On Wed, Apr 17, 2024 at 03:50:52PM +0800, Shiyang Ruan wrote:
> The length of Physical Address in General Media Event Record/DRAM Event
> Record is 64-bit, so the field mask should be defined as such length.
> Otherwise, this causes cxl_general_media and cxl_dram tracepoints to
> mask off the upper-32-bits of DPA addresses. The cxl_poison event is
> unaffected.
>
> If userspace was doing its own DPA-to-HPA translation this could lead to
> incorrect page retirement decisions, but there is no known consumer
> (like rasdaemon) of this event today.
>
So, an invalid DPA is emitted in the trace event log and that could
lead to 'incorrect page retirement decisions...'
> Fixes: d54a531a430b ("cxl/mem: Trace General Media Event Record")
> Cc: <stable@xxxxxxxxxxxxxxx>
> Cc: Dan Williams <dan.j.williams@xxxxxxxxx>
> Cc: Davidlohr Bueso <dave@xxxxxxxxxxxx>
> Cc: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> Cc: Ira Weiny <ira.weiny@xxxxxxxxx>
> Signed-off-by: Shiyang Ruan <ruansy.fnst@xxxxxxxxxxx>
> ---
> drivers/cxl/core/trace.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/cxl/core/trace.h b/drivers/cxl/core/trace.h
> index e5f13260fc52..cdfce932d5b1 100644
> --- a/drivers/cxl/core/trace.h
> +++ b/drivers/cxl/core/trace.h
> @@ -253,7 +253,7 @@ TRACE_EVENT(cxl_generic_event,
> * DRAM Event Record
> * CXL rev 3.0 section 8.2.9.2.1.2; Table 8-44
> */
> -#define CXL_DPA_FLAGS_MASK 0x3F
> +#define CXL_DPA_FLAGS_MASK 0x3FULL
> #define CXL_DPA_MASK (~CXL_DPA_FLAGS_MASK)
>
> #define CXL_DPA_VOLATILE BIT(0)
This works but I'm thinking this is the time to convene on one
CXL_EVENT_DPA_MASK for both all CXL events, rather than having
cxl_poison event be different.
I prefer how poison defines it:
cxlmem.h:#define CXL_POISON_START_MASK GENMASK_ULL(63, 6)
Can we rename that CXL_EVENT_DPA_MASK and use for all events?
--Alison
> --
> 2.34.1
>
