[U-Boot] [PATCH v1] Refactor linker-generated arrays

[Albert ARIBAUD]

Hi Andreas,

On Mon, 04 Feb 2013 13:41:09 +0100, "Andreas Bießmann"
<andreas.devel at googlemail.com> wrote:

> Hi Albert,
> 
> On 02.02.2013 18:02, Albert ARIBAUD wrote:
> > Refactor linker-generated array code so that symbols
> > which were previously linker-generated are now compiler-
> > generated. This causes relocation records of type
> > R_ARM_ABS32 to become R_ARM_RELATIVE, which makes
> > code which uses LGA able to run before relocation as
> > well as after.
> > 
> > Note: this affects more than ARM targets, as linker-
> > lists span possibly all target architectures, notably
> > PowerPC.
> > 
> > Signed-off-by: Albert ARIBAUD <albert.u.boot at aribaud.net>
> 
> Tested on avr32. The patch seems to work (basic shell testing), however
> it generates an aliasing warning:
> 
> ---8<---
> abiessmann at azuregos % PATH=$AVR32_PATH:$PATH BUILD_DIR=/tmp/build_avr32
> MAKEALL_LOGDIR=/tmp/LOG BUILD_NCPUS=4 BUILD_NBUILDS=4 ./MAKEALL atstk1002
> Configuring for atstk1002 board...
>    text	   data	    bss	    dec	    hex	filename
>  116315	   8972	 211900	 337187	  52523	/tmp/build_avr32/atstk1002/u-boot
> env_callback.c: In function 'find_env_callback':
> env_callback.c:47: warning: dereferencing pointer 'clbkp' does break
> strict-aliasing rules
> env_callback.c:44: note: initialized from here
> env_callback.c:46: note: initialized from here
> --->8---
> 
> I think it has something to do with tha fact that you re-cast the
> anonymous struct 'start' here:
> 
> ---8<---
> #define ll_entry_start(_type, _list)
> ({
> 	static struct {} start __aligned(4) __attribute__((unused,
> 		section(".u_boot_list_2_"#_list"_1")));
> 	(_type *)&start;
> })
> --->8---

Thanks Andreas for bringing this to my attention -- I'm surprised that
there's only one such warning, though, as there are may such casts. I'll
get my hands on a 4.4 compiler and try to find a way to cleanly avoid
the warning.

> I think other gcc-4.4 users will see the same error. Currently I have no
> time to dive into this.

Don't bother. I'll do the diving. :)

> Best regards
> 
> Andreas Bießmann

Amicalement,
-- 
Albert.