在 2024/2/10 14:49, Dan Williams 写道:
Shiyang Ruan wrote:
When a poison event is received, driver uses GET_POISON_LIST command
to get the poison list. Now driver has cxl_mem_get_poison(), so
reuse it and add a parameter 'bool report', report poison record to MCE
if set true.
If the memory error record has the poison event, why does the poison
list need to be retrieved by the kernel? I would expect it is sufficient
to just report the single poison event and leave it to userspace to
react to that event and retrieve more data if it wants.
The GMER has only physical address field, no range/length of the POISON,
we can't get the poison range from the single event record. Since the
POISON range is injected by one command, one GMER is sent to driver, we
have to use GET_POISON_LIST command to get the length.
--
Thanks,
Ruan.
