linux-kernel - Re: [PATCH 04/15] ftrace: pack event structures.

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: <AANLkTikKYaxLi-3a-0bdsbQR3scFzhn0txktNd+QPVa3@mail.gmail.com>
Date:	Thu, 10 Mar 2011 15:21:51 -0800
From:	David Sharp <dhsharp@...gle.com>
To:	Steven Rostedt <rostedt@...dmis.org>
Cc:	David Miller <davem@...emloft.net>, linux-kernel@...r.kernel.org,
	mrubin@...gle.com
Subject: Re: [PATCH 04/15] ftrace: pack event structures.

On Wed, Mar 9, 2011 at 7:18 AM, Steven Rostedt <rostedt@...dmis.org> wrote:
> On Tue, 2011-03-08 at 22:39 -0800, David Miller wrote:
>> From: Steven Rostedt <rostedt@...dmis.org>
>> Date: Tue, 08 Mar 2011 18:30:03 -0500
>>
>> > I'm going back through this patch set and looking to incorporate it into
>> > 2.6.39. But I know how touchy sparc is with "packed" data structures. Do
>> > you see this patch as hurting sparc?
>> >
>> > I wonder if we should change this to something like:
>> >
>> > #ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
>> > # define do_event_packed __attirbute__((packed))
>> > #else
>> > # define do_event_packed
>> > #endif
>> >
>> > and use "do_event_packed" instead?
>>
>> I think you should elide packed at all costs, and instead tell the compiler
>> what your intentions are by exposing the real types using unions or
>> similar instead of hiding it in an opaque way behind final char[] arrays.
>
> The problem here is not the issue of char[] but because of the way
> ftrace's header was 12 bytes and caused everything that had a 8byte word
> use another 4 bytes to align it.

Not only that, but those extra "4 bytes to align it" have a good
chance of in fact causing the 8 byte word to become *un*aligned, so
it's really just wasted space.

> As I stated in another email, I think
> removing the "int lock_depth" from the ftrace header should solve this.
>
> David (Sharp), the reason for the "packed" was because of that extra 4
> bytes in the ftrace header, right? So removing the lock_depth should fix
> the issues that you've seen?

Yes, I think that should have the same effect. Of course the next time
we come along and remove other useless fields like 'flags' and
'preempt_count', a similar condition could arise.

(Note I don't necessarily think that 'flags' and 'preempt_count' are
useless; although I don't know what they're used for, or why they need
to be in every entry. fwiw, we've taken trace_entry down to 4 bytes of
{unsigned short type; unsigned short pid;}, and we even have plans to
completely remove pid.)

>
> -- Steve
>
>
>
>
--
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/