From: Toshiki Fukasawa <t-fukasawa@vx.jp.nec.com>
To: David Hildenbrand <david@redhat.com>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
"dan.j.williams@intel.com" <dan.j.williams@intel.com>
Cc: Toshiki Fukasawa <t-fukasawa@vx.jp.nec.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"akpm@linux-foundation.org" <akpm@linux-foundation.org>,
"mhocko@kernel.org" <mhocko@kernel.org>,
"adobriyan@gmail.com" <adobriyan@gmail.com>,
"hch@lst.de" <hch@lst.de>,
"longman@redhat.com" <longman@redhat.com>,
"sfr@canb.auug.org.au" <sfr@canb.auug.org.au>,
"mst@redhat.com" <mst@redhat.com>,
Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>,
Junichi Nomura <j-nomura@ce.jp.nec.com>
Subject: Re: [RFC PATCH v2] mm: initialize struct pages reserved by ZONE_DEVICE driver.
Date: Mon, 9 Sep 2019 05:48:34 +0000 [thread overview]
Message-ID: <6a99e003-e1ab-b9e8-7b25-bc5605ab0eb2@vx.jp.nec.com> (raw)
In-Reply-To: <40a1ce2e-1384-b869-97d0-7195b5b47de0@redhat.com>
On 2019/09/06 19:35, David Hildenbrand wrote:
> On 06.09.19 12:02, Toshiki Fukasawa wrote:
>> Thank you for your feedback.
>>
>> On 2019/09/06 17:45, David Hildenbrand wrote:
>>> On 06.09.19 10:09, Toshiki Fukasawa wrote:
>>>> A kernel panic is observed during reading
>>>> /proc/kpage{cgroup,count,flags} for first few pfns allocated by
>>>> pmem namespace:
>>>>
>>>> BUG: unable to handle page fault for address: fffffffffffffffe
>>>> [ 114.495280] #PF: supervisor read access in kernel mode
>>>> [ 114.495738] #PF: error_code(0x0000) - not-present page
>>>> [ 114.496203] PGD 17120e067 P4D 17120e067 PUD 171210067 PMD 0
>>>> [ 114.496713] Oops: 0000 [#1] SMP PTI
>>>> [ 114.497037] CPU: 9 PID: 1202 Comm: page-types Not tainted 5.3.0-rc1
>>>> [ 114.497621] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.11.0-0-g63451fca13-prebuilt.qemu-project.org 04/01/2014
>>>> [ 114.498706] RIP: 0010:stable_page_flags+0x27/0x3f0
>>>> [ 114.499142] Code: 82 66 90 66 66 66 66 90 48 85 ff 0f 84 d1 03 00 00 41 54 55 48 89 fd 53 48 8b 57 08 48 8b 1f 48 8d 42 ff 83 e2 01 48 0f 44 c7 <48> 8b 00 f6 c4 02 0f 84 57 03 00 00 45 31 e4 48 8b 55 08 48 89 ef
>>>> [ 114.500788] RSP: 0018:ffffa5e601a0fe60 EFLAGS: 00010202
>>>> [ 114.501373] RAX: fffffffffffffffe RBX: ffffffffffffffff RCX: 0000000000000000
>>>> [ 114.502009] RDX: 0000000000000001 RSI: 00007ffca13a7310 RDI: ffffd07489000000
>>>> [ 114.502637] RBP: ffffd07489000000 R08: 0000000000000001 R09: 0000000000000000
>>>> [ 114.503270] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000240000
>>>> [ 114.503896] R13: 0000000000080000 R14: 00007ffca13a7310 R15: ffffa5e601a0ff08
>>>> [ 114.504530] FS: 00007f0266c7f540(0000) GS:ffff962dbbac0000(0000) knlGS:0000000000000000
>>>> [ 114.505245] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>> [ 114.505754] CR2: fffffffffffffffe CR3: 000000023a204000 CR4: 00000000000006e0
>>>> [ 114.506401] Call Trace:
>>>> [ 114.506660] kpageflags_read+0xb1/0x130
>>>> [ 114.507051] proc_reg_read+0x39/0x60
>>>> [ 114.507387] vfs_read+0x8a/0x140
>>>> [ 114.507686] ksys_pread64+0x61/0xa0
>>>> [ 114.508021] do_syscall_64+0x5f/0x1a0
>>>> [ 114.508372] entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>> [ 114.508844] RIP: 0033:0x7f0266ba426b
>>>>
>>>> The first few pages of ZONE_DEVICE expressed as the range
>>>> (altmap->base_pfn) to (altmap->base_pfn + altmap->reserve) are
>>>> skipped by struct page initialization. Some pfn walkers like
>>>> /proc/kpage{cgroup, count, flags} can't handle these uninitialized
>>>> struct pages, which causes the error.
>>>>
>>>> In previous discussion, Dan seemed to have concern that the struct
>>>> page area of some pages indicated by vmem_altmap->reserve may not
>>>> be allocated. (See https://lore.kernel.org/lkml/CAPcyv4i5FjTOnPbXNcTzvt+e6RQYow0JRQwSFuxaa62LSuvzHQ@mail.gmail.com/)
>>>> However, arch_add_memory() called by devm_memremap_pages() allocates
>>>> struct page area for pages containing addresses in the range
>>>> (res.start) to (res.start + resource_size(res)), which include the
>>>> pages indicated by vmem_altmap->reserve. If I read correctly, it is
>>>> allocated as requested at least on x86_64. Also, memmap_init_zone()
>>>> initializes struct pages in the same range.
>>>> So I think the struct pages should be initialized.>
>>>
>>> For !ZONE_DEVICE memory, the memmap is valid with SECTION_IS_ONLINE -
>>> for the whole section. For ZONE_DEVICE memory we have no such
>>> indication. In any section that is !SECTION_IS_ONLINE and
>>> SECTION_MARKED_PRESENT, we could have any subsections initialized. >
>>> The only indication I am aware of is pfn_zone_device_reserved() - which
>>> seems to check exactly what you are trying to skip here.
>>>
>>> Can't you somehow use pfn_zone_device_reserved() ? Or if you considered
>>> that already, why did you decide against it?
>>
>> No, in current approach this function is no longer needed.
>> The reason why we change the approach is that all pfn walkers
>> have to be aware of the uninitialized struct pages.
>
> We should use the same strategy for all pfn walkers then (effectively
> get rid of pfn_zone_device_reserved() if that's what we want).
True, but this patch replaces "/proc/kpageflags: do not use uninitialized
struct pages". If we initialize the uninitialized struct pages, no pfn walker
will need to be aware of them.
>
>>
>> As for SECTION_IS_ONLINE, I'm not sure now.
>> I will look into it next week.
>
> SECTION_IS_ONLINE does currently not apply to ZONE_DEVICE and due to
> sub-section support for ZONE_DEVICE, it cannot easily be reused.
>
It seems that SECTION_IS_ONLINE and SECTION_MARKED_PRESENT can be used to
distinguish uninitialized struct pages if we can apply them to ZONE_DEVICE,
but that is no longer necessary with this approach.
Thanks,
Toshiki Fukasawa
next prev parent reply other threads:[~2019-09-09 6:00 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-06 8:09 [RFC PATCH v2] mm: initialize struct pages reserved by ZONE_DEVICE driver Toshiki Fukasawa
2019-09-06 8:45 ` David Hildenbrand
2019-09-06 10:02 ` Toshiki Fukasawa
2019-09-06 10:35 ` David Hildenbrand
2019-09-09 5:48 ` Toshiki Fukasawa [this message]
2019-09-09 7:46 ` David Hildenbrand
2019-09-09 8:10 ` David Hildenbrand
2019-09-09 11:53 ` Dan Williams
2019-09-09 12:05 ` David Hildenbrand
2019-09-10 9:21 ` Dan Williams
2019-09-10 10:10 ` David Hildenbrand
2019-09-17 2:34 ` Toshiki Fukasawa
2019-09-17 7:13 ` David Hildenbrand
2019-09-17 9:32 ` Toshiki Fukasawa
2019-09-17 10:20 ` David Hildenbrand
2019-09-18 2:16 ` Toshiki Fukasawa
2019-09-17 15:49 ` Waiman Long
2019-09-17 16:21 ` Qian Cai
2019-09-17 17:04 ` Waiman Long
2019-09-17 20:23 ` David Hildenbrand
2019-09-18 2:28 ` Toshiki Fukasawa
2019-09-18 7:30 ` David Hildenbrand
2019-09-10 14:01 ` Michal Hocko
2019-09-10 14:53 ` Dan Williams
2019-09-10 17:35 ` Michal Hocko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=6a99e003-e1ab-b9e8-7b25-bc5605ab0eb2@vx.jp.nec.com \
--to=t-fukasawa@vx.jp.nec.com \
--cc=adobriyan@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=dan.j.williams@intel.com \
--cc=david@redhat.com \
--cc=hch@lst.de \
--cc=j-nomura@ce.jp.nec.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=longman@redhat.com \
--cc=mhocko@kernel.org \
--cc=mst@redhat.com \
--cc=n-horiguchi@ah.jp.nec.com \
--cc=sfr@canb.auug.org.au \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).