Re: [PATCH v8 2/3] mm/memory_hotplug: split memmap_on_memory requests across memblocks

[Verma, Vishal L]

On Fri, 2023-11-03 at 09:43 -0700, fan wrote:
> On Wed, Nov 01, 2023 at 04:51:52PM -0600, Vishal Verma wrote:
> > 
[..]
> >  
> > +static void __ref remove_memory_blocks_and_altmaps(u64 start, u64 size)
> > +{
> > +       unsigned long memblock_size = memory_block_size_bytes();
> > +       u64 cur_start;
> > +
> > +       /*
> > +        * For memmap_on_memory, the altmaps were added on a per-memblock
> > +        * basis; we have to process each individual memory block.
> > +        */
> > +       for (cur_start = start; cur_start < start + size;
> > +            cur_start += memblock_size) {
> > +               struct vmem_altmap *altmap = NULL;
> > +               struct memory_block *mem;
> > +
> > +               mem = find_memory_block(pfn_to_section_nr(PFN_DOWN(cur_start)));
> > +               WARN_ON_ONCE(!mem);
> > +               if (!mem)
> > +                       continue;
> > +
> > +               altmap = mem->altmap;
> > +               mem->altmap = NULL;
> > +
> > +               remove_memory_block_devices(cur_start, memblock_size);
> 
> Is cur_start always aligned to memory_block_size_bytes? If not, the
> above function will return directly, is that a issue?
> 
Hi Fan,

Thanks for taking a look and the review (btw v9 is the latest revision
of these).

I think we're okay because the create side would've adding this memory
in the first place as it too does an alignment check for
memory_block_size_bytes.

Thanks
Vishal