Re: [PATCH v3 0/5] HWpoison: further fixes and cleanups

[Aristeu Rozanski]

Hi Oscar, Naoya,

On Mon, Sep 14, 2020 at 12:15:54PM +0200, Oscar Salvador wrote:
> The important bit of this patchset is patch#1, which is a fix to take off
> HWPoison pages off a buddy freelist since it can lead us to having HWPoison
> pages back in the game without no one noticing it.
> So fix it (we did that already for soft_offline_page [1]).
> 
> The other patches are clean-ups and not that important, so if anything,
> consider patch#1 for inclusion.
> 
> [1] https://patchwork.kernel.org/cover/11704083/

I found something strange with your and Naoya's hwpoison rework. We have a
customer with a testcase that basically does:

	p1 = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
	p2 = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);

	madvise(p1, size, MADV_MERGEABLE);
	madvise(p2, size, MADV_MERGEABLE);

	memset(p1, 'a', size);
	memset(p2, 'a', size);

	madvise(p1, size, MADV_SOFT_OFFLINE);

	madvise(p1, size, MADV_UNMERGEABLE);
	madvise(p2, size, MADV_UNMERGEABLE);


where size is about 200,000 pages. It works on a x86_64 box (with and without the
hwpoison rework). On ppc64 boxes (tested 3 different ones with at least 250GB memory)
it fails to take a page off the buddy list (page_handle_poison()/take_page_off_buddy())
(madvise MADV_SOFT_OFFLINE returns -EBUSY). Without the hwpoison rework the test passes.

Possibly related is that ppc64 takes a long time to run this test and according
perf, it spends most of the time clearing pages:

	  17.15%  ksm_poison  [kernel.kallsyms]  [k] copypage_power7
	  13.39%  ksm_poison  [kernel.kallsyms]  [k] clear_user_page
	   8.70%  ksm_poison  libc-2.28.so       [.] __memset_power8
	   8.63%  ksm_poison  [kernel.kallsyms]  [k] opal_return
	   6.04%  ksm_poison  [kernel.kallsyms]  [k] __opal_call
	   2.67%  ksm_poison  [kernel.kallsyms]  [k] opal_call
	   1.52%  ksm_poison  [kernel.kallsyms]  [k] _raw_spin_lock
	   1.45%  ksm_poison  [kernel.kallsyms]  [k] opal_flush_console
	   1.43%  ksm_poison  [unknown]          [k] 0x0000000030005138
	   1.43%  ksm_poison  [kernel.kallsyms]  [k] opal_console_write_buffer_space
	   1.26%  ksm_poison  [kernel.kallsyms]  [k] hvc_console_print
(...)

I've run these tests using mmotm and mmotm with this patchset on top.

Do you know what might be happening here?

-- 
Aristeu