Re: [PATCH v4][mmc/host]:Blackfin SD Host Controller Driver

[Andrew Morton <akpm@linux-foundation.org>]

On Wed, 25 Nov 2009 17:25:20 -0500
Mike Frysinger <vapier.adi@gmail.com> wrote:

> >> +static void sdh_start_cmd(struct sdh_host *host, struct mmc_command *cmd)
> >> +{
> >> + __ __ unsigned int sdh_cmd;
> >> + __ __ unsigned int stat_mask;
> >> +
> >> + __ __ dev_dbg(mmc_dev(host->mmc), "%s enter cmd: 0x%p\n", __func__, cmd);
> >> + __ __ WARN_ON(host->cmd != NULL);
> >> + __ __ host->cmd = cmd;
> >> +
> >> + __ __ sdh_cmd = 0;
> >> + __ __ stat_mask = 0;
> >> +
> >> + __ __ sdh_cmd |= cmd->opcode;
> >> +
> >> + __ __ if (cmd->flags & MMC_RSP_PRESENT) {
> >> + __ __ __ __ __ __ sdh_cmd |= CMD_RSP;
> >> + __ __ __ __ __ __ stat_mask |= CMD_RESP_END;
> >> + __ __ } else
> >> + __ __ __ __ __ __ stat_mask |= CMD_SENT;

(ftso gmail)

> > Arguably wrong from a coding-style POV and looks weird IMO. __Adds a bit
> > of risk that subsequent coders will think they're writing in python adn
> > will add bugs.
> 
> i dont really get what you're referring to here.  the code in question
> looks ifne to me, and i dont see anything "python-esque" about it

Code like

	if (expr) {
		line1;
		line2;
	} else
		line3;

looks odd and can cause people to later add bugs along the lines of

	if (expr) {
		line1;
		line2;
	} else
+		line3a;
		line3;

this is particularly the case if line3 is long, or is preceded by a
comment.  This has happened in the past.

Adding the missing braces reduces the risk that this will occur.