Re: [PATCH] staging: rtl8188eu: remove dead code in do-while conditional step

[Connor Kuehl]

On 9/23/19 1:38 PM, Larry Finger wrote:
> On 9/23/19 2:48 PM, Connor Kuehl wrote:
> > The local variable 'bcmd_down' is always set to true almost immediately
> > before the do-while's condition is checked. As a result, !bcmd_down
> > evaluates to false which short circuits the logical AND operator meaning
> > that the second operand is never reached and is therefore dead code.
> >
> > Addresses-Coverity: ("Logically dead code")
> >
> > Signed-off-by: Connor Kuehl <connor.kuehl@xxxxxxxxxxxxx>
> > ---
> > Â drivers/staging/rtl8188eu/hal/rtl8188e_cmd.c | 3 +--
> > Â 1 file changed, 1 insertion(+), 2 deletions(-)
> >
> > diff --git a/drivers/staging/rtl8188eu/hal/rtl8188e_cmd.c 
> > b/drivers/staging/rtl8188eu/hal/rtl8188e_cmd.c
> > index 47352f210c0b..a4b317937b23 100644
> > --- a/drivers/staging/rtl8188eu/hal/rtl8188e_cmd.c
> > +++ b/drivers/staging/rtl8188eu/hal/rtl8188e_cmd.c
> > @@ -48,7 +48,6 @@ static u8 _is_fw_read_cmd_down(struct adapter 
> > *adapt, u8 msgbox_num)
> > Â static s32 FillH2CCmd_88E(struct adapter *adapt, u8 ElementID, u32 
> > CmdLen, u8 *pCmdBuffer)
> > Â {
> > ÂÂÂÂÂ u8 bcmd_down = false;
> > -ÂÂÂ s32 retry_cnts = 100;
> > ÂÂÂÂÂ u8 h2c_box_num;
> > ÂÂÂÂÂ u32 msgbox_addr;
> > ÂÂÂÂÂ u32 msgbox_ex_addr;
> > @@ -103,7 +102,7 @@ static s32 FillH2CCmd_88E(struct adapter *adapt, 
> > u8 ElementID, u32 CmdLen, u8 *p
> > ÂÂÂÂÂÂÂÂÂ adapt->HalData->LastHMEBoxNum =
> > ÂÂÂÂÂÂÂÂÂÂÂÂÂ (h2c_box_num+1) % RTL88E_MAX_H2C_BOX_NUMS;
> > -ÂÂÂ } while ((!bcmd_down) && (retry_cnts--));
> > +ÂÂÂ } while (!bcmd_down);
> > ÂÂÂÂÂ ret = _SUCCESS;
>
> This patch is correct; however, the do..while loop will always be 
> executed once, thus you could remove the loop and the loop variable 
> bcmd_down.

Ah, yes! That makes sense, good catch.

> @greg: If you would prefer a two-step process, then this one is OK.

I'll do whichever is preferred. I'm happy to NACK this and send a v2 
with the dead code and loop removed or I can send a separate patch based 
on this one to remove the loop.

Thank you,

Connor

> Larry