From: "Eric W. Biederman" ebiederm@xmission.com
[ Upstream commit 8b9c6b28312cc51a87055e292b11c5aa28f0c935 ]
The ia64 handling of failure to setup a signal frame has been trying to set overlapping fields in struct siginfo since 2.3.43. The si_pid and si_uid fields are stomped when the si_addr field is set. The si_code of SI_KERNEL indicates that si_pid and si_uid should be valid, and that si_addr does not exist.
Being at odds with the definition of SI_KERNEL and with nothing to indicate that this was a signal frame setup failure there is no way for userspace to know that si_addr was filled out instead.
In practice failure to setup a signal frame is rare, and si_pid and si_uid are always set to 0 when si_code is SI_KERNEL so I expect no one has looked closely enough before to see this weirdness. Further the only difference between force_sigsegv_info and the generic force_sigsegv other than the return code is that force_sigsegv_info stomps the si_uid and si_pid fields.
Remove the bug and simplify the code by using force_sigsegv in this case just like other architectures.
Fixes: 2.3.43 Cc: Tony Luck tony.luck@intel.com Cc: Fenghua Yu fenghua.yu@intel.com Cc: linux-ia64@vger.kernel.org Acked-by: Tony Luck tony.luck@intel.com Signed-off-by: "Eric W. Biederman" ebiederm@xmission.com Signed-off-by: Sasha Levin sashal@kernel.org --- arch/ia64/kernel/signal.c | 50 ++++++++++----------------------------- 1 file changed, 12 insertions(+), 38 deletions(-)
diff --git a/arch/ia64/kernel/signal.c b/arch/ia64/kernel/signal.c index d1234a5ba4c5..01fc133b2e4c 100644 --- a/arch/ia64/kernel/signal.c +++ b/arch/ia64/kernel/signal.c @@ -231,37 +231,6 @@ rbs_on_sig_stack (unsigned long bsp) return (bsp - current->sas_ss_sp < current->sas_ss_size); }
-static long -force_sigsegv_info (int sig, void __user *addr) -{ - unsigned long flags; - struct siginfo si; - - clear_siginfo(&si); - if (sig == SIGSEGV) { - /* - * Acquiring siglock around the sa_handler-update is almost - * certainly overkill, but this isn't a - * performance-critical path and I'd rather play it safe - * here than having to debug a nasty race if and when - * something changes in kernel/signal.c that would make it - * no longer safe to modify sa_handler without holding the - * lock. - */ - spin_lock_irqsave(¤t->sighand->siglock, flags); - current->sighand->action[sig - 1].sa.sa_handler = SIG_DFL; - spin_unlock_irqrestore(¤t->sighand->siglock, flags); - } - si.si_signo = SIGSEGV; - si.si_errno = 0; - si.si_code = SI_KERNEL; - si.si_pid = task_pid_vnr(current); - si.si_uid = from_kuid_munged(current_user_ns(), current_uid()); - si.si_addr = addr; - force_sig_info(SIGSEGV, &si, current); - return 1; -} - static long setup_frame(struct ksignal *ksig, sigset_t *set, struct sigscratch *scr) { @@ -295,15 +264,18 @@ setup_frame(struct ksignal *ksig, sigset_t *set, struct sigscratch *scr) * instead so we will die with SIGSEGV. */ check_sp = (new_sp - sizeof(*frame)) & -STACK_ALIGN; - if (!likely(on_sig_stack(check_sp))) - return force_sigsegv_info(ksig->sig, (void __user *) - check_sp); + if (!likely(on_sig_stack(check_sp))) { + force_sigsegv(ksig->sig, current); + return 1; + } } } frame = (void __user *) ((new_sp - sizeof(*frame)) & -STACK_ALIGN);
- if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return force_sigsegv_info(ksig->sig, frame); + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) { + force_sigsegv(ksig->sig, current); + return 1; + }
err = __put_user(ksig->sig, &frame->arg0); err |= __put_user(&frame->info, &frame->arg1); @@ -317,8 +289,10 @@ setup_frame(struct ksignal *ksig, sigset_t *set, struct sigscratch *scr) err |= __save_altstack(&frame->sc.sc_stack, scr->pt.r12); err |= setup_sigcontext(&frame->sc, set, scr);
- if (unlikely(err)) - return force_sigsegv_info(ksig->sig, frame); + if (unlikely(err)) { + force_sigsegv(ksig->sig, current); + return 1; + }
scr->pt.r12 = (unsigned long) frame - 16; /* new stack pointer */ scr->pt.ar_fpsr = FPSR_DEFAULT; /* reset fpsr for signal handler */