linux-kernel - Re: [PATCH 0/5] Persist printk buffer across reboots.

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:	Tue, 13 Mar 2012 03:14:21 -0400
From:	Avery Pennarun <apenwarr@...il.com>
To:	David Miller <davem@...emloft.net>
Cc:	akpm@...ux-foundation.org, josh@...htriplett.org,
	paulmck@...ux.vnet.ibm.com, mingo@...e.hu, a.p.zijlstra@...llo.nl,
	fdinitto@...hat.com, hannes@...xchg.org, olaf@...fle.de,
	paul.gortmaker@...driver.com, tj@...nel.org, hpa@...ux.intel.com,
	yinghai@...nel.org, linux-kernel@...r.kernel.org,
	linux-mm@...ck.org
Subject: Re: [PATCH 0/5] Persist printk buffer across reboots.

On Tue, Mar 13, 2012 at 2:50 AM, David Miller <davem@...emloft.net> wrote:
> The idea is that you call prom_retain() before you take a look at what
> physical memory is available in the kernel, and the firmware takes
> this physical chunk out of those available memory lists upon
> prom_retain() success.

This sounds like exactly the API I would have wanted, however:

1) It's only available in arch/sparc so I can't test my patch if I try
to use it;
2) There's nobody that calls it so it might not work;
3) I don't understand the API so I'm not really confident that
reserving memory this way will actually prevent it from being seen by
the kernel.

In short, I think I would screw it up.

On the other hand, as written it seems like my code would also work on
sparc, and would work with more than one kind of memory area if more
than one module chose to use this technique.  (ie. Since the prober
actually reserves memory, the next prober would necessarily reserve a
different bit of memory, and as long as you're using the same kernel
as before and you do all reservations before enabling interrupts, you
should get consistent results.)

I suppose I could move the actual probe-and-allocate code somewhere
(bootmem.c?  memblock.c?) and add a 'name' parameter to it which is
ignored in the generic implementation.  Then someone could write an
arch-specific implementation.  Would that work?  If so, please
recommend which file to put the generic implementation in :)

Thanks,

Avery
--
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/