From 6ded5f3848bfd3227ee208aa38f8bf8d7209d4e3 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 21 Apr 2014 18:28:02 +0200 Subject: [PATCH] uprobes/x86: Don't use arch_uprobe_abort_xol() in arch_uprobe_post_xol() 014940bad8e4 "uprobes/x86: Send SIGILL if arch_uprobe_post_xol() fails" changed arch_uprobe_post_xol() to use arch_uprobe_abort_xol() if ->post_xol fails. This was correct and helped to avoid the additional complications, we need to clear X86_EFLAGS_TF in this case. However, now that we have uprobe_xol_ops->abort() hook it would be better to avoid arch_uprobe_abort_xol() here. ->post_xol() should likely do what ->abort() does anyway, we should not do the same work twice. Currently only handle_riprel_post_xol() can be called twice, this is unnecessary but safe. Still this is not clean and can lead to the problems in future. Change arch_uprobe_post_xol() to clear X86_EFLAGS_TF and restore ->ip by hand and avoid arch_uprobe_abort_xol(). This temporary uglifies the usage of autask.saved_tf, we will cleanup this later. Signed-off-by: Oleg Nesterov Reviewed-by: Jim Keniston --- arch/x86/kernel/uprobes.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c index b3c2a92cce6c..2efb93f96030 100644 --- a/arch/x86/kernel/uprobes.c +++ b/arch/x86/kernel/uprobes.c @@ -759,22 +759,24 @@ int arch_uprobe_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs) struct uprobe_task *utask = current->utask; WARN_ON_ONCE(current->thread.trap_nr != UPROBE_TRAP_NR); + current->thread.trap_nr = utask->autask.saved_trap_nr; if (auprobe->ops->post_xol) { int err = auprobe->ops->post_xol(auprobe, regs); if (err) { - arch_uprobe_abort_xol(auprobe, regs); + if (!utask->autask.saved_tf) + regs->flags &= ~X86_EFLAGS_TF; /* - * Restart the probed insn. ->post_xol() must ensure - * this is really possible if it returns -ERESTART. + * Restore ->ip for restart or post mortem analysis. + * ->post_xol() must not return -ERESTART unless this + * is really possible. */ + regs->ip = utask->vaddr; if (err == -ERESTART) return 0; return err; } } - - current->thread.trap_nr = utask->autask.saved_trap_nr; /* * arch_uprobe_pre_xol() doesn't save the state of TIF_BLOCKSTEP * so we can get an extra SIGTRAP if we do not clear TF. We need @@ -819,9 +821,8 @@ int arch_uprobe_exception_notify(struct notifier_block *self, unsigned long val, /* * This function gets called when XOL instruction either gets trapped or - * the thread has a fatal signal, or if arch_uprobe_post_xol() failed. - * Reset the instruction pointer to its probed address for the potential - * restart or for post mortem analysis. + * the thread has a fatal signal. Reset the instruction pointer to its + * probed address for the potential restart or for post mortem analysis. */ void arch_uprobe_abort_xol(struct arch_uprobe *auprobe, struct pt_regs *regs) { -- 2.30.2