2021-07-23 12:35:21

by Miles Chen

Subject: [PATCH v3] mm/sparse: clarify pgdat_to_phys

Clarify pgdat_to_phys() by testing if
pgdat == &contig_page_data when CONFIG_NUMA=n.

We only expect contig_page_data in such case, so we
use &contig_page_data directly instead of pgdat.

No functional change intended when CONFIG_BUG_VM=n.

Comment from Mark [1]:
"
... and I reckon it'd be clearer and more robust to define
pgdat_to_phys() in the same ifdefs as contig_page_data so
that these, stay in-sync. e.g. have:

| #ifdef CONFIG_NUMA
| #define pgdat_to_phys(x) virt_to_phys(x)
| #else /* CONFIG_NUMA */
|
| extern struct pglist_data contig_page_data;
| ...
| #define pgdat_to_phys(x) __pa_symbol(&contig_page_data)
|
| #endif /* CONIFIG_NUMA */
"

[1] https://lore.kernel.org/linux-arm-kernel/[email protected]/

Cc: Mike Rapoport <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: David Hildenbrand <[email protected]>
Signed-off-by: Miles Chen <[email protected]>

--

Change since v1:
Thanks for Mike's comment, check if pgdat == &contig_page_data,
so it is clearer that we only expect contig_page_data when
CONFIG_NUMA=n.

Change since v2:
use VM_BUG_ON() to avoid runtime checking when CONFIG_BUG_VM=n
---
mm/sparse.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/mm/sparse.c b/mm/sparse.c
index 6326cdf36c4f..d13d831f88a5 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -348,7 +348,8 @@ size_t mem_section_usage_size(void)
static inline phys_addr_t pgdat_to_phys(struct pglist_data *pgdat)
{
#ifndef CONFIG_NUMA
- return __pa_symbol(pgdat);
+ VM_BUG_ON(pgdat != &contig_page_data);
+ return __pa_symbol(&contig_page_data);
#else
return __pa(pgdat);
#endif
--
2.18.0

2021-07-23 12:56:54

by David Hildenbrand

Subject: Re: [PATCH v3] mm/sparse: clarify pgdat_to_phys

On 23.07.21 14:33, Miles Chen wrote:
> Clarify pgdat_to_phys() by testing if
> pgdat == &contig_page_data when CONFIG_NUMA=n.
>
> We only expect contig_page_data in such case, so we
> use &contig_page_data directly instead of pgdat.
>
> No functional change intended when CONFIG_BUG_VM=n.
>
> Comment from Mark [1]:
> "
> ... and I reckon it'd be clearer and more robust to define
> pgdat_to_phys() in the same ifdefs as contig_page_data so
> that these, stay in-sync. e.g. have:
>
> | #ifdef CONFIG_NUMA
> | #define pgdat_to_phys(x) virt_to_phys(x)
> | #else /* CONFIG_NUMA */
> |
> | extern struct pglist_data contig_page_data;
> | ...
> | #define pgdat_to_phys(x) __pa_symbol(&contig_page_data)
> |
> | #endif /* CONIFIG_NUMA */
> "
>
> [1] https://lore.kernel.org/linux-arm-kernel/[email protected]/
>
> Cc: Mike Rapoport <[email protected]>
> Cc: Mark Rutland <[email protected]>
> Cc: David Hildenbrand <[email protected]>
> Signed-off-by: Miles Chen <[email protected]>
>
> --
>
> Change since v1:
> Thanks for Mike's comment, check if pgdat == &contig_page_data,
> so it is clearer that we only expect contig_page_data when
> CONFIG_NUMA=n.
>
> Change since v2:
> use VM_BUG_ON() to avoid runtime checking when CONFIG_BUG_VM=n
> ---
> mm/sparse.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/mm/sparse.c b/mm/sparse.c
> index 6326cdf36c4f..d13d831f88a5 100644
> --- a/mm/sparse.c
> +++ b/mm/sparse.c
> @@ -348,7 +348,8 @@ size_t mem_section_usage_size(void)
> static inline phys_addr_t pgdat_to_phys(struct pglist_data *pgdat)
> {
> #ifndef CONFIG_NUMA
> - return __pa_symbol(pgdat);
> + VM_BUG_ON(pgdat != &contig_page_data);
> + return __pa_symbol(&contig_page_data);
> #else
> return __pa(pgdat);
> #endif
>

Reviewed-by: David Hildenbrand <[email protected]>

--
Thanks,

David / dhildenb

2021-07-24 07:50:59

by Mike Rapoport

Subject: Re: [PATCH v3] mm/sparse: clarify pgdat_to_phys

On Fri, Jul 23, 2021 at 08:33:42PM +0800, Miles Chen wrote:
> Clarify pgdat_to_phys() by testing if
> pgdat == &contig_page_data when CONFIG_NUMA=n.
>
> We only expect contig_page_data in such case, so we
> use &contig_page_data directly instead of pgdat.
>
> No functional change intended when CONFIG_BUG_VM=n.
>
> Comment from Mark [1]:
> "
> ... and I reckon it'd be clearer and more robust to define
> pgdat_to_phys() in the same ifdefs as contig_page_data so
> that these, stay in-sync. e.g. have:
>
> | #ifdef CONFIG_NUMA
> | #define pgdat_to_phys(x) virt_to_phys(x)
> | #else /* CONFIG_NUMA */
> |
> | extern struct pglist_data contig_page_data;
> | ...
> | #define pgdat_to_phys(x) __pa_symbol(&contig_page_data)
> |
> | #endif /* CONIFIG_NUMA */
> "
>
> [1] https://lore.kernel.org/linux-arm-kernel/[email protected]/
>
> Cc: Mike Rapoport <[email protected]>
> Cc: Mark Rutland <[email protected]>
> Cc: David Hildenbrand <[email protected]>
> Signed-off-by: Miles Chen <[email protected]>

Acked-by: Mike Rapoport <[email protected]>

>
> --
>
> Change since v1:
> Thanks for Mike's comment, check if pgdat == &contig_page_data,
> so it is clearer that we only expect contig_page_data when
> CONFIG_NUMA=n.
>
> Change since v2:
> use VM_BUG_ON() to avoid runtime checking when CONFIG_BUG_VM=n
> ---
> mm/sparse.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/mm/sparse.c b/mm/sparse.c
> index 6326cdf36c4f..d13d831f88a5 100644
> --- a/mm/sparse.c
> +++ b/mm/sparse.c
> @@ -348,7 +348,8 @@ size_t mem_section_usage_size(void)
> static inline phys_addr_t pgdat_to_phys(struct pglist_data *pgdat)
> {
> #ifndef CONFIG_NUMA
> - return __pa_symbol(pgdat);
> + VM_BUG_ON(pgdat != &contig_page_data);
> + return __pa_symbol(&contig_page_data);
> #else
> return __pa(pgdat);
> #endif
> --
> 2.18.0
>

--
Sincerely yours,
Mike.