Re: cleanup PT_GNU_STACK size handling

[Nathan Sidwell <nathan_sidwell at mentor dot com>]

On 10/24/12 09:27, Alan Modra wrote:
> Please fix.  The empty-aligned failures are likely due to lack of
> support for -z stack-size, the assembly failures due to lack of a bare
> "nop" instruction.

I've been getting connectivity problems to CVS all day, so unable to even 
generate a diff.

Here are updated files which work for the following targets:

arc-elf bfin-uclinux frv-elf frv-uclinux i370-linux i686-pc-linux-gnu ia64-linux 
lm32-uclinux sh2-uclinux tic6x-uclinux


The change to stack.s is pretty obvious.  The empty-aligned.d change attempts to 
ignore any GNU_STACK segment, but I'm a little worried it'll also fail to detect 
something else that should cause a test fail.

Alternatively, I'll be back towards the end of next week.

nathan
--
Nathan Sidwell

Attachment: stack.s
Description: Text document

#source: empty-aligned.s
#ld: -T empty-aligned.t
#readelf: -l --wide
#xfail: "hppa64-*-*"
#notarget: frv-*-*linux*

#...
Program Headers:
 +Type +Offset +VirtAddr +PhysAddr +FileSiz +MemSiz +Flg +Align
 +LOAD +0x[0-9a-f]+ 0x[0-9a-f]+ 0x[0-9a-f]+ 0x[0-9a-f]+ 0x[0-9a-f]+ [RWE ]+ +0x[0-9a-f]+
! +LOAD .*
#...
 +Segment Sections\.\.\.
 +00 +.text 
#pass