New: there exists a wrong return value of function meye_probe()
A message that is not in the archive (sent off-list or not yet received): [email protected]/
---------- Forwarded message ----------
From: <[email protected]>
Date: Mon, May 18, 2015 at 3:19 AM
Subject: [Bug 98541] New: there exists a wrong return value of function meye_probe()
To: [email protected]
https://bugzilla.kernel.org/show_bug.cgi?id=98541
Bug ID: 98541
Summary: there exists a wrong return value of function
meye_probe()
Product: Drivers
Version: 2.5
Kernel Version: 3.19.1
Hardware: All
OS: Linux
Tree: Mainline
Status: NEW
Severity: normal
Priority: P1
Component: PCI
Assignee: [email protected]
Reporter: [email protected]
Regression: NoIn function meye_probe() at drivers/media/pci/meye/meye.c:1592, the call to video_register_device() in line 1748 may return a negative error code, and thus function meye_probe() will return the value of variable ret. And, the function meye_probe() will return 0 at last when it runs well. However, when the call to video_register_device() in line 1748 return a negative error code, the value of ret is 0. So the function meye_probe() will return 0 to its caller functions when it runs error because of the failing call to video_register_device(), leading to a wrong return value of function meye_probe().
The related code snippets in meye_probe() is as following. meye_probe @@ drivers/media/pci/meye/meye.c:1592
1592static int meye_probe(struct pci_dev *pcidev, const struct pci_device_id *ent)
1593{
...1672 if ((ret = pci_enable_device(meye.mchip_dev))) {
1673 v4l2_err(v4l2_dev, "meye: pci_enable_device failed\n");
1674 goto outenabledev;
1675 }
...
1748 if (video_register_device(meye.vdev, VFL_TYPE_GRABBER,
1749 video_nr) < 0) {
1750 v4l2_err(v4l2_dev, "video_register_device failed\n");
1751 goto outvideoreg;
1752 }
...
1761outvideoreg:
...1781outnotdev:
1782 return ret;
1783}
Generally, when the call to video_register_device() fails, the return value of caller functions should be different from another return value set when the call to video_register_device() succeeds, like the following codes in another file.cx8800_initdev @@ drivers/media/pci/cx88/cx88-video.c:1300 1300static int cx8800_initdev(struct pci_dev *pci_dev,
1301 const struct pci_device_id *pci_id)
1302{
...1493 err = video_register_device(dev->video_dev,VFL_TYPE_GRABBER,
1494 video_nr[core->nr]);
1495 if (err < 0) {
1496 printk(KERN_ERR "%s/0: can't register video device\n",
1497 core->name);
1498 goto fail_unreg;
1499 }
... 1545fail_unreg: 1546 cx8800_unregister_video(dev); 1547 free_irq(pci_dev->irq, dev); 1548 mutex_unlock(&core->lock); 1549fail_core: 1550 vb2_dma_sg_cleanup_ctx(dev->alloc_ctx); 1551 core->v4ldev = NULL; 1552 cx88_core_put(core,dev->pci);
1553fail_free:
1554 kfree(dev);
1555 return err;
1556}
Thank you
RUC_Soft_Sec
--
You are receiving this mail because:
You are watching the assignee of the bug.