On 24.04.24 10:54:22, Dan Williams wrote:
> Robert Richter wrote:
> > Adding a pr_info() when successfully adding a CFMWS memory range.
> >
> > Suggested-by: Alison Schofield <alison.schofield@xxxxxxxxx>
> > Signed-off-by: Robert Richter <rrichter@xxxxxxx>
> > ---
> > drivers/acpi/numa/srat.c | 10 +++++++++-
> > 1 file changed, 9 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/acpi/numa/srat.c b/drivers/acpi/numa/srat.c
> > index e3f26e71637a..c62e4636e472 100644
> > --- a/drivers/acpi/numa/srat.c
> > +++ b/drivers/acpi/numa/srat.c
> > @@ -338,8 +338,11 @@ static int __init acpi_parse_cfmws(union acpi_subtable_headers *header,
> > * found for any portion of the window to cover the entire
> > * window.
> > */
> > - if (!numa_fill_memblks(start, end))
> > + if (!numa_fill_memblks(start, end)) {
> > + pr_info("CEDT: memblk extended [mem %#010Lx-%#010Lx]\n",
> > + (unsigned long long) start, (unsigned long long) end - 1);
>
> This looks like pr_debug() material to me.
This should have the same log level as the message below and/or its
corresponding SRAT message. CEDT mem blocks wouldn't be reported
otherwise only because a smaller (overlapping) entry was registered
before. That is why I used pr_info here.
>
> > return 0;
> > + }
> >
> > /* No SRAT description. Create a new node. */
> > node = acpi_map_pxm_to_node(*fake_pxm);
> > @@ -354,8 +357,13 @@ static int __init acpi_parse_cfmws(union acpi_subtable_headers *header,
> > pr_warn("ACPI NUMA: Failed to add memblk for CFMWS node %d [mem %#llx-%#llx]\n",
> > node, start, end);
> > }
> > +
> > node_set(node, numa_nodes_parsed);
> >
> > + pr_info("CEDT: Node %u PXM %u [mem %#010Lx-%#010Lx]\n",
> > + node, *fake_pxm,
> > + (unsigned long long) start, (unsigned long long) end - 1);
> > +
>
> This makes sense to mirror the SRAT pr_info().
I evaluated this.
SRAT shows this:
pr_info("SRAT: Node %u PXM %u [mem %#010Lx-%#010Lx]%s%s\n",
node, pxm,
(unsigned long long) start, (unsigned long long) end - 1,
hotpluggable ? " hotplug" : "",
ma->flags & ACPI_SRAT_MEM_NON_VOLATILE ? " non-volatile" : "");
There is no direct mapping of SRAT memory affinity flags (acpi-6.5
spec, table 5.59) and something in the CFMWS entry (cxl-3.1, table
9-22). There is no "hotplug" flag and the "non-volatile" part would be
ambiguous, as some logic must be defined to handle the "volatile" and
"persistent" Window Restrictions. Since the CFMWS restrictions are not
used at all by the kernel my conclusion was to just dropped the flag
for the CEDT info.
Note there is a mapping defined for CDAT DSMAS and SRAT entries, see
CDAT 1.03 spec, Table 4.
-Robert
