Re: [PATCH 1/7] mm: move array mem_section init code out of memory_present()

[Baoquan He <bhe@redhat.com>]

On 04/02/24 at 11:19am, Mike Rapoport wrote:
> On Tue, Mar 26, 2024 at 02:11:27PM +0800, Baoquan He wrote:
> > When CONFIG_SPARSEMEM_EXTREME is enabled, mem_section need be initialized
> > to point at a two-dimensional array, and its 1st dimension of length
> > NR_SECTION_ROOTS will be dynamically allocated. Once the allocation is
> > done, it's available for all nodes.
> > 
> > So take the 1st dimension of mem_section initialization out of
> > memory_present()(), and put it into memblocks_present() which is a more
> > appripriate place.
> > 
> > Signed-off-by: Baoquan He <bhe@redhat.com>
> > ---
> >  mm/sparse.c | 26 +++++++++++++-------------
> >  1 file changed, 13 insertions(+), 13 deletions(-)
> > 
> > diff --git a/mm/sparse.c b/mm/sparse.c
> > index aed0951b87fa..46e88549d1a6 100644
> > --- a/mm/sparse.c
> > +++ b/mm/sparse.c
> > @@ -226,19 +226,6 @@ static void __init memory_present(int nid, unsigned long start, unsigned long en
> >  {
> >  	unsigned long pfn;
> >  
> > -#ifdef CONFIG_SPARSEMEM_EXTREME
> > -	if (unlikely(!mem_section)) {
> > -		unsigned long size, align;
> > -
> > -		size = sizeof(struct mem_section *) * NR_SECTION_ROOTS;
> > -		align = 1 << (INTERNODE_CACHE_SHIFT);
> > -		mem_section = memblock_alloc(size, align);
> > -		if (!mem_section)
> > -			panic("%s: Failed to allocate %lu bytes align=0x%lx\n",
> > -			      __func__, size, align);
> > -	}
> > -#endif
> > -
> >  	start &= PAGE_SECTION_MASK;
> >  	mminit_validate_memmodel_limits(&start, &end);
> >  	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION) {
> > @@ -267,6 +254,19 @@ static void __init memblocks_present(void)
> >  	unsigned long start, end;
> >  	int i, nid;
> >  
> > +#ifdef CONFIG_SPARSEMEM_EXTREME
> > +	if (unlikely(!mem_section)) {
> 
> With the new calling sequence sparse_init() -> memblocks_present() ->
> allocate mem_section, mem_section here will be NULL, so this if is not
> really relevant. We might want to add WARN_ON_ONCE(mem_section) just in
> case to catch multiple calls to sparse_init().

Sorry for late reply.

Hmm, I am not very eager to add WARN_ON_ONCE(mem_section) becasue as you
said it's not very relevant. Calling sparse_init() miltiple times is not
acceptable, while it may not be good to let checking mem_section in
memblocks_present() to catch it.

> 
> > +		unsigned long size, align;
> > +
> > +		size = sizeof(struct mem_section *) * NR_SECTION_ROOTS;
> > +		align = 1 << (INTERNODE_CACHE_SHIFT);
> > +		mem_section = memblock_alloc(size, align);
> > +		if (!mem_section)
> > +			panic("%s: Failed to allocate %lu bytes align=0x%lx\n",
> > +			      __func__, size, align);
> > +	}
> > +#endif
> 
> > +
> >  	for_each_mem_pfn_range(i, MAX_NUMNODES, &start, &end, &nid)
> >  		memory_present(nid, start, end);
> >  }
> > -- 
> > 2.41.0
> > 
> 
> -- 
> Sincerely yours,
> Mike.
>