Re: Bug#556790: gcc-4.3: unrecognizable insn on alpha
On 02/12/09 18:18, Matt Turner wrote:
On Tue, Dec 1, 2009 at 3:08 PM, Michael Cree<email@example.com> wrote:
On 18/11/2009, at 9:31 AM, Kurt Roeckx wrote:
I recently started seeing several cases of gcc giving an error
message like this:
packet-l2tp.c:1680: error: unrecognizable insn:
(insn 419 86 84 9 packet-l2tp.c:1585 (set (reg:DI 200)
(ne:SI (reg:DI 76 [ prephitmp.1208 ])
(const_int 0 [0x0]))) -1 (nil))
packet-l2tp.c:1680: internal compiler error: in extract_insn, at
Would this be http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42113 ??
Hmm. The patch that _uncovered_ (not caused) the regression was
It was included in>4.3.4 (that is, not in 4.3.4), 4.4.2, and the 4.5
branch, so I don't think it could be 42113, but it certainly couldn't
hurt to test.
From the Debian gcc-4.3 Changelog:
gcc-4.3 (4.3.4-2) unstable; urgency=low
[ Aurelien Jarno ]
* patches/kbsd-gnu-ada.dpatch: add Ada support on kfreebsd-amd64.
* rules.conf: enable Ada on kfreebsd-amd64.
[ Matthias Klose ]
* Update to SVN 20090826 from the gcc-4_3-branch.
- Fixes: PR target/41019, PR target/8603 (closes: #161432),
PR rtl-optimization/41033, PR target/40906, PR target/40577,
PR fortran/40847, PR fortran/40851, PR c/41046, PR debug/39706.
PR middle-end/41094, PR target/40718, PR fortran/41102.
* Avoid underscores in doc-base document id's to workaround a
So the patch that uncovered 42113 has been included in Debian's gcc
since version 4.3.4-2.