| Message ID | 1445051470.3014.1.camel@ingics.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 28b5df1838b357c9e3e8eba02f684df3c0db05b3 |
| Headers | show |
On Sat, Oct 17, 2015 at 11:11:10AM +0800, Axel Lin wrote: > Signed-off-by: Axel Lin <axel.lin@ingics.com> > --- > I think the intention of wm8904_readable_register is to return false for > undocumented registers, but current code returns true for all cases in > wm8904_readable_register. > Please review if this patch is correct or not. > Thanks. Ok so I have been through every register access in the driver (yes that was a bit boring) and it looks like the driver only accesses a register that isn't marked as readable once. But this looks to be a bug. WM8904_EQ_REGS is set to 25, but there are only 24 registers in the EQ. This appears to be a mistake caused by the fact the registers start numbering from 1 rather than 0. I will send a patch to fix this small bug and this patch looks fine to me: Reviewed-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com> Unfortunately I don't have hardware to test the patch, but I have reviewed it pretty carefully and am happy. Thanks, Charles
diff --git a/sound/soc/codecs/wm8904.c b/sound/soc/codecs/wm8904.c index b783743..f48b9cb 100644 --- a/sound/soc/codecs/wm8904.c +++ b/sound/soc/codecs/wm8904.c @@ -312,7 +312,7 @@ static bool wm8904_readable_register(struct device *dev, unsigned int reg) case WM8904_FLL_NCO_TEST_1: return true; default: - return true; + return false; } }
Signed-off-by: Axel Lin <axel.lin@ingics.com> --- I think the intention of wm8904_readable_register is to return false for undocumented registers, but current code returns true for all cases in wm8904_readable_register. Please review if this patch is correct or not. Thanks. sound/soc/codecs/wm8904.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)