Re: [PATCH 3/6] mips: call find_vma with the mmap_sem held

[Andreas Herrmann <herrmann.der.user@xxxxxxxxxxxxxx>]

On Sat, Apr 19, 2014 at 07:26:28PM -0700, Davidlohr Bueso wrote:
> Performing vma lookups without taking the mm->mmap_sem is asking
> for trouble. While doing the search, the vma in question can be
> modified or even removed before returning to the caller. Take the
> lock (exclusively) in order to avoid races while iterating through
> the vmacache and/or rbtree.
> 
> Updates two functions:
>   - process_fpemu_return()
>   - cteon_flush_cache_sigtramp()
> 
> This patch is completely *untested*.
> 
> Signed-off-by: Davidlohr Bueso <davidlohr@xxxxxx>
> Cc: Ralf Baechle <ralf@xxxxxxxxxxxxxx>
> Cc: linux-mips@xxxxxxxxxxxxxx

Tested-by: Andreas Herrmann <andreas.herrmann@xxxxxxxxxxxxxxxxxx>


Thanks,

Andreas

> ---
>  arch/mips/kernel/traps.c | 2 ++
>  arch/mips/mm/c-octeon.c  | 2 ++
>  2 files changed, 4 insertions(+)
> 
> diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
> index 074e857..c51bd20 100644
> --- a/arch/mips/kernel/traps.c
> +++ b/arch/mips/kernel/traps.c
> @@ -712,10 +712,12 @@ int process_fpemu_return(int sig, void __user *fault_addr)
>  		si.si_addr = fault_addr;
>  		si.si_signo = sig;
>  		if (sig == SIGSEGV) {
> +			down_read(&current->mm->mmap_sem);
>  			if (find_vma(current->mm, (unsigned long)fault_addr))
>  				si.si_code = SEGV_ACCERR;
>  			else
>  				si.si_code = SEGV_MAPERR;
> +			up_read(&current->mm->mmap_sem);
>  		} else {
>  			si.si_code = BUS_ADRERR;
>  		}
> diff --git a/arch/mips/mm/c-octeon.c b/arch/mips/mm/c-octeon.c
> index f41a5c5..05b1d7c 100644
> --- a/arch/mips/mm/c-octeon.c
> +++ b/arch/mips/mm/c-octeon.c
> @@ -137,8 +137,10 @@ static void octeon_flush_cache_sigtramp(unsigned long addr)
>  {
>  	struct vm_area_struct *vma;
>  
> +	down_read(&current->mm->mmap_sem);
>  	vma = find_vma(current->mm, addr);
>  	octeon_flush_icache_all_cores(vma);
> +	up_read(&current->mm->mmap_sem);
>  }
>  
>  
> -- 
> 1.8.1.4
> 
>