From: Walter Wu <walter-zh.wu@mediatek.com>
To: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Vlastimil Babka <vbabka@suse.cz>,
Alexander Potapenko <glider@google.com>,
Dmitry Vyukov <dvyukov@google.com>,
Matthias Brugger <matthias.bgg@gmail.com>,
Andrew Morton <akpm@linux-foundation.org>,
"Martin Schwidefsky" <schwidefsky@de.ibm.com>,
Will Deacon <will@kernel.org>,
Andrey Konovalov <andreyknvl@google.com>,
Arnd Bergmann <arnd@arndb.de>,
Thomas Gleixner <tglx@linutronix.de>,
Michal Hocko <mhocko@kernel.org>, Qian Cai <cai@lca.pw>,
<linux-kernel@vger.kernel.org>, <kasan-dev@googlegroups.com>,
<linux-mm@kvack.org>, <linux-arm-kernel@lists.infradead.org>,
<linux-mediatek@lists.infradead.org>, <wsd_upstream@mediatek.com>
Subject: Re: [PATCH v2 0/2] mm/kasan: dump alloc/free stack for page allocator
Date: Tue, 10 Sep 2019 20:46:15 +0800 [thread overview]
Message-ID: <1568119575.24886.20.camel@mtksdccf07> (raw)
In-Reply-To: <a7863965-90ab-5dae-65e7-8f68f4b4beb5@virtuozzo.com>
On Tue, 2019-09-10 at 13:50 +0300, Andrey Ryabinin wrote:
>
> On 9/9/19 4:07 PM, Vlastimil Babka wrote:
> > On 9/9/19 10:24 AM, walter-zh.wu@mediatek.com wrote:
> >> From: Walter Wu <walter-zh.wu@mediatek.com>
> >>
> >> This patch is KASAN report adds the alloc/free stacks for page allocator
> >> in order to help programmer to see memory corruption caused by page.
> >>
> >> By default, KASAN doesn't record alloc and free stack for page allocator.
> >> It is difficult to fix up page use-after-free or dobule-free issue.
> >>
> >> Our patchsets will record the last stack of pages.
> >> It is very helpful for solving the page use-after-free or double-free.
> >>
> >> KASAN report will show the last stack of page, it may be:
> >> a) If page is in-use state, then it prints alloc stack.
> >> It is useful to fix up page out-of-bound issue.
> >
> > I still disagree with duplicating most of page_owner functionality for the sake of using a single stack handle for both alloc and free (while page_owner + debug_pagealloc with patches in mmotm uses two handles). It reduces the amount of potentially important debugging information, and I really doubt the u32-per-page savings are significant, given the rest of KASAN overhead.
> >
> >> BUG: KASAN: slab-out-of-bounds in kmalloc_pagealloc_oob_right+0x88/0x90
> >> Write of size 1 at addr ffffffc0d64ea00a by task cat/115
> >> ...
> >> Allocation stack of page:
> >> set_page_stack.constprop.1+0x30/0xc8
> >> kasan_alloc_pages+0x18/0x38
> >> prep_new_page+0x5c/0x150
> >> get_page_from_freelist+0xb8c/0x17c8
> >> __alloc_pages_nodemask+0x1a0/0x11b0
> >> kmalloc_order+0x28/0x58
> >> kmalloc_order_trace+0x28/0xe0
> >> kmalloc_pagealloc_oob_right+0x2c/0x68
> >>
> >> b) If page is freed state, then it prints free stack.
> >> It is useful to fix up page use-after-free or double-free issue.
> >>
> >> BUG: KASAN: use-after-free in kmalloc_pagealloc_uaf+0x70/0x80
> >> Write of size 1 at addr ffffffc0d651c000 by task cat/115
> >> ...
> >> Free stack of page:
> >> kasan_free_pages+0x68/0x70
> >> __free_pages_ok+0x3c0/0x1328
> >> __free_pages+0x50/0x78
> >> kfree+0x1c4/0x250
> >> kmalloc_pagealloc_uaf+0x38/0x80
> >>
> >> This has been discussed, please refer below link.
> >> https://bugzilla.kernel.org/show_bug.cgi?id=203967
> >
> > That's not a discussion, but a single comment from Dmitry, which btw contains "provide alloc *and* free stacks for it" ("it" refers to page, emphasis mine). It would be nice if he or other KASAN guys could clarify.
> >
>
> For slab objects we memorize both alloc and free stacks. You'll never know in advance what information will be usefull
> to fix an issue, so it usually better to provide more information. I don't think we should do anything different for pages.
>
> Given that we already have the page_owner responsible for providing alloc/free stacks for pages, all that we should in KASAN do is to
> enable the feature by default. Free stack saving should be decoupled from debug_pagealloc into separate option so that it can be enabled
> by KASAN and/or debug_pagealloc.
Thanks your suggestion.
We will send the patch v3 as described above.
next prev parent reply other threads:[~2019-09-10 12:46 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-09 8:24 [PATCH v2 0/2] mm/kasan: dump alloc/free stack for page allocator walter-zh.wu
2019-09-09 13:07 ` Vlastimil Babka
2019-09-10 10:50 ` Andrey Ryabinin
2019-09-10 11:53 ` Vlastimil Babka
2019-09-10 12:45 ` Walter Wu
2019-09-10 12:46 ` Walter Wu [this message]
2019-09-10 9:53 ` Walter Wu
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=1568119575.24886.20.camel@mtksdccf07 \
--to=walter-zh.wu@mediatek.com \
--cc=akpm@linux-foundation.org \
--cc=andreyknvl@google.com \
--cc=arnd@arndb.de \
--cc=aryabinin@virtuozzo.com \
--cc=cai@lca.pw \
--cc=dvyukov@google.com \
--cc=glider@google.com \
--cc=kasan-dev@googlegroups.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mediatek@lists.infradead.org \
--cc=linux-mm@kvack.org \
--cc=matthias.bgg@gmail.com \
--cc=mhocko@kernel.org \
--cc=schwidefsky@de.ibm.com \
--cc=tglx@linutronix.de \
--cc=vbabka@suse.cz \
--cc=will@kernel.org \
--cc=wsd_upstream@mediatek.com \
/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).