sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling buggrafted

alb_break_llc() decides whether to break LLC preference during active
load balance. It does so by testing that every runnable fair task on the
source rq prefers its LLC:

	env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable

But the two counters cover different sets. nr_pref_llc_running is updated
in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued,
so it follows queued tasks. h_nr_runnable is updated in set_delayed()/
clear_delayed() and drops delay-dequeued tasks.

So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted
in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks,
alb_break_llc() returns false, and active balance is free to pull a task
off its preferred LLC. Active balance only moves runnable tasks, and this
is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB
skips the per-task test in can_migrate_task(). The runnable set is the one
we want.

Fix it on the counter side. A task should be counted in
nr_pref_llc_running exactly while it is both queued on its preferred LLC
(pref_llc_queued) and runnable (!sched_delayed). Define that membership
once in task_pref_llc_runnable(), and adjust the counter only through
pref_llc_running_inc()/pref_llc_running_dec() from the four sites that
change either input: account_llc_enqueue(), account_llc_dequeue(),
set_delayed() and clear_delayed(). Gating every update on the same
predicate keeps the delay, wake and dequeue paths from double-counting
or underflowing; see the comments at those sites for the ordering.

nr_llc_running and sd->llc_counts are not touched and stay on queued
semantics.

Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC")
Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/
Reported-by: Zhan Xusheng <zhanxusheng@xiaomi.com>
Suggested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Cc: <stable@kernel.org> # v7.2.x
Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com

1 files changed, 387 insertions, 0 deletions

diff --git a/include/ras/ras_event.h b/include/ras/ras_event.h
new file mode 100644
index 000000000..fdb785fa4
--- /dev/null
+++ b/include/ras/ras_event.h
@@ -0,0 +1,387 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM ras
+#define TRACE_INCLUDE_FILE ras_event
+
+#if !defined(_TRACE_HW_EVENT_MC_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_HW_EVENT_MC_H
+
+#include <linux/tracepoint.h>
+#include <linux/edac.h>
+#include <linux/ktime.h>
+#include <linux/pci.h>
+#include <linux/aer.h>
+#include <linux/cper.h>
+
+/*
+ * MCE Extended Error Log trace event
+ *
+ * These events are generated when hardware detects a corrected or
+ * uncorrected event.
+ */
+
+/* memory trace event */
+
+#if defined(CONFIG_ACPI_EXTLOG) || defined(CONFIG_ACPI_EXTLOG_MODULE)
+TRACE_EVENT(extlog_mem_event,
+	TP_PROTO(struct cper_sec_mem_err *mem,
+		 u32 err_seq,
+		 const guid_t *fru_id,
+		 const char *fru_text,
+		 u8 sev),
+
+	TP_ARGS(mem, err_seq, fru_id, fru_text, sev),
+
+	TP_STRUCT__entry(
+		__field(u32, err_seq)
+		__field(u8, etype)
+		__field(u8, sev)
+		__field(u64, pa)
+		__field(u8, pa_mask_lsb)
+		__field_struct(guid_t, fru_id)
+		__string(fru_text, fru_text)
+		__field_struct(struct cper_mem_err_compact, data)
+	),
+
+	TP_fast_assign(
+		__entry->err_seq = err_seq;
+		if (mem->validation_bits & CPER_MEM_VALID_ERROR_TYPE)
+			__entry->etype = mem->error_type;
+		else
+			__entry->etype = ~0;
+		__entry->sev = sev;
+		if (mem->validation_bits & CPER_MEM_VALID_PA)
+			__entry->pa = mem->physical_addr;
+		else
+			__entry->pa = ~0ull;
+
+		if (mem->validation_bits & CPER_MEM_VALID_PA_MASK)
+			__entry->pa_mask_lsb = (u8)__ffs64(mem->physical_addr_mask);
+		else
+			__entry->pa_mask_lsb = ~0;
+		__entry->fru_id = *fru_id;
+		__assign_str(fru_text);
+		cper_mem_err_pack(mem, &__entry->data);
+	),
+
+	TP_printk("{%d} %s error: %s physical addr: %016llx (mask lsb: %x) %sFRU: %pUl %.20s",
+		  __entry->err_seq,
+		  cper_severity_str(__entry->sev),
+		  cper_mem_err_type_str(__entry->etype),
+		  __entry->pa,
+		  __entry->pa_mask_lsb,
+		  cper_mem_err_unpack(p, &__entry->data),
+		  &__entry->fru_id,
+		  __get_str(fru_text))
+);
+#endif
+
+/*
+ * Hardware Events Report
+ *
+ * Those events are generated when hardware detected a corrected or
+ * uncorrected event, and are meant to replace the current API to report
+ * errors defined on both EDAC and MCE subsystems.
+ *
+ * FIXME: Add events for handling memory errors originated from the
+ *        MCE subsystem.
+ */
+
+/*
+ * Hardware-independent Memory Controller specific events
+ */
+
+/*
+ * Default error mechanisms for Memory Controller errors (CE and UE)
+ */
+TRACE_EVENT(mc_event,
+
+	TP_PROTO(const unsigned int err_type,
+		 const char *error_msg,
+		 const char *label,
+		 const int error_count,
+		 const u8 mc_index,
+		 const s8 top_layer,
+		 const s8 mid_layer,
+		 const s8 low_layer,
+		 unsigned long address,
+		 const u8 grain_bits,
+		 unsigned long syndrome,
+		 const char *driver_detail),
+
+	TP_ARGS(err_type, error_msg, label, error_count, mc_index,
+		top_layer, mid_layer, low_layer, address, grain_bits,
+		syndrome, driver_detail),
+
+	TP_STRUCT__entry(
+		__field(	unsigned int,	error_type		)
+		__string(	msg,		error_msg		)
+		__string(	label,		label			)
+		__field(	u16,		error_count		)
+		__field(	u8,		mc_index		)
+		__field(	s8,		top_layer		)
+		__field(	s8,		middle_layer		)
+		__field(	s8,		lower_layer		)
+		__field(	long,		address			)
+		__field(	u8,		grain_bits		)
+		__field(	long,		syndrome		)
+		__string(	driver_detail,	driver_detail		)
+	),
+
+	TP_fast_assign(
+		__entry->error_type		= err_type;
+		__assign_str(msg);
+		__assign_str(label);
+		__entry->error_count		= error_count;
+		__entry->mc_index		= mc_index;
+		__entry->top_layer		= top_layer;
+		__entry->middle_layer		= mid_layer;
+		__entry->lower_layer		= low_layer;
+		__entry->address		= address;
+		__entry->grain_bits		= grain_bits;
+		__entry->syndrome		= syndrome;
+		__assign_str(driver_detail);
+	),
+
+	TP_printk("%d %s error%s:%s%s on %s (mc:%d location:%d:%d:%d address:0x%08lx grain:%d syndrome:0x%08lx%s%s)",
+		  __entry->error_count,
+		  mc_event_error_type(__entry->error_type),
+		  __entry->error_count > 1 ? "s" : "",
+		  __get_str(msg)[0] ? " " : "",
+		  __get_str(msg),
+		  __get_str(label),
+		  __entry->mc_index,
+		  __entry->top_layer,
+		  __entry->middle_layer,
+		  __entry->lower_layer,
+		  __entry->address,
+		  1 << __entry->grain_bits,
+		  __entry->syndrome,
+		  __get_str(driver_detail)[0] ? " " : "",
+		  __get_str(driver_detail))
+);
+
+/*
+ * ARM Processor Events Report
+ *
+ * This event is generated when hardware detects an ARM processor error
+ * has occurred. UEFI 2.6 spec section N.2.4.4.
+ */
+#define APEIL "ARM Processor Err Info data len"
+#define APEID "ARM Processor Err Info raw data"
+#define APECIL "ARM Processor Err Context Info data len"
+#define APECID "ARM Processor Err Context Info raw data"
+#define VSEIL "Vendor Specific Err Info data len"
+#define VSEID "Vendor Specific Err Info raw data"
+TRACE_EVENT(arm_event,
+
+	TP_PROTO(const struct cper_sec_proc_arm *proc,
+		 const u8 *pei_err,
+		 const u32 pei_len,
+		 const u8 *ctx_err,
+		 const u32 ctx_len,
+		 const u8 *oem,
+		 const u32 oem_len,
+		 u8 sev,
+		 int cpu),
+
+	TP_ARGS(proc, pei_err, pei_len, ctx_err, ctx_len, oem, oem_len, sev, cpu),
+
+	TP_STRUCT__entry(
+		__field(u64, mpidr)
+		__field(u64, midr)
+		__field(u32, running_state)
+		__field(u32, psci_state)
+		__field(u8, affinity)
+		__field(u32, pei_len)
+		__dynamic_array(u8, pei_buf, pei_len)
+		__field(u32, ctx_len)
+		__dynamic_array(u8, ctx_buf, ctx_len)
+		__field(u32, oem_len)
+		__dynamic_array(u8, oem_buf, oem_len)
+		__field(u8, sev)
+		__field(int, cpu)
+	),
+
+	TP_fast_assign(
+		if (proc->validation_bits & CPER_ARM_VALID_AFFINITY_LEVEL)
+			__entry->affinity = proc->affinity_level;
+		else
+			__entry->affinity = ~0;
+		if (proc->validation_bits & CPER_ARM_VALID_MPIDR)
+			__entry->mpidr = proc->mpidr;
+		else
+			__entry->mpidr = 0ULL;
+		__entry->midr = proc->midr;
+		if (proc->validation_bits & CPER_ARM_VALID_RUNNING_STATE) {
+			__entry->running_state = proc->running_state;
+			__entry->psci_state = proc->psci_state;
+		} else {
+			__entry->running_state = ~0;
+			__entry->psci_state = ~0;
+		}
+		__entry->pei_len = pei_len;
+		memcpy(__get_dynamic_array(pei_buf), pei_err, pei_len);
+		__entry->ctx_len = ctx_len;
+		memcpy(__get_dynamic_array(ctx_buf), ctx_err, ctx_len);
+		__entry->oem_len = oem_len;
+		memcpy(__get_dynamic_array(oem_buf), oem, oem_len);
+		__entry->sev = sev;
+		__entry->cpu = cpu;
+	),
+
+	TP_printk("cpu: %d; error: %d; affinity level: %d; MPIDR: %016llx; MIDR: %016llx; "
+		  "running state: %d; PSCI state: %d; "
+		  "%s: %d; %s: %s; %s: %d; %s: %s; %s: %d; %s: %s",
+		  __entry->cpu,
+		  __entry->sev,
+		  __entry->affinity, __entry->mpidr, __entry->midr,
+		  __entry->running_state, __entry->psci_state,
+		  APEIL, __entry->pei_len, APEID,
+		  __print_hex(__get_dynamic_array(pei_buf), __entry->pei_len),
+		  APECIL, __entry->ctx_len, APECID,
+		  __print_hex(__get_dynamic_array(ctx_buf), __entry->ctx_len),
+		  VSEIL, __entry->oem_len, VSEID,
+		  __print_hex(__get_dynamic_array(oem_buf), __entry->oem_len))
+);
+
+/*
+ * Non-Standard Section Report
+ *
+ * This event is generated when hardware detected a hardware
+ * error event, which may be of non-standard section as defined
+ * in UEFI spec appendix "Common Platform Error Record", or may
+ * be of sections for which TRACE_EVENT is not defined.
+ *
+ */
+TRACE_EVENT(non_standard_event,
+
+	TP_PROTO(const guid_t *sec_type,
+		 const guid_t *fru_id,
+		 const char *fru_text,
+		 const u8 sev,
+		 const u8 *err,
+		 const u32 len),
+
+	TP_ARGS(sec_type, fru_id, fru_text, sev, err, len),
+
+	TP_STRUCT__entry(
+		__array(char, sec_type, UUID_SIZE)
+		__array(char, fru_id, UUID_SIZE)
+		__string(fru_text, fru_text)
+		__field(u8, sev)
+		__field(u32, len)
+		__dynamic_array(u8, buf, len)
+	),
+
+	TP_fast_assign(
+		memcpy(__entry->sec_type, sec_type, UUID_SIZE);
+		memcpy(__entry->fru_id, fru_id, UUID_SIZE);
+		__assign_str(fru_text);
+		__entry->sev = sev;
+		__entry->len = len;
+		memcpy(__get_dynamic_array(buf), err, len);
+	),
+
+	TP_printk("severity: %d; sec type:%pU; FRU: %pU %s; data len:%d; raw data:%s",
+		  __entry->sev, __entry->sec_type,
+		  __entry->fru_id, __get_str(fru_text),
+		  __entry->len,
+		  __print_hex(__get_dynamic_array(buf), __entry->len))
+);
+
+#ifdef CONFIG_PCIEAER
+/*
+ * PCIe AER Trace event
+ *
+ * These events are generated when hardware detects a corrected or
+ * uncorrected event on a PCIe device. The event report has
+ * the following structure:
+ *
+ * char * dev_name -	The name of the slot where the device resides
+ *			([domain:]bus:device.function).
+ * u32 status -		Either the correctable or uncorrectable register
+ *			indicating what error or errors have been seen
+ * u8 severity -	error severity 0:NONFATAL 1:FATAL 2:CORRECTED
+ */
+
+#define aer_correctable_errors					\
+	{PCI_ERR_COR_RCVR,	"Receiver Error"},		\
+	{PCI_ERR_COR_BAD_TLP,	"Bad TLP"},			\
+	{PCI_ERR_COR_BAD_DLLP,	"Bad DLLP"},			\
+	{PCI_ERR_COR_REP_ROLL,	"RELAY_NUM Rollover"},		\
+	{PCI_ERR_COR_REP_TIMER,	"Replay Timer Timeout"},	\
+	{PCI_ERR_COR_ADV_NFAT,	"Advisory Non-Fatal Error"},	\
+	{PCI_ERR_COR_INTERNAL,	"Corrected Internal Error"},	\
+	{PCI_ERR_COR_LOG_OVER,	"Header Log Overflow"}
+
+#define aer_uncorrectable_errors				\
+	{PCI_ERR_UNC_UND,	"Undefined"},			\
+	{PCI_ERR_UNC_DLP,	"Data Link Protocol Error"},	\
+	{PCI_ERR_UNC_SURPDN,	"Surprise Down Error"},		\
+	{PCI_ERR_UNC_POISON_TLP,"Poisoned TLP"},		\
+	{PCI_ERR_UNC_FCP,	"Flow Control Protocol Error"},	\
+	{PCI_ERR_UNC_COMP_TIME,	"Completion Timeout"},		\
+	{PCI_ERR_UNC_COMP_ABORT,"Completer Abort"},		\
+	{PCI_ERR_UNC_UNX_COMP,	"Unexpected Completion"},	\
+	{PCI_ERR_UNC_RX_OVER,	"Receiver Overflow"},		\
+	{PCI_ERR_UNC_MALF_TLP,	"Malformed TLP"},		\
+	{PCI_ERR_UNC_ECRC,	"ECRC Error"},			\
+	{PCI_ERR_UNC_UNSUP,	"Unsupported Request Error"},	\
+	{PCI_ERR_UNC_ACSV,	"ACS Violation"},		\
+	{PCI_ERR_UNC_INTN,	"Uncorrectable Internal Error"},\
+	{PCI_ERR_UNC_MCBTLP,	"MC Blocked TLP"},		\
+	{PCI_ERR_UNC_ATOMEG,	"AtomicOp Egress Blocked"},	\
+	{PCI_ERR_UNC_TLPPRE,	"TLP Prefix Blocked Error"}
+
+TRACE_EVENT(aer_event,
+	TP_PROTO(const char *dev_name,
+		 const u32 status,
+		 const u8 severity,
+		 const u8 tlp_header_valid,
+		 struct pcie_tlp_log *tlp,
+		 const char *bus_type),
+
+
+	TP_ARGS(dev_name, status, severity, tlp_header_valid, tlp, bus_type),
+
+	TP_STRUCT__entry(
+		__string(	dev_name,	dev_name	)
+		__field(	u32,		status		)
+		__field(	u8,		severity	)
+		__field(	u8, 		tlp_header_valid)
+		__array(	u32, 		tlp_header, PCIE_STD_MAX_TLP_HEADERLOG)
+		__string(	bus_type,	bus_type	)
+	),
+
+	TP_fast_assign(
+		__assign_str(dev_name);
+		__assign_str(bus_type);
+		__entry->status		= status;
+		__entry->severity	= severity;
+		__entry->tlp_header_valid = tlp_header_valid;
+		if (tlp_header_valid) {
+			int i;
+
+			for (i = 0; i < PCIE_STD_MAX_TLP_HEADERLOG; i++)
+				__entry->tlp_header[i] = tlp->dw[i];
+		}
+	),
+
+	TP_printk("%s %s Bus Error: severity=%s, %s, TLP Header=%s\n",
+		__get_str(dev_name), __get_str(bus_type),
+		__entry->severity == AER_CORRECTABLE ? "Corrected" :
+			__entry->severity == AER_FATAL ?
+			"Fatal" : "Uncorrected, non-fatal",
+		__entry->severity == AER_CORRECTABLE ?
+		__print_flags(__entry->status, "|", aer_correctable_errors) :
+		__print_flags(__entry->status, "|", aer_uncorrectable_errors),
+		__entry->tlp_header_valid ?
+			__print_array(__entry->tlp_header, PCIE_STD_MAX_TLP_HEADERLOG, 4) :
+			"Not available")
+);
+#endif /* CONFIG_PCIEAER */
+#endif /* _TRACE_HW_EVENT_MC_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>