From: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: industrypack-devel@lists.sourceforge.net,
linux-kernel@vger.kernel.org,
Jens Taprogge <jens.taprogge@taprogge.org>,
Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Subject: [PATCH 3/3] ipack: split ipack_device_register() in several functions
Date: Fri, 8 Mar 2013 09:21:47 +0100 [thread overview]
Message-ID: <1362730907-10047-3-git-send-email-siglesias@igalia.com> (raw)
In-Reply-To: <1362730907-10047-1-git-send-email-siglesias@igalia.com>
One function is ipack_device_init(). If it fails, the caller should execute
ipack_put_device().
The second function is ipack_device_add that only adds the device. If
it fails, the caller should execute ipack_put_device().
Then the device is removed with refcount = 0, as device_register() kernel
documentation says.
ipack_device_del() is added to remove the device.
Signed-off-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
---
drivers/ipack/carriers/tpci200.c | 14 +++++++++++++-
drivers/ipack/ipack.c | 24 +++++++++++++----------
include/linux/ipack.h | 39 ++++++++++++++++++++++++++++----------
3 files changed, 56 insertions(+), 21 deletions(-)
diff --git a/drivers/ipack/carriers/tpci200.c b/drivers/ipack/carriers/tpci200.c
index 0246b1f..c276fde 100644
--- a/drivers/ipack/carriers/tpci200.c
+++ b/drivers/ipack/carriers/tpci200.c
@@ -480,6 +480,7 @@ static void tpci200_release_device(struct ipack_device *dev)
static int tpci200_create_device(struct tpci200_board *tpci200, int i)
{
+ int ret;
enum ipack_space space;
struct ipack_device *dev =
kzalloc(sizeof(struct ipack_device), GFP_KERNEL);
@@ -495,7 +496,18 @@ static int tpci200_create_device(struct tpci200_board *tpci200, int i)
+ tpci200_space_interval[space] * i;
dev->region[space].size = tpci200_space_size[space];
}
- return ipack_device_register(dev);
+
+ ret = ipack_device_init(dev);
+ if (ret < 0) {
+ ipack_put_device(dev);
+ return ret;
+ }
+
+ ret = ipack_device_add(dev);
+ if (ret < 0)
+ ipack_put_device(dev);
+
+ return ret;
}
static int tpci200_pci_probe(struct pci_dev *pdev,
diff --git a/drivers/ipack/ipack.c b/drivers/ipack/ipack.c
index bdac7f6..a26f371 100644
--- a/drivers/ipack/ipack.c
+++ b/drivers/ipack/ipack.c
@@ -228,7 +228,7 @@ static int ipack_unregister_bus_member(struct device *dev, void *data)
struct ipack_bus_device *bus = data;
if (idev->bus == bus)
- ipack_device_unregister(idev);
+ ipack_device_del(idev);
return 1;
}
@@ -420,7 +420,7 @@ out:
return ret;
}
-int ipack_device_register(struct ipack_device *dev)
+int ipack_device_init(struct ipack_device *dev)
{
int ret;
@@ -429,6 +429,7 @@ int ipack_device_register(struct ipack_device *dev)
dev->dev.parent = dev->bus->parent;
dev_set_name(&dev->dev,
"ipack-dev.%u.%u", dev->bus->bus_nr, dev->slot);
+ device_initialize(&dev->dev);
if (dev->bus->ops->set_clockrate(dev, 8))
dev_warn(&dev->dev, "failed to switch to 8 MHz operation for reading of device ID.\n");
@@ -448,19 +449,22 @@ int ipack_device_register(struct ipack_device *dev)
dev_err(&dev->dev, "failed to switch to 32 MHz operation.\n");
}
- ret = device_register(&dev->dev);
- if (ret < 0)
- kfree(dev->id);
+ return 0;
+}
+EXPORT_SYMBOL_GPL(ipack_device_init);
- return ret;
+int ipack_device_add(struct ipack_device *dev)
+{
+ return device_add(&dev->dev);
}
-EXPORT_SYMBOL_GPL(ipack_device_register);
+EXPORT_SYMBOL_GPL(ipack_device_add);
-void ipack_device_unregister(struct ipack_device *dev)
+void ipack_device_del(struct ipack_device *dev)
{
- device_unregister(&dev->dev);
+ device_del(&dev->dev);
+ ipack_put_device(dev);
}
-EXPORT_SYMBOL_GPL(ipack_device_unregister);
+EXPORT_SYMBOL_GPL(ipack_device_del);
void ipack_get_device(struct ipack_device *dev)
{
diff --git a/include/linux/ipack.h b/include/linux/ipack.h
index def91fd..1888e06 100644
--- a/include/linux/ipack.h
+++ b/include/linux/ipack.h
@@ -207,19 +207,38 @@ int ipack_driver_register(struct ipack_driver *edrv, struct module *owner,
void ipack_driver_unregister(struct ipack_driver *edrv);
/**
- * ipack_device_register -- register an IPack device with the kernel
- * @dev: the new device to register.
+ * ipack_device_init -- initialize an IPack device
+ * @dev: the new device to initialize.
*
- * Register a new IPack device ("module" in IndustryPack jargon). The call
- * is done by the carrier driver. The carrier should populate the fields
- * bus and slot as well as the region array of @dev prior to calling this
- * function. The rest of the fields will be allocated and populated
- * during registration.
+ * Initialize a new IPack device ("module" in IndustryPack jargon). The call
+ * is done by the carrier driver. The carrier should populate the fields
+ * bus and slot as well as the region array of @dev prior to calling this
+ * function. The rest of the fields will be allocated and populated
+ * during initalization.
*
- * Return zero on success or error code on failure.
+ * Return zero on success or error code on failure.
+ *
+ * NOTE: _Never_ directly free @dev after calling this function, even
+ * if it returned an error! Always use ipack_put_device() to give up the
+ * reference initialized in this function instead.
+ */
+int ipack_device_init(struct ipack_device *dev);
+
+/**
+ * ipack_device_add -- Add an IPack device
+ * @dev: the new device to add.
+ *
+ * Add a new IPack device. The call is done by the carrier driver
+ * after calling ipack_device_init().
+ *
+ * Return zero on success or error code on failure.
+ *
+ * NOTE: _Never_ directly free @dev after calling this function, even
+ * if it returned an error! Always use ipack_put_device() to give up the
+ * reference initialized in this function instead.
*/
-int ipack_device_register(struct ipack_device *dev);
-void ipack_device_unregister(struct ipack_device *dev);
+int ipack_device_add(struct ipack_device *dev);
+void ipack_device_del(struct ipack_device *dev);
void ipack_get_device(struct ipack_device *dev);
void ipack_put_device(struct ipack_device *dev);
--
1.7.10.4
next prev parent reply other threads:[~2013-03-08 8:22 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-02-26 9:03 [PATCH] ipack: add missing put_device() after device_register() failed Samuel Iglesias Gonsalvez
2013-02-26 22:28 ` Dmitry Torokhov
2013-02-27 9:00 ` Samuel Iglesias Gonsálvez
2013-03-08 8:21 ` [PATCH 1/3] ipack: avoid double free on device->id Samuel Iglesias Gonsalvez
2013-03-08 8:21 ` [PATCH 2/3] ipack: add ipack_get_device() ipack_put_device() Samuel Iglesias Gonsalvez
2013-03-08 8:21 ` Samuel Iglesias Gonsalvez [this message]
2013-03-08 17:47 ` [PATCH 1/3] ipack: avoid double free on device->id Greg Kroah-Hartman
2013-03-08 18:11 ` Samuel Iglesias Gonsálvez
2013-03-08 19:36 ` Greg Kroah-Hartman
2013-03-11 7:58 ` Samuel Iglesias Gonsálvez
2013-03-08 18:11 ` Samuel Iglesias Gonsálvez
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1362730907-10047-3-git-send-email-siglesias@igalia.com \
--to=siglesias@igalia.com \
--cc=gregkh@linuxfoundation.org \
--cc=industrypack-devel@lists.sourceforge.net \
--cc=jens.taprogge@taprogge.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®