Re: [PATCH v2 1/3] powerpc/pseries: Parse control memory access error

On 8/24/21 12:09 PM, Michael Ellerman wrote:

Hi Ganesh,

Some comments below ...

Ganesh Goudar [off-list ref] writes:

quoted

Add support to parse and log control memory access
error for pseries.

Signed-off-by: Ganesh Goudar <redacted>
---
v2: No changes in this patch.
---
  arch/powerpc/platforms/pseries/ras.c | 21 +++++++++++++++++++++
  1 file changed, 21 insertions(+)

diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
index 167f2e1b8d39..608c35cad0c3 100644
--- a/arch/powerpc/platforms/pseries/ras.c
+++ b/arch/powerpc/platforms/pseries/ras.c

@@ -80,6 +80,7 @@ struct pseries_mc_errorlog {
  #define MC_ERROR_TYPE_TLB		0x04
  #define MC_ERROR_TYPE_D_CACHE		0x05
  #define MC_ERROR_TYPE_I_CACHE		0x07
+#define MC_ERROR_TYPE_CTRL_MEM_ACCESS	0x08

...

quoted

  
+#define MC_ERROR_CTRL_MEM_ACCESS_PTABLE_WALK	0
+#define MC_ERROR_CTRL_MEM_ACCESS_OP_ACCESS	1

Where do the above values come from?

It is from latest PAPR that added support for control memory error.

quoted

+
  static inline u8 rtas_mc_error_sub_type(const struct pseries_mc_errorlog *mlog)
  {
  	switch (mlog->error_type) {

@@ -112,6 +116,8 @@ static inline u8 rtas_mc_error_sub_type(const struct pseries_mc_errorlog *mlog)
  	case	MC_ERROR_TYPE_ERAT:
  	case	MC_ERROR_TYPE_TLB:
  		return (mlog->sub_err_type & 0x03);
+	case	MC_ERROR_TYPE_CTRL_MEM_ACCESS:
+		return (mlog->sub_err_type & 0x70) >> 4;

Can you add to the comment above sub_err_type explaining what these bits are.

Sure, for other errors it is explained in pseries_mc_errorlog definition, ill add it there.

quoted

  	default:
  		return 0;
  	}

@@ -699,6 +705,21 @@ static int mce_handle_err_virtmode(struct pt_regs *regs,
  	case MC_ERROR_TYPE_I_CACHE:
  		mce_err.error_type = MCE_ERROR_TYPE_ICACHE;
  		break;
+	case MC_ERROR_TYPE_CTRL_MEM_ACCESS:
+		mce_err.error_type = MCE_ERROR_TYPE_RA;
+		if (mce_log->sub_err_type & 0x80)

This appears many times in the file.

Can we add eg. MC_EFFECTIVE_ADDR_PROVIDED?

ok, thanks.

quoted

+			eaddr = be64_to_cpu(mce_log->effective_address);
+		switch (err_sub_type) {
+		case MC_ERROR_CTRL_MEM_ACCESS_PTABLE_WALK:
+			mce_err.u.ra_error_type =
+				MCE_RA_ERROR_PAGE_TABLE_WALK_LOAD_STORE_FOREIGN;

That name is ridiculously long, but I guess that's not your fault :)
We can fix it up in a later patch.

quoted

+			break;
+		case MC_ERROR_CTRL_MEM_ACCESS_OP_ACCESS:
+			mce_err.u.ra_error_type =
+				MCE_RA_ERROR_LOAD_STORE_FOREIGN;
+			break;
+		}
+		break;

cheers