This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/51212] ICE: verify_flow_info failed: BB 3 can not throw but has an EH edge with -fgnu-tm -fnon-call-exceptions and transaction_callable
- From: "aldyh at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Wed, 21 Dec 2011 16:28:12 +0000
- Subject: [Bug middle-end/51212] ICE: verify_flow_info failed: BB 3 can not throw but has an EH edge with -fgnu-tm -fnon-call-exceptions and transaction_callable
- Auto-submitted: auto-generated
- References: <bug-51212-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51212
Aldy Hernandez <aldyh at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot |aldyh at gcc dot gnu.org
|gnu.org |
--- Comment #2 from Aldy Hernandez <aldyh at gcc dot gnu.org> 2011-12-21 16:28:12 UTC ---
The problem here is that with -fnon-call-exceptions, the load from *p may trap,
but when we instrument the store, we have lost the landing pad information.
We can save and attach the exception information to the instrumented load
(_ITM_RU*), but currently all the TM load/store builtins are marked as
no-throw. Perhaps we can get rid of the no-throw bit when
-fnon-call-exceptions is enabled.
Maybe something like this:
1. Get rid of the NOTHROW bits from the TM load/store builtins.
2. Attach the EH information to the TM load/store calls.
3. Split the BB accordingly when the TM load/store calls are not the last in
the BB (because we have added a cast after the instrumentation).
How does this sound Richard?