Re: [Qemu-devel] [PULL 07/23] exec.c: Relax restrictions on watchpoint length and alignment

[Max Filippov]

Hi Peter,

On Fri, Sep 12, 2014 at 6:23 AM, Peter Maydell <address@hidden> wrote:
> The current implementation of watchpoints requires that they
> have a power of 2 length which is not greater than TARGET_PAGE_SIZE
> and that their address is a multiple of their length. Watchpoints
> on ARM don't fit these restrictions, so change the implementation
> so they can be relaxed.
>
> Signed-off-by: Peter Maydell <address@hidden>
> Reviewed-by: Richard Henderson <address@hidden>
> ---
>  exec.c            | 44 +++++++++++++++++++++++++++++++-------------
>  include/qom/cpu.h |  2 +-
>  linux-user/main.c |  3 +--
>  3 files changed, 33 insertions(+), 16 deletions(-)
>
> diff --git a/exec.c b/exec.c
> index 7dddcc8..f3e7fb6 100644
> --- a/exec.c
> +++ b/exec.c
> @@ -582,12 +582,10 @@ int cpu_watchpoint_insert(CPUState *cpu, vaddr addr, 
> vaddr len,
>  int cpu_watchpoint_insert(CPUState *cpu, vaddr addr, vaddr len,
>                            int flags, CPUWatchpoint **watchpoint)
>  {
> -    vaddr len_mask = ~(len - 1);
>      CPUWatchpoint *wp;
>
> -    /* sanity checks: allow power-of-2 lengths, deny unaligned watchpoints */
> -    if ((len & (len - 1)) || (addr & ~len_mask) ||
> -            len == 0 || len > TARGET_PAGE_SIZE) {
> +    /* forbid ranges which are empty or run off the end of the address space 
> */
> +    if (len == 0 || (addr + len - 1) <= addr) {

With this change it's no longer possible to set one-byte-long watchpoint,
xtensa testsuite fails because of that.
I guess you meant 'len == 0 || (addr + len - 1) < addr' ?

-- 
Thanks.
-- Max