Capacity is stranded when CFMWS regions are not aligned to block size.
On x86, block size increases with capacity (2G blocks @ 64G capacity).
Use CFMWS base/size to report memory block size alignment advice.
Suggested-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Gregory Price <gourry@gourry.net>
---
drivers/acpi/numa/srat.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/acpi/numa/srat.c b/drivers/acpi/numa/srat.c
index 44f91f2c6c5d..a24aff38c465 100644
--- a/drivers/acpi/numa/srat.c
+++ b/drivers/acpi/numa/srat.c
@@ -14,6 +14,7 @@
#include <linux/errno.h>
#include <linux/acpi.h>
#include <linux/memblock.h>
+#include <linux/memory.h>
#include <linux/numa.h>
#include <linux/nodemask.h>
#include <linux/topology.h>
@@ -338,12 +339,26 @@ static int __init acpi_parse_cfmws(union acpi_subtable_headers *header,
{
struct acpi_cedt_cfmws *cfmws;
int *fake_pxm = arg;
- u64 start, end;
+ u64 start, end, align, size;
int node;
cfmws = (struct acpi_cedt_cfmws *)header;
start = cfmws->base_hpa;
- end = cfmws->base_hpa + cfmws->window_size;
+ size = cfmws->window_size;
+ end = cfmws->base_hpa + size;
+
+ /* Align memblock size to CFMW regions if possible */
+ for (align = SZ_64T; align >= SZ_256M; align >>= 1) {
+ if (IS_ALIGNED(start, align) && IS_ALIGNED(size, align))
+ break;
+ }
+
+ if (align >= SZ_256M) {
+ if (memory_block_advise_max_size(align) < 0)
+ pr_warn("CFMWS: memblock size advise failed\n");
+ } else {
+ pr_err("CFMWS: [BIOS BUG] base/size alignment violates spec\n");
+ }
/*
* The SRAT may have already described NUMA details for all,
--
2.43.0
On 29.10.24 21:20, Gregory Price wrote:
> Capacity is stranded when CFMWS regions are not aligned to block size.
> On x86, block size increases with capacity (2G blocks @ 64G capacity).
>
> Use CFMWS base/size to report memory block size alignment advice.
>
> Suggested-by: Dan Williams <dan.j.williams@intel.com>
> Signed-off-by: Gregory Price <gourry@gourry.net>
> ---
> drivers/acpi/numa/srat.c | 19 +++++++++++++++++--
> 1 file changed, 17 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/acpi/numa/srat.c b/drivers/acpi/numa/srat.c
> index 44f91f2c6c5d..a24aff38c465 100644
> --- a/drivers/acpi/numa/srat.c
> +++ b/drivers/acpi/numa/srat.c
> @@ -14,6 +14,7 @@
> #include <linux/errno.h>
> #include <linux/acpi.h>
> #include <linux/memblock.h>
> +#include <linux/memory.h>
> #include <linux/numa.h>
> #include <linux/nodemask.h>
> #include <linux/topology.h>
> @@ -338,12 +339,26 @@ static int __init acpi_parse_cfmws(union acpi_subtable_headers *header,
> {
> struct acpi_cedt_cfmws *cfmws;
> int *fake_pxm = arg;
> - u64 start, end;
> + u64 start, end, align, size;
> int node;
>
> cfmws = (struct acpi_cedt_cfmws *)header;
> start = cfmws->base_hpa;
> - end = cfmws->base_hpa + cfmws->window_size;
> + size = cfmws->window_size;
> + end = cfmws->base_hpa + size;
> +
> + /* Align memblock size to CFMW regions if possible */
> + for (align = SZ_64T; align >= SZ_256M; align >>= 1) {
> + if (IS_ALIGNED(start, align) && IS_ALIGNED(size, align))
> + break;
> + }
Are there maybe some nice tricks bi-tricks to avoid the loop and these
hardcoded limits? :)
align = 1UL << __ffs(start | end));
Assuming "unsigned long" is sufficient in this code (64bit) and "start |
end" will never be 0.
--
Cheers,
David / dhildenb
On Wed, Oct 30, 2024 at 11:40:08AM +0100, David Hildenbrand wrote:
> On 29.10.24 21:20, Gregory Price wrote:
> > Capacity is stranded when CFMWS regions are not aligned to block size.
> > On x86, block size increases with capacity (2G blocks @ 64G capacity).
> >
> > Use CFMWS base/size to report memory block size alignment advice.
> >
> > Suggested-by: Dan Williams <dan.j.williams@intel.com>
> > Signed-off-by: Gregory Price <gourry@gourry.net>
> > ---
> > drivers/acpi/numa/srat.c | 19 +++++++++++++++++--
> > 1 file changed, 17 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/acpi/numa/srat.c b/drivers/acpi/numa/srat.c
> > index 44f91f2c6c5d..a24aff38c465 100644
> > --- a/drivers/acpi/numa/srat.c
> > +++ b/drivers/acpi/numa/srat.c
> > @@ -14,6 +14,7 @@
> > #include <linux/errno.h>
> > #include <linux/acpi.h>
> > #include <linux/memblock.h>
> > +#include <linux/memory.h>
> > #include <linux/numa.h>
> > #include <linux/nodemask.h>
> > #include <linux/topology.h>
> > @@ -338,12 +339,26 @@ static int __init acpi_parse_cfmws(union acpi_subtable_headers *header,
> > {
> > struct acpi_cedt_cfmws *cfmws;
> > int *fake_pxm = arg;
> > - u64 start, end;
> > + u64 start, end, align, size;
> > int node;
> > cfmws = (struct acpi_cedt_cfmws *)header;
> > start = cfmws->base_hpa;
> > - end = cfmws->base_hpa + cfmws->window_size;
> > + size = cfmws->window_size;
> > + end = cfmws->base_hpa + size;
> > +
> > + /* Align memblock size to CFMW regions if possible */
> > + for (align = SZ_64T; align >= SZ_256M; align >>= 1) {
> > + if (IS_ALIGNED(start, align) && IS_ALIGNED(size, align))
> > + break;
> > + }
>
> Are there maybe some nice tricks bi-tricks to avoid the loop and these
> hardcoded limits? :)
>
> align = 1UL << __ffs(start | end));
>
> Assuming "unsigned long" is sufficient in this code (64bit) and "start |
> end" will never be 0.
>
This will work, if start | end is < 256MB, the ACPI table is invalid by
definition since either the block itself is <256MB or the size is 0 (which
is nonsense). So yeah i can simplify here.
Ack. will push v5 once i get KLP to clear another warning.
> --
> Cheers,
>
> David / dhildenb
>
On Wed, Oct 30, 2024 at 11:40:08AM +0100, David Hildenbrand wrote:
> On 29.10.24 21:20, Gregory Price wrote:
> > Capacity is stranded when CFMWS regions are not aligned to block size.
> > On x86, block size increases with capacity (2G blocks @ 64G capacity).
> >
> > Use CFMWS base/size to report memory block size alignment advice.
> >
> > Suggested-by: Dan Williams <dan.j.williams@intel.com>
> > Signed-off-by: Gregory Price <gourry@gourry.net>
> > ---
> > drivers/acpi/numa/srat.c | 19 +++++++++++++++++--
> > 1 file changed, 17 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/acpi/numa/srat.c b/drivers/acpi/numa/srat.c
> > index 44f91f2c6c5d..a24aff38c465 100644
> > --- a/drivers/acpi/numa/srat.c
> > +++ b/drivers/acpi/numa/srat.c
> > @@ -14,6 +14,7 @@
> > #include <linux/errno.h>
> > #include <linux/acpi.h>
> > #include <linux/memblock.h>
> > +#include <linux/memory.h>
> > #include <linux/numa.h>
> > #include <linux/nodemask.h>
> > #include <linux/topology.h>
> > @@ -338,12 +339,26 @@ static int __init acpi_parse_cfmws(union acpi_subtable_headers *header,
> > {
> > struct acpi_cedt_cfmws *cfmws;
> > int *fake_pxm = arg;
> > - u64 start, end;
> > + u64 start, end, align, size;
> > int node;
> > cfmws = (struct acpi_cedt_cfmws *)header;
> > start = cfmws->base_hpa;
> > - end = cfmws->base_hpa + cfmws->window_size;
> > + size = cfmws->window_size;
> > + end = cfmws->base_hpa + size;
> > +
> > + /* Align memblock size to CFMW regions if possible */
> > + for (align = SZ_64T; align >= SZ_256M; align >>= 1) {
> > + if (IS_ALIGNED(start, align) && IS_ALIGNED(size, align))
> > + break;
> > + }
>
> Are there maybe some nice tricks bi-tricks to avoid the loop and these
> hardcoded limits? :)
>
> align = 1UL << __ffs(start | end));
>
> Assuming "unsigned long" is sufficient in this code (64bit) and "start |
> end" will never be 0.
>
I don't think 0 itself is necessarily invalid, but it would be strange.
I can look a bit.
> --
> Cheers,
>
> David / dhildenb
>
© 2016 - 2026 Red Hat, Inc.