On 07/02/2013 04:38 AM, Borislav Petkov wrote:
On Mon, Jul 01, 2013 at 09:08:59PM +0530, Naveen N. Rao wrote:
If the firmware indicates in GHES error data entry that the error threshold
has exceeded for a corrected error event, then we try to soft-offline the
page. This could be called in interrupt context, so we queue this up similar
to how we handle memory failure scenarios.
Signed-off-by: Naveen N. Rao <[email protected]>
---
drivers/acpi/apei/ghes.c | 12 ++++++++++
include/linux/mm.h | 1 +
mm/memory-failure.c | 53 ++++++++++++++++++++++++++++++----------------
3 files changed, 48 insertions(+), 18 deletions(-)
diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
index fcd7d91..5a630ed 100644
--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -429,6 +429,18 @@ static void ghes_do_proc(struct ghes *ghes,
mem_err);
#endif
#ifdef CONFIG_ACPI_APEI_MEMORY_FAILURE
+ if (sec_sev == GHES_SEV_CORRECTED &&
+ (gdata->flags & CPER_SEC_ERROR_THRESHOLD_EXCEEDED)
&&
+ (mem_err->validation_bits &
CPER_MEM_VALID_PHYSICAL_ADDRESS)) {
+ unsigned long pfn;
+ pfn = mem_err->physical_addr >> PAGE_SHIFT;
+ if (pfn_valid(pfn))
+ soft_memory_failure_queue(pfn, 0, 0);
+ else
+ pr_warning(FW_WARN GHES_PFX
+ "Invalid address in generic error data:
%#lx\n",
+ mem_err->physical_addr);
+ }
Yuck, this looks like BIOS code.
Can we carve out this into a function and do
void function(.. )
{
#ifdef CONFIG_ACPI_APEI_MEMORY_FAILURE
<code at 1st indentation, much more readable>
#endif
}
so that we can nicely call it from ghes_do_proc()?
Sure.
if (sev == GHES_SEV_RECOVERABLE &&
sec_sev == GHES_SEV_RECOVERABLE &&
mem_err->validation_bits &
CPER_MEM_VALID_PHYSICAL_ADDRESS) {
diff --git a/include/linux/mm.h b/include/linux/mm.h
index e0c8528..f9907d2 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1787,6 +1787,7 @@ enum mf_flags {
};
extern int memory_failure(unsigned long pfn, int trapno, int flags);
extern void memory_failure_queue(unsigned long pfn, int trapno, int flags);
+extern void soft_memory_failure_queue(unsigned long pfn, int trapno, int
flags);
extern int unpoison_memory(unsigned long pfn);
extern int sysctl_memory_failure_early_kill;
extern int sysctl_memory_failure_recovery;
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index ceb0c7f..50caefd 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1222,6 +1222,7 @@ struct memory_failure_entry {
unsigned long pfn;
int trapno;
int flags;
+ bool soft_offline;
Why a new bool? This flags int looks nice above. :)
D'uh! I considered that, but I can't recall why I chose not to use that!
Let me redo this patch.
Thanks,
Naveen
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
