http://bugs.winehq.org/show_bug.cgi?id=22387
Summary: Wine fails to work when compiled with GCC 4.5.0 and certain GCC flags Product: Wine Version: 1.1.43 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: critical Priority: P2 Component: comctl32 AssignedTo: wine-bugs@winehq.org ReportedBy: t.artem@mailcity.com
Created an attachment (id=27400) --> (http://bugs.winehq.org/attachment.cgi?id=27400) Backtrace
I have just compiled Wine with the following CFLAGS:
-march=native -O2 -pipe -ftree-vectorize (which on my PC translate to -march=atom -mcx16 -msahf -maes -mpclmul -mpopcnt -msse4.2 --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=4096 -mtune=core2)
using recently released GCC 4.5.0.
Now all non-native applications fail to start.
http://bugs.winehq.org/show_bug.cgi?id=22387
Marcus Meissner marcus@jet.franken.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |marcus@jet.franken.de
--- Comment #1 from Marcus Meissner marcus@jet.franken.de 2010-04-17 06:35:42 --- yet another alignment issue.
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #2 from Artem S. Tashkinov t.artem@mailcity.com 2010-04-17 06:47:44 --- Created an attachment (id=27401) --> (http://bugs.winehq.org/attachment.cgi?id=27401) disas CreateMappedBitmap; disas; disas; disas
http://bugs.winehq.org/show_bug.cgi?id=22387
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |DUPLICATE
--- Comment #3 from Alexandre Julliard julliard@winehq.org 2010-04-17 08:52:36 --- gcc bug.
*** This bug has been marked as a duplicate of bug 22316 ***
http://bugs.winehq.org/show_bug.cgi?id=22387
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #4 from Vitaliy Margolen vitaliy@kievinfo.com 2010-04-17 10:35:24 --- Closing dup.
http://bugs.winehq.org/show_bug.cgi?id=22387
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|comctl32 |-unknown Severity|critical |normal
--- Comment #5 from Vitaliy Margolen vitaliy@kievinfo.com 2010-04-17 10:35:47 --- http://bugs.winehq.org/page.cgi?id=fields.html#bug_severity
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #6 from Dmitry Timoshkov dmitry@codeweavers.com 2010-04-19 22:58:39 --- In future please find a better place to file this kind of bugs. If you believe that if someone miscompiles Wine code that's a Wine bug, then something is very wrong with your logic.
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #7 from Artem S. Tashkinov t.artem@mailcity.com 2010-04-20 02:36:36 --- Dmitry, I vividly remember bug 2488 which was indeed a wine's problem, rather than GCC's one, so I posted this bug thinking that there's a subtle bug in Wine's code.
If you tell me to refrain from posting GCC related bugs I'll do so.
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #8 from Dmitry Timoshkov dmitry@codeweavers.com 2010-04-20 03:11:25 --- (In reply to comment #7)
Dmitry, I vividly remember bug 2488 which was indeed a wine's problem, rather than GCC's one, so I posted this bug thinking that there's a subtle bug in Wine's code.
Of course it's very nice that you remember a bug you reported 6 years ago.
However it's pretty strange that you have filed bug 22377 and bug 22385 which have been resolved as invalid with direct link to the gcc bug, and then you report this bug 22387. Marcus in the Comment #1 clearly said once again what's the source of the problem, and even after that your memory can't recognize some "vaguely similar" symptom of the problem?
If you tell me to refrain from posting GCC related bugs I'll do so.
Just a suggestion though, if you prefer for some reason to use your own custom compilation flags (although it's been stated many times that it's not recommended) with a bleeding edge gcc, then you could at least take some efforts to investigate the problem on your own, not dump it on somebody else's shoulders right away.
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #9 from Artem S. Tashkinov t.artem@mailcity.com 2010-04-20 03:26:52 --- (In reply to comment #8)
(In reply to comment #7)
Point taken - I'm going back to GCC 4.2.4 which has never caused any problems for me.
BTW, I'm not using any extreme GCC flags, for the last few years I compile all sources either with:
C Compiler flags: -march=pentium2 -O2 -pipe -ftree-vectorize C++ Compiler flags: -march=pentium2 -O2 -pipe -ftree-vectorize Linker flags: -Wl,-O1 -Wl,--hash-style=gnu -Wl,--enable-new-dtags
or with:
C Compiler flags: -march=native -O2 -pipe -ftree-vectorize C++ Compiler flags: -march=native -O2 -pipe -ftree-vectorize Linker flags: -Wl,-O1 -Wl,--hash-style=gnu -Wl,--enable-new-dtags
Anyway, it looks like GCC is kinda degrading because these innocuous flags cause some much stir, however Linux kernel works nicely having compiled with the aforementioned flags (albeit without -ftree-vectorize).
http://bugs.winehq.org/show_bug.cgi?id=22387
Dan Kegel dank@kegel.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dank@kegel.com
--- Comment #10 from Dan Kegel dank@kegel.com 2010-11-07 07:04:07 CST --- Artem, does gcc-4.5.1 behave any better for you? If not, does -mstackrealign help?
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #11 from Artem S. Tashkinov t.artem@mailcity.com 2010-11-07 07:27:54 CST --- Created an attachment (id=31777) --> (http://bugs.winehq.org/attachment.cgi?id=31777) crash log
No, all applications still crash and I'm using not yet releases GCC 4.5.2 from SVN ;)
I will now try to use -mstackrealign flag.
http://bugs.winehq.org/show_bug.cgi?id=22387
--- Comment #12 from Artem S. Tashkinov t.artem@mailcity.com 2010-11-07 07:42:31 CST --- OK, -mstackrealign solves this problem.
PS Meanwhile I just don't compile wine using -ftree-vectorize any longer.