powerpc/hmi: Fix kernel hang when TB is in error state.

On TOD/TB errors timebase register stops/freezes until HMI error recovery
gets TOD/TB back into running state. On successful recovery, TB starts
running again and udelay() that relies on TB value continues to function
properly. But in case when HMI fails to recover from TOD/TB errors, the
TB register stay freezed. With TB not running the __delay() function
keeps looping and never return. If __delay() is called while in panic
path then system hangs and never reboots after panic.

Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

diff --git a/arch/powerpc/include/asm/opal-api.h b/arch/powerpc/include/asm/opal-api.h
index e1d118ac61dc8ddbc6204f1757683f4a95264d35..234fde15b37cde5a0dfb9ef84b0b46d10f18e04c 100644 (file)
--- a/arch/powerpc/include/asm/opal-api.h
+++ b/arch/powerpc/include/asm/opal-api.h
@@ -209,6 +209,7 @@
 #define OPAL_SENSOR_GROUP_ENABLE               163
 #define OPAL_PCI_GET_PBCQ_TUNNEL_BAR           164
 #define OPAL_PCI_SET_PBCQ_TUNNEL_BAR           165
+#define OPAL_HANDLE_HMI2                       166
 #define        OPAL_NX_COPROC_INIT                     167
 #define OPAL_XIVE_GET_VP_STATE                 170
 #define OPAL_LAST                              170
@@ -635,6 +636,15 @@ struct OpalHMIEvent {
        } u;
 };
 
+/* OPAL_HANDLE_HMI2 out_flags */
+enum {
+       OPAL_HMI_FLAGS_TB_RESYNC        = (1ull << 0), /* Timebase has been resynced */
+       OPAL_HMI_FLAGS_DEC_LOST         = (1ull << 1), /* DEC lost, needs to be reprogrammed */
+       OPAL_HMI_FLAGS_HDEC_LOST        = (1ull << 2), /* HDEC lost, needs to be reprogrammed */
+       OPAL_HMI_FLAGS_TOD_TB_FAIL      = (1ull << 3), /* TOD/TB recovery failed. */
+       OPAL_HMI_FLAGS_NEW_EVENT        = (1ull << 63), /* An event has been created */
+};
+
 enum {
        OPAL_P7IOC_DIAG_TYPE_NONE       = 0,
        OPAL_P7IOC_DIAG_TYPE_RGC        = 1,

diff --git a/arch/powerpc/include/asm/opal.h b/arch/powerpc/include/asm/opal.h
index 4e978d4dea5ce446e6a359d0b4e9127d98e746c9..4cc37e708bc732285c510da2cb71d415b120499e 100644 (file)
--- a/arch/powerpc/include/asm/opal.h
+++ b/arch/powerpc/include/asm/opal.h
@@ -203,6 +203,7 @@ int64_t opal_set_param(uint64_t token, uint32_t param_id, uint64_t buffer,
 int64_t opal_sensor_read(uint32_t sensor_hndl, int token, __be32 *sensor_data);
 int64_t opal_sensor_read_u64(u32 sensor_hndl, int token, __be64 *sensor_data);
 int64_t opal_handle_hmi(void);
+int64_t opal_handle_hmi2(__be64 *out_flags);
 int64_t opal_register_dump_region(uint32_t id, uint64_t start, uint64_t end);
 int64_t opal_unregister_dump_region(uint32_t id);
 int64_t opal_slw_set_reg(uint64_t cpu_pir, uint64_t sprn, uint64_t val);
@@ -359,6 +360,7 @@ int opal_power_control_init(void);
 extern int opal_machine_check(struct pt_regs *regs);
 extern bool opal_mce_check_early_recovery(struct pt_regs *regs);
 extern int opal_hmi_exception_early(struct pt_regs *regs);
+extern int opal_hmi_exception_early2(struct pt_regs *regs);
 extern int opal_handle_hmi_exception(struct pt_regs *regs);
 
 extern void opal_shutdown(void);

diff --git a/arch/powerpc/include/asm/time.h b/arch/powerpc/include/asm/time.h
index 54bf7e68a7e190228f5268356c541bac8f16c726..57e968413d1e0863782e30cd29be75422fcaae0e 100644 (file)
--- a/arch/powerpc/include/asm/time.h
+++ b/arch/powerpc/include/asm/time.h
@@ -36,6 +36,8 @@ extern unsigned long ppc_proc_freq;
 extern unsigned long ppc_tb_freq;
 #define DEFAULT_TB_FREQ                125000000UL
 
+extern bool tb_invalid;
+
 struct div_result {
        u64 result_high;
        u64 result_low;

diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c
index 6ef32472ee1d9d792e221f5f2fb54264e460a03f..325d60633dfa44b69defadfcfbc4461d210f7ada 100644 (file)
--- a/arch/powerpc/kernel/time.c
+++ b/arch/powerpc/kernel/time.c
@@ -150,6 +150,8 @@ EXPORT_SYMBOL_GPL(ppc_proc_freq);
 unsigned long ppc_tb_freq;
 EXPORT_SYMBOL_GPL(ppc_tb_freq);
 
+bool tb_invalid;
+
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
 /*
  * Factor for converting from cputime_t (timebase ticks) to
@@ -459,6 +461,13 @@ void __delay(unsigned long loops)
                                diff += 1000000000;
                        spin_cpu_relax();
                } while (diff < loops);
+       } else if (tb_invalid) {
+               /*
+                * TB is in error state and isn't ticking anymore.
+                * HMI handler was unable to recover from TB error.
+                * Return immediately, so that kernel won't get stuck here.
+                */
+               spin_cpu_relax();
        } else {
                start = get_tbl();
                while (get_tbl() - start < loops)

diff --git a/arch/powerpc/platforms/powernv/opal-call.c b/arch/powerpc/platforms/powernv/opal-call.c
index 7cba0d5da3ff778b75e659d3480a5a337dde14b7..36c8fa3647a2f368a7ee205c936f28f919b4167d 100644 (file)
--- a/arch/powerpc/platforms/powernv/opal-call.c
+++ b/arch/powerpc/platforms/powernv/opal-call.c
@@ -220,6 +220,7 @@ OPAL_CALL(opal_sensor_read,                 OPAL_SENSOR_READ);
 OPAL_CALL(opal_get_param,                      OPAL_GET_PARAM);
 OPAL_CALL(opal_set_param,                      OPAL_SET_PARAM);
 OPAL_CALL(opal_handle_hmi,                     OPAL_HANDLE_HMI);
+OPAL_CALL(opal_handle_hmi2,                    OPAL_HANDLE_HMI2);
 OPAL_CALL(opal_config_cpu_idle_state,          OPAL_CONFIG_CPU_IDLE_STATE);
 OPAL_CALL(opal_slw_set_reg,                    OPAL_SLW_SET_REG);
 OPAL_CALL(opal_register_dump_region,           OPAL_REGISTER_DUMP_REGION);

diff --git a/arch/powerpc/platforms/powernv/opal.c b/arch/powerpc/platforms/powernv/opal.c
index 737c51d634801d3ed168897cf90729cd98eeec72..f2b063b027f0adb5c5064ee133f2ab355f788797 100644 (file)
--- a/arch/powerpc/platforms/powernv/opal.c
+++ b/arch/powerpc/platforms/powernv/opal.c
@@ -614,6 +614,27 @@ int opal_hmi_exception_early(struct pt_regs *regs)
        return 0;
 }
 
+int opal_hmi_exception_early2(struct pt_regs *regs)
+{
+       s64 rc;
+       __be64 out_flags;
+
+       /*
+        * call opal hmi handler.
+        * Check 64-bit flag mask to find out if an event was generated,
+        * and whether TB is still valid or not etc.
+        */
+       rc = opal_handle_hmi2(&out_flags);
+       if (rc != OPAL_SUCCESS)
+               return 0;
+
+       if (be64_to_cpu(out_flags) & OPAL_HMI_FLAGS_NEW_EVENT)
+               local_paca->hmi_event_available = 1;
+       if (be64_to_cpu(out_flags) & OPAL_HMI_FLAGS_TOD_TB_FAIL)
+               tb_invalid = true;
+       return 1;
+}
+
 /* HMI exception handler called in virtual mode during check_irq_replay. */
 int opal_handle_hmi_exception(struct pt_regs *regs)
 {

diff --git a/arch/powerpc/platforms/powernv/setup.c b/arch/powerpc/platforms/powernv/setup.c
index 14befee4b3f1eae867947e9282c7b064cc1974a8..3cf40f689aace855ed1d00c4e61fedf8b045db0a 100644 (file)
--- a/arch/powerpc/platforms/powernv/setup.c
+++ b/arch/powerpc/platforms/powernv/setup.c
@@ -401,7 +401,10 @@ static void __init pnv_setup_machdep_opal(void)
        /* ppc_md.system_reset_exception gets filled in by pnv_smp_init() */
        ppc_md.machine_check_exception = opal_machine_check;
        ppc_md.mce_check_early_recovery = opal_mce_check_early_recovery;
-       ppc_md.hmi_exception_early = opal_hmi_exception_early;
+       if (opal_check_token(OPAL_HANDLE_HMI2))
+               ppc_md.hmi_exception_early = opal_hmi_exception_early2;
+       else
+               ppc_md.hmi_exception_early = opal_hmi_exception_early;
        ppc_md.handle_hmi_exception = opal_handle_hmi_exception;
 }