Skip to content

Commit

Permalink
opal/hmi: Add a new opal_handle_hmi2 that returns direct info to Linux
Browse files Browse the repository at this point in the history 
It returns a 64-bit flags mask currently set to provide info
about which timer facilities were lost, and whether an event
was generated.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Stewart Smith <stewart@linux.ibm.com>
  • Loading branch information
@ozbenh @stewartsmith
ozbenh authored and stewartsmith committed Apr 17, 2018
1 parent a9d92e2 commit 88abbe2
Show file tree
Hide file tree
Showing 3 changed files with 90 additions and 46 deletions.
127 changes: 82 additions & 45 deletions core/hmi.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -291,7 +291,7 @@ static int setup_scom_addresses(void)
return 0;
}

static int queue_hmi_event(struct OpalHMIEvent *hmi_evt, int recover)
static int queue_hmi_event(struct OpalHMIEvent *hmi_evt, int recover, uint64_t *out_flags)
{
size_t num_params;

Expand All @@ -314,6 +314,8 @@ static int queue_hmi_event(struct OpalHMIEvent *hmi_evt, int recover)
*/
num_params = ALIGN_UP(sizeof(*hmi_evt), sizeof(u64)) / sizeof(u64);

*out_flags |= OPAL_HMI_FLAGS_NEW_EVENT;

/* queue up for delivery to host. */
return _opal_queue_msg(OPAL_MSG_HMI_EVT, NULL, NULL,
num_params, (uint64_t *)hmi_evt);
Expand Down Expand Up @@ -409,7 +411,7 @@ static bool decode_core_fir(struct cpu_thread *cpu,
}

static void find_core_checkstop_reason(struct OpalHMIEvent *hmi_evt,
bool *event_generated)
uint64_t *out_flags)
{
struct cpu_thread *cpu;

Expand All @@ -435,16 +437,14 @@ static void find_core_checkstop_reason(struct OpalHMIEvent *hmi_evt,
hmi_evt->u.xstop_error.xstop_reason = 0;
hmi_evt->u.xstop_error.u.pir = cpu->pir;

if (decode_core_fir(cpu, hmi_evt)) {
queue_hmi_event(hmi_evt, 0);
*event_generated = 1;
}
if (decode_core_fir(cpu, hmi_evt))
queue_hmi_event(hmi_evt, 0, out_flags);
}
}

static void find_capp_checkstop_reason(int flat_chip_id,
struct OpalHMIEvent *hmi_evt,
bool *event_generated)
uint64_t *out_flags)
{
struct capp_info info;
struct phb *phb;
Expand Down Expand Up @@ -496,8 +496,7 @@ static void find_capp_checkstop_reason(int flat_chip_id,

hmi_evt->severity = OpalHMI_SEV_NO_ERROR;
hmi_evt->type = OpalHMI_ERROR_CAPP_RECOVERY;
queue_hmi_event(hmi_evt, 1);
*event_generated = true;
queue_hmi_event(hmi_evt, 1, out_flags);

return;
}
Expand All @@ -506,7 +505,7 @@ static void find_capp_checkstop_reason(int flat_chip_id,

static void find_nx_checkstop_reason(int flat_chip_id,
struct OpalHMIEvent *hmi_evt,
bool *event_generated)
uint64_t *out_flags)
{
uint64_t nx_status;
uint64_t nx_dma_fir;
Expand Down Expand Up @@ -564,8 +563,7 @@ static void find_nx_checkstop_reason(int flat_chip_id,
xscom_write(flat_chip_id, nx_dma_engine_fir, PPC_BIT(38));

/* Send an HMI event. */
queue_hmi_event(hmi_evt, 0);
*event_generated = true;
queue_hmi_event(hmi_evt, 0, out_flags);
}

/*
Expand Down Expand Up @@ -623,7 +621,7 @@ static void dump_scoms(int flat_chip_id, const char *unit, uint32_t *scoms)

static void find_npu2_checkstop_reason(int flat_chip_id,
struct OpalHMIEvent *hmi_evt,
bool *event_generated)
uint64_t *out_flags)
{
struct phb *phb;
struct npu *p = NULL;
Expand Down Expand Up @@ -714,13 +712,12 @@ static void find_npu2_checkstop_reason(int flat_chip_id,
hmi_evt->u.xstop_error.u.chip_id = flat_chip_id;

/* Marking the event as recoverable so that we don't crash */
queue_hmi_event(hmi_evt, 1);
*event_generated = true;
queue_hmi_event(hmi_evt, 1, out_flags);
}

static void find_npu_checkstop_reason(int flat_chip_id,
struct OpalHMIEvent *hmi_evt,
bool *event_generated)
uint64_t *out_flags)
{
struct phb *phb;
struct npu *p = NULL;
Expand All @@ -733,7 +730,7 @@ static void find_npu_checkstop_reason(int flat_chip_id,

/* Only check for NPU errors if the chip has a NPU */
if (PVR_TYPE(mfspr(SPR_PVR)) != PVR_TYPE_P8NVL)
return find_npu2_checkstop_reason(flat_chip_id, hmi_evt, event_generated);
return find_npu2_checkstop_reason(flat_chip_id, hmi_evt, out_flags);

/* Find the NPU on the chip associated with the HMI. */
for_each_phb(phb) {
Expand Down Expand Up @@ -783,22 +780,22 @@ static void find_npu_checkstop_reason(int flat_chip_id,
hmi_evt->u.xstop_error.u.chip_id = flat_chip_id;

/* The HMI is "recoverable" because it shouldn't crash the system */
queue_hmi_event(hmi_evt, 1);
*event_generated = true;
queue_hmi_event(hmi_evt, 1, out_flags);
}

static void decode_malfunction(struct OpalHMIEvent *hmi_evt)
static void decode_malfunction(struct OpalHMIEvent *hmi_evt, uint64_t *out_flags)
{
int i;
uint64_t malf_alert;
bool event_generated = false;
uint64_t malf_alert, flags;

flags = 0;

if (!setup_scom_addresses()) {
prerror("Failed to setup scom addresses\n");
/* Send an unknown HMI event. */
hmi_evt->u.xstop_error.xstop_type = CHECKSTOP_TYPE_UNKNOWN;
hmi_evt->u.xstop_error.xstop_reason = 0;
queue_hmi_event(hmi_evt, false);
queue_hmi_event(hmi_evt, false, out_flags);
return;
}

Expand All @@ -811,22 +808,23 @@ static void decode_malfunction(struct OpalHMIEvent *hmi_evt)
if (malf_alert & PPC_BIT(i)) {
xscom_write(this_cpu()->chip_id, malf_alert_scom,
~PPC_BIT(i));
find_capp_checkstop_reason(i, hmi_evt, &event_generated);
find_nx_checkstop_reason(i, hmi_evt, &event_generated);
find_npu_checkstop_reason(i, hmi_evt, &event_generated);
find_capp_checkstop_reason(i, hmi_evt, &flags);
find_nx_checkstop_reason(i, hmi_evt, &flags);
find_npu_checkstop_reason(i, hmi_evt, &flags);
}
}

find_core_checkstop_reason(hmi_evt, &event_generated);
find_core_checkstop_reason(hmi_evt, &flags);

/*
* If we fail to find checkstop reason, send an unknown HMI event.
*/
if (!event_generated) {
if (!(flags & OPAL_HMI_FLAGS_NEW_EVENT)) {
hmi_evt->u.xstop_error.xstop_type = CHECKSTOP_TYPE_UNKNOWN;
hmi_evt->u.xstop_error.xstop_reason = 0;
queue_hmi_event(hmi_evt, false);
queue_hmi_event(hmi_evt, false, &flags);
}
*out_flags |= flags;
}

static void wait_for_cleanup_complete(void)
Expand Down Expand Up @@ -911,7 +909,7 @@ static int get_split_core_mode(void)
* - SPR_TFMR_TB_RESIDUE_ERR
* - SPR_TFMR_HDEC_PARITY_ERROR
*/
static void pre_recovery_cleanup_p8(uint64_t hmer)
static void pre_recovery_cleanup_p8(uint64_t hmer, uint64_t *out_flags)
{
uint64_t tfmr;
uint32_t sibling_thread_mask;
Expand Down Expand Up @@ -940,11 +938,19 @@ static void pre_recovery_cleanup_p8(uint64_t hmer)
*/
lock(&hmi_lock);
tfmr = mfspr(SPR_TFMR);
if (!(tfmr & SPR_TFMR_TB_VALID))
*out_flags |= OPAL_HMI_FLAGS_TB_RESYNC;
if (tfmr & SPR_TFMR_DEC_PARITY_ERR)
*out_flags |= OPAL_HMI_FLAGS_DEC_LOST;
if (!(tfmr & (SPR_TFMR_TB_RESIDUE_ERR | SPR_TFMR_HDEC_PARITY_ERROR))) {
unlock(&hmi_lock);
return;
}

/* Tell OS about a possible loss of HDEC */
if (tfmr & SPR_TFMR_HDEC_PARITY_ERROR)
*out_flags |= OPAL_HMI_FLAGS_HDEC_LOST;

/* Gather split core information. */
split_core_mode = get_split_core_mode();
threads_per_core = cpu_thread_count / split_core_mode;
Expand Down Expand Up @@ -1015,7 +1021,7 @@ static void pre_recovery_cleanup_p8(uint64_t hmer)
* - SPR_TFMR_TB_RESIDUE_ERR
* - SPR_TFMR_HDEC_PARITY_ERROR
*/
static void pre_recovery_cleanup_p9(uint64_t hmer)
static void pre_recovery_cleanup_p9(uint64_t hmer, uint64_t *out_flags)
{
uint64_t tfmr;
int threads_per_core = cpu_thread_count;
Expand Down Expand Up @@ -1043,6 +1049,10 @@ static void pre_recovery_cleanup_p9(uint64_t hmer)
*/
lock(&hmi_lock);
tfmr = mfspr(SPR_TFMR);
if (!(tfmr & SPR_TFMR_TB_VALID))
*out_flags |= OPAL_HMI_FLAGS_TB_RESYNC;
if (tfmr & SPR_TFMR_DEC_PARITY_ERR)
*out_flags |= OPAL_HMI_FLAGS_DEC_LOST;
if (!(tfmr & (SPR_TFMR_TB_RESIDUE_ERR | SPR_TFMR_HDEC_PARITY_ERROR))) {
unlock(&hmi_lock);
return;
Expand All @@ -1068,6 +1078,10 @@ static void pre_recovery_cleanup_p9(uint64_t hmer)
if ((*(this_cpu()->core_hmi_state_ptr) & CORE_THREAD_MASK) == 0)
*(this_cpu()->core_hmi_state_ptr) &= ~HMI_STATE_CLEANUP_DONE;

/* Tell OS about a possible loss of HDEC */
if (tfmr & SPR_TFMR_HDEC_PARITY_ERROR)
*out_flags |= OPAL_HMI_FLAGS_HDEC_LOST;

/*
* Clear TB and wait for other threads to finish its cleanup work.
*/
Expand Down Expand Up @@ -1098,12 +1112,12 @@ static void pre_recovery_cleanup_p9(uint64_t hmer)
wait_for_cleanup_complete();
}

static void pre_recovery_cleanup(uint64_t hmer)
static void pre_recovery_cleanup(uint64_t hmer, uint64_t *out_flags)
{
if (proc_gen == proc_gen_p9)
return pre_recovery_cleanup_p9(hmer);
return pre_recovery_cleanup_p9(hmer, out_flags);
else
return pre_recovery_cleanup_p8(hmer);
return pre_recovery_cleanup_p8(hmer, out_flags);
}

static void hmi_exit(void)
Expand Down Expand Up @@ -1135,7 +1149,8 @@ static void hmi_print_debug(const uint8_t *msg, uint64_t hmer)
}
}

int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
static int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt,
uint64_t *out_flags)
{
struct cpu_thread *cpu = this_cpu();
int recover = 1;
Expand All @@ -1145,7 +1160,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
* In case of split core, some of the Timer facility errors need
* cleanup to be done before we proceed with the error recovery.
*/
pre_recovery_cleanup(hmer);
pre_recovery_cleanup(hmer, out_flags);

lock(&hmi_lock);
/*
Expand Down Expand Up @@ -1178,15 +1193,15 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
if (hmi_evt) {
hmi_evt->severity = OpalHMI_SEV_NO_ERROR;
hmi_evt->type = OpalHMI_ERROR_PROC_RECOV_DONE;
queue_hmi_event(hmi_evt, recover);
queue_hmi_event(hmi_evt, recover, out_flags);
}
}
if (hmer & SPR_HMER_PROC_RECV_ERROR_MASKED) {
handled |= SPR_HMER_PROC_RECV_ERROR_MASKED;
if (hmi_evt) {
hmi_evt->severity = OpalHMI_SEV_NO_ERROR;
hmi_evt->type = OpalHMI_ERROR_PROC_RECOV_MASKED;
queue_hmi_event(hmi_evt, recover);
queue_hmi_event(hmi_evt, recover, out_flags);
}
hmi_print_debug("Processor recovery Done (masked).", hmer);
}
Expand All @@ -1195,7 +1210,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
if (hmi_evt) {
hmi_evt->severity = OpalHMI_SEV_NO_ERROR;
hmi_evt->type = OpalHMI_ERROR_PROC_RECOV_DONE_AGAIN;
queue_hmi_event(hmi_evt, recover);
queue_hmi_event(hmi_evt, recover, out_flags);
}
hmi_print_debug("Processor recovery occurred again before"
"bit2 was cleared\n", hmer);
Expand All @@ -1206,7 +1221,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)

hmi_print_debug("Malfunction Alert", hmer);
if (hmi_evt)
decode_malfunction(hmi_evt);
decode_malfunction(hmi_evt, out_flags);
}

/* Assert if we see Hypervisor resource error, we can not continue. */
Expand All @@ -1218,7 +1233,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
if (hmi_evt) {
hmi_evt->severity = OpalHMI_SEV_FATAL;
hmi_evt->type = OpalHMI_ERROR_HYP_RESOURCE;
queue_hmi_event(hmi_evt, recover);
queue_hmi_event(hmi_evt, recover, out_flags);
}
}

Expand All @@ -1237,7 +1252,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
hmi_evt->severity = OpalHMI_SEV_ERROR_SYNC;
hmi_evt->type = OpalHMI_ERROR_TFAC;
hmi_evt->tfmr = tfmr;
queue_hmi_event(hmi_evt, recover);
queue_hmi_event(hmi_evt, recover, out_flags);
}
}
if (hmer & SPR_HMER_TFMR_PARITY_ERROR) {
Expand All @@ -1250,7 +1265,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)
hmi_evt->severity = OpalHMI_SEV_FATAL;
hmi_evt->type = OpalHMI_ERROR_TFMR_PARITY;
hmi_evt->tfmr = tfmr;
queue_hmi_event(hmi_evt, recover);
queue_hmi_event(hmi_evt, recover, out_flags);
}
}

Expand All @@ -1273,7 +1288,7 @@ int handle_hmi_exception(uint64_t hmer, struct OpalHMIEvent *hmi_evt)

static int64_t opal_handle_hmi(void)
{
uint64_t hmer;
uint64_t hmer, dummy_flags;
struct OpalHMIEvent hmi_evt;

/*
Expand All @@ -1286,8 +1301,30 @@ static int64_t opal_handle_hmi(void)
hmi_evt.version = OpalHMIEvt_V2;

hmer = mfspr(SPR_HMER); /* Get HMER register value */
handle_hmi_exception(hmer, &hmi_evt);
handle_hmi_exception(hmer, &hmi_evt, &dummy_flags);

return OPAL_SUCCESS;
}
opal_call(OPAL_HANDLE_HMI, opal_handle_hmi, 0);

static int64_t opal_handle_hmi2(__be64 *out_flags)
{
uint64_t hmer, flags;
struct OpalHMIEvent hmi_evt;

/*
* Compiled time check to see size of OpalHMIEvent do not exceed
* that of struct opal_msg.
*/
BUILD_ASSERT(sizeof(struct opal_msg) >= sizeof(struct OpalHMIEvent));

memset(&hmi_evt, 0, sizeof(struct OpalHMIEvent));
hmi_evt.version = OpalHMIEvt_V2;

hmer = mfspr(SPR_HMER); /* Get HMER register value */
handle_hmi_exception(hmer, &hmi_evt, &flags);
*out_flags = cpu_to_be64(flags);

return OPAL_SUCCESS;
}
opal_call(OPAL_HANDLE_HMI2, opal_handle_hmi2, 1);
8 changes: 8 additions & 0 deletions include/opal-api.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -769,6 +769,14 @@ 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_NEW_EVENT = (1ull << 63), /* An event has been created */
};

enum {
OPAL_P7IOC_DIAG_TYPE_NONE = 0,
OPAL_P7IOC_DIAG_TYPE_RGC = 1,
Expand Down
1 change: 0 additions & 1 deletion include/opal-internal.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -82,7 +82,6 @@ extern void opal_del_host_sync_notifier(bool (*notify)(void *data));
* Opal internal function prototype
*/
struct OpalHMIEvent;
extern int handle_hmi_exception(__be64 hmer, struct OpalHMIEvent *hmi_evt);
extern int occ_msg_queue_occ_reset(void);

extern unsigned long top_of_ram;
Expand Down

0 comments on commit 88abbe2

Please sign in to comment.