public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Dan Carpenter <error27@gmail.com>
To: Greg KH <gregkh@suse.de>
Cc: Rahul Ruikar <rahul.ruikar@gmail.com>,
	David Brownell <dbrownell@users.sourceforge.net>,
	nm127@freemail.hu, linux-usb@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH] usb: gadget: goku_udc: Fix error path
Date: Tue, 5 Oct 2010 18:55:34 +0200	[thread overview]
Message-ID: <20101005165534.GJ5692@bicker> (raw)
In-Reply-To: <20101005132005.GB3548@suse.de>

This is based on an initial patch by Rahul Ruikar.

The goku_remove() function can be called before device_register() so it
can call device_unregister() improperly.  Also if the call to 
device_register() fails we need to call put_device().

As I was changing the error handling in goku_probe(), I noticed that
the label was "done" but actually if the function succeeds we return
earlier.  I renamed the error path to "err" instead of "done."

Reported-by: Rahul Ruikar <rahul.ruikar@gmail.com>
Signed-off-by: Dan Carpenter <error27@gmail.com>

diff --git a/drivers/usb/gadget/goku_udc.c b/drivers/usb/gadget/goku_udc.c
index 1088d08..7232018 100644
--- a/drivers/usb/gadget/goku_udc.c
+++ b/drivers/usb/gadget/goku_udc.c
@@ -1744,7 +1744,8 @@ static void goku_remove(struct pci_dev *pdev)
 				pci_resource_len (pdev, 0));
 	if (dev->enabled)
 		pci_disable_device(pdev);
-	device_unregister(&dev->gadget.dev);
+	if (dev->registered)
+		device_unregister(&dev->gadget.dev);
 
 	pci_set_drvdata(pdev, NULL);
 	dev->regs = NULL;
@@ -1774,7 +1775,7 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (!pdev->irq) {
 		printk(KERN_ERR "Check PCI %s IRQ setup!\n", pci_name(pdev));
 		retval = -ENODEV;
-		goto done;
+		goto err;
 	}
 
 	/* alloc, and start init */
@@ -1782,7 +1783,7 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (dev == NULL){
 		pr_debug("enomem %s\n", pci_name(pdev));
 		retval = -ENOMEM;
-		goto done;
+		goto err;
 	}
 
 	spin_lock_init(&dev->lock);
@@ -1800,7 +1801,7 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	retval = pci_enable_device(pdev);
 	if (retval < 0) {
 		DBG(dev, "can't enable, %d\n", retval);
-		goto done;
+		goto err;
 	}
 	dev->enabled = 1;
 
@@ -1809,7 +1810,7 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (!request_mem_region(resource, len, driver_name)) {
 		DBG(dev, "controller already in use\n");
 		retval = -EBUSY;
-		goto done;
+		goto err;
 	}
 	dev->got_region = 1;
 
@@ -1817,7 +1818,7 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (base == NULL) {
 		DBG(dev, "can't map memory\n");
 		retval = -EFAULT;
-		goto done;
+		goto err;
 	}
 	dev->regs = (struct goku_udc_regs __iomem *) base;
 
@@ -1833,7 +1834,7 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 			driver_name, dev) != 0) {
 		DBG(dev, "request interrupt %d failed\n", pdev->irq);
 		retval = -EBUSY;
-		goto done;
+		goto err;
 	}
 	dev->got_irq = 1;
 	if (use_dma)
@@ -1844,13 +1845,16 @@ static int goku_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	create_proc_read_entry(proc_node_name, 0, NULL, udc_proc_read, dev);
 #endif
 
-	/* done */
 	the_controller = dev;
 	retval = device_register(&dev->gadget.dev);
-	if (retval == 0)
-		return 0;
+	if (retval) {
+		put_device(&dev->gadget.dev);
+		goto err;
+	}
+	dev->registered = 1;
+	return 0;
 
-done:
+err:
 	if (dev)
 		goku_remove (pdev);
 	return retval;

  parent reply	other threads:[~2010-10-05 16:56 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-10-03 19:59 [RESEND/PATCH] usb: gadget: goku_udc: Fix error path Rahul Ruikar
2010-10-04 12:22 ` Dan Carpenter
2010-10-04 13:24   ` Rahul Ruikar
2010-10-04 14:10     ` Dan Carpenter
2010-10-05 13:20   ` Greg KH
2010-10-05 16:53     ` Dan Carpenter
2010-10-05 16:55     ` Dan Carpenter [this message]
  -- strict thread matches above, loose matches on Subject: below --
2010-10-03 14:10 [PATCH] " Rahul Ruikar
2010-10-03 16:17 ` Dan Carpenter
2010-10-04 10:46 ` Sergei Shtylyov
2010-10-04 10:50   ` Rahul Ruikar

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=20101005165534.GJ5692@bicker \
    --to=error27@gmail.com \
    --cc=dbrownell@users.sourceforge.net \
    --cc=gregkh@suse.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=nm127@freemail.hu \
    --cc=rahul.ruikar@gmail.com \
    /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