Bert Wesarg wrote: > Hi all, > > I have a problem with this retry algorithm of > hwloc_linux_foreach_proc_tid(). For example with the > hwloc_linux_get_pid_cpubind() function. hwloc_linux_get_pid_cpubind() > should collect all affinity mask from the threads. But if the retry is > triggered and the new tid list does not have a tid anymore which had > an affinity mask which is not a subset of the collected one in the > first run. The end result is inaccurate. >
Thanks for the report, I applied a slightly different patch to SVN. Brice