On Fri, Jul 26, 2019 at 01:27:32PM +0200, Anders Roxell wrote:
> When fall-through warnings was enabled by default, d93512ef0f0e
> ("Makefile: Globally enable fall-through warning"), we could see the
> following warnings was starting to show up. However, this was originally
> introduced in commit 6ee33c2712fc ("ARM: hw_breakpoint: correct and
> simplify alignment fixup code"). Commit d968d2b801d8 ("ARM: 7497/1:
> hw_breakpoint: allow single-byte watchpoints on all addresses") was
> written with the intent to allow single-byte watchpoints on all
> addresses but forgot to move 'case 1:' down below 'case 2:'.
>
> ../arch/arm/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
> ../arch/arm/kernel/hw_breakpoint.c:609:7: warning: this statement may fall
> through [-Wimplicit-fallthrough=]
> if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
> ^
> ../arch/arm/kernel/hw_breakpoint.c:611:3: note: here
> case 3:
> ^~~~
> ../arch/arm/kernel/hw_breakpoint.c:613:7: warning: this statement may fall
> through [-Wimplicit-fallthrough=]
> if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
> ^
> ../arch/arm/kernel/hw_breakpoint.c:615:3: note: here
> default:
> ^~~~~~~
>
> Rework so 'case 1:' are next to 'case 3:' and also add '/* Fall through
> */' so that the compiler doesn't warn about fall-through.
>
> Cc: stable@xxxxxxxxxxxxxxx # v3.16
> Fixes: 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")
> Signed-off-by: Anders Roxell <anders.roxell@xxxxxxxxxx>
> ---
> arch/arm/kernel/hw_breakpoint.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm/kernel/hw_breakpoint.c b/arch/arm/kernel/hw_breakpoint.c
> index af8b8e15f589..c14d506969ba 100644
> --- a/arch/arm/kernel/hw_breakpoint.c
> +++ b/arch/arm/kernel/hw_breakpoint.c
> @@ -603,15 +603,17 @@ int hw_breakpoint_arch_parse(struct perf_event *bp,
> case 0:
> /* Aligned */
> break;
> - case 1:
> case 2:
> /* Allow halfword watchpoints and breakpoints. */
> if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
> break;
> + /* Fall through */
> + case 1:
Why are you moving the 'case 1:' here? AFAICT, your patch now rejects
byte-aligned watchpoints of length 2.
Will
