Contributors: 3
Author Tokens Token Proportion Commits Commit Proportion
Yazen Ghannam 1454 84.39% 5 71.43%
Muralidhara M K 267 15.50% 1 14.29%
Linus Torvalds 2 0.12% 1 14.29%
Total 1723 7 


// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * AMD Address Translation Library
 *
 * umc.c : Unified Memory Controller (UMC) topology helpers
 *
 * Copyright (c) 2023, Advanced Micro Devices, Inc.
 * All Rights Reserved.
 *
 * Author: Yazen Ghannam <Yazen.Ghannam@amd.com>
 */

#include "internal.h"

/*
 * MI300 has a fixed, model-specific mapping between a UMC instance and
 * its related Data Fabric Coherent Station instance.
 *
 * The MCA_IPID_UMC[InstanceId] field holds a unique identifier for the
 * UMC instance within a Node. Use this to find the appropriate Coherent
 * Station ID.
 *
 * Redundant bits were removed from the map below.
 */
static const u16 umc_coh_st_map[32] = {
	0x393, 0x293, 0x193, 0x093,
	0x392, 0x292, 0x192, 0x092,
	0x391, 0x291, 0x191, 0x091,
	0x390, 0x290, 0x190, 0x090,
	0x793, 0x693, 0x593, 0x493,
	0x792, 0x692, 0x592, 0x492,
	0x791, 0x691, 0x591, 0x491,
	0x790, 0x690, 0x590, 0x490,
};

#define UMC_ID_MI300 GENMASK(23, 12)
static u8 get_coh_st_inst_id_mi300(struct atl_err *err)
{
	u16 umc_id = FIELD_GET(UMC_ID_MI300, err->ipid);
	u8 i;

	for (i = 0; i < ARRAY_SIZE(umc_coh_st_map); i++) {
		if (umc_id == umc_coh_st_map[i])
			break;
	}

	WARN_ON_ONCE(i >= ARRAY_SIZE(umc_coh_st_map));

	return i;
}

/* XOR the bits in @val. */
static u16 bitwise_xor_bits(u16 val)
{
	u16 tmp = 0;
	u8 i;

	for (i = 0; i < 16; i++)
		tmp ^= (val >> i) & 0x1;

	return tmp;
}

struct xor_bits {
	bool	xor_enable;
	u16	col_xor;
	u32	row_xor;
};

#define NUM_BANK_BITS	4
#define NUM_COL_BITS	5
#define NUM_SID_BITS	2

static struct {
	/* UMC::CH::AddrHashBank */
	struct xor_bits	bank[NUM_BANK_BITS];

	/* UMC::CH::AddrHashPC */
	struct xor_bits	pc;

	/* UMC::CH::AddrHashPC2 */
	u8		bank_xor;
} addr_hash;

static struct {
	u8 bank[NUM_BANK_BITS];
	u8 col[NUM_COL_BITS];
	u8 sid[NUM_SID_BITS];
	u8 num_row_lo;
	u8 num_row_hi;
	u8 row_lo;
	u8 row_hi;
	u8 pc;
} bit_shifts;

#define MI300_UMC_CH_BASE	0x90000
#define MI300_ADDR_CFG		(MI300_UMC_CH_BASE + 0x30)
#define MI300_ADDR_SEL		(MI300_UMC_CH_BASE + 0x40)
#define MI300_COL_SEL_LO	(MI300_UMC_CH_BASE + 0x50)
#define MI300_ADDR_SEL_2	(MI300_UMC_CH_BASE + 0xA4)
#define MI300_ADDR_HASH_BANK0	(MI300_UMC_CH_BASE + 0xC8)
#define MI300_ADDR_HASH_PC	(MI300_UMC_CH_BASE + 0xE0)
#define MI300_ADDR_HASH_PC2	(MI300_UMC_CH_BASE + 0xE4)

#define ADDR_HASH_XOR_EN	BIT(0)
#define ADDR_HASH_COL_XOR	GENMASK(13, 1)
#define ADDR_HASH_ROW_XOR	GENMASK(31, 14)
#define ADDR_HASH_BANK_XOR	GENMASK(5, 0)

#define ADDR_CFG_NUM_ROW_LO	GENMASK(11, 8)
#define ADDR_CFG_NUM_ROW_HI	GENMASK(15, 12)

#define ADDR_SEL_BANK0		GENMASK(3, 0)
#define ADDR_SEL_BANK1		GENMASK(7, 4)
#define ADDR_SEL_BANK2		GENMASK(11, 8)
#define ADDR_SEL_BANK3		GENMASK(15, 12)
#define ADDR_SEL_BANK4		GENMASK(20, 16)
#define ADDR_SEL_ROW_LO		GENMASK(27, 24)
#define ADDR_SEL_ROW_HI		GENMASK(31, 28)

#define COL_SEL_LO_COL0		GENMASK(3, 0)
#define COL_SEL_LO_COL1		GENMASK(7, 4)
#define COL_SEL_LO_COL2		GENMASK(11, 8)
#define COL_SEL_LO_COL3		GENMASK(15, 12)
#define COL_SEL_LO_COL4		GENMASK(19, 16)

#define ADDR_SEL_2_BANK5	GENMASK(4, 0)
#define ADDR_SEL_2_CHAN		GENMASK(15, 12)

/*
 * Read UMC::CH::AddrHash{Bank,PC,PC2} registers to get XOR bits used
 * for hashing.
 *
 * Also, read UMC::CH::Addr{Cfg,Sel,Sel2} and UMC::CH:ColSelLo registers to
 * get the values needed to reconstruct the normalized address. Apply additional
 * offsets to the raw register values, as needed.
 *
 * Do this during module init, since the values will not change during run time.
 *
 * These registers are instantiated for each UMC across each AMD Node.
 * However, they should be identically programmed due to the fixed hardware
 * design of MI300 systems. So read the values from Node 0 UMC 0 and keep a
 * single global structure for simplicity.
 */
int get_umc_info_mi300(void)
{
	u32 temp;
	int ret;
	u8 i;

	for (i = 0; i < NUM_BANK_BITS; i++) {
		ret = amd_smn_read(0, MI300_ADDR_HASH_BANK0 + (i * 4), &temp);
		if (ret)
			return ret;

		addr_hash.bank[i].xor_enable = FIELD_GET(ADDR_HASH_XOR_EN,  temp);
		addr_hash.bank[i].col_xor    = FIELD_GET(ADDR_HASH_COL_XOR, temp);
		addr_hash.bank[i].row_xor    = FIELD_GET(ADDR_HASH_ROW_XOR, temp);
	}

	ret = amd_smn_read(0, MI300_ADDR_HASH_PC, &temp);
	if (ret)
		return ret;

	addr_hash.pc.xor_enable = FIELD_GET(ADDR_HASH_XOR_EN,  temp);
	addr_hash.pc.col_xor    = FIELD_GET(ADDR_HASH_COL_XOR, temp);
	addr_hash.pc.row_xor    = FIELD_GET(ADDR_HASH_ROW_XOR, temp);

	ret = amd_smn_read(0, MI300_ADDR_HASH_PC2, &temp);
	if (ret)
		return ret;

	addr_hash.bank_xor = FIELD_GET(ADDR_HASH_BANK_XOR, temp);

	ret = amd_smn_read(0, MI300_ADDR_CFG, &temp);
	if (ret)
		return ret;

	bit_shifts.num_row_hi = FIELD_GET(ADDR_CFG_NUM_ROW_HI, temp);
	bit_shifts.num_row_lo = 10 + FIELD_GET(ADDR_CFG_NUM_ROW_LO, temp);

	ret = amd_smn_read(0, MI300_ADDR_SEL, &temp);
	if (ret)
		return ret;

	bit_shifts.bank[0] = 5 + FIELD_GET(ADDR_SEL_BANK0, temp);
	bit_shifts.bank[1] = 5 + FIELD_GET(ADDR_SEL_BANK1, temp);
	bit_shifts.bank[2] = 5 + FIELD_GET(ADDR_SEL_BANK2, temp);
	bit_shifts.bank[3] = 5 + FIELD_GET(ADDR_SEL_BANK3, temp);
	/* Use BankBit4 for the SID0 position. */
	bit_shifts.sid[0]  = 5 + FIELD_GET(ADDR_SEL_BANK4, temp);
	bit_shifts.row_lo  = 12 + FIELD_GET(ADDR_SEL_ROW_LO, temp);
	bit_shifts.row_hi  = 24 + FIELD_GET(ADDR_SEL_ROW_HI, temp);

	ret = amd_smn_read(0, MI300_COL_SEL_LO, &temp);
	if (ret)
		return ret;

	bit_shifts.col[0] = 2 + FIELD_GET(COL_SEL_LO_COL0, temp);
	bit_shifts.col[1] = 2 + FIELD_GET(COL_SEL_LO_COL1, temp);
	bit_shifts.col[2] = 2 + FIELD_GET(COL_SEL_LO_COL2, temp);
	bit_shifts.col[3] = 2 + FIELD_GET(COL_SEL_LO_COL3, temp);
	bit_shifts.col[4] = 2 + FIELD_GET(COL_SEL_LO_COL4, temp);

	ret = amd_smn_read(0, MI300_ADDR_SEL_2, &temp);
	if (ret)
		return ret;

	/* Use BankBit5 for the SID1 position. */
	bit_shifts.sid[1] = 5 + FIELD_GET(ADDR_SEL_2_BANK5, temp);
	bit_shifts.pc	  = 5 + FIELD_GET(ADDR_SEL_2_CHAN, temp);

	return 0;
}

/*
 * MI300 systems report a DRAM address in MCA_ADDR for DRAM ECC errors. This must
 * be converted to the intermediate normalized address (NA) before translating to a
 * system physical address.
 *
 * The DRAM address includes bank, row, and column. Also included are bits for
 * pseudochannel (PC) and stack ID (SID).
 *
 * Abbreviations: (S)tack ID, (P)seudochannel, (R)ow, (B)ank, (C)olumn, (Z)ero
 *
 * The MCA address format is as follows:
 *	MCA_ADDR[27:0] = {S[1:0], P[0], R[14:0], B[3:0], C[4:0], Z[0]}
 *
 * Additionally, the PC and Bank bits may be hashed. This must be accounted for before
 * reconstructing the normalized address.
 */
#define MI300_UMC_MCA_COL	GENMASK(5, 1)
#define MI300_UMC_MCA_BANK	GENMASK(9, 6)
#define MI300_UMC_MCA_ROW	GENMASK(24, 10)
#define MI300_UMC_MCA_PC	BIT(25)
#define MI300_UMC_MCA_SID	GENMASK(27, 26)

static unsigned long convert_dram_to_norm_addr_mi300(unsigned long addr)
{
	u16 i, col, row, bank, pc, sid;
	u32 temp;

	col  = FIELD_GET(MI300_UMC_MCA_COL,  addr);
	bank = FIELD_GET(MI300_UMC_MCA_BANK, addr);
	row  = FIELD_GET(MI300_UMC_MCA_ROW,  addr);
	pc   = FIELD_GET(MI300_UMC_MCA_PC,   addr);
	sid  = FIELD_GET(MI300_UMC_MCA_SID,  addr);

	/* Calculate hash for each Bank bit. */
	for (i = 0; i < NUM_BANK_BITS; i++) {
		if (!addr_hash.bank[i].xor_enable)
			continue;

		temp  = bitwise_xor_bits(col & addr_hash.bank[i].col_xor);
		temp ^= bitwise_xor_bits(row & addr_hash.bank[i].row_xor);
		bank ^= temp << i;
	}

	/* Calculate hash for PC bit. */
	if (addr_hash.pc.xor_enable) {
		temp  = bitwise_xor_bits(col  & addr_hash.pc.col_xor);
		temp ^= bitwise_xor_bits(row  & addr_hash.pc.row_xor);
		/* Bits SID[1:0] act as Bank[5:4] for PC hash, so apply them here. */
		temp ^= bitwise_xor_bits((bank | sid << NUM_BANK_BITS) & addr_hash.bank_xor);
		pc   ^= temp;
	}

	/* Reconstruct the normalized address starting with NA[4:0] = 0 */
	addr  = 0;

	/* Column bits */
	for (i = 0; i < NUM_COL_BITS; i++) {
		temp  = (col >> i) & 0x1;
		addr |= temp << bit_shifts.col[i];
	}

	/* Bank bits */
	for (i = 0; i < NUM_BANK_BITS; i++) {
		temp  = (bank >> i) & 0x1;
		addr |= temp << bit_shifts.bank[i];
	}

	/* Row lo bits */
	for (i = 0; i < bit_shifts.num_row_lo; i++) {
		temp  = (row >> i) & 0x1;
		addr |= temp << (i + bit_shifts.row_lo);
	}

	/* Row hi bits */
	for (i = 0; i < bit_shifts.num_row_hi; i++) {
		temp  = (row >> (i + bit_shifts.num_row_lo)) & 0x1;
		addr |= temp << (i + bit_shifts.row_hi);
	}

	/* PC bit */
	addr |= pc << bit_shifts.pc;

	/* SID bits */
	for (i = 0; i < NUM_SID_BITS; i++) {
		temp  = (sid >> i) & 0x1;
		addr |= temp << bit_shifts.sid[i];
	}

	pr_debug("Addr=0x%016lx", addr);
	pr_debug("Bank=%u Row=%u Column=%u PC=%u SID=%u", bank, row, col, pc, sid);

	return addr;
}

/*
 * When a DRAM ECC error occurs on MI300 systems, it is recommended to retire
 * all memory within that DRAM row. This applies to the memory with a DRAM
 * bank.
 *
 * To find the memory addresses, loop through permutations of the DRAM column
 * bits and find the System Physical address of each. The column bits are used
 * to calculate the intermediate Normalized address, so all permutations should
 * be checked.
 *
 * See amd_atl::convert_dram_to_norm_addr_mi300() for MI300 address formats.
 */
#define MI300_NUM_COL		BIT(HWEIGHT(MI300_UMC_MCA_COL))
static void retire_row_mi300(struct atl_err *a_err)
{
	unsigned long addr;
	struct page *p;
	u8 col;

	for (col = 0; col < MI300_NUM_COL; col++) {
		a_err->addr &= ~MI300_UMC_MCA_COL;
		a_err->addr |= FIELD_PREP(MI300_UMC_MCA_COL, col);

		addr = amd_convert_umc_mca_addr_to_sys_addr(a_err);
		if (IS_ERR_VALUE(addr))
			continue;

		addr = PHYS_PFN(addr);

		/*
		 * Skip invalid or already poisoned pages to avoid unnecessary
		 * error messages from memory_failure().
		 */
		p = pfn_to_online_page(addr);
		if (!p)
			continue;

		if (PageHWPoison(p))
			continue;

		memory_failure(addr, 0);
	}
}

void amd_retire_dram_row(struct atl_err *a_err)
{
	if (df_cfg.rev == DF4p5 && df_cfg.flags.heterogeneous)
		return retire_row_mi300(a_err);
}
EXPORT_SYMBOL_GPL(amd_retire_dram_row);

static unsigned long get_addr(unsigned long addr)
{
	if (df_cfg.rev == DF4p5 && df_cfg.flags.heterogeneous)
		return convert_dram_to_norm_addr_mi300(addr);

	return addr;
}

#define MCA_IPID_INST_ID_HI	GENMASK_ULL(47, 44)
static u8 get_die_id(struct atl_err *err)
{
	/*
	 * AMD Node ID is provided in MCA_IPID[InstanceIdHi], and this
	 * needs to be divided by 4 to get the internal Die ID.
	 */
	if (df_cfg.rev == DF4p5 && df_cfg.flags.heterogeneous) {
		u8 node_id = FIELD_GET(MCA_IPID_INST_ID_HI, err->ipid);

		return node_id >> 2;
	}

	/*
	 * For CPUs, this is the AMD Node ID modulo the number
	 * of AMD Nodes per socket.
	 */
	return topology_amd_node_id(err->cpu) % topology_amd_nodes_per_pkg();
}

#define UMC_CHANNEL_NUM	GENMASK(31, 20)
static u8 get_coh_st_inst_id(struct atl_err *err)
{
	if (df_cfg.rev == DF4p5 && df_cfg.flags.heterogeneous)
		return get_coh_st_inst_id_mi300(err);

	return FIELD_GET(UMC_CHANNEL_NUM, err->ipid);
}

unsigned long convert_umc_mca_addr_to_sys_addr(struct atl_err *err)
{
	u8 socket_id = topology_physical_package_id(err->cpu);
	u8 coh_st_inst_id = get_coh_st_inst_id(err);
	unsigned long addr = get_addr(err->addr);
	u8 die_id = get_die_id(err);

	pr_debug("socket_id=0x%x die_id=0x%x coh_st_inst_id=0x%x addr=0x%016lx",
		 socket_id, die_id, coh_st_inst_id, addr);

	return norm_to_sys_addr(socket_id, die_id, coh_st_inst_id, addr);
}