Re: [PATCH v3] usb: musb: remove unused variable 'devctl'

[Sergei Shtylyov <sergei.shtylyov@gmail.com>]

On 11/24/20 12:13 PM, Sergei Shtylyov wrote:

[...]
>> From: Min Guo <min.guo@mediatek.com>
>>
>> Remove unused 'devctl' variable to fix compile warnings:
>>
>>      drivers/usb/musb/musbhsdma.c: In function 'dma_controller_irq':
>>      drivers/usb/musb/musbhsdma.c:324:8: warning: variable 'devctl' set
>>      but not used [-Wunused-but-set-variable]
>>
>> Signed-off-by: Min Guo <min.guo@mediatek.com>
>> ---
>> changes in v3
>> suggested by Greg Kroah-Hartman:
>> Add a comment.
>>
>> changes in v2
>> suggested by Alan Stern:
>> Add void before musb_read to indicate that the register MUSB_DEVCTL
>> was intended to be read and discarded.
>> ---
>>   drivers/usb/musb/musbhsdma.c | 8 +++++---
>>   1 file changed, 5 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/usb/musb/musbhsdma.c b/drivers/usb/musb/musbhsdma.c
>> index 0aacfc8be5a1..2a345b4ad015 100644
>> --- a/drivers/usb/musb/musbhsdma.c
>> +++ b/drivers/usb/musb/musbhsdma.c
> [...]
>> @@ -336,7 +334,11 @@ irqreturn_t dma_controller_irq(int irq, void *private_data)
>>                           < musb_channel->len) ?
>>                       "=> reconfig 0" : "=> complete");
>>   -                devctl = musb_readb(mbase, MUSB_DEVCTL);
>> +                /*
>> +                 * Some hardware may need to read the
>> +                 * MUSB_DEVCTL register once to take effect.
>> +                 */
>> +                (void)musb_readb(mbase, MUSB_DEVCTL);
> 
>    Hm, forcibly reading DevCtl in the DMA driver... sounds quite nonsensically. Lemme take a look...

   Indeed, prior to commit c418fd6c01fbc5516a2cd1eaf1df1ec86869028a ("usb: gadget: musb: fix short
isoc packets with inventra dma") the DevCtl read was done in order to check the DevCtl.HostMode bit --
this test was removed but the DevCtl read was left intact... so my vote goes for deleting both the
read and the variable...

> [...]

MBR, Sergei