Use of iret when returning from exec system call

Question

I noticed that at the end of the start_thread function, which is called after most of the work of exec is done, there is a call to force_iret :

    static void
    start_thread_common(struct pt_regs *regs, unsigned long new_ip,
        unsigned long new_sp,
        unsigned int _cs, unsigned int _ss, unsigned int _ds)
    {
         loadsegment(fs, 0);
         loadsegment(es, _ds);
         loadsegment(ds, _ds);
         load_gs_index(0);
         regs->ip       = new_ip;
         regs->sp       = new_sp;
         regs->cs       = _cs;
         regs->ss       = _ss;
         regs->flags        = X86_EFLAGS_IF;
         force_iret();
    }    

I presume that this is done to ensure that that sysexit is not used to return to user space. So why does iret have to be used when returning from exec ?

Answer 1

This function modifies registers that sysret / sysexit would not restore.

Here's arch/x86/include/asm/thread_info.h :

/*
 * Force syscall return via IRET by making it look as if there was
 * some work pending. IRET is our most capable (but slowest) syscall
 * return path, which is able to restore modified SS, CS and certain
 * EFLAGS values that other (fast) syscall return instructions
 * are not able to restore properly.
 */
#define force_iret() set_thread_flag(TIF_NOTIFY_RESUME)