker <paul.gortmaker@xxxxxxxxxxxxx> Cc: Ross Lagerwall <ross.lagerwall@xxxxxxxxxx> Cc: xen-devel@xxxxxxxxxxxxxxxxxxxx Cc: linux-kernel@xxxxxxxxxxxxxxx Cc: linux-pci@xxxxxxxxxxxxxxx Cc: Anthony Liguori <aliguori@xxxxxxxxxx> Cc: KarimAllah Ahmed <karahmed@xxxxxxxxx> Signed-off-by: Juergen Gross <jgross@xxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- arch/x86/include/asm/xen/events.h | 11 ++++++ arch/x86/pci/xen.c | 2 - arch/x86/xen/enlighten.c | 21 ++++++++---- arch/x86/xen/smp.c | 2 + arch/x86/xen/time.c | 5 ++ drivers/xen/events/events_base.c | 26 ++++++++++----- drivers/xen/platform-pci.c | 64 ++++++++++++++++++++++++++++++++++++++ include/xen/xen.h | 3 + 8 files changed, 117 insertions(+), 17 deletions(-) --- a/arch/x86/include/asm/xen/events.h +++ b/arch/x86/include/asm/xen/events.h @@ -20,4 +20,15 @@ static inline int xen_irqs_disabled(stru /* No need for a barrier -- XCHG is a barrier on x86. */ #define xchg_xen_ulong(ptr, val) xchg((ptr), (val)) +extern int xen_have_vector_callback; + +/* + * Events delivered via platform PCI interrupts are always + * routed to vcpu 0 and hence cannot be rebound. + */ +static inline bool xen_support_evtchn_rebind(void) +{ + return (!xen_hvm_domain() || xen_have_vector_callback); +} + #endif /* _ASM_X86_XEN_EVENTS_H */ --- a/arch/x86/pci/xen.c +++ b/arch/x86/pci/xen.c @@ -447,7 +447,7 @@ void __init xen_msi_init(void) int __init pci_xen_hvm_init(void) { - if (!xen_feature(XENFEAT_hvm_pirqs)) + if (!xen_have_vector_callback || !xen_feature(XENFEAT_hvm_pirqs)) return 0; #ifdef CONFIG_ACPI --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -137,6 +137,8 @@ struct shared_info xen_dummy_shared_info void *xen_initial_gdt; RESERVE_BRK(shared_info_page_brk, PAGE_SIZE); +__read_mostly int xen_have_vector_callback; +EXPORT_SYMBOL_GPL(xen_have_vector_callback); static int xen_cpu_up_prepare(unsigned int cpu); static int xen_cpu_up_online(unsigned int cpu); @@ -1521,7 +1523,10 @@ static void __init xen_pvh_early_guest_i if (!xen_feature(XENFEAT_auto_translated_physmap)) return; - BUG_ON(!xen_feature(XENFEAT_hvm_callback_vector)); + if (!xen_feature(XENFEAT_hvm_callback_vector)) + return; + + xen_have_vector_callback = 1; xen_pvh_early_cpu_init(0, false); xen_pvh_set_cr_flags(0); @@ -1860,7 +1865,9 @@ static int xen_cpu_up_prepare(unsigned i xen_vcpu_setup(cpu); } - if (xen_pv_domain() || xen_feature(XENFEAT_hvm_safe_pvclock)) + if (xen_pv_domain() || + (xen_have_vector_callback && + xen_feature(XENFEAT_hvm_safe_pvclock))) xen_setup_timer(cpu); rc = xen_smp_intr_init(cpu); @@ -1876,7 +1883,9 @@ static int xen_cpu_dead(unsigned int cpu { xen_smp_intr_free(cpu); - if (xen_pv_domain() || xen_feature(XENFEAT_hvm_safe_pvclock)) + if (xen_pv_domain() || + (xen_have_vector_callback && + xen_feature(XENFEAT_hvm_safe_pvclock))) xen_teardown_timer(cpu); return 0; @@ -1915,8 +1924,8 @@ static void __init xen_hvm_guest_init(vo xen_panic_handler_init(); - BUG_ON(!xen_feature(XENFEAT_hvm_callback_vector)); - + if (xen_feature(XENFEAT_hvm_callback_vector)) + xen_have_vector_callback = 1; xen_hvm_smp_init(); WARN_ON(xen_cpuhp_setup()); xen_unplug_emulated_devices(); @@ -1954,7 +1963,7 @@ bool xen_hvm_need_lapic(void) return false; if (!xen_hvm_domain()) return false; - if (xen_feature(XENFEAT_hvm_pirqs)) + if (xen_feature(XENFEAT_hvm_pirqs) && xen_have_vector_callback) return false; return true; } --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -765,6 +765,8 @@ static void __init xen_hvm_smp_prepare_c void __init xen_hvm_smp_init(void) { + if (!xen_have_vector_callback) + return; smp_ops.smp_prepare_cpus = xen_hvm_smp_prepare_cpus; smp_ops.smp_send_reschedule = xen_smp_send_reschedule; smp_ops.cpu_die = xen_cpu_die; --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -432,6 +432,11 @@ static void xen_hvm_setup_cpu_clockevent void __init xen_hvm_init_time_ops(void) { + /* vector callback is needed otherwise we cannot receive interrupts + * on cpu > 0 and at this point we don't know how many cpus are + * available */ + if (!xen_have_vector_callback) + return; if (!xen_feature(XENFEAT_hvm_safe_pvclock)) { printk(KERN_INFO "Xen doesn't support pvclock on HVM," "disable pv timer\n"); --- a/drivers/xen/events/events_base.c +++ b/drivers/xen/events/events_base.c @@ -1314,6 +1314,9 @@ static int rebind_irq_to_cpu(unsigned ir if (!VALID_EVTCHN(evtchn)) return -1; + if (!xen_support_evtchn_rebind()) + return -1; + /* Send future instances of this interrupt to other vcpu. */ bind_vcpu.port = evtchn; bind_vcpu.vcpu = xen_vcpu_nr(tcpu); @@ -1647,15 +1650,20 @@ void xen_callback_vector(void) { int rc; uint64_t callback_via; - - callback_via = HVM_CALLBACK_VECTOR(HYPERVISOR_CALLBACK_VECTOR); - rc = xen_set_callback_via(callback_via); - BUG_ON(rc); - pr_info("Xen HVM callback vector for event delivery is enabled\n"); - /* in the restore case the vector has already been allocated */ - if (!test_bit(HYPERVISOR_CALLBACK_VECTOR, used_vectors)) - alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR, - xen_hvm_callback_vector); + if (xen_have_vector_callback) { + callback_via = HVM_CALLBACK_VECTOR(HYPERVISOR_CALLBACK_VECTOR); + rc = xen_set_callback_via(callback_via); + if (rc) { + pr_err("Request for Xen HVM callback vector failed\n"); + xen_have_vector_callback = 0; + return; + } + pr_info("Xen HVM callback vector for event delivery is enabled\n"); + /* in the restore case the vector has already been allocated */ + if (!test_bit(HYPERVISOR_CALLBACK_VECTOR, used_vectors)) + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR, + xen_hvm_callback_vector); + } } #else void xen_callback_vector(void) {} --- a/drivers/xen/platform-pci.c +++ b/drivers/xen/platform-pci.c @@ -42,6 +42,7 @@ static unsigned long platform_mmio; static unsigned long platform_mmio_alloc; static unsigned long platform_mmiolen; +static uint64_t callback_via; static unsigned long alloc_xen_mmio(unsigned long len) { @@ -54,6 +55,51 @@ static unsigned long alloc_xen_mmio(unsi return addr; } +static uint64_t get_callback_via(struct pci_dev *pdev) +{ + u8 pin; + int irq; + + irq = pdev->irq; + if (irq < 16) + return irq; /* ISA IRQ */ + + pin = pdev->pin; + + /* We don't know the GSI. Specify the PCI INTx line instead. */ + return ((uint64_t)0x01 << 56) | /* PCI INTx identifier */ + ((uint64_t)pci_domain_nr(pdev->bus) << 32) | + ((uint64_t)pdev->bus->number << 16) | + ((uint64_t)(pdev->devfn & 0xff) << 8) | + ((uint64_t)(pin - 1) & 3); +} + +static irqreturn_t do_hvm_evtchn_intr(int irq, void *dev_id) +{ + xen_hvm_evtchn_do_upcall(); + return IRQ_HANDLED; +} + +static int xen_allocate_irq(struct pci_dev *pdev) +{ + return request_irq(pdev->irq, do_hvm_evtchn_intr, + IRQF_NOBALANCING | IRQF_TRIGGER_RISING, + "xen-platform-pci", pdev); +} + +static int platform_pci_resume(struct pci_dev *pdev) +{ + int err; + if (xen_have_vector_callback) + return 0; + err = xen_set_callback_via(callback_via); + if (err) { + dev_err(&pdev->dev, "platform_pci_resume failure!\n"); + return err; + } + return 0; +} + static int platform_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) { @@ -92,6 +138,21 @@ static int platform_pci_probe(struct pci platform_mmio = mmio_addr; platform_mmiolen = mmio_len; + if (!xen_have_vector_callback) { + ret = xen_allocate_irq(pdev); + if (ret) { + dev_warn(&pdev->dev, "request_irq failed err=%d\n", ret); + goto out; + } + callback_via = get_callback_via(pdev); + ret = xen_set_callback_via(callback_via); + if (ret) { + dev_warn(&pdev->dev, "Unable to set the evtchn callback " + "err=%d\n", ret); + goto out; + } + } + max_nr_gframes = gnttab_max_grant_frames(); grant_frames = alloc_xen_mmio(PAGE_SIZE * max_nr_gframes); ret = gnttab_setup_auto_xlat_frames(grant_frames); @@ -123,6 +184,9 @@ static struct pci_driver platform_driver .name = DRV_NAME, .probe = platform_pci_probe, .id_table = platform_pci_tbl, +#ifdef CONFIG_PM + .resume_early = platform_pci_resume, +#endif }; static int __init platform_pci_init(void) --- a/include/xen/xen.h +++ b/include/xen/xen.h @@ -38,7 +38,8 @@ extern enum xen_domain_type xen_domain_t */ #include <xen/features.h> #define xen_pvh_domain() (xen_pv_domain() && \ - xen_feature(XENFEAT_auto_translated_physmap)) + xen_feature(XENFEAT_auto_translated_physmap) && \ + xen_have_vector_callback) #else #define xen_pvh_domain() (0) #endif