[PATCH 2/4] EDAC/i10nm: Retrieve and print retry_rd_err_log registers for HBM

From: Tony Luck
Date: Fri Jul 22 2022 - 19:34:07 EST


From: Qiuxu Zhuo <qiuxu.zhuo@xxxxxxxxx>

An HBM memory channel is divided into two pseudo channels. Each
pseudo channel has its own retry_rd_err_log registers. Retrieve and
print retry_rd_err_log registers of the HBM pseudo channel if the
memory error is from HBM.

Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@xxxxxxxxx>
Signed-off-by: Tony Luck <tony.luck@xxxxxxxxx>
---
drivers/edac/skx_common.h | 4 ++
drivers/edac/i10nm_base.c | 84 +++++++++++++++++++++++++++++++--------
2 files changed, 71 insertions(+), 17 deletions(-)

diff --git a/drivers/edac/skx_common.h b/drivers/edac/skx_common.h
index 70ec4f41911b..dbf8e458ad2b 100644
--- a/drivers/edac/skx_common.h
+++ b/drivers/edac/skx_common.h
@@ -158,7 +158,11 @@ struct res_config {
int sad_all_offset;
/* Offsets of retry_rd_err_log registers */
u32 *offsets_scrub;
+ u32 *offsets_scrub_hbm0;
+ u32 *offsets_scrub_hbm1;
u32 *offsets_demand;
+ u32 *offsets_demand_hbm0;
+ u32 *offsets_demand_hbm1;
};

typedef int (*get_dimm_config_f)(struct mem_ctl_info *mci,
diff --git a/drivers/edac/i10nm_base.c b/drivers/edac/i10nm_base.c
index 6cf50ee0b77c..976d8e8a4d1b 100644
--- a/drivers/edac/i10nm_base.c
+++ b/drivers/edac/i10nm_base.c
@@ -77,18 +77,20 @@ static int retry_rd_err_log;

static u32 offsets_scrub_icx[] = {0x22c60, 0x22c54, 0x22c5c, 0x22c58, 0x22c28, 0x20ed8};
static u32 offsets_scrub_spr[] = {0x22c60, 0x22c54, 0x22f08, 0x22c58, 0x22c28, 0x20ed8};
+static u32 offsets_scrub_spr_hbm0[] = {0x2860, 0x2854, 0x2b08, 0x2858, 0x2828, 0x0ed8};
+static u32 offsets_scrub_spr_hbm1[] = {0x2c60, 0x2c54, 0x2f08, 0x2c58, 0x2c28, 0x0fa8};
static u32 offsets_demand_icx[] = {0x22e54, 0x22e60, 0x22e64, 0x22e58, 0x22e5c, 0x20ee0};
static u32 offsets_demand_spr[] = {0x22e54, 0x22e60, 0x22f10, 0x22e58, 0x22e5c, 0x20ee0};
+static u32 offsets_demand_spr_hbm0[] = {0x2a54, 0x2a60, 0x2b10, 0x2a58, 0x2a5c, 0x0ee0};
+static u32 offsets_demand_spr_hbm1[] = {0x2e54, 0x2e60, 0x2f10, 0x2e58, 0x2e5c, 0x0fb0};

-static void __enable_retry_rd_err_log(struct skx_imc *imc, int chan, bool enable)
+static void __enable_retry_rd_err_log(struct skx_imc *imc, int chan, bool enable,
+ u32 *offsets_scrub, u32 *offsets_demand)
{
u32 s, d;

- if (!imc->mbase)
- return;
-
- s = I10NM_GET_REG32(imc, chan, res_cfg->offsets_scrub[0]);
- d = I10NM_GET_REG32(imc, chan, res_cfg->offsets_demand[0]);
+ s = I10NM_GET_REG32(imc, chan, offsets_scrub[0]);
+ d = I10NM_GET_REG32(imc, chan, offsets_demand[0]);

if (enable) {
/* Save default configurations */
@@ -115,21 +117,39 @@ static void __enable_retry_rd_err_log(struct skx_imc *imc, int chan, bool enable
d &= ~RETRY_RD_ERR_LOG_EN;
}

- I10NM_SET_REG32(imc, chan, res_cfg->offsets_scrub[0], s);
- I10NM_SET_REG32(imc, chan, res_cfg->offsets_demand[0], d);
+ I10NM_SET_REG32(imc, chan, offsets_scrub[0], s);
+ I10NM_SET_REG32(imc, chan, offsets_demand[0], d);
}

static void enable_retry_rd_err_log(bool enable)
{
+ struct skx_imc *imc;
struct skx_dev *d;
int i, j;

edac_dbg(2, "\n");

list_for_each_entry(d, i10nm_edac_list, list)
- for (i = 0; i < I10NM_NUM_IMC; i++)
- for (j = 0; j < I10NM_NUM_CHANNELS; j++)
- __enable_retry_rd_err_log(&d->imc[i], j, enable);
+ for (i = 0; i < I10NM_NUM_IMC; i++) {
+ imc = &d->imc[i];
+ if (!imc->mbase)
+ continue;
+
+ for (j = 0; j < I10NM_NUM_CHANNELS; j++) {
+ if (imc->hbm_mc) {
+ __enable_retry_rd_err_log(imc, j, enable,
+ res_cfg->offsets_scrub_hbm0,
+ res_cfg->offsets_demand_hbm0);
+ __enable_retry_rd_err_log(imc, j, enable,
+ res_cfg->offsets_scrub_hbm1,
+ res_cfg->offsets_demand_hbm1);
+ } else {
+ __enable_retry_rd_err_log(imc, j, enable,
+ res_cfg->offsets_scrub,
+ res_cfg->offsets_demand);
+ }
+ }
+ }
}

static void show_retry_rd_err_log(struct decoded_addr *res, char *msg,
@@ -140,12 +160,24 @@ static void show_retry_rd_err_log(struct decoded_addr *res, char *msg,
u32 corr0, corr1, corr2, corr3;
u64 log2a, log5;
u32 *offsets;
- int n;
+ int n, pch;

if (!imc->mbase)
return;

- offsets = scrub_err ? res_cfg->offsets_scrub : res_cfg->offsets_demand;
+ if (imc->hbm_mc) {
+ pch = res->cs & 1;
+
+ if (pch)
+ offsets = scrub_err ? res_cfg->offsets_scrub_hbm1 :
+ res_cfg->offsets_demand_hbm1;
+ else
+ offsets = scrub_err ? res_cfg->offsets_scrub_hbm0 :
+ res_cfg->offsets_demand_hbm0;
+ } else {
+ offsets = scrub_err ? res_cfg->offsets_scrub :
+ res_cfg->offsets_demand;
+ }

log0 = I10NM_GET_REG32(imc, res->channel, offsets[0]);
log1 = I10NM_GET_REG32(imc, res->channel, offsets[1]);
@@ -163,10 +195,24 @@ static void show_retry_rd_err_log(struct decoded_addr *res, char *msg,
log0, log1, log2, log3, log4, log5);
}

- corr0 = I10NM_GET_REG32(imc, res->channel, 0x22c18);
- corr1 = I10NM_GET_REG32(imc, res->channel, 0x22c1c);
- corr2 = I10NM_GET_REG32(imc, res->channel, 0x22c20);
- corr3 = I10NM_GET_REG32(imc, res->channel, 0x22c24);
+ if (imc->hbm_mc) {
+ if (pch) {
+ corr0 = I10NM_GET_REG32(imc, res->channel, 0x2c18);
+ corr1 = I10NM_GET_REG32(imc, res->channel, 0x2c1c);
+ corr2 = I10NM_GET_REG32(imc, res->channel, 0x2c20);
+ corr3 = I10NM_GET_REG32(imc, res->channel, 0x2c24);
+ } else {
+ corr0 = I10NM_GET_REG32(imc, res->channel, 0x2818);
+ corr1 = I10NM_GET_REG32(imc, res->channel, 0x281c);
+ corr2 = I10NM_GET_REG32(imc, res->channel, 0x2820);
+ corr3 = I10NM_GET_REG32(imc, res->channel, 0x2824);
+ }
+ } else {
+ corr0 = I10NM_GET_REG32(imc, res->channel, 0x22c18);
+ corr1 = I10NM_GET_REG32(imc, res->channel, 0x22c1c);
+ corr2 = I10NM_GET_REG32(imc, res->channel, 0x22c20);
+ corr3 = I10NM_GET_REG32(imc, res->channel, 0x22c24);
+ }

if (len - n > 0)
snprintf(msg + n, len - n,
@@ -420,7 +466,11 @@ static struct res_config spr_cfg = {
.sad_all_devfn = PCI_DEVFN(10, 0),
.sad_all_offset = 0x300,
.offsets_scrub = offsets_scrub_spr,
+ .offsets_scrub_hbm0 = offsets_scrub_spr_hbm0,
+ .offsets_scrub_hbm1 = offsets_scrub_spr_hbm1,
.offsets_demand = offsets_demand_spr,
+ .offsets_demand_hbm0 = offsets_demand_spr_hbm0,
+ .offsets_demand_hbm1 = offsets_demand_spr_hbm1,
};

static const struct x86_cpu_id i10nm_cpuids[] = {
--
2.35.3