> Perhaps I'm misunderstanding what you mean, but isn't this what is dealt > with in ia64_poke: if you try and update a process's user RBS via > ptrace, it writes to the kernel RBS instead? How do you manage to make > the user RBS newer - are you somehow updating it via shared memory? (In > which case I would say don't do that ;)) This is a step on the path to getting rid of the ia64_poke kludges in ptrace. It has been under discussion for a long time (years now). When this is in place, ia64's special code for PTRACE_PEEKDATA et al will go away and enable other desired cleanups in that code. This also fixes the longstanding problem that debuggers cannot reliably use /proc/pid/mem to read all memory on ia64 (as they do on all other machines). Because they do not get true data for user RBS regions, debuggers now special-case ia64 and stick to using ptrace for memory access, crippling the performance of some debugger uses (especially gdb's gcore) compared to Linux on other machines. > IMHO copying the RBS back and forth seems like an awfully expensive > workaround for something that would better be handled at the time that > you actually need to write to the RBS. Been there, done that, that's how it is now and those who did it in the first place agreed it was a bad can of worms and this is the new plan. Thanks, Roland - To unsubscribe from this list: send the line "unsubscribe linux-ia64" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
|