[PATCH] MIPS/GAS/test: Assorted ECOFF fixes

[PATCH] MIPS/GAS/test: Assorted ECOFF fixes

[Maciej W. Rozycki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1121 bytes --]

Hi,

 Here are some assorted ECOFF fixes for the GAS testsuite.  Apart from the 
obvious fixes they include changes like -mabi=32 to -32 GAS option changes 
(the former is ELF-specific), explicit ISA specifiers for `objdump' as 
ECOFF is not capable enough to store some settings and explicit symbol 
reference removals as ECOFF tends to replace local symbols at the 
beginning with section names in dumps.

 I'll see if I can do more as most if not all of these failures are simple 
obvious mistakes coming from the unawareness of ECOFF peculiarities and no 
testing at the same time; and then some tests should be limited to ELF. 
Not sure I'll be able to fit in the 2.21 timeframe though.

2010-10-03  Maciej W. Rozycki  <macro@linux-mips.org>

	gas/testsuite/
	* gas/mips/at-1.d: Handle ECOFF relocations.
	* gas/mips/eret-1.d: Handle ECOFF.
	* gas/mips/eret-2.d: Likewise.
	* gas/mips/eret-3.d: Likewise.
	* gas/mips/lb.d: Handle an ECOFF addend.
	* gas/mips/mips.exp: Only run "aent" for ELF.

 OK to apply?

  Maciej

binutils-2.20.51-20100925-mips-gas-test-ecoff.patch
[Patch attached compressed due to its size.]

[-- Attachment #2: Type: APPLICATION/X-BZIP2, Size: 1903 bytes --]

Re: [PATCH] MIPS/GAS/test: Assorted ECOFF fixes

[Richard Sandiford]

"Maciej W. Rozycki" <macro@linux-mips.org> writes:
> 2010-10-03  Maciej W. Rozycki  <macro@linux-mips.org>
>
> 	gas/testsuite/
> 	* gas/mips/at-1.d: Handle ECOFF relocations.
> 	* gas/mips/eret-1.d: Handle ECOFF.
> 	* gas/mips/eret-2.d: Likewise.
> 	* gas/mips/eret-3.d: Likewise.
> 	* gas/mips/lb.d: Handle an ECOFF addend.
> 	* gas/mips/mips.exp: Only run "aent" for ELF.

Everything apart from the lb.d change is OK.  I think an ecoff@ override
is more appropriate for the lb.d test.

Richard

Re: [PATCH] MIPS/GAS/test: Assorted ECOFF fixes

[Maciej W. Rozycki]

On Sun, 10 Oct 2010, Richard Sandiford wrote:

> > 	gas/testsuite/
> > 	* gas/mips/at-1.d: Handle ECOFF relocations.
> > 	* gas/mips/eret-1.d: Handle ECOFF.
> > 	* gas/mips/eret-2.d: Likewise.
> > 	* gas/mips/eret-3.d: Likewise.
> > 	* gas/mips/lb.d: Handle an ECOFF addend.
> > 	* gas/mips/mips.exp: Only run "aent" for ELF.
> 
> Everything apart from the lb.d change is OK.  I think an ecoff@ override
> is more appropriate for the lb.d test.

 This is exactly the same situation as with ld.d and I'd prefer to take 
the same approach especially as here we should be able to avoid any ECOFF 
overrides at all.  I've taken the change out so as the failure keeps 
sticking out as a reminder even though it only made the single remaining 
offset consistent with the rest.  I have applied the rest of the changes, 
thanks.

  Maciej