2023-05-11 04:14:50

by Rabara, Niravkumar L

Subject: [PATCH v3 2/2] EDAC/altera: Check previous DDR DBE during driver probe

From: Niravkumar L Rabara <[email protected]>

Add DDR DBE check during driver probe to notify user if previous
reboot cause by DDR DBE and print DBE error related information.

Signed-off-by: Niravkumar L Rabara <[email protected]>
---
drivers/edac/altera_edac.c | 29 ++++++++++++++++++++++++-----
1 file changed, 24 insertions(+), 5 deletions(-)

diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c
index 8b31cd54bdb6..04c0675adc8c 100644
--- a/drivers/edac/altera_edac.c
+++ b/drivers/edac/altera_edac.c
@@ -2159,6 +2159,7 @@ static int altr_edac_a10_probe(struct platform_device *pdev)
#ifdef CONFIG_64BIT
{
int dberror, err_addr;
+ struct arm_smccc_res result;

edac->panic_notifier.notifier_call = s10_edac_dberr_handler;
atomic_notifier_chain_register(&panic_notifier_list,
@@ -2168,11 +2169,29 @@ static int altr_edac_a10_probe(struct platform_device *pdev)
regmap_read(edac->ecc_mgr_map, S10_SYSMGR_UE_VAL_OFST,
&dberror);
if (dberror) {
- regmap_read(edac->ecc_mgr_map, S10_SYSMGR_UE_ADDR_OFST,
- &err_addr);
- edac_printk(KERN_ERR, EDAC_DEVICE,
- "Previous Boot UE detected[0x%X] @ 0x%X\n",
- dberror, err_addr);
+ /* Bit-31 is set if previous DDR UE happened */
+ if (dberror & (1 << 31)) {
+ /* Read previous DDR UE info */
+ arm_smccc_smc(INTEL_SIP_SMC_READ_SEU_ERR, 0,
+ 0, 0, 0, 0, 0, 0, &result);
+
+ if (!result.a0) {
+ edac_printk(KERN_ERR, EDAC_DEVICE,
+ "Previous DDR UE:Count=0x%X,Address=0x%X,ErrorData=0x%X\n"
+ , (unsigned int)result.a1
+ , (unsigned int)result.a2
+ , (unsigned int)result.a3);
+ } else {
+ edac_printk(KERN_ERR, EDAC_DEVICE,
+ "INTEL_SIP_SMC_SEU_ERR_STATUS failed\n");
+ }
+ } else {
+ regmap_read(edac->ecc_mgr_map, S10_SYSMGR_UE_ADDR_OFST,
+ &err_addr);
+ edac_printk(KERN_ERR, EDAC_DEVICE,
+ "Previous Boot UE detected[0x%X] @ 0x%X\n",
+ dberror, err_addr);
+ }
/* Reset the sticky registers */
regmap_write(edac->ecc_mgr_map,
S10_SYSMGR_UE_VAL_OFST, 0);
--
2.25.1

2023-05-23 13:18:16

by dinh.linux

Subject: Re: [PATCH v3 2/2] EDAC/altera: Check previous DDR DBE during driver probe

Hi,

> From: Niravkumar L Rabara <[email protected]>
>
> Add DDR DBE check during driver probe to notify user if previous
> reboot cause by DDR DBE and print DBE error related information.
>
> Signed-off-by: Niravkumar L Rabara <[email protected]>
> ---
> drivers/edac/altera_edac.c | 29 ++++++++++++++++++++++++-----
> 1 file changed, 24 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c
> index 8b31cd54bdb6..04c0675adc8c 100644
> --- a/drivers/edac/altera_edac.c
> +++ b/drivers/edac/altera_edac.c
> @@ -2159,6 +2159,7 @@ static int altr_edac_a10_probe(struct
> platform_device *pdev)
> #ifdef CONFIG_64BIT
> {
> int dberror, err_addr;
> + struct arm_smccc_res result;
>
> edac->panic_notifier.notifier_call = s10_edac_dberr_handler;
> atomic_notifier_chain_register(&panic_notifier_list,
> @@ -2168,11 +2169,29 @@ static int altr_edac_a10_probe(struct
> platform_device *pdev)
> regmap_read(edac->ecc_mgr_map, S10_SYSMGR_UE_VAL_OFST,
> &dberror);
> if (dberror) {
> - regmap_read(edac->ecc_mgr_map, S10_SYSMGR_UE_ADDR_OFST,
> - &err_addr);
> - edac_printk(KERN_ERR, EDAC_DEVICE,
> - "Previous Boot UE detected[0x%X] @ 0x%X\n",
> - dberror, err_addr);
> + /* Bit-31 is set if previous DDR UE happened */
> + if (dberror & (1 << 31)) {
> + /* Read previous DDR UE info */
> + arm_smccc_smc(INTEL_SIP_SMC_READ_SEU_ERR, 0,
> + 0, 0, 0, 0, 0, 0, &result);
> +
> + if (!result.a0) {
> + edac_printk(KERN_ERR, EDAC_DEVICE,
> + "Previous DDR UE:Count=0x%X,Address=0x%X,ErrorData=0x%X\n"
> + , (unsigned int)result.a1
> + , (unsigned int)result.a2
> + , (unsigned int)result.a3);
> + } else {
> + edac_printk(KERN_ERR, EDAC_DEVICE,
> + "INTEL_SIP_SMC_SEU_ERR_STATUS failed\n");
> + }
> + } else {
> + regmap_read(edac->ecc_mgr_map, S10_SYSMGR_UE_ADDR_OFST,
> + &err_addr);
> + edac_printk(KERN_ERR, EDAC_DEVICE,
> + "Previous Boot UE detected[0x%X] @ 0x%X\n",
> + dberror, err_addr);
> + }
> /* Reset the sticky registers */
> regmap_write(edac->ecc_mgr_map,
> S10_SYSMGR_UE_VAL_OFST, 0);
> --
> 2.25.1
>
>

I think it would make sense if you combined the 1st patch. This patch
cannot survive on its own without the defines.

Dinh