linux-kernel - Re: [PATCH 1/2] ramoops: use pstore interface

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]

Message-ID: <CAGXu5jJfhLjBN4hEUNv5DaivJJaX5=15+JaC0VEON+mfDSeF+A@mail.gmail.com>
Date:	Thu, 17 Nov 2011 10:10:00 -0800
From:	Kees Cook <keescook@...omium.org>
To:	Chen Gong <gong.chen@...ux.intel.com>
Cc:	linux-kernel@...r.kernel.org, Greg Kroah-Hartman <gregkh@...e.de>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Arnd Bergmann <arnd@...db.de>,
	Nicolas Pitre <nicolas.pitre@...aro.org>,
	Marco Stornelli <marco.stornelli@...il.com>,
	Paul Gortmaker <paul.gortmaker@...driver.com>
Subject: Re: [PATCH 1/2] ramoops: use pstore interface

On Wed, Nov 16, 2011 at 9:35 PM, Chen Gong <gong.chen@...ux.intel.com> wrote:
> 于 2011/11/17 5:25, Kees Cook 写道:
>> Instead of using /dev/mem directly, use the common pstore infrastructure
>> to handle Oops gathering and extraction.
>> [...]
>> +       /* Explicitly only take the first part of any new crash.
>> +        * If our buffer is larger than kmsg_bytes, this can never happen,
>> +        * and if our buffer is smaller than kmsg_bytes, we don't want the
>> +        * report split across multiple records. */
>> +       if (part != 1)
>> +               return -ENOSPC;
>
> why only one part is accepted? You are afraid about your filename style?

The logic in ramoops doesn't expect to have a split-up report. Since
pstore doesn't limit reports to kmsg_bytes in size (it actually splits
reports on pstore_info.bufsize) this is a non-issue, but in the case
that a platform defines very small ramoops record sizes, I didn't want
the extra stuff written to additional records. If ramoops gains real
record headers ever, this can change, of course. In the meantime, it
should be defensive.

>> +       /* Only a single ramoops area allowed at a time, so fail extra
>> +        * probes.
>> +        */
>> +       if (cxt->max_count)
>> +               goto fail3;
>
> Should be fail4
> [...]
> In some situations fail4 maybe hits max_count != 0, so here max_count should
> be cleared. I think you should rearrange the logic in this function
> carefully.

Ah, thanks for the catch. All the error targets got messed up. I'll
fix them and name them instead of using numbers.

>> +       /* TODO(kees): It shouldn't be possible to remove ramoops since
>> +        * pstore doesn't support unregistering yet. When it does, remove
>> +        * this early return and add the unregister where noted below.
>> +        */
>> +       return -EBUSY;
>
> This style is not reasonable. Maybe it should have a better wrap.

I'm not sure I understand what you mean. It's wrapped roughly to
column 75 already. What would be better for this comment? Or did you
mean I shouldn't have unreachable code?

> BTW, you need to update Documentation/ramoops.txt

Ah! Yes, thanks for the reminder.

-Kees

-- 
Kees Cook
ChromeOS Security
--
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/