| Message ID | e70f7b9de7e601b9e4a6fedad8eaf64d304b1637.1571326276.git.robin.murphy@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | arm64: kprobes: Drop open-coded exception fixup | expand |
Hi Robin, On Thu, 17 Oct 2019 16:31:42 +0100 Robin Murphy <robin.murphy@arm.com> wrote: > The short-circuit call to fixup_exception() from kprobe_fault_handler() > poses a problem now that the former wants to consume the fault address > too, since the common kprobes API offers us no way to pass it through. > Fortunately, however, it works out to be unnecessary: Thank you for pointing it out! > > - uaccess instructions themselves are not probeable, so at most we > should only ever expect to take a fixable fault from the pre or post > handlers. Right. If it is not fixable, we should handle it as a kernel bug. (to avoid it we can use probe_kernel_read() in kprobe handler) > - the pre and post handler run with preemption disabled, thus for any > fault they may cause, an unhandled return from kprobe_page_fault() > will proceed directly to __do_kernel_fault() thanks to the > faulthandler_disabled() check. OK, this is reasonable. > - __do_kernel_fault() will immediately call fixup_exception() unless > we're in an EL1 instruction abort, and if we've somehow taken one of > those on what we think is the middle of a uaccess routine, then the > world is already very on fire. OK, this looks good to me. Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Thank you! > > Thus we can reasonably drop the call from kprobe_fault_handler() and > leave uaccess fixups to the regular flow. > > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > --- > arch/arm64/kernel/probes/kprobes.c | 7 ------- > 1 file changed, 7 deletions(-) > > diff --git a/arch/arm64/kernel/probes/kprobes.c b/arch/arm64/kernel/probes/kprobes.c > index c4452827419b..422fbd5c6c55 100644 > --- a/arch/arm64/kernel/probes/kprobes.c > +++ b/arch/arm64/kernel/probes/kprobes.c > @@ -334,13 +334,6 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, unsigned int fsr) > */ > if (cur->fault_handler && cur->fault_handler(cur, regs, fsr)) > return 1; > - > - /* > - * In case the user-specified fault handler returned > - * zero, try to fix up. > - */ > - if (fixup_exception(regs)) > - return 1; > } > return 0; > } > -- > 2.21.0.dirty >
diff --git a/arch/arm64/kernel/probes/kprobes.c b/arch/arm64/kernel/probes/kprobes.c index c4452827419b..422fbd5c6c55 100644 --- a/arch/arm64/kernel/probes/kprobes.c +++ b/arch/arm64/kernel/probes/kprobes.c @@ -334,13 +334,6 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, unsigned int fsr) */ if (cur->fault_handler && cur->fault_handler(cur, regs, fsr)) return 1; - - /* - * In case the user-specified fault handler returned - * zero, try to fix up. - */ - if (fixup_exception(regs)) - return 1; } return 0; }
The short-circuit call to fixup_exception() from kprobe_fault_handler() poses a problem now that the former wants to consume the fault address too, since the common kprobes API offers us no way to pass it through. Fortunately, however, it works out to be unnecessary: - uaccess instructions themselves are not probeable, so at most we should only ever expect to take a fixable fault from the pre or post handlers. - the pre and post handler run with preemption disabled, thus for any fault they may cause, an unhandled return from kprobe_page_fault() will proceed directly to __do_kernel_fault() thanks to the faulthandler_disabled() check. - __do_kernel_fault() will immediately call fixup_exception() unless we're in an EL1 instruction abort, and if we've somehow taken one of those on what we think is the middle of a uaccess routine, then the world is already very on fire. Thus we can reasonably drop the call from kprobe_fault_handler() and leave uaccess fixups to the regular flow. Signed-off-by: Robin Murphy <robin.murphy@arm.com> --- arch/arm64/kernel/probes/kprobes.c | 7 ------- 1 file changed, 7 deletions(-)