[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <612eedc5-0f28-7a38-abad-4b9573208875@wanadoo.fr>
Date: Mon, 19 Dec 2016 07:19:43 +0100
From: Marion & Christophe JAILLET <christophe.jaillet@...adoo.fr>
To: linus.walleij@...aro.org, baohua@...nel.org
Cc: linux-gpio@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
kernel-janitors@...r.kernel.org
Subject: [RFC] Question about freeing of resource in 'atlas7_pinmux_probe()',
in file 'drivers/pinctrl/sirf/pinctrl-atlas7.c'
Hi,
while playing with coccinelle, a missing 'of_node_put()' triggered in
'atlas7_pinmux_probe()', in file 'drivers/pinctrl/sirf/pinctrl-atlas7.c'.
/* The sd3 and sd9 shared all pins, and the function select by
* SYS2PCI_SDIO9SEL register
*/
sys2pci_np = of_find_node_by_name(NULL, "sys2pci");
if (!sys2pci_np)
return -EINVAL;
ret = of_address_to_resource(sys2pci_np, 0, &res);
if (ret) <------------- missing of_node_put(sys2pci_np);
return ret;
pmx->sys2pci_base = devm_ioremap_resource(&pdev->dev, &res);
if (IS_ERR(pmx->sys2pci_base)) {
of_node_put(sys2pci_np); <------------- added by commit
151b8c5ba1eb
return -ENOMEM;
}
Looking at the history of this file, I found a recent commit that added
another missing of_node_put (see above).
Adding missing 'of_node_put()' in error handling paths is fine, but in
this particular case, I was wondering if one was not also missing in the
normal path?
In such a case, I would revert 151b8c5ba1eb and propose something like:
/* The sd3 and sd9 shared all pins, and the function select by
* SYS2PCI_SDIO9SEL register
*/
sys2pci_np = of_find_node_by_name(NULL, "sys2pci");
if (!sys2pci_np)
return -EINVAL;
ret = of_address_to_resource(sys2pci_np, 0, &res);
if (ret) {
of_node_put(sys2pci_np);
return ret;
}
of_node_put(sys2pci_np);
pmx->sys2pci_base = devm_ioremap_resource(&pdev->dev, &res);
if (IS_ERR(pmx->sys2pci_base))
return -ENOMEM;
Thanks for your comment,
best regards,
CJ
Powered by blists - more mailing lists