From: Rob Gardner <rob.gardner@xxxxxxxxxx> Date: Tue, 22 Dec 2015 21:16:06 -0700 > When an interrupt (such as a perf counter interrupt) is delivered > while executing in user space, the trap entry code puts ASI_AIUS in > %asi so that copy_from_user() and copy_to_user() will access the > correct memory. But if a perf counter interrupt is delivered while the > cpu is already executing in kernel space, then the trap entry code > will put ASI_P in %asi, and this will prevent copy_from_user() from > reading any useful stack data in either of the perf_callchain_user_X > functions, and thus no user callgraph data will be collected for this > sample period. An additional problem is that a fault is guaranteed > to occur, and though it will be silently covered up, it wastes time > and could perturb state. > > In perf_callchain_user(), we ensure that %asi contains ASI_AIUS > because we know for a fact that the subsequent calls to > copy_from_user() are intended to read the user's stack. > > Signed-off-by: Rob Gardner <rob.gardner@xxxxxxxxxx> > Signed-off-by: Dave Aldridge <david.j.aldridge@xxxxxxxxxx> I applied this, but modified it slightly. We have shorthand helpers "get_fs()" and "set_fs()" for doing this kind of work, and as you can see in arch/sparc/kernel/process_64.c and elsewhere this is the canonical way to adjust the %asi value in these kinds of circumstances. Also, set_fs() updates the thread's current_ds value properly. Notice that NGmemcpy.S uses the TI_CURRENT_DS value via the RESTORE_ASI() macro. So if that's not set properly, the %asi restore won't be done properly. -- To unsubscribe from this list: send the line "unsubscribe sparclinux" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
