This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

New PRs for various failing test-results (arm, d10v, frv, h8300, iq2000, m32c, m32r, mcore, mips, mn10300, sh, v850)


Hi.  I made myself an autotester for src/sim (starting a
test-round for any gdb check-in), so I have a need for
reasonably current binutils that "works", as many/most simulator
tests require binutils in order to run the testsuite.  I hoped
to use clean test-results for the sim targets as the "works"
indicator.  From the binutils sources for my binutils autotester
(the one that tests that binutils for cris-elf passes cleanly),
when the results are clean I make candidate tarballs, and pick
them up in the sim autotester, which does an extra sanity check
before updating its binutils sources.  I believe Geoff Keating
does (or did) something similar in his gcc autotester.

The baseline results aren't impressive.  Of these targets, all
having simulators in src/sim (though not all having executable
sim testsuites):
 arm-elf cris-elf d10v-elf sparc-elf frv-elf h8300-elf iq2000-elf
 m32c-elf m32r-elf m68hc11-elf mcore-elf mips-elf mn10300-elf
 ppc-elf sh-elf sh64-elf v850-elf
very few pass the binutils testsuites.  As the extra binutils
sanity test, I have settled for just checking that h68hc11-elf
and ppc-elf pass cleanly. :(

In a feeble attempt to make this test-round not entirely wasted
and perhaps to improve the situation, I invested a little more
time and created sourceware bugzilla PRs for the failing tests,
hosted on i686-pc-linux-gnu:

arm-elf:
3795 gas
3796 ld

d10v-elf
3797 gas
3798 ld

frv-elf
3799 ld

h8300-elf
3800 gas

iq2000-elf
3801 ld

m32c-elf
3802 binutils
3803 gas
3804 ld

m32r-elf
3805 ld

mcore-elf
3806 gas

mips-elf
3807 binutils
3808 gas
3809 ld

mn10300-elf
3810 gas

sh-elf, sh64-elf
3811 gas

v850-elf
3812 gas

Those maintainers first mentioned for the target (where any
specific maintainer is listed) and that I don't clearly recall
having seen the last month on this list, are CC'ed.  Feel free
to ignore.  NB: I really don't think I'll be fixing the tests
myself.  If you have a go, pretty please make sure your targets
pass on a "64-bit host" (the sim autotester is hosted on one); I
know some targets fail because of that.  Film at 11.

brgds, H-P


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]