uprobes: longjmp / special-mapping fixes
Ingo, please pull from
git://git.kernel.org/pub/scm/linux/kernel/git/oleg/misc uprobes/core
Based on tip:perf/core
Currently ret-probes can't work (the application will likely crash) if the probed function does not return, and this is even documented in handle_trampoline(). This tries to make the first step to fix the problem, assuming that the probed functions use the same stack.
Also, xol_add_vma() doesn't use install_special_mapping() correctly, and we can name the xol vma which currently looks like anon mapping.
Oleg Nesterov (14):
uprobes: Introduce get_uprobe()
uprobes: Introduce free_ret_instance()
uprobes: Send SIGILL if handle_trampoline() fails
uprobes: Change prepare_uretprobe() to use uprobe_warn()
uprobes: Change handle_trampoline() to find the next chain beforehand
uprobes: Export struct return_instance, introduce arch_uretprobe_is_alive() uprobes/x86: Reimplement arch_uretprobe_is_alive()
uprobes: Change handle_trampoline() to flush the frames invalidated by longjmp()
uprobes: Change prepare_uretprobe() to (try to) flush the dead frames
uprobes: Add the "enum rp_check ctx" arg to arch_uretprobe_is_alive() uprobes/x86: Make arch_uretprobe_is_alive(RP_CHECK_CALL) more clever
uprobes: fix the usage of install_special_mapping()
uprobes: use vm_special_mapping to name the xol vma
uprobes: fix the waitqueue_active() check in xol_free_insn_slot()
arch/x86/kernel/uprobes.c | 9 ++ include/linux/uprobes.h | 17 ++++ kernel/events/uprobes.c | 228 ++++++++++++++++++++++++++------------------- 3 files changed, 156 insertions(+), 98 deletions(-)