Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
Horms [EMAIL PROTECTED] schrieb: Hi Erhard, your initial bug report states that after pluging in your cammera, vol_id is executed (presumably the result of a kernel-generated hotplug event) and it segfaults. After this happens your camera is unusable until after you reinsert it. Could you please describe what happens after you reinsert it. Hi, as long as usbmount is in place to execute vol_id, the I/O errors occur again. With until I plug it in again I referred to the fact the device isn't readable anymore after vol_id ran. The flash drive can be accessed normally when usbmount / vol_id is disabled, so I guess the problem is something that vol_id does but mount does not. I did not try to format the drive so far. Regards, Erhard -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Fri, Aug 12, 2005 at 10:50:41AM +0200, Erhard Schultchen wrote: Horms [EMAIL PROTECTED] schrieb: Hi Erhard, your initial bug report states that after pluging in your cammera, vol_id is executed (presumably the result of a kernel-generated hotplug event) and it segfaults. After this happens your camera is unusable until after you reinsert it. Could you please describe what happens after you reinsert it. Hi, as long as usbmount is in place to execute vol_id, the I/O errors occur again. With until I plug it in again I referred to the fact the device isn't readable anymore after vol_id ran. The flash drive can be accessed normally when usbmount / vol_id is disabled, so I guess the problem is something that vol_id does but mount does not. I did not try to format the drive so far. Thanks. Do you have any idea what exactly usbmount / vol_id is trying to mount. And if so, does trying to do that manually put the camera in a state where you can't mount anything until you reconnect? -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Tue, Aug 09, 2005 at 09:09:04PM +0200, Marco d'Itri wrote: close 322205 -1 retitle -1 can't deal with a broken FAT file system reassign -1 linux-2.6 thanks On Aug 09, Erhard Schultchen [EMAIL PROTECTED] wrote: I'm using usbmount in conjunction with udev to mount usb devices on plug in. However, my digital camera does not like /sbin/vol_id. vol_id segfaults and the kernel reports plenty i/o errors like these: OK, it should not segfault. Can you get a stack trace from the binary with debugging symbols? end_request: I/O error, dev sda, sector 250584 Buffer I/O error on device sda, logical block 250584 From this point on, the camera isn't readable any more no matter which sector (same error as above) until I plug it in again. My camera is a But this looks like the kernel reacting badly to a damaged file system, so I'm reassigning a copy of this bug to the kernel package. Hi Marco, I am unclear on what behaviour you expect from the kernel if the file system or media is damaged. Also, as an aside, the kernel appears to be reporting itself as 2.6.12.2, which is not a Debian kernel. -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Aug 11, Horms [EMAIL PROTECTED] wrote: I am unclear on what behaviour you expect from the kernel if the file system or media is damaged. Maybe deal more gracefully with this and not require replugging the camera? -- ciao, Marco signature.asc Description: Digital signature
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Aug 11, Horms [EMAIL PROTECTED] wrote: Is it the case that a) the flash does indeed contain a corupted file system, or b) vol_id's bogusness (or something else) is causing a mount request for something that is invalid, and from there the kernel' can't mount anything on the device, even things that are valid? vol_id just reads the superblock, so I think it's a. BTW, vol_id doesn't have a man page, is there some documentation for it somewhere else. No, but it's also a very simple program. -- ciao, Marco signature.asc Description: Digital signature
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Thu, Aug 11, 2005 at 12:45:22PM +0200, Marco d'Itri wrote: On Aug 11, Horms [EMAIL PROTECTED] wrote: Is it the case that a) the flash does indeed contain a corupted file system, or b) vol_id's bogusness (or something else) is causing a mount request for something that is invalid, and from there the kernel' can't mount anything on the device, even things that are valid? vol_id just reads the superblock, so I think it's a. If its a) and the file system is corrupted, then the kernel isn't going to be able to read it. I'm really not clear on what you expect to happen. BTW, vol_id doesn't have a man page, is there some documentation for it somewhere else. No, but it's also a very simple program. Ok, does it do anything other than read the superblock? -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Aug 11, Horms [EMAIL PROTECTED] wrote: If its a) and the file system is corrupted, then the kernel isn't going to be able to read it. I'm really not clear on what you expect to happen. I understand that usually it was readable. BTW, vol_id doesn't have a man page, is there some documentation for it somewhere else. No, but it's also a very simple program. Ok, does it do anything other than read the superblock? For a FAT file system, it will also try to read the root directory. -- ciao, Marco signature.asc Description: Digital signature
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Thu, Aug 11, 2005 at 01:21:07PM +0200, Marco d'Itri wrote: On Aug 11, Horms [EMAIL PROTECTED] wrote: If its a) and the file system is corrupted, then the kernel isn't going to be able to read it. I'm really not clear on what you expect to happen. I understand that usually it was readable. Hi Erhard, your initial bug report states that after pluging in your cammera, vol_id is executed (presumably the result of a kernel-generated hotplug event) and it segfaults. After this happens your camera is unusable until after you reinsert it. Could you please describe what happens after you reinsert it. BTW, vol_id doesn't have a man page, is there some documentation for it somewhere else. No, but it's also a very simple program. Ok, does it do anything other than read the superblock? For a FAT file system, it will also try to read the root directory. Hi Marco, Please provide some documenation for vol_id. It is a binary in /usr/sbin and thus should have a man page. But in lieu of that, could you please briefly describe what it is supposed to do. Does it just try and determine the valid partitions on a device, information wich can subsequently be passed through to mount? -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Aug 12, Horms [EMAIL PROTECTED] wrote: Please provide some documenation for vol_id. It is a binary in /usr/sbin and thus should have a man page. But in lieu of that, could you please briefly describe what it is supposed to do. Does it just try and determine the valid partitions on a device, information wich can subsequently be passed through to mount? It first tries to recognize the filesystem type, and then to read its volume label. -- ciao, Marco signature.asc Description: Digital signature
Bug#322220: Bug#322205: udev: vol_id crashes on usb storage device
On Fri, Aug 12, 2005 at 04:40:52AM +0200, Marco d'Itri wrote: On Aug 12, Horms [EMAIL PROTECTED] wrote: Please provide some documenation for vol_id. It is a binary in /usr/sbin and thus should have a man page. But in lieu of that, could you please briefly describe what it is supposed to do. Does it just try and determine the valid partitions on a device, information wich can subsequently be passed through to mount? It first tries to recognize the filesystem type, and then to read its volume label. Thanks -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]