[U-Boot] [PATCH 1/2] ext4: Rename block group descriptor table from gd to bgd

[Graeme Russ]

Hi Simon,

On Thu, Oct 4, 2012 at 7:44 AM, Graeme Russ <graeme.russ at gmail.com> wrote:
> Hi Simon,
>
> On Oct 4, 2012 6:58 AM, "Simon Glass" <sjg at chromium.org> wrote:
>>
>> Hi Graeme,
>>
>> On Wed, Oct 3, 2012 at 1:47 PM, Graeme Russ <graeme.russ at gmail.com> wrote:
>> > Hi Simon,
>> >
>> > On Oct 4, 2012 6:40 AM, "Simon Glass" <sjg at chromium.org> wrote:
>> >>
>> >> Hi Tom,
>> >>
>> >> On Wed, Oct 3, 2012 at 1:04 PM, Tom Rini <trini at ti.com> wrote:
>> >> > -----BEGIN PGP SIGNED MESSAGE-----
>> >> > Hash: SHA1
>> >> >
>> >> > On 10/03/12 12:53, Simon Glass wrote:
>> >> >
>> >> >> On x86 machines gd is unfortunately a #define, so we should avoid
>> >> >> using gd for anything. This patch changes uses of gd to bgd so that
>> >> >> ext4fs can be used on x86.
>> >> >>
>> >> >> Signed-off-by: Simon Glass <sjg at chromium.org>
>> >> >
>> >> > Is there any way to change x86 to not be using a #define for gd?
>> >>
>> >> I wasn't brave enough to look hard at that, although Graeme is on copy
>> >> and will know. It is actually using inline assembly to access this
>> >> special variable.
>> >
>> > Isn't 'gd' used by everyone (global data)? I fail to see how this ever
>> > worked.
>>
>> Well only x86 uses a #define for it, so other archs cause no problem.
>> It means that we can't use 'gd' as a symbol anywhere in U-Boot. I
>> suppose the only sensible use is a structure member, as here.
>
> Ah, I see - and I don't see a quick and easy way out. Let me look a bit
> deeper...

I remember now... commit 9e6c572ff03cda84c88663b23c7157d8b1f275ac
explains why the #define gd came about:

"Use the base address of the 'F' segment as a pointer to the global data
structure. By adding the linear address (i.e. the 'D' segment address) as
the first word of the global data structure, the address of the global data
relative to the 'D' segment can be found simply, for example, by:

fs movl 0, %eax

This makes the gd 'pointer' writable prior to relocation (by reloading the
Global Desctriptor Table) which brings x86 into line with all other arches

NOTE: Writing to the gd 'pointer' is expensive (but we only do it
twice) but using it to access global data members (read and write) is
still fairly cheap"

The other option was rather ugly - create gd_get() and gd_set() inline
functions and replace all instances of gd-> in all U-Boot source with
gd_get()-> or gd_set(foo). I don't think it would have made any
difference to code size, but the amount of code touched would have
been massive.

The only other option I can think of is to change gd into something
much less likely to be used as a symbol (__gd for example), but again,
the patch to do so would be huge

I'm open to alternatives

Regards,

Graeme