| Message ID | 20190330005900.17282-2-richard.henderson@linaro.org |
|---|---|
| State | Superseded |
| Headers | show |
| Series | target/arm: Implement ARMv8.5-BTI for linux-user | expand |
On Sat, 30 Mar 2019 at 00:59, Richard Henderson <richard.henderson@linaro.org> wrote: > > The value of btype for syscalls is CONSTRAINED UNPREDICTABLE, > so we need to make sure that the value is 0 before clone, > fork, or syscall return. > > The value of btype for signals is defined, but it does not make > sense for a SIGILL handler to enter with the btype set as for > the indirect branch that caused the SIGILL. > > Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> thanks -- PMM
diff --git a/linux-user/aarch64/cpu_loop.c b/linux-user/aarch64/cpu_loop.c index d75fd9d3e2..f5cce4769d 100644 --- a/linux-user/aarch64/cpu_loop.c +++ b/linux-user/aarch64/cpu_loop.c @@ -85,6 +85,13 @@ void cpu_loop(CPUARMState *env) switch (trapnr) { case EXCP_SWI: + /* + * The state of BTYPE on syscall entry is CONSTRAINED + * UNPREDICTABLE. The real kernel will need to tidy this up + * as well. Do this before syscalls so that the value is + * correct on return from syscall (especially clone & fork). + */ + env->btype = 0; ret = do_syscall(env, env->xregs[8], env->xregs[0], diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index f84a9cf28a..078873c4a5 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -506,10 +506,11 @@ static void target_setup_frame(int usig, struct target_sigaction *ka, + offsetof(struct target_rt_frame_record, tramp); } env->xregs[0] = usig; - env->xregs[31] = frame_addr; env->xregs[29] = frame_addr + fr_ofs; - env->pc = ka->_sa_handler; env->xregs[30] = return_addr; + env->xregs[31] = frame_addr; + env->pc = ka->_sa_handler; + env->btype = 0; if (info) { tswap_siginfo(&frame->info, info); env->xregs[1] = frame_addr + offsetof(struct target_rt_sigframe, info);
The value of btype for syscalls is CONSTRAINED UNPREDICTABLE, so we need to make sure that the value is 0 before clone, fork, or syscall return. The value of btype for signals is defined, but it does not make sense for a SIGILL handler to enter with the btype set as for the indirect branch that caused the SIGILL. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- v4: Do not clear btype in signal frame. --- linux-user/aarch64/cpu_loop.c | 7 +++++++ linux-user/aarch64/signal.c | 5 +++-- 2 files changed, 10 insertions(+), 2 deletions(-) -- 2.17.1