----- On Jul 14, 2020, at 5:30 PM, carlos carlos@xxxxxxxxxx wrote: > On 7/14/20 9:19 AM, Mathieu Desnoyers wrote: >> Is there an arch-agnostic way to get the thread pointer from user-space code ? >> That >> would be needed by all rseq critical section implementations. > > Yes, and no. We have void *__builtin_thread_pointer (void), but > few architectures implement the builtin so we'd have to go through > a round of compiler updates and backports. All targets know how to > access the thread pointer because the compiler has to generate > IE-mode accesses to the TLS variables. Practically speaking, I suspect this would mean postponing availability of rseq for widely deployed applications for a few more years ? I can very well see end users upgrading their kernel and using an early-adoption library to use rseq today, but requiring to upgrade the entire toolchain will likely postpone adoption to many years from now. It would be good to start getting feedback from rseq users so we can progress on the system call feature development. Unfortunately everything has been in a stand-still for the past years due to lack of rseq registration coordination in user-space. Thanks, Mathieu -- Mathieu Desnoyers EfficiOS Inc. http://www.efficios.com