[Bug target/106282] [10/11/12/13 Regression] m68k: Problem with thread-local storage and -mcpu=5206 since r9-2326-gede9446c26a929

["jsm28 at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106282

Joseph S. Myers <jsm28 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |10.5
            Summary|m68k: Problem with          |[10/11/12/13 Regression]
                   |thread-local storage and    |m68k: Problem with
                   |-mcpu=5206 since            |thread-local storage and
                   |r9-2326-gede9446c26a929     |-mcpu=5206 since
                   |                            |r9-2326-gede9446c26a929
                 CC|                            |jsm28 at gcc dot gnu.org

--- Comment #2 from Joseph S. Myers <jsm28 at gcc dot gnu.org> ---
My glibc bot has been failing for ColdFire since some time between commits
79d38dd46e6b07e5a90ab25df1438eb0918eb475 and
f56d48b2471c388401174029324e1f4c4b84fcdb, with an assembler error building
libgomp (affinity-fmt.c) that looks just like the one in this bug.

affinity-fmt.s: Assembler messages:
affinity-fmt.s:4062: Error: syntax error -- statement `lea
(gomp_tls_data@TLSLE+8,%a0),%a0' ignored

https://sourceware.org/pipermail/libc-testresults/2022q4/010394.html

Although that's from October 2022, I expect it's the same underlying issue, and
just was latent building libgomp until some other change exposed it there.

Note: soft-float ColdFire now fails earlier in my bot, failing to build glibc
with bug 103370 (also an assembler error, also probably latent for some time
before being exposed by an unrelated change).