Re: [PATCH v4] mm: add replace_page_cache_page() function

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Jan 19, 2011 at 8:28 AM, Andrew Morton
<akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Tue, 18 Jan 2011 12:18:11 +0100
> Miklos Szeredi <miklos@xxxxxxxxxx> wrote:
>
>> +int replace_page_cache_page(struct page *old, struct page *new, gfp_t gfp_mask)
>> +{
>> +     int error;
>> +     struct mem_cgroup *memcg = NULL;
>
> I'm suspecting that the unneeded initialisation was added to suppress a
> warning?
>
> I removed it, and didn't get a warning.  I expected to.
>
> Really, uninitialized_var() is better.  It avoids adding extra code
> and, unlike "= 0" it is self-documenting.
>
>> +     VM_BUG_ON(!PageLocked(old));
>> +     VM_BUG_ON(!PageLocked(new));
>> +     VM_BUG_ON(new->mapping);
>> +
>> +     /*
>> +      * This is not page migration, but prepare_migration and
>> +      * end_migration does enough work for charge replacement.
>> +      *
>> +      * In the longer term we probably want a specialized function
>> +      * for moving the charge from old to new in a more efficient
>> +      * manner.
>> +      */
>> +     error = mem_cgroup_prepare_migration(old, new, &memcg, gfp_mask);
>> +     if (error)
>> +             return error;
>> +
>> +     error = radix_tree_preload(gfp_mask & ~__GFP_HIGHMEM);
>> +     if (!error) {
>> +             struct address_space *mapping = old->mapping;
>> +             pgoff_t offset = old->index;
>> +
>> +             page_cache_get(new);
>> +             new->mapping = mapping;
>> +             new->index = offset;
>> +
>> +             spin_lock_irq(&mapping->tree_lock);
>> +             __remove_from_page_cache(old);
>> +             error = radix_tree_insert(&mapping->page_tree, offset, new);
>> +             BUG_ON(error);
>> +             mapping->nrpages++;
>> +             __inc_zone_page_state(new, NR_FILE_PAGES);
>> +             if (PageSwapBacked(new))
>> +                     __inc_zone_page_state(new, NR_SHMEM);
>> +             spin_unlock_irq(&mapping->tree_lock);
>> +             radix_tree_preload_end();
>> +             page_cache_release(old);
>> +             mem_cgroup_end_migration(memcg, old, new, true);
>
> This is all pretty ugly and inefficient.
>
> We call __remove_from_page_cache() which does a radix-tree lookup and
> then fiddles a bunch of accounting things.
>
> Then we immediately do the same radix-tree lookup and then undo the
> accounting changes which we just did.  And we do it in an open-coded
> fashion, thus giving the kernel yet another code site where various
> operations need to be kept in sync.
>
> Would it not be better to do a single radix_tree_lookup_slot(),
> overwrite the pointer therein and just leave all the ancilliary
> accounting unaltered?

I agree single radix_tree_lookup but accounting still is needed since
newpage could be on another zone. What we can remove is just only
mapping->nrpages.


-- 
Kind regards,
Minchan Kim
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [Samba]     [Device Mapper]     [CEPH Development]
  Powered by Linux