Bug 434397
| Summary: | cmucl borked on f9/gcc43 | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Jesse Keating <jkeating> |
| Component: | cmucl | Assignee: | Rex Dieter <rdieter> |
| Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 9 | CC: | dcantrell, green, jonstanley |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | 19e-0.3.pre2 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2008-07-08 03:53:31 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 432425 | ||
|
Description
Jesse Keating
2008-02-22 16:08:48 UTC
scratch build 482739 Well this build ran a little on the long side :). Canceled it, downloaded the generated SRPM to my box and tried building it. Lisp took 100% of the CPU immediately when it started building, and attaching an strace to the lisp process gave me this pile of joy: --- SIGSEGV (Segmentation fault) @ 0 (0) --- sigreturn() = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- sigreturn() = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- sigreturn() = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- sigreturn() = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- The same build on F8 (both in mock on the same machine) resulted in pretty much similar results, except this was repeated in build.log *** Sigsegv in page not marked as write protected I'd venture this may have something to with the same problems sbcl has (see bug #434401). gcc43/kernel bug? sbcl devs pointed me to: "GCC 4.3.0 exposes a kernel bug" http://lwn.net/Articles/272048/ (I'm not currently an lwn subscriber, so can't access the article, yet). I just forwarded you a link to the LWN article (but I haven't read it yet). I also saw the problem on F8, which doesn't use gcc43, so I'm not inclined to believe that it's a gcc43 issue. But hey, you never know! As I understand it: unpatched kernel + gcc43-derived buildroot = boom joy, also would appear f8+ builds are broken due to kernel-header changes in
/usr/include/asm/unistd.h which now contains:
#ifdef __i386__
#include "unistd_32.h"
#else
#include "unistd_64.h"
#endif
which induces:
gcc -m32 -rdynamic -Wstrict-prototypes -Wall -O2 -g -DGENCGC -DLINKAGE_TABLE -O2
-g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector
--param=ssp-buffer-size=4 -m32 -march=i386 -mtune=generic
-fasynchronous-unwind-tables -D__NO_CTYPE -D_GNU_SOURCE -I. -I../../src/lisp -I-
-I/usr/X11R6/include -DGENCGC -DLINKAGE_TABLE -c -o lisp.o ../../src/lisp/lisp.c
...
In file included from /usr/include/sys/syscall.h:25,
from ../../src/lisp/Linux-os.h:25,
from ../../src/lisp/os.h:41,
from ../../src/lisp/lisp.c:25:
/usr/include/asm/unistd.h:2:25: error: unistd_32.h: No such file or directory
doh, looks like -I- is more to blame here. Tried out cmucl-19e-pre1, and this works better and gets further, but now fails in some asm in ../../src/lisp/x86-assem.S ../../src/lisp/x86-assem.S: Assembler messages: ../../src/lisp/x86-assem.S:84: Error: suffix or operands invalid for `fnstsw' See: http://koji.fedoraproject.org/koji/buildinfo?buildID=42817 (fwiw, cmucl-19e-pre1 builds fine on/for f8) Changing version to '9' as part of upcoming Fedora 9 GA. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping builds as of 2008-07-03. |