binutils 2.20 regressions (20090930)

[Joel Sherrill]

Tristan Gingold wrote:
> Hi,
>
> after the recent commits, I did a new batch of check.  See the results  
> below.
> Unless I am wrong I don't see anything blocking.  I plan to do the  
> 2.20 release this week (after the PE
> linker fix).
>   
Will Alan's fix for the powerpc-rtems failures be included?  I suspect
it will also knock out some of the m32c failures.

What does the mips jal failure indicate?  Is that serious?

I would like to rerun all the RTEMS targets after that is committed.

Current RTEMS target results are:

h8300-rtems4.10 has no failures.
i386-rtems4.10 has no failures.
lm32-rtems4.10 has no failures.
m32c-rtems4.10 has 8 failures.
    FAIL: Check --gc-section
    FAIL: Check --gc-section/-q
    FAIL: NOCROSSREFS 3
    FAIL: ld-scripts/empty-address-2a
    FAIL: ld-scripts/empty-address-2b
    FAIL: ld-scripts/empty-address-3a
    FAIL: ld-scripts/empty-address-3b
    FAIL: ld-scripts/empty-address-3c
m32r-rtems4.10 has no failures.
m68k-rtems4.10 has no failures.
mips-rtems4.10 has 1 failure.
    FAIL: MIPS jal
powerpc-rtems4.10 has 4 failures.
    FAIL: Check --gc-section
    FAIL: Check --gc-section/-q
    FAIL: S-records
    FAIL: S-records with constructors
sh-rtems4.10 has no failures.
sparc-rtems4.10 has no failures.