On 6/27/2011 8:56 PM, Todd Poynor wrote:
On Mon, Jun 27, 2011 at 06:33:06PM +0200, Benoit Cousson wrote:
...
+ r = clk_get_sys(dev_name(&od->pdev.dev), clk_alias);
+ if (!IS_ERR(r)) {
+ pr_warning("omap_device: %s: %s already exist\n",
+ dev_name(&od->pdev.dev), clk_alias);
I believe a clk_put(r) is appropriate here.
Appropriate I don't know, but useless for sure :-)
This clk_put is a no-op for every ARM platforms (I found one exception).
I do not know why it was originally done like that, but that api is
clearly not a put/get kind of API.
I'm OK to add that, but I think it is a little bit misleading.
+ return;
+ }
+
+ r = omap_clk_get_by_name(clk_name);
+ if (IS_ERR(r)) {
+ pr_err("omap_device: %s: omap_clk_get_by_name for %s failed\n",
+ dev_name(&od->pdev.dev), clk_name);
+ return;
+ }
+
+ l = clkdev_alloc(r, clk_alias, dev_name(&od->pdev.dev));
+ if (!l) {
+ pr_err("omap_device: %s: clkdev_alloc for %s failed\n",
+ dev_name(&od->pdev.dev), clk_alias);
And here.
No, it is not needed in that case because the omap_clk_get_by_name is
not using the clk_get API.
+ return;
+ }
+
+ clkdev_add(l);
And here.
Not needed either, no clk_get used.
Benoit
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
