[PATCH] mm: allow deferred page init for vmemmap only

[PATCH] mm: allow deferred page init for vmemmap only

[Pavel Tatashin]

It is unsafe to do virtual to physical translations before mm_init() is
called if struct page is needed in order to determine the memory section
number (see SECTION_IN_PAGE_FLAGS). This is because only in mm_init() we
initialize struct pages for all the allocated memory when deferred struct
pages are used.

My recent fix exposed this problem, because it greatly reduced number of
pages that are initialized before mm_init(), but the problem existed even
before my fix, as Fengguang Wu found.

Since FLATMEM is already disallowed for deferred struct pages, it makes
sense to allow deferred struct pages only on systems with
SPARSEMEM_VMEMMAP.

The problems are discussed in these threads:
http://lkml.kernel.org/r/20180418135300.inazvpxjxowogyge@wfg-t540p.sh.intel.com
http://lkml.kernel.org/r/20180419013128.iurzouiqxvcnpbvz@wfg-t540p.sh.intel.com
http://lkml.kernel.org/r/20180426202619.2768-1-pasha.tatashin@oracle.com

Fixes: 3a80a7fa7989 ("mm: meminit: initialise a subset of struct pages if CONFIG_DEFERRED_STRUCT_PAGE_INIT is set")
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
---
 mm/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/Kconfig b/mm/Kconfig
index d5004d82a1d6..1cd32d67ca30 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -635,7 +635,7 @@ config DEFERRED_STRUCT_PAGE_INIT
 	bool "Defer initialisation of struct pages to kthreads"
 	default n
 	depends on NO_BOOTMEM
-	depends on !FLATMEM
+	depends on SPARSEMEM_VMEMMAP
 	help
 	  Ordinarily all struct pages are initialised during early boot in a
 	  single thread. On very large machines this can take a considerable
-- 
2.17.0

Re: [PATCH] mm: allow deferred page init for vmemmap only

[Michal Hocko]

On Wed 09-05-18 15:17:13, Pavel Tatashin wrote:
> It is unsafe to do virtual to physical translations before mm_init() is
> called if struct page is needed in order to determine the memory section
> number (see SECTION_IN_PAGE_FLAGS). This is because only in mm_init() we
> initialize struct pages for all the allocated memory when deferred struct
> pages are used.
> 
> My recent fix exposed this problem, because it greatly reduced number of
> pages that are initialized before mm_init(), but the problem existed even
> before my fix, as Fengguang Wu found.
> 
> Since FLATMEM is already disallowed for deferred struct pages, it makes
> sense to allow deferred struct pages only on systems with
> SPARSEMEM_VMEMMAP.

This doesn't really explain why CONFIG_SPARSMEM or DISCONTIG has the
problem.

> The problems are discussed in these threads:
> http://lkml.kernel.org/r/20180418135300.inazvpxjxowogyge@wfg-t540p.sh.intel.com
> http://lkml.kernel.org/r/20180419013128.iurzouiqxvcnpbvz@wfg-t540p.sh.intel.com
> http://lkml.kernel.org/r/20180426202619.2768-1-pasha.tatashin@oracle.com

Using references is good but there should be a highlevel summary in the
chaneglog as well.

> Fixes: 3a80a7fa7989 ("mm: meminit: initialise a subset of struct pages if CONFIG_DEFERRED_STRUCT_PAGE_INIT is set")
> Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
> ---
>  mm/Kconfig | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/Kconfig b/mm/Kconfig
> index d5004d82a1d6..1cd32d67ca30 100644
> --- a/mm/Kconfig
> +++ b/mm/Kconfig
> @@ -635,7 +635,7 @@ config DEFERRED_STRUCT_PAGE_INIT
>  	bool "Defer initialisation of struct pages to kthreads"
>  	default n
>  	depends on NO_BOOTMEM
> -	depends on !FLATMEM
> +	depends on SPARSEMEM_VMEMMAP
>  	help
>  	  Ordinarily all struct pages are initialised during early boot in a
>  	  single thread. On very large machines this can take a considerable
> -- 
> 2.17.0
> 

-- 
Michal Hocko
SUSE Labs

Re: [PATCH] mm: allow deferred page init for vmemmap only

[Pavel Tatashin]

> This doesn't really explain why CONFIG_SPARSMEM or DISCONTIG has the
> problem.

Hi Michal,

Thank you for reviewing this patch. I sent out a version two of this patch,
with expanded explanation of the problem.

Thank you,
Pavel