Re: [PATCH v7 00/12] Support non-lru page migration

[Joonsoo Kim]

On Thu, Jun 16, 2016 at 07:09:32PM +0900, Minchan Kim wrote:
> On Thu, Jun 16, 2016 at 05:42:11PM +0900, Sergey Senozhatsky wrote:
> > On (06/16/16 15:47), Minchan Kim wrote:
> > > > [..]
> > > > > > this is what I'm getting with the [zsmalloc: keep first object offset in struct page]
> > > > > > applied:  "count:0 mapcount:-127". which may be not related to zsmalloc at this point.
> > > > > > 
> > > > > > kernel: BUG: Bad page state in process khugepaged  pfn:101db8
> > > > > > kernel: page:ffffea0004076e00 count:0 mapcount:-127 mapping:          (null) index:0x1
> > > > > 
> > > > > Hm, it seems double free.
> > > > > 
> > > > > It doen't happen if you disable zram? IOW, it seems to be related
> > > > > zsmalloc migration?
> > > > 
> > > > need to test more, can't confidently answer now.
> > > > 
> > > > > How easy can you reprodcue it? Could you bisect it?
> > > > 
> > > > it takes some (um.. random) time to trigger the bug.
> > > > I'll try to come up with more details.
> > > 
> > > Could you revert [1] and retest?
> > > 
> > > [1] mm/compaction: split freepages without holding the zone lock
> > 
> > ok, so this is not related to zsmalloc. finally manged to reproduce
> > it. will fork a separate thread.
> 
> The reason I mentioned [1] is that it seems to have a bug.
> 
> isolate_freepages_block
>   __isolate_free_page
>     if(!zone_watermark_ok())
>       return 0;
>   list_add_tail(&page->lru, freelist);
> 
> However, the page is not isolated.
> Joonsoo?

Good job!
I will fix it soon.

Thanks.