[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20221013100943.GA1505152@u2004>
Date: Thu, 13 Oct 2022 19:09:43 +0900
From: Naoya Horiguchi <naoya.horiguchi@...ux.dev>
To: Oscar Salvador <osalvador@...e.de>
Cc: linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>,
Miaohe Lin <linmiaohe@...wei.com>,
David Hildenbrand <david@...hat.com>,
Mike Kravetz <mike.kravetz@...cle.com>,
Yang Shi <shy828301@...il.com>,
Muchun Song <songmuchun@...edance.com>,
Jane Chu <jane.chu@...cle.com>,
Naoya Horiguchi <naoya.horiguchi@....com>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v6 4/4] mm/hwpoison: introduce per-memory_block hwpoison
counter
On Thu, Oct 13, 2022 at 10:33:45AM +0200, Oscar Salvador wrote:
> On Fri, Oct 07, 2022 at 10:07:06AM +0900, Naoya Horiguchi wrote:
> > From: Naoya Horiguchi <naoya.horiguchi@....com>
> >
> > Currently PageHWPoison flag does not behave well when experiencing memory
> > hotremove/hotplug. Any data field in struct page is unreliable when the
> > associated memory is offlined, and the current mechanism can't tell whether
> > a memory block is onlined because a new memory devices is installed or
> > because previous failed offline operations are undone. Especially if
> > there's a hwpoisoned memory, it's unclear what the best option is.
> >
> > So introduce a new mechanism to make struct memory_block remember that
> > a memory block has hwpoisoned memory inside it. And make any online event
> > fail if the onlining memory block contains hwpoison. struct memory_block
> > is freed and reallocated over ACPI-based hotremove/hotplug, but not over
> > sysfs-based hotremove/hotplug. So the new counter can distinguish these
> > cases.
> >
> > Signed-off-by: Naoya Horiguchi <naoya.horiguchi@....com>
> > Reported-by: kernel test robot <lkp@...el.com>
>
> I glanzed over it and looks good overall.
> Have a small question though:
Thank you for looking.
>
> > @@ -864,6 +878,7 @@ void remove_memory_block_devices(unsigned long start, unsigned long size)
> > mem = find_memory_block_by_id(block_id);
> > if (WARN_ON_ONCE(!mem))
> > continue;
> > + num_poisoned_pages_sub(-1UL, memblk_nr_poison(mem));
>
> Why does num_poisoned_pages_sub() have to make this distinction (!-1 == -1)
> for the hot-remove stage?
The first argument is used to find memory_block including the given pfn.
And in the above context remove_memory_block_devices() already has the
pointer "mem", so recalcurating it looked to me not necessary. Moreover,
this code is about to free the memory_block so updating the counter inside
it can be avoided. This is just a tiny optimization, and there can be
better option.
Thanks,
Naoya Horiguchi
Powered by blists - more mailing lists