From: Arnd Bergmann <arnd@arndb.de> Sent: Monday, August 24, 2020 11:34 AM

On Mon, Aug 24, 2020 at 6:48 PM Michael Kelley [off-list ref] wrote:

quoted

 /*
+ * This function is invoked via the ACPI clocksource probe mechanism. We
+ * don't actually use any values from the ACPI GTDT table, but we set up
+ * the Hyper-V synthetic clocksource and do other initialization for
+ * interacting with Hyper-V the first time.  Using early_initcall to invoke
+ * this function is too late because interrupts are already enabled at that
+ * point, and hv_init_clocksource() must run before interrupts are enabled.
+ *
+ * 1. Setup the guest ID.
+ * 2. Get features and hints info from Hyper-V
+ * 3. Setup per-cpu VP indices.
+ * 4. Initialize the Hyper-V clocksource.
+ */
+
+static int __init hyperv_init(struct acpi_table_header *table)
+{
+       struct hv_get_vp_registers_output result;
+       u32     a, b, c, d;
+       u64     guest_id;
+       int     i, cpuhp;
+
+       /*
+        * If we're in a VM on Hyper-V, the ACPI hypervisor_id field will
+        * have the string "MsHyperV".
+        */
+       if (strncmp((char *)&acpi_gbl_FADT.hypervisor_id, "MsHyperV", 8))
+               return -EINVAL;
+
+       /* Setup the guest ID */
+       guest_id = generate_guest_id(0, LINUX_VERSION_CODE, 0);
+       hv_set_vpreg(HV_REGISTER_GUEST_OSID, guest_id);
+
+       /* Get the features and hints from Hyper-V */
+       hv_get_vpreg_128(HV_REGISTER_FEATURES, &result);
+       ms_hyperv.features = result.as32.a;
+       ms_hyperv.misc_features = result.as32.c;
+
+       hv_get_vpreg_128(HV_REGISTER_ENLIGHTENMENTS, &result);
+       ms_hyperv.hints = result.as32.a;
+
+       pr_info("Hyper-V: Features 0x%x, hints 0x%x, misc 0x%x\n",
+               ms_hyperv.features, ms_hyperv.hints, ms_hyperv.misc_features);
+
+       /*
+        * If Hyper-V has crash notifications, set crash_kexec_post_notifiers
+        * so that we will report the panic to Hyper-V before running kdump.
+        */
+       if (ms_hyperv.misc_features & HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE)
+               crash_kexec_post_notifiers = true;
+
+       /* Get information about the Hyper-V host version */
+       hv_get_vpreg_128(HV_REGISTER_HYPERVISOR_VERSION, &result);
+       a = result.as32.a;
+       b = result.as32.b;
+       c = result.as32.c;
+       d = result.as32.d;
+       pr_info("Hyper-V: Host Build %d.%d.%d.%d-%d-%d\n",
+               b >> 16, b & 0xFFFF, a, d & 0xFFFFFF, c, d >> 24);
+
+       /* Allocate and initialize percpu VP index array */
+       hv_vp_index = kmalloc_array(num_possible_cpus(), sizeof(*hv_vp_index),
+                                   GFP_KERNEL);
+       if (!hv_vp_index)
+               return -ENOMEM;
+
+       for (i = 0; i < num_possible_cpus(); i++)
+               hv_vp_index[i] = VP_INVAL;
+
+       cpuhp = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN,
+                       "arm64/hyperv_init:online", hv_cpu_init, NULL);
+       if (cpuhp < 0)
+               goto free_vp_index;
+
+       hv_init_clocksource();
+       if (hv_stimer_alloc())
+               goto remove_cpuhp_state;
+
+       hyperv_initialized = true;
+       return 0;
+
+remove_cpuhp_state:
+       cpuhp_remove_state(cpuhp);
+free_vp_index:
+       kfree(hv_vp_index);
+       hv_vp_index = NULL;
+       return -EINVAL;
+}
+TIMER_ACPI_DECLARE(hyperv, ACPI_SIG_GTDT, hyperv_init);

I think this has come up before, and I still don't consider it an acceptable
hack to hook platform initialization code into the timer code.

Please split out the timer into a standalone driver in drivers/clocksource
that can get reviewed by the clocksource maintainers.

I see two related topics here.  First, the Hyper-V clocksource driver is
drivers/clocksource/hyperv_timer.c.  The code is architecture independent
and is used today on the x86 side and for ARM64 in this patch series.  A few
architecture specific calls are satisfied by code under arch/x86, and in this
patch series, under arch/arm64.  Is there some aspect of this driver that
needs reconsideration?  I just want to make sure to understand what you
are getting at.

Second is the question of where/how to do Hyper-V specific initialization.
I agree that hanging it off the timer initialization isn't a great approach.
Should I add a Hyper-V specific initialization call at the appropriate point
in the ARM64 init sequence?  The x86 side has some structure for handling
multiple hypervisors, and the Hyper-V initialization code naturally plugs into
that structure.  I'm certainly open to suggestions on the best way to handle
it for ARM64.

Michael

      Arnd