Fwd: [PATCH] em7210: fix memory size

Fwd: [PATCH] em7210: fix memory size

[Dan Williams]

[ replied to wrong mailing list ]


---------- Forwarded message ----------
From: Dan Williams <dan.j.williams@intel.com>
Date: Sun, Aug 30, 2009 at 4:31 PM
Subject: Re: [PATCH] em7210: fix memory size
To: Arnaud Patard <arnaud.patard@rtp-net.org>
Cc: linux-arm-kernel at lists.arm.linux.org.uk


On Wed, Jul 29, 2009 at 4:56 AM, Arnaud Patard<arnaud.patard@rtp-net.org> wrote:
>
> The default memory module size on the ss4000e is 256M but Redboot has
> been hardcoded to pass 512M to the kernel through ATAG. As getting a new
> redboot is more than unlikely, it has to be fixed in the kernel
> (unfortunately). This patch does that.
>

I am on the fence on this one. ?On the one hand it makes sense in that
the kernel is usable without a "mem=256M..." command line option on
default configured platforms. ?Users that have upgraded the memory can
pass a "mem=512M..." to use their extra memory. ?However, it still
seems that this will appear as a regression to any current 512M
platforms that are not passing a command line option, and a "don't
care" to everyone who must be passing "mem=256M..." to get a
functional platform.

Hmm... ?can you say a bit more about this comment "solving this issue will help
also people porting the d-i on the device." [1] That might help tip the balance.

Thanks,
Dan

[1]: http://marc.info/?l=linux-arm-kernel&m=124810119114783&w=2

Fwd: [PATCH] em7210: fix memory size

[Arnaud Patard (Rtp)]

Dan Williams <dan.j.williams@intel.com> writes:
Hi,

> [ replied to wrong mailing list ]
>
>
> ---------- Forwarded message ----------
> From: Dan Williams <dan.j.williams@intel.com>
> Date: Sun, Aug 30, 2009 at 4:31 PM
> Subject: Re: [PATCH] em7210: fix memory size
> To: Arnaud Patard <arnaud.patard@rtp-net.org>
> Cc: linux-arm-kernel at lists.arm.linux.org.uk
>
>
> On Wed, Jul 29, 2009 at 4:56 AM, Arnaud Patard<arnaud.patard@rtp-net.org> wrote:
>>
>> The default memory module size on the ss4000e is 256M but Redboot has
>> been hardcoded to pass 512M to the kernel through ATAG. As getting a new
>> redboot is more than unlikely, it has to be fixed in the kernel
>> (unfortunately). This patch does that.
>>
>
> I am on the fence on this one. ?On the one hand it makes sense in that
> the kernel is usable without a "mem=256M..." command line option on
> default configured platforms. ?Users that have upgraded the memory can
> pass a "mem=512M..." to use their extra memory. ?However, it still
> seems that this will appear as a regression to any current 512M
> platforms that are not passing a command line option, and a "don't
> care" to everyone who must be passing "mem=256M..." to get a
> functional platform.
>
> Hmm... ?can you say a bit more about this comment "solving this issue will help
> also people porting the d-i on the device." [1] That might help tip the balance.

I've asked more details and I'm going to try to sum up the point. I hope
I won't make any mistake as I'm not familiar with d-i internals at
all. Please bear with me.