On Thu, Apr 19, 2018 at 5:57 PM, Burakov, Anatoly <anatoly.bura...@intel.com > wrote:
> On 19-Apr-18 3:48 PM, Arnon Warshavsky wrote: > >> Copy on the commit message and volatile. >> >> Regarding the new function defunct_and_remain_in_endless_loop () >> I don't think I can put that in a separate patch without breaking the >> current patch independence. >> > > How so? > > Just leave some panic instances in there for thread-related stuff and fix > them up in the next patch. > > Also, i'm not sure sending threads into an infinite loop on panic is such > a good idea. You might want to look at Olivier's approach [1] to creating > threads, using pthread_barriers and pthread_kill/cancel. > > This does warrant a separate patch now :) Thanks Anatoly Going into the infinite loop looked like the lesser evil at the time , but I guess I was too eager to get rid of as many panics on this path as I could. Given the bw I will need to properly handle it, and the fact that there are still some more panic calls in the code I will withdraw this specific removal inside the thread and handle it for the next build.
