Re: [PATCH] memblock: Make sure reserved.regions is freed really

[Yinghai Lu]

On Mon, Jun 18, 2012 at 4:09 PM, Tejun Heo <tj@xxxxxxxxxx> wrote:
> Hey, Yinghai.
>
> On Mon, Jun 18, 2012 at 03:58:52PM -0700, Yinghai Lu wrote:
>> > Hmmm... nice catch but I think it's a bit complex and ugly.  In this
>> > case, we *know* that the region isn't gonna be split.  Maybe a better
>> > option is to add something like the following?
>>
>> how do you know __pa(memblock.reserved.regions) would not be merged
>> with other entries
>> in reserved.regions?
>>
>> >
>> > void memblock_remove_region_by_ptr(struct memblock_type *type,
>> >                                   struct memblock_region *r)
>> > {
>> >        WARN_ON(/* make sure @r is inside @type->regions */);
>> >        memblock_remove_region(type, r - type->regions);
>> > }
>>
>> you want to change
>
> Heh, you're right.  My suggestion was completely bonkers.  Sorry about
> that.  Hmm.... how about separating out removal pre-expansion and
> doing it before calling into free?  ie. something like
>
> static int memblock_prepare_for_isolation(struct memblock_type *type)
> {
>        /* we'll create at most two more regions */
>        while (type->cnt + 2 > type->max)
>                if (memblock_double_array(type) < 0)
>                        return -ENOMEM;
>        return 0;
> }
>
> int __init_memblock memblock_free_reserved_regions(void)
> {
>        int ret;
>
>        if (memblock.reserved.regions == memblock_reserved_init_regions)
>                return 0;
>
>        ret = memblock_prepare_for_isolation(&memblock.reserved);
>        if (ret)
>                return ret;
>
>        ret = memblock_free(__pa(memblock.reserved.regions),
>                sizeof(struct memblock_region) * memblock.reserved.max);
>        WARN_ON(ret || memblock.reserved.regions has changed);
>        return ret;
> }

yes. that looks good. but that WARN_ON looks wrong, could just drop that line.

Please send complete version to Andrew or Linus.

Thanks

Yinghai
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/