The ghes driver reports errors with 'unknown label' even if the actual
DIMM label is known, e.g.:
EDAC MC0: 1 CE Single-bit ECC on unknown label (node:0 card:0
module:0 rank:1 bank:0 col:13 bit_pos:16 DIMM location:N0 DIMM_A0
page:0x966a9b3 offset:0x0 grain:1 syndrome:0x0 - APEI location:
node:0 card:0 module:0 rank:1 bank:0 col:13 bit_pos:16 DIMM
location:N0 DIMM_A0 status(0x0000000000000400): Storage error in
DRAM memory)
Fix this by using struct dimm_info's label string in error reports:
EDAC MC0: 1 CE Single-bit ECC on N0 DIMM_A0 (node:0 card:0 module:0
rank:1 bank:515 col:14 bit_pos:16 DIMM location:N0 DIMM_A0
page:0x99223d8 offset:0x0 grain:1 syndrome:0x0 - APEI location:
node:0 card:0 module:0 rank:1 bank:515 col:14 bit_pos:16 DIMM
location:N0 DIMM_A0 status(0x0000000000000400): Storage error in
DRAM memory)
The labels are initialized by reading the bank and device strings from
DMI. Now, the label information can also read from sysfs. E.g. a
ThunderX2 system will show the following:
/sys/devices/system/edac/mc/mc0/dimm0/dimm_label:N0 DIMM_A0
/sys/devices/system/edac/mc/mc0/dimm1/dimm_label:N0 DIMM_B0
/sys/devices/system/edac/mc/mc0/dimm2/dimm_label:N0 DIMM_C0
/sys/devices/system/edac/mc/mc0/dimm3/dimm_label:N0 DIMM_D0
/sys/devices/system/edac/mc/mc0/dimm4/dimm_label:N0 DIMM_E0
/sys/devices/system/edac/mc/mc0/dimm5/dimm_label:N0 DIMM_F0
/sys/devices/system/edac/mc/mc0/dimm6/dimm_label:N0 DIMM_G0
/sys/devices/system/edac/mc/mc0/dimm7/dimm_label:N0 DIMM_H0
/sys/devices/system/edac/mc/mc0/dimm8/dimm_label:N1 DIMM_I0
/sys/devices/system/edac/mc/mc0/dimm9/dimm_label:N1 DIMM_J0
/sys/devices/system/edac/mc/mc0/dimm10/dimm_label:N1 DIMM_K0
/sys/devices/system/edac/mc/mc0/dimm11/dimm_label:N1 DIMM_L0
/sys/devices/system/edac/mc/mc0/dimm12/dimm_label:N1 DIMM_M0
/sys/devices/system/edac/mc/mc0/dimm13/dimm_label:N1 DIMM_N0
/sys/devices/system/edac/mc/mc0/dimm14/dimm_label:N1 DIMM_O0
/sys/devices/system/edac/mc/mc0/dimm15/dimm_label:N1 DIMM_P0
Since dimm_labels can be rewritten, that label will be used in a later
error report:
# echo foobar >/sys/devices/system/edac/mc/mc0/dimm0/dimm_label
# # some error injection here
# dmesg | grep foobar
[ 751.383533] EDAC MC0: 1 CE Single-bit ECC on foobar (node:0 card:0
module:0 rank:1 bank:259 col:3 bit_pos:16 DIMM location:N0 DIMM_A0
page:0x8c8dc74 offset:0x0 grain:1 syndrome:0x0 - APEI location:
node:0 card:0 module:0 rank:1 bank:259 col:3 bit_pos:16 DIMM
location:N0 DIMM_A0 status(0x0000000000000400): Storage error in DRAM
memory)
Signed-off-by: Robert Richter <[email protected]>
---
v4:
* dimm->label: Only update dimm->label in if bank/device is found in
the SMBIOS table, this keeps current behavior for machines that do
not provide this information.
* e->location: Keep current behavior how e->location is written.
* e->label: Use dimm->label if a DIMM was found by its handle and
"unknown memory" otherwise. This aligns with the edac_mc
implementation.
Signed-off-by: Robert Richter <[email protected]>
---
drivers/edac/ghes_edac.c | 37 ++++++++++++++++++++++++++-----------
1 file changed, 26 insertions(+), 11 deletions(-)
diff --git a/drivers/edac/ghes_edac.c b/drivers/edac/ghes_edac.c
index cb3dab56a875..9a6a055ab624 100644
--- a/drivers/edac/ghes_edac.c
+++ b/drivers/edac/ghes_edac.c
@@ -87,16 +87,29 @@ static void ghes_edac_count_dimms(const struct dmi_header *dh, void *arg)
(*num_dimm)++;
}
-static int get_dimm_smbios_index(struct mem_ctl_info *mci, u16 handle)
+static struct dimm_info *find_dimm_by_handle(struct mem_ctl_info *mci, u16 handle)
{
struct dimm_info *dimm;
mci_for_each_dimm(mci, dimm) {
if (dimm->smbios_handle == handle)
- return dimm->idx;
+ return dimm;
}
- return -1;
+ return NULL;
+}
+
+static void dimm_setup_label(struct dimm_info *dimm, u16 handle)
+{
+ const char *bank = NULL, *device = NULL;
+
+ dmi_memdev_name(handle, &bank, &device);
+
+ /* both strings must be non-zero */
+ if (bank && *bank && device && *device) {
+ snprintf(dimm->label, sizeof(dimm->label),
+ "%s %s", bank, device);
+ }
}
static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg)
@@ -179,9 +192,7 @@ static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg)
dimm->dtype = DEV_UNKNOWN;
dimm->grain = 128; /* Likely, worse case */
- /*
- * FIXME: It shouldn't be hard to also fill the DIMM labels
- */
+ dimm_setup_label(dimm, entry->handle);
if (dimm->nr_pages) {
edac_dbg(1, "DIMM%i: %s size = %d MB%s\n",
@@ -228,7 +239,6 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
memset(e, 0, sizeof (*e));
e->error_count = 1;
e->grain = 1;
- strcpy(e->label, "unknown label");
e->msg = pvt->msg;
e->other_detail = pvt->other_detail;
e->top_layer = -1;
@@ -345,7 +355,7 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
p += sprintf(p, "bit_pos:%d ", mem_err->bit_pos);
if (mem_err->validation_bits & CPER_MEM_VALID_MODULE_HANDLE) {
const char *bank = NULL, *device = NULL;
- int index = -1;
+ struct dimm_info *dimm;
dmi_memdev_name(mem_err->mem_dev_handle, &bank, &device);
if (bank != NULL && device != NULL)
@@ -354,13 +364,18 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
p += sprintf(p, "DIMM DMI handle: 0x%.4x ",
mem_err->mem_dev_handle);
- index = get_dimm_smbios_index(mci, mem_err->mem_dev_handle);
- if (index >= 0)
- e->top_layer = index;
+ dimm = find_dimm_by_handle(mci, mem_err->mem_dev_handle);
+ if (dimm) {
+ e->top_layer = dimm->idx;
+ strcpy(e->label, dimm->label);
+ }
}
if (p > e->location)
*(p - 1) = '\0';
+ if (!*e->label)
+ strcpy(e->label, "unknown memory");
+
/* All other fields are mapped on e->other_detail */
p = pvt->other_detail;
p += snprintf(p, sizeof(pvt->other_detail),
--
2.20.1
On Thu, May 28, 2020 at 12:13:06PM +0200, Robert Richter wrote:
> The ghes driver reports errors with 'unknown label' even if the actual
> DIMM label is known, e.g.:
>
> EDAC MC0: 1 CE Single-bit ECC on unknown label (node:0 card:0
> module:0 rank:1 bank:0 col:13 bit_pos:16 DIMM location:N0 DIMM_A0
> page:0x966a9b3 offset:0x0 grain:1 syndrome:0x0 - APEI location:
> node:0 card:0 module:0 rank:1 bank:0 col:13 bit_pos:16 DIMM
> location:N0 DIMM_A0 status(0x0000000000000400): Storage error in
> DRAM memory)
>
> Fix this by using struct dimm_info's label string in error reports:
>
> EDAC MC0: 1 CE Single-bit ECC on N0 DIMM_A0 (node:0 card:0 module:0
> rank:1 bank:515 col:14 bit_pos:16 DIMM location:N0 DIMM_A0
> page:0x99223d8 offset:0x0 grain:1 syndrome:0x0 - APEI location:
> node:0 card:0 module:0 rank:1 bank:515 col:14 bit_pos:16 DIMM
> location:N0 DIMM_A0 status(0x0000000000000400): Storage error in
> DRAM memory)
>
> The labels are initialized by reading the bank and device strings from
> DMI. Now, the label information can also read from sysfs. E.g. a
> ThunderX2 system will show the following:
>
> /sys/devices/system/edac/mc/mc0/dimm0/dimm_label:N0 DIMM_A0
> /sys/devices/system/edac/mc/mc0/dimm1/dimm_label:N0 DIMM_B0
> /sys/devices/system/edac/mc/mc0/dimm2/dimm_label:N0 DIMM_C0
> /sys/devices/system/edac/mc/mc0/dimm3/dimm_label:N0 DIMM_D0
> /sys/devices/system/edac/mc/mc0/dimm4/dimm_label:N0 DIMM_E0
> /sys/devices/system/edac/mc/mc0/dimm5/dimm_label:N0 DIMM_F0
> /sys/devices/system/edac/mc/mc0/dimm6/dimm_label:N0 DIMM_G0
> /sys/devices/system/edac/mc/mc0/dimm7/dimm_label:N0 DIMM_H0
> /sys/devices/system/edac/mc/mc0/dimm8/dimm_label:N1 DIMM_I0
> /sys/devices/system/edac/mc/mc0/dimm9/dimm_label:N1 DIMM_J0
> /sys/devices/system/edac/mc/mc0/dimm10/dimm_label:N1 DIMM_K0
> /sys/devices/system/edac/mc/mc0/dimm11/dimm_label:N1 DIMM_L0
> /sys/devices/system/edac/mc/mc0/dimm12/dimm_label:N1 DIMM_M0
> /sys/devices/system/edac/mc/mc0/dimm13/dimm_label:N1 DIMM_N0
> /sys/devices/system/edac/mc/mc0/dimm14/dimm_label:N1 DIMM_O0
> /sys/devices/system/edac/mc/mc0/dimm15/dimm_label:N1 DIMM_P0
>
> Since dimm_labels can be rewritten, that label will be used in a later
> error report:
>
> # echo foobar >/sys/devices/system/edac/mc/mc0/dimm0/dimm_label
> # # some error injection here
> # dmesg | grep foobar
> [ 751.383533] EDAC MC0: 1 CE Single-bit ECC on foobar (node:0 card:0
> module:0 rank:1 bank:259 col:3 bit_pos:16 DIMM location:N0 DIMM_A0
> page:0x8c8dc74 offset:0x0 grain:1 syndrome:0x0 - APEI location:
> node:0 card:0 module:0 rank:1 bank:259 col:3 bit_pos:16 DIMM
> location:N0 DIMM_A0 status(0x0000000000000400): Storage error in DRAM
> memory)
>
> Signed-off-by: Robert Richter <[email protected]>
> ---
> v4:
>
> * dimm->label: Only update dimm->label in if bank/device is found in
> the SMBIOS table, this keeps current behavior for machines that do
> not provide this information.
>
> * e->location: Keep current behavior how e->location is written.
>
> * e->label: Use dimm->label if a DIMM was found by its handle and
> "unknown memory" otherwise. This aligns with the edac_mc
> implementation.
>
> Signed-off-by: Robert Richter <[email protected]>
> ---
> drivers/edac/ghes_edac.c | 37 ++++++++++++++++++++++++++-----------
> 1 file changed, 26 insertions(+), 11 deletions(-)
Yap, looks good. I'll queue it after the merge window.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
On 02.06.20 17:48:43, Borislav Petkov wrote:
> On Thu, May 28, 2020 at 12:13:06PM +0200, Robert Richter wrote:
> > v4:
> >
> > * dimm->label: Only update dimm->label in if bank/device is found in
> > the SMBIOS table, this keeps current behavior for machines that do
> > not provide this information.
> >
> > * e->location: Keep current behavior how e->location is written.
> >
> > * e->label: Use dimm->label if a DIMM was found by its handle and
> > "unknown memory" otherwise. This aligns with the edac_mc
> > implementation.
> >
> > Signed-off-by: Robert Richter <[email protected]>
> > ---
> > drivers/edac/ghes_edac.c | 37 ++++++++++++++++++++++++++-----------
> > 1 file changed, 26 insertions(+), 11 deletions(-)
>
> Yap, looks good. I'll queue it after the merge window.
Great, thanks.
-Robert