linux-kernel - Re: [PATCH v2] resource: make sure requested range is included in the root range

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] [day] [month] [year] [list]

Message-ID: <CAErSpo4=1mVJ0K9fnwAG8+ABqEOdi54GNnbU+Yf7bbCqfDsXkg@mail.gmail.com>
Date:	Fri, 13 Jul 2012 09:34:23 -0600
From:	Bjorn Helgaas <bhelgaas@...gle.com>
To:	Octavian Purdila <octavian.purdila@...el.com>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org, Ram Pai <linuxram@...ibm.com>
Subject: Re: [PATCH v2] resource: make sure requested range is included in the
 root range

On Fri, Jul 13, 2012 at 2:38 AM, Octavian Purdila
<octavian.purdila@...el.com> wrote:
> On Thu, Jul 12, 2012 at 10:02 PM, Bjorn Helgaas <bhelgaas@...gle.com> wrote:
>
>> Why don't you fix this right where the problem occurs, in
>> __reserve_region_with_split(), with something like this:
>>
>>         if (end > conflict->start && conflict->start > start)
>>                 __reserve_region_with_split(root, start,
>> conflict->start-1, name);
>>         if (start < conflict->end && conflict->end < end && )
>>                 __reserve_region_with_split(root, conflict->end+1, end, name);
>>
>
> I am not sure this will behave properly in the overlapping cases.
> Consider start=10, end=150, root->start=100, root->end=200.
>
> In this case only the first condition above will be true (150 > 100 &&
> 100 > 10) and we will try to reserve  [10, 99] - which will return
> nicely. But we will not reserve [100, 150].

OK, I see your point -- I was unclear on the semantics of
reserve_region_with_split().  Apparently it is supposed to reserve
both pieces, even if it crosses the edge of an existing region.

> But more important is the fact that by fixing the issue here we won't
> be able to log the error and give a chance to upper layer fixing the
> problem.
>
>>> +               if (start > root->end || end < root->start) {
>>> +                       abort = 1;
>>> +                       pr_err("unable to fix request, aborting it\n");
>>
>> This message doesn't contain any useful information (range/root/etc).
>>
>
> Good point, I will remove it.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/