[PATCH] Make GDBJITRegistrar thread-safe
jim.kearney at oracle.com
Thu Mar 13 11:03:07 PDT 2014
BTW, the second link there is a different change, only related insofar
as it also breaks threaded MCJIT in some cases.
On 03/13/2014 01:49 PM, Ivan Puzyrevskiy wrote:
> Jim pointed out to me that patch was approved some time ago but it
> didn’t make its way to the trunk. See:
> Andy, if this patch would be approved can I ask you to commit it?
> Neither me nor Jim has commit rights.
> On 13 марта 2014 г., at 21:45, Kaylor, Andrew <andrew.kaylor at intel.com
> <mailto:andrew.kaylor at intel.com>> wrote:
>> That looks good as an intermediate fix.
>> In the long term I think the GDB registration should probably be
>> moved out of RuntimeDyldELF (maybe to an event listener), but these
>> changes would probably be nice to have even then.
>> *From:*llvm-commits-bounces at cs.uiuc.edu
>> <mailto:llvm-commits-bounces at cs.uiuc.edu>[mailto:llvm-commits-bounces at cs.uiuc.edu]*On
>> Behalf Of*Ivan Puzyrevskiy
>> *Sent:*Thursday, March 13, 2014 4:45 AM
>> *To:*llvm-commits at cs.uiuc.edu <mailto:llvm-commits at cs.uiuc.edu>
>> *Cc:*jim.kearney at oracle.com <mailto:jim.kearney at oracle.com>
>> *Subject:*[PATCH] Make GDBJITRegistrar thread-safe
>> Hi there.
>> While investigating a crash in GDBJITRegistrar I have stumbled
>> uponhttp://llvm.org/bugs/show_bug.cgi?id=15750. In my setting there
>> were multiple instances of MCJIT operating in different threads.
>> Patch from Bugzilla applies cleanly and solves the issue. I have
>> attached a patch against r203780, yet full credit goes to Jim Kearney
>> (jim.kearney at oracle.com <mailto:jim.kearney at oracle.com>), original
>> author of the patch.
>> Any chances to see this in the trunk or 3.5 release?
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu <mailto:llvm-commits at cs.uiuc.edu>
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-commits