Re: [PATCH v5 2/3] x86, mpx: hook #BR exception handler to allocate bound tables

[Dave Hansen <dave.hansen@intel.com>]

On 02/23/2014 05:27 AM, Qiaowei Ren wrote:
> +static bool allocate_bt(unsigned long bd_entry)
> +{
> +	unsigned long bt_size = 1UL << (MPX_L2_BITS+MPX_L2_SHIFT);
> +	unsigned long bt_addr, old_val = 0;
> +
> +	bt_addr = sys_mmap_pgoff(0, bt_size, PROT_READ | PROT_WRITE,
> +			MAP_ANONYMOUS | MAP_PRIVATE | MAP_POPULATE, -1, 0);
> +	if (bt_addr == -1) {
> +		pr_err("L2 Node Allocation Failed at L1 addr %lx\n",
> +				bd_entry);
> +		return false;
> +	}
> +	bt_addr = (bt_addr & MPX_L2_NODE_ADDR_MASK) | 0x01;
> +
> +	user_atomic_cmpxchg_inatomic(&old_val,
> +			(long __user *)bd_entry, 0, bt_addr);
> +	if (old_val)
> +		vm_munmap(bt_addr & MPX_L2_NODE_ADDR_MASK, bt_size);
> +
> +	return true;
> +}
> +
> +bool do_mpx_bt_fault(struct xsave_struct *xsave_buf)
> +{
> +	unsigned long status;
> +	unsigned long bd_entry, bd_base;
> +	unsigned long bd_size = 1UL << (MPX_L1_BITS+MPX_L1_SHIFT);
> +
> +	bd_base = xsave_buf->bndcsr.cfg_reg_u & MPX_BNDCFG_ADDR_MASK;
> +	status = xsave_buf->bndcsr.status_reg;
> +
> +	bd_entry = status & MPX_BNDSTA_ADDR_MASK;
> +	if ((bd_entry >= bd_base) && (bd_entry < bd_base + bd_size))
> +		return allocate_bt(bd_entry);
> +
> +	return false;
> +}

Can you talk a little bit about what the design is here?  Why does the
kernel have to do the allocation of the virtual address space?  Does it
really need to MAP_POPULATE?  bt_size looks like 4MB, and that's an
awful lot of memory to eat up at once.  Shouldn't we just let the kernel
demand-fault this like everything else?