linux-kernel - Re: [PATCH 2/5] mm, devm_memremap_pages: handle errors allocating final devres action

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 21 May 2018 17:07:49 -0700
From:   Dan Williams <dan.j.williams@...el.com>
To:     Andrew Morton <akpm@...ux-foundation.org>
Cc:     stable <stable@...r.kernel.org>, Christoph Hellwig <hch@....de>,
        Jérôme Glisse <jglisse@...hat.com>,
        Logan Gunthorpe <logang@...tatee.com>,
        Linux MM <linux-mm@...ck.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/5] mm, devm_memremap_pages: handle errors allocating
 final devres action

[ resend as the last attempt dropped all the cc's ]

On Mon, May 21, 2018 at 4:10 PM, Andrew Morton
<akpm@...ux-foundation.org> wrote:
> On Mon, 21 May 2018 15:35:24 -0700 Dan Williams <dan.j.williams@...el.com> wrote:
>
>> The last step before devm_memremap_pages() returns success is to
>> allocate a release action to tear the entire setup down. However, the
>> result from devm_add_action() is not checked.
>>
>> Checking the error also means that we need to handle the fact that the
>> percpu_ref may not be killed by the time devm_memremap_pages_release()
>> runs. Add a new state flag for this case.
>>
>> Cc: <stable@...r.kernel.org>
>> Fixes: e8d513483300 ("memremap: change devm_memremap_pages interface...")
>> Cc: Christoph Hellwig <hch@....de>
>> Cc: "Jérôme Glisse" <jglisse@...hat.com>
>> Cc: Logan Gunthorpe <logang@...tatee.com>
>> Signed-off-by: Dan Williams <dan.j.williams@...el.com>
>
> Why the cc:stable?  The changelog doesn't describe the end-user-visible
> impact of the bug (it always should, for this reason).

True, I should have included that, one of these years I'll stop making
this mistake.

>
> AFAICT we only go wrong when a small GFP_KERNEL allocation fails
> (basically never happens), with undescribed results :(
>

Here's a better changelog:

---
The last step before devm_memremap_pages() returns success is to
allocate a release action to tear the entire setup down. However, the
result from devm_add_action() is not checked.

Checking the error also means that we need to handle the fact that the
percpu_ref may not be killed by the time devm_memremap_pages_release()
runs. Add a new state flag for this case.

Without this change we could fail to register the teardown of
devm_memremap_pages(). The likelihood of hitting this failure is tiny
as small memory allocations almost always succeed. However, the impact
of the failure is large given any future reconfiguration, or
disable/enable, of an nvdimm namespace will fail forever as subsequent
calls to devm_memremap_pages() will fail to setup the pgmap_radix
since there will be stale entries for the physical address range.